Theory and Operation of The Fourdrinier Paper Machine

THEORY AND OPERATION
OF THE FOURDRINIER
PAPER MACHINE
THEORY AND OPERATION
OF THE FOURDRINIER
PAPER MACHINE
BY
G. H. NUTTALL, M.A. (Cantab.)
LONDON:
S. C. PHILLIPS & CO. LTD.
© G. H. Nuttall
PRINTED IN GREAT BRITAIN BY

ADLARD & SON LTD., BARTHOLOMEW PRESS, DORKING
CONTENTS
Foreword ix
PART 1: THE WET-END FLOW SYSTEM
11 Introduction 3
CHAPTER lA: THEORETICAL CONSIDERATIONS.
lA. 1 Flow of Fibre Suspensions . 6
lA.2 Influence of Wet-end Design on Operational Stability 15
CHAPTER lB: OPERATING FACTORS AFFECTING PERFORMANCE.
lB. 1 Retention on the Wire 31
lB.2 Fresh Stuff Flow and Mixing 35
lB. 3 Backwater and Whitewater Systems 38
lB.4 Breast Box Approach System 40
lB. 5 Breast Box Design 46
1B. 6 Slice Design 54
1B. 7 Aeration of the Stock 59
lB. 8 Hydraulic Disturbances 63
CHAPTER lC: RUNNING THE WET-END FLOW SYSTEM.
1 C. 1 Daily Operation 66
lC.2 Maintenance 76
1C.3 Control of Substance . 80
1C .4 Practical Points . 87
References 95
PART 2: SCREENS AND CLEANERS
21 Introduction 101
CHAPTER 2A: THEORETICAL CONSIDERATIONS.
2A. l Theory of Screening . 105
2A. 2 Theory of Cleaning . 117
2A. 3 Assessing Screen and Cleaner Efficiency 129
CHAPTER 2B: OPERATING FACTORS AFFECTING SCREENS AND CLEANERS.
2B. I Position in the Flow System 135
28. 2 Open Screening Equipment 140
2B . 3 Closed Pressure Screens 144
28. 4 Early Cleaning Equipment . 150
2B.5 Cylindrical Cleaners . 152
2B. 6 Cyclone Cleaners 154
CHAPTER 2C: RUNNING SCREENS AND CLEANERS.
2C. I Daily Operation 171
2C. 2 Maintenance 172
2C. 3 Practical Points 17 5
References 180
v
CONTENTS
PART 3: THE WIRE SECTION
31 Introduction 185
CHAPTER 3A: THEORETICAL CONSIDERATIONS.
3A. l Drainage in the Table Roll Section 187
3A. 2 Forming the Sheet on the Wire 195
3A. 3 Drainage at the Suction Boxes 209
3A .4 Conditions for Couching . 219
CHAPTER 3B: OPERATING FACTORS AFFECTING THE WIRE SECTION.
3B.1 Early Drainage Conditions. 226
3B. 2 Table Rolls and other Dewatering Devices 232
3B. 3 Shake 237
3B . 4 Suction Boxes 241
3B. 5 The Dandy 247
3B. 6 Couching . 252
3B. 7 The Wire and Showers 262
3B. 8 Breast Box Stock Consistency and Temperature 268
38. 9 Machine Speed, Substance and Stuff Treatment 269
3B.10 Equipment 271
CHAPTER 3C: RUNNING THE WIRE SECTION.
3C. 1 Daily Operation 275
3C. 3 Practical Points 286
References 296
PART 4: THE PRESS SECTION

41 Introduction 303
CHAPTER 4A: GENERAL THEORY OF PRESSING.
4A. 1 Early Theory 306
4A.2 Modern Theory of Pressing 310
CHAPTER 4B: OPERATING FACTORS AFFECTING PRESS PERFORMANCE.
4B. l Press Load and Nip Width 318
48.2 Felts and Felt Cleaning 320
4B. 3 Felt and Paper Runs . 325
4B .4 Suction Roll Conditions 332
4B. 5 Machine Speed and Draw Control 335
4B. 6 Moisture Content of Paper entering Press 337
4B. 7 Paper Properties 339
4B. 8 Temperature 340
4B. 9 Equipment 342
CHAPTER 4C: RUNNING THE PRESS SECTION.
4C. I Daily Operation 345
4C. 3 Cross-web Variations 355
4C. 4 Practical Points 360
References 366
vi
CONTENTS
PART 5: THE DRYING SECTION AND CALENDERS

SI Introduction 371
CHAPTER SA: THEORETICAL CONSIDERATIONS
SA.1 Removal of Water in Drying 373
SA. 2 Performance of the Drying Section 387
SA. 3 Effect of Drying on Paper Properties 394
SA.4 Calendering 411
CHAPTER SB: OPERATING FACTORS AFFECTING DRYING AND
CALENDERING.
SB. 1 Steam Heating and Condensate System 422
SB. 2 Felts and Felt Drying 434
SB. 3 Conditions inside the Dryers 444
SB.4 Ventilation 453
SB. 5 Auxiliary Drying Methods . 461
5B.6 Machine Speed and Draw Control 471
SB. 7 Paper Properties 475
SB.8 The M.G. Cylinder 476
SB. 9 Calenders 481
SB. 10 Equipment 487
CHAPTER SC: RUNNING THE DRYERS AND CALENDERS.
SC. 1 Daily Operation 497
SC. 2 Maintenance 506
SC. 3 Control of Moisture Content 515
SC .4 Control of Thickness . 528
SC. 5 Practical Points 535
References 546
PART 6: PRODUCTION CONTROL

61 Introduction 553
CHAPTER 6A: MONITORING MACHINE PERFORMANCE.
6A. 1 Analysis of Downtime 556
6A.2 Performance Data 561
6A. 3 Monitoring Quality . 570
CHAPTER 6B: IMPROVING MACHINE PERFORMANCE.
6B. l General Survey . 581
6B. 2 Financial Considerations 586
6B. 3 Performance under Different Operating Conditions 600
References 613
Index 615
Vll
FOREWORD
This book has been written primarily for persons who are, or hope to be,
directly concerned with the production of paper or with technical aspects
of papermaking. It assumes some basic knowledge and a fair degree of
familiarity with paper machines. The intention has been to provide a
comprehensive background of information which can be readily referred
to, but is at the same time sufficiently readable to serve as a text book for
advancing knowledge of the process. In scope it fills a gap between general
introductory books on papermaking and the large reference works
which are mainly concerned with giving a detailed description of the
various types of equipment available. When supplemented with practical
experience and an intimate knowledge of the particular machine under his
charge, it should assist the operator and supervisor to use the equipment
in a systematic and informed manner and obtain the best performance
from it.
The book is divided into six parts. The first five of these deal with
separate sections of the Fourdrinier paper machine: 1, the wet-end flow
system; 2, screens and cleaners; 3, the wire part; 4, the press section;
and 5, the dryers and calenders. Each of these parts is divided into three
main chapters. The first chapter describes theoretical aspects of the process
as derived from the results of research work. This material is presented as
simply as possible, but to avoid undue length it is assumed that the reader
has some familiarity with graphical presentation and a basic knowledge
of physics. The approach adopted is essentially descriptive rather than
analytical and is not intended as a detailed and exhaustive survey. Rather
the object is to give a general background of fundamental principles so
far as they are known, and then discuss their relevance and importance
to efficient operation.
The second chapter of each part concerns operational aspects of the
paper machine and details the various factors which can affect and upset
the process. Particularly in these chapters the results of experimental
work, for the most part reported in papermaking literature, have been
freely drawn on and interpreted in the light of theoretical considerations.
An attempt is made to place into perspective the influence of different
operational variables associated with running a paper machine in order
to indicate how best various quality requirements can be fulfilled and
the whole operation brought under more efficient control. Inevitably
this involves making comparisons between different designs of equipment
but to ensure that such remarks are as impartial as possible reliance is
placed almost entirely on published reports and experimental comparisons;
any comments based on personal experience or manufacturers' pamphlets
are carefully identified.
The third chapter of each part is concerned entirely \vith practical
lX
FOREWORD
aspects of paper machine operation. The value and use of process instru-
mentation and controls are discussed, maintenance requirements are
detailed, and the main tasks of machine crews are described with particular
emphasis on the problems of running the machine to produce a consistent
product. In this respect the requirements for making periodic comprehen-
sive checks through the machine system are given special attention
because it is considered that these provide an essential background of
information for successful long-term operation. There is also some dis-
cussion of the basic practical tasks involved in machine operation but in
this no attempt is made to be thorough and only certain common, and
in most cases to the experienced man doubtless commonplace, principles
are stated. The usual topics of start-up and shut-down routines are covered
briefly and consideration is also given to discussing what to pay attention
to when operation is proceeding smoothly. Practical knowledge must be
acquired by actually running a paper machine, and the remarks in these
particular sections give only a general background against which to relate
the characteristics and peculiarities of any individual machine.
Part 6 is concerned with methods of monitoring the performance of
paper machines and describes the operating data required if adequate
financial control is to be achieved. This subject, which comes under the
broad heading of Production Control, is one that has been rather neglected
especially in relation to making a comparison of the profitability of
running different grades on a machine, and indeed of running a single
grade at different speeds. These highly important questions are discussed
in some detail, and other equally interesting topics such as the costing
of improvements in machine operation and assessing the economic benefit
from installing new equipment are also covered.
This book, then, is concerned with running a paper machine to achieve
the best possible performance. The machine itself, and the ancillary
equipment and instrumentation associated with it, are not described in
any detail except for a few less well-known items where a brief description
appeared necessary for clarification. If the reader requires information
about the design and basic operation of papermaking machinery he
should refer to the listed references or to one of the several standard text
books in which this subject is already well covered. Units throughout are
English, unless otherwise stated.
Most of the contents have already appeared in The Paper Maker during
the period from September 1962 to April 1967. For publication in the
present book form the material has been substantially revised and brought
up-to-date to include references appearing in papermaking literature up
to the end of 1966. References earlier than 1950 have generally not been
consulted except in a few specific cases, but when appropriate use has been
made of information in several books, particularly the following:-
R.H. Clapperton and W. Henderson. Modern Paper Making. Waverley

Book Co. Ltd. 1952.
J. N. Stephenson (ed.). Pulp and Paper Manufacture, Vol. III. McGraw
Hill Book Co. Inc. 1953.
x
l'OREWORD
H. Hardman and E. J. Cole. Paper Making Practice. Manchester

University Press, 1960.
J. Mardon, R. G. Arklie, A. Mclnnis, and R. C. Buser. Paper Machine
Crew Operating Manual. Lockwood Trade Journal Co. Inc. 1961.
C. E. Libby (ed.). Pulp and Paper Science and Technology, Vol. II.
Paper. McGraw Hill Book Co. Ltd. 1962.
G. Gavelin. Fourdrinier Papermaking-Stock Supply through Presses.
Lockwood Trade Journal Co. Inc. 1963.
To the authors of these books, and of the many references quoted in

the body of this book, due acknowledgement is given.
Lastly the author must record his tremendous debt to the many people
who, directly or indirectly, have contributed to this book. This applies
first to the Editor and Associate Editor of The Paper Maker, Stuart Don
and Jack Elliott, who by their interest and encouragement over the years
transformed a plan for a few articles on papermaking into a full-scale
project with the present book as its objective. Then, as regards the material,
the author wishes to record his deep gratitude to those people from whom
he has learnt so much of the real practice of papermaking, particularly
the Dixon family and colleagues at the Grimsby mill of Peter Dixon and
Son Ltd., especially Martin Armitage, Dick Arnett, Bill Harrod and Bill
Ross. The greater proportion of information in the book concerning the
actual practical operation of paper machines has come from day-to-day
discussion of production problems with associates at Dixons.
Finally, a special debt of gratitude is due to three people who have
devoted much time and effort to reading through parts of the manuscript
and making numerous suggestions and corrections: Jim Burton, Works
Manager at Culter Mills, Peterculter, Aberdeenshire, who provided much
information and support in the early days of preparation; Tony
Truman, Assistant Research Manager at Kearsley Mill, Stoneclough,
Lancashire, who has closely vetted the theoretical and technological
sections, and in the process revealed a considerable talent as a proof
reader; and Bob Fyfe, Chief Papermaker at Oughtibridge Mill, Yorkshire,
whose more recent forthright comments created trouble with corrections at
the printers. The author is also grateful to two persons who made
valuable additions to Part 6 of the book: Alan Marriott of Kimberley-
Clark, Ltd., Larkfield, Kent and Ian Kenworthy of the British Paper and
Board Industry Research Association, now the Paper and Board, Printing
and Packaging Industries Research Association, at Kenley, Surrey. To
these friends must go a great deal of credit for this book. It is a reflection
on the bond of interest which throws members of the paper industry to-
gether that they were willing to spend so much of their leisure time in the
arduous and difficult task of reading the material through with a critical
eye.
xi
PART 1
THE WET-END FLOW SYSTEM

INTRODUCTION
11 The Fourdrinier machine as a process for making paper may con-
veniently be regarded as commencing at the machine chest. Up to this
point the stuff has been suitably prepared from pulp by beating, refining
and blending of different furnishes, and as a general rule the fibre receives
no further treatment. The machineman takes over complete responsibility
for the process at the machine chest, his primary object then being to form
and dry the sheet of paper; the characteristics of that sheet are largely
determined by the treatment the fibre has received in the preparation
stage so the major requirement of the paper machine system is to produce
the sheet at the desired substance and moisture content with, in the general
sense of the terms, a suitable finish and as good a formation as possible.
By the time the fibre suspension has reached the slice the most critical
aspects of this task are settled, certainly with regard to substance and to a
large extent also with regard to formation. It is this important part of the
process, from the machine chest to the slice, that forms the subject of
Part 1 of this book.
It is not intended to cover completely the various pieces of equipment
which are found between the machine chest and the slice; in particular,
the cleaning and screening of paper stock is not dealt with and consistency
regulation is only touched upon in passing. On the other hand the back-
water system, in its complete sense, and to a lesser degree the broke system
are considered, as also is the question of retention in the wire section.
Each of these subjects has some bearing on the manner in which the stock
feeding the breast box and slice is constituted and as such can affect the
making of the sheet and the substance of the paper. The various factors
affecting the substance of the paper both along and across the machine are
all treated in detail but the formation of the sheet is only considered in so
far as it is influenced directly by the condition of the stock leaving the slice.
(Thus, the characteristics and angle of flow of stock from the slice are
dealt with but the conditions on the wire following impingement are left
to Part 3 ; the overall consistency and condition of the backwater passing
through the wire are considered but not the details of drainage).
The general function of the wet-end flow system is to receive stuff from
the machine chest at a relatively high consistency, dilute it and lead it to
the slice in such a manner that a satisfactory sheet can be made on the
wire. The total dry weight of the paper must be correct and steady and this
is dependent on the various flows and consistencies which constitute the
wet-end and also on the general retention conditions on the wire. The
profile of the dry weight of the paper across the machine must also be
even and this depends almost entirely on uniformity of the fibre flow and
velocity from the slice; the efficiency with which this is achieved is in turn
3
11 .1 THE WET-END FLOW SYSTEM
dependent essentially on the approach flow and breast box design. Apart
from these requirements the flow leaving the slice should ideally carry
fibres evenly dispersed throughout its volume, not flocculated in any way
nor (except for some special paper property) aligned in any preferential
direction; this is generally considered one of the most difficult objectives
to achieve and requires just the right degree of small-scale turbulence at the
slice. With all these conditions adequately fulfilled the task of forming
the sheet, though still very dependent on having the correct velocity relation
between the slice flow and the wire and on suitable controlling of the
drainage conditions on the wire, is well on the way to completion.
11 . 1 Terminology
It is necessary to use several terms in the following pages which are neither
adequately defined nor used in exactly the same sense in different paper-
making circles. There is also a difference between the common terminology
in use in this country and on the continent of North America. For this
reason the author decided to standardize on the same terms throughout
and these have been selected with the object of avoiding confusion and
ambiguity in the mind of the reader.
'Head box' is used in papermaking literature to refer both to a high-level
box for constant-gravity feed of stuff to the machine, and also as the name
of the box situated above the slice. This term will therefore be avoided
and for the first function 'stuff box' will be adopted ('service box' has the
same meaning, but the terms 'mixing box' or 'regulating box' are only
used when, in addition to providing a constant head, the thick stuff is
diluted in the box); for the second function the old term 'breast box' will
be used since this to the papermaker is more familiar than the alternative
name 'flow box.'
In systems using a pump for mixing and providing a pressure for feeding
the breast box the term 'mixing pump' will be used in preference to 'fan
pump' or 'stock pump' since it gives a clear indication of the position of
the pump in the wet-end flow system and emphasizes its function as a
replacement of the mixing box; it also allows 'fan pump' or 'feed pump'
to be applied to a second pump when this is used to take stock from the
cleaners to feed the breast box.
In the breast box one piece of equipment has received a plethora of
designations: 'evener roll,' 'distributor roll,' 'holey roll,' 'monkey roll,'
'hog roll,' and 'perforated roll' are all in common use. The latter term,
'perforated roll,' seems to strike a balance between the familiar and the
presumptuous and will be the one adopted.
Finally, 'backwater system' will be used to refer to the flow of stock
drawn through the wire when it is used directly on the machine; 'white-
water system' will be used to describe the excess flow which is pumped
out of the wire system either to a temporary storage tank for feeding
consistency regulators, beaters or hydrapulpers in the preparation plant
or directly to a clarifier or save-all. 'Stuff' refers to the fibre suspension in
the machine chest at a consistency of 3 per cent. or more; 'stock' to the
4
THE WET-END FLOW SYSTEM 11.1
diluted suspension feeding the breast box at consistencies usually under 1
per cent.
Consistencies are always understood to refer to bone-dry values, never
air-dry since this can too easily lead to error and confusion. Loading or
ash content is most conveniently represented as a percentage of the total
solids.
5
CHAPTER lA
THEORETICAL CONSIDERATIONS
lA.1 FLOW OF FIBRE SUSPENSIONS
Though seldom a subject of direct concern to the papermaker, it is desir-
able to have some idea of the manner in which fibre suspensions flow. This
is particularly important for understanding the formation of fibre floccula-
tion and also for a critical examination of the approach flow and breast
box design. The general arrangement and sizing of pipes and channels at
the wet-end is frequently rather crude when looked at from the viewpoint
of papermaking requirements and often unnecessary power loss and
aeration may occur; a knowledge of the particular characteristics of the
flow of fibre suspensions as compared with other liquids is helpful when
considering such aspects as these.
It is proposed firstly to discuss briefly the way in which water flows and
then to describe the modification caused by the presence of fibres and
the manner in which individual fibres are carried in the flow. The question
of frictional resistance to flow and the development and breaking down
of fibre floes will next be dealt with, leading to a discussion of the flow
velocities desirable for the various parts of the wet-end system. Brief
mention will also be made of the relationship between pressure and flow
velocity with, finally, a few points on centrifugal pumps.
lA. 1 1 Effect of velocity on the pattern of water flow

At very low velocities the flow of water in a pipe is streamlined. This
implies that each particle of water follows directly behind the one in front
and keeps in exactly the same relative position in the cross-section of the
pipe, all the flow streams being parallel to the pipe axis. On entering a pipe
the velocities of the flow streams are at first uniform throughout the pipe
cross-section. But at the wall of the pipe friction affects the velocity of the
flow and slippage occurs between one streamline and the next. The region
in which this happens is known as the boundary layer and as the water
travels down the pipe this layer gradually grows to influence more and
more of the streamlines. Eventually the boundary layer converges to the
centre of the pipe and embraces the entire flow; under this condition the
flow pattern is in equilibrium and the velocity of the streamline varies
ideally in a parabolic relation from zero at the wall of the pipe to a maxi-
mum in the centre. In this type of flow, known as laminar flow, the pressure
drop or friction loss along the pipe is proportional to the velocity. Fig. 1.1
illustrates the changing velocity profile and growing boundary layer in the
pipe.
As the quantity of water :flowing through the pipe is increased the
laminar pattern gradually begins to break down. Instead of the flow
keeping strictly parallel to the pipe, eddies begin to form in the centre
6
THE WET-END FLOW SYSTEM lA.l 2
of the pipe so that the velocity there fluctuates from one moment to the
next. With further increase in the average flow velocity the influence of the
eddies grows and the flow pattern passes through a transitional stage until
it becomes fully turbulent. In this condition, which is the most usual in all
practical engineering applications, the velocity of flow, though fluctuating
from one instant to the next, is practically the same on average throughout
the pipe cross-section with only a slight increase from the wall to the centre.
The true boundary layer where the flow is still effectively laminar becomes
confined to a narrow region close to the pipe wall. Because of the energy
loss in eddies the pressure loss along the pipe is much greater than in
laminar flow and becomes approximately proportional to the square of the
velocity.
These changing patterns of flow in a pipe apply to all homogeneous
fluids and the point at which laminar flow begins to break down is known
as the critical velocity; this velocity varies inversely with the diameter of
........___ l!>ouMt>A~Y L.A.YtR
~IkTIE~{f- V;L.OC\iY P~OFll..lS

Fig. 1. 1. Growth of boundary layer and change in velocity profile of water entering a
pipe at very low velocity
the pipe and the density of the fluid and is proportional to the viscosity.
A dimensionless number formed by taking the product of the velocity, pipe
diameter and density of the fluid and dividing by the viscosity has the
same value of approximately 2,000 at the critical velocity for all fluids
flowing in a pipe. This expression, termed the Reynold's number, is very
useful for comparing the behaviour of different fluids under different
conditions and predicting when the flow pattern changes.
IA. 1 2 Modification to flow pattern produced by fibre

When fibre is introduced into water the flow patterns that have been
described are greatly modified. The work of Robertson and Mason
(26, 32, 37) contributed most to an understanding of this and what follows
derives largely from their work. Even at very low consistencies of less than
0·05 per cent. the presence of the fibre destroys the homogeneity of the
fluid and considerably restricts the freedom of movement of individual
water particles. This effect is most marked at very low velocities through a
pipe when the whole mass moves along the pipe almost as a solid body.
This is known as plug flow and in this state there is virtually no relative
movement over the whole cross-section of the pipe, the velocity at all points
being the same apart from a small boundary layer at the pipe wall where
the velocity drops to zero. The boundary layer is free of fibres except at
extremely low velocities where small floes of fibres appear to gather in
the layer and roll along the wall. This form offlow is very similar to that for
7
lA.l 2 THE WET-END FLOW SYSTEM
water entering a pipe at low velocity before the boundary layer has spread
inward in the manner described above, and it may be considered that the
presence of fibre permanently inhibits the normal development of the
boundary layer.
As the flow velocity is increased a point is reached where the fibre-free
wall layer becomes unstable and a turbulent annulus is formed round the
plug. Further increase of the velocity causes progressive enlargement of
this annulus and disintegration of the plug until a point is reached where
the flow becomes completely turbulent.
Apart from changing the pattern of flow with increasing velocity the
presence of fibre substantially alters the frictional resistance and pressure
w
':: (00
t: so
0
0
t 10
§"' s
lU
Q(
~
'11
~ 2
°'
0.
o. 2 5 Co 2o
FL.OW VELOCITY Fr. P£1't S!CONO.
Fig. 1. 2. Pressure loss at different flow velocities and consistencies in a pipe
loss for flow in a pipe. At low velocities the pressure loss is greater for
fibre suspensions than for water and the higher the consistency, the greater
the difference becomes. However, with increasing pipe velocity the presence
of fibre also appears to delay the onset of completely turbulent conditions
in which the pressure loss increases with velocity at a much higher rate,
with the result that a point is reached where the pressure losses are actually
lower than for pure water. With further velocity increases in the turbulent
region the energy losses occurring continue to be slightly lower than for
water and this is thought to be due to a reduction in the scale of turbulence
caused by the presence of fibres. Under normal papermaking conditions
velocities in pipelines are in the turbulent region and the head loss is
roughly proportional to the square of the flow velocity as with water; for
consistencies under about 1 per cent. the general characteristics of flow
in the turbulent region are in fact very similar to water.
The relationship between pressure loss and flow velocity is illustrated
in Fig. 1. 2, which shows the sort of results that have been obtained for
8
THE WET-END FLOW SYSTEM lA.1 3
fibre suspensions at different consistencies. The curves are drawn from
some graphs given by Bonnington (35) in a survey of frictional losses in
pipes carrying pulp. It will be observed that the pressure loss over part of
the curves actually shows a decrease with increasing velocity; this is
characteristic of fibre suspensions and is considered to be due to friction
changes in the plug flow region as the layer close to the pipe wall becomes
fibre-free and alters in thickness. However, a precise relation with the
changing patterns of flow described above has not yet been satisfactorily
obtained. The actual velocities and pressure losses depicted in Fig. 1. 2
can only be taken as illustrative because both Robertson and Mason (32)
and Bonnington (35) emphasize that pipe flow data from various sources,
though extensive, is highly contradictory. Changes in pressure losses under
different conditions do not agree when the Reynolds numbers are the
same, as with water, but are apparently influenced considerably by the
characteristics of the fibres in suspension. The peculiarities of fibre flow
which can lead to such phenomena as stapling and collecting on sharp
edges and the tendency to settle and coalesce at low velocities also compli-
cate the picture. The formation of floes of fibres has a particularly import-
ant relation with the flow pattern and this will now be discussed.
lA. 1 3 Flow of individual fibres and flocculation

In an experiment using rigid and uniform acrylic fibres at low concentra-
tion in a tube, Baines and Nichol (22) established that in slow laminar
flow the fibres have a tendency to align close to the streamlines with small
oscillations about this orientation. With the onset of turbulence at greater
flow velocities, fibre orientation in the flow becomes random, though near
to the wall of the tube a preference for alignment along the wall persists.
In a region of contraction with flow acceleration the fibres show a tendency
to align with the flow even at turbulent velocities. Under all conditions the
fibres travel at the same velocity as the flow.
These findings are in broad agreement with earlier results by other
workers who have used papermaking fibres under different experimental
conditions, though some authors, notably Van den Akker (16), earlier
considered that fibre alignment in laminar flow does not occur. In par-
ticular, the work of Mason and Robertson (13, 14, 15) showed that in the
presence of a simple velocity gradient fibres tend to align more with
flow while showing twisting and bending motions round themselves; in
turbulent conditions, however, the orientation is random. In normal
papermaking, turbulent conditions apply in pipe flow but at the slice the
acceleration of flow produces some preferential alignment in the direction
of flow; this has been clearly demonstrated by Wrist (74) who found that
a degree of fibre orientation in the machine direction exists in the slice
jet and is dependent on the consistency of the stock and, to a lesser extent,
the type of fibre and design of the slice itself.
The work of Mason and Robertson mentioned above was directed
primarily towards gaining an understanding of flocculation and it is in
this connection that some interesting facts emerged which have since been
9
lA. l 3 THE WET-END FLOW SYSTEM
verified by Steenberg (92). It appears that flocculation is primarily a

mechanical phenomenon occurring when conditions of flow are such that
individual fibres are lead to contact each other in a relatively gentle manner.
The conditions producing mechanical entanglement are those where mild
shear forces exist which allow slow relative motion between individual
fibres; while slowly passing each other fibres then have a tendency to
entangle and adhere together forming a 'doublet' (the number of doublets
formed increasing with time) and some of these may in turn collect further
fibres resulting in a growing size of floe. It was also shown that the greater
the fibre concentration and the longer the length and greater the flexibility
of individual fibres, the more flocculation tends to occur.
This phenomenon is associated with the increase in network strength of
the suspension which occurs in these conditions, i.e., in the ability of the
fibre mass to resist breaking up under the influence of tensile and shear
forces. When subjected to disrupting forces a network of fibres yields in
regions where the consistency happens by random chance to be lower
than the average; such regions are fewer when the suspension is at a higher
concentration or more uniform in structure, in which case in given condi-
tions of shear, the floes generated will be greater in size.
While promoting flocculation by entanglement of fibres, shear forces
act at the same time as a disrupting factor, pulling the floes apart by causing
relative motion between different parts of the floe. In fact, the life of a
fibre doublet was shown in the work of Mason and Robertson to be
inversely proportional to the shear rate, i.e. the greater the shear rate, the
quicker floes break up. For this reason, individual floes reach a maximum
size in given conditions of turbulence, and also the number of floes appears
to reach a state of dynamic equilibrium in which the rate of formation and
disruption becomes equal.
From this explanation it is apparent that conditions in a chest where
fibres are kept in suspension by mechanical shear can easily lead to floccu-
lation. The degree of flocculation in pipe flow at low velocities depends
very much on the change in flow pattern as the plug flow pertaining under
these conditions gradually breaks down. Where flow becomes turbulent,
however, the increasing shear forces which occur will rapidly produce
dispersion.
In one interesting experiment, using consistencies up to 0·7 per cent.,
Mason and Robertson confirmed by an optical method of measuring
flocculation in a tube that in turbulent flow floe dispersion is promoted
by increased flow velocity and hindered by increased consistency. Chemicals
such as deacetylated Karaya gum were also investigated and shown to
reduce flocculation; their effect was considered to be due to the deposition
of a lubricating film on individual fibres. Retention aids and whitewater
flocculants such as Sveen glue, activated silica sol, and sodium aluminate
have the opposite effect of increasing flocculation but this is thought to
be largely an electrolytic action.
The degree of flocculation in stock is ultimately of greatest importance
in the slice jet. At this point there are many· difficulties in obtaining a
satisfactory measure of flocculation though optical methods similar to
10
THE WET-END FLOW SYSTEM lA.14
those employed by Mason and Robertson have been successfully used
by Wrist (74) and, on an actual paper machine, by Beveridge and Bridge
(80). Correlation with the formation properties of paper is necessary before
this work can be usefully interpreted but the technique is likely to prove
valuable in future investigations.
lA. 1 4 Desirable flow velocities at the wet-end

If flow velocities of fibre suspensions are low there may be a tendency for
flocculation to occur and the fibres may also begin to settle. In addition,
because of the increased likelihood of regions where the flow is practically
stationary, air bubbles will gather increasingly on the surface producing
foam and slime accumulations. On the other hand, if flow velocities are
high undue pressure losses will be sustained and there is an increasing
chance of pressure fluctuations with surges in flow due to unstable flow
separation at bends and valves in the pipework and the presence of excessive
eddying in channels and boxes. Further, suction of air into the stock
becomes more likely and this can also be very detrimental.
For flow in pipelines the ideal is probably a mildly turbulent condition
since this will best avoid the various disadvantages of extremes of low and
high velocity. Unfortunately data available for a reasonable estimation
of the appropriate velocities under different conditions is, as has already
been emphasised, contradictory, and the critical velocities for onset of
turbulence are very dependent on fibre characteristics. However, for
normal pipelines carrying stock feeding the breast box a reasonable
compromise is a flow velocity of between 5 and 10 ft. per second. The
relevant flow for pipework is readily calculated from either of these
formulae:
FI . . . ft I _ 0·49 x Flow in gallons/min.
ow m pipe m · sec. - (Diameter in ins.) 2
_ 7·66 x B.D. production in long tons/24 hrs.

- (Diameter in ins.) 2 x Consistency
The latter term assumes no loss or re-circulation of fibre.
For higher consistencies, to achieve the same flow condition requires
higher velocities: for 3 per cent. stock between 15 and 20 ft. per second is
probably the velocity region which would be needed. But in practice this
would lead to very high friction losses and it is not usual to plan for a
velocity over about 10 ft. per second. For the stuff pipe feeding a mixing
pump, however, it is worth considering another aspect. In this case a
predominant requirement is that the flow of fibre down the pipe should
be constant. With the pressure difference along the pipe and across the
stuff control valve constant, as under given conditions they will be to all
intents and purposes, the flow velocity is altered only by a change in
consistency producing a different frictional resistance along the pipe.
Since a greater consistency increases the pressure loss, the flow is neces-
sarily reduced; careful choice of the average velocity down the pipe can
11
lA.1 5 THE WET-END FLOW SYSTEM
cause the increase in consistency to be closely balanced by the consequent

drop in flow, with a net effect on the total flow of fibre which becomes
negligible. To achieve this condition it is necessary that an increase in
consistency produces a proportionate increase in the total pressure drop.
In most installations the stuff control valve will account for over 80 per
cent. of the total pressure drop so that the pipe loss must increase by a
proportion roughly four or five times larger. It appears from graphs
similar to that shown in Fig. 1. 2 that this order of change in pressure loss
with consistency only occurs in the laminar flow region where the velocity
required is rather lower than that producing turbulent conditions. In most
cases velocities of about 4 to 5 ft. per second will be fairly close to the
ideal; this is lower than is often prefered but the pipe run will generally
be vertical and short enough to make it unlikely that settling or flocculation
become troublesome. Also with the normal system of feeding from a
stuff box it is not necessary to have the box very high to achieve this order
of velocity and a further advantage is that the lower pressure loss from
pipe friction means that the flow control valve is more sensitive.
For the breast box and pit, particularly the latter, the average flow
velocity will be much lower than in other parts of the system and will be
dependent on the varying cross-sectional area below the surface level. The
flow patterns at different velocities described for pipes cannot be extended
to flow under these conditions and fuller discussion is left to the appropriate
section later.
IA. 1 5 Relation between pressure and flow

In liquid flow of any kind if pressure losses due to wall friction and
turbulence are ignored application of the principle of conservation of
energy shows that the total energy compounded of potential energy above
a datum line, kinetic energy of flow, and static pressure energy must remain
constant. Expressed for convenience in equivalent feet head of liquid this
becomes Z + v2/2g + P constant, where Z is the height above the
datum line in feet, v is the velocity in ft./sec., g = 32 and P is the static
pressure (relative to atmospheric pressure) in equivalent feet head of the
liquid, as would be indicated by an ordinary pressure gauge. This simple
formula is very useful when considering changing hydrodynamic conditions
and it is necessary to appreciate the implications of it if the significance
of pressure gauge readings is to be completely understood.
In the first place it is important to note that a change in level, Z, with
everything else constant produces a change in the pressure, p: thus, a rise
of 23 ft. in a pipeline with stock flowing through will reduce the reading
of a pressure gauge by 10 p.s.i. (1 p.s.i. = 2·3 ft. head water). For a change
in velocity, such as occurs with changing area of the cross-section of flow,
a similar alteration occurs in the static pressure. For example, an increase
in velocity from 8 ft./sec. to 32 ft./sec. caused by halving the diameter of
a pipe will produce a decrease of pressure equivalent to 15 ft. of head or
fr5 p.s.i. This phenomenon is used, of course, in orifice and venturi flow
meters where measurement of the difference in pressure at two different
12
flow velocities enables the quantity of flow to be determined. If the cross-
section is reduced considerably producing a large increase in flow velocity
the static pressure is also considerably reduced and can become lower
than atmospheric. This principle is usefully used in the various types of
ejector. Normally, however, such a condition is undesirable for two reasons.
Expansion of the cross-section after the constriction can cause the flow to
separate and leave the wall of the pipe, producing highly unstable conditions
with pressure surges down-stream and excessive turbulence. Also if, due
to high velocity in the constriction, the pressure in the fluid falls below the
vapour pressure at the prevailing temperature, the flow boils and bubbles
form which are carried with the stream and collapse when a region of
higher pressure is reached; this is known as cavitation and leads to the
formation of pressure waves in the fluid as well as producing serious
mechanical damage to the pipe wall. Cavitation is more likely to be trouble-
some at the suction side of pumps but excessive throttling of valves can
also be a cause.
If a flow of water emerges from an orifice into air the static pressure of
the liquid becomes zero. Reference to the formula shows that this must
be accompanied by an increase in velocity compared to the velocity just
inside the orifice where some static pressure can be sustained. If the flow
is to keep intact, this in turn implies that the cross-sectional area must
diminish in order that the quantity is unaltered. This reduction in the
area taken up by the flow is readily observed as a narrowing in the width
of a jet issuing from a hole in a tank full of water, and is partly the cause
of a growth in instability in the flow as the jet gets further from the hole.
A similar phenomenon occurs when stock issues from a slice and one
practical significance is that the velocity of the flow at the point of leaving
the slice (and also, to a slightly lesser extent, at the point of meeting the
wire) is lower than the theoretical velocity which would be achieved if all
the pressure head behind the slice were turned into kinetic energy of
velocity. The velocity is, therefore, not given simply by the equation
v2/2g = h or v = y2gh, where h is the head behind the slice, but a
coefficient less than unity, C, must be introduced to take account of the
contraction in the jet and also of a slight friction loss, giving v = Cy2gh.
The value of C depends entirely on the geometrical character of the opening:
in vertical slices it is between 0·6 and 0·7 but in projection slices it is close
to unity. It must also be realised that up to the point where the jet is
fully contracted the 'vena contracta,' the velocity is continually increasing
so that the velocity of the jet when it lands on the wire is dependent to
some extent on the distance it travels. For the purpose of calculation it is
normally convenient to refer to the velocity at the orifice rather than the
vena contracta in order that calculation of the volume discharge involves a
straightforward multiplication by the actual dimensions of the slice
opening.
In practice the general equation relating velocity and pressure requires
modification due to the fact that energy losses occur from heat generation
in turbulence and friction at the pipe walls. This effectively reduces the
value of the constant in the equation as the flow progresses through the
13
1A.l 6 THE WET-END FLOW SYSTEM
system. Pressure losses of water in pipework under different conditions

are fairly well known and it has already been mentioned that in the normal
turbulent flow the loss increases proportionally to the square of the
velocity. Bends, particularly when sharp, tee-junctions, orifices, and
especially valves which are used to govern the flow by alteration of the
pressure drop across their opening all usually contribute greater head
losses in the system than simple pipe losses. For fibre suspensions there is
comparatively little data on losses in pipe fittings though at consistencies
above 1 per cent. they are likely under normal conditions to be two or
three times those for water. Perforated rolls, evener plates and other
equipment often found in a breast box all contribute a pressure loss which
affects the effective head behind the slice when they are situated close to
the slice; this will be considered in greater detail later.
1A. 1 6 Centrifugal pumps

Centrifugal pumps of varying design are the most common type used for
providing pressure at the mixing pump. Fig. 1. 3 shows the usual sort of
Or -:.
w
~
>-
(j
...:
u.
0 "Z
0. 0
~ <(
.:i u w
:x;
~
Or:'
0
u:
IJ..
:c w 00
HEAD
IZS 100 50
100 So 4o
75 60 30
50 4.o 2o
25 2o lo
0 0 0
0 1,000 2,000 3,000 4,ooo 5 1 000
DISCHAQQw GA~~Oft.IS Pi~ MIN.
Fig. 1. 3. Performance of a typical centrifugal pump with varying water flow
alteration in performance with varying water flow of this type of pump.

When pumping fibre suspension, particularly at stuff consistencies, these
characteristics are modified considerably and the pressure head and
efficiencies are both reduced with increasing consistency. The air content
in the suspension also affects the performance appreciably and in extreme
14
cases the pump can, of course, cease to function; for this reason it is
always preferable to have the suction side of the pump drowned to reduce
the ease with which air can separate out in the pump body.
Partly because of the paucity of reliable data on stuff pump performance
the manufacturers have a strong tendency to be on the safe side and supply
pump impellers and motors which are amply over-sized. The initial expense
is therefore higher than necessary but this is a negligible factor compared
to the increased cost of power consumption resulting from a lower running
efficiency, and the necessity for heavy throttling on the pump discharge
to dissipate a large proportion of the power generated (this increased cost
can easily equal the initial cost of the pump within a few months). When
faced with an over-powerful pump the simplest answer is usually to
substitute a smaller size impeller. Re-circulation of a proportion of the
flow permits a reduced pressure drop across the main flow control valve,
but apart from the risk of instability if high recirculation flows are used
this procedure generally increases power consumption dependent on
whereabouts on the pump performance curves the normal running condi-
tion is located.
These remarks assume that the motor is a.c. and that control of the
flow need not be too critical. At the mixing pump it is far more important
to secure a smooth flow and this is closely connected with the method
used to control the flow, a topic dealt with in more detail in lC.1 4. It
can be noted here, however, that for this particular application a variable
speed drive from a d.c. motor has much to commend it if the supply is
available, and the additional expense involved in this is very soon covered
by the large power saving possible.
lA.2 INFLUENCE OF WET-END DESIGN ON OPERATIONAL

STABILITY
It is often thought that if the flow and the consistency of fresh stuff entering
the mixing box or pump at the wet-end is closely controlled then, ignoring
moisture differences in the finished paper, the substance will also be closely
controlled provided the machine speed and draw are stable. While this is
largely true it is only correct up to a point and in this section it is proposed
to discuss the various factors which, at least on a theoretical level, can
affect the substance of the paper even when the flow and consistency of
fresh stuff are perfectly steady. In many paper machine systems some of the
points about to be discussed are unlikely to have any great significance,
but it is important nonetheless to be aware of the factors that could have
a possible influence. Many long-term or cyclic weight fluctuations which
defy elimination can be due to unsuspected variations in the backwater
and whitewater system and it is the effect of these, and of deliberate
alterations made by machinemen in the course of their work, that is now
considered. It is likely in the near future that the use of dynamic models
of the wet-end flow system in conjunction with analogue and on-line
digital computers will bring considerable progress in this field, see for
example references 76 and 87.
15
lA.21 THE WET-END FLOW SYSTEM
1A.2 1 Backwater system balance at the wet-end

It is common practice nowadays to use backwater from the wire table rolls
and suction boxes to dilute the fresh stuff before it passes to the breast
box. It is proposed firstly to consider this particular flow, which for con-
venience will be referred to as the backwater circuit. This circuit can be
thought to have inputs of water and fibre in the form of fresh stuff, and
also to a lesser extent in the form of whitewater showers, together with an
input of fresh water in breast box and cleaner showers, sealing glands,
wire sprays in the main pit, etc. This must be balanced by the water and
fibre leaving in the paper passing from the couch, together with the excess
backwater which is pumped out of the system and any reject from the
cleaners. The excess backwater can be removed either separate from or
Fig. 1 . 4. Simple wet-end flow system. Flow volume figures printed above, consistencies
below, for each flow
together with the trim from the wire and the wire sprays above the couch
pits, but this does not affect the balance in the backwater system, only in
the whitewater system and this will be considered later.
It follows that any change occurring in any of these inputs or outputs
can affect the fibre and water leaving the system in the paper web and as
such can affect the dry weight of the paper. A blocked whitewater shower or
a change in the cleaner reject flow to drain could, for example, affect the
weight. Apart from the fresh stuff most flows of this nature should be
sufficiently steady in practice to be neglected, at least over a relatively short
period; the volume of water leaving in the paper will also have a negligible
effect on the overall balance, so that the excess backwater should normally
have a steady flow when the fresh stuff is also steady. However, the fibre
volume leaving in the excess backwater flow will be influenced by any
change that may occur in the backwater circuit consistency and this must
affect the weight of the paper. Such a change may be deliberate, as when
the machineman observes that the sheet appears wetter or freer and alters
the amount of 'water on the wire,' i.e. the quantity of backwater in relation
to fresh stuff used for mixing, or it may be accidental, as when retention
conditions on the wire change. Often, an attempt is made to reduce the
influence of these factors by using backwater from the suction boxes
preferentially for the excess flow since this is usually at a much lower
consistency than the table roll backwater. This expedient, when efficiently
carried out, effectively reduces weight variations of a permanent nature
due to disturbances in the backwater circuit. However, it is not often
realised that considerable transient weight changes are still possible.
16
THE WET-END FLOW SYSTEM 1A.2 2
This can best be illustrated by taking a specific example and for this
purpose the simple system fairly typical of a newsprint machine shown in
Fig. 1.4 will be used. Flow volume figures are printed above for each flow
and may be taken as gallons per minute, consistency figures are printed
below. Thus a fresh stuff flow of 300 gallons per minute at 3 per cent.
B.D. consistency is diluted at the mixing box or pump by 2,400 gallons per
minute of backwater at 0·3 per cent. This produces a flow of 2,700 gallons
per minute at 0·6 per cent. consistency at the slice. Retention, defined here
in the terms of the consistencies of breast box stock and backwater, is
50 per cent. The volume of water leaving in the paper is 50 gallons per
minute at 17·5 per cent. consistency and the excess backwater (assumed
to be from the suction boxes) is 250 gallons per minute at a consistency
of 0· 1 per cent. In terms of solids it is seen that 90 lb. per minute of fresh
stuff yield 87·5 lb. per minute of paper with 2·5 lb. per minute of excess
backwater flow.
1A. 2 2 Alteration in retention conditions

Suppose, firstly, that retention conditions on the wire suddenly alter from
50 per cent. to 33t per cent. The fresh stuff flow remains unchanged, and
it is assumed that the effect on the excess backwater consistency, composed
largely of suction box water, is negligible. At the time when the retention
drops a greater percentage of fibre passes through the wire; it follows that
less fibre is couched off the wire and the consistency at the couch reduces
together with the dry weight of the paper. This is shown in Fig. 1. 5a. The
backwater consistency, being higher now, raises the breast box consistency
and at the second pass, Fig. 1. 5b, raises all the consistencies at the wire
and begins to correct the weight of the paper. Fig. 1. 5c shows the situation
at the third pass and successive passes can be shown to cause a stepped
exponential approach in the backwater and breast box consistencies to
the final equilibrium conditions shown in Fig. 1. 5d.
The ultimate effect of the change in retention is seen, by comparing
Figs. 1.4 and 1. 5d, to be an increase from 0·3 per cent. to 0·55 per cent.
in the backwater consistency and an increase from 0·6 per cent. to 0·82
per cent. in the breast box stock consistency. The weight of the paper has
returned from the equivalent of 63·5 solids when the retention first dropped
to its original value of 87·5 solids, confirming that in equilibrium conditi-
tions if the input and excess backwater fibre are unaltered then the total
output is also unchanged. The reduction in fibre passing over the couch
at the time the retention dropped, up to the final return to the previous
value, has provided the additional fibre in circulation in the backwater
system. The reduction in weight refers, of course, to dry weight and if the
moisture content of the paper were not controlled the effect on the total
substance of the paper may be even greater due to the moisture and dry-
weight following each other.
This exercise shows clearly that alteration in retention conditions on
the wire will have an effect on the dry weight until equilibrium conditions
are re-established. An instantaneous change in retention would require a
17
IA.22 THE WET-END FLOW SYSTEM
Fig. 1. Sa. Retention in system shown in Fig. 1.4 suddenly reduced to 33t %
---
Fig. 1 . Sb. Conditions after second pass round backwater circuit
~So 77.5
IS·S% S.1.1os
9 2so
Q•/°/o
Fig. 1. Sc. Conditions after third pass round backwater circuit
------ 2700
Q ..z:~:i~
t~4;;%
300 -
0·p% O•!>!>,%
Fig. 1. 5d. Final equilibrium conditions with new retention
length of time for correction which increased with the volume of the back-
water system and decreased with the flow of backwater used for mixing.
In practice retention changes would normally be fairly gradual so the
ultimate effect on the weight may, after all, be negligible. This depends
entirely on the relative time scales involved.
It is also clear from this example that the retention on the wire closely
governs the consistency of the backwater system. The system is self-
stabilising and accounts for the fact that the backwater and breast box
stock can alter appreciably in character while the same substance of paper
is being made; in fact, both backwater and breast box consistency adjust
themselves continually to the changing retention conditions on the wire.
This is one reason, incidentally, why attempts to relate breast box stock
18
properties, such as consistency, loading content or freeness, to the paper
quality are invariably unsuccessful. For instance, Bienkiewicz and Hendry,
et al. (2, 3, 4) reported tests showing that wetness of fresh stuff on a machine
making tissue was 25 deg. S.R. while in the breast box it was 65 deg. S.R.;
incidentally the wetness at the couch had returned to 25 deg. S.R. These
authors found large variations in breast box consistency and wetness tests
with no apparent alteration in fresh stuff or operating conditions and
emphasised that breast box tests could not be used for control purposes
though readings of this nature can be very useful for detecting and assessing
changes in the retention conditions on the wire. It is worth noting in this
respect that to a close approximation the consistency of the breast box
stock and the backwater keep closely in step, i.e. the difference between
the two consistencies remains approximately the same as the retention
alters; hence a measure of either consistency would, for day-to-day running,
be all that is necessary to assess retention changes on the wire. Retention
is affected in practice by numerous factors, and in particular by the breast
box consistency itself, so there will be some interaction between these two
variables; the factors affecting retention will receive fuller treatment in 18. I.
Other changes occurring in the backwater circuit will have a similar
transient effect on the dry weight. For instance it is common practice for the
reject flow from pressure-screens to be passed to a secondary screen from
where the accepted stock goes into the main wire pit; in a similar way the
accepted stock from the second stage of a battery of cyclones might also
be taken to the main wire pit. If for one reason or another either flow to the
main pit changes in character, this in turn could affect the backwater and
the weight of the paper until equilibrium conditions were re-established.
For example, a drop in consistency (with consequent rise in breast box
consistency) would have the same type of effect as an increase in retention
on the wire; the weight would rise temporarily but would finally return
to its original position when the backwater consistency had stabilised at its
new, slightly lower value.
It is important to note in the example just given the assumption that
there is no change in fibre content of the excess backwater. If this were not
the case, in addition to the transient weight change observed (which would
still be substantially the same) the final equilibrium conditions would also
be altered. Suppose, for example, that excess backwater consistency rose
to 0·15 per cent., then the weight of paper would be 86·25 lb. per minute
in final equilibrium conditions instead of the original 87·5, in other words
the substance is lower and there has been a permanent weight change.
To avoid permanent as opposed to transient weight changes it is desirable
to make the excess backwater consistency as low as possible. Hence the
superiority of arranging for excess backwater to be primarily from suction
boxes rather than a mixture of tray and suction box water from a single
silo, a point discussed further in lA. 2 4.
The most important change in the backwater system that occurs in
practice is the deliberate one when the machineman adjusts the backwater
flow to suit the conditions on the wire, in particular the position of the
dry-line. This will frequently be occasioned in the first place by a change
19
e>~Mfi:IT e,o.I( ~~00
~00 n
r~cSH. 5TUFf- - - - - O· 54%
n So. 9"·o
PA?!l:'t
~ '~H~ $01.10,
3·o~
l't..
t "\'
.:2
oA.Cl<.WA.T!~ ;;;:JOOO
~ooole,A.CKWA.TE:R~
0·27% E:xce.ss
__,.....
e>ACKWATE~
2So
0·3'7'o ool 0/o
Fig. 1. 6a. Flow of backwater in Fig. 1.4 suddenly increased to 3,000 gals. p.m.
So 9l·c::»
9-----...- 18·2% Soctos
___,,._ 2so
o•/%
Fig. 1. 6b. Conditions after second pass round backwater circuit.
Fig. 1. 6c. Conditions after third pass round backwater circuit
So 87·S
()---">- lc·S% Sc:u..1os
7--411- Z:,o
o·I%
Fig. 1. 6d. Final equilibrium conditions with new backwater flow
in retention conditions on the wire but ignoring this point for the moment,
the result simply of changing the volume of backwater in circulation is
worth studying in more detail.
lA. 2 3 Alteration in backwater flow

Using the flow system in Fig. 1.4 once again, suppose this time that the
backwater flow is altered from 2,400 to 3,000 gallons per minute. Fig. 1. 6a
shows the situation at the first pass, Figs. 1. 6b and 1. 6c at the second and
third passes respectively, and Fig. 1. 6d the final equilibrium conditions.
In this case the consistencies of the backwater and breast box stock have
decreased stepwise exponentially to their final equilibrium values which
are 0·25 per cent. instead of0·3 per cent. for the backwater and 0·5 per cent.
instead of 0·6 per cent. for the breast box stock. The weight of the paper
20
has risen initially to the equivalent of 96-0 solids, dropping with each
subsequent pass until the original weight is again reached.
These changes are seen to be the inverse to those occurring when reten-
tion decreased. It is interesting to note that decreased retention would
in practice lead to the dry line coming back on the wire (because the
drainage through the wire would invariably be faster) and this would
lead the machineman to add more backwater in the mixing box or pump.
Hence, provided the corrective action were performed quickly enough,
the tendency for the weight of the paper to reduce transiently when the
retention dropped would be offset by the machineman increasing the
backwater flow, thereby producing a compensatory transient increase in
weight.
It is worth drawing attention to one point of practical importance.
When no continuous measure of substance is available, alteration of the
backwater flow near the top of a roll should be avoided otherwise the
weight reading will be unrepresentative. Most machinemen are aware
that altering the backwater dilution flow affects the substance temporarily
but the intuitive idea of what happens is misleading. Since increasing the
flow of backwater means a reduction in the consistency of the breast box
this commonly leads to the notion that the substance of the paper will be
temporarily reduced; in fact, as seen above, the opposite effect occurs.
It is apparent from these examples that a change in the consistency of
the breast box stock and the backwater follows necessarily from a change
in retention or an alteration to the volume of backwater used for mixing.
Under equilibrium conditions these variables can easily be shown to be
related, and Figs. 1 . 7a and 1 . 7b depict characteristic curves of the
relationship for different retentions between the volume flow at the slice
and, respectively, the consistency in the breast box and in the backwater.
If the normal running condition of Fig. 1. 4 is at position A, a drop in
retention with the same backwater flow as in Figs. 1 . 5 would lead to a
new equilibrium at position B. Similarly a change in the volume of back-
water flow at the same retention, as in Figs. 1. 6, would lead to position C.
lA. 2 4 Stability of the paper machine flow system

In the two previous sections attention has been given to changes that can
occur in the substance of the paper when retention on the wire alters or
when the backwater mixing flow is altered, usually to take care of a change
in retention or drainage rate. These changes will evidently affect not only
the dry weight but also, because they involve a selective drainage particu-
larly of short fibres and loading, the characteristics of the finished sheet of
paper. To minimize these disturbances it is not sufficient simply to control
the flow and consistency of the fresh stuff, the drainage characteristics of the
fresh stuff must also be closely controlled. This is primarily a matter for
careful stock preparation, which is not the topic under present discussion,
but it is also dependent on the whitewater and the broke system since the
addition of both whitewater and broke earlier in the preparation plant has
an important influence on the fresh stuff drainage.
2 21
I
\
\
\
\
\
\
\
"' " ' ,...,,..12
::::::--=:==::- a
a
s
Zs%
So"o
R. "' 75%
.._~~~~~~~~~~~~~~'R xloo%
VOLUME: AT SL.lC&.
Fig. 1. 7a. Relation between breast-box consistency and volume of backwater flow for
different retentions (R) on the wire
R : 25,o
R. ..
R -a:
So'•
1sr.
VO\..UME AT S\..lC:.~.
Fig. 1. 7b. Relation between backwater consistency and volume of backwater flow for
different retentions (R) on the wire
It is not proposed here to enter into any lengthy discussion of the pros
and cons of different systems for dealing with whitewater and broke.
These depend to a large extent on the type of paper manufactured and on
whether the machine makes a variety of papers during a short space of
time. The author intends only to state three broad principles which seem
desirable if stability is to be achieved.
Firstly, it is important that the excess backwater flow entering the
whitewater system is as steady as possible. Any variation in flow rate
22
implies unsteady addition of fresh stuff or, possibly, whitewater or fresh
water, and.this should be corrected at source. Any variation in consistency
means that alterations to the dry weight of the paper of a permanent nature
are occurring (as opposed to the transient effects which were considered
above). Stability in consistency of the excess backwater flow is therefore
important and is usually achieved to some extent by ensuring that the
flow is composed of suction box and suction couch backwater since these
are lower in consistency and less variable than table roll backwater. This
can be arranged by allowing the main pit to collect all the backwater from
the table rolls and suction boxes but designing the pit so that the latter
overflows preferentially. The danger here is, of course, that the success
with which this is achieved is not readily assessed. A preferable alternative
is to control the level of the main pit by addition of the required quantity
of suction box water, probably with just a small overflow to remove scum
and froth that may be suspended on the surface.
The consistency of fresh stuff entering the backwater circuit is generally
of the order of 3 per cent. On the wire, drainage usually reduces the con-
sistency of the web to this value sometime after the first or second suction
box. Thus, although in most systems a small quantity of whitewater or
fresh water is also added to the backwater circuit in addition to the fresh
stuff, normally the backwater drained from the table rolls will not be
sufficient by itself to dilute the fresh stuff in the backwater circuit. The
consistency of backwater from the first suction box is generally higher
than in the other boxes (see, for example, Bennett (10)), especially the
'dry' ones nearer the couch; as the flow from the first box is equal to that
from the last few table rolls, and can be almost as much as from all the
other suction boxes put together, this flow alone should often be more
than sufficient to make up the deficiency in the main pit. In this event it
would be preferable to arrange the backwater make-up to be selected in
this way and so take advantage of the higher consistency of the first suction
box backwater; it would in any case seem worthwhile to ensure that the
main pit deficiency is made up preferentially by flow from this source.
A second requirement for achieving stability is that any fluctuations that
do occur in the backwater excess flow should have their influence in other
parts of the system minimized. It would be helpful in this respect if the
disturbances were confined, so far as possible, at least to the wire circuit.
On many machines excess flow composed of suction box backwater is used
directly on some wire sprays, on screening and cleaning equipment (par-
ticularly, cyclone installations and secondary tremor screens) or in the
breast box, and also for sealing the suction box and suction couch pumps.
In such applications as these, for any variation in the consistency of the
excess backwater flow it should not require long before new equilibrium
conditions are established in the wire circuit. On many machines with a
closed system, however, it is also the practice to use excess backwater
directly for dilution in consistency regulators, beaters, hydrapulpers, etc.,
in the stock preparation system. This use, though convenient, does carry
with it the danger of extending disturbances in the wire section throughout
the whole preparation system and will considerably increase the time
23
needed to establish a new equilibrium. It is therefore preferable to take all

the excess backwater flow not used directly at the wet-end of the machine
straight to some form of save-all, clarifier or thickener. The thick stock
from this could then be treated as broke and returned to the system at a
suitable point (preferably at a steady rate to reduce the possibility of ·
affecting retention on the wire, since it will be composed largely of fine
fibres and loading); the clarified whitewater, at a very low consistency,
should then hardly alter in character or consistency and using it for dilution
would cause little or no disturbance to be transmitted to the stock prepara-
tion system. In some modern systems this is elaborated by having two
save-alls or clarifiers in series; whitewater from the first simple clarifier is
used for dilution in the preparation and broke system, thinner whitewater
from the second clarifier is used for wire showers and suction roll lubri-
cating water. Alternatively a disc filter provides cloudy water for the first
purpose and clear for the second. When clear water is used in wire sprays
and collected in the machine pit, the tray water from under the wire going
to a separate silo, then this spray water should return direct to the disc
filter to avoid disturbances in the backwater excess flow.
It may be noted in passing that in completely closed systems, since the
water in the web at the couch (70 per cent. to 90 per cent. moisture content)
is always more than the water added in the furnish, a quantity of water is
required for make-up; in practice it is generally the case that more than
this quantity is added in fresh-water sprays at the wire and couch, pump
seals, cutting jets, etc., producing some excess. Part of this may be cleaner
reject flow and the remainder would be taken to waste from the clarified
whitewater chest, which should have sufficient capacity to act as an adequate
reservoir for sudden demands on the machine.
The third principle desirable for stability is to keep the excess backwater
and the broke system completely separate; wet broke in the form of trim
off the wire, or the whole width of web from the wire or presses during a
break, should be diluted with clarified whitewater at the lowest consistency
available. The reason for this separation is twofold; firstly, it is not desir-
able to allow any broke to mix with the whitewater system since variation
in the quantity of broke added, whether before or after clarification,
will affect equilibrium conditions in the stock preparation and eventually
in the wire circuit; secondly, if whitewater containing little fines or loading
is used for dilution of wet broke then the character of the resulting mixture
will be as close as possible to that of the fresh stuff and when added in the
preparation system at any point (and it is becoming increasingly common
for this to be the machine chest) it should have an indistinguishable effect
on drainage characteristics compared to the fresh stuff flow to the machine.
Even when, as is obviously desirable, the wet broke is consistency controlled
and carefully metered from a storage chest, occasions are bound to arise
when there is a considerable excess of broke and it becomes necessary
to increase the flow; at such times this use of clarified whitewater for dilu-
tion will be most beneficial in reducing the possibility of a vicious circle
of broke and breaks arising on the machine.
One difficulty of realizing this in practice is that usually it becomes
24
necessary to have a separate broke thickening system and whitewater
tank. This is because it is not easy to arrange a compromise between
having sufficient force of whitewater available during a break at the couch
or presses for adequately breaking up the wet web, and at the same time
preventing excessive dilution below the 3 per cent. or so suitable if the
wet broke is to be returned direct to the preparation system. With machines
having an open draw the necessity for a shower on all the time to wash off
the sheet at a break makes it impossible to prevent excessive dilution of
the broke. However, on pick-up machines powerful high pressure deluge
sprays which only come into operation across the width of the machine
during a break are meeting with some success in this respect though an
added difficulty can be encountered in pumping broke at this relatively
high consistency for any distance. With the machine working normally
the small amount of trim can be washed off the wire edges and fed to a
separate small thickener.
Dry broke creates a rather different problem. Due to the fact that it
needs some refining or defibrating before it is suitable for return to the
system this necessarily alters its character somewhat from wet broke; it
would appear preferable for dry broke to be stored, consistency regulated
and metered into the preparation system separately. The manner in which
this is accomplished need not differ in principle from that described for
the wet broke system. For certain types of paper, particularly wet-
strengthened and coated, such a procedure is imperative anyway.
With growing recognition of the difficulties that can occur at the wet-
end of a paper machine due to unsuspected disturbances in the whitewater
and broke systems, it seems increasingly likely that something along the
line of the principles detailed above will eventually come to be regarded
as necessary. Certainly machines laid down in recent years have, in different
ways, shown much more care and forethought in the arrangement of their
systems than is apparent on many older machines.
1A. 2 5 Degree of closure of whitewater systems

So far all that has been said applies primarily to completely closed back-
water and whitewater systems. On many machines it is common practice
deliberately to add fresh water somewhere in the system. The economic
advantages of reducing fibre loss have long since closed up the backwater
fresh stuff dilution circuit completely except for a very few specialized
machines where some fresh water addition in the main backwater pit is
still necessary for a specific purpose. The degree of closure of the back-
water and whitewater system is usually defined as the difference from
100 per cent. of the percentage ratio between the whitewater flow to drain
and the flow through the slice; where all excess flow from the suction
boxes goes to drain the degree of closure would be low by modern terms,
between 60 per cent. and 80 per cent. Sometimes the excess suction box
backwater is taken to a clarifier and only the thickened stock retained;
this reduces fibre loss but may still be regarded as a relatively open white-
\Vater system since fresh water make-up will be necessary. Such a system
25
will generally be used on a machine requiring frequent colour changes

where a more comprehensive whitewater system with dilution in stock
preparation would never have a chance to reach equilibrium and in the
short duration of a normal run would succeed only in disturbing the colour
balance. A system slightly more closed will involve use of clarified white-
water for sprays on the machine and, perhaps, in a consistency regulator
close to the machine. A completely 100 per cent. closed system with no
whitewater overflow to drain can only be found where fresh water usage
in the system is sufficiently low for some make-up to be needed; this is more
likely when using a dry furnish (e.g. 10 per cent. moist bales) or in a hot
system when there will be more evaporation of water (though the quantity
of water involved in this is not likely to be of much practical significance
compared to other losses). For rag and esparto mills and integrated wood
pulp and paper mills the paper machine whitewater system can not be
considered separate from the pulp preparation system; in the case of the
integrated type of mill conservation of fibre and heat makes it most
important that the high whitewater surplus on the paper machine is made
available in grinders and other parts of the pulp preparation system.
The effect of closing up a backwater and whitewater system has been
studied at length by Bienkiewicz, Mardon et al. (2, 3, 5). The advantages
in economy of fibre and loading and lower heat losses are apparent but
closure leads to a progressive buildup in the amount of fibre debris and
loading particles in the system. Also the reduced flow to drain, though
beneficial for mills with an effluent problem, leads to a greater proportion
of fine fibres in the paper once equilibrium conditions are established
(more of the fine fibres in the furnish are retained in the sheet) and this
alters the characteristics of the paper. In some cases this can be very
undesirable; for example it is reported that with glassine opacity is increased.
Closing the system has other results which can be very detrimental to
efficient running of the machine. The quantity of dispersed pitch and slime
in the system grows and foam can become increasingly troublesome. With
constant addition of alum, adsorption of alumina on fibres leaves more
free sulphate ions circulating in the backwater and the sulphuric acid in
the system steadily rises. The pH falls due to this and less alum would be
used to keep it steady; this makes the system increasingly unstable to any
variation in fresh water addition and changes the chemical conditions
affecting retention of loading, dyes, size and wet-strengthening resins.
These factors make it necessary in many systems to have a sizable fresh
water addition and this should always be as far from the wire backwater
circuit as possible to reduce disturbing effects; certainly in normal cases
fresh water should never be added in the backwater pit and usually the
whitewater tank is the obvious place. Build-up of fine fibres and loading
in the backwater as the system is closed will also reduce retention on the
wire leading to an increased breast box consistency. Occasionally this
may be so excessive (e.g. with groundwood furnishes) as to impair the
ability of the machineman to form the sheet with the quantity of water
he is able to drain from the wire. In this event it could become necessary
to open up the system.
26
Another factor which is closely related to the degree of closure of the
backwater. and whitewater system is the time required for equilibrium to
be reached throughout the system. Hendry (4) has reported that the
backwater loading consistency of a tissue machine with a highly closed
system required five or six days to reach equilibrium, and when making
coloured paper the hue of the backwater continually changed due to the
fine fibres which accumulated being preferentially dyed by one or other
constituent. If backwater or breast box consistencies are tested at regular
intervals from start-up it will frequently be found, particularly with
machines making papers carrying a high proportion of fines and loading,
that the consistencies steadily rise throughout a period of days. This is
directly due to the reduction in retention occurring with the increasing
proportion of fines and loading in the backwater circuit.
Calculations by the authors quoted above showed that equilibrium
conditions from start-up are established in an exponential manner (similar
to the examples illustrating the effect of changing retention and backwater
flow given earlier). If a proportion of fresh water is used in place of back-
water for dilution in the mixing pump or box then the final equilibrium
conditions are altered and, as expected, the greater the proportion of
fresh water added the lower the equilibrium consistencies in the backwater
circuit. Addition of a small quantity of fresh water in this way to a com-
pletely closed system has a greater effect on the equilibrium consistencies
than subsequent additions. At start-up, the backwater consistency will
rise to within approximately 5 per cent. of final equilibrium after three
passes and within 1 per cent. after five passes round the circuit (allowing
for the fact that the retention will be lower at first when the breast box
consistency is lower); thus if the capacity of the main pit, piping, cleaners
and breast box is, for example, equivalent to four minutes of flow through
the slice, then the substance of the paper should be within 5 per cent. about
12 minutes after the fresh stuff valve is open and within 1 per cent. after
about 20 minutes. Subsequent changes in the backwater circuit and
establishment of final equilibrium will depend, of course, on the design
of the whitewater system and on many machines true equilibrium may
never be reached within normal working runs.
lA. 2 6 Formation and wire conditions

Though only indirectly connected with the wet-end flow system one topic
is conveniently considered in the present section under the heading of
theory-that of the relationship between the slice jet velocity and opening
and the wire speed. The formation of the sheet of paper on the wire is
closely determined by the relative speed of the wire and the jet from the
slice at the point of impingement. This relationship (usually termed the
efflux ratio) will be discussed in greater detail elsewhere and for the
moment it is only necessary to stress that under given conditions for each
wire velocity there must be, within fairly close limits, a slice jet velocity
which is most suitable to get the best formation. Normally these two
velocities are very close though for some types of machine with an upsloping
27
wire and for certain papers requiring a special fibre orientation (e.g.
twisting paper) the velocities can be quite different.
The jet velocity is directly controlled by the pressure head behind the
slice and this in turn is governed by the level in the breast box (due account
being taken of air pressure difference in air-loaded boxes). If the coefficient
of discharge of the slice is known it is possible, as shown in lA. 1 5, to
calculate the head appropriate for any slice jet velocity; however, any
loss in head due to friction at the slice walls and also across breast box
fittings in front of the slice, such as perforated rolls or evener plates, will
affect the calculation. Pressure loss across a perforated roll is smaller than
across evener plates, Attwood and Alderson (49) found it to be approxi-
mately 0-4 in. compared to 2 in. for a 45 in. head, but the loss will alter
with position in the breast box and also with the velocity of flow. In
theory, then, the head at the slice can be related to the wire speed by a
general formula which takes the form
h v (1 - p/100)
3 [ 60C
= 16 ]2 + oq
where h is the head in inches, V the wire velocity in ft. per min., C the
coefficient of discharge, p = the optimum percentage by which the jet
speed should be slower than the wire speed for the best formation, and
oq is equivalent head loss in inches across breast box fittings in the slice.
In practice, neither C nor oq are known to any degree of accuracy (according
to Beveridge and Bridge (80) C is not even constant on any particular
machine but varies significantly with changes in stock consistency and
other characteristics). Alsop can only in the first instance be guessed at,
so it is really only feasible to use this relationship as a guide. For that
purpose when C is close to unity, as in a projection slice, the simplified
formula usually quoted in one form or another, h = V2 /19,200, is probably
adequate. The head appropriate to different wire speeds (or speeds measured
elsewhere along the machine if allowance is made for the normal difference
due to draw), as calculated from either of these formulae, may be presented
in a suitable form to the machineman who then has a useful basis from
which to learn, by experience and by experiment, the difference between
the calculated and running heads which usually gives the best results
under any particular conditions. Alternatively, the calculated jet velocity
for the head measured at the slice can be compared with the actual velocity
of the wire. This has the advantage of providing a linear scale, whereas
differences between heads increase with the square of the velocity; thus
it is then easier to compare the normal percentage velocity difference or
the efflux ratio at different speeds.
While the wire speed determines the pressure head required behind the
slice, the quantity of backwater used for diluting the fresh stuff determines
the width of opening irr the slice. This is evident because if the velocity
of the slice jet is fixed the volume can only be varied by adjusting the slice
opening. Thus, strictly speaking, whenever the machineman requires an
alteration in the volume of backwater in circulation in the wire circuit
this should be accomplished by moving the slice up or down. If the level
28
in the breast box is not automatically controlled then a further adjustment
becomes necessary to bring the level back to its original value, thereby
leaving the pressure head at the slice unaltered (at least within practical
limits). In practice it is common for alteration of the slice lip to be avoided
as much as possible because of the difficulties which may be encountered
with cross-level changes (see lB. 6 3); consequently the quantity of back-
water in circulation is only altered effectively by changing the level in the
breast box, i.e., the velocity of the jet. While such a procedure may be a
reasonable compromise for very small changes it is very important that
the machineman realizes that by this action he is not only altering the
amount of 'water on the wire,' he is also affecting the formation to some
degree. The obvious danger is that the two effects become confused and
the formation alters to such an extent that it affects the drainage charac-
teristics which the machineman is trying to correct. A vicious circle may
well be set up until the whole wet-end is in a completely chaotic state.
Sometimes the only way of avoiding this is to put limits to the extent by
which the machineman is permitted to adjust the slice pressure head so
that under unusual conditions he reaches a point when the slice opening
has to be altered and any adverse cross-level effects from this have to be
faced up to and corrected. Occasionally drainage on the wire is so limited,
and the slice flow so low, that the slice opening would be too narrow for
a satisfactory sheet to be made without every small irregularity in the slice
lips appearing in the paper; under such conditions formation would be
sacrificed to obtain a satisfactory cross-level by opening up the slice,
allowing the head to drop. It would then be especially valuable to set a
limit on the difference tolerated between the required and running head
or velocity to prevent formation and other paper qualities deteriorating
too much.
When making an alteration in machine speed to suit general running
conditions the machineman will make a corresponding change in the
fresh stuff to keep the substance the same and will also alter the breast box
level (and hence the flow through the slice) to keep the correct relation
between the slice jet velocity and the new wire speed. Provided the slice
opening is not altered then, for all practical significance, the effect of these
alterations is simply to reduce all the flow velocities round the backwater
circuit proportionately. The consistency of the breast box and backwater
is unaltered, which is precisely the condition required.
When altering the machine speed to effect a relatively small substance
change for the same grade of paper the position is different. In this case
a familiar situation is one in which the speed is altered by an amount
which is roughly inversely proportional to the substance change; the gross
production at the reel-up is then unchanged, as is the quantity of water
drained at the wire and presses, and evaporated in the drying section.
By this means the machine is kept running at what is presumably an
optimum speed to suit the drainage and steam capacity available. For this
type of change when the fresh stuff flow remains effectively untouched
it is necessary, in order to keep the same flows and consistencies in the
backwater circuit at the new speed, that in addition to altering the head
29
IA.2 6 THE WET-END FLOW SYSTEM
at the slice, the slice opening must also be adjusted proportionately to the
velocity change. In fact if the main flow valve feeding the breast box is
left untouched while the slice is opened sufficiently to obtain the new head
required, then the desired condition is reached.
For larger changes of substance other factors, in particular altered reten-
tion characteristics on the wire, affect the validity of the foregoing. These
aspects are intimately tied to the particular machine and paper concerned
Breast box flow of 2430 reduced from 2700 by 10 % drop in slice jet velocity.
Fig. 1. Sa. Conditions in Fig. 1. 4 altered to take care of a machine speed reduction
of 10 % with same substance paper, i.e. solids flow at couch also reduced by 10 %. Slice
opening unchanged
Breast box flow of 2700 unchanged: 10 % drop in slice jet velocity balanced by approxi-
mately 10 % increase in slice opening.
9--17.5% s ......
So 87·5
--- 2So
O•lfo
Fig. 1. Sb. Conditions altered to take care of substance increase of 10 % and corre-
sponding speed decrease of 10 %, i.e. solids flow at couch unchanged. Slice opening
increased by 10 %
and are not discussed further. For clarification of the two points raised
in the preceding paragraphs, the wet-end conditions used for illustration
in Fig. 1. 4 are shown in Figs. 1. Sa and 1. 8b as they would be at 10 per
cent. reduced speed, without and with an accompanying substance change.
30
CHAPTER 18
OPERATING FACTORS AFFECTING PERFORMANCE
OF THE WET-END FLOW SYSTEM
1B .1 RETENTION ON THE WIRE

The importance of maintaining steady retention conditions on the wire
has been considered in some detail from a theoretical standpoint. It is
appropriate now to set down the various factors which affect retention
in order to assess what parts of the wet-end it is desirable to control.
Some of these factors affect not only retention but also the general drainage
conditions on the wire, in fact of course the two are closely connected.
Drainage and its effect on formation will form a later subject for discussion
but it is emphasized that factors which affect retention can disturb not
only the substance and fines retained in the paper but also the whole
formation, and the effect on the latter can be quite as important for
maintaining satisfactory paper quality.
18. 1 1 Composition of the furnish and the effect of beating

The constituents of the furnish and the treatment given to it in the prepara-
tion plant set the whole character of the wet-end flow system. Retention
on the wire is closely affected by the degree of beating or refining since
development of fibrillation and the increased proportion of fibre debris
changes both the composition of the stock and the matting conditions
on the wire. Retention will also be affected by the proportions of short
to longer fibre pulps mixed in the furnish and by the addition particularly
of loadings and dry broke. It is not possible to make any useful generali-
zations about the effect of composition of the furnish on retention, so
much depends on the particular circumstances. In all mills, however, an
attempt is made to keep the composition of the furnish steady so that
the quality of the paper remains uniform, and this will also necessarily
help to keep retention conditions steady.
It is almost universal practice to use the results of freeness or wetness
tests to assess the degree of beating or refining given to individual pulps.
In fact, an appreciable amount of research is directed at the present time
towards developing a satisfactory continuous freeness instrument with
which it is hoped to control refiners and in some cases this is already being
attempted (57, 63, 64, 75, 93). This procedure seems to the author to be
satisfactory only in so far as the freeness test, duly corrected for consistency
and temperature variation, is an adequate measure of the beaten condition
of any particular pulp and correlates with development of the final proper-
ties of the paper made from it. Originally the test was designed to predict
how a furnish would drain on the wire and subsequent work showed that
31
lB.11 THE WET-END FLOW SYSTEM
in the case of individual types of pulp the results obtained from the test
are closely related to the duration of beating. But in many circumstances,
especially where large quantities of fines are involved, the test can give
spurious results. The main concern when using any beating equipment
must be to control the manner and degree of beating to that found to be
most suitable for the paper to be made, and in some circumstances other
measurements, for example refiner load or temperature rise, may well be
better than freeness for this purpose.
While the use of the freeness test for control in beating of individual
pulps may prove very useful in ensuring uniformity of treatment, freeness
results for the final furnish going to the machine must be used with far
greater care. They seem in this case to have two purposes: to act as a control
on the total beaten condition of the furnish, as for individual pulps, and
to predict and keep steady the drainage condition on the wire. If their
use is primarily to control the beaten condition of the fresh stuff then the
same restrictions apply as for the practice of taking the tests on individual
pulps; in this case, however, because the proportions of different pulps,
dry and wet broke, whitewater fines, etc. will alter from time to time, it is
much less likely that freeness figures on the final furnish will bear any
but a very rough relation to development of the paper properties and so
their use to control beating of one or other constituents of the furnish
earlier in the stock preparation system must be carefully justified. It is
interesting in this respect that Mardon et al. (5) quote that the power used
in refining a tissue stock was reduced from 670 to 460 k.w.h. per ton within
24 hours from start-up to keep the wetness constant at 480 deg. S.R.; this
illustrates well the effect of fines building up in the whitewater and altering
the apparent wetness. Also for the same machine Hendry et al. (4) found
that addition of backwater fines to refined stock had the effect of making
it appear slower, from 23 deg. S.R. to 33 deg. S.R. for 10 per cent. addition
and to 50 deg. S.R. for 20 per cent. addition by weight. The effect also
depended to a slight degree on whether the backwater originated from a
tray collecting the breast roll and the discharge from some dandy-type
table rolls or from a second tray draining some solid table rolls, which
shows how easily the freeness test on a composite furnish can be influenced.
On the other hand, if the use of freeness tests on the machine furnish
is primarily to give a guide to forthcoming drainage conditions on the
wire, then their value is also not very likely to be great. This is due to
several well-known limitations of the freeness and wetness testers as a
means of simulating drainage (consistency, suction and other factors differ
from those on the wire), and also because drainage on the wire is dependent
as much on the backwater condition as on fresh stuff (see lB. 1 2). Thus,
freeness tests on the machine furnish can usually at best serve only to
indicate the uniformity with which the whole system is functioning and a
rise or fall in test figures could be caused by a variety of factors other
than beating and refining. The danger of a continuous freeness tester on
machine furnish or in measuring and reporting freeness figures on a routine
basis is that the results come to be used as a criterion for the degree of
treatment the furnish has received and any difficulties at the wet-end of
32
the machine may too readily be traced to a variation in the freeness
figures; in time the refillers may come to be run more to suit the wire
conditions of the machine than the properties of the paper manufactured
while the compromise needed should, of course, be much more the other
way.
A further use of the freeness test has been reported by Tousignant and
Madgett (79), who installed a continuous freeness tester on the breast
box stock of a newsprint machine as a means of assessing stability of the
system. Variations in the freeness record obtained in this way were found
to have a connection with draw at the couch (presumably due mainly to
changes in moisture content of the sheet at the couch resulting from
different rates of drainage on the wire). General fluctuations in the freeness
record were compared with various records obtained in the stock prepara-
tion and blending process in an endeavour to detect the sources of disturb-
ance, and alteration to the whitewater dilution consistency, broke addition
flow rate, proportion of sulphite pulp to groundwood in the furnish, and
other aspects of the system were made in order to try to reduce variations
in the freeness record. The main problem in using this sort of approach is
that, although fluctuations in drainage rate on the wire originate at least
in part due to differences in the composition of the machine furnish, these
are likely to be detected much easier by direct assessment of the furnish
itself, rather than by taking the flow box stock which is dependent also
on various conditions on the wire and is in any case part of a self-stabilizing
system. Further, for the intention of providing a means of assessing
stability, a more satisfactory (and simpler) approach would be to measure
flow from one or more table rolls; this has the additional merit of giving a
direct measure of drainage rate variability on the wire, rather than relying
on a complicated simulation device.
Returning to discussion of the effects of the furnish on retention condi-
tions on the wire, it is evident that the greater the proportion of fibre
debris and loading in the furnish the lower will be the retention. Fines and
loading, particularly the latter, predominate in the backwater and can
virtually control the drainage conditions on the wire. For instance the
following results are typical for one newsprint machine (fibre fractionation
figures obtained in a Bauer Mc.Nett Classifier and corrected for ash
content):-
Ash on Long fibres Short fibres
Consistency solids >48 <100
% % mesh mesh
Fresh stuff 3·0 8 43 42
Breast box stock .. 0·6 22 24 67
Backwater (main pit) 0·3 40 trace 97
Whitewater (suction box) 0· 11 57 trace 97
It is important to note particularly the high loading content (china clay)

in the backwater and even higher percentage in the whitewater. The reten-
tion of loading on the wire is under 10 per cent. whereas the overall reten-
tion is 55 per cent.; even fine fibres less than 200 mesh have a retention of
about 40 per cent. Similar very low retentions for loadings (11 per cent.-13
33
1B.1 2 THE WET-END FLOW SYSTEM
per cent. for diatomaceous earth, 4 per cent. to 6 per cent. for talc and
china clay) have been reported by Hendry (4), while results given by
Bennett (10) for a newsprint machine also show a clay retention of only
about 10 per cent. Under these conditions it is apparent that only a slight
alteration in the drainage conditions could have a large effect on the
loading circulating in the backwater which, as with consistency variations,
would have a substantial (though transient) effect on the quantity of
loading retained in the paper until new equilibrium conditions were
established.
lB . 1 2 Drainage conditions on the wire

The retention conditions on the wire, apart from being governed by the
composition of the machine furnish as discussed above, are affected by
other factors which alter the character of drainage on the wire. The most
important under this heading are the consistency of the slice stock, the
substance of paper (i.e. thickness of the fibre mat on the wire), machine
speed, temperature, wire mesh, and aeration of the stock. Each of these
will be discussed briefly.
The interaction between consistency of stock in the breast box and
retention on the wire has already been mentioned in some detail; apart
from the consistency being essentially governed by the retention it is
generally considered that, other things being equal, a lower breast box
consistency will reduce retention. It is difficult, however, to isolate effects
on retention caused solely by the breast box consistency because other
conditions, in particular the speed of the machine and substance of the
paper, are likely to be altered at the same time. In practice the normal
running variations in consistency probably have only a small effect on
retention; Bennett (10) reported that in a series of trials on a newsprint
machine breast box consistency did not have a significant effect on general
drainage conditions, though Hendry (4) on a tissue machine obtained a
significance relation such that retention decreased with reduced consistency.
These workers are in agreement, however, as to the effect of wire speed,
an increase of which reduces retention; this has also been reported for
retention of loading by Hansen (1), working on an experimental machine,
though the determination of the results and manner in which the back-
water system was operated in this work are not clear from the article.
Hansen also reported an increased loading retention at a higher substance
and this was confirmed for overall retention conditions by Hendry.
The direct effect of temperature of the breast box stock on retention
was not found to be significant in the work reported by Bennett though
it is likely that the increased drainage rate at a higher temperature would
normally cause the breast box consistency to be reduced and this could
lead to a lower retention. A finer wire mesh obviously improves retention
and normally on any machine the mesh used is as small as possible com-
patible with adequate drainage being possible and a reasonable wire
life. In a similar way a wire becoming made up will alter retention con-
ditions. Aeration of the breast box stock reduces the rate of drainage
34
THE WET-END FLOW SYSTEM 1B.2
on the wire and probably affects retention for this reason. However,
aeration is considerably more important for its effect on other conditions
at the wet-end and will be dealt with separately.
IB. 1 3 Chemical conditions

It is becoming increasingly common for chemical additives to be put in
at the wet-end of the paper machine rather than in the preparation system,
Liquid alum, loading, dyes and wet-strength additives may all be added
in the backwater pit or the mixing pump and this procedure has the merit
of permitting a closer and quicker control of their effect on the finished
paper. Rosin sizing may also be added in this way though it is more usual
to add it earlier in the stock preparation system to allow sufficient time
and achieve better conditions for precipitation to be more closely con-
trolled.
Retention can be affected by the chemical condition of the backwater;
for example, Bennett (10) reported that a lower pH increased retention,
while Hendry (4) found that urea-formaldehyde wet-strength resin addition
increased retention. Some chemicals such as Sveen glue and activated
silica sol are used expressly to improve retention of fibre and loading,
though Hansen (1) considered the effect of Sveen glue was confined
almost entirely to the loading, and calcium chloride added after refining
has also been reported to have this effect (71).
Generally it is more important to control the chemical condition of the
backwater, particularly the pH, for reasons other than that of retention.
A low pH accentuates the likelihood of corrosion, particularly with an
aerated stock; this is because air bubbles in the stock provide carbon
dioxide and oxygen which combine with hydrogen ions collected on the
metal surface as a result of dissociation of water-removal of these ions
allows further hydrogen ions from the metal to form causing further
corrosion. Apart from damage to metal, corrosion products may be
transferred into the stock solution producing discoloration and black
specks in the paper. The pH is also thought to have an effect on floccula-
tion tendencies though there is some doubt about this; on the other hand
some deflocculating agents apparently do not function at all with alum
present and their addition may even worsen flocculation.
lB.2 FRESH STUFF FLOW AND MIXING

Flow of fresh fibre to the machine must be as steady as possible and this
implies constant consistency and constant flow. By suitable choice of
pipe size the effect on the total fibre flow of variations in consistency can
be minimized (see IA. 1 4) but this can only be regarded as a palliative.
Both consistency and flow of fresh stuff require careful control. The
manner in which this is achieved depends on the machine set-up and broadly
three methods can be distinguised: the old-fashioned mixing box with or
without separate consistency regulator, the combined mixing box and
35
lB.2 1 THE WET-END FLOW SYSTEM
consistency regulator, and the more modern mixing pump with or without
stuff box and a separate regulator.
lB. 2 1 Consistency regulation

It is not proposed to enter into any detail about the problems and limita-
tions of present methods of consistency regulation. Suffice it to stress
that none of the popular instruments available at present regulates con-
sistency: in one way or another all types depend on measuring a form of
viscosity of the pulp and any factor which affects viscosity, the proportion
of fines, freeness or temperature, affects the working of the regulator even
when the true consistency is unaltered. One instrument which actually
does measure consistency is of the balanced U-tube type often used for
specific gravity measurement of slurries, but the difference in density of
fibre and water is so small that the sensitivity of this device has to be very
great for it to be useful and problems are likely to occur, particularly with
aeration.
Despite the limitations of consistency regulators there has been a
tendency in recent years to use more of them in series through the stock
preparation plant, on the theory that each will help to smooth out further
the irregularities left by the previous regulator. But there is a distinct
danger in this procedure, particularly if a regulator is used finally on the
mixed stuff feeding the paper machine; the uniformity of consistency from
earlier regulators could, in fact, be worsened because of the distorting
effects on viscosity caused by differing proportions of broke, loading, etc.
In addition, variation in the whitewater consistency and composition
used for dilution could have a similar effect on the regulator, and this
incidentally is another reason, apart from the question of keeping stable
retention conditions on the wire, why it is preferable to use clarified
whitewater in the stock preparation system. Consistency regulators can
also be be bedevilled by poor mixing and too long a time-lag between
the point of dilution and the measurement; generally the best system
is one in which the dilution takes place at the inlet side of a pump and the
sampling as close as practical to the discharge side. The dilution water
pressure upstream of the control valve should be constant to avoid offset
in the control point.
It is very important that a check is kept on the performance of a regulator
by taking samples at intervals for accurate laboratory determination of
consistency. The frequency with which this is done is a matter for individual
decision but it is useful to take sufficient to enable a mean figure and a
measure of the spread of individual readings to be calculated each week
or, on a machine with frequent changes, at least for a typical making
which recurs frequently. In this way trends in the average, and any undue
variation associated with clogging viscosity tubes, mechanical friction,
etc., are readily detected and the instrument can then be rigorously checked
when appropriate. In some mills consistency determinations are taken on a
routine basis and, for example, Moore and Walsh (66) reported that they
use hourly tests to adjust the set point of a De Zurik regulator. Provided
36
THE WET-END FLOW SYSTEM lB.22
the known accuracy of the test figures is used to set limits within which
the control plant is not adjusted (to avoid making alterations for test
differences which are not significant), this seems to be a worthwhile though
perhaps rather tedious approach; in this particular case, the authors
used the consistency regulator signal output to adjust the control point
of a magnetic flow meter in the fresh stuff line in an attempt to keep a
constant fibre flow to the machine.
lB . 2 2 Fresh stuff flow control and mixing
In the older type of mixing box and also in the type incorporating a
consistency regulator, the basic control of the flow of fresh stuff and back-
water is usually through gate valves; the pressure behind the gate valves
is kept constant by having overflows from levels which may also be
adjustable. This design, even when improved by using an orifice instead
of a gate valve to introduce the stuff, suffers from the relative coarseness
of the adjustments and the fact that the volume flow is dependent on the
head behind the valve (which will vary despite the overflow), and on the
consistency and other flow characteristics including viscosity. Further,
mixing is often achieved by cascading fresh stuff and backwater through
valves sited opposite one another in a compartment and this must induce
considerable aeration. With the type of mixing box which includes a
consistency regulator it is sometimes the practice to adjust the set-point
of the regulator to produce small changes in substance; while certainly
practicable it is difficult to determine whether this works better than altering
the volume flow, though the method may have advantages when the fresh
stuff control valve is crude.
On faster machines it is usual to use a pump for mixing, and fresh stuff
flows to the pump from a stuff box giving adequate head. Though it is
not always possible, the inlet to the suction side of the pump should be
drowned and trumpet shaped to give a smooth flow entry for the backwater.
On some machines where the mixing pump is above the backwater pit,
a non-return valve may be used at the bottom of the inlet pipe to allow
the pump to be primed; this arrangement is not always satisfactory
because unless regularly cleaned the valve is a frequent source of trouble
from sticking and slime growth, and with age gets stiffer and provides a
strong resistance to the flow thereby affecting the suction head. A suction
line diverted from the suction box pump can be used to prime the mixing
pump very quickly and is generally a more satisfactory solution.
The flow control valve in the stuff line must be capable of sensitive
adjustment and located well below the stuff box to prevent the development
of pressures below atmospheric on the downstream side. Most stuff valves
are of the plate type and are not sufficiently sensitive for making small
changes in substance; there is some merit in having a small by-pass flow for
this purpose though this often introduces some practical problems. The
Egger diaphragm regulating-valve which keeps a circular opening for all
positions rather like the iris of a camera has recently been claimed to have
some advantages, particularly in giving a consistent performance and in
keeping clean (43). There should be a separate cut-off gate valve to allow
37
the fresh stuff flow to be stopped without altering the position of the
control valve.
Opinion differs as to whether the fresh stuff should be introduced
directly to the mixing pump inlet by means of a suitably placed pipe entry
or to the main pit close to the point where backwater is drawn into the
mixing pump suction. The main advantages of the latter arrangement are
that priming the pump is easier and there is less possibility of air inclusion
when the stuff is turned on. In addition there is probably less risk of poor
mixing and stratification with entry to the pit than may occur in a pipe
joining in a tee-branch to the suction side of the pump; also if the mixing
pump is suddenly stopped the stuff passing down the stuff box before the
valve is closed is less likely to cause trouble if it flows to the pit and cannot
lodge in the pump suction line.
To keep the flow of stuff steady it is important that the pressure difference
between the stuff box and the point of entry to the backwater is constant.
With a pipe joining to the suction side of the pump changes in the pump
speed or discharge control valve position (for drawing more or less back-
water for dilution) will alter slightly the suction at the bottom of the stuff
pipe. On some machines speeding up the mixing pump can be seen to
produce a drop in the overflow level of the stuff box for this reason. If the
stuff pipe enters the main pit close to the suction pipe entry at a point
where the flow velocity is reasonably low, and if the level of the pit is
adequately controlled, this change in suction will be avoided.
Recently there has been a tendency to abandon the stuff box as a means
of feeding fresh stuff to the machine in favour of using a magnetic flow
meter to control the flow in a pipe taken directly from the machine chest.
The magnetic flow meter, though difficult to calibrate for such large
volumes, is more accurate and reliable than a simple orifice or venturi
meter, especially at fresh stuff consistencies, and some success has been
reported with this instrument (20). The main advantages claimed are
avoidance of aeration in the stuff box and elimination of overflow, which
reduces power consumption and removes a potential slime collecting line.
Also a visual record of the flow and a relatively accurate means of setting
the substance is provided. The control point for the substance can, in
fact, be cascade-controlled by a signal from a consistency regulator,
thereby helping to keep the fibre flow steady irrespective of changes in
consistency. The success of the method must depend entirely on the
accuracy of the magnetic flow meter and there is not as yet much data
available to show how the meter is influenced by aeration, temperature
changes and other variables, nor have there been any reports comparing
the long-term uniformity of substance at the reel-up when using the meter in
place of more conventional methods.
lB.3 BACKWATER AND WHITEWATER SYSTEMS

There are almost as many methods of arranging the backwater circuit and
whitewater flows as there are machines. General principles have been
discussed earlier and only a few special points will be made here.
38
THE WET-END FLOW SYSTEM lB.3 2
18. 3 1 Backwater pit and the flow control valve
Backwater discharging from the table rolls is usually collected in trays and
lead to a main pit situated either directly under the wire part or at the
side of the machine. Occasionally the backwater may be allowed simply
to drop through the return side of the wire into a pit. The primary consider-
ation here is to prevent excessive aeration in collecting the backwater and
in this respect the use of trays and carefully designed chutes is really
imperative. In the pit itself it is desirable to arrange compartments to
channel the flow in a suitable manner. It is difficult to state any general
principles but the main aim must be to prevent any dead spots and ensure
a smooth and sufficient flow rate to obviate any tendency to settle (1 ·0 to
1·5 feet per second is one estimate of the velocity required for this).
Important also in this respect is the volume of the pit, which should not
be too large otherwise apart from the flow being stagnant the time taken
to reach equilibrium condition at start-up and after a disturbance will be
longer. The pit should slope slightly down towards a point where the
drain is situated so that heavy contaminants can settle during running.
In level-controlled pits a small overflow is usually desirable to remove
accumulated froth.
While the flow of fresh stuff is normally controlled by a valve in the stuff
line, the flow of backwater dilution in the enclosed type of system is
adjusted by controlling the whole flow usually with a valve situated on the
discharge side of the mixing pump. The opening of this flow control valve
is determined in the first place by the amount of dilution required for the
stock feeding the machine; however, with a particular slice opening the
head in the breast box will vary with the flow passing through the valve,
so it is usual in more modern installations to control the head from the
flow control valve, either manually or automatically. This has lead to the
need for a more sensitive valve than the usual butterfly type and in this
respect particularly the Egger and Beloit valves are reputed to permit
more accurate and consistent setting and to be well suited for automatic
control. When the slice head automatically controls the flow control
valve it is desirable to have a reset or integral operation incorporated to
avoid offset in the valve position. Other control systems are in use,
especially where the mixing pump motor is A.C. and it is not possible to
reduce the flow as in a D.C. motor; these are dealt with in lC. l 4.
18. 3 2 Fresh water addition and the whitewater system

On machines where addition of fresh water to the backwater pit is a
deliberate policy, it is important to keep this flow closely under control.
If this. is not done the breast box consistency and temperature will fluctuate
and retention may vary with changes in the chemical condition, particularly
pH, of the backwater flow, thereby affecting the substance and composition
of the paper. On most machines fresh water is only added deliberately in
the whitewater tank where its influence will be much smaller but it is still
important for the same reasons to control carefully the fresh water actually
used on the machine. In a completely closed whitewater system excessive
39
18.4 THE WET-END FLOW SYSTEM
introduction of fresh water will increase the overfl.ov> to drain. In this

respect pump seals can admit a surprising quantity of fresh water and
should be set and controlled with a rotameter whenever possible. If
fresh water sprays are necessary on the wire it is advantageous to collect
and drain off this water separately from the rest of the system. Where fresh
water sprays are used directly on the stock the temperature of the water
should be preferably near that of the stock.
The use of whitewater, clarified or otherwise, for sprays, vacuum pump
sealing, etc., on the machine should be measured, even if only at intervals,
to prevent the system altering in balance over a period and gradually
producing changes which are difficult to detect. In fact, this applies equally
well to the whole whitewater balance which should also be checked in much
greater detail than it normally is on most machines. On existing machines
this task is unfortunately often very difficult though it is always most useful
to know the total fresh water consumption in the ft.ow system and also the
flows of thick and thin stock to drain. On machines built recently it is
commonplace to organize the whitewater system with considerable care
and provide flow measuring instruments at these points as well as to
measure the individual whitewater flows used in consistency regulators,
hydrapulpers, etc. (35, 36, 51, 58, 61).
On some types of whitewater save-all a certain quantity of longer-fibred
'sweetener' stock is required to make a mat for retaining the shorter fibres
predominating in the whitewater; in this case the save-all may actually be
incorporated in the machine fresh stuff feed system, i.e. the thickened
stock immediately becomes part of the fresh stuff flow. There seems to be
considerable danger here that variations in whitewater consistency and
flow (being mainly fines) will alter the fibre-length balance of the fresh
stuff which would then produce instability in retention conditions on the
wire. If this type of disturbance is to be adequately avoided, it is preferable
to have a separate tank to receive the thickened stock from the save-all,
and meter it at a constant rate into the fresh stuff flow or machine chest.
1B. 4 BREAST BOX APPROACH SYSTEM

With increasing machine speeds more attention has been devoted to the
design of the approach system to the breast box. The old method whereby
stock entered through one or more holes at the bottom of the breast box
began to prove inadequate mainly because of the difficulty experienced in
keeping the substance of the paper level across the web. It was found
that permanent velocity differences existed at points across the slice
causing some places to be heavier than others, and this can only be corrected
within narrow limits without causing undesirable waves on the wire and
distortion of the slice. But a more difficult and more common problem than
this is the general instability in the uniformity of flow across the machine
which a poor approach system can produce, and this can only be alleviated
by filling the breast box itself with various devices in an attempt to smooth
out the flow. The influence of the approach flow on formation depends
primarily on the degree of turbulence and flocculation of the stock on
40
entering the breast box but this is more important when considering the
design of the breast box and slice. As a general principle it may be stated
that a moderate amount of turbulence is desirable in the approach system
but the most important requirement is to spread the flow as evenly as
possible, without leaving persistent cross-machine flows, so that it enters
the breast box proper at uniform velocity all across the machine.
lB. 4 1 General approach system

The design of the approach piping to the breast box is governed to a large
extent by the layout of the machine and the stock cleaners and screens
in use immediately before the flow enters the breast box. With open strainers
on slow machines the flow is usually collected into a common pipe. On
faster machines the flows from several open strainers or from the enclosed
type of screen are lead to the breast box in separate pipes. With such a
system as this the immediate difficulty is to keep the flow matched in each
breast box inlet pipe. Pressure gauges or visual assessment are not really
satisfactory for this purpose; inlet velocity differences have been found to
persist to the slice and fluctuating cross-flows which react with one another
and form vortices at the front wall of the box can occur for the same
reason. Also the width of the screening equipment is frequently wider than
the machine, which makes even entry from each screen more difficult to
design. Because of this the trend on faster machines seems to be towards
a return to using the common pipe into which each screen discharges before
the approach system proper begins; sometimes the pressure remaining
after screening is inadequate for the speed of the machine and a second
pump may be used to take stock into the breast box.
The use of the newer enclosed screens has permitted the whole wet-end
system to be enclosed and this is generally considered to reduce the risk
of aeration while permitting a closer control on the flows in the system.
On faster machines these advantages are considerable though aeration
can still be a major problem due to the instability of flow that can occur
when air is present in stock pipelines. This will be discussed in greater
detail in lB. 7, but it may be noted that it is very important to have all
pipe runs in the approach system sloping upwards (at an angle of at least
5 deg. to 10 deg.), with air vents at top points preceding a downward
slope if this cannot be avoided. It is also considered profitable to slope
bends gradually to prevent the build-up of excessive eddies which may
vary in size and cause the main flow to fluctuate in velocity. For really sharp
bends a specially designed hydrodynamic shape is desirable. This require-
ment is particularly necessary close to the breast box but the effect of
surges caused by a disturbance anywhere in the system has been shown to
persist right through to the slice (see 1B. 8).
lB. 4 2 Comparison of different approach systems

The inadequacies of any- design for spreading the flow of stock from a
single pipe to the full width of the machine prior to entering the breast
41
box proper are shown up much more at higher speeds where inequal.ities
in flow from the slice are more obvious and troublesome. For this reason
most of the work done on this particular problem, as in fact with breast
box and slice design also, has been carried out on comparatively fast
machines or on models designed to simulate behaviour of fast machines.
With slower speed machines the main problem is generally to avoid dead
spots in the approach system which could gather slime and froth, while
at the same time keeping the stock adequately in suspension and preventing
flocculation by using a perforated roll at the point of inlet to the breast
box; the flow velocities are generally lower than on faster machines
(especially in the breast box), and inequalities in the approach have a
greater chance to even out.
Work on approach systems has also generally been associated with
investigating the breast box and slice design and it is usually difficult to
separate the two. Two principal methods have been used. Firstly, the
flows have actually been measured using various specially designed
instruments to overcome the lack of sensitivity of the simple pitot tube,
but at the same time disturbing the flow as little as possible; from this
the variations in velocity and positions of eddies and cross flows have
been studied. This technique has been confined more to experiments on
model approach systems. Secondly, the pressure of the flow as it leaves
the slice has been measured and the variations have been related to the
approach flow piping. Investigations reported in the literature have been
undertaken mainly by Mardon and his colleagues (12, 17, 24, 40, 53, 77,
89), but reports have also been given by Reitzel (46), Baines et al. (22),
Nelson (55), and Muller-Rid and his colleagues (45, 54, 72). A summary
has also appeared in an article by Brewden and Locking (50).
Despite all this it is not easy to disentangle from the wealth of data
anything of really solid worth. So much of the work has necessarily been
undertaken in conditions where it was impossible to check easily the
effect of making alterations, and it is risky to generalize too much from
improvements reported for individual machines. What follows represents
a brief summary of the ,characteristics reasonably attributable to the
various major designs which, for convenience of identification, are sketched
in figures 1 . 9.
18. 4 3 Summary of approach design characteristics

The simplest improvement over the straightforward discharge of the pipe
into the bottom of the breast box is the use of a manifold giving several
inlets, Fig. 1. 9a. In this case the pipe approaches the breast box centrally,
or less commonly from one side, and terminates in a short length of pipe
under the box from which smaller branches all of the same diameter are
taken off as smoothly as possible. This appears to have the same inadequacy
as the use of separate discharges from a number of screens; it is very diffi-
cult to ensure uniformity of flow in each inlet and discrepancies persist
through to the slice. Nevertheless, this is probably still one of the most
commonly used approach systems and it seems in practice, with the aid
42
of suitable baffles, perforated rolls, and evener plates in the breast box,
to give a satisfactory performance on slower machines.
The manifold inlet, as with other types, is sometimes associated with a
large 'explosion' chamber into which the individual pipes discharge before
entering the breast box proper, this is intended to help even out dis-
crepancies in the approach flow. However, this idea has been generally
condemned and it is claimed to be impossible to design such a chamber
on sound hydrodynamic principles; in practice it is likely to cause con-
siderable instability in the flow due to the formation of large vortices
which periodically break away to produce a disturbance.
A development from the manifold inlet design has involved the use of
a continuous sub-division of the main pipe into smaller and smaller pipes
spreading across the width of the machine, Fig. 1 . 9b. Though logical in
conception this design seems to have proved disastrous in practice, apart
from being a nightmare to keep clean. It is, of course, extremely difficult
to ensure that flow at a tee-junction splits evenly and periodic surges
from side to side can easily occur even, it seems, when some form of
streamlining is attempted. With a multitude of splits of this type these
surges can apparently become well-nigh uncontrollable.
Fig. 1. 9a. Simple manifold approach
Fig. 1. 9b. Multiple manifold approach

43
Fig. 1. 9c. Wedge flow spreader with guide vanes
---
Fig. 1 . 9d. Cross flow distributor approach
Fig. 1. 9e. Tapered side-inlet approach

44
A different approach to the problem is represented by the inlet design
involving a gradual change from the main pipe into a narrow wedge equal
in width to the machine, Fig. 1. 9c. The main problem here is having
sufficient space available to allow a reasonably gradual change in cross-
section. If the angle of spread is too large (Van der Meer (17), put it at
12 deg., although this has been disputed and 7 deg. seems to be accepted
in other industries with a similar problem), then undue turbulence occurs
even if the stock is accelerated slightly by carefully reducing the cross-
sectional area of the inlet. A perforated roll situated at the inlet to the
breast box assists in reducing this turbulence and wider angles seem possible
then. Nevertheless, to obtain sufficient length for the gradual change in
cross-section it is often necessary to double the inlet over once or twice.
This should always be done in a gradual upward direction (requiring
height which may not be available) or the whole effect of the design will
be easily destroyed by cascading and eddies in the flow. A final point is
that this type of inlet, though it has proved quite popular and has been
foun.d better than other types particularly by Nelson (55), can be very
hard to keep clean and slime is reported to gather easily at points on the
wall where flow separation occurs. Guide vanes, as shown in the figure,
inserted close to the point where the flow first begins to open out are
reported to be helpful in keeping the flow steady and bleed-offs at the
side have also been used.
Another completely different type of approach system received much
attention some years ago and had several strong advocates. This involves
splitting the flow into two pipes which enter at right angles from both
sides of machine into a specially designed taper system, Fig. 1 . 9d. It is
known as the cross-flow inlet and several types have been tried. By arrang-
ing for the rate of taper from one side to be balanced by the rate of enlarge-
ment into the opposite side it was expected that the joint ft.ow up into the
breast box would be equal across the width. At first the end of the taper
from one side was arranged to allow an overflow into the entry of the
opposite taper, but it was found in practice that the flow tended to go in
the opposite direction to that intended-from the wide end to the narrow
end. This led to the use of an overflow from both sides which was taken
back to the mixing pump inlet in the pit. The quantity overflowing has
an influence on the performance of the inlet and allows some correction
to be made, but despite alteration of this and other parts of the flow it
appears that it is difficult to remove three positions of higher pressure
at the slice, in the middle and close to both sides of the machine, and to
prevent the same sort of pressure fluctuations from side to side that have
been observed in other types of approach system. Muller-Rid and his
colleagues found the original cross-flow concept unsatisfactory for these
reasons, but with extensive experimental work on models later claimed to
have developed the arrangement to give much more improved results in
full-scale installations; their design also incorporates what they call
'tubular distributors' to even out fluctuations and this seems to be a
development of the explosion chamber idea.
Finally, a fairly recent newcomer to approach system designs is the
45
18.5 THE WET-END FLOW SYSTEM
one-sided inlet with a controlled recirculation and a gradual tapering

from one side of the machine to the other, Fig. 1. 9e. Discharge into the
breast box is effected upwards through either a conical plug or a large
number of holes or short pipes the diameter of which is carefully selected
in relation to their length. This type, though requiring careful calculation
of the taper section, is designed to give a cross-web profile which remains
very stable when changes are made to the head or backwater flow, a
feature which has been confirmed in a number of model and full-scale
experimental studies. But the use of an overflow of 15 per cent. or more
of the flow, which either requires· a separate pump or is returned to the
mixing pump inlet in the pit, means a substantial increase in pumping costs.
Also a sensitive flow control valve is needed to adjust the pressure differ-
ence between the ends of the taper though in practice this does not actually
appear to be too critical. A large overflow may also have undesirable
possibilities in that any instability in this flow could produce fluctuations
in consistency in the system which would affect the substance as changes in
retention do.
In sum, it appears fairly easy to design an approach flow system that
gives a passable performance so that with some ingenuity from the
machineman satisfactory paper can be made; but it is not possible yet to
say that any particular design is outstanding, particularly for faster
machines where the situation is more critical. The three types that appear
to give least trouble are the manifold, the gradual enlargement, and the
one-sided inlet with a taper flow; the latter, in particular, has been the
subject of considerable development in recent years and is finding much
favour for new machines, particularly in the form which involves a rect-
angular taper and perforated plate. Other types, particularly when a split
in flow is required, do not appear to be so stable. As a rule it seems wise
to ensure that all flows into the breast box are in an upward direction with
a slight acceleration produced by narrowing the cross-section (too great
an acceleration can produce backing-up eddies which will be most trouble-
some). Otherwise the only other obvious necessities are avoidance of spots
where slime can collect and designing to permit easy cleaning.
18. 5 BREAST BOX DESIGN

If the approach system were so well designed that the flow velocity is
uniform over the whole width of the machine, with no cross-currents and
only a relatively small degree of turbulence sufficient to keep the individual
fibres in random suspension without flocculation, then the breast box would
serve only to provide sufficient head at the slice. As it is, the shortcomings
of the approach system necessitate a relatively large reservoir of stock in
the breast box and a multitude of devices to help even out unwanted
disturbances and poor uniformity. There have been many new designs for
breast boxes appearing in post-war years and, as with the approach
system, most experimental work has been carried out on faster machines.
Essentially the same techniques have been used in investigating breast
box design and in addition the use of models and methods of visually
46
THE WET-END FLOW SYSTEM 18.5 l
observing the flow have been tried (see for example, Sergeant (6) and Att-
wood and Alderson (49)).
It is, however, even more difficult to extract from the numerous reports
data which may be safely generalized, particularly since the breast box is
judged mainly by its ability to smooth flow fluctuations and the magnitude
of these depend very much on how well the whole wet-end system functions.
What follows represents a brief summary of what the author considers to
be the most important design points and is largely drawn, except where
stated to the contrary, from the same articles as those quoted on page 42
for the approach system. For convenience an arbitrary division between
low, medium and high-speed machine breast boxes has been taken since the
speed, and hence the head required behind the slice, governs the general
design of a breast box more than any other factor.
1B . 5 l Low-speed design
For very low-speed machines the old dam breast box is still in common
use, yet there has been very little work of value published about it. The
head required behind the final slice may need theoretically to be only a
matter of an inch or two in order to give a velocity appropriate to the low
speed of the wire, and at such low heads there exists an appreciable
difference in velocity between the upper and lower planes of the stock
issuing under the slice. Under these conditions formation depends less on
controlling the level in the breast box, as for faster machines, than on the
position of the apron and the shake but this topic will be deferred until
drainage and formation conditions on the wire are considered. The
position of one or more earlier slices in the breast box does not seem to
be very critical and their purpose appears to be mainly to help spread the
flow evenly across the machine (since the approach system is generally
rudimentary) and to allow the head behind the final slice to be adjusted.
A row of adjustable rods of suitable shape may extend across the box in
the main compartment to help spread the flow evenly and induce a small
amount of turbulence to keep the stock in suspension; bricks and lumps
of wood can similarly, though less systematically, be employed. A recent
article (59) has gone into the subject in rather more detail and may be
referred to for further information.
For speeds much greater than 200 to 300 ft. per minute the comparative
crudity of this simple design shows up and it is more usual to find a
single slice, of a more developed type (see lB. 6), and a breast box which
is more elaborate in construction. Nevertheless relatively low heads are
still required to suit machine speeds under about 500 ft. per min. and
create a difficulty if the volume of stock in the box is not to be too small
for adequate time to be available for spreading and evening out the flow.
Where a perforated roll is used it is preferable to have it submerged
to avoid floes developing; this may necessitate curving the bottom of the
box upwards to the slice. Apart from this the problems encountered are
essentially the same as for medium-speed boxes and these will now be
considered.
47
18.5 2 THE WET-END FLOW SYSTEM
lB. 5 2 Medium-speed design and general comments

The general principle adopted in most straightforward breast boxes is to
direct the flow entering the box upwards to a level which gives sufficient
head behind the slice and to divide this flow from the slice by a baffle
extending from the bottom of the box up to somewhere near the surface
level. This serves to reduce the effect of inlet disturbances on the pressure
at the slice and often the flow may be directed under and over a further
pair of baffles for the same purpose. But it is not desirable to have the
length of the box too great because this will lead to large volumes of stock
moving at a comparatively slow velocity in the box and it then becomes
easier for dead spots, settling, aeration and flocculation to occur given
sufficient time. If the approach system is satisfactory a single baffi.e dividing
the upward entry flow and downward flow to the slice should be all that
is necessary. A moderate amount of small-scale turbulence is suitable
though in appearance the flow should be fairly streamlined and there
should be no eddies; it can be useful in this respect to build out the edge
of baffi.es on the trailing side into a streamlined P shape because where the
stock flows under or over the edge at relatively higher velocities eddies
tend to form in the lee. The flow down to the slice is preferably accelerated
slightly for better control and to keep the turbulence scale down to a
minimum; this may be achieved by humping the bottom of the box so that
the cross-sectional area can be narrowed slightly approaching the slice,
but provision may then be needed for draining the box when the machine
is shut. Especially with a vertical front wall to the box, at fairly low heads
small vortices may form when there are excessive cross flows and this is
very detrimental to formation at the point on the wire opposite the vortex.
Numerous designs which follow, more or less, these precepts are on the
market and a useful comparison of the main varieties may be found in
reference 85.
The necessity of having the breast box easily accessible for cleaning is
well recognized yet it is surprising how often with faster speed designs
that it proves difficult in practice to accomplish this without practically
dismantling the box. Joins, welds and gasketed strips have to be carefully
finished off or they readily form a place for collecting slime and fibre
lumps which cause holes in the paper when they break away. Wooden
boxes also easily collect slime and cast slivers so that most non-metal
breast boxes nowadays are lined in copper, rubber or thermosetting plastic.
Polished stainless steel is more commonly used for breast boxes though
plastic, in particular unplasticized P.V.C. strengthened by layers of resin-
bonded glass-fibre, gives a smoother surface and is easier to keep clean
than a similar-priced finish in steel. The upper surfaces of completely
closed boxes, and behind the top lip of a projection slice, are also likely
to get very dirty, especially when the stock is aerated; occasionally stock
is bled off upwards at suitable points where air could collect but unless
air gulps at the slice this may cause more trouble by disturbing the flow
than if the air is allowed to remain. Some boxes have an overflow for the
removal of excess stock together with froth on the surface but sprays are
generally considered better for this purpose.
48
Many breast boxes possess a large dump valve for diverting the full flow
to the pit for a quick inspection of the wire part without interfering with
the flow through the mixing pump and cleaners, and this requires careful
attention if it is neither to clog up with fibre nor act as a slitpe trap. There
is some merit in siting the valve outside the box proper but this depends
on the general layout. One useful solution is to insert a specially-designed
dump valve (streamlined to avoid interfering with the flow) followed by a
gate valve in the main approach pipeline; if these are pneumatically operated
it is then convenient to arrange for a single switch to close the gate valve
and simultaneously open the dump valve, diverting the flow into the back-
water pit instead of the breast box (a helpful adjunct is to have the fresh
stuff valve automatically closed at the same time to prevent thickening
up the pit).
lB. 5 3 Breast box equipment

The perforated roll has proved a very successful device for assisting in
dispersing the fibres and inducing a degree of small-scale turbulence, while
at the same time damping down larger eddies. It can be used with advantage
for evening out where the flow enters the box and it then becomes, in
effect, part of the approach system. The perforated roll is also used very
frequently just in front of the slice where its main purpose is to impart
a final small agitation to the fibres in an endeavour to spread them evenly
throughout the flow, particularly in a vertical direction; the roll is also
thought to help reduce any tendency to flocculation due to the high shear
forces experienced as the stock is accelerated through the holes, and this
has been confirmed in experimental work by Marden et al. (91). This par-
ticular reference presents a thorough-going analysis of machine experiences
with perforated rolls in addition to experimental results, and is drawn on
considerably in what follows. As the perforated roll nearest the slice has
the greatest influence on the characteristics of stock in the slice jet, attention
is concentrated on this particular application.
There are so many variables involved with a perforated roll that it is
extremely difficult to reach any general conclusions on its performance.
Mardon distinguishes between two basic effects of the roll: the hydro-
dynamic which relates to the smoothness of flow from the slice, and the
papermaking which relates to the degree of deflocculation and absence of
fibre clumps resulting from using the roll. Often it is found that require-
ments for hydrodynamic stability and for papermaking are directly oppo-
site, and choice of dimensions and other features of the roll becomes a
compromise based on inadequate detailed information. The variables
associated with perforated rolls are mainly the position relative to the
slice, hole diameter, open area, and speed and direction of rotation.
Each of these will be discussed briefly.
If a perforated roll is too near the slice streaks and other disturbances
are more apparent on the wire. These disturbances, known generally as
the 'wake effect' of the roll, are enhanced when there is discharge from the
breast roll. Mardon quotes an empirical formula for the minimum distance
49
behind the slice at which to place the roll to avoid the wake effect, and
gives several examples illustrating its application. The formula indicates
that the distance required between the roll and slice increases for larger
hole diameter, smaller roll diameter, smaller open area and greater flow
velocity through the roll. Elsewhere a useful rule-of-thumb guide for the
minimum distance is given as the equivalent of twenty hole diameters.
From the hydrodynamic viewpoint the smaller the hole diameter the
less likely it is that any wake effect occurs. But Mardon found that smaller
holes produce less deflocculation downstream of the roll, though this
effect also depends on the velocity of flow through the roll and at lower
velocities is not so prominent. He advises that generally the smallest holes
practicable should be used, though to avoid other disturbances referred
to below it may be necessary to use holes up to 1 in. in diameter. A com-
promise is needed with the flow velocity because although when this is
lower it is less likely to cause the wake effect, it also gives more time for
reflocculation to occur downstream of the roll. Mardon found evidence
for the existence of a critical velocity above which the stock would flow
for a considerable distance in a defl.occulated state; for a mixture of soft-
wood and hardwood kraft this is about l ·O feet per second.
The questions of open area of the perforated roll and speed of rotation
are ones that have received the attention of several workers. Bennett (29)
carried out experiments using air as a medium (even allowing for the
appropriate scale factors some care must obviously be taken before
extending these results to a fibrous suspension), and found that although
the usual design of perforated roll is good for levelling non-uniform flow,
the degree of solidity should not be greater than 50 per cent. or instability
is likely to occur. This confirmed an earlier statement by Van der Meer
(17) who advised use of a roll more than half open for inducing micro-
turbulence, otherwise uncontrolled merging of streams on the downstream
side of the roll creates large scale turbulence. Nelson (55) on the other
hand has reported that a 60 per cent. solid roll was good at evening
fluctuations, which is in direct contradiction. Marden has presented the
most detailed comparison of rolls with different open area and his con-
clusions are that rolls away from the slice should generally have 40 per
cent. open area, whilst those just ahead of the slice are best made with about
48 per cent. open area unless long fibres are involved in which case the
open area is best reduced to 43 per cent. or lower to lessen the risk of
stapling of fibres across the land gaps.
Rotation of the roll is important from the point of view of keeping it
clean and although Bennett found that rate of rotation had no effect there
is now evidence that the flow is disturbed if rotation is too great. Wrist
(74) found that a slow rotation can produce large floes that pass unbroken
through the slice while too fast a speed of rotation increased turbulence
and instability of the flow; he considers that a variable speed drive for a
roll at the slice position provides a valuable control on the performance.
Mardon again in two papers (83, 91) investigated extensively the effect
both of speed and direction of rotation. He reaches the conclusion that
speed is not too critical so long as it is fast enough to avoid stapling of
50
THE WET-END FLOW SYSTEM IB.5 3
fibres, but not so high that hydrodynamic disturbances occur and show
up as deep ridges in the flow from the slice. It is usual to rotate the roll
against the flow at the bottom to reduce the possibility of approach
system disturbances passing straight to the slice. This also reduces the
severity of fibre clumps caused by fibre 'scooping' on the trailing edge of
the holes and then breaking away on the downstream side. 'Scooping' of
fibres is worse at higher speeds of rotation (in contrast to stapling effects)
and with smaller diameter holes.
Various constructional details of perforated rolls are important. The
gap between the rolls and the walls of the breast box must be even and
narrow otherwise unstable flow conditions can occur and too much stock
can by-pass the roll without meeting shear forces in the holes. When a roll
is situated in the body of the box it is often considered preferable to keep
the level just above the top of the roll in order to reduce any by-pass flow
but prevent the possibility of aeration; on the other hand many paper-
makers prefer to have the roll just above the level so that the presence of
fibre on the land areas can be seen. Holes must be pitched in a spiral
pattern to equalize their effect across the machine, and according to
Mardon it is preferable for rigidity to make the roll from a drilled tube
and have a reasonable wall thickness radiused off to not greater than
rl2 in. Discs in the roll assist in damping down cross-currents and can be
spaced as little as 1!- inches apart; they should be perforated to allow some
cross-flow to take place. Construction of the ends of the roll is particularly
important and in this respect there has been some difference of opinion
regarding how close to the wall the holes in the roll should extend. Some
have considered that leaving a solid space at the end will cause the flow
on the downstream side to rush into the wails thereby overcoming the
slow boundary layer which can reduce substance at the edges. However,
apart from the fact that the roll should be far enough back from the slice
to eliminate influences of this sort, all that seems likely to happen is that
undue turbulence will be created at the edges making substance there
variable. The roll should in fact be identical in construction along its
entire length and various designs conforming to this requirement are given
by Mardon.
Other devices have been used to even out flow and damp down turbulence
in the breast box, including stave rolls (constructed from a number of
long thin rods stretching across the width of the box and held together
at intervals by discs), pear-shaped deflectors, 'homogenizers' (angled plate
fins on a tubular shaft) and evener plates. Reitzel (45), has described the
use of stave rolls in the approach system of a specially designed box where
he claimed that they prevented alignment of fibres in a converging flow
and were easier to manufacture and keep clean than a perforated roll.
The stave roll could not, however, be used near to the slice otherwise
marking of the sheet occurred and in a more recent modified breast box
designed on similar lines this type of roll has apparently not been used.
Evener plates or flow eveners are in general use to reduce cross currents
close to the slice in the breast box; the trouble with these is that channelling
is likely to occur producing permanent streaks in the paper and it is
51
normally considered inadvisable to use flow evener plates without following

them with a perforated roll to overcome this effect. Alternatively, Wrist
has claimed that temporarily narrowing the passage by using a stationary
cylinder or solid projections on the walls close to the slice is more beneficial
in this position though other reports on this device have indicated that
excessive turbulence in the slice jet may be induced (82). The edges of the
individual plates should be rounded to prevent stapling of fibres and it is
essential that the plates fit snugly into the breast box walls.
lB. 5 4 High-speed boxes and designs for a wide range of speeds

The pressure required behind the slice for machine speeds of 1,500 ft.
per min. is equivalent to about 12 feet head and the design of a satisfactory
breast box to give heads of this order presents many difficulties. The air-
loaded box has been very successful in providing a way out of this difficulty
and many types are now in use. The basic advantage of an air-loaded box
is that the lower stock level gives a more controllable flow to the slice.
On the other hand, air-loaded boxes are more sensitive to pressure changes
in the approach system, and although the actual stock level in the box
alters less than for an open box working under similar conditions, the
change in total head at the slice (and hence of stock velocity from the slice)
is much greater than with an air-loaded box.
The most common device for stabilizing the operation of an air-loaded
box is the so-called Hornbustel hole which allows a mixture of stock and
air to be continually bled off from the side of the box. The stock level
in the box stabilizes at the height of the hole and effectively prevents the
water/air surface from the violent fluctuation in level which occurs in a
completely closed and uncontrolled system. To obtain a greater head at
the slice, all that is necessary is to increase flow from the mixing pump
and a new equilibrium condition is eventually reached with the same
stock level but greater air pressure in the box. It is a valuable asset to be
able to adjust the level of stock in the box in order that velocity of flow
to the slice can be varied though frequently the range of machine speeds
and breast box consistency is sufficiently narrow to make alteration of
the level unnecessary. With a simple hole in the side of the box adjusting
of level is not possible, but there are several simple modifications providing
this facility and these include placing the hole eccentrically in a rotatable
flange, using a vertical tube adjustable in height in place of a simple hole,
or employing a float system operating on a needle valve to regulate the
air bleed-off.
In an analysis of the stability of different types of air-loaded box control
systems, Marden et al. (90) points out that with the Hornbustel hole the
slice head change resulting from a given increase in stock flow to the
breast box can be considerable. This is because if the stock level rises,
air pressure must first build-up in order to reduce the level to its former
position, thus magnifying the original head change. Hunting is very easy
to obtain if the size of the hole and setting of the air input pressure are not
52
carefully matched. In practice some difficulty can be experienced in obtain-
ing a smooth bleed-off which does not gulp periodically and cause a
violent change in head.
A different approach to the problem of achieving stability of air-loading
is to control the level in the breast box by automatic adjustment of the
main stock flow control valve on the discharge side of the mixing pump.
(Alternatively control can be on a bypass round the main stock flow valve
or on a recirculation line round the mixing pump). To detect the level
position in an air-loaded box a differential pressure cell arrangement is
used. The total head at the slice is measured and controlled by regulating
the air pressure in the box either by direct adjustment of a valve on the
supply line or by adjustment of a separate air bleed-off also on the supply
line or on the box itself. With this arrangement, if for example a greater
head is required at the slice, then the air pressure can be increased, thus
depressing the stock level causing a compensating action on the main stock
valve which opens until the additional quantity necessary to maintain the
level is reached.
According to Mardon, this system is not stable because although an
increase in flow to the box causes level to rise thus producing a closing
action on the flow control valve, air pressure will also reduce due to the
control action initiated by the rise in total head, and this will tend to
counteract the first action; thus hunting is readily set up. A better system
for stability is to control the total head by adjusting the main stock flow
valve (as. is in effect done with an open box), and control the level by
adjusting air flow (possibly retaining the Hornbustel hole for this purpose).
Provided a three-term controller is used on the stock flow valve, this
system can give very good stability if properly set up. Although in theory
it should be possible with either of these systems to run at very low machine
speeds by applying a vacuum instead of pressure to the air chamber this
does not in practice appear to work very satisfactorily.
Another design which has met with some success and which departs
completely from the air-loaded pattern is the flow nozzle. In this type
there is no large reservoir of stock or air-chamber giving a cushioning action
in the breast box, but instead the flow is direct from the pump through
the approach system to the slice and always under pressure. A refreshing
and frank account of the difficulties encountered in taking a design of this
type from the drawing board to the machine is given in the paper by
Reitzel mentioned above. Apart from designing the nozzle to permit
adequate cleaning and avoiding horizontal upper surfaces on which slime
can collect, one of the main problems in practice is to prevent hydraulic
fluctuations from reaching the slice when there is no natural damping
available as in normal breast boxes. However, this difficulty has been
overcome adequately even without use of a surge tank which has been
advocated by several writers, and it is very likely that on faster machines
a modification of this type of box could prove useful. There are also some
designs for tissue manufacture where the sheet is partly formed on the
wire within the box itself and also the 'frozen-flow' box in which stock is
conducted through a large number of 4 in. copper or plastic tubes extending
3 53
well into the box, but these are too specialized or too little known about
for consideration here.
One advantage of both the air-loaded box and the flow nozzle is that
both are readily adapted for varying machine speeds. The difficulty here
is often to maintain adequate depth in the box at low speeds while pre-
venting sluggishness in the flow. The problem of depth is solved satis-
factorily in air-loaded boxes though the flow will tend to become too slow
for low machine speeds; in the flow nozzle this is apparently not so much
of a difficulty. On boxes of conventional design an attempt is sometimes
made to run at lower speeds and at the same time to keep up the level of
stock in the box by interposing in the section of the box leading to the
slice a means of throttling the flow; this may be in the form of rubber
restrictors or possibly by means of a specially designed sliding gate valve.
This solution suffers from the considerable disadvantage that it is almost
impossible to avoid interfering with the flow pattern at a critical point
and the uniformity of the sheet is liable to suffer; also the presence of air
in the stock can very easily cause trouble.
1B. 6 SLICE DESIGN

Adjustment of the slice across the machine is the only way in which
deficiencies of flow uniformity in the approach flow system and breast
box can be removed, and once stock has flowed through the slice no further
correction to the dry weight of the paper across the machine can be made.
Apart from this highly important aspect, the slice opening in relation to
the pressure head determines the volume flow, and hence the consistency
of the breast box stock, while the relative position of the lips is critical
for assisting in formation of the sheet on the wire in the early stages.
This latter topic will not be dealt with in detail, since it is more approp-
riately considered as a part of the wire section in Part 3, and only the
general character of the slice jet itself will be covered here.
1B . 6 1 Vertical slices
With the old double-straight slice no proper adjustment of the cross-
machine substance is possible (unless sticking pieces of paper underneath
the slice is considered adequate). For that reason an adjustable form of
vertical slice has replaced the rigid slice on most older machines and this
allows the substance to be altered more easily across the machine. More
advanced designs of breast box for medium speed machines are also
sometimes fitted with vertical slices, but they have not received as good
a reception as the projection slice. The main reason appears to be that
adjustment of the slice blade in a vertical direction introduces stresses
which can easily lead to buckling of the blade; this is particularly the case
where large temperature changes occur and in addition it is generally
necessary to change the blade on the slice at fairly frequent intervals.
The vertical slice is usually situated directly above the end of the bottom
lip or apron and this gives a jet which is roughly horizontal. The coefficient
54
of contraction is, however, much lower than for a projection slice (about
0·7 compared to 0·95 or more) which means that higher heads behind the
slice are necessary. Although the jet is probably more stable close to the
slice the large increase in velocity up to the vena contracta will make
the actual velocity of impact with the wire for a given head very sensitive
to the width of the gap over which the jet travels after leaving the slice
orifice. It is generally considered that the substance across the machine can
be adjusted more accurately with a vertical slice and in a report by
Miiller-Rid and Pausch (45) this was confirmed; these authors tried out
both a projection slice and a vertical slice on a machine and both tests for
uniformity of substance and formation across the sheet and the subjective
assessments of the machine crews agreed that adjustment was achieved
more rapidly and accurately with the vertical slice.
The same authors report similar tests using a combination of the projec-
tion and vertical slice. In this design a thin blade is attached vertically
to the end of the upper lip of a projection slice and protrudes a fraction
of an inch below the lip; this blade is used for adjusting cross-machine
differences rather than the upper lip itself. This modification proved more
successful than either the straightforward projection or vertical slices and
was chosen as the most suitable method of adjusting the slice. It is now
being used increasingly on new machines and the blade has been christened
a 'profile bar' or, with a curious sense of inappropriateness, a 'spoiler'.
The main difficulty encountered with this idea is that fibre tends to collect
on the edge of the vertical slice; with long-fibred stock it becomes hardly
practicable to use it. It would appear also that the extent by which the
vertical blade protrudes below the upper· lip is critical and it should
ideally only skim off the boundary layer, where the velocity is lower; if
the blade extends too far into the gap formed by the main lips turbulence
can be expected at a critical point in the jet and this would be most
undesirable.
1B . 6 2 Projection slice
The projection slice is by far the most common in use and despite its
vvell known weaknesses it is generally considered the most reliable type.
A substantial amount of work has been done in investigating flow from
this type of slice and this will be drawn upon in what follows; apart from
the results of Milller-Rid and his colleagues mentioned above, this work
has been reported, except where stated, in the papers already listed on page
42 when describing approach flow systems.
First, a few general remarks on the projection slice. It is particularly
important that the slice edges, especially the upper lip, are kept clean, and
this is much easier if they are rounded slightly. Another reason for avoiding
sharp edges is that instability of the flow is more likely to occur due to the
sudden separation of flow at the end of the lips, but on the other hand if
the bevelling is too great air is liable to gulp back causing a disturbing
jump in the jet. The inner surfaces of the slice lips should be very smooth
and it is important that there should not be a sudden discontinuity where
55
the lips are attached to the breast box. With the upper lip this is
usually difficult to avoid since the lip is arranged to pivot about the
end point; however, in the paper by Reitzel an attempt to overcome
this by using a flexible joining strip is described and appears to have been
successful.
In some types of projection slice, notably the Van de Carr, it is usual
to slope the wall of the breast box down to the slice. Though this seems
an eminently sensible idea to encourage a smooth flow to the slice and at
lower heads should reduce the risk of vortices forming (as occurs with
vertical walls), the sloping front wall has met with some criticism. It is
apparently difficult to predict the degree of slope required and Mardon
and Van der Meer (12) in particular indicate from their experience a
strong preference for the vertical wall. For stable flow the angle between
the lips should, according to Van der Meer (17) be between 10 deg. and
20 deg.
1B. 6 3 Adjustment of the upper lip
The position and operation of the individual jackscrews on the upper lip
is very important since it is their efficient functioning that governs how
easily a uniform substance across the web is achieved. The flow issuing
from any particular point across the web spreads slightly and intermingles
with adjacent streams; this has been shown clearly by Sergeant (6) who
added a thin stream of dyed fibres just after the slice and found a distribu-
tion in the reel-up spreading over a width of several inches, and also by
Mason et al. (14) who used radioactive-tagged fibre in a similar manner
and obtained a normal distribution in the reel-up spreading over 3-t in.
to 5 in., less than in Sergeant's work but this was on a faster machine.
Normally the jack-screws are arranged at 5 in. to 6 in. intervals, which
should therefore be adequate for adjusting the substance accurately.
Unfortunately, as is well known, alteration of one screw causes stresses
in the blade which alters the position of adjacent screws. Cuffrey (30) has
investigated this effect in some detail and found, for example, that a drop
in one screw giving the equivalent of a 4·5 per cent. weight change caused
the adjacent screws (set at 4 in. intervals) on either side to drop also,
though to a lesser extent; however, as far as three screws away there actually
occurred an increase in substance equivalent to as much as 3 per cent.
Though severe there was no doubt that this effect was due mainly to the
slice blade flexing and only partly to diversion of the flow outwards at the
slice and on to the wire on either side of the restriction. Some other experi-
ments reported by Cuffey are illuminating in illustrating the difficulties that
face machinemen in adjusting cross-level on a normal projection slice. In
one case individual jack-screws were adjusted from front to back of the
machine; when this was completed it was found that the front settings
were no longer correct by a comparatively large margin. Even when set
as near as possible during a shut period, when water was run through the
slice under the usual running pressure of under 20 in. head the slice opening
increased more in the centre than at the sides necessitating the resetting
of screws at the sides by as much as 15 thou. Some weaknesses in the
56
usual worm gear and offset cam arrangement for lifting the whole lip
were also found; for instance, the front side dropped much more than the
back when the overall slice gap was reduced and this would have produced
a lighter sheet at the front compared to the back. For small movements
of the whole lip the back of the slice could actually go in the opposite
direction to the front depending on the direction from which previous
movement was made.
Machinemen are generally aware of the lack of reliability in adjusting the
whole of the slice gap and know that doing this is liable to upset the cross-
level substance. For this reason there is a definite reluctance to alter the
slice gap when a change in the amount of backwater used for dilution is
required and the head behind the slice is allowed to alter instead. As
already discussed, this is a bad practice except possibly for very small
changes and it would appear most desirable to effect an improvement in
design of the mechanism for altering the slice lip, possibly along the lines
suggested by Reitzel in his paper.
Another problem associated with the slice is that of obtaining uniform
substance right to the edges. In practice, because of the slowing down of
flow at the edge walls of the breast box and slice, the substance tends to be
light at the edges and some undesirable bow waves can be formed on the
wire. A common practice to overcome this is to lift the slice slightly
at the edges. This would appear a more satisfactory solution than attempt-
ing to influence the rate of flow at the edges of the slice either by leaving
out holes in the perforated roll at the edges, locating the deckle a little
in from the slice edge, or attempting to divert some of the flow towards
the edge by means of a small obstruction projecting from the side into
the slice. All of these methods are frequently used but may be liable to
induce too much turbulence, making the setting critical and giving a
variable performance. Another popular method of overcoming the edge
problem is to bleed off a small quantity of stock at the edges of the slice;
however, some recent work by Mardon and Wahlstrom (53) indicates
that the bleed-off may have the opposite effect to that desired-closing
the flows off could actually cause an increase in pressure, and hence in
substance, at the slice edges, rather than the decrease expected. The position
of the deckles should coincide with the slice edges to minimize turbulence
and this is one very strong reason why even on very slow machines it is
usually better to avoid changing deckles and instead to run a trim down
into a hog pit. Even if flow from the slice were perfect it would not prove
possible to form a perfect sheet right to the edges of the deckle because
stock will tend to climb up the edge inducing a wave across the wire. The
main object of the machineman should be to confine these disturbances
to as narrow an area as possible so that low-quality paper at the edges is
removed in the trim; this must be done by trial and error for each machine
by observing the effect of the adjustments available on the dry-line and the
general smoothness of flow at the edges.
During a shut period it is common practice to set the slice by means of
a taper gauge, but this is inadequate in several respects. Even if the setting
is done with sufficient care to get the gap as accurate as possible all the way
57
across, once the slice gets heated with stock and pressure forces up the
middle more than the edges, the gap between the lips across the machine
may be extremely variable. To all intents and purposes this implies that
the substance also will be variable in the same way since at any point
across the machine this depends essentially on the width of the gap
(Attwood and Alderson (49) exhibited this rather neatly by using a
suspension of polystyrene pellets in water through a model slice-the
number collected at any part of the slice followed almost exactly the width
under pressure of the slice gap). Perhaps the only method which achieves
a reasonable degree of accuracy is to study flow from the slice with water
heated to the normal running temperature and working with the usual
head behind the slice. Provided sufficient time is allowed for the whole slice
to achieve normal operating temperature (and this may be sometime if
the slice and breast box has cooled down) observation and measurement
of the volume flows across the slice will enable the jack-screws to be
adjusted closely to normal operating conditions as well as exhibiting other
deficiencies in the flow. But this procedure can only be done with the wire
removed and even then measurement of the flow at different points across
the machine is far from easy to manage, so provided an adequate and
rapid means of assessing the cross-web substance profile is available to the
machineman (this is considered in more detail in lC. 3 4) then it will
generally be more satisfactory to continue with a relatively crude prepara-
tion and then to adjust the jack-screws as quickly as possible once the
machine is under way.
lB. 6 4 Character of the slice jet

Some detailed studies have been made of the character of the jet from a
projection slice in relation to the position of the lips. Mardon and his
colleagues have presented a wealth of data obtained on different machines
in the course of their investigations and have exhibited how pressure
varies through the thickness of a slice jet; their general findings in this
respect have also been confirmed by Attwood and Alderson. Close to
the lips the pressure, and hence the velocity, of the ft.ow is lower due to
boundary layer friction. This area of low velocity extends only a short
distance from each lip but is nevertheless thought to be responsible for
some of the instability and fluttering of the jet that has been observed on
faster machines. Small obstructions in the slice are less liable to create
turbulence when the boundary layer is thin; the boundary layer thickness
is increased by roughness of the slice lips but the presence of fibre helps
to reduce it and makes the jet more stable. Apart from this effect, which
is confined closely to the top and bottom of the slice, the pressure through
the thickness of the jet is pretty well the same except for a slight increase
towards the bottom of the jet due presumably to the slightly greater
pressure there. Further details can be obtained on this topic in an article
by Mardon and Shoumatoff (25).
The positions of the top lip and the bottom lip or apron in relation to
each other and to the wire are very important. Normally the apron is
58
horizontal and the top lip angled downwards, though in some cases the
apron may also be directed downwards when rapid drainage is desired.
If the top lip terminates directly over the end of the apron then the jet
as a whole has a slight downward velocity. The further the top lip is moved
back in relation to the bottom lip the more horizontal becomes the jet,
at least up to a point where instability may occur due to prolonged drag
on the apron. Normally it is considered that the more horizontally the
jet hits the wire the less will be the disturbance, and it has been recom-
mended that a good compromise is achieved by having the top lip back
from the apron by a distance equivalent to twice the slice opening; accord-
ing to Nelson (55) this also has the advantage that the jet angle is not so
sensitive to variations in slice opening. If the top lip terminates above the
apron, a horizontal jet can only be obtained by sloping the apron upwards.
On some machines the lack of drainage capacity necessitates removing the
water as quickly as possible and in this case it is usual to direct the jet
downwards to give greater impact and initial drainage. The position of
the end of the apron controls the point at which the stock hits the wire
although, of course, this is also affected by moving the top lip (if this is
moved back, up to a certain point the jet will carry further down the wire).
If rapid drainage is desired the jet is made to hit the wire close to top
dead centre of the breast roll which will then remove a lot of water, but
generally speaking this will reduce the chance of getting a clean formation
because the jet is violently disrupted by the suction forces of the breast
roll before it has settled even slightly on the wire. Since it is the formation
which is affected primarily by this it is appropriate to leave this topic for
closer consideration in Part 3.
lB.7 AERATION OF THE STOCK

It has long been recognized that an excessive amount of air in the paper
stock at the wet-end is undesirable, mainly because froth is more likely to
become troublesome under those conditions and in some installations
there may be difficulties with depriming at the mixing pump. During the
last few years, however, it has become apparent that aeration causes many
more effects than these and most of them are detrimental in some way to
efficient running of the wet-end. The development of a simple apparatus
for measuring the volume of air in a sample of paper stock by Boadway
(23) has assisted in stimulating investigations into this problem and since
then there have been several useful reports of work in which either the
instrument of his design or a modification of it has been used. In addition
on several machines application of the deculator, a device which deaerates
stock very efficiently, and other methods of de-aeration have shown up
differences in behaviour of the stock caused by the removal of air.
lB. 7 1 Nature and effects of air in paper stock

There is some difference of opinion regarding the precise definitions to be
applied to distinguish between the various ways in which air is held in
suspension in paper stock. In descriptive terms the analysis used by
59
Gavelin (11), Jacobsson (19) and other workers seems most convenient
and in this the air is divided into three main classes: free, residual and
dissolved. Free air represents air which will settle out and bubble to the
surface if the stock were allowed to settle; it is primarily responsible for
foam and reduces in quantity with depth. The definition is a rough one
in that if the stock were allowed to stand the volume of air remaining would
reduce for a long time; thus Gavelin found that some breast box stock
retained 0·25 per cent. of air after one minute but only 0· l per cent. after
a much longer time. Although free air can disturb flow on the wire and
seems to accentuate pitch and slime trouble, it is not thought likely to
affect drainage and formation directly.
Residual air does not settle out and may be thought of as being inti-
mately in contact with individual fibres. It is considered to originate largely
from mechanical dispersion of free air in refining and other sources of
vigorous agitation but may amount to only 0· 1 per cent. by volume of the
stock; however, even though this seems a very low percentage, it still
represents a substantial proportion of the volume occupied by the fibre
in breast box stock which, if the consistency were 0·4 per cent. would be
only four times as great. The effect of air in this form is more critical with
regard to drainage on the wire where the presence of air in capilliaries will
affect the water flow and retard drainage. It also can be expected to decrease
the effective specific gravity of fibres, affecting their settling rate and
increasing any tendency towards flocculation in quiescent regions because
of the different rates of settling or even of buoyancy between individual
fibres. Surface tension forces in air bubbles contacting fibres that have
coalesced will increase the strength with which they are held together and
this also will increase the degree of flocculation.
Dissolved air is regarded as air which is held within individual fibres
and derives primarily from the fibres themselves in the form of air and
carbon-dioxide. It is not regarded as important unless, due mainly to
beating action, fibres become saturated with air and bubbles then take the
form more of residual air on the surface of the fibres.
The apparatus designed by Boadway for measuring air in paper stock
can easily be used with a little practice and subjects a fixed volume of the
stock to a given vacuum; the change in volume due to expansion of the
air is measured and the percentage of air in the stock can then readily be
calculated. The determination includes both free and residual air in the
stock but not air in the dissolved or combined state. In practice the
amount of residual air at the wet-end of a paper machine (as assessed by
other methods) does not appear to alter substantially except possibly at
the mixing pump where the percentage of free air may increase at the
expense of residual air and of residual at the expense of dissolved air; by
far the biggest volume of air, and the cause of most variation in air content,
is due to the free air. The sources of this free air and its relation to foam
development will now be discussed.
lB. 7 2 Free air and its causes
A certain amount of free air may be expected in the fresh stuff from the
60
THE WET-END FLOW SYSTEM IB.7 3
preparation system but most is developed at the wet-end itself. The main
sources, not in order of importance, may be listed as due to cascading
flows (in mixing box, approach system and wire tray discharge), falling
through the wire (especially if trays are not used and the backwater falls
direct to a wire pit), insuction of air in mixing pump glands, open pipe
discharge into tanks (these should be below the surface level), insuction
in vortex cleaners with an open reject orifice, and excessive eddies and
vortices in tanks, breast box, etc. open to the atmosphere or in channels
and half-full pipes. Careful design can obviate most of these causes and
the whole backwater system should be organized with these requirements
very much in view; this has been done, for instance, on a machine started
recently where a specific objective in the design stage was to site chests
and pipe lines to minimize cascading while the main whitewater chest
surface was made as large as possible to help air to settle out (48).
In addition to these more obvious causes of aeration some recent work
by Mardon et al. (40, 88) has brought to light other less suspected sources.
The air content in breast box stock was found, for instance, to depend on
the depth of backwater in the wire pit, the shallower the depth the greater
the aeration. In addition discontinuities in velocity and pressure in pipes,
valves, tee branches, etc. were shown to affect the air content considerably.
Apart from this, the separation of air into a separate phase produces
surges in the flow which are detrimental to stability and Mardon demon-
strated that reduction of free air in the system helped to lessen the effect
of these disturbances on the weight of the paper. This topic will be dealt
with in greater detail in lB. 8.
Brecht and Kirchner (44) determined the air content of stocks for a
variety of papers and found an enormous variation from as little as 0· 1
per cent. for parchment to 4 per cent. for newsprint and 4·5 per cent. for
tissue. The latter figures are of the same order as those reported by other
workers mentioned earlier. The air percentage in the stuff chest of one
machine was 2 per cent. while that of the table-roll backwater was 11
per cent., thus confirming that backwater is the main source of free air.
It was also found that different pulps have a different propensity for
entraining air under mechanical agitation: for instance, groundwood fibres
were particularly prone to collect air. Drainage rates were slower with
more aerated stock while laboratory handsheets exhibited a reduced wet
and dry strength.
Reduction of air in stock has other important effects besides those
already mentioned and of these lessening the development of foam is one
of the most important.
lB. 7 3 Foam at the wet-end

The use of devices for removing air in the approach system of the paper
machine has demonstrated that foam does not form unless air is present
in the stock. Nevertheless on a machine which is plagued with this common
source of trouble it is not immediately helpful to the machineman to be
told that the stock is becoming too aerated and the design of the wet-end
61
needs overhauling, as will almost certainly be the case. Various defoaming

agents, usually containing a sulphonated oil, are available, but their use
should be regarded as an uneconomic stop-gap. Normally pulsating or
rotating sprays in the breast box are most useful for preventing the foam
growing to such an extent that it breaks away and the paper becomes
affected; skimming off the surface in the breast box and backwater pits
also helps though there are usually practical difficulties.
Foam becomes worse under certain chemical conditions such as are
caused by insufficient washing of a pulp, excess alkalinity, high pitch content
pulps, high rosin size concentration or when lime salts exist in the water
(4 7). Higher temperatures may also be expected to increase the likelihood
of foam because the air then settles out easier. Fillers do not give much
foam except, according to Brecht and Kirchner (62), satin white, and even
in this case foam did not occur at a suitable pH. The same workers
developed an experimental technique designed to assess how readily
pulps took up air; this involves blowing a given volume of air through
the pulp suspension with simultaneous agitation and then measuring the
subsequent air content. Basic and acidic dyes both usually caused extensive
take-up of air and also produced a foam which was comparatively stable,
i.e. would not disperse so readily once formed. Direct dyes, however, were
mainly without effect. A large number of sizing agents all affected the
development and stability of foam. The effect of the circulation water
pH was also investigated though this cannot easily be extended to machine
conditions; it was found that air was absorbed most easily at a pH of 5·5,
and was at a minimum at a pH of 4·5, though the foam tended to become
more stable the lower the pH.
lB. 7 4 Removal of air in paper stock

The deculator has proved very efficient for removing air in the stock and
its design and use has been described in detail by Jacobsson (19) and others
(8, 38, 42). The device involves the spraying of jets of stock onto plates
in a cylindrical tank under a vacuum which is within O· l in. to O· 3 in.
Hg of the boiling pressure corresponding to the stock temperature. This
succeeds in removing the free air and most of the residual air in the stock.
It is obviously necessary to treat all the breast box flow and not just the
fresh stuff because it is the backwater used for dilution that contains the
higher percentage of air. Thus a rather expensive installation is required
and apart from the vacuum pump .it is frequently necessary to use two
stock pumps, one to mix the stuff and backwater and provide pressure
to the de-aerator (and in some cases also the stock cleaners), the next to
provide pressure in the normal way to the breast box.
Another form of de-aerator is the Vorvac system which is claimed to
remove air from the stock in association with a type of cleaner developed
from the older Vortrap design. Also 'eductors' can be fitted to the reject
orifices of cyclone cleaners for the same purpose. Both these devices
originated from the observation that the high rotational forces and general
turbulence developed in the cleaners produced an increase in aeration in
62
the stock. Their use may well reduce this effect considerably but to the
author's knowledge no experimental data has been made available to
show whether the amount of air left in the stock is of the same order as
with a device working on the principle used in the deculator.
The result of de-aerating stock in a deculator is very marked. Foam is
absent and no showers are necessary in the breast box; the suspension
settles quickly in the absence of turbulence leaving water on the surface.
Slime grows in the usual places but almost no fibres adhere to it and the
slime appears to stick to surfaces more firmly; this reduces the likelihood
of lumps breaking away. The slice jet is smoother and on the wire the stock
looks more glassy in appearance and drains faster than usual (this has
been confirmed experimentally by Gavelin (11) using a specially designed
drainage tester). Quicker drainage on the wire enables a better sheet to
be formed or the machine speed to be increased; in the former case a
reduction in porosity, increased smoothness and better printability of
newsprint have been reported. These advantages seem at first sight to be
considerable, though as with all devices like this economic justification is
not easy without a knowledge of the likely improvement on any particular
machine. On most machines troubled with foam or lacking drainage
capacity on the wire it will be more satisfactory in the first instance to
examine the wet-end critically, bearing in mind the points raised above,
to detect and remove sources of aeration.
It is appropriate to mention here that bubbles on the wire of the type
which are frequently 'broken' by means of steam jets are not considered
to be due directly to aeration, though they are genera11y accentuated when
stock is aerated. The bubbles are too large to originate from air actually
in the suspension and come mainly from vortices in the breast box, stock
jump on the wire, and accumulations of air in the breast box which become
large bubbles and are then emitted from the slice. Air can also get sucked
in between the slice and the wire and there are various devices designed
to alleviate this.
1B. 8 HYDRAULIC DISTURBANCES

Hydraulic disturbances in the main backwater circuit can have a direct
effect on the substance of the paper since any fluctuation in pressure which
reaches the slice produces a variation in flow rate. Such fluctuations can
be conveniently divided into two categories: those caused by a sudden
pulse of pressure which produces a surge that eventually reaches the slice,
and those caused by a vibration in the system which sets up a wave of
continuously oscillating pressure and produces a cycle in substance of the
paper of the same frequency. On faster machines these disturbances can
show up very clearly in the paper and under certain conditions may be
so extreme as to prevent paper being made; it is, therefore, important to
be aware of their causes and the remedies available.
lB . 8 1 Flow surges
Marden and his colleagues (40) made a comprehensive investigation of
63
hydraulic oscillations in the approach system and breast box of a fast

machine and measured fluctuations in pressure of a frequency less than
one cycle per second; these oscillations were compared with beta-ray
traces along the machine and high-speed cine records of the wire and
fluttering of the dry line in order to assess their effect on the paper.
Generally the frequency of the vibrations was found to diminish from the
mixing pump onwards, though the amplitude increased.
As a result of this work surges in flow with their accompanying oscilla-
tions were found to be very dependent on the percentage of free air in the
stock, the lower the percentage the less the magnitude of the oscillations.
The surges originated where abrupt changes in pressure occurred at the
mixing pump, well-throttled valves, tee-branches, etc., where air was able
to separate out-a similar type of phenomenon to cavitation but dependent
more on instability of the flow than development of a very low static
pressure.
Some suppression of these surges was found possible by using perforated
orifice plates in the flow system after bends and valves where separation
of flow was likely to occur. Also flexible couplings and a tank to absorb
the surges immediately in front of the breast box should prove beneficial.
The question of using a special surge tank has also been considered by
Baines (34) who approached the subject mathematically and concluded
that flow surges should be considerably reduced in magnitude provided
the natural period of oscillation of the tank were much greater than that
of the surge.
lB . 8 2 Vibrations in pressure
Apart from oscillations set up due to surges in the flow, vibrations occur
in the system which are transmitted as pressure waves in a manner similar
to the transmission of sound vibrations in air. The essential difference
between the vibrational type of disturbance and the flow surge is that in
the former case the pressure can be affected both upstream and down-
stream, whereas a surge affects only downstream of the source of dis-
turbance. Pressure oscillations of this vibrational type are not affected by
air in the stock and originate primarily from mechanical vibrations in
the system.
Besides Mardon, several other workers, including Cuffey (30, 70) and
Reitzel (42), have investigated this type of oscillation in an endeavour to
track down the regular and rapid fluctuations in substance of the paper
which are produced as a direct result and which, when severe, are shown up
as 'barring' on the wire. Some pressure vibrations are initiated at the mixing
pump but according to Mardon these tend to die out as the flow progresses
through the system. Vibrating screens are a particularly frequent source
of pressure oscillation and if they get into synchronization can be so
bad that stock on the wire shows a most prominent barring effect; it may
be necessary to take specific steps to prevent this happening by coupling
the screens out-of-phase instead of using separate motors. In a similar
way the impellers in enclosed screens, at least under certain conditions
64
and at certain speeds of rotation, can create pressure pulses and in many
installations have been found to produce a cyclic pattern of the same
frequency in the substance of the paper.
Cuffey has emphasized how difficult it is to trace the source of these
vibrations and has found that they can even be transmitted from one
machine to another. Forming boards and table rolls may set up a resonance
which affects the substance while in some cases the whole breast box and
slice lips may vibrate; other sources of vibration which have been observed
to affect the substance are the shake and rotating machinery of different
types. In association with Ingram (70) Cuffey described one investigation
in some detail and it is apparent from this work that the process of elimi-
nating vibrations causing flow disturbances is a very complex task.
Furthermore there does not appear to be an easy means of easing their
severity if the source cannot be traced; according to Baines (34) a surge
tank would reduce oscillations of this type, particularly a shallow one with
a large area, but this is not a practical structure for most machines and
would not prevent vibrations being induced in the breast box itself.
A slightly different approach to the problem of detecting and tracing
the source of cyclic variations in substance, caused not only by vibrations
but also from other sources, has been developed by the British Paper and
Board Research Association. This technique involves the measurement of
several variables on the machine and simultaneous recording on a multi-
channel galvanometer. High-frequency substance variations beyond the
sensitivity of a beta-ray gauge on the machine are measured using a
specially-developed instrument which can sense changes in optical density
of the web. Subsequent analysis of the different records enables relation-
ships between the variables to be examined and, in particular, any cycles
occurring can be evaluated and correlated. Apart from its use in isolating
the sources of substance oscillations this method is, of course, valuable for
general investigations to discover how one part of a system affects another.
65
CHAPTER lC
RUNNING THE WET-END FLOW SYSTEM
lC.l DAILY OPERATION

As with all parts of the paper machine, the problem of keeping adequate
control on the performance and general efficiency with which the wet-end
flow system operates may be considered from two broad aspects. First,
it is necessary to look at the question from the standpoint of day-to-day
working. What measurements are essential and what measurements can be
regarded as important or simply as useful for the machineman to have in
order that he can best keep a check on the running conditions; what
variables should be controlled whenever possible? Second, the problem of
longer-term operation must be dealt with. What maintenance and opera-
tional checks appear desirable to ensure that general conditions in the flow
system do not alter over a period of time to the detriment of running
efficiency?
Each of these aspects will now be considered in detail. The question of
substance control, both along and across the machine, is such an important
subject that for convenience it will be treated separately. Finally, there is
a section dealing with practical aspects of running the wet-end flow system.
1 C . 1 1 Essential measurements
The formation and general quality of the paper cannot be expected to
keep uniform over a period unless the making conditions on the wire are
the same. The relation between the wire and the slice jet velocities, together
with the quantity of backwater diluting the fresh stuff, are both most
important factors in determining the conditions of formation and drainage
on the wire, at least on all but very slow paper machines. It is, therefore,
essential to have a measure of a property related to the slice jet velocity,
preferably the pressure behind the slice, and of the wire speed and the
width of opening of the slice, each of which govern these factors.
Pressure behind the slice is conveniently measured from an appropriate
water or air-purged tapping, or a diaphragm at the side of the breast box.
It is not particularly important that an absolute or even a completely
representative value of the pressure is obtained, so long as the measure-
ment follows closely with the overall effective pressure behind the slice.
Alternatively, in open breast boxes it is equally satisfactory to use the
surface level in the box to give a measure of the head above the slice.
Pressure in the box itself will be greater than the pressure head behind the
slice but so long as flow rates do not vary considerably for any particular
head, i.e. a wide range of slice opening is not used, then the relatively small
head difference due to pressure loss in the breast box fittings will not vary
significantly.
66
THE WET-END FLOW SYSTEM lC.11
Wire speed can be measured from a special tachometer on the driving
roll or couch. On all machines the speed will be measured at some point,
but this is usually either at the main driving section in the cylinders or at
the reel-up. The wire speed may be inferred from a measurement off some
other part of the machine provided the draw difference does not alter
to any extent; on many machines, however, draw differences can vary
the equivalent of several per cent. of the machine speed and this will
render the measurement too inaccurate to serve for the purpose of com-
parison with the slice jet speed.
In practice it is the ratio or difference between the slice jet velocity and
the wire speed which is important so a precise determination of either
figure is unnecessary so long as the measurement bears an accurate and
consistent relation to the true value. The presentation of the two appro-
priate figures to the machineman has already been discussed in lA. 1 5
and it is normally preferable to work in terms of velocities rather than
heads, i.e. to transform the slice pressure or head determination into an
equivalent velocity (for which the simple conversion formula is adequate).
The slice pressure conversion can be performed automatically in the same
way in which differential pressure measurements are converted to give a
flow reading, and the resulting velocity can be conveniently displayed on
a recorder along with the wire velocity. Even more satisfactory, there is
no difficulty in obtaining a differential signal to allow the velocity difference
or the ratio between the velocities to be shown on a separate digital
indicator or recorder (84); this enables the machineman to see at a glance
the effect of any unsuspected or deliberate change in operating conditions.
The positions of two pens giving a double trace on a single recorder do not
show small differences very clearly and use of a differential record in this
way will improve the value of the measurements appreciably.
Usually some form of measurement of the width of opening of the slice
can be obtained from the lifting mechanism. This is, unfortunately, not
usually very satisfactory. Backlash in the mechanical system and alteration
of the gap under pressure contribute to reducing the usefulness of what is
usually a relatively crude reading anyway. Ideally it would be preferable
to measure the opening of the slice lips directly and, if the bottom lip can
be assumed rigidly fixed as is usually the case, this effectively means
measuring the position of the top lip relative to a fixed datum point by
means of a suitable micrometer. It is also useful to have a measure at
several points across the slice, even opposite every jackscrew, of the extent
to which the slice is being distorted relative to the top lip as a whole. In
many ways this latter facility is of greater importance than a single average
reading for the slice opening because it assists in attaining cross-web
substance uniformity. The basic reason for measuring the average slice
opening is to assess the volume flow of stock since, for a given head, this
depends entirely on the slice opening. However, as will be mentioned
shortly, it is also possible to do this by taking a direct flow measurement
in the approach system.
Finally, in the category of essential information can be included several
pressure measurements in the system. Gauges are generally adequate for
67
lC. l 2 THE WET-END FLOW SYSTEM
this purpose and siting depends entirely on the organization of the wet-end.
Familiar positions in the approach system will include the mixing pump
discharge (this may be a combined pressure/suction gauge when priming
is done by application of vacuum) and before and after cleaning equip-
ment; also, of course, pressure gauges will be available on such positions
as the discharge side of the spray pump and excess whitewater pump.
These gauges are often seen in poor condition, especially where the system
is subject to considerable hydraulic vibration and the sensing element
is not damped sufficiently; perhaps more than anything else it is the sign
of an efficient instrument department and attentive machine crews when
they are kept well serviced and regularly checked.
1C. 1 2 Important measurements
The measurements listed above are all essential because upon them depends
the machineman's ability to set conditions at the wet-end accurately and
consistently both during a making and from one making to the next.
There are a number of other measurements which assist the machineman
to keep a check on the wet-end conditions and these may be loosely
divided into important or simply useful in value. Important measurements
are of breast box stock volume flow, temperature, pH and the flow in the
various chemical and steam supply lines.
The volume flow of stock to the breast box is unfortunately not easily
determined and this measurement is rarely seen on a paper machine,
though there has been at least one report of its application (33). Yet as
a complement to the rough slice opening indication it has considerable
potential value. Under particular slice pressure and wire speed conditions
(and assuming, of course, that the fresh stuff flow and consistency are
adequately regulated), the breast box flow corresponds precisely to the
volume of backwater being used for dilution and hence will give both an
accurate idea of drainage conditions on the wire and permit the machine-
man to reproduce previous conditions more exactly. The backwater
dilution flow at start-up is frequently set in the first place by comparing the
value in previous makings of such things as the power used by the mixing
pump, the gate opening in a simple mixing box, and the position of the main
throttling valve; afterwards the position of the dry line on the wire and the
vacuum on the suction boxes are the main criteria for assessing adjustments
to the quantity of dilution water required and it is by no means easy from
these alone to ensure that repeatable making conditions are obtained.
For one thing the volume of dilution water can be increased considerably
(keeping the slice pressure constant), thereby lowering appreciably the
consistency of stock in the breast box, yet most of the extra water drains
through the wire at the table rolls and the position of the dry-line and
suction box vacuum may alter only very slightly. The formation and
making conditions on the wire may have altered completely at the lower
consistency yet there will be nothing in the general running of the machine
at the wet-end to indicate this. Thus, used intelligently, a measurement of
the breast box flow would be very helpful.
It is possible to measure this flow adequately by means of differential
68
THE WET-END FLOW SYSTEM lC.1 2
pressure readings across a normal orifice plate with filtered water purging
of the tappings. When there is fear of the usual concentric orifice clogging,
a segmental orifice open only in the bottom half or a flow nozzle can be
used. The reading need not be absolutely correct, nor accurately calibrated,
so long as it responds with adequate sensitivity to the sort of changes in
flow met with in practice. Some pressure loss will occur across an orifice
plate or flow nozzle but this would necessitate an important increase in
power only if the mixing pump and throttling valve were at all times closely
regulated to use as little power as possible (in most cases they are not).
A venturi meter has a much lower pressure loss than either an orifice plate
or flow nozzle but, particularly for larger machines, would become very
large to cope with the flows involved and would need an extremely lengthy
straight run of piping. The venturi meter is also said to be more sensitive
to viscosity changes though this is more relevant for thick stock. The
alternative, though expensive, is to use a magnetic flow meter; it would
be more important in this case to ensure that variation in aeration of the
stock did not have a significant effect on the reading though the degree of
turbulence does not affect this instrument as it may with other types.
Whatever the method chosen, to obtain full value a recorder for the flow
would be necessary. With the tapered side inlet approach system the over-
flow would have to be taken into account and a second flow measurement
(shown on the same recorder for convenience) becomes necessary; this
adds to the complexity and expense of obtaining satisfactory breast box
flow measurement.
The temperature of the stock, measured most conveniently in the pit,
is an important reading to have available primarily because of its influence
on drainage rate on the wire. It is becoming more and more usual to heat
stock, either at the machine or in the preparation system, to obtain the
advantage of quicker drainage, and in this event the temperature is of
greater importance. It is easy to allow temperatures to get higher because
this may make the wire conditions less critical but with higher tempera-
tures heat losses are greater and a point must be reached where this
procedure, apart from being uncomfortable for working in, becomes
uneconomic. Also, at start-up it is important to get the temperature up to
normal as quickly as possible. For both these reasons a temperature
measurement, shown either on a simple indicator or preferably linked to
a recorder, is most useful.
On many machines the heat capacity of the system is sufficiently great
for temperatures to remain reasonably steady once equilibrium has been
reached. In this case some steam may be applied through closed coils in
the main wire or backwater pit at the start-up but not afterwards. How-
ever, if steam is in constant use for keeping up the temperature of the
wet-end stock, the flow of steam possibly being automatically controlled
from the temperature reading, then it is useful in addition to have a
measure of the total steam used during any desired period such as a
making. A straightforward record of the steam consumed over comparable
periods will enable a simple check on seasonal and long-term variations
to be kept and, in particular, excessive use of steam will be readily observed.
69
lC.1 3 THE WET-END FLOW SYSTEM
A knowledge of the pH of the wet-end flow system is important for

many reasons that have already been discussed in some detail. In par-
ticular, maintaining the pH steady contributes enormously to keeping
pitch trouble in bounds. pH may be measured at a number of points in
the system, commonly in the main wire or backwater pit, the breast box
or strainers, and the reading is normally recorded. For this instrument
particularly it is highly important that adequate maintenance is applied
to the electrodes, especially when rosin is used in the system as this tends
to coat them. Electrodes may need to be cleaned and checked every day
or even once a shift if the reading is not to wander significantly.
Associated with a pH meter in the wet-end flow system it is usual to
have a liquid alum feed fairly close upstream to the point of pH measure-
ment, possibly in the mixing pump inlet or the backwater pit. This may be
the only source of alum in the system or it may be used solely for a final and
relatively rapid trimming of the pH. The alum feed is often set manually
unless the system has a wide variability, in which case automatic control
from the pH recorder may be justified. Either way, it is extremely useful
to have a measure of the total quantity of alum used during each making
for both operational and accounting purposes, and also possibly to have
a record of the alum flow. Alum must of course be provided at a controlled
uniform density. Especially when running a low pH the consumption of
alum for small reductions in the pH increases rapidly and can rise to
phenomenal quantities if careful inspection does not take place at regular
intervals. A variable stroke metering pump for delivery of the alum or a
rotameter arrangement coupled to a constant flow control are probably
the best devices for regulating flow, though care is necessary to avoid
blockage of the line. Alternatively a small magnetic flowmeter can be
utilized.
Other additives to the wet-end such as wet-strengthening resins, loadings,
dyes etc., can be treated in the same way as the alum, and metering pumps,
rotameters or flowmeters can be used to advantage for measuring the
flow in each of the supply lines and controlling each to a pre-set proportion
relative to the fresh-stuff flow. Probably for these additives an indication
of the flow with an integrating meter is adequate. Fresh water entering
the machine system should be metered, and in this case a recorder is
useful for showing up instabilities in the system. The same applies to
whitewater effluent for which an integrated reading in conjunction with
an average consistency figure (this can be obtained using a special device
which samples the flow at regular intervals) is valuable for determination
of fibre losses from the flow system.
lC .1 3 Useful measurements
Apart from the more important measurements mentioned above there
are three other readings that can provide useful information to help run
the wet-end systematically: these are the consistency of the breast box
stock, the volume flow of the fresh stuff, and the flows of fresh water or
whitewater at points where they may affect the system balance, such as in
cleaners, breast box, or wire sprays mixing with the main backwater.
70
The consistency of the breast box stock has been referred to in some
detail and it has been shown that the main backwater circuit settles at an
equilibrium consistency which depends essentially on the retention and
drainage conditions on the wire. Used preferably in conjunction with a
measurement of stock flow to the breast box, this consistency reading
would be very helpful for tracing unsuspected disturbances in the system.
Any alteration to any of the wet-end conditions-the pH, temperature,
loading additions, cleaner reject flow, table roll assembly, forming board
position, etc.-may have an effect on drainage conditions, and this would
be immediately apparent from a breast box stock consistency record since
the consistency will vary with change in any of the wet-end input or output
flows or consistencies, or in the retention on the wire.
Though it would require considerable experience to interpret correctly
fluctuations in breast box consistency, the author feels confident that an
accurate and consistent measurement of this property would be of im-
measurable value in trouble-shooting at the wet-end. Unfortunately at
the present time no really satisfactory instrument is available which is not
sensitive to numerous other changes in conditions. Several approaches
are being made similar to those discussed in lB. 2 1, but the problem with
consistencies under 1 per cent. is even greater than for fresh stuff. Other
methods such as use of a device for detecting small changes in optical
properties have been tried but have so far not met with much success.
This is one problem that warrants intensive research and it is to be hoped
that a suitable instrument will become available soon. With an on-line
computer the breast box consistency can be calculated by means of a
material balance based on data from the fresh stuff and excess whitewater
flows together with the substance of the paper. This provides an alterna-
tive approach which is very useful.
Measurement of the flow of fresh stuff, presumed to be consistency
controlled, can be particularly useful for comparison with a continuous
record of the substance of the paper. Also it is possible to see whether
changes in certain wet-end conditions, e.g. of the backwater dilution flow,
can effect the flow of fresh stuff. This flow measurement poses the same
sort of problems on a smaller scale as measurement of the breast box stock
flow although, being at a relatively high consistency, in this case the mag-
netic flowmeter is almost certainly the most suitable device. This particular
instrument is more likely to be installed when the fresh stuff line runs
direct from the stuff chest to the mixing pump; the flow rate will then be
automatically controlled and the flowmeter becomes an integral part of
the system, not simply a measuring device, and so is easier to justify.
Apart from providing a straightforward record of the fresh stuff flow,
the flowmeter permits small alterations to the substance to be made with
considerably greater accuracy than can normally be achieved simply by
altering the main or by-pass stuff valve position based on experience;
substance changes of the same grade of paper should also be made more
exactly. Other applications involve use of the flow reading for control
purposes and these will be considered in more detail in the next section.
A flow measurement could be applied with some value to the excess
71
tC.14 THE WET-END FLOW SYSTEM
whitewater flow. It is hardly possible to specify this in detail because its

location would depend essentially on the machine set-up. There will
always be a certain quantity of whitewater flowing out of the system,
even if only to drain, and a record of this could be of assistance in charac-
terizing the stability of the machine and in trouble-shooting. As already
discussed in some detail, any change in input or output from the wet-end
will show up in the excess whitewater flow and even if this is measured
simply by recording flow over a weir or flume it could be of great potential
service.
Finally, measurement of the flow of all important fresh and whitewater
lines is almost as important as incorporating pressure gauges. Indicating
rotameter-type instruments can generally be adapted for this purpose and
at the usual consistencies of whitewater lines no serious difficulty with
longer fibres sticking to the rotameter float should be encountered. Making
up or wear of sprays can affect the flow through them considerably, even
for the same pressure gauge setting, and when the spray is invisible, as
for instance in some closed breast boxes or a suction roll, this can become a
source of trouble that is difficult to trace. Small rotameters are also advis-
able on pump sealing gland lines where a worn gland can admit a surprising
quantity of fresh water to the system. If it is not considered necessary to
install rotameters on individual fresh water lines at least the total flow of
fresh water to wire guide, dandy spray, roll sprays, cutter jets, suction box
seals, etc. should be measured or checked at intervals.
1C. 1 4 Control applications
There are a number of control applications which can with advantage be
installed in the wet-end flow system. By far the most important is the level
in the breast box or effective head at the slice but other applications such
as control of the fresh stuff flow, pit levels, temperature, pH, etc. will be
briefly mentioned.
The level of stock in the breast box or the total head at the slice in an
air-loaded box regulates the pressure behind the slice and hence the
velocity of the slice jet, so it is vital that there should be as little variation
in this as possible. On slower machines the approach system may be
sufficiently stable for a straightforward stock flow control valve, once set,
to keep the slice head adequately steady. But on faster machines it is
generally advisable to control the head. This is done most efficiently
by a three-term control on the main flow valve: the integral action is
necessary because flow demand or supply pressure changes produce offset
in the control operation, i.e. the head is then actually controlled at a value
different to that required; the derivative action provides for a rapid
reaction to disturbances in the system and reduces the time needed to effect
a correction.
Instead of applying control to the main stock flow valve, which becomes
very large on fast machines, there are two alternative systems of control
that can be used. These are particularly useful in conjunction with an a.c.
mixing pump motor when the flexibility of operation provided by a d.c.
motor is absent. Firstly, control can be applied to a valve on a relatively
72
small recirculation line round the mixing pump (normally with such an
arrangement the recirculation would not return directly to the suction
side of the pump but to the backwater pit or silo). With this system the
main flow valve opening is rarely altered, indeed a straight-forward air-
operated gate valve which is either fully open or closed is sometimes used,
and the smaller control valve on the recirculation line provides a more
sensitive means of regulating flow to the breast box. The other control
system that can be used involves a by-pass line, the main stock pipeline
being divided in two, one large branch containing a hand valve and the
other smaller branch in parallel containing the control valve. This too has
the advantage of providing a more sensitive control on flow to the breast
box than a single valve in the main stock line. However, Hasu (81), in an
analysis of the performance of these two systems, has shown that the
sensitivity of control achieved with a recirculation valve arrangement is
much less dependent on the total flow than when control is on a ·by-pass
valve, especially when the pump curve is steep as is usually the case with
mixing pumps. The recirculation method thus appears preferable.
It is most important that the position of the flow control valve is
recorded, usually on the same chart as the level or slice head, because this
trace gives more evidence of what the controller is doing than the near
straight line that will probably be drawn for the head measurement when
the controller is properly adjusted. With such a controller if the speed of
response is adequate it is only necessary at start-up to set the head required
to give the desired flow velocity from the slice and this is then reached
automatically provided the mixing pump gives sufficient pressure.
In some cases this control of slice head in the breast box is elaborated
and the set-point of the controller is arranged to be governed by the speed
of the wire (following the theoretical relationship between wire and slice
jet velocity). The machineman sets the desired relationship between the
wire speed and head at the slice in a similar way to that discussed earlier
for manual operation, and the controller then takes care of any changes
in wire speed; if the machine is speeded up a rise in head will automatically
follow. Eastwood and Gade (31) have described such an application in
which a pneumatic speed measurement obtained from the couch shaft
is squared and the signal used to set the control point of the breast box
level; a small positive or negative bias is set by the machineman and a
manual cut-out is available if necessary. The authors state that this arrange-
ment has proved most helpful at start-up by reducing breaks and spoilt
paper caused when increasing the speed. However, in the opinion of the
author this will normally be unnecessarily elaborate; it cannot be easy to
justify such a control solely for speed changes since it is not difficult for
the machineman, in the course of making all the other alterations usually
necessary, to reset the breast box level control point to suit the new speed
according to an appropriate chart or graph. (Even when this operation is
done automatically it may be necessary anyway to alter the effiux ratio).
It could perhaps be argued that the effect on the substance of machine
speed fluctuations due to frequency changes or other disturbances could
be reduced by using such a control; however, even if such fluctuations
73
lC.14 THE WET-END FLOW SYSTEM
could be followed quickly enough for their effect to be reduced it would

probably be more sensible to try first to eliminate these at source if they
are definitely known to be detrimental to the machine operation.
One particular advantage of the head controller is that alterations to
the volume of backwater used for dilution can be accomplished in the
correct way, i.e. without changing the slice jet velocity. If, for example,
it is desired to increase the amount of dilution water, the slice can be
slowly opened up. As the gap widens so the head in the breast box will
fall but the controller will correct this automatically by opening up the
flow control valve to preserve the level. Provided the operation of the slice
mechanism does not affect the cross-level (as it so often does), this is the
ideal way of regulating the amount of dilution water used on the wire.
The level in the stuff box may only be controlled in the sense that an
overflow is provided, though a level controller may in some cases be used
to throttle the supply line from the fresh stuff chest and this has the
advantage of avoiding an overflow. With either a stuff box feed to the
mixing pump or a direct line from the stuff chest it is possible to have a
control on the flow of fresh stuff by using a fl.owmeter to regulate the main
valve. Not only should the flow be shown on a recorder but also the
position of the stuff control valve. The set-point of this controller would
then govern the substance of the paper and allow a fine control. Provision
of an interlock to close the stuff valve if the mixing pump stops or the
dump valve is opened is an easy but valuable accessory, though a preferable
method which avoids unnecessary alteration of the fresh stuff valve is to
incorporate a second cut-off valve for this purpose.
Paterson and Hazeldine (41) have described an elaboration of this simple
technique of fresh stuff flow control in which a signal from a consistency
regulator is used in conjunction with the flowmeter signal to control the
stuff valve; small deviations of flow or consistency from the desired set-
points are added together to give proportional control of the valve and
this to a first order of change effectively controls the product of the flow
and consistency, i.e. the flow of fibre to the machine. Even further, a speed
signal from the machine sets the datum line for fibre quantity, but in the
author's opinion this particular refinement is unnecessary for the same
reasons given above with reference to speed control of the slice head.
Though the idea behind control of the product of flow and consistency is an
obvious one, one would like to know whether in fact the short-term sub-
stance fluctuations were less than would be the case with a simple flow
controller. Certainly the fibre quantity as determined by this arrangement
could not be used as an absolute setting for paper of any given substance
and a manual override adjusted from substance tests at the dry-end of the
machine is an integral part of the set-up. Bearing in mind the shortcomings
of consistency regulators the author considers that this system would
require careful evaluation with a continuous substance measurement at the
dry-end and it may even be found that the short-term weight keeping has
worsened when compared to a straightforward flow control system. An
arran~ement described by Moore and Walsh (66) in which the flowmeter
set-point is adjusted by a consistency regulator signal to achieve constant
74
fibre flow is open to a similar criticism; in their case the consistency regu-
lator is checked hourly against bone-dry determinations and the set-point
altered if necessary, but it could be argued that this will be less effective than
altering the set-point from determination of the substance of the paper
since, after all, the whole point of the exercise is to help keep the dry weight
steady. These ideas will be considered in further detail in 1C. 3.
Control of the slice opening from position of the dry-line on the wire
has been reported by Tousignant and Madgett (79). In this arrangement the
position of the dry-line is detected by measuring the flow of backwater
from the suction box situated close to the usual position of the line.
Changes in this flow are then used to control small movements of the slice
at timed intervals (the breast box head presumably being automatically
controlled). Two forms of control have been attempted, a direct one with
backwater flow controlling slice opening and a more refined system in
which the overall vacuum on the suction boxes is first controlled within
limits before making slice adjustments. Such a control is theoretically
sound, and in fact achieves automatically the important operation of
altering the backwater dilution flow according to drainage rate on the
wire which is one of the most important functions of the machineman.
But such elaboration would only seem justifiable if the dry-line were to
fluctuate appreciably and as this would be symptomatic of considerable
variations in stock preparation and wire drainage conditions it should be
more profitable to seek these out and correct them. Further, a positive
argument against using such a control is that it is liable to upset the flow
conditions in the backwater circuit, disturbing equilibrium and substance
of the paper to a greater extent than exists naturally. It may well be
preferable from the point of view of maintaining overall dry weight
stability to permit reasonable fluctuations in dry-line position, requiring the
machineman to make only relatively larger, but much less frequent, adjust-
ments to the slice opening when a definite shift in drainage rate has
occurred.
Other control applications have already been mentioned in the approp-
riate place and little more need be said about them. Control of the main
pit level by regulating the addition of suction box backwater is a simple
task and may be preferable to the customary overflow method when it is
desired to keep the consistency of the excess flow as low and as steady as
possible. Neither a recorder nor indicator should really be necessary unless
evident fluctuations occur that require tracing. The excess whitewater
pumped from the machine wet-end to a save-all must be flow-controlled
to keep the collecting pit level steady, and this particular flow can with
value be recorded, even if only indirectly from flow-control valve move-
ment, in order to show up instabilities in the whole wet-end system. The
level of pits receiving broke at the wet and dry ends of the machine can also
be controlled; in this case the level is normally governed either by varying
the volume of broke recirculating round the extraction pump, or by having
a separate large pump which comes into operation only when the level
rises above a certain point (as during a break). Consistency of broke leaving
the pits is regulated by varying the addition of whitewater.
75
Temperature of the stock can be controlled from the steam flow to

heating coils in the main pit either continuously or just at start-up, and
pH in a similar way from the alum flow. Though variation in pH is
frequently small enough during normal running to remove the necessity
of a controller, it should always be remembered that closer control may
make it possible to run the wet-end at a higher pH and thereby reduce
alum consumption appreciably. If the upstream pressure varies, addition
of other chemicals can also be flow-controlled when necessary and in some
cases it may be justifiable to control the flow of such chemicals propor-
tionally to the main flow of fresh stuff. For instance, Van Derveer (68) has
described a system in which the addition of loading, colour, and Sveen
glue are automatically controlled into the mixing pump inlet from a signal
received from a master magnetic flowmeter measuring the total stuff to
the machine.
lC.2 MAINTENANCE
Maintenance of the wet-end flow system, so far as the papermaker is
concerned, amounts almost entirely to one thing, keeping the system clean.
The only mechanical parts which require routine inspection, the mixing
pump, valves, breast box equipment, and slice, are primarily of engineering
rather than papermaking responsibility though the demarcation line varies
from one mill to the next. Setting the slice and checking the condition of
the all-important slice lips are definitely the responsibility of the paper-
making department, though how often this is done thoroughly, possibly
by checking the opening with a taper gauge, depends on how sensitive
the cross-machine weight profile is and how frequently the jackscrews are
adjusted. Also occasionally the position of the lower slice lip relative to the
wire should be checked for evenness across the machine in order to show
up any long-term deflection or wear of the lip and breast roll.
Regarding cleaning it is only possible to say that each week all pipework,
boxes, pits, etc. should be thoroughly hosed down with high-pressure jets;
in addition, periodically the whole system will require a thorough clean-out
using heated water to which caustic soda or a chemical compound specially
designed for the purpose is added. The frequency with which this is done
depends entirely on mill conditions and, in particular, on the water used.
When scale formation occurs it is often difficult to remove chemically
without affecting some metal parts and if wires become coated scouring
with acid shortens their life; thorough and regular cleaning with high-
pressure jets is very important in such conditions.
Apart from general cleaning there are two related subjects which may
conveniently be treated under the heading of maintenance though they
have just as much relevance to day-to-day operation. These are the preven-
tion of slime and pitch formation.
1C . 2 1 Prevention of slime
In many mills, particularly those using groundwood or waste paper, keep-
ing down accumulations of slime in the system can be quite a problem.
76
The most common treatment is to add shock dosages of a slimicide,
usually an organomercurial such as phenyl mercuric acetate and ethyl
mercuric phosphate, or a chlorinated phenol such as phenylphenol,
pentachlorophenol and their salts. The quantity added to the system and
frequency of dose depends usually on the recommendations of the manu-
facturer and may be quite an expensive business.
The value of these slimicides has always been hotly debated and it is
only comparatively recently that any reliable evidence has become avail-
able. The most painstaking examination of this problem has been given
in a book by Rathman and reviewed by Russell (73). In carefully controlled
experiments extending over a period of eight years Rathman obtained no
evidence at all that the use of disinfectants gave a demonstrable improve-
ment in the slime situation; in a mill in which he worked neither the degree
of infection nor the number of breaks caused by slime decreased with
greater dosages. The effect of a slimicide was only apparent while an
effective concentration remained in the system and in practice the concen-
tration from a shock dose diminished so rapidly that the number of organ-
isms was rapidly restored. Rathman concluded that it would be too costly to
maintain an effective dose and that shock-dosing has little to commend it.
Another interesting fact that emerged from this work was that seasonal
variations in temperature of the system were shown to cause a statistically
significant effect on the degree of infection. The degree of infection appeared
to rise with increases of temperature of the backwater up to a temperature
of about 65 deg. C., above which it began to fall. However, Rathman
emphasized the difficulty of isolating a phenomenon of this sort using
spot checks and it would not be easy for a mill to study the effects of
different conditions on the slime in the system; even so when this is a
serious problem it may warrant the making of careful observations, possibly
with specially designed containers with which the growth of slime can be
more accurately assessed.
Slime forms in the presence of air when bacteria and fungi in the system
grow on fibre and alumina debris collected in such places as the stock
surface level in boxes, particularly when they are made of wood or con-
crete. It is probably worsened when the air content of the stock is high and
certainly deaerated stock helps considerably to prevent slime becoming
troublesome on the machine because fibre is then less likely to adhere to
the slime already deposited. Long fibres, loadings and sticky pitch con-
stituents all worsen the slime situation by speeding up growth of the
accretion, and corrosion is more rapid where slime is deposited.
The most effective method of preventing slime is undoubtedly good
housekeeping. Efficient wash-ups and the pumping round of hot alkali,
possibly with a slimicide added, will remove existing slime deposits and
prevent a slime growth over the shut period which may break away when
the machine is started again. If this needs to be done more frequently than
the machine is normally shut the system should be arranged so that cleaning
is possible throughout the wet-end, including the stuff box and stuff pipe
system, without drawing water from the machine chest. The suction box
and whitewater pipes on all machines are particularly liable to slime growth
77
and special provision may be necessary to keep this part of the system
adequately clean.
1C.2 2 Prevention of pitch

The effects of pitch on the paper machine have much in common with those
of slime and it is often difficult with small dispersed deposits on the wire
to tell the difference between the two without a simple laboratory check.
As in the case of slime, when deposited earlier in the stock system pitch
is liable to break away in lumps causing holes and breaks in the paper.
In addition pitch is more liable to affect the presses and the wire part of
the machine. On the presses pitch gathers under the doctor blades and may
break away in lumps and affect the sheet; on the wire it gradually reduces
drainage, producing light spots in the paper, and may also collect on the
dandy, table rolls, forming boards, and suction boxes (on the latter it can
accumulate sufficiently to raise the wire and cause loss of vacuum).
Although pulps can be checked in various ways for their propensity to
deposit pitch, there are other sources of pitch-like substances in the mill
and in the long run the only reliable information comes from keeping
accurate records of its occurrence on the machine. Deposition of pitch
is affected by the pH and temperature of the backwater and especially in
a closed system careful control of these helps to keep the free pitch particles
dispersed and encourage their precipitation onto fibres; this reduces the
tendency towards agglomeration of the particles in the system where a
slight disturbance to the fine state of equilibrium could lead to a sudden
heavy deposition. Often opening up the whitewater system slightly appears
to help the situation by making the equilibrium less sensitive and reducing
the concentration in the system.
Attempts are sometimes made to control pitch by additions of loadings
such as diatomaceous earth and hydrous magnesium silicates, which
provide a suitable surface on to which the pitch particles may deposit,
and also by using calgon and detergents to act as dispersing agents to
maintain the pitch in a colloidal state. But apart from other disadvantages
these methods are often rather expensive for continual use though,
undoubtedly, some improvement can be obtained. There are also mechan-
ical methods of pitch removal including a series of metal baffles placed in
the thick stuff after beating when the dissolved pitch content is greatest.
Unfortunately, however, this whole problem is far from being satisfactorily
solved and in every mill the appearance of pitch on the machine is greeted
with much anxiety.
lC.2 3 Long-term records

The importance of obtaining long-term records to act as standards with
which to compare conditions during trouble-shooting or when making a
deliberate alteration to the machine is fully discussed in Part VI; this
procedure is especially important for the wet-end flow system. During
normal runs of one or more of the major grades of paper produced on the
78
machine, when equilibrium has been reasonably established the whole
system should be carefully analyzed in addition to noting the readings of
all measurements already displayed on the machine. Flows, consistencies,
loading content, fibre classification tests, strength and porosity measure-
ments on standard handsheets, and also air content determinations should
be performed on samples taken at all the main points including the back-
water pit, fresh stuff, cleaner and screen rejects, breast box stock, suction
box backwater, and whitewater excess. In addition, any of the properties
already discussed which are not measured on a routine or continuous
basis, such as the slice opening, pH, temperature, alum and other chemical
addition rate, etc., should be checked.
The measurement of flows at the wet-end depends essentially on the
accessibility and quantity involved. Smaller flows up to 30 or 40 gallons
per minute are conveniently measured, either by the simple bucket and
stop watch or with a portable container coupled to a rotameter (the latter
is more suitable where short-term variation in flow occurs and is less
influenced by the operator, but normally requires more room). Enclosed
flows can often be determined with reasonable accuracy by introduction
of a concentrated salt solution at a known steady rate and measurement
of the diluted concentration in samples drawn downstream; some care is
required in general technique and the method of chemical titration, how-
ever, especially when the flow is part of a closed circuit. But it is not usually
possible to determine all the flows, particularly those for the fresh stuff,
the main backwater circuit, and the whitewater excess flow, when special
provision has not been made.
Comparing consistency results for the input of fresh stuff and backwater
with the output from the mixing pump or box normally enables the main
large flows to be assessed with fair accuracy. At least one flow must be
known for this method to be applied and the fresh stuff is usually the one
most easily measured. If no flow determination is possible then one must
be calculated; this may be either the breast box flow determined from the
slice pressure and opening (made as accurate as possible by using the exact
formulae quoted earlier) or the fresh stuff flow determined from the bone-
dry production, fresh stuff consistency and estimates of the loss of fibre
at cleaners and in excess whitewater. Both calculations have their limita-
tions and agreement of the main flows calculated in the two ways depends
essentially on how representative and accurate are the various measure-
ments; it should, however, be possible to get well within 5 per cent.
Comparison (by determining ash contents of consistency pads) ofloading
input and output at the same points provides a cross-check on the reliability
of the calculations and again agreement should be possible to within 5
per cent.; in a relatively stable system it is also quite possible, with careful
sampling and analysis, to obtain agreement in the quantity of long or short
fibres entering and leaving the mixing pump to within about 10 per cent.
by weight of the total fibre, despite the relative lack of precision of devices
like the Bauer McNett Classifier.
The Sankey diagram is a useful device for illustrating the flows of water,
fibre or loading throughout the system, but it is only really suitable for
79
general purposes of comparison with other machines or to examine the

outcome of altering the system. For comparison of conditions on the
same machine running the same grade of paper but taken on different
occasions the Sankey diagram is not sufficiently precise. It is preferable
for characterizing the backwater circuit to relate the slice opening with
breast box consistency and loading content on a graph and also to calculate
the retention of fibre and loading on the wire.
Measurements across the machine of the precise position of the slice
lips with regard to each other and to the breast roll are important for their
influence on formation and drainage on the wire. Tests for freeness, etc.,
and possibly also quality checks on handsheets made from the fresh stuff
will, of course, be useful, though it is presumed that the establishment
of long-term records extends to the preparation process and this informa-
tion will be more relevant to that section.
lC.3 CONTROL OF SUBSTANCE

The substance of paper is the most important single property and one of
the main concerns of the machineman is to ensure that it is correct.
Fluctuations in the substance from one roll to the next and within a roll
must be reduced to a minimum, and at the same time the substance at
different points across the sheet must be uniform. Since the wet-end flow
system and, in particular, the flow of fresh stuff to the machine is respon-
sible for determining the substance of the paper, it is appropriate to consider
this question now in some detail.
1C.3 1 General comments

The problem of maintaining correct substance is intimately tied up with
that of keeping the correct percentage of moisture in the sheet. The sub-
stance of paper is defined as the basis weight in the standard humidified
atmosphere; thus if a 100 gram per sq. metre paper has an equilibrium
moisture content of 10 per cent. but comes off the paper machine at 6
per cent., then the making weight should, strictly speaking, be 96 g.s.m.
The difficulty, of course, is that both the humidified moisture content and
that of the paper at the reel-up varies, so the actual making weight to be
aimed at differs continually. Ideally the dry weight of the paper should be
the standard and differences in humidified moisture content averaged and
kept as small as possible by making a product of consistent structure; in
the example above, for instance, the making weight would be 90 g.s.m.
bone-dry.
The traditional method of determining the substance at the reel-up is to
tear off a number of layers, trim the edges with a template and weigh on a
quadrant balance with a suitably calibrated scale. The shortcomings of
this operation are fairly obvious and the main limitation is, of course,
that the result of weighing a single sample, or at best the average of
two or more samples across the sheet, is taken as representative of the
whole roll though it is in fact only a minute part of that roll that happens
80
to have been made just before the roll reached maximum size. The problem
of differing degrees of moisture in the paper as made on the roll is partly
overcome by accident, since in the interval between extracting the sample
and weighing the paper, moisture is gained or, in a very few cases, lost as
the moisture level approaches equilibrium with the atmosphere. This
process occurs very rapidly (a sample of newsprint at 6 per cent. moisture
rises to 7 per cent. in less than a minute and will be within 0·5 per cent.
moisture of the final equilibrium value of between 9 per cent. and 10
per cent. after only 10 minutes or so). Also the rate of change is more
rapid the greater the initial difference from the final equilibrium moisture;
hence a considerable degree of equalization of moisture naturally occurs
before testing. Unfortunately, however, the equilibrium moisture of the
paper varies with humidity and temperature of the atmosphere and the
conditions in the machine house or in a laboratory close by where the
weighing has to be accomplished with reasonable speed may be very differ-
ent from the standard (65 per cent. humidity, 20 deg. C.).
Normally, however, these limitations to sample weighings rarely seem
to cause much concern, but the main reason is probably that standards of
uniformity and accuracy are relatively low. With a growing consciousness
of the advantages to be gained from keeping a tighter control on the sub-
stance, the lower margin of allowance necessary to be within specification
and the increased uniformity of other paper properties which occurs when
the substance is more uniform, it has become more important to improve
methods of substance control. The application of quality control has been
encouraged in post-war years and undoubtedly much improvement occurs
when such schemes are introduced because they serve to draw attention to
the importance of keeping the substance, and for that matter other proper-
ties, closer to standard; in practice, an improvement probably occurs more
often than not because the machineman is encouraged to leave the process
alone more and, in particular, to adjust the stuff valve less frequently.
The degree of success with which the substance is regulated depends in
large measure on the sensitivity of the stuff valve. Unfortunately it is
usually impossible to calibrate the average valve satisfactorily because the
relatively small alterations in flow required for adequate control of the
substance are lost in the backlash of the valve movement. A valve positioner
is helpful but this degree of elaboration will normally be thought of only in
connection with some form of automatic control from a beta-ray gauge.
An alternative which can work very successfully is to utilize a small
parallel flow of fresh stuff additional to the main one which is then only
altered for large substance changes; this parallel flow can be altered
visually in a mixing box system or by a suitably calibrated valve regulating
a separate stuff line which goes either to the mixing pump or, preferably
and more easily, to the inlet of the pump in the main pit.
1C . 3 2 Continuous measurement of the substance

The lack of knowledge of what is happening to the substance in the middle
of a roll, except for what the experienced machineman can assess very
81
1C.3 2 THE WET-END FLOW SYSTEM
approximately from visual evidence of changes in the suction box vacuum

and draws, makes some form of on-machine continuous measurement of
tremendous importance. The beta ray gauge has succeeded in filling this
gap and despite the earlier limitations and inaccuracies of the technique
much development work has enabled the electronic stability and sensitivity
of the instrument to be greatly increased. For recording changes in sub-
stance of the paper over a short period the beta-ray gauge is now extremely
valuable, though for long-term regulation of the substance certain pre-
cautions in its use are necessary.
In the first place it is important to realize that a really exact measurement
of the substance, in the strictly defined sense of the term, is virtually impos-
sible. Even when inaccuracies in the instrument reading due to temperature
variation in the air gap, static electricity, different scattering coefficients
produced by a change in formation or in the pass-line of the sheet between
the source and ionization chamber, and several other minor factors are
eliminated or allowed for, the reading is still of paper in a variable state
different from what it will stabilize at in standard atmospheric conditions.
The paper is stretched, so the area presented for measurement contains
less fibre than a sheet of corresponding area drawn from the roll. But more
important, as the moisture content on the machine varies, so will the
apparent substance, and in contrast to off-machine measurement by
sampling there is in this case no equalization effect. This latter basic limita-
tion of the beta-ray gauge is satisfactorily removed only if the instrument
is used in conjunction with a moisture meter; in this case a measure of the
dry weight of the sheet becomes available and this will be far more valuable
for correct adjustment of the stuff valve.
Apart from the limitations imposed by the inferential nature of the
instrument reading and the condition of the paper, the fixed beta-ray gauge
measures only a portion of the width of the paper web so that another
effect must be considered, that of variation across the sheet. This subject
will be treated in more detail in lC. 3 4, but the point of importance
here is that if the reading is obtained from only one position across the
web there must be the possibility of a bias in comparison with the average
substance of the sheet unless the substance profile is exceptionally well
controlled. This can be overcome to some extent by using a moving beta-
ray gauge to give integrated readings taken from a series of rapid cross-
web traverses, but this procedure loses the advantage of continuity of
record in the machine direction which is so valuable an asset of the gauge.
Opinions differ over the best way of countering these difficulties and the
solution depends partly on the relative variability along and across the
web, and partly on the particular requirements and speed of the machine.
When the cross-web substance and moisture profile remains reasonably
stable over a long period, possibly the simplest effective method of overall
control would require continuous substance measurement on one portion
of the web, punctuated at intervals by a traverse to check the cross-web
substance level.
It does seem essential, however, that the zero-deviation position of the
beta-ray gauge reading used for setting the required running substance for
82
THE WET-END FLOW SYSTEM 1C.3 3
any particular grade of paper, while being based initially on a test calibra-
tion or data from previous runs, should also be adjusted during the making
to match results obtained from some direct standardized weighing of
the paper. In other words a fundamental comparison should be available
to permit detection of any long-term drift or bias in the beta-ray reading;
this standard may, for instance, be a quickly humidified test weighing done
at intervals on a large sample and compared to the gauge reading at the
time of extraction, or it may be a direct comparison of yardage and weight
of jumbo roll or a slit reel compared to an integrated reading taken over
the appropriate interval. The basis of comparison does not matter so long
as the determination can be done with sufficient accuracy to serve the
purpose of standardizing the beta-ray gauge; suitable statistically deter-
mined error limits would be provided to indicate when action is necessary,
though the equipment may well prove sufficiently accurate and reliable
for the procedure eventually to become redundant.
lC.3 3 Automatic control with a beta-ray gauge

Several attempts to control substance automatically from a beta-ray gauge
on the machine have been reported (7, 9, 18, 28), but the case most fully
documented has been given by Attwood (60). The primary difficulty in
regulating the stuff valve from the beta-ray gauge reading is the long time
lag involved before the effect of an alteration is observed. Continuous
process control with a long time lag is not very satisfactory and Attwood's
method of overcoming this difficulty was to integrate the reading for a
period of 1t minutes, compare this with a standard to initiate action, then
pause for 2-t minutes giving time for the effect to be registered at the dry-
end before repeating the cycle. Allowance was made for inaccuracies and
random variation in the system by requiring the integrated reading to
be outside a dead zone before action was initiated. Movement of the valve,
which was fitted for accurate adjustment with a special design of positioner,
was made proportional to the deviation outside the dead zone. This
procedure apparently proved very effective and, though it could be further
refined in several ways especially in association with on-line computer
control, illustrates the basic approach for a discontinuous form of control
that overcomes the long time-lag.
Several other ideas have been tried with the same objective of endeavour-
ing to overcome the time-lag. One is that the speed of the machine is altered
instead of the stuff valve to correct the substance of the paper. This is fre-
quently easier to arrange than alteration of the stuff valve and will make
the response quicker and possibly finer; it is, however, practicable only
for very small corrections on slower machines because otherwise the power
required to continually accelerate and decelerate the machine is in danger
of becoming excessive. In addition objections can be raised that with
this sort of control the machine cannot be run at the maximum possible
economic speed all the time and if the degree of correction were at all
high, say above 1 per cent. or 2 per cent., then the relative change in slice
83
jet speed and wire speed might affect formation and other conditions on
the machine may be upset.
It has also been proposed that the substance should be controlled from a
beta-ray gauge reading at the presses, or possibly even from the suction
box vacuum. Both of these measurements would be affected by varying
water content of the sheet and, in the case of the latter, by the porosity
of the sheet, but the set-point of the control in either case would come from
a beta-ray gauge reading at the dry-end. The idea is that faster fluctuations
in the substance are removed by the more rapid response gained from
situating the measuring point at the wet-end, while over the long-term the
substance of the paper is controlled from the dry-end. An alternative
arrangement is to measure both substance and moisture at the presses, for
which a non-contacting type of moisture meter is needed, and compute the
dry weight of the paper from these two measurements.
Though these schemes sound plausible in theory they have several
unattractive features and may well prove awkward to manage successfully.
Apart from the practical difficulties of using electronic equipment at the
wet-end and the uncertainty of how much other factors will influence the
efficiency of control, the gain in response time may not be very large because
on many machines it can take longer for fibre to travel from the stuff valve
to the suction boxes or presses, than from the press to the dry-end. In this
event the reduction in time-lag would hardly be great enough to warrant
the added complexity of instrumentation and it is probably only on
board or heavy-substance machines that the additional time required for
the sheet to travel between the presses and reel-up can usefully be
eliminated.
In the author's opinion the time-lag between the correcting signal to
the stuff valve and detection at the beta-ray gauge need not anyway present
a difficulty provided the control is not expected to cope with rapid dis-
turbances. Variation in the flow of fresh fibre to the machine is after all
the main source of longer-term variation in substance, at least of a per-
manent nature as opposed to the transient changes discussed in lA. 2, so it
is logical for this to be the control point. Provided adequate control of
conditions is exercised in the preparation plant, and the flow and con-
sistency of the fresh stuff is regulated as closely as possible, then a beta-ray
gauge controlling the stuff valve should be able to smooth out variations
of the order of several minutes very successfully. This presumes, as stressed
earlier, that there is an appropriate standard to which the gauge reading
can, if necessary, be compared at intervals, and also that moisture in the
paper is either measured simultaneously and allowed for or is adequately
controlled and therefore assumed to be sufficiently constant. Even then a
wise precaution is to provide means to attract the attention of the operators
when the automatic control adjusts the stuff valve beyond certain set
limits, in order to prevent the substance going haywire in the event of a
breakdown in the system. More rapid oscillations and fluctuations with
periods of up to two or three minutes due to vibrations and speed variations
in the drive could not possibly be corrected by this sort of control and
must be tackled at source.
84
1C . 3 4 Control of substance across the sheet
Setting the substance uniformly across the machine is extremely important
because any position which is heavier will also reach the dry-end containing
more moisture and the combined effect of this is to produce a damp place
in the roll. Apart from this a heavy streak may have experienced different
drying stresses from the rest of the web and will receive a different finish
at the calenders; in extreme cases excessive moisture at the calenders may
cause blackening of the paper in a heavy streak and the web may crease
and break. All these effects are detrimental to uniformity of product
across the web and with thicker papers cut into sheets uneven basis weight
also produces cockling. Any defect in the approach flow system, breast
box design and equipment, or slice setting can cause unevenness, but apart
from this the machineman may deliberately run one position different in
dry weight from the rest of the web to remedy trouble on the wire, presses,
or in the drying section. A ridge in the wire giving poor drainage may
necessitate running lighter in that position, as may a deficiency in a wet
or dry felt which impedes water removal. The edges of the sheet, par-
ticularly the front side away from the drive gears which obstruct ingoing
air flow on a machine with an open drying section, may be run heavier
to counteract the tendency to become overdry. In all cases, however,
when the substance is uneven across the web for any length of time trouble
is likely to occur sooner or later. For example, a heavy streak will wear and
plug a new press felt faster because of the greater quantity of water
extracted, so that as the felt ages it is less able to cope with the water and
the presence of the streak becomes more apparent.
Particularly with faster machines, it is of paramount importance that a
level roll is produced otherwise it will be very difficult for the roll to be
slit satisfactorily. In their endeavours to achieve this machine crews will
often adjust the substance to overcome deficiencies of drying or calender-
ing. This is thoroughly bad practice and can only be condoned as an
essentially provisional and temporary expedient. The most satisfactory
approach in the long run is to set the substance as evenly as possible, then
correct for any deficiencies in moisture profile (this pre-supposes ideally a
means of correcting moisture unevenness at any position across the web
apart from altering press load to overcome a general tendency to be
damper or drier in the middle compared to the edges of the roll). Finally the
air blowers on the calenders can help to even up paper thickness across the
roll.
There are many methods by which an attempt is made to control the
substance across the machine from samples extracted from finished rolls,
but the difficulty of moisture pick-up is even harder to overcome in this
case because of the time which will be required to complete the testing.
Weighing small squares of the paper is particularly prone to this error
unless sufficient time is allowed before testing for all the strips to reach
approximate moisture equilibrium. The off-machine beta-ray profiler
provides a very useful continuous cross-web substance curve from a strip,
or preferably a number of strips taken together, and can be quicker and
therefore less likely to be subject to moisture variation. When suitably
4 85
1C.3 4 THE WET-END FLOW SYSTEM
determined control lines are drawn on the profile curves and the position
of jack-screws is marked, the machineman can easily interpret the result
and set the substance across the sheet more evenly than in any other way.
The main difficulty with this method, as with others dependent on testing
a strip taken from the machine roll, is the time and energy involved in
performing the operation; if the machineman is to adjust the jack-screws
conscientiously, then check on the adjustment, and do this as often as is
necessary to get the substance across the sheet as level as possible, a great
deal of work will be required. Furthermore, the occasions when the greatest
change in the cross-web profile is likely to occur, and when checking the
substance across the sheet can be most useful, are at start-up and during
alterations in substance, speed, and breast-box flow conditions; unfor-
tunately, it is at precisely these occasions that crews and laboratory testing
personnel are busiest and requests for a series of profiles may not be well
received.
For a machine with a very stable cross-substance profile which hardly
alters from one shift to the next this may nevertheless be an adequate
arrangement. An on-machine beta-ray gauge capable of traversing across
the web will perform the same job more easily though ultimately, especially
for faster machines, the combined beta-ray gauge and moisture meter
traversing across the machine at the reel-up is coming into more general
use. This instrument has the supreme advantage that the true dry-fibre
weight curve across the machine can be exhibited in an appropriate form
and with it the moisture profile and also, possibly, the total substance
profile (though if the moisture profile is adequately controlled, perhaps
automatically, the dry-fibre weight becomes superfluous). The machineman
should then succeed in adjusting the jack-screws far more efficiently,
particularly when adjustments can be made on the basis of a number of
such profiles obtained at suitable intervals and preferably displayed one
above the other on an XY recorder; persistent heavy or light peaks can
be distinguished more accurately from the random fluctuations which must
exist in any single cross machine profile. Because of the time involved in
making each traverse some machine-direction variation must be included
in the profile drawn by any on-machine substance measuring device;
however, provided action is taken from visual examination of a number of
such profiles the fluctuations from this cause will be averaged out. In
exceptional cases it would be perfectly feasible to reduce the average
cross-web variation below that of the machine-direction to an extent
dependent on the traversing speed and relative degree of fluctuation in the
two directions, i.e. on the general stability and characteristics of the system.
The effect of alteration to the jack-screws, or of the whole slice, cleaning
the slice lips, adjustment of flow at the deckles, etc., can also be quickly
displayed using this technique so that accurate adjustment and correction
becomes much easier.
There have been one or two reports of the use of traversing beta-ray
gauges (18, 28, 56, 86) in which details of the traversing speeds, their use
for machine direction substance control, and other general points are given,
but as yet it is not easy to generalize on details. The author believes that
86
THE WET-END FLOW SYSTEM 1C.4
after initial adjustment it should not be necessary on most machines to
traverse the sheet more frequently than about twice an hour though the
machineman should, of course, be able to obtain a profile at any time
when he requires one; if considerable cross-substance variation can occur in
the sheet at a quicker rate than this then the approach flow and breast box
system is likely to be relatively unstable and rather than trying to chase
the cross-machine substance fluctuations it would be more satisfactory to
go to the root of the trouble. With this frequency of traverse the same
instrum~nt could, as mentioned above, satisfactorily be used for providing
a machine direction record, possibly with automatic control, provided the
reading were locked for the short duration necessary to make a traverse.
lC.4 PRACTICAL POINTS

The wet-end system of a paper machine, probably more than any other
part, is individual to the machine. For that reason a discussion of practical
aspects, as opposed to general principles, is of limited value. Even so there
are several points which seem worth noting and these are dealt with below.
1C.4 1 General considerations affecting start-up conditions

The ability displayed by a machineman when getting his machine going
after a shut period is put very high in assessment of his overall merit and
skill. The machineman is well aware of this and also, of course, realises
that it is only too easy to compare his time to get the sheet through to the
reel-up with that of his colleagues on the other shifts. The result of this is
seen all too often in paper mills; paper is made at the reel-up in good time,
only to be tom up for hours afterwards because it is inferior in quality.
The fault does not lie entirely with the machineman for often the beater-
man, in his haste to fill the machine chests, skimps treatment in the beater
or hydrapulper and is not so particular about the results of routine
laboratory tests on the stuff. Nevertheless, each beaterman and machineman
has his own routine for starting a machine and it would be an unwise
person who attempted to interfere with this unnecessarily. A routine is
gradually worked out between them which, provided they are reasonably
capable and do not work in competition with or isolation from one
another, reaches a point where it is at least adequate for starting the machine
in reasonable time.
The trouble always comes when the stock is run on the wire and attempts
are made to get the web through the presses and drying section; one
Monday morning everything will go smoothly, next time there are breaks
galore at the couch and everywhere else on the machine, and even when a
reel has been built up without too many breaks it is only useful for broke
because of the poor quality. Some blame can no doubt be attached to
variations in pulp quality yet on most machines, when everything is running
smoothly, surprisingly wide swings in pulp preparation can be tolerated
without ruining the paper quality or producing an excessive number of
87
tC.41 THE WET-END FLOW SYSTEM
breaks. Also, perhaps more frequently than is necessary in many mills,

conditions are not set ready for the particular paper to be made as well as
they might be. This, of course, applies to the whole machine and it goes
without saying that all machine crews should know, from records of
previous operation they keep either on paper or mentally, precisely where
to position the stuff valve, slice, presses, draws, etc. to ensure they will be
close to the eventual settings required. Apart from such more obvious
faults, however, it is frequently the case that inadequate preparation is
made to start up as close to normal running condition as possible.
In older fine paper mills where treatment in beaters is a lengthy process
stuff will be kept stored in the preparation system up to and including the
machine chest, at least provided the shut is not longer than a normal
week-end. In such systems, particularly when recirculation is confined
to a simple backwater system with no hog-pit and with fresh water in general
use elsewhere, the normal running condition on the machine is rapidly
achieved with regard to fines recirculation; the time for establishment of
equilibrium in the main backwater circuit has already been discussed in
1A .2 5. In fact, with frequent colour changes and quality changes, the
efficiency of operation of such a machine hinges very much on being able
to settle down rapidly to the usual speed and general running conditions.
Contrast this with the position of a modern fast machine where it is
customary to run out the chest and empty the system for a clean-up. If
at the start-up fresh water is used until the whitewater system commences
to operate, then the whole character of the preparation system will change
over a period lasting many hours. At worst this might simply necessitate
frequent alterations to the flow of backwater used for diluting the fresh
stuff to take care of fluctuations in wetness on the wire, and also changes
to the stuff valve to keep the substance correct. But with increasing
recirculation and build-up of fines in the whitewater system it is probable
that refining will require alteration as well and it may well prove difficult
to keep the paper quality consistent over this period. To alleviate this it
would, where no colour changes are involved, be well worth retaining the
whitewater tank full over a normal shut period, even if this incurred the
problem of slime build-up and necessitated use of a small agitator at start-
up or, possibly, during the whole shut period. By this means it should be
possible to reduce considerably the fresh water usage at start-up and ensure
a quicker approach in the preparation stage to normal running.
It is often the case with faster machines making newsprint, tissue, and
kraft papers that the quality is likely to suffer if insufficient broke is avail-
able. This presents a similar problem to the fines in whitewater but cannot
be solved so easily as few mills regularly like to keep broke tanks full over
a normal week-end period. On such machines it is general to have hog pits
collecting the sheet washed off the wire so considerable broke is usually
generated right at the start; this, if returned with a minimum of treatment
straight to a broke chest and thence to the machine chest, will greatly
assist in establishing equilibrium reasonably quickly so far as broke
characteristics are concerned. It is, though, probably more satisfactory to
use a dry-end hydrapulper or broke pulper to break down trim or reels
88
stripped from the previous week and ensure that this is done early enough
to permit the usual supply of broke to be available for metering in at the
addition point.
On all machines a more familiar start-up problem is to obtain normal
running conditions with regard to temperature and chemical condition
of the stock, especially pH. On larger machines it may take many hours
to reach the usual running temperature and pH and this all adds to the
problem of keeping conditions and quality consistent. In particular, it
usually necessitates starting up at a slower speed than normal and it may
take two or three shifts before speed is finally got up. During all this time
the machine crews have to keep extremely alert to ensure that everything
is kept under control and it is not surprising that the general quality of
paper made during this period, if not actually broke, is of lower standard
or differs in some important respect from the usual quality.
With regard to temperature the remedy at first sight seems simple.
Steam condensing pipes in the machine backwater pit will enable normal
running temperature to be attained at start-up, while adequate supply of
heat to the preparation plant should likewise ensure that the fresh stuff
is not so cold as to affect adversely the stock temperature at the most
critical point, the breast box. Unfortunately it is precisely at this time that
steam demand in the whole mill is highest with the drying cylinders
usualJy drawing as much steam as it is possible to get hold of. Even so the
situation may well be worth examining in many mills where temperature
is known to be a major factor in requiring a slow start-up in machine speed.
By appropriate metering of steam and more careful setting of valves it
may be practicable to achieve a better compromise than the existing
one.
The same difficulty should not apply to alum supply whether by batch
in powdered form or from a central liquid supply plant. On machines
where the procedure is to use alum in the beaters with perhaps only a
trimming device on the machine itself, the problem of pH fluctuations at
start-up is not likely with adequate control to be serious. But when the
sole alum control is in the machine backwater pit it may be difficult at
first to keep pace with the variations when the machine starts. A pH
recorder and controller is invaluable for this but even then it may be more
satisfactory to add alum temporarily in the preparation plant, especially
when fresh water is used at start-up. The method adopted must, of course,
depend on the lay-out but it is well worth considering dosing the white-
water tank itself when this is first filled with fresh water for use in diluting
the various parts of the preparation system.
lC .4 2 Start-up of the machine

The first step on any machine is to fill the machine backwater pit with
fresh water and begin flushing out the system through the mixing box or
pump, cleaners, screens and breast box. This water may at first go to drain
but eventually it can be passed through a dump valve back to the back-
water pit to be recirculated round for some time. On some machines,
89
depending on the lay-out, the water will pass through the slice and onto
the wire and this is necessary eventually anyway if the slice is to be cleaned;
the wire may be crawled round and washed off while this is being done
or it may be stationary with the debris collecting on the wire to be washed
off at intervals. When necessary, the temperature of the water will be
gradually raised during this time and circulation will help to heat up the
whole system. With high temperatures care is necessary to avoid damaging
the wire, especially if the sprays are cold at first. Raising the temperature
to near normal running is advantageous not only from the point of view
of drainage of the stock but also for the slice lips which may expand to
nearer their eventual shape and will not then be liable to affect the cross-
level substance after start-up.
By the time the machine is ready to start the pH of the water should
also have been regulated to an appropriate value. Screen and wire sprays,
perforated roll drive and seals, shake, breast box air-loading compressor,
hog-pit agitator, vacuum pumps, broke pump, etc. are all started and
checked. Then the machine chest pump is started and, in the absence of an
automatic control, the throttling valve is manually adjusted to set a small
overflow from the stuff box once the box has filled. In the meantime the
mixing pump speed and the main valve governing the flow of backwater
should have been set manually or automatically to give approximately a
head at the slice appropriate to the starting-up speed of the wire. The stuff
valve or gate is slowly opened and the dry-line and vacuum gauge on the
suction boxes (which should be visible from the valve) are carefully watched;
with enclosed systems, it is advantageous to have a separate stuff cut-off
valve apart from the substance regulating valve for this purpose so that the
latter need not be subjected to violent movement from its normal working
position.
Until the backwater consistency has built up to near normal in three or
four circuits the substance will be very light and sometimes an attempt is
made to compensate for this by opening the stuff valve wider for a short
period. There is, however, little point in rushing this part of the procedure
and the necessary delay can be usefully devoted to many tasks. Amongst
these may be mentioned setting of the dry-line evenly by adjusting the
jack-screws (which should reduce the alteration necessary once the sub-
stance across the web has been checked at the dry-end), ensuring the head is
up for the wire speed (particularly if the stock is cold and drainage slower at
start-up there may be a tendency to run with the head too low and the
poor formation resulting can produce trouble when feeding through the
presses and drying section), setting the edges as well as possible (particularly
important when there is little or no trim), and general inspection of the
wire part which is dealt with in detail in Part 3.
When sufficient whitewater is available those sprays normally using
whitewater will be turned over from fresh water. By the time the vacuum
on the suction boxes and couch is adjusted, the dandy lowered if required,
and weight is applied to the jacketed couch roll or the presser roll lowered
to normal running position, conditions in the backwater circuit should be
close to equilibrium. With the stuff valve checked back if necessary to its
90
usual setting the sheet on the wire will then be ready for passing through
to the presses.
On modern machines it is becoming customary to ease the task of the
machineman at start-up and speed the procedure by grouping the various
switches and controls in convenient positions which allow him to keep a
continual eye on the situation; it is surprising how much dashing about
the average machineman has to do on an older machine when he should
have ample time to watch and adjust the more important and vulnerable
aspects of the wire part. The logical conclusions of the process of grouping
is to have a single control panel at the wet-end with the starting buttons
placed in a suitable sequence. Such an arrangement has been described
in some detail by Eastwood and Cade (31) with a complete start-up
sequence of operations included, from turning on fresh water to the
showers and closing the wire pit drain valve right through to opening of
the stuff valve. One advantage of this is that safeguards are easily built
into the sequence to ensure, for example, that steam heating is not allowed
into the pit before an adequate level of water has been reached. Eastwood
and Gade also describe two further refinements, the use of a graphic
panel to depict the key parts of the process and a sequence timer with
which the whole start-up routine could be performed automatically. It
may be interesting to note that the graphic panel was eventually considered
of doubtful value to the operators while some machinemen preferred not
to use the automatic start-up. However, the benefit in having the switches
in sequence on a single panel was thought to be definitely worth the capital
expense involved and the additional cost on top of that for making a
graphic panel and incorporating the automatic sequence timer was
marginal.
lC.4 3 Shut-down
For an ordinary scheduled shut-down the machine chest will usually be
run down as low as possible before stopping the chest pump. The substance
of the sheet will immediately begin to drop as the head in the stuff box
falls and at some point shortly after this the sheet is broken down at the
couch. The wire is slowed, stuff valve closed, mixing pump stopped, and
sprays switched over to fresh water. If fibre is retained the hog-pit and
main pit may be washed out through the save-all, but normally they will
be washed to drain through the dump valve and pits. The wire is finally
stopped, the sprays turned off, and other parts of the wet-end, the screens,
perforated rolls, etc., closed down.
Cleaning of the wet-end is usually accomplished with a high-pressure
water jet to which special nozzles may be attached for various jobs. All
inspection plates and drains should be opened and pipework and breast
box parts thoroughly cleaned in this way, the more so when slime collects
on the machine. If the wire has been cut off a more complete wash-up
may be undertaken by pumping a caustic solution, possibly with a little
insecticide added, right through the main flow system, a procedure adopted
more commonly on faster machines with a great deal of enclosed boxes
and pipework. The solution is placed in the backwater pit, heated if
91
necessary, and pumped round through cleaners, screens, breast box and
slice; periodically the machine chest may also be filled and the chest pump
and stuff box cleaned in the same way, as may the broke system, whitewater
system, particularly sprays, and the preparation plant when necessary.
In every case when the cleaning is finished the caustic solution must be
drained out and sufficient fresh water should be circulated to flush the
system clear.
For an emergency shut when the wire has to be stopped immediately,
the stuff valve must be closed as quickly as possible and the sprays turned
to fresh water to avoid marking the wire with whitewater jets. On machines
with a dump valve to the main pit this can be opened; alternatively the
mixing pump is stopped. When these steps have been taken to make the
wire safe and prevent excessive accumulation of stock or water in the
system which would make starting-up difficult, attention can be given to
the other normal details of shutting and to the trouble in hand. It is at
such times as this that a central control panel is invaluable, and use of
pneumatically operated valves enables the process of stopping the wire
to be accomplished rapidly by pressing a single button (see lB. 5 2).
1C.4 4 Changing quality and substance
More than with any other part of the paper machine, a change in quality
affects the running of the wet-end and is dependent on the design of the
system. On slower machines a colour change will rarely necessitate a
complete shut-down and clean out; such machines are designed to facilitate
changing and customarily have two machine chests so that one may be
cleaned out and filled with the new furnish while the other is run down and
cleaned out afterwards when the machine is under way again. In some
cases it may be quite possible and economical to run through from one
colour to another, tearing up paper made in the transitional stage; but
often the whitewater system will only be rudimentary, with none leaving
the machine house for dilution purposes earlier in stock preparation plants
and in this case for a difficult colour change it is only necessary to empty
the wire pit and save-all and fill up with fresh water. Rapid colour changes
are facilitated by adding dyes and broke recovered from a save-all as late
in the system as possible consistent with thorough mixing, an adequate
residence time, and compatibility with alum addition; when this can be
arranged after the machine chest, i.e. all fresh stuff from the preparation
plant is uncoloured, then whether it is necessary to wash out the wet-end
or simply run through, the time for changing is low, matching is rapid,
and the risk of loss due to overcolouring in preparation is avoided.
But on faster machines with an elaborate whitewater system a colour
change may necessitate draining out the whole system, whitewater, broke
and all, exactly as for a week-end shut. Not until the system has been
thoroughly cleaned out can the new colour be run. Normally, of course,
this is obviously uneconomical and on such machines changes in furnish
are confined to differing proportions of pulps and varying degrees of
refining. In this case the machine chest can be run down as far as possible
before changing the furnish but the machine is never stopped and the
92
paper made during the transition period is torn up. The economics of these
different types of change depend primarily on the time involved and the
characteristics of the machine; what may be satisfactory to suit a rare
contingency in the order book would often be most unprofitable as a regular
routine. Certainly any regular change-over warrants careful planning.
For a straightforward change in substance the situation is much easier.
On slow machines with variable deckle the machineman will determine
the alteration to speed necessary to take care of the change in substance,
and at the same time an alteration in stuff valve or gate position may be
necessary to accommodate the change in deckle. The machine speed may
be fixed so that the production per foot width will be about the same while
the stuff valve is altered pro rata with the deckle. Precisely the same applies
to faster machines with no deckle change but in this case the stuff valve
will probably not be touched.
The general validity of these remarks depends essentially on the degree
of substance change required. On fine machines the substance may be
increased fourfold and this obviously requires other considerations to be
taken into account when compared to a change of a few g.s.m. on a fast news
machine. To take the latter case first, no change in treatment of the furnish
should be necessary and if drainage or drying capacity is assumed to be the
limiting factor then production is kept about the same and the only
alterations required to obtain the new substance are appropriate adjustment
of the machine speed and head at the slice. With the stuff valve unaltered in
position the drainage and drying demand is effectively unaltered and, as
examined in the theoretical section, the consistencies at the wet-end are also
unaltered if the change in head is effected by opening up the slice gap. In
practice the head will probably be altered by changing the backwater
circulation and for a small substance change this is perfectly satisfactory.
Effecting such a change is a straightforward procedure involving (for a
substance increase) lowering of the head, which will bring back the dry-
line, followed by a drop in the machine speed, the same steps repeated as
often as necessary to avoid upsetting the machine. For a substance decrease
the reverse procedure is used.
If, on the other hand, machine speed is limited and for a decrease in
substance production must be reduced, then a proportional change to the
stuff valve would be necessary. In these circumstances this is probably the
only alteration that would be made until the new position of the dry-line
and vacuum in the suction boxes indicated the desirability of altering the
backwater circulation.
For a more considerable change in substance the wetness of the furnish
will be altered and then the changes on the machine must be more depen-
dent on gradual alteration of conditions to keep the dry-line and suction
box vacuum steady. The slice gap will invariably need alteration in addition
to the other changes detailed above and its final position will depend more
than anything else on the relative difference in drainage on the wire due
to the alteration in substance and beating. Especially on machines making
such large changes in substance an appreciable saving is possible with the
aid of a beta-ray gauge at the dry-end to monitor the effect of the various
93
adjustments made during the change-over. Also, as at a start-up, the normal

positions of the valves, slice, gauges, etc. for the new substance should be
known from past running and this will help the machineman to achieve
the desired conditions quickly. It is in this sphere especially that on-line
computers which can carry out a grade-change automatically and in a
minimum time are presenting a high rate of economic return.
lC .4 5 Checking the wet-end flow system during running
The importance of the various gauges and recorders which can be found
through the wet-end flow system has been dealt with in detail in lC.1 and
their use to the machineman in keeping a check on general running condi-
tions requires no further elaboration. The majority of these measurements
serve either to show whether running conditions differ from normal or
to allow more accurate setting of the conditions than would otherwise be
the case. The flow of stock to the breast box, or in a rougher but more
common way, the gap between the slice lips are particularly important
when used in association with the position of the dry-line since they give the
machineman a systematic indication of whether drainage is satisfactory
or whether the stock is working wetter or freer on the wire. Similarly a
slice head measurement compared to wire speed, by assisting in setting the
relative speeds of jet and wire, draws attention to a highly important aspect
of formation which it is difficult to judge by eye except on slow machines.
These features apart, inspection at the wet-end so far as the flow system
is concerned largely comes down to such matters as checking that the
sprays are functioning adequately, that pressures in the pipe system are
satisfactory, chemical, loading, and dye addition is steady, perforated roll
drive is running, and the whitewater system and save-all are in order.
The presence of slime, pitch, foam, and fibre clumps in the whole flow
system will obviously be noted and steps taken to eradicate the trouble by,
if necessary, shutting the machine for a wash-up; prevention of these
nuisances has been considered in 1C.2, and removal of slime and pitch
from the wire part where they are most troublesome will be dealt with in
Part 3. Levels in pits, tanks, stuff box, and breast box require checking,
whether they are controlled or not; the liquid surface and walls of the
breast box, when observable (which should always be the case), whould
be examined to ensure the sprays are keeping down froth and scum, also
that no violent disturbance is present on the surface which could affect
the evenness of slice discharge. The slice itself should always be scrupulously
clean and the effect of dirt and fibre clumps lodged under the lips is readily
detected in the flow on the wire; the edges of the slice can also be affected
by a build-up of lumps of fibre and require careful inspection. Keeping the
substance both across and along the web steady usually occupies a fair
proportion of the machineman's time, especially on a machine which is
poorly controlled and relatively unstable, and this topic has already been
discussed in detail. Otherwise the attention of the machineman is directed
mainly to the wire part and, though in practice the flow system and the
wire part are inspected together, further details of the latter will be deferred
till Part 3.
94
REFERENCES
(l) Hansen, E.: •The Distribution of Filler in Paper.' TAPPI 34, 4, 180. April, 1951.
(2) Bienkiewicz, B., Mardon, J., and Robinson, W. F. E.: 'Studies in Backwater
Systems.' Part !-'Theoretical Principles of Backwater Systems, Operation and
Methods for Material Balance Construction.' Proc. Tech. Sec. 33, 2, 383. May,
1952.
(3) Bienkiewicz, B., Mardon, J., Robinson, W. F. E., and Scruton, H.: 'Studies in
Backwater Systems.' Part II-'Examples of Closed Systems in Operation.'
Proc. Tech. Sec. 33, 2, 395. May, 1952.
(4) Hendry, I. F., Mardon, J., Scruton, H., and Wilkinson, D. M.: 'Drainage on a
Tissue Machine Wire Table.' Proc. Tech. Sec. 33, 3, 603. December, 1952.
(5) Mardon, J., Robinson, W. F. E., and Wilkinson, D. M.: 'Studies in Backwater
Systems.' Part III-'Operation of Closed System: Paperrnaking Considerations.'
Proc. Tech. Sec. 33, 2, 408. May, 1952.
(6) Sergeant, S. V.: 'Some Flow Box Investigations.' Proc. Tech. Sec. 33, 1, 49.
February, 1952.
(7) Chartrand, J. Y.: 'Automatic Control with the Betameter.' P. and P. Mag. Can.
54, 5, 109. April, 1953.
(8) Duskin, A. W. and Jenkins, W. B.: 'The Deculator on a Kraft Paper Machine.'
TAPP! 36, 11, 490. November, 1953.
(9) Wilson, G. E.: 'The Betameter in a Basis-Weight Control Application.' P. and
P. Mag. Can. 54, 3, 224. Convention Issue, 1953.
(l 0) Bennett, W. E.: 'Quantitative Studies of Water Removal by Table Rolls.' TAPP I
37, 11, 534, November, 1954.
(11) Gavelin, G.: 'Some Effects of Gases on Properties of Fibre Suspensions.' P. and
P. Mag. Can. 55, 3, 191. Convention Issue, 1954.
(12) Mardon, J. and Van Der Meer, W.: 'Flocculation, Formation and Headbox
Design from the Aspect of Turbulent Motion.' P. and P. Mag. Can. 55, 9, 107.
October, 1954.
(13) Mason, S. G.: 'Fibre Motions and Flocculation.' TAPPI 37, 11, 494. November,
1954.
(14) Mason, S. G., Robertson, A. A., and Allen, G. A.: 'Turbulent Diffusion in
Fourdrinier Machines.' P. and P. Mag. Can. 55, 8, 97. August, 1954.
(15) Robertson, A. A. and Mason, S. G.: 'Flocculation in Flowing Pulp Suspensions.'
P. and P. Mag. Can. 55, 3, 263, Convention Issue. 1954.
(16) Van Den Akker: 'Flow in Fibre Suspensions and the Effects of Turbulence.'
TAPPI 37, 11, 489. November, 1954.
(17) Van Der Meer, W.: 'Hydraulics of Flowbox and Slice.' TAPPI 37, 11, 502.
November, 1954.
(18) Bearer, G. F.: 'Automatic Control of Basis Weight Using Radioisotopes.' TAPP I
38, 1, 139A. January, 1955.
(19) Jacobsson, J. J.: 'De-aerating of Paper Stock for the Paper Machine.' Proc.
Tech. Sec. 36, 2, 387. May, 1955.
(20) Landesman, A. L.: 'The Role of Flow Control in Basis Weight and Caliper
Control.' TAPPI 38, 5, 148A. May, 1955.
(21) Mardon, J., Gavelin, G., and Logan, K. C.: 'The Effect of Some Machine Charac-
teristics on Paper Uniformity.' P. and P. Mag. Can. 56, 3, 275. Convention
Issue, 1955.
(22) Baines, W. D. and Nicholl, C. I. H.: 'Hydrodynamics of the Fourdrinier Wet-End.
I. Fibre Orientation in Pipe Flow.' P. and P. Mag. Can. 57, 6, 119. May, 1956.
(23) Boadway, J. D.: 'Gas in Papermaking Stock.' P. and P. Mag. Can. 57, 3, 185.
Convention Issue, 1956.
(24) Mardon, J., O'Blenes, G. and Ryan, J. A.: 'Hydraulic Measurements in Paper
Machine Headbox Design and Operation.' Svensk Pap. 59, 12, 429. June, 1956.
95
(25) Mardon, J. and Shoumatoff, N.: 'Flocculation, Formation and Headbox Design.
Part IV.' P. and P. Mag. 57, 3, 305. Convention Issue, 1956.
(26) Robertson, A. A. and Mason, S. G.: 'Hydrodynamics of the Fourdrinier Wet-
End. II Pipe Flow and Fibre Flocculation.' P. and P. Mag. Can. 57, 6, 121.
May, 1956.
(27) Underwood, N. C. and Shipstone, B. A.: 'Reduction of Basis Weight Variation
on a Newsprint Machine.' Canadian Tech. Sec. Proc. p. 510, 1956.
(28) Van Horne: 'Evaluation of Beta Gauge Control System.' TAPP! 39, 12, 193A.
December, 1956.
(29) Bennett, H. W.: 'The Flow Characteristics of Distributor Rolls and Perforated
Plates.' TAPPI 40, 12, 978. December, 1957.
(30) Cuffey, W. H.: 'Some Factors Involved in Basis Weight Uniformity.' TAPPI
40, 6, 190A. June, 1957.
(31) Eastwood, G. G. M. and Gade, M. F.: 'Instrumentation for the Automatic
Start-Up of the Wet End of a Paper Machine.' TAPPI 40, 6, 403. June, 1957.
(32) Robertson, A. A. and Mason, S. G.: 'The Flow Characteristics of Dilute Fibre
Suspensions. TAPPI 40, 5, 326. May, 1957.
(33) Anon: 'Instrumentation and Automatic Control at Aylesford.' Paper Maker 86,
5, 50. November, 1958.
(34) Baines, W. D.: 'The Influence of a Simple Surge Tank on Pressure Oscillations
in a Piping System.' P. and P. Mag. Can. 59, 3, 177. Convention Issue, 1958.
(35) Bonnington, S. T.: 'The Frictional Losses of Paper Stock in Pipes.' Proc. Tech.
Sec. 39, 2, 257. June, 1958.
(36) Evans, J. C. W.: 'Hamilton Completes Major Expansion with many Interesting
Features.' Paper Trade Journal 142, 41, 38. December 22, 1958.
(37) Forgacs, 0. L., Robertson, A. A., and Mason, S. G.: 'The Hydrodynamical
Behaviour of Paperrnaking Fibres.' P. and P. Mag. Can. 59, 5, 117. May, 1958.
(39) Mc.Nown, J. S. and Spengos, A. C.: 'A Preliminary Study of the Instability of
the Stock on the Wire of a Fourdrinier Paper Machine.' TAPPI 41, I, 9. January,
1958.
(40) Mardon, J., O'Blenes, G., and Wahlstrom, P. B.: 'The Hydrodynamics of Paper
Machine Headbox Approach Piping.' P. and P. Mag. Can. 59, 4, 139. April,
1958.
(41) Paterson, H. A. and Hazeldine, E.: 'Mixed Stock Control System at Mersey
Paper Co. Ltd.' P. and P. Mag. Can., 59, 3, 279. Convention Issue, 1958.
(42) Reitzel, N. M.: 'An Investigation of Hydraulic Pressure Pulsations in Wet-End
Piping of a Paper Machine.' P. and P. Mag. Can. 59, 3, 184. Convention Issue,
1958.
(43) Anon: 'The Egger Diaphragm Regulating Valve.' Das Papier 13, 11/12, 268.
June, 1959. (In German).
(44) Brecht, W., and Kirchner, U.: 'The Air Content of Paper Stock.' Woch. fur Pap.
87, 8, 295. April, 1959. (In German). Summary in Paper Maker 138, 2, 60.
August, 1959.
(45) Mliller-Rid, W. and Pausch, G.: 'Studies on the Slice of a Newsprint Machine.'
Woch. fur Pap. 87, 11/12, 478. 1959 (In German). Summary in Paper Technology
1, 1, 11 February, 1960.
(46) Reitzel, N. M.: 'A Wide Range Pressure Inlet: Design and Operation.' TAPPI
42, 2, 137. February, 1959.
(47) Ure, R. A., Nuttall, G. H., and 'Lay Boy.': Keeping Down Froth on a Paper
Machine.' Paper Maker 138, 1, 36. July, 1959.
(48) Anon: 'British Made Machine Installed in German Mill.' Paper Trade Review
153, 21, 1899. May 26, 1960.
(49) Attwood, D. and Alderson, J. P.: 'The Use of a Large Model to Study the Flow
Pattern of an Open Flow Box.' Paper Technology 1, 6, 653. December, 1960.
(50) Brewden, H. A., and Locking, B.: 'New Developments in Paper Machine Headbox
Design and Equipment.' Paper Trade Journal 144, 31, 36; 32, 24; 33, 38. August
1, 8 and 15, 1960.
(51) Gardner, R. C., and LeBreton, L.: 'Instrumentation of a Paper Machine.' Paper
Technology 1, 3, T82. June, 1960.
96
(52) Howarth, J. and Banks, W. A.: 'Machinery for Automation.' Paper Technology
1, 3, T112. June, 1960.
(53) Mardon, J., and Wahlstrom, B.: 'The Characterisation of some Modern Head-
boxes by the use of Hydraulic Measurements.' Svensk Pap. 63, 20, 716. October
31, 1960.
(54) Muller-Rid, W. and Pausch, G.: 'The Effect of Operating Conditions on Sheet
Formation.' Das Papier 14, 12, 647. 1960. (In German). Summary in Paper
Technology 2, 2, 186. April, 1961.
(55) Nelson, H. C.: 'Paper Machine Inlet Performance in Relation to the Fourdrinier
Wire.' TAPP! 43, 4, 330. April, 1960.
(56) Smith, B. W.: 'Scanning Basis Weight and Moisture Gauges Systems on Paper
Machines.' TAPP! 43, 3, 226. March, 1960.
(57) William, D. J.: 'A Continuous Freeness Recorder for Pulp Suspensions.' TAPPI
43, 7, 613. July, 1960.
(58) Anon: 'New Machine boosts Fine Paper output in latest Wassau Expansion.'
Paper Trade Journal 145, 40, 20. October 2, 1961.
(59) Anon: 'How much do you know about Knife Slices?' Paper Trade Journal 145,
51, 39. December 18, 1961.
(60) Attwood, D.: 'The use of a Beta-Gauge for Automatic Basis Weight Control.'
Paper Technology 2, 1. T19. February, 1961.
(61) Barker, E. F.: 'Unique Bruderhaus Organisation Installs New Paper Machine
in Own Mill.' Paper Trade Journal 145, l, 20. January 2, 1961.
(62) Brecht, W. and Kirchner, U.: 'The Air Content of Pulp Suspension.' Das Papier
15, lOa, 625. 1961. (In German.)
(63) Chedomir, G. A.: 'Continuous Freeness Recording Controller.' TAPPI 44, 6,
220A. June, 1961.
(64) Harman, G. L.: 'Continuous Freeness Measurement and Control.' Paper Trade
Journal 145, 34, 54. August 21, 1961.
(65) MacLaurin, M. I.: 'Instrumentation-its Success and Delusions.' Svensk Pap.
64, 10, 394. May 31, 1961.
(66) Moore, M., and Walsh, E. P.: 'Discussion of some News Machine Wet-End
Operating Variables.' P. and P. Mag. Can. 62, 5, T-293. May, 1961.
(67) Parker, J. D.: 'Recent Theoretical Work in the Area of Sheet Formation.'
TAPP! 44, 4, 162A. April, 1961.
(68) Van Derveer, P. D.: 'New Machine Boosts Nekoosa-Edward's Fine Paper
Production.' Paper Trade Journal 145, 9, 1-8. February 27, 1961.
(69) Vercoe, J. E.: 'Fourdrinier Maintenance-Key to Machine Efficiency and Wire
Life.' Paper Trade Journal 145, 14, 36. April 3, 1961.
(70) Cuffey, W. H., and Ingram, G.: 'Cyclic Machine Direction Basis Weight Varia-
tion.' P. and P. Mag. Can. 63, 4, T-203. April, 1962.
(71) Mac Gugan, I. C.: 'Calcium Chloride-An Effective Stock Drainage Aid.' Paper
Trade Journal 146, 2, 22. January 8, 1962.
(72) Millier-Rid, W., Stark, A., and Pausch, G.: 'European Thoughts on Paper
Machine Stock Distributors and Stock Inlets.' Paper Trade Journal 146, 1, 26.
January l, 1962.
(73) Rathman, H.: 'Slime Formation and Slime Control in Paper Mills.' Book Review
by Russell, P. in Paper Technology 3, 3, 252. June, 1962.
(74) Wrist, P. E.: 'Dynamics of Sheet Formation on the Fourdrinier Machine.'
Formation and Structure of Paper. Transactions of the September 1961 Oxford
Symposium, p. 839. B.P. & B.M.A. Tech. Sec. publication, 1962.
(75) Anon: 'Round-up Report on Three Continuous Freeness Testers.' Paper Trade
Journal 147, 10, 38. March 11, 1963.
(76) Beecher, A. E.: 'Dynamic Modelling Techniques in the Paper Industry.' T APPI
46, 2, 117. February, 1963.
(77) Mardon, J., Manson, D. W., Wilder, J. E., et al.: 'The Design of Manifold
Systems for Paper Machine Headboxes.' Part I 'Branching Pipe Manifolds.'
P. and P. Mag. Can. 64, 2, T-35. February, 1963. Part II 'Taper Flow Manifolds.'
TAPPI 46, 3, 172 March, 1963.
(78) Robinson, D. E.: 'The Detection of Pulsation and Consistency Variation in Jet
Inlet Flow.' P. and P. Mag. Can. 64, 3, T-169. March, 1963.
97
PART 2
SCREENS AND CLEANERS

INTRODUCTION
21 An essential part of the operation of making paper on the Fourdrinier
is removal from the stock of impurities and tangled mats of fibre which
spoil the appearance of the sheet, are a source of breaks, and in some
cases increase wear of the wire and calender rolls. The growing standards
of converters and greater speeds of converting and printing machinery
ensure that this function is continually increasing in importance, even
though pulp itself, whether imported or prepared at the mill, has been
improved appreciably in cleanliness to meet the needs of the papermaker.
A measure of the value now attached to efficient purification of stock is
that even newsprint machines are being equipped more and more with
both screens and cleaners, despite the heavy increase in capital and run-
ning costs.
This part of the book is concerned with the design and operation of
equipment used at the wet end of a paper machine for the removal of
unwanted material in the stock. Equipment for the same purpose is, of
course, used in pulp mills where it is generally similar in design and in
most cases can be readily interchanged with units in the paper mill. How-
ever, certain models are normally found only in pulp mills, particularly
those concerned with the coarse screening and cleaning of pulp or having
some specialized function such as the magnetic separation of ferrous
particles. Also some models are designed primarily for use on particular
types of pulp, esparto, rag, wood, etc., and some are intended for use at
relatively high consistencies more usual in the pulp making process.
Attention is confined here solely to equipment commonly in use on the
paper machine, which implies that it is used on diluted stock between
the mixing pump or box and the breast box at a consistency normally
under 1 per cent.
The operation of a piece of screening or cleaning equipment is closely
related to the actual design. Though two models may have precisely the
same function and have been designed with removal of the same kinds of
impurity in mind, their actual construction, performance, and running
can be totally different. For this reason in what follows it has been found
convenient to divide the chapter which deals with operating factors
affecting screens and cleaners according to the various main categories of
equipment available. The chapter on theory is divided in a straightforward
manner between screening and cleaning as a whole, with an additional
piece on methods of assessing the efficiency of removal of impurities. The
final chapter on the running of screens and cleaners follows the pattern
adopted throughout the book for each section of the Fourdrinier.
21 . 1 Different types of equipment

It will be as well, at the outset, to be clear on what is meant by 'screens'
and 'cleaners'. The two terms are often used loosely, particularly in the
101
21.1 SCREENS AND CLEANERS
expression 'screening equipment' which is frequently meant to cover

both types. Confusion arises too because a screen can be said to clean
stock in the sense that it removes certain types of impurity, while a cleaner
also acts to some extent as a screen.
The distinction becomes reasonably clear if a screen is thought of as
a piece of equipment which separates essentially by size, while a cleaner
separates by density. On this basis to 'screen' means to direct the paper
stock through small holes or slots and thereby keep back anything that
is not small enough to pass through. To 'clean' means in practice to subject
the stock to forces, usually centrifugal, which cause particles of different
density to go in different directions so that a separation can be effected.
Most screens and cleaners also tend to break down or deflocculate tangled
clumps of fibres: screens by the action of a vibrating or pulsating mech-
anism, cleaners because of the heavy shear forces that occur in the flow.
Screens come in a variety of types which for convenience can be divided
as follows (the more common makes are given in parentheses):
flat diaphragm (Watford, Jonsson, Sherbrooke, Parker)
rotary outward flow (Wandel, Walpole)
rotary inward flow (Bird, Voith, Leith-Walk, Watford)
enclosed pressure (Selectifier, Centriscreen, Finckh, Lamort, Dura-
finer)
Cleaners can be divided into:
sand traps or riffiers
rotating basket (Erkensator, Purifuge, Centrifiner)
cylindrical (Dirtec, Vortrap, Vorject, Voith Tube Separator, Hydra-
clone)
cyclone (Centricleaner, Berg, Radiclone, Hy-Cleaner)
A great number of these models are now obsolete, though many are
still in use so brief comment on them will be made in the appropriate place.
There are in addition to those listed above several rather specialized
designs, for example high-frequency rotating screens (e.g. Vibrotor) used
before dilution with backwater (but nevertheless at a consistency as low
as 1 per cent.) on kraft machines where cleanliness is of relatively little
importance; such models will be mentioned only insofar as details of
their operation have some bearing on the more general types employed
on paper machines.
Other types of equipment find their way into paper milis but are, strictly
speaking, designed specifically for use on pulp and are not incorporated
as part of the paper machine system itself. This applies to models such as
the Centriffier, a bulk trash remover commonly used in waste paper stock
preparation. This type also is not dealt with in what follows.
Finally, it is as well to clarify terminology. In this work the flow to
any screen or cleaner will be referred to as the 'inlet flow'. At the screen
or cleaner this divides into an 'accept flow' containing (it is hoped) fewer
impurities in relation to fibre, and a 'reject flow' containing a greater
proportion of impurities. The reject flow is normally either re-treated in
further screening or cleaning equipment or led to drain. Inlet, accept and
102
SCREENS AND CLEANERS 21.2
reject flow refers to the whole flow; occasionally it will be necessary to
refer to the individual constituents and then the terms 'fluid', 'fibre' or
'dirt' flow will be used. 'Efficiency' used in a general sense means simply
the ability of a screen or cleaner to separate impurities from fibre. It will
be necessary later to define particular aspects of this more carefully.
21. 2 Choice of equipment

The wide range of screening and cleaning equipment, and the variety of
quite different principles of operation, makes it extremely difficult to
select the most appropriate for any particular situation. Because of the
problems involved in assessing the efficiency of screens and cleaners
when they are actually in operation, there is little to guide the papermaker
to a sound decision and as often as not a selection is made on the basis
of capital required, running costs, and reports of easy operation and
maintenance. Such information as is available on the comparative per-
formance of the various models, and what applications they appear best
suited for, is presented below in an attempt to clarify the position in this
respect.
No installation or alteration of cleaning and screening equipment
should ever be undertaken without having a clear idea of what impurities
(using the term in a general sense) are present in the stock, and what
standard of cleanliness is being aimed at. Impurities can take a multitude
of forms:
(a) pure contaminants within the pulp (shives, esparto roots and dust,
undisintegrated bark, pitch, unbleached clumps of fibre, undis-
solved lumps, etc.)
(b) other contamination of the pulp (grinder grit, boiler scale, sawdust,
rust particles, etc.)
(c) external contamination in the paper mill (dirt off bales, especially
grit and coal, bale wire, sand, sisal hairs, fresh water impurities,
loading impurities, rubber and plastic pieces, bitumen, airborne
dirt particularly boiler fly ash, etc.)
(d) internal contamination (pipe scale, rust particles, wood chips,
lumps of cement, beater and refiner metal, slime, oil, spanners, etc.)
(e) stock contamination (uncleaned pieces of broke, tangles and
strings of fibre, poorly dyed clumps, unrefined lumps, etc.).
No single cleaner or screen can be expected to remove all of these; in
any case certain types of contaminant are always more serious on a parti-
cular machine than others. Then again, the motive in wanting certain
impurities removed can vary: on one machine it may be purely a question
of obtaining as clean a sheet as possible, on another of reducing breaks
on the paper machine itself or afterwards during converting, and on yet
another of promoting wire and calender roll life by getting rid of abrasive
materials in the stock. In other cases it may be a matter mainly of over-
coming persistent blotches in the sheet which are caused by poor drying
round large particles; in bad circumstances such damp patches are dis-
103
torted in the calenders by a miniature creping effect or can be stamped

out completely to leave a hole.
The task required of a cleaner or screen on any particular machine will
frequently alter appreciably according to the furnish and grade being run.
Also the standard of cleanliness demanded varies enormously from
minimal on a grade such as kraft wrapper, important on newsprint and
printing papers, high on bleached papers, up to crucial on specialized
grades such as photographic base. It is thus unfair to expect any new
piece of equipment to solve all difficulties arising from impurities, parti-
cularly if these have not been carefully enumerated beforehand and some
idea obtained of just what has to be removed from the stock. Recognition
of the enemy by analysis of impurities found in the paper and in the stock
should always be the first step, if only because in many cases such impuri-
ties are produced in the mill and it can be far cheaper and more practicable
to prevent them entering the system in the first place.
104
CHAPTER 2A
2A. 1 THEORY OF SCREENING
No detailed theory of screening paper stock has yet been formulated,
nor has any investigation been reported which gives fundamental data
that could form a basis for a theory of screening. Details of the performance
of one or two models of screen with respect to particular types of con-
taminant have appeared from time to time, but to the author's knowledge
there has been no attempt under controlled conditions to discover the
effect of the various variables involved in the screening action. These
include such factors as throughput, rejection rate, hole variables (size,
shape, thickness), and oscillation characteristics, and their effect on dif-
ferent types of contaminant under different stock conditions of fibre,
consistency, wetness, and so forth. Nevertheless, certain general informa-
tion on the screening action is known and this is presented below.
2A. 1 1 The screening action; basic ideas

The action of screening is essentially one of size discrimination, but the
extent of this discrimination when applied to paper stock is by no means
easy to predict. Some understanding of what is involved can however be
built up from certain elementary considerations.
The simplest example of this form of separation is the passage of dry,
solid spherical particles falling under gravity through a sieve. Ideally
each particle of a diameter smaller than the diameter of the sieve hole
(or the side of the square formed in a mesh) should pass through and
the remainder be rejected. But in practice this is modified by the manner
in which the sieving takes place and it is instructive to consider one or
two aspects of this.
In the first place, a particle of small diameter falling onto the sieve will
be unlikely to pass straight through. If it were extremely small, the chance
of immediate passage will be equivalent to the open area of the screen,
i.e. if this is 75 per cent. then the particle will have a 75 per cent. chance
of passing straight through. Larger particles will have less chance than
this of being accepted because the greater contact area they present must,
to permit direct passage through, correspond with a free space in the
sieve; it is evident that this only occurs when the centre of the particle
meets the sieve in such a way that it is over a free space and at the same
time more than a radius length from the edge of the space. For example,
if the sieve contains circular holes it is easy to show that the probability
of acceptance of a spherical particle at a first pass can be expressed as
h (rs rp) 2
p = rs2 '
105
2A.12 SCREENS AND CLEANERS
where rp is the radius of the particle, rs the radius of holes in the sieve,
and h the proportion of open area. Under these conditions, for a sieve
with 75 per cent. open area the relationship between probability of first
passage and radius of the particle is illustrated in Fig. 2. 1.
In practice this probability would be slightly greater due to the possi-
bility of a particle glancing the edge of a hole but nevertheless passing
through. More important though is the effect on this probability of further
attempts (after bouncing) to pass through the sieve. An example of this
l·o
.ri...
PQ01'o~T-10N oF OP,,... A~EA~ 1!"4 S111vE.

- - - - - -·- - - - --- ---
OL-~~--lL...-~~--l~~~--1~~-=~~
0 o·2S O·S 0·75 l·o
i<ATlO e>t:TWEE-N PA.Q.TlCLE SIZ.! 6. ~OLE Yp/Y::.

Fig. 2.1. Relationship in an ideal situation between spherical particle size and proba-
bility of acceptance through a 75 per cent. open sieve with circular holes
for particles one quarter and one half the size of holes in the same sieve
is shown in Fig. 2. 2. This illustrates the growth of probability of acceptance
with repeated attempts to pass the sieve.
Two points of importance emerge from these simple examples. Firstly,
in an ideal screening operation the chance of a spherical particle being
accepted increases appreciably as its size decreases below the size of the
hole in the sieve. Secondly, this probability increases also with the number
of passes of the sieve attempted by the particle, or put another way with
the average residence time of the particle on the sieve. Given sufficient
time, the probability of any particle less than the hole size being accepted
becomes a near certainty.
2A. 1 2 The screening action; particle interaction

So far attention has been confined to the screening action on a single
particle. To take this a stage further it is necessary to consider the situation
106
SCREENS AND CLEANERS 2A.l 2
when there is a flow of various sizes of particles being screened as a con-
tinuous process. What effect is particle interaction likely to have on
screening efficiency?
It is evident for a start that the density of particles falling on the sieve
will affect the performance. The less the average space between particles,
the more likely their path is to be interrupted. However, this in itself
should not significantly alter the probability of an individual particle
passing through at any time because interaction between particles can be
presumed random. Density will affect the performance more directly
insofar as particles larger than the hole size tend to cover or lodge in the
!&J
~
~07s
...
Q.
~
-"'(
u. o·S
()
)>..
t
:! o·Zs
d}
<(
d)
(}
If. " ~.....1.-~-!.--!'-=T-".:--:!::--~--

l 2 .a 4 G. 7 /13
NuMe,ER OF ATTEMPTS To PAS!> SHWt:, n-.
Fig. 2.2. Relationship in an ideal situation between probability of acceptance of
spherical particles one quarter and one half the size of holes in a 75 per cent. open
sieve with the number of attempts to pass the sieve
holes and block them, thus reducing the area available for free passage;
the greater the density of a given flow, the more will larger particles gather
in the region of the sieve to restrict passage of smaller particles. The
extent of this interference will evidently depend on the proportion of
large particles in the flow, in other words on the size distribution, and on
the average length of time the particles are allowed to remain in the vicinity
of the sieve plate before passing out as reject.
Taken together, these points illustrate several fundamental considera-
tions in screening. Firstly, it is obviously necessary to agitate the flow of
particles either by vibration of the screen itself or, alternatively, of the
medium carrying the particles. Without some form of agitation the
holes of the screen will rapidly clog up. Secondly, the density and size
distribution of particles will affect the time that can be allowed for screen-
ing: if this is too long the region next to the screen will pack down with
larger particles and gradually inhibit flow altogether, even with reasonable
107
2A. l 3 SCREENS AND CLEANERS
agitation. As residence time on the screen determines the average number

of passes a particle can attempt and hence the probability of it getting
through, it is evident that the greater the density and the higher the
proportion of larger particles, the lower the average probability of accept-
ance becomes due to the necessity of reducing residence time to a level
where the screen continues to work.
It is at this point that recourse to experiment becomes necessary. In
any practical case, to find the best working conditions it would be necessary
to investigate the effect on the performance of varying residence time and
different forms of vibration. As residence time is increased it can be
expected that performance will at first increase but then, as the point of
clogging is approached, begin to decrease. This latter point would define
the maximum flow of particles the· screen could cope with in the conditions
pertaining.
Further considerations inevitably arise when the carrying medium
itself has a density comparable to the particles, i.e. in practical terms when
it is water rather than air. This introduces hydrodynamic complications
because the flow of water will channel through the holes of the screen,
tending to carry with it the smaller particles. This of course should be
advantageous. Also the transmission to the particles of vibrations applied
to the screen will be more effective in a denser medium.
2A . 1 3 Screening action on paper stock

The screening action considered so far has one important feature; although
there is always a certain probability of particles smaller than the holes
passing through, it is physically impossible for any larger particles to pass.
In screening terms this means that the cut-off is very sharp and exclusion
of all particles above a certain size (that of the holes) is a certainty. This
will always apply to rigid particles in any screen, but immediately it
becomes possible for the shape of the particle to distort or flex, the whole
character of screening changes.
Exclusion of large particles generally 'has far more importance in any
screening operation and if the cut-off is not sharp it is of special interest
to know the size at which the probability of larger particles passing
becomes negligible. This is a feature of any screening operation of flexible
particles that can only be determined experimentally because of the
numerous factors that come into play: the degree of flexibility of the
particles, the head above the screen (which affects the pressure producing
distortion of shape), the elasticity of the particles, and so on.
With fibres and shives in paper stock, the element of flexibility overrides
most others. The shape of a fibre can alter at will, especially when beating
has imparted a greater flexibility, and it is also known that fibres are com-
pressible in their cross-section. In theory it should be possible for any
fibre to pass through a hole of a diameter comparable to its cross-sectional
area, which in practical terms implies that all fibres would pass through
the holes used in commercial screens.
The reasons this does not occur are basically twofold: the element of
108
SCREENS AND CLEANERS 2A.l 3
probability introduced by the fact that only a proportion of fibres will
approach a hole endways on, and the interaction of fibres caused by their
tendency to entangle mechanically. An analysis of the first point has been
presented by Tirado (24) who calculated that the probability of a rigid
fibre or shive of small cross-sectional area passing through a hole during
screening is governed by its length in such a way that a minimum of one-
third will be accepted whatever the length; when the length is less than
the diameter of the holes this fraction increases. Interaction of fibres
reduces the probability of acceptance when it results in their being en-
tangled into a relatively large mass; the vibration of a screen plate should
assist in practice to break down such clumps of fibres, but if this does
not occur it is preferable anyway that they are not accepted. Interaction
of fibres is also affected by consistency, being of course more likely when
this is high, so fibres have a greater probability of acceptance at a lower
consistency.
To summarize, screening efficiency on paper stock must by its nature
represent an extremely imperfect operation. Ideally all individual fibres
should be accepted and all impurities above a given size held back. But
because of the limited capacity of a screen it is not possible to allow
adequate residence time for each fibre to have sufficient opportunity to
pass through a hole. So a proportion must be rejected, together with
unbroken clumps of fibres which, though representing good fibre, it is
necessary to remove anyway.
Likewise the probability of rejection of different impurities is governed
by numerous factors, most important of which are the size, rigidity, and
shape of the impurities, and the residence ·time on the screen plate; in
most cases residence time is preferably kept short to reduce the chances
an impurity has of getting through the holes. Solid particles larger in
size than the diameter of holes in the screen are certain to be rejected;
but other particles will always have some measure of probability of getting
through with the accepted fibre.
A typical curve relating impurity size to probability of rejection is
shown in Fig. 2. 3. The minimum rejection for very small particles is
seen to be effectively equal to the volume of flow in the reject as a propor-
tion of the inlet flow (here about 5 per cent.), and as particle size increases
complete probability of rejection is eventually approached. The slope of
the upgoing part of the curve is a measure of the sharpness of cut-off:
the steeper this is the better from a screening efficiency viewpoint. The
50 per cent. probability point (A) is often taken as a convenient measure
of efficiency of rejection for comparative purposes.
A final element is the shape of hole: this is frequently in the form of a
slot rather than a circular hole. It is evident that for the same open area
this gives a smaller minimum cross-section, which implies that the cut-off
point for rigid spherical particles will be lower. On the other hand rigid
particles tending to be fiat in shape will have some possibility of acceptance
so long as their maximum cross-sectional diameter is less than the width
of the slot, whereas with circular holes if the diameter of the hole is smaller
than the maximum diameter of the particle passage is impossible. Using
109
2A.l 4 SCREENS AND CLEANERS
slots instead of circular holes thus favours removal of spherical-shaped

particles, but makes it more likely that fiat shapes, especially in the form
of shives, are accepted. This is of importance, as will be seen later, when
considering the relative efficiencies of current types of enclosed screen,
which generally have circular holes in the screen plates, compared to the
older open rotary types which invariably have slots.
2A. 1 4 Screening capacity

Whatever the conditions of screening, the most fundamental factor con-
trolling performance is obviously the size of holes in the screen plate.
The ideal situation is to choose a diameter of hole which will cut off
A. Is THe.. PARTlcl..E..
SIZE WITH So,% C~ANCE

OF l<E.JE.CTION
I
I
OL---------------7'-------------------
A
PA~TlcLe. StZ.E ----'> 1Ncsu:A.s1NG.
Fig. 2. 3. Typical relationship between probability of rejection in a screen and size of

impurity
impurities above a size which is well under that acceptable for the machine
and paper. Unfortunately, this is rarely possible because of the effect of
hole size on the capacity of a screen.
Even if the overall open area of a screen plate is kept constant, reduction
in the diameter of holes brings about a rapid drop in the flow that any
particular screen can cope with. This is primarily due to the greater
tendency for holes to get clogged with impurities, coupled with a dewater-
ing effect as fibres clump together at the entrance to holes. Vibration
reduces this tendency, which is one reason why high-frequency vibrating
screens have greater capacity, but there must always be some limitation
110
SCREENS AND CLEANERS 2A.I 5
to flow dependent on the nature of the screen and composition of the
stock it has to deal with.
Of particular importance in this respect is the manner in which impurities
are removed from the screen plate in the reject flow. If this reject flow is
effectively uncontrolled, as in most flat screens, reducing the hole size
(other things being equal) has the effect of increasing the volume of
reject. If the reject flow is controlled by some valve setting, reducing the
hole size need only affect the volume insofar as pressure distribution in
the screen may be altered; but the effect on the screen operation is to
increase the power needed for the vibrating action and, in an enclosed
system, the pressure loss across the screen plates. Eventually, if hole
size is reduced sufficiently the screen either overflows if it is of the open
type or seizes up if enclosed.
So inevitably some compromise is necessary in choosing hole size.
The possibility of clogging or overflowing must be avoided, and where
reject flow is effectively uncontrolled it is essential to be able to set the
screen to keep this to a small amount. This is because once the point is
reached where the screen can function satisfactorily no increase in effici-
ency occurs by having a large reject flow: the ability to remove impurities
is virtually unchanged because conditions at the actual site of the holes
are the same, and all that happens is that a greater quantity of good fibre
is passed to the reject. When screening is to take place at a higher con-
sistency, or with long fibres that are more likely to entangle and become
stapled between holes, then obviously the compromise must err more to a
larger size of hole.
This point is clearer when the case is considered where reject flow can
be closely regulated. A certain minimum reject flow is of course essential
to ensure that impurities are cleared from the screen plate and any tendency
to clog or overflow is obviated. But it is a fallacy to believe that greater
efficiency necessarily ensues from increasing this reject flow further. The
basic operation desired of a screen is to separate out particles above a
given size for a minimum loss of good fibre. In other words any alteration
can only be deemed to lead to greater efficiency if it produces a lower
proportion of impurities in the accept flow compared to fibre. It is rare
that increasing reject flow has this effect: certainly a greater number of
impurities may appear in the reject, but the proportional increase in rela-
tion to the fibre rejected is almost always lower so the cost of separation
is proportionately higher and the efficiency consequently lower. Since
increasing reject flow generally reduces efficiency in this way, it is only
really permissible if the extra loss is worthwhile in order to reduce the
overall number of impurities occurring at a particular time.
2A. 1 5 The secondary screen
So far, attention has been confined to the action of a single screen. In
practice this may of course comprise several identical units placed in
parallel to give sufficient capacity to treat the required flow. Whether there
are one or more individual units, when these treat the main stock flow
line they are henceforth termed the primary screen.
111
The reject flow from the primary screen can comprise between about
2 per cent. and 10 per cent. by volume of the inlet flow. This contains
too much good fibre to be allowed to go to waste, especially as the reject
flow is almost always at a slightly higher consistency than the inlet, so
an essential feature of any screening system is the arrangement for treating
the reject flow in some secondary screen. This has a vital role to play in
determining the overall efficiency of the system and yet, as Steenberg has
put it (4), 'Most manufacturers are happy to recommend other types of
screen for the secondary stage, provided their own design is used for the
first stage.'
Secondary screening equipment should be designed to cope with the
lower flows and higher impurity concentrations involved. The lower
capacity demanded generally necessitates the use of an entirely different
PA1<T1c1..E Size__,,... 1Nc~E..o..s1Nc;;.

FIG. 2.4a
Fig. 2. 4. Effect on overall primary screen efficiency of using secondary screens of
different efficiency
type of screen and it is unfortunately the case, as will be seen later, that
the type commonly used for this purpose is excellent for fulfilling the duty
of reclaiming fibre but only at the expense of having an extremely low
efficiency of impurity removal.
Normally the secondary screen accept is returned to the main stock
flow at a point ahead of the inlet to the primary screen. The efficiency of
the secondary screen is thus highly important in determining the overall
screening efficiency of the system. It is therefore interesting to examine
the effect of different secondary screen e.fficiencies on the overall perform-
ance. This is shown in Figs. 2. 4a, b and c.
The primary screen is here assumed to have a typical relationship
between the probability of rejection and the incoming particle size (Fig.
112
SCREENS AND CLEANERS 2A.1 5
2. 4a). This can be considered to apply for a first pass of stock containing
an even distribution of particle sizes. The relationship between probability
of rejection and incoming particle size in the secondary screen (Fig. 2. 4b)
is assumed to be higher in curve A, i.e. the efficiency of the secondary
5£CONOARY 5c~~!N
z l·o A. Moret. ,f(ici~nl-:
0 t;. S.aMci. e{fici~ncy
tJ o.S C. l~s cz,{(ici~nl"'
,
Ill
uJ
<t
O•W
u..
0
PA.QTlCLE S1z.i= __,,,_ INCR.E.A.SING.

FrG. 2.4b
Pi<1MAR.Y Scg,e.eN.
ovao..lf a.ff i cie.ncy.
PActT1cLE Size: ---tNc..Rt:.As1Ni;.
FIG. 2.4c
screen is greater than the primary, to be the same in curve B, and lower in
curve C. The resulting effect on the overall efficiency of the primary screen
when a secondary screen having these different efficiencies is added to the
system is then illustrated in Fig. 2.4c, in which for comparison the original
primary screen curve is shown dotted. (The overall probability of rejection
for any given particle size can be readily shown to approach a value given
113
by the expression ps/(1 - p + ps) where p, s are the probabilities of

rejection in the primary and secondary screens, respectively). Reject flow
is taken at 10 per cent. of the inlet in both screens and fibre reject from
each system is the same, i.e. the efficiencies being considered are effectively
those relative to fibre. The 50 per cent. probability points are marked
with a circle.
These graphs show up several interesting points. Firstly, it is clear that
however good the secondary screen efficiency, recirculating its accept to
the primary screen inlet creates a reduction in the overall performance.
This is particularly the case with smaller particle sizes where in fact little
separation at all occurs and the probabilities of rejection are governed
almost entirely by the split to accept and reject in volume flow from the
system. For the larger particle sizes the efficiency of the primary screen
largely determines the overall efficiency, but for particle sizes in between,
the relative efficiency of the secondary screen obviously has a profound
influence on the overall performance.
When a relatively high-efficiency secondary screen is used, curve A,
the probability of rejection of large particles in the primary screen is
hardly altered and the 50 per cent. probability size is only very slightly
increased. This is certainly the most desirable state of affairs for screening
because by this means fibre is saved for a negligible reduction in the
efficiency with which the primary screen removes larger sizes of impurity.
In this respect it is interesting to note that Tirado (24), in one of the few
attempts to analyse this problem, advocated that the hole size of a second-
ary screen should be equal to the maximum width of impurity acceptable
whereas the primary screen hole size should equal the maximum length
of impurity acceptable. This could well provide a useful guide for laying
down screening equipment, since it ensures that the effective efficiency of
the whole system is that of the primary screen, and the secondary screen
only mars this efficiency to a minor degree.
Using a secondary screen of similar efficiency to the primary, curve B,
causes a noticeable reduction in the probability of rejecting the larger
sizes of particles, and with a poor secondary screen it is clear that the
50 per cent. probability point is moved to an appreciably higher size of
particle. It is not possible to give quantitative values to these curves, which
are purely for illustration. But they do emphasize the disadvantages of
using secondary screens especially of the older type that often have an
efficiency approaching virtually nil; their only purpose is then in fact to
save fibre and, by returning to the primary screen inlet practically all the
impurities originally separated out, a build-up in concentration occurs
which eventually means that the impurities are taken into the accept flow.
In this event only those particles for which the probability of rejection
in the primary screen is almost 100 per cent. will be removed. This gives
a sharp cut-off in particle size rejected, perhaps the only desirable result,
but as it must occur at a level so high as to command near certainty of
rejection it implies all but the largest particles will pass to the accept.
Efficient secondary screen operation is thus seen to be of great import-
ance and a reasonable conclusion to draw is that the efficiency should be
114
SCREENS AND CLEANERS 2A.16
at least equal to that of the primary unit. The performance of a typical
secondary screen can often be improved by diluting the primary reject
flow to a much lower consistency than the main inlet flow; this is generally
a feasible proposition as the basic flow involved is small and the size of
the secondary screen still need· not be great. Backwater or whitewater is
used for this purpose and the dilution can be direct into the primary
reject flow or through sprays which assist in the secondary screen opera-
tion. But taken by and large it is probably fair to say that the vast majority
of screening installations in paper mills are giving only a fraction of the
performance they could if more attention were devoted to the design and
operation of the secondary units.
2A. I 6 Treatment of secondary screen reject flow

The reject from a secondary screen may contain such a negligible quantity
of useful fibre and a very high proportion of impurities that it is worthless
and can be taken to drain. On the other hand an efficient secondary screen
system could well require quite a high reject flow wl:iich of necessity
carries with it too high a quantity of good fibre to waste. There are then
various courses of action possible.
In the first place this secondary reject flow can be returned to an early
part of the preparation system in the hope that passage through beaters
or refiners will reduce the size of impurities to a level that can be accepted
in the paper (this applies primarily to the fibrous impurities, which screens
are more efficient at removing, not metallic for which other considerations,
in particular that of wear on moving parts in the system, have to be taken
into account). Before doing this it may be necessary to have a coarse
screen that separates out really large impurities that could cause damage
or are impossible to break down small enough, but this need involve
only a negligible fibre loss. The ultimate effect of recirculating rejects
must be to increase slightly the proportion of fine impurities in the sheet
though, depending on the efficiency of the primary screen, removal of
larger impurities should not be unduly affected. This is probably acceptable
where the general appearance of cleanliness in a sheet is not so important
compared to its ability to withstand some converting process such as
printing in which weak points at the site of large shives in the web can
cause breaks.
An alternative procedure possible in mills where several machines are
running different grades is to use secondary screen reject from one machine
on another making a very coarse grade. Where several machines are on
similar grades it is possible to channel together all secondary screen
rejects (possibly also including the rejects from cleaning systems) to a
single tertiary screen. As the inlet flow involved will be quite small, even
when diluted well down, this screen can be of a relatively fine hole size
and operate in such a way as to allow a large residence time on the screen
plate. In this way it should be possible to recover most of the good fibre
with a high efficiency of separation.
Secondary screen rejects often contain high proportions of shives,
115
2A.1 7 SCREENS AND CLEANERS
especially of course when groundwood is used in the furnish. This is one

indication that a primary screen is working well because removal of shives,
being basically a question of separation by size, is very much the function
of a screening system. In appropriate cases, a secondary screen reject
flow containing many shives could be returned to the pulping plant, or
alternatively it could be treated either continuously or by batch in a special
kind of defibrator designed to break up shives. In some cases adequate
treatment may be possible in the normal beaters or refiners of the prepara-
tion plant, as mentioned above, though the danger here is that larger
shives start appearing in the paper and to overcome this the beaterman
is instructed to treat the stuff harder, a course which may not be appro-
priate for other reasons.
2A. 1 7 Return of secondary screen accept flow

Normal screening procedure is to return the accept flow from a secondary
screen to join the inlet flow to the primary screen. Addition can be either
direct to the screen inlet, or to a convenient backwater pit at a point
chosen to ensure the flow enters the mixing pump, or into the appropriate
compartment of a mixing box. There is nothing to choose between these
positions and in practice the one most convenient to suit the system is
selected.
It is definitely not advisable to allow the secondary screen accept to
join the accept stock from the primary screen. The reason for this has
been most clearly enunciated by Steenberg and Almin (4, 5) in their
analysis of the coupling of screens. Streams of different cleanliness, i.e.
containing different ratios of impurities to fibre in the stock, should not
be mixed if the best screening efficiency is desired. The inlet flow to the
secondary screen has a much higher content of impurities than the primary
screen inlet (since it carries all those rejected), hence when the efficiency
of both screens is similar the secondary screen accept can be expected
to contain about the same impurity content as the primary inlet and so
this is the position to return it to. It is of course possible that the secondary
screen is so efficient it produces an accept flow with as low a dirt concen-
tration as the primary accept flow, but this contingency is so unlikely
with present-day secondary screens as to be ignored.
Steenberg also pointed out there is no objection to mixing streams of
different consistency, and in fact where applicable this is useful to minimize
the overall water flow and, thereby, pumping costs of the system.
Where a screening system is too large for the stock flow there are two
alternatives; either a proportion of the flow can be recirculated to fill
the screen, or the screen can. be reduced in capacity by blanking off
appropriate portions. (If the secondary screen is over-capacity it may be
possible to increase the primary reject.) In theory the preferable alternative
will be the one that best adheres to the rule that only flows of similar
cleanliness should be mixed. In practice, depending on the type of screening
system involved, the alternative cheapest to run will be adopted, provided
the cleanliness level in the paper is satisfactory.
116
Occasionally the efficiency of the secondary screen can be so low that
when coupled with a barely adequate primary screen efficiency the level
of impurity removal is unacceptable. Short of making a complete change
in equipment there is one course which is sometimes advocated to help
improve this situation, that is to return the whole secondary screen
accept to the preparation system in the hope that further refining will
reduce the size of the impurities to a more acceptable level. This is similar
in principle to the re-treating of screen rejects earlier in the system as
mentioned above, but an important difference arises due to the relatively
large flow involved.
If the wet-end flow system, and particularly the reject flow from the
primary screen, is stable in volume and consistency, then leading the
secondary screen accept away from the machine system need have no
deleterious effect on the stability of substance of the paper. But if flows
are liable to fluctuate and consistencies to change, then removal from the
wet-end of a fair proportion of the total fibre in circulation could create
problems in substance variation and may not be desirable for this reason.
It is particularly hazardous to attempt to re-route secondary screen accept
flow from one position (say the primary screen inlet) to another (say
returning to the preparation system) while the machine is running because
it will be found that this operation causes a complete change to the sub-
stance. These and similar matters have already been discussed in more
detail in lA. 2.
2A.2 THEORY OF CLEANING

Whereas the operation of a screen depends on separating particles of
different size, cleaning equipment operates by separating particles of
different specific gravity. Some overlapping of function inevitably occurs
because heavier impurities will often also be large in size, but by and large
the nature of particles separated in the two pieces of equipment is quite
distinct. The normal screen will not remove small heavy impurities such
as grit, coal, scale, rust, and sand, except insofar as there is some region
in the screen where a natural settlement has the opportunity to collect
a proportion of these particles. Similarly the normal cleaner will not
remove large clumps of fibres and shive, undissolved pieces of broke,
sawdust, sisal hairs, etc., except insofar as some of these naturally follow
the flow to reject.
In this section the principles of separation by specific gravity are dis-
cussed first. Then, because cleaners of the cyclone type are always operated
in several stages to reduce fibre loss, the theory of coupling these together
is dealt with.
2A. 2 1 Separation by specific gravity

Separation of particles in a fluid by specific gravity depends on two basic
factors. The first is the application of external forces which have a different
5 117
effect on the fluid and on the particles, causing a relative motion between
the two. With conventional cleaning equipment this is always achieved
by constraining the fluid to rotate in a circular motion, thus applying
centrifugal forces. The second factor is resistance to motion through the
fluid which the particles experience. This is primarily dependent on shape,
and the way this affects drag on the particle, but is also affected by whether
the flow is streamline or turbulent.
Analysis in any particular situation of the relative motion of particles
in a fluid can be extremely complicated, but as an example the equation
which applies for spherical particles in a streamline flow under the influence
of a simple centrifugal field is u = v2 (Ps - p) d 2/18vr where u is the
relative velocity between particle and fluid in equilibrium conditions,
v is the velocity of flow of the fluid in a circle radius r, Ps and p are respec-
tively the densities of the particle and fluid, d the diameter of the particle,
and v is the coefficient of viscosity of the fluid. This equation shows
immediately why separation occurs in accordance with specific gravity
for the greater the value of Ps, the higher is the value of u. Size of the
particle is also important, for the relative velocity u increases in proportion
to d 2, effectively the cross-sectional area of the particle.
Two other points are worth noting from this equation. Firstly, the
separation effect is proportional to v2 , the square of the velocity of the
fluid. This shows up the tremendous advantage of high velocity of circula-
tion and explains why greater inlet pressure to a cleaner always brings
about an improvement in performance. Secondly, the separation effect
increases when the fluid flows in a circle of smaller radius, r. This means
that, other things being equal and in particular velocity being maintained,
the smaller the diameter of the walls constraining the fluid into a circular
motion the better. Hence the reason why smaller diameter cyclone cleaners
are more efficient. It may be observed further that an increase in viscosity,
v, of the carrying fluid decreases the separation effect; in practice this
becomes relevant as consistency increases. Practical examples illustrating
these points will be given in 2A.24 and 2B.65.
It is not difficult to arrange for fluid to be constrained to flow in a
circle. All that is required is a container constructed with cylindrical
symmetry in order to promote rotation; the force of this rotation is then
encouraged by injecting the flow at a tangent to the outer wall. Older
types of cleaner worked on a different principle which involved rotating
the whole outer wall to impart rotation to the fluid, but the inadequacies
of this from a constructional point of view are fairly self-evident and here
attention is confined essentially to the type of cleaner that depends for
its operation only on force induced by inlet pressure on the fluid.
Fluid injected into a cylindrically-shaped container in this way can
readily be drawn off at a point near the axis, leaving a small reject flow
to carry off heavier particles concentrated against the outer wall to which
they are thrown in accordance with the principle discussed above. There
must be an inward flow of fluid from the tangential inlet close to the
outer wall towards the central axis where it leaves the container; for a
particle to travel outwards to the outer wall the relative velocity (u in the
118
equation) must exceed this natural inward velocity of the fluid. The
typical fluid flow in a cleaner is shown in Fig. 2. 5.
2A. 2 2 Type of vortex

The important question with the flow in any cylindrically-shaped container
is, what is the relationship at various points within the container between
the velocity in the direction of rotation (inward and axial motion are
Fig. 2. 5. Illustration of the main flow in a typical cylindrically-symmetrical cleaner
generally relatively small) and the distance from the central axis, in other
words between v and r in the equation quoted above. This is particularly
vital because the separating effect has been seen to be proportional to
v2 /r. If, as the fluid flows inward from the wall and r decreases, the velocity
v increases as well, then it is evident that v2/r increases so the effect of
centrifugal force builds up as the flow passes inwards and the separating
effect increases. But if on the other hand v decreases as the central axis is
119
approached, it may be that v2 /r, and therefore the separating effect,

actually diminishes. The result of this would be that unless a particle
were subjected to sufficient centrifugal force immediately on entering
for it to receive an impetus starting it towards the outer wall, the oppor-
tunity to avoid being drawn into the accept flow would be passed as
thereafter the centrifugal force would diminish.
These two extremes are met in practice in the form of what are known
as an open or free vortex, and a closed or forced vortex. In the free vortex
the velocity v is inversely proportional to the radius r, i.e. v k/r, k
constant; in other words angular momentum is conserved and the fluid
rotates faster and faster the nearer it gets to the centre. The separating
effect in this case is proportional to 1 /r 3, and so increases tremendously
as r decreases. In the forced vortex fluid rotation is like a solid and the
velocity v is directly proportional to the radius r, v = kr. Here the separat-
ing effect is proportional to r which implies that a decrease in r reduces
its strength.
It is evident that it is more desirable for flow in a cleaner to take the
form of a free vortex than a forced, and fortunately fluid has a natural
tendency to form a free vortex. In practice the flow falls somewhere
between the two, because the free vortex is modified by several factors
such as inner friction and viscosity, the drag of the containing wall, and
lack of complete smoothness in the flow. If the general relationship between
v and r is expressed as vrn = constant (n + 1 free, - 1 forced vortex)
then as an example it may be mentioned that various workers have
estimated n at between 0·5 and 0·8 in cyclone cleaners. But in fact n
appears to be dependent on r in such a way that there is a transition from
a more free vortex near the wall to almost a forced vortex at the centre
along the cyclone axis. The whole matter is thus rather complicated.
2A. 2 3 Secondary flows
So far attention has been confined to the circular or tangential motion
which is the predominating one in any cylindrically-shaped cleaner. But
in addition to this there are two other important secondary motions and
it is upon these that much of the efficiency of a cleaner depends. These
secondary motions are essentially a function of the geometry of the cleaner
and both affect the initial separation induced by centrifugal force.
The basic problem is to arrange for all the heavy impurities thrown to
the outer wall of a cleaner to go one way, and all the fibre to go another.
Usually this is solved by using an orifice sited in the top of the container
over the central axis for removal of the accept flow, and a second orifice
at the bottom, which is either also concentric with the axis or tangential
to the wall, for removing the reject flow. Direct flow from inlet to accept
is generally minimized by some form of cover or 'vortex finder' round
the accept orifice. The reject flow can be direct from the bottom of a
cone (as in a cyclone cleaner) or from a specially designed baffle arrange-
ment; in some cleaners the reject flow is to a fixed vessel which gradually
accumulates rejected material, but more commonly, at least in a primary
cleaning stage, there is continuous evacuation.
120
The main secondary flow brings about this separation to accept and
reject; for a cyclone cJeaner it is shown in Fig. 2. 6 (i). The reject flow is
almost entirely from the boundary layer close to the wall, a comparatively
slow-moving downward flow which retains the larger impurities thrown
into it. Any disturbance to the smoothness of this flow (such as would
be caused by undue roughness of the wall) induces impurities to be swept
upwards into the accept flow. In some models of cyclone, elutriation water
Acee.PT.
INl..ET
(i) (ii)
Fig. 2.6. Diagram illustrating secondary flows in a cyclone cleaner: (i) separating
accept and reject; (ii) double-eddy motion
is injected tangentially near the apex of the cone to dilute the flow and
reduce fibre rejected; there must be a distinct danger in this of disturbing
the smooth flow of impurities down the wall unless the velocity of entry
is carefully matched to the natural rotational velocity in the cyclone at
that point.
Superimposed on this essential secondary flow is another which can
take various forms depending on the shape and design of the cleaner.
This flow is invariably deleterious to cleaner operation because it spoils
the straightforward tangential spin and separation of accept and reject
flows which set the main level of efficiency. Unfortunately little is known
about the form this flow can take except that it is difficult to prevent;
141
baffles placed inside the cleaner to reduce its intensity invariably seem to
have the opposite effect. To the author's knowledge a thorough investiga-
tion has only been made for the cyclone cleaner where optical observations
of aluminium and perspex spheres flowing in water through a glass
cyclone first brought the existence of this particular secondary flow to
light.
The form this takes in a cyclone is shown in Fig. 2. 6 (ii). There is first
of all a direct flow from the inlet down the side of the vortex finder sur-
rounding the accept and straight into the accept flow. This is a particular
source of inefficiency because in this region hardly any separation occurs
at all. The other flow is a double-eddy motion which is probably closely
related to the cyclone shape and can be expected to take a completely
different form in a wholly cylindrical cleaner. This double-eddy, apart
from being wasteful in energy, is also undoubtedly a source of inefficiency
because it disturbs the cleanness of separation to the accept and reject in
the main flow.
Another feature of the flow which has been observed in cyclone cleaners,
and may also be present in other types if the spin is strong enough, is a
central column free of water (this is shown dotted in Fig. 2. 6). This
column is effectively at zero static pressure and is of a diameter which
is dependent on the size of the accept and reject orifices. It does not contri-
bute in any way to the efficiency of separation in a cyclone, being purely
a result of the extremely high velocities reached near the axis. But unfortu-
nately, being at near zero pressure, an insuction of air takes place into this
column when the reject orifice is open to atmosphere. This adds to the
free air in the flow, presenting in the case of paper stock an undesirable
feature. On the other hand, the existence of the low-pressure column
can also be expected to cause air already in the inlet flow in a free form to
boil out. Efficient removal of air from this column can thus be expected
not only to prevent insuction but to reduce to some extent the free
air already in the flow. This will be discussed more fully in 2B. 63.
2A. 2 4 Examples of separation achieved

Most investigations into the operation of cleaning equipment have been
with the cyclone type, and some of the results obtained will now be pre-
sented to illustrate the effect of varying conditions. The purpose of this
is to show the practical effect of those changes considered earlier from
the purely theoretical standpoint. The basis of comparison is the percentage
of particles in the inlet that pass out in the reject. This is analogous to
the probability of rejection in screening equipment and it will be noted
that there is a basic similarity to the shape of curves presented in Figs.
2.3 and 2.4.
Fig. 2. 7 shows the typical effect on efficiency of rejection with increasing
particle size for three different sizes of cyclone cleaner. The curves relate
to spherical-shaped, heavy particles. The relatively rapid increase from a
minimum level of rejection for very small particles (determined by the
122
volume percentage of flow in the reject) up to 100 per cent. removal
indicates the relatively sharp cut-off in size which occurs, an advantageous
feature of cyclone cleaners. The reason for this is partly the relatively
efficient nature of the separating flow in a cyclone and partly the depend-
ence of the separating effect on the square of the particle diameter, as
discussed in 2A . 2 1.
Increasing the diameter of the cyclone with other conditions constant,
in particular pressure at the inlet, reduces the efficiency of the cyclone for
all levels of particle size. The effect is quite marked and has been con-
firmed in countless different experiments. It results entirely from the fact
lo
z
0
- So
}-
\J
.,
ut
Ill
Cl cOo
h.
()
t4o
z
u
u.: Zo
u.
w
PARTlc1..e Size__.,.... IN.c.Re.A.saNa
Fig. 2. 7. Effect on efficiency of rejection of near-spherical particles in different dia-

meters of cyclone cleaner
that, though velocity at the cyclone wall is the same in the larger cyclone
for equal pressure at the inlet, the radius of rotation is larger and this
reduces the separating effect proportionally.
The effect of increasing specific gravity (not shown) is to shift the whole
curve to the left. Grit and iron particles are rejected more efficiently than
ash and sand of the same size because the latter have lower specific gravity.
Fig. 2. 8 shows a different aspect on the rejection efficiency of cyclone
separation, the effect of particle shape. If instead of being near-spherical,
particles are of a flattened shape, then the curious effect illustrated occurs;
increasing overall size of the particle leads first to a normal increase in
rejection efficiency but then to a sudden tapering off and fairly rapid
decrease in efficiency. The precise reason why this occurs is unknown,
though since particle shape has a close influence on the drag coefficient
it is hardly surprising that the curve is different from that for spherical
123
particles. But exactly why a higher proportion of large flat particles pass
through to the accept has not yet been explained.
The fact that it happens, however, is highly important when the rejection
of shive-like material is considered, especially as this effect is very pro-
nounced for material with as low a specific gravity as shives. Brecht (19)
has estimated that it applies very strongly when the ratio of cross-section
length to thickness exceeds about six to one, while Kemp and Rance (26)
have put it at about five to one. Fibres of course exceed this ratio, which
loo
l
0 So
...u
IJ
1
fol.I
~
,0
la.
0
'.)... 4i:>
u
z
~
u 2o
lL
lL
UJ
0
Fig. 2. 8. Effect on efficiency of rejection for flat particles of different sizes
is perhaps a fortunate reason why a higher percentage of these do not

appear in the reject; in this respect it is interesting to note that those
fibres found in the last stage of a typical cyclone system do tend to be
short and fine rather than long, a confirmation of the effect of particle
shape. This topic will be returned to in 28. 6.
2A. 2 5 Coupling of cyclones together

With many types of cleaner it is customary to utilize as a secondary stage
a specially-designed low capacity model similar to but smaller than those
used in the primary stage. In this case the same general principles apply
as for screens where it was seen that a secondary unit should be at least
as efficient as the primary if overall performance is not to be severely
prejudiced.
With the cyclone cleaner it is more common for the same size to be
used throughout and as a consequence three and even four stages of reject
treatment may be needed to reduce the final fibre loss to an acceptable
124
degree. Although at first sight this appears to lead to a cumbersome
arrangement, it does ensure that there is no marked decrease in efficiency
of the secondary and later units. The whole problem of how such large
numbers of individual units are to be coupled together is then very import-
ant. This will now be discussed; the information presented has been drawn
from theoretical discussions presented by Steenberg and Almin (4, 5),
Nuttall and Hendry (18), Brecht et al. (19), and Corte (20).
In the first place, the capacity of an individual cyclone unit is relatively
small and to cope with the sort of flows common to paper machine stock
systems a large number linked in parallel are needed. For the purposes of
discussion such an arrangement can be considered equivalent to a single
treatment of stock with a common inlet, accept, and reject flow. The
primary accept flow could be led to a further unit of cyclones to establish
an even greater degree of cleanliness, but the author has heard of no mill
where this has in fact been done, presumably because of the heavy pumping
costs involved in double treatment. So effectively the most important
consideration is the manner of dealing with the reject flow. This is done in
what are commonly called a series of 'stages'.
The reject flow from the primary or first stage of cyclones should be
taken direct, after appropriate dilution (see next section) to the second
stage; the accept from the second stage is then returned to join the untreated
stock in the first stage inlet. The reject flow from the second stage may
contain a small enough quantity of fibre to be manageable, but normally
the same treatment is repeated in a third stage. In this case the accept
flow should return to the inlet of the second stage. Further stages can be
added and in each case the accept flow should return to the inlet of the
previous stage. The resulting arrangement is illustrated in Fig. 2. 9 and
this has been shown to be the most economic in terms of securing minimum
pumping costs and low fibre loss from the final stage.
It would appear at first sight that the addition of successive stages
should reduce the fibre loss from the final stage reject to as small as
desired. Although re-treatment of reject flows in a further stage after
the second does affect the overall efficiency of impurity removal (just as
was seen from the case when a secondary screen is introduced) Nuttall
and Hendry have shown that when the efficiency of a single pass is high,
as it is for heavier and larger dirt particles passing through cyclones, then
the change in efficiency of the whole system resulting from addition of a
third or fourth stage is in fact so small as to be negligible. This applies
even if the efficiency of the final stage is reduced considerably to minimize
fibre loss, e.g. by using elutriation water or an enclosed reject collecting
chamber. As an example, their work indicates that even if the third stage
of a cyclone system has only 30 per cent. efficiency instead of 90 per cent.
as in the first stage, the overall efficiency in terms of impurity removal
relative to fibre in the primary accept drops from only 88·9 per cent. to
86·6 per cent.
But treatment of rejects in too many stages is not practicable for two
main reasons. Firstly a certain flow is required to fill a single cyclone at
normal inlet pressure, so the size of the final stage is limited. It would,
125
of course, be possible by diluting successive reject flows more and more

to keep adding cyclones in stages. But when this is done it is found that
the actual reduction in fibre loss taking place becomes less and less because
beyond a certain point the more stock is diluted the greater is the percent-
age loss of fibre (see Fig. 2. 17 on page 166).
The second reason that the number of reject stages has to be limited
is due to the selective rejection of shorter fibres and fines that occurs.
\...,u.. e.-r
011..UT\ON
FINAl.. ReJ =.c:r.

Fig. 2. 9. Arrangement of cyclones in a three-stage system
This has the effect of increasing the percentage loss of fibre in successive
stages, even when the inlet consistency and the percentage volume flow
to the reject are unchanged; in other words the ratio of the reject consist-
ency to the inlet consistency, the 'thickening factor', increases. For example,
for an inlet consistency of 0·8 per cent. in each stage, the reject consistency
of the first stage might be 1·6 per cent., of the second 3·0 per cent., and
126
of the third stage 4· 5 per cent. This increase in thickening factor thus
diminishes more and more the advantage in reducing fibre loss gained by
adding a further stage.
For these reasons the number of stages usual in a system of cyclones
is normally three and only very occasionally four. Even then as much as
50 per cent. of the fibre entering the final stage can be rejected. To add a
further stage would cause so little further reduction in fibre loss that the
added capital cost and increase in pumping costs (an extra pump always
being necessary for each stage) cannot generally be justified.
2A . 2 6 Planning a cyclone installation

The arrangement of a system of cyclones into an appropriate number of
stages requires careful planning to determine how many units are needed
in each stage and the number of stages it will be economical to use. The
inlet pressure for running must first be decided on the basis of power
cost versus efficiency for the first stage (see 2B. 6 5), and from this the rated
throughput of a single cyclone under normal operating conditions can be
obtained from the manufacturer's specification.
Next it is necessary to have some idea of the percentage of flow which
will go to the reject. This depends entirely on the size of the reject orifice,
the vacuum applied if any, and whether there is any restriction to the
accept or reject flow. There is always a minimum reject size below which
operation of individual cyclones is not smooth, either because the flow
becomes uneven or erratic or because the hole is so small that it gets
partially or wholly blocked too easily. When planning an installation care
is taken to assume a reject flow appreciably larger than this minimum
because even if the cyclones are set up initially with the reject orifice
diameter at the minimum tolerable, as wear of the cyclone body occurs
the reject widens giving a higher flow. In this respect it is worth noting
that any cyclone works most efficiently, i.e. with minimum impurities
to fibre in the accept flow, when the reject diameter is as small as practi-
cable. So as reject orifices widen a cyclone system becomes less efficient,
largely because of the greater fibre loss resulting from the final stage. It
is therefore wise to avoid planning an installation that will be able to
cope with very large reject flows lest the gradual increase in fibre loss
over a long period goes unnoticed.
Finally, the likely degrees of thickening in successive stages must be
known or estimated for the average reject diameter it is intended to use.
If no comparable furnish has been treated before, experimental work
may be needed to obtain reliable enough information on this point.
But nowadays most manufacturers have sufficient data for an adequate
range of furnishes and beating treatments to make this unnecessary.
With all this information the numbers of cyclone units in each stage
and the number of economic stages can be determined, though the task
is not easy to do well. Even the expert generally relies on trial and error
though formulae have been developed by Nuttall and Hendry which
enable the job to be done systematically. Addition of dilution water
127
(usually from the machine pit) is necessary for each stage but the first
one to keep the inlet consistency down to a level where efficient separation
occurs; this is normally arranged so that inlet consistency is about the
same or lower in each stage, allowance being made of course for the
consistency of the dilution water which can be high with furnishes con-
taining short fibres and loadings. Fortunately, though calculation of the
numbers of cyclone units in this way is necessarily approximate, there is
considerable latitude in practice because more or less dilution water can
always be added in any stage to adjust the flow to suit the number of
cyclones. But reasonably accurate determination is still desirable other-
wise each stage will be installed oversize and in operation one or more
units may have to be blocked off and the pump run at an inefficient level
of efficiency.
The capacity required in the first stage depends on the flow of stock to
the breast box and the amount recirculated from the other stages. Both
are subject to much error in measurement or calculation and it frequently
happens that the capacity well exceeds the actual flow to be treated. To
keep the cyclone units full and working efficiently, it is then necessary
either to block off a sufficient number of units to obtain a reasonable
inlet pressure, or to recirculate a proportion of the flow. The more appro-
priate course depends on the demands of cleanliness for the paper:
recirculating increases the quantity of impurities removed, but also in-
creases the percentage of fibre lost from the final stage. According to
Nuttall and Hendry this occurs in such a way that if a proportion p of
the accept flow is recycled, the impurity content decreases by a factor
(1 p) but the fibre reject increases in proportion to 1/(1 - p). Also
pumping costs increase with recirculation, certainly when a variable-
speed pump motor is used for the first stage; even when the flow is simply
controlled by a valve on the discharge of a fixed speed pump the pumping
costs can also increase with recirculation depending on whereabouts on
the pump characteristic curves the running conditions fall.
Occasionally the cyclone installation is installed in the flow system
with a separate pump to the main mixing pump. This has the advantage
that power is not consumed if at any time it is possible to run the machine
without the cleaners operating. When there is some sort of buffer between
the two pumps, a separate chest or some other piece of equipment such
as a level-controlled deculator, this procedure is perfectly satisfactory and
gives a more flexible system. But on the other hand with a completely
closed system, operation of two pumps in series can present certain compli-
cations, especially with regard to stability, and automatic control of the
flow to the breast box is then advisable.
2A . 2 7 Minimizing fibre loss

It is possible to re-use fibre from the final stage of a system of cyclones
in one of a number of ways similar to those described for treating screen
rejects. This is not normally done, however, because it is found in practice
that the proportion of impurities to fibre in a final stage is so high that
128
re-use is never contemplated. Furthermore, the fibre that is present is
generally very short and of low brightness, and the impurities often of a
highly abrasive type which would create rapid wear if they were recircu-
lated to an earlier part of the system and allowed to build up. Because of
this it is all the more important to plan a cyclone installation to have as
low a fibre loss as possible.
Ways of achieving this have already been described and in particular
the limitations of adding more than three or four stages have been ex-
plained. Fibre loss can still even in a well-planned installation amount to
I per cent. or more of fibre passing through the installation, which means
an even higher percentage loss when this is expressed in terms of produc-
tion (because a proportion of fibre passing through the installation comes
from the backwater). So other methods of fibre loss reduction have been
sought. Most of them depend on using a different or a modified cyclone
unit or units for the final stage. These modifications, discussed in 2B. 6 3,
sometimes take the form of adding elutriation water close to the reject
orifice to reduce consistency there, or they can involve a smaller cyclone
or one with some special arrangement on the reject which allows the
fibre loss to be controlled.
Any of these modifications invariably not only reduces fibre rejected
but also to some extent the efficiency of impurity removal. It is rarely
possible to do the one without the other. However, it has already been
seen that the efficiency of a final stage can be reduced appreciably without
materially reducing the cleanliness of the main stock accept. So provided
adaptations to the final stage are not noticeable in terms of increasing
impurities in the paper, and this should normally be the case, then any
course of action along these lines is to be recommended.
One suggestion put forward by Nuttall and Hendry is that a proportion
of the final reject should be returned to the inlet of the final stage. Where
capacity of the final stage allows this, and the resulting increase in thicken-
ing factor does not offset the effect as eventually it will do if too much is
recycled in this way, then this seems to be a satisfactory and simple means
of achieving a reduction in fibre loss. If a proportion p of the final stage
reject is recirculated direct to the final stage inlet, then a reduction ap-
proximately proportional to (1 - p) in the fibre loss can be expected for
negligible change to the overall dirt removal efficiency of the system.
2A. 3 ASSESSING SCREEN AND CLEANER EFFICIENCY

In the preceding discussion of screens and cleaners, a great deal was
said about the 'efficiency' with which particles are rejected. This is a
simple term to use, but it can be interpreted in several ways. When new
screening or cleaning equipment is installed it is natural to expect some
assessment of its 'efficiency' to be made, if only to ensure it is operating
satisfactorily. How to do this in practice requires about the most pain-
staking experiment that a mill laboratory is likely to undertake. This
section is devoted to a few notes on this whole question of efficiency,
how it should be defined, and how it can be measured.
129
2A. 3 1 Definitions of efficiency

The difficulty about finding a satisfactory definition of efficiency is that
the performance of any screen or cleaner depends not only on how well
it operates in different conditions, which is what is of most interest, but
that it depends also on the type of impurity concerned. To be clear about
this, consider first the curves in Figs. 2. 3 and 2. 7 which show, for a screen
and a cleaner respectively, how the probability or efficiency with which a
particle passes out in the reject flow depends closely on the size of the
particle. A really comprehensive assessment of efficiency demands a
series of curves of this type covering different specific gravities and operat-
ing conditions.
This would be time-consuming and in most cases unnecessary because
the curves would all be similar. There are exceptions, as for example
with cyclone cleaners when the shape of a particle strongly influences the
curve. In this case a full assessment over a range of particle sizes is vital.
But otherwise from typical curves a number of parameters may be used
with which it is possible to extrapolate to other conditions without actually
performing the tests. These parameters are useful, especially for com-
paring different types of equipment.
Most common of the parameters in use is the size of particle which
has an even or 50 per cent. chance of leaving in the reject flow. As an
alternative, where particles of the same size but differing specific gravity
are considered, the specific gravity with a 50 per cent. chance of rejection
can be taken, but this cannot sensibly be applied to screens where specific
gravity has little influence on separation. An alternative and less easily
defined parameter is to take the largest size of particle within some common
grouping (for example between standard mesh sizes) which appears in
the accept flow, i.e. the largest size of particle which falls below 100 per
cent. reject efficiency.
Each of these parameters have their place, the 50 per cent. ones being
particularly useful for comparing performance. The maximum size of
impurity appearing in the accept is of importance for certain specific
papers, e.g. photographic base, and also from a purely visual standpoint
since a single large speck of dirt is far more readily noticed than a number
of smaller specks of equivalent total size. A simplification of this parameter
is to consider only the efficiency with which particles over a given size
are rejected. This too has a strong practical justification and can be related
to visual terms or to such things as the likelihood of causing a break on
the machine. It is certainly the easiest to assess, and probably for this
reason is a common one to use.
An equally important aspect of efficiency is the sharpness with which
separation occurs. A screen or cleaner which gives a very gradual increase
in probability of rejection as particle size increases, curve A in Fig. i. 10,
will generally be less desirable than one with a very sharp cut-off, curve B,
even though the cut-off occurs at a relatively large size. This is because
it is frequently more valuable to be certain of excluding impurities above
a certain size than to remove a higher percentage of smaller particles,
130
SCREENS AND CLEANERS 2A.3 l
an extension of the point made in the previous paragraph. The difficulty
here is that the 50 per cent. probability size as a measure of rejection
efficiency is inadequate, as comparison of the two curves A and B shows:
the 50 per cent. particle size is in fact smaller for A than B and so the former
might be thought the better though in practice the latter is almost always
preferable. This is one argument for preferring the definition of efficiency as
maximum size in the accept, and for the same reason the simpler definition
using rejection of some relatively large standard size of particle is also
satisfactory.
These then are some definitions of efficiency which are perfectly adequate
as they stand for assessing the removal of particles to the reject. But for
Fig. 2.10. Curves relating probability of rejection to particle size for equipment with
a poor (A) and a sharp (B) cut-off. Curve B will usually be preferable because it is
more efficient in removing the large particles in the shaded portion
papermaking what is important is the rate of removal relative to fibre.

The efficiency of a cyclone cleaner, defined in any of the terms above,
can be made to increase simply by opening up the reject orifice, which of
course causes the quantity of fibre rejected to increase also. There will
then be a lower number of particles of any particular size passing in a
given time to the accept, i.e. a higher absolute efficiency of rejection,
but relative to the quantity of fibre in the accept (and hence to area of
paper) there may well be more particles and hence true efficiency will be
lower.
The point here is that the efficiency of rejecting particles, however
measured, must take account of the quantity of fibre in the reject. If the
131
latter is always kept the same, or effectively the same as when the reject
volume is kept constant though small changes in consistency may occur
under different conditions, then efficiencies of particle rejection can be
compared directly. Otherwise a correction is necessary and this amounts
to defining efficiency as (reject per cent. of impurity reject per cent. of
fibre) x 100/ (100 - reject per cent. of fibre).
Alternatively, when the method of assessing efficiency is defined in
terms of the number of particles above a particular size, then for the same
inlet conditions a direct comparison can be made of the actual numbers
occurring in a fixed volume of accept flow provided each figure is multiplied
by a correction factor which allows for any difference in the actual flow
rate of fibre in the accept as determined from flow and consistency measure-
ments.
2A. 3 2 Measuring rejection efficiency

There have been surprisingly few reports giving details of experimental
as opposed to general machine tests of screens and cleaners, and most
of these have referred to cyclone cleaners. This is unfortunate because
reliable data is so hard to obtain from equipment in mill operation and
it is rarely practicable to investigate the effect of changing operating
conditions. Above all the opportunity of adding a controlled type of
impurity to obtain a really accurate measure of rejection efficiency is
not possible in a mill installation.
For any sort of development work and for comparing the efficiency of a
particular screen and cleaner at different operating pressures, consistency
hole size, and so on, it is essential to be able to detect and measure some
representative type of impurity when it is mixed in with the normal paper
stock. Various techniques have been devized for this, including the dyeing
of sawdust to represent shive-like particles, radioactive tagging of shives,
adding sand to the stock, or adding various other readily recognized dirt
particles. Counts on a suitable number of handsheets, together with
measurement of flows and consistency for the inlet and reject are then
adequate for determination of fibre and impurity reject percentages; the
usual difficulties of dirt counting need not apply in experimental work
of this nature since the size of impurity and its identification are chosen
to minimize complications. In the case of sand or loadings, ashing the
handsheets gives a convenient measure of the quantity of impurities in
any flow.
Normally, however, such tests are outside the range of a mill where
deliberate contamination of stock, particularly in a form that is readily
identified, is out of the question. It is sometimes possible to make sense
out of dirt counts done on handsheets made from stock sampled from .
relevant points of the system, but the difficulties are phenomenal. Use of
the modified T APPI dirt chart is helpful in reducing operator errors in
dirt counting and for checking different size groups of impurity, but
generally it is found that far too many counts are needed to obtain reliable
figures and a count on the same sheets can vary enormously from one
132
SCREENS AND CLEANERS 2A.3 3
operator to another. An alternative is to compare handsheets with a
standard set of carefully protected sheets containing various degrees of
dirt of the kind normally encountered in the grade of paper concerned,
but this allows only a general assessment applicable to the range of dirt
present in the system at any particular time and at best can give only a
rough comparison.
2A . 3 3 Assessing the worth of a mill installation
A mill installing new screening or cleaning equipment is primarily interested
in knowing if the money has been well spent. It is not adequate simply
to examine rejects for the type of impurity present because this is notori-
ously unreliable, depending as it does on the amount of dirt in the stock
system at the time of examination (and this obviously can fluctuate from
day to day) and on the amount of fibre associated in the reject with the
impurities. A comparison over a long period is essential and this can be
done in several ways.
Firstly samples of the actual paper produced can be examined and a
direct comparison made of the dirt content before and after installation.
Provided the installation does not coincide with some marked seasonal
change in the amount of impurities present, e.g. in pulp, slime, etc., a
reasonably reliable picture should emerge in a relatively short time if
any significant improvement has occurred. Sheets for this purpose must
be carefully protected in plastic bags, and counted at the same time in
randomized order to eliminate bias. A development of this approach is
to use a laboratory dirt counter such as the PAPRIC or even a dirt counter
on the machine. Provided the limitations of these instruments are accepted,
they can provide a more objective evaluation particularly in the case of a
laboratory tester where sheets obtained over a period can be tested one
after the other and instrument and operator error thereby minimized.
A less rigorous comparison is sometimes possible simply by comparing
the number of breaks occurring or reels rejected due to dirt and allied
causes such as blotches in the web. Similarly, sheets rejected for dirtiness
in the salle or on a cutter with an electronic dirt-detection sorter can be
used. With new cleaning equipment when improved separation of abrasive
particles in pulp and loadings is expected, comparison of wire life and of
the period between grinding of calender rolls can provide a useful indica-
tion of improved efficiency.
Shive removal is particularly important in some grades and determina-
tion of this is very difficult. Counting handsheets is especially hopeless
when dealing with shives and other methods have been proposed. One is
to screen samples from inlet and reject in a Bauer McNett fractionator,
taking the weight of the fraction greater than 14 mesh as representative
of the quantity of objectionable shives present (though it necessarily also
contains fibre). Another similar method is the British Standard using the
Sommerville fractionator. A third that has been reported is to use a mesh
of 5 to the inch in a fiat screen under standard conditions, continuing
the screening to exhaustion and collecting the rejects for weighing. Un-
fortunately, as Robinson and Kingsnorth (29) have reported, these various
133
methods often give contradictory results due to differences in the type

of shive separated out. A careful comparison by Macmillan et al. (35)
has confirmed that counting or weighing techniques for assessing shives
in newsprint are unreliable. These workers also reported that other
attempts relying on detection of shives by measuring gloss or using
X-rays and monochromatic light of various wave-lengths including ultra-
violet gave no marked contrast between shives and fibre; the conclusion
reached was that as yet there was no really satisfactory means of assessing
shive content accurately and in particular neither the 14 mesh retained-
fibre in the Bauer McNett nor the fiat screen technique appeared much
use because too many shives which could in practice be troublesome
passed through.
The most suitable method of assessment also depends on the size of
the shive that it is thought necessary or desirable to remove. In respect of
printing papers it is worth noting the work that has been done by Hopkins
(32), Macmillan (35), and Sears (36), and their respective colleagues.
Their reports show conclusively that shives are a major source of web
breaks in newsprint and the likelihood of a break is greater the larger the
size of shive. Such breaks can occur at tensions well below the average
tensile strength of the paper because a long, thin shive, particularly one
set in the cross-direction, provides a weakness due to the poor bonding
of fibres on to its surface.
To obtain concrete data on the size of shive responsible, a means of
assessing the 'runability' of the paper is needed and various types of
simulation tests have been tried. Running at high tension on a winder was
used by Hopkins et al. as a simple test for this purpose and they found that
50 per cent. of breaks occurring were attributable to groundwood particles.
of l ·5 sq. mm. or larger in area and 90 per cent. to particles 0·7 sq. mm.
or more in area. Following a different approach Sears et al. developed an
instrument for subjecting a 16 in. reel of paper to increasing strain and
showed by this means the extent to which shives can be responsible for
breaks (of 3,200 breaks, all but 45 had the fracture passing through one
or more shives); this work indicated that most of the shives responsible
were over about 3·5 to 4 mm. in length and of a width up to half the
thickness of the paper. The importance of this line of research to newsprint
and similar grades is very evident and more reports on this subject can
be expected.
134
CHAPTER 2B
OPERATING FACTORS AFFECTING SCREENS AND
CLEANERS
2B .1 POSITION IN THE FLOW SYSTEM
For discussion of operating factors affecting screens and cleaners it will
be convenient to differentiate between the main groups of equipment
available and treat each in a separate section. This course is adopted
because the behaviour of screening and cleaning devices is invariably
individual to the particular operating principle involved, and the effect
on performance of variables such as stock consistency, stock flow, impuri-
ties in the stock, and so on is in turn largely dependent on the particular
model concerned.
First, however, there are several general points concerning screens and
cleaners that are relevant whatever the models concerned. These relate
to the method of feeding stock to a unit and dealing with the accept and
reject flows, and also to the question of where in the wet-end flow system,
and in what order, screens and cleaners should be used. These subjects
will now be considered.
2B . 1 1 Position of screening and cleaning equipment

It is common practice when applying screens and cleaners to the treatment
of paper stock to choose a position in the wet-end flow system between
the mixing box or pump and the breast box. There are two reasons for
this. Firstly, in this position the consistency of the stock after dilution
with backwater is generally under 1 per cent., and it is only possible to
remove relatively small impurities at a consistency as low as this.
Secondly, the removal of impurities takes place immediately ahead of the
paper machine proper, so reducing the possibility of later contamination.
The only exception to this is with certain speciality grades where the
consistency of stock entering the breast box needs to be as low as O· l
to 0·2 per cent. In this case the extremely high flows involved make
screening and cleaning immediately ahead of the breast box an expensive
business and little is gained in added purity by performing the operation
at such a low consistency. Consequently an initial dilution from normal
refining consistencies to the region of 0·5 to 1 per cent. is sometimes
carried out for the purpose of screening and cleaning, while further
dilution for making the sheet follows in a second mixing pump. The
main disadvantage in this is that there is no protection against any impuri-
ties introduced in the second dilution, so scrupulous attention to cleanliness
of machine pit, pipework and the second mixing pump becomes highly
important.
135
2B.11 SCREENS A ND CLEANERS
Feed to the rotary or flat type of screen is frequently by open chute

from a simple mixing box where fresh fibre is mixed with backwater
either over adjustable weirs or by valves. The reject goes to a secondary
screen where it may be diluted by sprays using machine pit water or
lower-consistency water from the vacuum boxes or a save-all. The accepted
stock from this secondary screen is normally returned to the main machine
pit or to a suitable compartment of the mixing box, though it can be taken
out of the machine system and returned to the stock preparation to help
break down impurities by further beating and refining. With pressure
screens a similar system is used except that the feed is from a mixing
pump and pipework is used exclusively.
With this typical sort of arrangement it is always useful if a line to
by-pass the screening equipment is available because frequently this will
allow repair work to be carried out on the screen without stopping the
machine. Whether or not this is feasible depends of course on how critical
an operation the screening is: on some machines it is impossible to make
a satisfactory sheet without screens even for a short time, but for coarser
papers it is often possible to run for a while and suffer a temporary in-
crease in shives in the paper. When a number of individual screening
units are coupled in parallel the position is simpler and provision should
always be made to allow a single unit to be isolated with the flow tempor-
arily accommodated by the remaining units. In some cases an additional
unit is installed and kept specifically as a standby; it should then rarely
be necessary to shut the machine for repair work on the screens. For the
same reason, provision for a temporary diversion round the secondary
unit is also advisable.
Turning to cleaning equipment, with the more modern types a pumped
system is always essential due to the relatively high feed pressures required.
As with screens the reject is taken to a secondary unit or (in the case of
cyclone cleaners) to a number of fibre-recovery stages. Some dilution is
generally needed to ensure that the secondary and later stage units are kept
full and to reduce the primary reject consistency which is normally some-
what higher than the inlet; this dilution can take place in level controlled
open troughs or the system can be completely enclosed and pressures
balanced by valves in the pipelines.
The accepted flow from the secondary unit or the second stage of a
cyclone installation can be joined to the main accept flow from the primary
cleaner, but normal practice is to return it for a second pass through the
primary unit (as with screening equipment). Accordingly, the secondary
accept flow is usually returned either to the main pit, or to the suction
side of the pump feeding the cleaners. Alternatively it can go to the feed
side of a specially designed box which contains two compartments, one
for feeding the cleaners and the other for receiving the cleaned stock (the
point of this arrangement is to permit some balancing recirculation over
a weir between the compartments). Such boxes are carefully designed for
ease of cleaning and due to the high pressures developed entry pipes
should always be well below the normal operating level to reduce aeration.
As with screening equipment, a by-pass round the primary cleaners,
136
SCREENS AND CLEANERS 2B.l 2
isolation of secondary and later stages of fibre recovery, and isolation of
individual units in parallel are valuable installation features.
2B .1 2 Using both screens and cleaners
Screens are very often the only type of equipment used ahead of a paper
machine, especially for the manufacture of coarser grades where quite
large contraries in the paper can be tolerated. Occasionally cleaners
(usually of the cyclone type) are used without any screening equipment,
but this practice is only found in integrated mills where pulp is well
screened prior to entering the paper mill. However, it is becoming in-
creasingly common to use both screens and cleaners, the former primarily
to remove fibrous impurities and strings, the latter to remove heavier
and smaller impurities. This poses one or two special problems.
Using screens and cleaners together makes no particular difference to
the requirements and operation of each individual system, except that the
doubling of protection makes it more feasible that one or other piece of
equipment can be shut down temporarily for cleaning and repair, hence
it is even more important to incorporate by-pass lines as discussed above.
When a dump line is installed to divert the stock flow from entering the
breast box, this should be immediately ahead of the breast box after
both screening and cleaning equipment in order to allow the wire to be
stopped or the breast box examined without stopping either screens or
cleaners.
The main question confronting a papermaker when both screens and
cleaners are required is: which order are they put in? At one time screening
equipment was always placed immediately ahead of the breast box,
mainly because it was felt that only in this position could full advantage
be taken of the deflocculating action of screens. Also screens have a
lower operating pressure, which can better suit entry into the breast box,
and there is more opportunity for free air in the stock to be separated
out and removed. On longer-fibred stocks these points still apply, though
with improved approach flow systems and breast box designs such functions
of a screen are probably less important than formerly. Nevertheless, it is
undoubtedly still the case that screening equipment is considered most
valuable when used to give the final removal of impurities of a more
fibrous nature and reduce clumps, strings, and balls of good fibre immedi-
ately before the flow enters the breast box and is deposited on the wire.
Thus, whenever flocculation or aeration are likely to be troublesome,
there is a particularly strong case for screens immediately preceding the
breast box.
However, with the advent of more efficient cleaners of the cyclone type
this sequence of cleaners first, screens second has been questioned. Cyclone
cleaners are prone to plugging of the quite small orifices used in individual
cyclone units, at least when these are of smaller diameter, so to prevent
this occurring it is considered preferable that screens are used first. By
this means those larger particles are removed that can cause plugging
or become trapped and create heavy wear of the cyclone body as they
whirl perpetually round inside.
137
2B.l 3 SCREENS AND CLEANERS
Installing screens before cleaners was a step taken somewhat hesitantly

at first, particularly because it was feared that the stock would be too
flocculated and full of air as it entered the breast box. One or two instal-
lations with both cleaners and screens were erected in such a way that
either could be used first in the flow system. But there is now sufficient
evidence available to show that at least with shorter-fibred stocks there is
no reason why cleaners should not be used immediately ahead of the
breast box. Perhaps the only precaution that may be advisable with
certain arrangements of cyclone cleaners is to make provision for prevent-
ing suction of air into the accept flow.
Two examples of installations where cleaning equipment has been
installed immediately ahead of the breast box are worth citing. Graham
(16) described a system put in on a newsprint machine in which cyclone
cleaners discharge straight to the breast box. These cyclones were equipped
with eductors on the reject nozzles to reduce aeration of the stock entering
the breast box, but otherwise no special precautions were taken. The
formation showed no noticeable change after installation, and there were
no obvious signs of increased flocculation of the stock.
A more recent example has been reported by Downey and Blake (31 ).
In this case two Selectifiers (with reject to a Jonsson flat screen, from
which accept goes to the wire pit and reject to the sewer) are followed by
a battery of eight 12 in. cyclones of the Bauer type (reject under 16-18 in.
Hg vacuum, accept from secondary stage to the wire pit and reject to
sewer) before entry to the breast box. Improvement in cleanliness of the
sheet was observed and no mention is made of any difficulties in operation.
But it should be added that following this installation two other machines
were re-built with the opposite sequence, i.e. screens following cleaners,
though no reasons for the change in order are given (possibly the decision
was governed by the use of a deculator in the system in place of a vacuum
tank on the cyclone rejects, with the consequent need for two stock
pumps instead of one).
2B .1 3 Effect on wet-end stability

In the normal position between the mixing pump and breast box, screens
and cleaners have some influence on operation of the wet-end which is
worth noting because occasionally trouble can occur with the stability
of substance-keeping or dry-line position as a direct result of their malfunc-
tion. This comes about since any cleaning and screening equipment creates
two kinds of changes in the composition and consistency of stock after it
has left the mixing pump. Firstly, a certain amount of water and fibre
leaves the equipment with impurities either as reject to drain or at any
rate out of the immediate wet-end flow system. Secondly, dilution in some
form is nearly always an essential feature of the operation so an addition
of water and fibre, sometimes appreciable in relation to the stock flow,
takes place.
Normally operation of the wet-end flow system is unaffected by this
extraction and addition of water and fibre. This is not just because the
138
SCREENS AND CLEANERS 2B.l 3
flows involved are small, though this is usually the case, but because
whatever difference they make to breast box consistency and even to
drainage conditions (as a result of a change in temperature or fibre charac-
teristics of the stock) is automatically compensated for, once equilibrium
of the backwater circuit has been reached, by appropriate setting at the
mixing pump of the flow of fresh stuff and of the quantity of backwater
used for dilution. To take an example, suppose a large quantity is needed
either of spray water in a screen or dilution water in a cleaner, and this
is at a very low consistency (perhaps extracted from the suction boxes).
Then the consistency of stock leaving the screen or cleaner will be lower
than entering it and this will in turn reduce consistency in the breast box.
But the machineman will set the re-circulation of backwater to give him
an appropriate dry-line position, so in fact he will automatically compen-
sate for the addition of low-consistency water that occurs after the mixing
pump. Less backwater will actually be circulated through the mixing
pump, where consistency will be higher than is required at the breast box,
the remaining dilution taking place in the screen or cleaner. In a similar
way, fibre rejected from the system is in effect compensated for by a slight
addition in fresh stuff.
A further point to note is that the fibre-length distribution at the mixing
pump may also be different from that in the breast box due, for example,
to the tendency for pressure screens selectively to reject longer fibres and
cyclones to reject shorter fibres. But as with consistency the fibre-length
distribution in the breast box will effectively stabilize at the same condition
it would have reached without the screen or cleaner operating.
The presence of a fibre reject and of dilution at a screen or cleaner thus
does not materially affect normal operation with regard to substance
setting and such factors as drainage conditions on the wire. But what
happens if either of these flows fluctuate for some reason? Consider the
flow of rejected fibre from a screen or cleaner. When this leaves the closed
backwater circuit (the flow from the main machine pit through the mixing
pump to the breast box) then fibre contained in it is effectively deducted
from the flow of fresh fibre to the machine system. Hence it has an effect
on the value of paper substance. Should a large increase in this flow occur
due to some disturbance, then the substance will be reduced. For this
reason it is unwise to operate with a large reject flow from screens or
cleaners when this leaves the immediate machine system, i.e. is taken
(perhaps via a secondary screen or cleaner) either to waste or to the stock
preparation or pulping system. There is no reason why a large reject flow
should not be operated from a primary screen or cleaner provided it
eventually returns to the backwater circuit.
As the usual course is to take the accepted flow from a secondary screen
or cleaner straight to the machine pit, leaving only a very small final
reject for disposal out of the system, in practice this should not normally
be a source of any significant substance instability. But large variations in
the primary reject flow, especially when rapid, can still affect the substance
level temporarily in a similar manner to other variations in flow at the
wet-end (as discussed in lA.2). An example of this is seen if the reject
139
2B.2 SCREENS AND CLEANERS
valves of pressure screens are opened too quickly for the purpose of
purging: a marked series of bars can sometimes be seen passing down the
wire and the head at the slice can drop momentarily, a disturbance that
easily causes a break. In practice, apart from this, temporary disturbances
due to variations in reject flow do not appear to affect operation.
Variation in the flow of spray or dilution water to screens and cleaners
can likewise affect substance stability when the source of dilution comes
from outside the immediate machine system. This applies even more to
chemical condition of the stock and especially to pH. Dilution from
sources other than machine backwater is frequently used for sprays, and
also when backwater consistency is considered too high for adequate
dilution, for example in cyclone cleaner fibre-recovery stages. The source
can then be fresh water or, when large flows are involved, whitewater
recovered in a fibre save-all is sometimes used. The effect on substance
stability caused by a sudden change in addition of dilution water from
outside the machine system will obviously not be so great as in the case
of fibre reject discussed above, because its main effect will be on consistency
levels while the effective fibre flows are relatively unaltered. An increase
in the dilution flow would ultimately reduce breast box consistency, leading
the machineman to compensate by reducing the quantity of backwater
used for dilution at the mixing pump. At the new equilibrium the substance
level should be unaltered except insofar as there has been any consequent
change in the consistency of excess backwater leaving the machine system.
On the other hand a fluctuating flow of dilution water would be undesir-
able because it could undoubtedly lead to transient changes in the sub-
stance.
2B. 2 OPEN SCREENING EQUIPMENT

Attention is now turned to discussing the various types of screening and
cleaning equipment used on stock in the paper mill. It is proposed first
to deal specifically with screening and for convenience this has been
somewhat arbitrarily divided into two sections. The present section
covers open screening equipment comprising the diaphragm or fiat screen
and the various types of rotary vibrating screens; the next section will
deal with the more modern enclosed pressure screens.
2B. 2 I Flat screens

The earliest type of screening equipment used in paper mills was probably
introduced primarily to remove the large clumps and strings of fibre that
appear in long-fibred stocks. Flat screens were the first to be installed
and the vibration necessary to prevent the screen plates clogging was
found to provide a useful defiocculating action on fibres. The fiat screen
as a primary screen is now· obsolete, but it is still in general use as a
secondary screen.
Over the years a multitude of different designs of flat screen have
appeared on the market. The most common type has a box at one end
140
SCREENS AND CLEANERS 2B.21
from which the stock flows on to a series of screen plates that may slope
upwards; accepted fibre passes through the plates to a vat underneath
while rejected material discharges over the end of the plates. Sprays and
scrapers are used to aid the process and help break down clumps of fibres
that clog the perforations or become stapled across two adjoining slots.
Flat screens differ mainly in regard to how the pulsation is applied: this
is either to the vat underneath by means of a cam and springs or by a
positive rocker mechanism, or to the plate itself by means of a motor
driving through an eccentric. The first type, which is the older, requires
a dam underneath in the vat to ensure that the accepted stock level is
high enough for the vibration to be transmitted to fibre on top of the
plates; the second type allows a much higher frequency of vibration and
is generally shorter in length. Other models of flat screen have appeared,
some with the flow upward from a vibrating vat through a curved screen
plate which then contained the accepted fibre until it was sucked out by
means of a simple water ejector system. However, none have found such
general acceptance as the main two types described.
Flat screens are not easy to operate or control efficiently. It is usually
possible to change the amplitude or frequency of the vibration, but this
rarely seems to make any obvious difference to performance. Usually
careful positioning of sprays on top is a more satisfactory means of
adjustment. Screen plates of the vat-vibrated type are commonly in tin.
thick bronze, chromium plate or stainless steel, and there are normally
several that can be removed individually. They are usually cut with 3 in.
to 4 in. long slots, which are 6 to 60 thou wide (normally 8-12 thou) and
et
have a downward taper, though relatively large diameter in. to -rlr- in.)
holes are also used. The slots are arranged from 4 to 8 to an inch in a
variety of patterns which are thought by the designers to promote efficiency.
The other type of flat screen usually has a single plate and the life of this
can be quite short due to the high frequency of vibration it is subjected to.
Regular cleaning is essential for all flat screens otherwise it will be found
that the flow of fibre over the end of the plate has increased to uneconomic
proportions; ease of access to the vat for cleaning underneath, and with the
upward-flowing type ease of turning over the plate to allow the impurities
to be removed from the vat and the underside of the plate cleaned, are as
important features as any to look for in flat screens.
Examination of the material passing over the end of a typical flat screen
gives a convincing indication of efficiency. When working properly, there
is always a very high proportion of large impurities embedded in a rela-
tively small amount of fibre. The impurities separated out obviously
depend on the stock, but shives, knots, unbroken pieces of broke and
Cellophane, etc., and tight strings of fibres are commonly seen mixed with
any really large objects (pieces of wood, rubber bands, and so forth)
that have found their way into the system. The only systematic examination
of the performance of a fiat screen is due to Lambert (34) who examined
one of the high-frequency vibration type with t in. holes which received
the reject from a number of primary screens working on newsprint stock.
Samples of the feed, accept and reject flows were taken and measured
141
for consistency, C.S.F., and fibre fractionation. It appeared from this work
that a relatively efficient separation occurred in the sense that the freeness
of the reject was very high compared to the inlet (496 compared to 61)
and there were five times as many fibres retained on a 14-mesh screen.
But when the low reject flow rate was taken into account the true efficiency
of separation (using the greater than 14-mesh fraction as a criterion of
'objectionable material') was in fact so low (0·3 per cent.) as to be negli-
gible. Confirmation of the low efficiency of flat screens can also be found
in a report by Downey and Blake (31) in which they calculated that 90
per cent. of shive passing from the reject of a closed screen was accepted
by a secondary flat screen.
Lambert considers that reducing the hole size would not improve this
situation as it would only produce a greater reject flow and lower capacity,
while not affecting the true efficiency. He concluded from his examination
that, at least for secondary screening of newsprint, it was difficult to
justify using a flat screen. One can question the definition of efficiency
used by Lambert, but it does seem that in practice a certain quantity of
larger shives and other large contraries will be removed, but the majority
will pass straight to the accept flow and thence back into the stock system.
Inevitably such material will eventually find its way into the paper so the
only merit in using a secondary screen seems to be to provide some means
of achieving relatively little fibre loss.
2B. 2 2 Rotary screens
As with flat screens, there have been numerous types of open rotary
screens marketed. The original models were outward flowing, i.e. stock
entered through a hollow journal into the inside of a rotating cylindrical
screen and passed out into the vat. But nowadays these are mostly used
on fine papers with a high content of long fibres and most models in use
are of the inward flowing type. In this case stock enters the vat over
rubber sealing strips or from the bottom, passes inside the cylinder, and
is discharged through one end which is sealed from the bearing. The
preference for the inward-flowing rotary screen is partly because it is
easier to clean and arrange a continuous reject flow, and partly a question
of mechanical design and reduction of maintenance which, with vibration
an essential feature, can become heavy.
Showers are always a vital part of the operation of any rotary screen
and these must be positioned to ensure that fibre is washed away from the
surface on the ingoing side and no part of the plate is missed because
of poor overlapping of individual jets (oscillating showers give the best
protection against this). With the outward-flowing type of screen a direct
spray on the outside of the plate pushes fibre sticking in the slots into
a trough inside the screen; with the inward-flowing model the sprays
are either internal, or glance the outside of the plate at a fine angle on the
upgoing side to avoid forcing fibre through as it is washed back down to the
vat. The cylindrical screen plates are in sections and invariably cut with
slots which can be anything from 8 to 35 or more thou wide (normally
18-20 thou) and which taper out in the direction of flow; they are made
142
of stainless steel, phosphor bronze, or monel, and are often chrome-plated,
or plastic-coated.
Various methods of imposing a vibration on the screen are used. With
the outward-flowing models, in some types the cylinder itself is shaken
from pivoted bearing arms or by means of a rotary-eccentric centre
bearing at one end, and in another type the vat itself is vibrated by being
supported on straps one end of which are fixed and the other attached to
a jogging lever. In yet another arrangement rotation of the cylinder
instead of being smooth is in jerks by a ratchet device. With the inward-
flowing models, either the whole vat is supported on :flexible mounts and
vibrated sideways from an eccentric shaft, or a semi-cylindrical plate or
perforated diaphragm underneath the main rotating cylinder is vibrated
up and down or from side to side.
Rejects from the outward-flowing rotary screens comprise light material
skimmed off the stock surface in the cylinder together with fibre and
impurities washed into the trough by sprays. From inward-flowing screens
either larger impurities settling in the bottom of the vat must be cleaned
out when the screen is stopped or· more usually there is a continuous
bleed-off from the lowest point at the bottom of the vat; lighter material
may be skimmed off the top in an overflow. The reject flow can be altered
within a wide range and is usually governed entirely by the capacity of
secondary screening equipment.
Although open rotary screens are rapidly becoming obsolete there has
been one interesting development of a new screen of this type designed
for a specific purpose. Robinson and Kingsnorth (29) tackled the problem
of removing long flexible hairs (sisal) from a chemical/groundwood stock
by building a rotary screen which is basically of the outward-flowing
type with sprays backwashing the screen plates into a trough inside the
cylinder having a discharge at one end. The main feature of this screen
is the use of a coarse wire mesh (six meshes to the inch) as this was found
by experiment to present the best efficiency: dyed sisal hairs between
about t in. to 1 in. length were rejected at over 90 per cent. efficiency,
though this level of efficiency had to be sacrified during modification of
the design to attain greater throughput with an acceptable fibre loss in
a full-scale version. The authors state that using an open-mesh screen
instead of the usual slotted type of plate for this particular purpose has
the advantage of giving a much greater throughput. The screen acts as a
tangling device for the long hairs while presenting a minimum of restriction
to the flow of fibres. This development represents a good example of the
best way to approach the removal of impurities from stock systems, by
considering the individual sources of impurity that must be taken out and
then deciding the most economical manner of achieving this end. Not
everyone of course will be prepared to go to the expense of developing
a new screen, nor is this likely to be at all necessary, but the point is
worth making that only sufficient screening or cleaning capacity should
be provided to do the job required for the particular grade of paper
concerned, otherwise it will invariably be the case that more power will
be needed, or more fibre lost, than is strictly necessary.
143
Lambert (34), in the work already referred to above, also checked the
performance of a bank of five rotary screens of the inward-flowing type
with a 25 thou slot. The C.S.F. of the reject flow was no different from
that of the inlet, though the consistency was slightly higher. The true
efficiency of separation, measured on the same basis as mentioned above
for the fiat screen, was higher but still only 2·8 per cent. for material
retained on a 14-mesh screen. Lambert comments that this shows the
screens gave virtually no separation.
If this result can be generalized (and there is no reason to expect other
screens to act more efficiently, at least with regard to the basis of measure-
ment used in this work) it would seem that open rotary screens can serve
little screening function except for the removal of really large impurities.
Of course in many cases the defiocculating effect of open screens can be
an invaluable asset, and for some stocks the prevention of large strings
of fibre passing to the breast box is highly important. But when it comes
to removing smaller shive-like material, the efficiency appears so low as
to make the expense totally uneconomical.
2B. 3 CLOSED PRESSURE SCREENS

Since the introduction of the totally-enclosed pressure screen, this type
has come to predominate on modern paper machines and in the vast
majority of cases is installed both on new machines and when re-building
the wet-end of old machines. Certainly for the larger machine making
relatively coarse grades, the pressure screen is far superior in design to
the open rotary screen, for reasons which will emerge in what follows.
Only on slower fine machines, especially with a long-fibred furnish,
may an open screen still often be preferable, mainly because any tendency
for clogging of the screen plates to occur can be immediately seen and
quickly remedied.
2B . 3 1 General details
In the closed pressure screen, stock enters at the top tangential to the wall
of the screen. It then passes downwards either into an annular region
between two cylindrical screen plates or into the centre of a single cylindri-
cal screen plate. Passage outward through the plate, and inwards also
when there are two plates, is assisted by means of impellers rotating on
a central shaft supported from below. There are usually two devices
collecting rejected material; one collects larger objects thrown out in the
initial entry to the screen and is emptied intermittently, the other represents
a bleed-off from the bottom of the screen consisting of fibre and material
that does not pass through the plates (this is almost always a continuous·
flow). Accepted flow leaves centrally from the screen at a point level in
height with the plates. One model is an exception to this and here the
flow passes inward through a single screen plate and then leaves tan-
gentially. Capacity of a single unit is from around 1,000 gallons per minute
upwards.
144
SCREENS AND CLEANERS 2B.3 I
One of the most critical features in the operation of enclosed screens
is the impeller arrangement. The head is of an aerofoil design and is set
so that the leading edge skims close to the screen plate with the gap
between the impeller head and the plate widening gradually towards the
trailing edge. Setting the gap between the impeller head and screen plate
(normally between ~- in. and :!- in.) and the angle formed between the head
and plate on the trailing side are both quite critical. The shape of the
impeller head is designed so that stock trying to pass through the plate
receives an initial shock assisting it in that direction, followed by a sub-
stantial suction (akin to that formed by a foil under a Fourdrinier wire)
that clears any fibre or dirt left lodging in the holes. This effectively reduces
any tendency for the screen to make up, though especially with longer-
fibred stock clogging still represents a hazard on many installations.
When clogging begins to take place, power required to drive the impeller
increases and above all the pressure drop across the screen rises appreci-
ably; eventually the screen seizes up altogether and it can then be a lengthy
operation to clean it, necessitating removal of the top cover (this is easiest
when hinged on a davit arm) and also the screen plates as well.
Adjustment of the clearances between impeller heads and plates can
sometimes reduce any tendency to clog, but on the other hand it can also
affect power consumption and in adverse circumstances promote hydraulic
pulsations (due mainly to flexing of the screen plate) which can show up
as a short cycle machine-direction substance variation down the wire.
Altering speed of rotation of the impellers also affects the running condition
of a screen, lower speed tending to reduce power demand and any hydraulic
pulsations but (due to poorer clearing of the plates) at the expense of a
higher pressure loss across the screen. Occasionally four-bladed impellers
are used instead of two-bladed and this too appears to reduce hydraulic
pulsations and is claimed to give a stronger screening action, however
power consumption rises as does the fibre discharged in the reject for a
given valve setting. When several enclosed screens are in parallel, it is
occasionally necessary to arrange the motors to be deliberately out of
phase otherwise there are periods lasting a few seconds when an intense
mechanical vibration is set up which has the same effect on the substance
as hydraulic pulsations from the impellers but is usually far more violent.
The large material thrown out on entering the screen accumulates either
in a collecting chamber in the form of a box or a straightforward section
of pipe. In both cases there are usually two valves, an upper and a lower,
the former of which is left open to allow the debris to settle under gravity
into the collecting chamber. Periodically, depending on how rapidly the
collecting chamber becomes filled, the upper valve is closed off and the
box emptied by opening the lower one. Before the upper valve is again
opened it is useful to be able to flush out and fill the chamber with fresh
water to prevent a sudden suction of air into the stock system. The
whole business of periodic cleaning out of the collection chamber can
be made automatic if the expense seems justified; in one arrangement, at
a pre-set frequency a special rotary three-way valve connected to a timing
device changes for a short time from the normal position (connecting the
145
2B.3 2 SCREENS AND CLEANERS
screen to the collection chamber) to a second position (connecting the

chamber to drain). Examination of the debris found when this part of
the screen is emptied often shows surprising objects.
The continuous reject is always at a much higher consistency than the
inlet flow. It is led to a secondary screen, usually of the flat type as when
using open rotary screens, and the flow is generally controlled to suit the
capacity of the secondary screen, being in the region of 3 per cent. to
8 per cent. or so of the inlet flow. Especially when the flow has to be kept
fairly low due perhaps to lack of capacity in the secondary screen, there is
a tendency for a blockage to occur in the valve controlling the flow. This
is not always easy to detect, particularly when there are a number of
screens in parallel and their reject lines join into a common manifold.
To avoid the possibility of this going undetected, a separate and visible
discharge is preferable and a purge line into the upstream side of the valve
can be useful. Alternatively one of a number of different designs of special
purging valve incorporating an automatic timing device can be used.
These valves are arranged to open a specified amount and close back to
their original position at regular intervals, and this effectively clears any
fibre built up in the throat of the valve. The only precaution needed in
their use is to avoid too large and rapid opening of the reject valve as
this can temporarily starve a screen to an extent that affects the substance
of the sheet and can cause a break (see the discussion in 2B .1 3.)
There are also facilities for removing a small bleed-off from the top
cover of a pressure screen. This is generally necessary to prevent air
building up under the cover and eventually gulping away into the accept
flow; also, if much air is allowed to collect the screen impellers begin to
create excessive turbulence and this magnifies hydraulic pulsations and
creates a rise in the power consumed. Although the air vent is sometimes
opened only at start-up and then closed, it is generally preferable for a
small flow to be removed continuously and taken to the secondary screen.
2B . 3 2 Comparison with open rotary screens
For practically all screening operations the enclosed pressure screen has
now replaced the open rotary screen. There are many reasons for this.
The initial capital cost and the power consumption when running are both
lower relative to screening capacity. The enclosed screen requires less
space and is better adapted to the higher stock pressures involved on fast
machines where a considerable head is needed at the slice and the breast
box is pressurized. Also, being enclosed and operated with an air bleed-off
reduces the possibility of trouble from foam and slime, and continues the
general trend from open surfaces (chutes, mixing box, etc.) to closed
systems (pipes, mixing pump, etc.) which has characterized development
in recent years. Enclosed screens also generally require lower maintenance
than the open rotary type because their impeller rotation is smooth and
it is not necessary to induce a vibration into the equipment; also there are
no showers which can often be a problem to keep working satisfactorily.
Finally, the reject flow from an enclosed screen is much more readily
controlled.
146
SCREENS AND CLEANERS 2B.3 3
This is a formidable list of advantages, so it is hardly surprising that
there has been a complete swing over to using enclosed screens. Practically
the only disadvantage of the enclosed screen is the relatively high con-
sistency of the reject (which as mentioned above can be troublesome to
ensure a steady flow) and the time needed to clean down if clogging of
the screen plate occurs.
So much for a comparison of operating features, but what about relative
performance? The consensus of opinion here is that enclosed screens
a,
,...:
'41
...J
.!:
0
t-
4
t-
\.J
..,
u.t
w
°' z
0
~
Cl
0
0 4
Fig. 2. 11. Effect of altering the reject flow from an enclosed screen on the reject/
inlet ratio for consistency and for the fraction retained on a 14-mesh screen (after
Salomon and Lukianovitch)
prove more efficient than open rotary screens in some respects but not
in others. Before discussing this in detail it will be as well to set down
what data is available on the performance of enclosed screens.
2B. 3 3 Performance of enclosed screens

Thickening of the reject flow in comparison with the inlet can be appreci-
able. Salomon and Lukianovitch (23) reported that the reject consistency
of an enclosed screen was nearly double that of the inlet with only a slight
drop in the ratio as the reject (or 'tailings') flow was increased, see Fig. 2 .11
(lower curve). They also found that there was a larger proportion of
147
longer fibres in the reject flow with a consequent increase in the percentage
of fines in the accept flow. Other reports (for example, reference 17) have
confirmed these observations and also indicate that the thickening ratio
increases at lower inlet consistencies. On a newsprint grade the proportion
of sulphite fibres in the inlet stock can be doubled in the reject, though the
proportion of loading particles remains unaltered.
Lambert (34), in the investigation already referred to in 2B. 2 1 and
2B. 2 2 with reference to open screening equipment, also checked the per-
formance of an enclosed screen on newsprint stock. He found that effi-
ciency of rejection for material kept back on a 14-mesh screen was 40 per
cent. This may be compared with 0·3 per cent. for a typical flat screen and
2·8 per cent. for a rotary screen. With a greater reject flow the percentage
of larger material rejected was even higher, though this could of course lead
to greater fibre loss from the system. Both the main types of pressure screen,
the Centriscreen and Selectifier, gave a similar performance in this respect.
A similar result was found by Salomon and Lukianovitch, and is also
shown in Fig. 2.11 (upper curve). They reported that the concentration
of material greater in size than a 14-mesh screen was six times as high
in the reject as the inlet with a 1 per cent. reject flow, dropping to a little
over four times for reject flow 5 per cent. of the inlet, the usual running
condition. The first of these results represents, in the terms used by
Lambert, an efficiency of about 12 per cent., but the second is over 40
per cent. which confirms his own figure. Macmillan (35) reported a reduc-
tion in shive content from 2·75 per cent. to 2 per cent. on a newsprint stock,
representing a rather lower efficiency of 27 per cent.
The implications of this data are that as a means of removing shive-like
material the enclosed screen is relatively efficient. Indeed inspection of
handsheets made from the reject flow from a screen operating on stock con-
taining groundwood furnish gives a ready indication of this. Operating
experience when enclosed screens have been substituted for old rotary
screens also amply confirms this. But there have been many reports that
such a change in equipment is also accompanied by an increase in the
appearance of relatively large particles of a more spherical shape.
Thus, Salomon and Lukianovitch recorded that when a pressure screen
was first used there were many holes in the sheet, especially after each
start up, due to particles of sand and pipe scale. Sloping the impeller
hydrofoils to give a downward motion and running with high reject
flow assisted in overcoming this difficulty, but many other mills have
met the same problem and not been so fortunate in finding a permanent
solution. Another report, by Hopkins et al. (32), compared the performance
of a number of machines using essentially the same furnish. Those with
pressure screens were excellent for removing long and thin slivers and
unbroken pieces of broke, but overall cleanliness of the sheet was lower.
2B . 3 4 Slots versus holes

The reason for poorer rejection in enclosed screens of heavier and more
spherical-shaped particles is thought to be due principally to the use of
148
circular holes in pressure screens as opposed to slots in rotary screens.
Although slotted screens are available for at least one of the enclosed
models, not a great deal is known about their performance. Theoretical
considerations already outlined in 2A. 1 3 show that the use of holes in a
screen plate instead of slots of a similar area would favour acceptance of
solid spherical particles and rejection of shive-like material. This would
seem to be precisely what happens when a pressure screen with holes
replaces a rotary one with slots. The fact that the open area corresponds,
a 62 thou hole to a 16 thou slot, a 79 thou hole to a 20 thou slot and so
on, does not affect the situation.
It is sometimes argued that although an enclosed screen has holes,
because of the fast rotation the view presented to the stock is of an elong-
ated hole or slot. In other words it is in effect no different to an open rotary
screen. This is patently fallacious because at the point which stock passes
through holes in the plate its velocity must be substantially parallel with
the direc.tion of the hole. In the boundary layer from where stock passing
through the holes is drawn, the chances of acceptance or rejection will be
governed primarily by the interaction of the various factors, including
particle shape and hole shape, which were discussed in 2A. 1.
Efficiency of rejection of smaller solid particles can be improved by
reducing the size of the holes. Normally holes are from 60 to over 100
thou diameter for coarser screening, but in some cases a diameter of
45 thou can be used. Unfortunately the smaller the hole size, the higher
is the power demand and pressure drop across the screen and the greater
is the tendency for clogging to occur. In practice the hole size is chosen
at a level governed mainly by the necessity of avoiding clogging.
It is clear that particles up to 60 thou in diameter can pass through
the typical pressure screen, whereas with an old open rotary screen having
20 thou slots, no solid round particles of a diameter less than that can
pass. If the greater size of particles accepted by the enclosed screen is
unsatisfactory, there is no alternative but to use the rotary screen and with
it the very low efficiency of shive removal.
With the growing demand for improved cleanliness of paper, and
particularly since work has been reported showing beyond doubt that
shives are a major source of breaks in the high-speed printing of news-
print, it has become more essential to ensure a high efficiency of shive
removal. This means using pressure screens, and if this gives an unaccept-
able level of particle impurity in the sheet (or at worst leads to trouble
with holes in the web) then cleaners become an essential adjunct. Fortu-
nately centrifugal cleaning equipment, especially of the cyclone type,
though often very inefficient at rejecting shives particularly the larger
ones that cause the trouble, is highly efficient with solid spherical particles.
Thus the two pieces of equipment complement one another especially
on groundwood furnish: cleaners to remove solid particles, grit, stubby
debris, etc., which spoil the sheet appearance and wear wires, suction
box surfaces, and calender rolls; screens to remove long, thin shives and
unbleached lumps which, though not so numerous relatively, probably
cause breaks more often.
6 149
2B .4 EARLY CLEANING EQUIPMENT

Early forms of cleaning equipment were of two types: the riffier or sand
table which relies entirely on natural settlement to separate out heavier
particles; and the enclosed rotating basket type of cleaner which represents
the first attempt to utilize the powerful separating potential of centrifugal
force. Each of these cleaners is now briefly dealt with.
2B. 4 1 Sand tables

Riffi.ers or sand tables are a straightforward application of an age-old
method of purification by settling. Stock is allowed to flow gently down
a slight slope which is usually interposed at intervals by baffies meant to
catch the heavier particles settling out. The whole flow has to be very
shallow and slow to permit a reasonable settling time, though not so
slow as to allow fibre and loading to settle also. This means having a very
large area available and is one of the main disadvantages of the sand
table, apart from exposing a large area of stock to further contamination
from air impurities. Baffies are 4 in. to 8 in. high and the whole floor is
sometimes covered in long-nap felt intended to help catch and hold the
dirt. Sometimes an electromagnet is placed across the trap to attract
metallic particles. Also boards are hung with their edges just below the
top surface to skim off floating scum, rubber, and so on. Regular cleaning
out is absolutely essential as once a fair layer of dirt is built up behind a
baffie it is very easily disturbed. Usually the table is arranged in parallel
banks each of which can be dammed off in turn for cleaning out.
The sand table is reasonably efficient at removing the larger and more
obvious contaminants that appear in paper stock from time to time, but
it is often difficult to get the flow right for removal of grit and sand. There
has always been some difference of opinion about just how deep the
flow should be and how the baffles should be set to produce a mild amount
of turbulence, sufficient to keep the fibres in suspension but not disturb
dirt collecting against the baffles. With wet-beaten stock at a fairly high
consistency the flow readily becomes sluggish and begins to divide into
channels of relatively fast-moving low-consistency stock separated by
dead areas from which fibre settles out and which become thicker as time
passes.
The sand table has been obsolete since pre-war days, yet curiously
enough the first real study of its performance did not appear until the
classic report of Chester (1) in 1950. In addition to a theoretical considera-
tion of the flow in a sand table, Chester carried out comparative experi-
ments on a laboratory model using ground coal particles in an unbeaten
bleached sulphite stock, assessing cleanliness by counting on standard
handsheets. The effect of variations in depth of the baffles, flow rate,
consistency, different shapes of baffle, temperature, beating, particle size,
and several other variables were all investigated. The best results appeared
with a table lined with old machine wet felt, and stock flowing at about
6·5 inches per second to a depth of one inch. This gave a mild degree of
150
turbulence just sufficient to prevent settling of fibres. Higher temperature
increased efficiency but beating caused a decrease.
The coal particles were observed to settle out in a fibre-free layer ad-
jacent to the bottom of the table. If flow was so slow as to be streamline
conditions for settling were inhibited by the fibre because this tended to
settle at the same time. Thus, some degree of turbulence must be induced
to promote settling and Chester examined the value of various methods
in common use. Baffles dipping into the surface caused a slight improve-
ment, as did vertical baffles under the surface. The traditional undulating
riffle was better than simple baffles, but ordinary felt appeared to be the
most efficient. Even with this there was hardly any reduction in the number
of small particles and overall efficiency was extremely low.
2B. 4 2 Rotating basket cleaners

The principle of the rotating basket cleaners, the first type to utilize
centrifugal force, depends on the rotation of an elaborate arrangement
of cylindrical containers and baffles. Stock introduced down a funnel at
the centre of the device accelerates in a tangential direction under the
influence of the revolving compartments and then has to pass under and
over a series of strategically-placed baffles designed to promote retention
of heavier particles flung against the walls of the compartments. In some
models light impurities are retained on specially-placed rings and in others
it is customary to have one large central distributor surrounded radially
by a number of individual cleaning units.
The whole operation of this type of cleaner depends on the building
up of thick layers of pulp in the various positions where impurities collect.
This is essential to ensure retention of impurities against the wall of the
compartment and reduce turbulence in that region, so aiding the capture
offurther impurities. But inevitably there is a tendency for more and more
pulp to build up into a solid lump until a large part of the device is rotating
as a relatively stable mass and further stock introduced tends to pass
straight through. It is then necessary to clean the whole contraption,
at which point the elaborate design of compartments and baffles becomes
a positive nuisance. Hosing out is often not practicable except for finishing
off because the pulp is so thickly embedded against the walls of the com-
partments that it all has to be dug out by hand. When starting up, con-
sistency of the accepted stock is at first low until a substantial layer of
fresh pulp has been built up, so when a change of unit is made while the
machine is running the substance can be badly affected.
These devices were probably about the most clumsy, useless, and waste-
ful ever to find their way into paper mills. Despite their high cost, capacity
is very low and there always has to be one spare because it is not possible
to run for long before the tedious business of cleaning out becomes
necessary. The cumbersome arrangement of compartments and the whole
size of the contraption limits the rotational speed that can be built up,
and so minimizes the benefit of centrifugal force. Maintenance is usually
heavy, fibre loss considerable, use of floor space excessive, and running
151
costs high. On a machine with frequent colour changes revolving basket

cleaners can account for a really excessive fibre loss and much wasted time
in cleaning. Perhaps the most charitable statement that can be made is
that they were probably more efficient than sand tables, but it is hardly
surprising that as a cleaning device their appearance in paper mills was
rapidly superseded by cleaners that utilized centrifugal force in a much
more simple and efficient fashion.
2B.5 CYLINDRICAL CLEANERS

There is no difference in principle between the operation of cylindrical
and cyclone cleaners, both rely entirely on the separating effect created
by high centrifugal force. In practice the main difference between the two
types is that the cyclone cleaner, by virtue of its shape, can discharge the
reject flow into free air (though as will be seen there is a growing tendency
for this facility not to be used) while the cylindrical cleaner has a special
device at the bottom of the cylinder designed either to channel the reject
off to a secondary cleaner or to effect a separation of rejected material
into a collecting bottle. The cylindrical cleaners are generally larger than
cyclones, so fewer units are needed to cope with a particular stock flow;
however, there are now both large-diameter cyclone cleaners and relatively
small-diameter cylindrical cleaners available, so this distinction too is
blurred. Information given below is drawn from a number of references,
in particular numbers 2, 11, 12 and 14.
2B . 5 1 General details
The cylindrical cleaner has a tangential inlet at the top of the cylinder
(this can be specially scrolled to assist the flow) into which stock is pumped
at a pressure of between 20 and 40 p.s.i. The tangential motion continues
down the cylinder until one or more baffles are met which turn the main
body of the flow inwards and upwards. This inner column emerges as
accept flow centrally from the top of the cylinder into a pipe, or in some
models it is arranged to emerge tangentially thus continuing the generally
circular motion imparted at the inlet. The baffles in the lower part of the
cylinder are of various arrangements designed to encourage the flow
from the outer wall of the cylinder to continue downwards, taking with
it such impurities as have been thrown outwards. Beneath the baffles
the cylinder tapers to a pipe section with a valve below which is a chamber
for collecting rejected material; additionally there can be a continuous
reject taken out centrally or tangentially from the side of the cylinder
and in some models elutriation water can be injected into the cylinder
for the purpose of reducing the consistency of this reject.
Cylindrical cleaners vary from as low as It in. diameter (capacity
around 20 gallons per minute) to 10 in. or over with a capacity of 1,000
gallons per minute and higher. The capacity increases with higher pressure
drop across the cleaner, either when inlet pressure is raised or accept
152
pressure reduced to a minimum of 1 to 2 p.s.i.; the pressure drop is norm-
ally 20 (in larger units) up to 40 p.s.i., above which the increase in separat-
ing efficiency is not considered worth the added pumping costs (for a
fuller discussion of this point see section 2B. 6 5 on cyclones). Material of
construction is stainless steel, bronze or cast iron with wear-resistant
lining; parts exposed to wear are often chrome-plated or in ceramic. In
some models designed to de-aerate stock a low inlet pressure of 10 p s.i.
or so may be used, vacuum is applied to the reject (which discharges con-
tinuously to a secondary cleaner), and the accept flow is usually taken direct
to the suction of a pump feeding the breast box.
When there is a continuous reject flow it is usually diluted with back-
water to a lower consistency before being pumped into a secondary cleaner.
This is generally similar in design to the primary unit though often smaller
in diameter and with a consequent lower capacity. Occasionally a tertiary
stage can be used, though with cylindrical cleaners this is rare. Often,
however, there is no continuous reject and material settling at the bottom
of the cleaner is simply allowed to accumulate in a collecting chamber
which is equipped with a window for easy observation. Even with a
continuous reject this same arrangement is also used, as it is on the
secondary cleaners. The collecting chamber slowly fills with impurities
and fibre, and must be occasionally emptied. The procedure for this is
to close the upper of a pair of isolating valves and dump the contents of
the chamber by opening a valve or plate at the bottom; rather than start
with air in the chamber, it is then preferable to fill up with water using
a fresh water line and air vent. This whole procedure can be made automatic
if this is thought necessary. Fibre loss is kept small by this means, generally
from 0·02 per cent. to 0·2 per cent. depending on the use or otherwise of
secondary cleaners and whether or not loading is present in the stock.
In one unique design rejected material settles into a chamber below which
is a screw arrangement which compacts and pushes out against a counter-
weight like a giant mincer. In another an automatic reject valve gives a
constant volume discharge from a vane rotor. There are also several other
special adaptations designed to allow some measure of control and provide
a continuous reject flow.
Because of the high velocities generated, cylindrical cleaners are vulner-
able to damage if large hard objects are allowed to enter the body. For
this reason some form of pre-tre3:tment is advisable to remove really
large impurities, nuts, beater tiles, bale wire, etc. Alternatively it is possible
to use the cleaners after screening equipment and immediately prior to
entering the breast box. Points regarding this have already been discussed
in 2B. l 2.
Some manufacturers advocate the use of cylindrical cleaners to handle
the rejects of a primary screen either of the open rotary or enclosed type.
This practice has nothing to recommend it because it has been seen that
there is ample evidence to indicate that impurities removed by cleaners are
of a completely different type to those removed by a screen. Large shives
efficiently separated out in the screen will simply pass straight through a
cleaner treating the rejects; the cleaner can in fact be expected to remove
153
only the heavy large dirt that is screened out by virtue of its size, together
with such small, heavy dirt as happens to pass out with it.
28 . 5 2 Running conditions
Despite the popularity of the cylindrical cleaner there have been extremely
few details published of experimental work on its operation. This is in
sharp contrast to the cyclone cleaner for which there is a wealth of data
available. Many operational reports have been made but hardly any can
be considered of useful general validity.
It is generally stated that efficiency of separation of heavy material is
improved by increasing the pressure drop, i.e. difference between inlet and
accept pressure, and by reducing consistency. So far as they go these are
reasonable points to make in the light of the mode of working of the
cylindrical cleaner, but reference to the effect of pressure and consistency
in cyclone cleaners which are likely to be basically similar in performance
will show that other factors need to be considered. In particular some
economic qualification is necessary to take regard of the increased pumping
cost needed when pressure is increased and consistency reduced, and also
the greater proportional fibre loss likely to occur at very low consistency.
One report on a cylindrical cleaner (15) commented on the effect of
furnish and consistency. With a long-fibred stock, efficiency of rejection·
of sand dropped from 90 per cent. at I per cent. consistency to 74 per cent.
at 0·5 per cent. consistency. On a newsprint stock, however, reducing
consistency appeared to make no alteration in the efficiency of sand re-
moval which remained at over 90 per cent. This was for unchanged inlet
pressure. Another report (8) gave some comparisons of the efficiency of
various models produced by the same manufacturer but nothing of general
validity can be deduced from this.
Unfortunately there appears to be no test results reported in which
cylindrical and cyclone cleaners can be compared on the same basis, an
unfortunate hiatus in available knowledge since this would provide most
interesting data. There have been reports of improvement brought about
when cyclones have replaced old cylindrical cleaners, but as the latter
would most likely have been of larger diameter than the cyclones replacing
them, and probably would have been of the early type with no secondary
cleaner or only a simple collecting chamber, it is hardly surprising that
the performance of cyclone cleaners appeared better. Whether a com-
parison of similar diameter cyclone and cylindrical cleaners, both having
a continuous reject system, would yield similar results for both is com-
pletely unknown.
28.6 CYCLONE CLEANERS

Since its introduction in the early 1950s, the cyclone cleaner (otherwise
called 'Centricleaner', 'Hydraulic cyclone', or 'Hydrocyclone') has grown
rapidly in popularity until today it is undoubtedly the most common
stock cleaner in use. This is despite the disadvantage, inherent in earlier
154
SCREENS AND CLEANERS 2B.6 l
installations, of excessive maintenance and a certain clumsiness of arrange-
ment. Since its inception a considerable amount of experimental work
has been reported and a great deal is now known of the behaviour and
peculiarities of cyclone cleaners when used on paper stock. This section
is therefore devoted to a fairly thorough examination of the data available,
and is drawn mainly from references 3, 6, 7, 8, 9, 10, 13, 16, 19, 22, 25,
26, 27, 28, 30 and 33.
2B . 6 l General details
The basic design of the cyclone cleaner and some general details with
regard to feed arrangements and positioning in the stock line have already
been described. Cyclones on the market, of whatever diameter, all follow
this basic design with minor differences in the relative diameter of inlet
and accept orifices, angle of cone, shape of the inlet (whether directly
tangential or spiral), depth of the vortex finder, and so forth. Little evidence
is available to show the superiority of one particular geometrical arrange-
ment over another though it is claimed that under comparable conditions
a higher cone angle gives a lower removal of fibre though capacity and
efficiency of removal of impurities is also lower. Apart from this, about
the only important distinctions which arise in practice relate to devices
attached to the reject orifice. These are dealt with later.
Cyclones are built from or lined inside with a variety of materials
including rubber, stainless steel, porcelain or ceramic, nylon, and other
vinyl plastics~ with particular attention being taken to make the bottom
of the cone as wear-resistant as possible because it is in this region that
rapid abrasion occurs, manifesting itself in particular as an uneven widen-
ing of the reject orifice diameter. Normal input to a cyclone battery is
from a pipe manifold (preferably tapering towards the end) with a pressure
in the region of 30 to 50 p.s.i.; the cyclones are usually suspended in a
double row on either side of the inlet manifold and tilted slightly from
vertical so that the reject tips are closer together. Pressure in the accept
is kept as low as possible, just sufficient to allow unimpeded exit from the
individual cyclones into a common manifold and thence either to a chest
or pump, or direct into the machine breast box. Isolating valves or cocks
on the inlet and accept pipes of each cyclone are advantageous to permit
easy maintenance. An alternative arrangement designed to remove air from
the stock as well as clean it involves the accept flow discharging direct
on to plates in an evacuated chamber and the rejects joined into a common
pipe which is also under vacuum; fo this case inlet pressure need not be
so high since it is the overall pressure drop from inlet to accept which
governs the general efficiency of working. Rejects from open cyclones
discharge as a spray of stock into a trough which is level-controlled by
addition of backwater and serves as feed for the second stage. Similar
arrangements serve succeeding stages. The common sizes of cyclone are
3 in., 6 in., and 12 in. though other diameters are available. Throughput
varies from 10 to 20 gallons per minute for the smaller units up to as high
as 1,000 gallons per minute and even higher for cyclones used in coarse
cleaning.
155
A relative newcomer to the category of cyclone cleaners is the Radiclone

which departs in several respects from the conventional design. In this
device 25 2-in. diameter cyclones are arranged radially in a horizontal
circle with apex pointing inwards, and a number of such circles are stacked
one on top of the other and enclosed in a large cylindrical container. The
whole container is divided up into annular compartments so that a single
pipe at the bottom serves as a common feed to all the cyclone units,
while there are similar compartments for the accept and reject flows and
(when used) the elutriation water. To allow the cyclone units to be examined
the cylindrical container is raised by a hydraulic jack and the units can
then be individually extracted with a special tool. Sealing of the whole
cylinder is by inflated rubber hoses; fins on the individual cyclones ensure
sealing between compartments.
Apart from the improvement in appearance over the conventional
cyclone lay-out, the Radiclone would appear to be advantageous in re-
quiring less room for a similar capacity, being totally enclosed, and
operating at a lower input pressure than usual (under 30 p.s.i.) which
means a saving in power costs. But the most important aspect of this
device, which is of especial value where the prime requirement is to attain
maximum cleanliness, is that being of such small diameter the effective
efficiency of removal of impurities is extremely high. Normally 2 in.
diameter cyclones could not be used on paper stock because the reject
orifices clog too easily. In the Radiclone this is neatly avoided by having
two inlet holes into each cyclone each of which is smaller in diameter
than the reject orifice; thus large impurities which would block off the
reject cannot enter the cyclone in the first place and as long as they remain
in the inlet compartment it is unlikely that any restriction to flow will
occur. An air bleed-off from the top of the reject compartment is also
provided. The usual arrangement of reject stages for successive fibre
recovery is adopted. From its design it is possible to imagine there being
trouble with cleaning if pitch and slime or large impurities become lodged
in the inner compartments, but this should in practice present little more
difficulty than keeping pipework clean and the usual chemical methods
would have to be relied on. Unfortunately, to date there have been no
operational reports dealing with the application of this device to paper
stock.
The principal advantage of cyclone cleaners lies in their simplicity of
construction and high efficiency, though they can have several disad-
vantages. Larger models have a small space requirement for a given
capacity, but for the smaller more efficient models a greater number of
individual units is required and in conventional installations this increases
the space needed. Capital cost may be slightly lower than for a com-
parable installation of cylindrical cleaners. Properly designed, a battery
of cyclones should present little trouble in running provided attention is
given to avoiding the possibility of plugging of reject orifices; if this is
likely to be troublesome with the diameter of cyclone required (i.e. 6 in.
or under with normal stocks) then it will be advisable to remove large
impurities beforehand either in a suitable piece of screening equipment
156
or in a special large diameter cyclone with a reject chamber closed to
reduce fibre loss to a minimum. Occasionally aeration and flocculation
of stock passing through cyclones is noticeable. Power demand is large
due to the high input pressure and each stage needs a pump, also rejection
of fibre and particularly of loading can be a nuisance on some grades;
these points are enlarged on later.
Despite these disadvantages, the superiority of cyclone cleaners in
regard to efficiency is unquestioned. Apart from the wealth of experi-
mental data that has been accumulated, numerous mill applications testi-
fying to their high efficiency can be quoted. For example, Downey and
Blake (31) assessed the dirt in paper made on three machines equipped
with large-diameter cyclones and compared it with other machines making
similar grades. The total number of specks counted using the T APPI
dirt count system was a third less than on the machines with no cyclones
and using the PAPRIC dirt counter the reduction in large specks was
between 30 per cent. and 75 per cent., and in small specks between 25
per cent. and 60 per cent. From another aspect, Graham (16) reported
that replacing old screens of the inward-flowing rotary type with cyclones
on a newsprint machine not only produced an obviously cleaner sheet
relatively free from specks and with a higher brightness, but removal of
grit particles permitted the frequency of wire change to be reduced from
12 to 18 days and the frequency of grinding calender rolls from four
weeks to eight months. Even with a higher machine speed there were less
breaks due to dirt and slime in the furnish. But as a matter of interest it
is also worth noting that in comparison with the old screens power con-
sumption was doubled, fibre loss increased, and capital cost 20 per cent.
higher than if replacement screens had been purchased.
2B . 6 2 General characteristics of cyclone separation

The consensus of opinion from numerous reports is that cyclones are
excellent for removal of small heavy impurities such as grit, sand and
scale, but relatively inefficient for removal of lighter, shive-like particles.
It is for this reason of course that it is becoming increasingly common
for screens to be used together with cyclone cleaners, since the former
are regarded as necessary for removal of the shive-like impurities. The
deficiency of a cyclone in this respect is due to the influence of particle
shape, a topic already touched upon in the theoretical section. It is worth
examining this feature a little more closely from the practical viewpoint.
A typical report on the behaviour of shives in cyclones comes from
Gavelin (25) who observed a definite tendency for thick, woody shives to
appear in the reject of cyclones treating groundwood while long, slender
shives were predominant in the accepted stock. Unfortunately, from the
point of view of breaks long slender shives especially of the larger variety
are more likely to be troublesome because when deposited in the sheet
in the cross direction tension on the web exerts its influence on the weak-
ened bonding over a greater width and is more likely to cause a tear.
Nevertheless some 50 per cent. of groundwood shives (separated according
157
to the standard method using a Somerville screen) were on average

removed by the cyclone.
This selectivity of rejection was seen to be manifested in other ways.
Tests on handsheets made from pulp that had passed through the cyclone
showed that they were harder, smoother, stronger (especially in respect
of tear and fold) and less bulky than handsheets made from the untreated
stock. This is also shown in the tendency for short fine fibres and fibre
debris to pass to the reject in preference to long, thin fibres, an advantage
over screening equipment where it is the longer fibres that tend to separate
to the reject. Normally a third stage reject will contain predominantly
short fibres of low brightness together with pieces of bark and shives.
Confirmation of Gavelin's results are found in several other reports.
For example, Andersson (6) noted that a 3 in. cyclone cleaner removed a
variety of relatively homogeneous dirt very well but not shive. Fibres
retained on a 20-mesh screen reduced from 60 per cent. in the feed to
only 8 per cent. in the final reject from a three-stage battery. Robinson
and Kingsnorth (29) found that cyclones rejected grinder grit, which is
usually less than 100 micron in size, with very high efficiency but shives
(assessed by the British Standard method) were removed with very low
efficiency and showed only a nil to 27 per cent. improvement.
These characteristics can cause some difference in performance on
different types of stock. Broadly speaking, it can be expected that the
greater the quantity of short, fine fibres in the stock the higher will be the
percentage of fibre rejected for a given reject orifice diameter. This in-
creased fibre reject is shown up not by a greater flow but by an increased
thickening factor, i.e. the consistency of the reject is higher. Differences
have also been observed between the behaviour of stocks beaten to dif-
ferent degrees.
2B . 6 3 Special reject devices
Even with a three-stage cyclone battery the fibre lost at the final stage
can be high and when loading is used, especially in a low-retention system
giving a high concentration in the breast box feed stock, the loss of this
can become economically crippling (rejection in a third stage can be
50 per cent. giving an overall loss which has been quoted at 10 per cent.
or more). It was realized quite early on that some means of controlling
the fibre lost in the reject flow (particularly from the final stage) would
be advantageous and several devices designed to do just this have appeared
on the market. Other devices attached to the reject are intended primarily
to prevent suction of air into the reject orifice and to remove air already
in the stock. In some cases the function of reducing fibre loss and preventing
air being sucked in are combined. These various devices will now be
discussed.
The simplest and earliest modification to the straightforward open
reject was to submerge the tip below the level of stock in the collecting
trough. This prevents suction of air through the reject into the accepted
stock and usually makes a small reduction in the reject of fibre (though in
some circumstances fibre reject can actually increase). But the efficiency of
158
impurity removal is also adversely affected and. with a submerged tip it is
difficult to see if there is any partial or complete plugging of the orifice
(though this latter disadvantage can be avoided by using a transparent
reject orifice.) One manufacturer claims to overcome the drop in efficiency
with a submerged orifice by using a special design of vortex finder which
opens out in the shape of a bell. However, Downey and Blake (31) com-
pared two cyclone installations on machines using the same furnish, one
having normal open rejects in the third stage and the other with the bell-
shaped vortex finder. Reject from the open orifice installation contained a
greater proportion of fine fibres and this is possibly an indication of lower
rather than higher efficiency when using the bell-shaped vortex finder.
A development of the submerged nozzle is to have the cyclone reject
tips enclosed and connected together to a manifold. Pressure in this can
then be regulated so that the flow is controllable and there can be no
suction of air into the accept flow. The disadvantage of reduced efficiency
still applies and the greater the pressure placed on the reject line the worse
this becomes. Sometimes the reject is closed off altogether and a collecting
chamber added at the bottom giving a similar arrangement to that fre-
quently used with cylindrical cleaners. This certainly cuts fibre loss to a
minimum but efficiency drops to only a small fraction of what it is with a
normal open orifice. Such a device is used only to give a very coarse pre-
liminary removal of large impurities.
In an endeavour to improve the performance of a cyclone Gavelin and
Sikstrom (25, 28) developed a special device for attaching to the reject.
This consists of a small chamber with one or more openings into which
is injected elutriation water (fresh water or whitewater of low consistency
can be used). The purpose of this is to dilute the reject flow, thereby
reducing fibre content, while keeping in the impurities. The arrangement
also prevents suction of air into the reject. The pressure of the elutriation
water needs careful regulation: if too high the fibre loss is increased, and if
too low the nozzle can plug. Optimum conditions appear to be with a reject
flow of about half the quantity of water injected and tests then indicated
a substantial reduction in fibre reject (more than one third) with no signi-
ficant decline in the removal of specks and shives. A commercial model is
now available and is frequently recommended for the final stage of a
battery of cyclones. In this position any deleterious effect on the efficiency
of removal of impurities should not be very important and the pressure
of injection water can be adjusted to minimize fibre loss and remove
any danger of plugging. It would not be advisable to use this device on
the primary stage because it is probable that changes in the injection
water pressure have a marked effect on the cyclone performance.
Various other methods are available for preventing suction of air into
the reject and some of these are also claimed to remove free air already
in the stock. All involve applying a suction to the reject orifices and are
used mainly for the first stage but can also be found on other stages. The
simplest device provides suction on a water ejector principle and is termed
an 'eductor'. The most elaborate has the reject discharging into a common
manifold to which a controlled vacuum is applied via a level-controlled
159
receiver and separator (similar to the arrangement used for suction boxes).
The expense of the latter system is relatively high and its likely performance
must be weighed against the desirability of actually removing air from
the stock in the first place. As a final degree of elaboration the deculator-
cleaner system with accepts discharging into a large tank under very
high vacuum has been developed. With this device, which definitely
removes a substantial portion of air in the stock, the rejects must also of
necessity be enclosed and joined to a common manifold. Applying suction
to the reject and accept lines does not apparently have any significant
effect on the efficiency of impurity removal, provided the pressure drop
from inlet to accept is maintained and the absolute pressure on the reject
line is kept similar to that on the accept.
28 . 6 4 Influence of reject diameter on performance

In the straightforward cyclone with open reject orifice, the general per-
formance is greatly influenced by the actual diameter of this orifice. This
is relevant at the planning stage of an installation and also because widen-
ing of an orifice with age due to gradual wear alters the character of the
reject. In this section the influence of reject diameter will be discussed.
Information presented here and in the following section represents a
consensus of work reported from several sources, in particular references
3, 7, 19, 21, 26, 28, 29. The data refers to a 3 in. diameter cyclone, though
the shape of the curves apply equally to cyclones of other diameters.
Efficiencies have been assessed experimentally using a variety of readily
identified particles such as dyed shives and sawdust, radioactively-tagged
shives, ash, rubber, sand and graded grit.
With a reasonably wide orifice, the reject flow discharges in a steady
spray from the bottom of the cone. As the orifice is reduced in diameter
a point is eventually reached where the spray become unstable and it
finally changes to an uneven and intermittent dribble. This occurs with an
orifice diameter of less than -rl>- in. with a 3 in. cyclone and i in. with a
6 in. cyclone. If reduced further, even this flow will eventually cease. It
is generally agreed that efficient smooth functioning of a cyclone is impaired
when the reject orifice is reduced below the point where the discharge
ceases to be in the form of an even spray; this marks the minimum accept-
able flow and is generally in the region of from 2 per cent. to 4 per cent. by
volume of the inlet flow depending on the geometry of the cyclone. Above
this minimum, the volume of the reject flow depends closely on the orifice
diameter and the relationship is such that the flow is approximately
proportional to the square of the diameter, see Fig. 2. 12.
When considering the efficiency with which various types of particles
are rejected in cyclone cleaners, it is common to obtain either figures in
the region of 80 per cent. to 90 per cent. or more, or quite low values.
The reason for this is the relatively sharp cut-off of the cyclone, i.e. the
sharp increase in efficiency that occurs once a given size of particle is
exceeded, as shown in Fig. 2. 7 on page 123. It is therefore obviously more
relevant that discussion of the effect on performance of making various
160
alterations is centred on changes in efficiency of rejection of a size of particle
which is already in the high region. Fig. 2. 12 shows such a typical relation-
ship between the reject orifice diameter and efficiency of rejection, the latter
being defined as the percentage of ingoing particles found in the reject flow.
As would be expected, efficiency increases as greater flow to the reject is
allowed.
In cleaning paper stock, efficiency must be assessed in terms of the
reduction in impurities in relation to fibre. In other words, efficiency must
loo
,:. Bo
\j
w
.....
w
~
g 4"
"''9
~
zUJ
v
et Zo
w
Q...
REJECT ORIFICE: OIAM.ETEQ, M~s.

Fig. 2. 12. Dependence of percentage number of particles and of flow volume to reject
on the size of the reject orifice for a typical 3-inch cyclone
be calculated as the reduction in particles passing out in the accept flow

per given quantity of fibre; in practice due to the relatively small change
in accept flow consistency under different operating conditions, the
efficiency can equally well be determined on the basis of the reduction
per given volume. If the efficiency curve in Fig. 2. 12 is modified to take
account of changes in the volume flow to reject, the curve in Fig. 2. 13
results; the difference between this based on a correct definition of efficiency
and that in Fig. 2 .12 which relates only to the total number of particles
going to the reject is seen to be relatively small except when higher reject
volumes are reached. This point can therefore be ignored without any sig-
161
nificant loss of accuracy though it should be noted that for lower efficiencies
the difference would become greater.
At first sight then it appears that efficiency of cleaning can always be
increased by opening up the reject orifice. While certainly true, in practice
account must of course also be taken of the fact that fibre in the reject flow
also increases as the orifice is opened up. Even with several stages of fibre
recovery in an installation the fibre lost from the final stage is dependent
on the fibre rejected from the previous stages, so increase in the reject
diameter effectively produces a similar increase in the fibre lost from the
system. A typical relationship between fibre rejected and reject orifice
diameter is shown in Fig. 2. 14. It will be noted that the rate of increase
~ loo
7
Q
9o
~
.,w
VJ
C1 So
u.
0
)-... 7o
\)
zUI
0
u.
L\.
w
0 z a lo
~EJ~CT
4
ORIFICE " DIAMET'E-R MMS.
Fig. 2. 13. Actual efficiency of rejection in relation to reject orifice diameter for the
same 3-inch cyclone
is not so great as for the actual volume of flow rejected and this is due to
the fact that the thickening factor, i.e. the value of the reject consistency
in relation to that of the inlet, varies as shown also in Fig. 2.14. The
thickening factor first shows a slight increase but beyond quite a low orifice
diameter further increase in size of the orifice causes a gradual diminishing
of the factor so that the rate of increase in fibre rejected is lower than the
rate of increase of volume rejected.
It is evident therefore that an increase in efficiency is gained only by
sacrificing a greater fibre loss. The relation is shown in Fig. 2. 15 and the
choice of whereabouts on this curve to operate obviously presents an
economic versus quality problem that can only be related to individual
machines. The rise in efficiency as the reject orifice is opened up from the
minimum flow possible is at first relatively high, whereas the increase in
fibre lost is much slower. It is reasonable on this basis to assume that most
cyclone installations have a fair margin of running conditions where the
reject orifice diameter can vary over a tolerable range without noticeably
162
."o
Jo=
u
.,
UJ
l.11
Ol
~ 4o
w
et
cQ
u:
u.
0
2o
0 2
I< EJECT
4
ORl
'°
FlCE OlAM.ETER.
a
MM&.
lo
Fig. 2.14. Dependence of thickening factor and of the reject percentage of fibre on
the reject flow from the same 3-inch cyclone
affecting efficiency. If a battery of cyclones is first installed with the size

of the reject close to the minimum possible, then a reasonable amount of
wear can be tolerated before the performance becomes economically
inefficient due to excessive fibre loss.
2B. 6 5 Influence of operating conditions on performance

There are two important operating conditions which closely affect the
performance of cyclones: the feed pressure and the stock consistency.
Both these variables affect performance not only in regard to the efficiency
of rejection of impurities but also in respect of operating costs to deal with
a given throughput of fibre. These points will now be considered.
163
Fig. 2. 16 shows for a 3 in. cyclone the relationship with other con-
ditions constant between input pressure and efficiency, and also between
pressure and throughput. It will ·be observed that both efficiency and
throughput increase with higher pressure but that both begin to level
out in the region of 40 to 50 p.s.i., especially the efficiency. The behaviour
of other characteristics is less certain as reports are conflicting in some
respects, but there is evidence that with increasing pressure the percentage
of flow in the reject diminishes while the thickening factor increases. The
net result of this is that the fibre passing out in the reject flow shows little
change with pressure.
Based on this it is generally recommended that little is gained from
running 3 in. cyclones above about 40 p.s.i. inlet pressure. For larger
loo
~
x()
9o
F
...,
u
...
at 80
u.
0
>- 7o
0
....z
0
ii:
II.
w ? 0
I
fo
Fle,Re.
io
Re:J ~c.Teo
I
3o
f"
I
4.o
Fig. 2.15. Relationship between efficiency of rejection of impurities and the fibre
lost in the reject for increasing reject flow
cyclones the optimum pressure is considered to be slightly higher though

of course the efficiency level for comparable spherically-shaped particles
would be much lower at any given operating pressure. Some observations
by Gavelin and Sikstrom (28) have indicated that at lower pressures
removal of specks from paper is less efficient (in confirmation of the
above), but that removal of shives is better than at higher pressures.
With regard to consistency of the stock pumped through a cyclone
battery, as this is increased a point is reached where efficiency of rejection
of impurities suddenly drops off. This point depends·-on the reject diameter,
the drop occurring at a lower consistency when reject diameter is smaller.
The changes in efficiency are presumably connected in some way with the
change in viscosity which occurs somewhere above 1 per cent. consistency
and with the influence on vortex strength of the viscosity, especially
164
towards the apex of the cone. A typical example of change in efficiency
with inlet consistency is shown in Fig. 2 .17.
Also shown in this figure is the change in the percentage of fibre rejected.
This is seen to have a minimum, the reason for which is somewhat obscure
but appears to result from the opposition of two changes which occur
c Zo
i
~
A..
...s: IS
~
(j
V)
....:J lo
a..
:r
\!]
:::l s
&
:r
.... 0
~ lo
z
()
5
UI
8o
~
Ql
u.
0
Go
>-
\)
z
w
\:) A.o
ii:
lL
UJ ?
0
lo 2o ?>o 4o So
INPuT PRESSURE ps1
Fig. 2. 16. Dependence of efficiency of rejection of impurities and also of throughput
on the input pressure to a 3-inch cyclone
with increasing consistency: the thickening factor decreases, with a level-

ling off above about l ·5 per cent. consistency, while the percentage of
flow going to the reject gradually increases with increasing inlet con-
sistency.
The presence of this minimum implies that under any given conditiQns
of input pressure, type of stock, cyclone geometry, and so on, there is an
optimum consistency which will keep loss of fibre to a minimum. This
seems to occur at a higher consistency than the point where efficiency
begins to diminish sharply. The consensus of opinion on this matter is
165
that consistencies should ideally be kept in the 0·6 per cent. to 0·9 per
cent. region for optimum performance, though this must vary with con-
ditions and some workers notably Brecht et al. (19) advocate lower
consistencies than this.
The consistency also has a bearing on the cost of treating a given quan-
tity of fibre because the higher the consistency the lower the pumping
loo
So
t
.,w <Oo
UI
at
...Q
C&J 4.o
~
~
-z.
...
u
ot
~ 2o
O·ZS O•S o·75 f·o 1·25 l·S

\N\..eT C.ONS\STENC.Y %.
Fig. 2. 17. Dependence of the efficiency of rejection of impurities and of the fibre
rejected on varying inlet consistency
costs. This point has been examined by Robinson and Kingsnorth (29)
in relation to the inlet pressure. Fig. 2. 18 from their work gives an illus-
tration of how the cost of pumping is related to consistency and pressure.
Ideally curves of this nature need to be considered in comparison with
those of efficiency of impurity removal under the same conditions.
In practice most machine systems do not permit running of stock to
the breast box at a consistency selected at will, and this is in fact governed
by drainage and other variables on the wire. Stock is normally fed to the
166
box at consistencies under 1 per cent., which from the foregoing appears
quite suitable. Only for machines using breast box stock at very low con-
sistencies, say below 0·3 per cent. or 0·4 per cent., does the fibre reject
begin to get excessive and the efficiency improve little with more dilution.
In such cases it can often be advisable to dilute to between 0·6 per cent.
IS
~
\j
flUl
a
.-i: 10
z
~
Ot
w
0.
"'
>-
i
5
0:
:i.
0 O·S
CONStSTE.NCY. %.
Fig. 2.18. Dependence of power requirements (assuming 50 per cent. overall pump
efficiency) on operating pressure and consistency (after Robinson & Kingsnorth)
and 0·8 per cent. or so for cleaning and follow this by a further dilution
before entering the breast box. This procedure would be economical
anyway to reduce the size of the installation required to clean a given
tonnage of fibre and to cut high-pressure pumping costs.
2B . 6 6 Choice of cyclone size

There are on the market a range of sizes of cyclone from 2 in. diameter
up to 12 in. diameter and greater. The choice of which size to use for any
specific application is one of the most important decisions that have to
be taken. The advantages of using a large size of cyclone can be listed as
follows:
(a) the initial cost is generally cheaper because a lower number of
units and less ancillary pipework, pressure gauges, isolating valves,
etc., are needed.
167
(b) there may well be a fewer number of stages needed and this reduces
running costs and saves the expense of an additional pump and
associated pipework, trough, etc.
(c) it is often not considered necessary to provide a higher inlet pressure
to larger cyclones, so that pumping costs are similar to those for
smaller cyclones.
(d) the larger reject orifice diameter implies less possibility of plugging
occurring.
(e) space requirement is lower.
Against these points, the principal advantage of smaller cyclones is their
undoubtedly superior efficiency in removing small, heavy, spherical-shaped
impurities, especially grit, ash and sand. In regard to shives, the matter
is more complicated for small cyclones are less efficient than large for
removing relatively large shives of a long and slender shape. This is of
importance with groundwood and other furnishes used for lower grades
of paper, though not if screens are used in conjunction with the cyclones.
Normally it would be argued that for efficient removal of shive-like material
screens are necessary anyway, as even larger sizes of cyclones have a
relatively poor separation efficiency for this type of material. Thus this
particular deficiency of small cyclones is relevant in only a few special
cases.
A second advantage to small cyclones is that fibre loss from an instal-
lation can generally be kept lower because of the possibility of reducing
the size of the final stage to treat a lower capacity than a single larger-
size cyclone. However, this point is not of great importance in practice
because there are other methods, as described above, of reducing fibre
loss from an installation and in any case a number of small size units
could be used just as effectively for this purpose in the final stage of a
battery of otherwise large cyclones. Also for any given stock the thickening
factor is higher for smaller cyclones and this tends to become more
disadvantageous in the final stage and also to offset the advantage of
smaller capacity.
Early work by Rastatter and Croup (3) established that the 3 in. cyclone
was economically preferable to larger sizes from the point of view of
obtaining maximum cleanliness at minimum cost. Consequently when
commercial models first became available it was this size that was normally
recommended for installation. Since then there has been a steady trend
towards using the larger sizes of cyclones until now the 12 in. diameter
is probably becoming the most popular. The reasons for this gradual
change hinge on the various advantages of larger cyclones as listed above,
but undoubtedly the most important factor has been avoidance of plug-
ging. This in itself has become even more imperative because of the
increasing interest in combining cyclones with devices to remove air that
necessitate an enclosed reject; in such cases a plugged reject orifice cannot
easily be detected and cleared so that no risk of this occurring can be
taken.
Batteries of 3 in. cyclones not only contain a large number of units but
in the vast majority of applications are a positive bugbear to keep running
168
SC'R.EENS AND CLEANE'R.S 2B.66
satisfactorily. A continual inspection of the reject tips is invariably needed
and can demand a great deal of the machineman's time. It is essential
to have a means of rapid isolation of individual cyclones (normally by
valves or cocks on each inlet and accept line) so that the reject tip can
be removed, cleaned and replaced as quickly as possible. Various attempts
to avoid this plugging have been made, but none have been wholly suc-
cessful. The most promising was the use of a soft rubber tip which distends
under pressure when the accept flow is temporarily closed off and thus
blows out debris accumulated at the bottom, but in practice wear of
the tip proved too much of a disadvantage. Assessment of the likelihood
of plugging in any situation is difficult though as a guide it is sometimes
said that this depends on the frequency of impurities greater in size than
about half the proposed reject orifice diameter (an impurity does not need to
exceed the orifice in size in order to cause blockage because fibre rapidly
collects round a particle in the vicinity of the reject and a partial blockage
can develop first before the reject stops completely). It is always of course
feasible to give stock a preliminary coarse cleaning to remove impurities
likely to cause plugging, though this inevitably increases the overall cost
for achieving higher cleanliness.
It has also been argued by the advocates of larger cyclones that so long
as the efficiency of removal of impurities is sufficiently high, there is no
virtue in using the smaller diameter cyclones with their attendant dis-
advantages. This may well be true in some cases but it would be unusual
for a mill to be able accurately to define a lower limit of dirtiness for its
paper below which it did not consider it worthwhile to go. It is safe to
say that most mills would prefer complete· freedom from blemishes in
their paper, and the smaller size of cyclone is more likely to achieve this.
Further in regard to such benefits as longer wire life or interval between
calender grinding, removal of really small grit probably confers as much
advantage as removal of the larger pieces. So it may equally well be
argued that if a mill is going to the expense of installing and running
cyclone cleaners it should demand the highest practicable efficiency.
For this the smaller the cyclone the better, so it is well worthwhile always
to look closely at the relative importance of the various disadvantages
to using the small size. In this respect the Radiclone design merits con-
sideration as it appears to overcome several of the disadvantages mentioned
above, especially that of plugging.
A final point regarding the use of small cyclones concerns paper stock
with a heavily-loaded furnish. Loadings with a large particle size, parti-
cularly china clay, are rejected at a comparatively high efficiency which
can mean that a heavy loss of 10 per cent. or more is sustained from the
final stage of a 3 in. cyclone installation. Also when there is a low retention
and breast box stock carries a high proportion of loading to fibre, the
concentration of loading from one stage to the next can become so high
that excessive dilution is needed, thereby increasing capital and running
costs as well as fibre and loading loss due to the larger second and third
stages required.
Where this is a problem it is often the case that large cyclones are the
169
only ones economical to run. To give some indication of the position in

this respect some data given by Jacobsson (21) are worth quoting. He
reports that rejection of such loadings as titanium dioxide (which has a
small particle size) is in practice negligible, but for finely graded china
clay the rejection in a single pass through a cyclone is of the order of
20 per cent. for a 3 in. diameter unit, 3 to 4 per cent. for a 6 in., and 1
per cent. for a 12 in. unit. If the clay is relatively coarse these percentages
can rise to 30 to 40 per cent. for a 3 in. cyclone, 6 per cent. for a 6 in.,
and 2 to 3 per cent. for a 12 in. unit. The advantages of going to larger
sizes of cyclone for stock heavily loaded with clay are thus evident.
170
CHAPTER 2C
RUNNING SCREENS AND CLEANERS
2C.1 DAILY OPERATION
In dealing with the running of screens and cleaners, the same procedure
is followed as in other Parts of the book. Measurements required by the
machineman for day-to-day operation are covered first, and this is followed
by a discussion of the longer-term maintenance needed. Finally, practical
aspects of running screens and cleaners are dealt with.
2C . 1 1 Measurements required
Practically all the measurements required for successful operation of
screens are straightforward indications of pressure. For flat and open
rotary screens it is important that the pressure of sprays is measured.
With enclosed screens measurement of the inlet pressure of the stock is
important, especially when there are a number of units in parallel and it is
necessary to adjust the flow through each to be similar. Equally it is
important to measure pressure on the outlet side of a screen.
The difference between these inlet and outlet pressures is a measure
under given flow conditions of resistance to passage of stock through
the screen. Should this difference begin to rise this is an indication that
the screen plate is beginning to clog and therefore gives the machineman
early warning that action is necessary. In some cases the risk of this
happening may be so high that a differential pressure gauge is considered
necessary. Developing this even further, because changes in the differential
pressure between inlet and outlet of an enclosed screen are relatively
gradual and early detection of any trend to clogging is vital, it is occasion-
ally thought advisable to link the reading to a recorder.
In a similar way measurement of pressures is the basis of successful
cleaner operation. With both the cylindrical and cyclone type the pressure
in the inlet manifold has to be measured and set to the recommended
running level. Pressure in the accept must also be measured to ensure no
restriction in the line is raising this above the low value usually required.
This is very important because the whole action of a centrifugal cleaner
depends on the pressure drop from inlet to accept. These measurements
are, of course, equally essential in each stage of reject treatment.
When the rejects from a number of cylindrical or cyclone cleaners are
connected together then measurement of pressure on this particular
manifold is necessary to ensure it does not rise above the recommended
minimum. This is especially important when the line is operated under
vacuum. The pressure of whitewater or fresh water used for dilution
between stages of an enclosed system must also be measured to permit
171
2C.l 2 SCREENS AND CLEANERS
flows to be balanced. This applies too to water used for elutriation or

vacuum ejector purposes.
Finally, the power required to operate a screen is of obvious importance,
especially for enclosed screens where the power to rotate the impeller
can be a useful indication of any tendency for the plate to clog or the level
of stock to be low inside the screen. The power used by pumps in cleaner
installations must also of course be measured.
2C. 1 2 Control applications

There are very few control applications vital to the operation of screens
and cleaners though they are becoming more common. The flow from the
reject of a primary enclosed screen to a secondary has to be kept steady
and this is one area where a number of automatic control mechanisms
have been introduced. Straightforward pressure control on the reject
valve would be possible but this method has not found favour because
the basic problem here is to prevent the valve itself from gradually becom-
ing plugged due to the relatively high consistency usual with an enclosed
screen reject. The most favoured approach has been to use a special type
of automatic valve which can be opened momentarily and then returned
to its original position; this is controlled by a conventional timing mechan-
ism which allows both the frequency and the period of opening to be set
as desired. There are various techniques used for this purpose, some of
which are applied also to the periodic opening of screen traps and the
collecting chambers of cylindrical cleaners. The same principle in reverse
has been used to close temporarily the accept line from cyclones in order
to raise pressure in the unit and blow out any debris collecting in the reject
orifices.
Occasionally control systems involving the pressure drop across cleaners
are used. These usually work on a re-circulation line from the primary
accept back to the inlet, this flow being regulated to keep the pressure
drop constant. With open-reject cyclone units discharging into troughs,
the level in these is kept steady by controlled addition of dilution water.
When vacuum is applied to the reject line of cleaners it is generally
governed by a straightforward automatic control, especially when a pump
is used. Level control in the vacuum receiver tank would also be usual.
Similarly vacuum and level controls are used when vacuum is applied to
the accept of a cleaning installation, as for example with the deculator-
cleaner system.
2C. 2 MAINTENANCE
2C . 2 1 General maintenance
Apart from general engineering maintenance and lubrication which will
be the responsibility of the engineering department, the main attention
the papermarker must give to screening equipment is to keep it clean and
check the screen plates for fracture and wear. Cleanliness is covered in
2C. 3. With regard to the screen plates, it is important that each is given
172
SCREENS AND CLEANERS 2C.2 l
a careful examination every shut period. With the open rotary and flat
type this involves little more than a visual examination and inspection
of holding bolts and sealing strips. To inspect the enclosed screen necessi-
tates removal of the cover and for this reason it is preferable to have the
cover fastened with swing bolts and suspended from a davit arm. Any
fractures or holes in the plate means immediate replacement to prevent
large impurities or fibre and slime lumps passing through.
Wear of screen plates occurs over a long period so that a regular three-
or six-monthly check of the size of slits and holes is a useful precaution.
Slots that start off at, say, 18-20 thou will widen steadily and the point at
which replacement is necessary, 25 thou or even 30 thou, must depend
on assessment of the deterioration in terms of performance. Several
positions should be checked at random in the screen for this purpose,
taking care that separate plates are each checked individually. With the
enclosed screen, holes in different parts of the compass should be checked
as wear is not always even.
Other points of maintenance of screens depend on the type in use.
Scrapers on certain types of flat screen must be checked periodically for
contact with the surface. On the inward-flowing rotary screen the sealing
strips at the ends of the screen should be examined regularly. These fre-
quently need quite a lot of attention to ensure a good fit. No bolts should be
missing from the plates. The screen plates themselves occasionally need a
thorough cleaning with acid or some other suitable chemical compound
depending on whether they are subject to corrosion or scale build-up.
With regard to centrifugal cleaners of the cyclone type, the main main-
tenance necessary is change of the reject orifice tips when these become
over-large. Generally, this becomes evident when fibre loss has increased
to an unacceptable degree, but it is worthwhile periodically to measure
the orifice diameter of several or all cyclone units and keep a record of
the readings. Wear takes place more rapidly in the later stages of a battery,
especially the final stage which of course is mainly responsible for the
fibre loss, so cyclones in this position require frequent checking depend-
ing on abrasiveness of the stock.
Occasionally a complete cyclone should be dismantled and inspected
because wear and scale build-up can affect the inside surface. This can
reach the stage where a regular spiral groove from the inlet to the reject
orifice is visible and there is no doubt that this impedes efficient removal
of impurities. Close to the reject orifice a number of deep circular grooves
are sometimes cut in the body of the cyclone where a large abrasive particle
has spun round for a long time. Once such a groove is started it tends to
collect and hold further particles, thus accelerating the wear. This is a
region where flow should be very smooth and is therefore particularly
sensitive from the point of view of efficiency. Generally all that can be
done in cases of bad wear is to re-line or replace the whole cyclone.
Cylindrical cleaners do not require a great deal of maintenance other
than occasional changing of the reject diaphragms which gradually wear.
Each unit should be dismantled at intervals to inspect for roughness and
pitting of the surface.
173
2C.22 SCREENS AND CLEANERS
2C. 2 2 Long-term records

Keeping a check on the long-term performance of screens and cleaners
is obviously important but unfortunately it is one of the most difficult
tasks to do with any accuracy. The main purpose of both screens and
cleaners is a negative one, to remove from the stock impurities which are
likely to give trouble and cause downtime from one cause or another.
Accordingly their efficiency depends in the first place on how big a job they
have to do at any particular time. A dirty batch of pulp can cause havoc
on the paper machine, but the screening and cleaning equipment may be
nonetheless functioning at the usual level of efficiency which just happens
not to be good enough to cope satisfactorily with this particular situation.
In a similar way there are generally seasonal variations in pulp and this
too can affect the apparent efficiency.
Ideally it would be preferable to be able to measure the efficiency at
intervals to provide long-term records of performance. But enough has
already been said about the problems of doing this with a mill installation
to make it obvious that any direct test is out of the question. Records
of such things as dirt counts on the web or on sheets, reels rejected or
sheets sorted out due to impurities, breaks of the web suspected as being
caused by lumps, and even of wire life where an abrasive stock is con-
cerned should all in theory give some indication of changing efficiency.
But these are also affected by so many other factors that in practice it
would appear highly unlikely that any direct change in efficiency of screens
or cleaners can be detected by this means.
Turning specifically to screens, provided the slit and hole sizes in plates
are measured regularly this would appear to be the most useful check
on likely performance variation that can be carried out. It is also useful
to take samples from the inlet and reject of screens, both primary and
secondary, to assess the consistency and loading changes that take place;
on some types of stock fibre fractionation and freeness figures can also
give valuable data. The reject flows should be measured, as should the
flow and consistency of any spray water. Together with measurements
of flow, consistency and so on taken throughout the wet-end flow system
(as discussed in lC. 2 3), a fairly complete picture of performance can be
built up. Comparison over a long period of results obtained in such
checks gives useful confirmation of any trends likely from wear of the
screen plates.
Similar checks are necessary for cylindrical and cyclone cleaners. Here
samples from the inlet and reject of each stage and of dilution water
should be taken and checked for consistency, loading and possibly also
for fibre fractionation. Flows through the system can be easily calculated
but a check on the actual flow of dilution water to each stage or the whole
installation is useful for confirmation. Where elutriation water is used,
the flow of this too should be measured.
Of particular importance is the flow from the final stage since this will
almost invariably represent loss from the system. Checks of the flow,
consistency and loading content of this loss should be carried out more
174
SCREENS AND CLEANERS 2C.3
regularly than would be usual simply for the purposes of long-term records
which are at present under discussion. At least once a week is advisable
if the loss is not suddenly to be discovered much higher than was thought.
It is also essential to check this loss on each major grade of paper made
because it can vary considerably, especially with loaded stock, from one
furnish to the next.
Sheets can be made from the reject of the first and later stages of a
cleaning installation and from the reject of a screen for the purpose of
visual examination of impurities. But this is not advisable because inter-
pretation of any differences observed is very tricky. Even with the same
grade of stock and, as near as can be arranged, the same flow and pressure
conditions, the reject can vary greatly in impurities. Certainly this is true
in regard to the actual quantity of impurities present, if only because this
depends closely on the amount of fibre being rejected at any given time.
About the only use such sheets can have is to illustrate at any time the
sort of impurities that are present in the stock. This is of course quite
different from giving some indication of efficiency.
2C.3 PRACTICAL POINTS

2C. 3 1 Start-up
The exact starting-up procedure for a screen depends on the method
used for the wet-end as a whole. If the machine wire pit is filled with fresh
water and this is pumped round the system first, fresh stuff being gradually
added when everything is already operating,· then few problems are likely
to arise. Whether open rotary or enclosed screens are used, these are
started immediately before the mixing pump discharge valve is opened.
Power consumption is then checked and where sprays are used these are
turned on. The vibration mechanism and sealing strips of an open screen
should be examined. Reject valves are set in their usual operating position.
With an enclosed screen the air bleed-off is opened.
As soon as stock starts coming through, the action of the screen must
be carefully watched. An open rotary screen should present a completely
clean appearance at the top with no areas where fibre is left hanging.
Also the level in the vat should be stable. With an enclosed screen the
only indication that all is well comes from checking the pressure drop
between the inlet and outlet. The reject flow to the secondary screen is ad-
justed as necessary and when a non-clogging valve is used this is switched on.
When the breast box stock has approached normal running conditions,
just before the web is fed over from the couch, screens should be given a
final check-over. This is to ensure that they are coping all right with the
increase in consistency that occurs as a growing quantity of fines re-circulate
in the backwater. The spray water is turned over either manually or auto-
matically to whitewater. When there are several screens in parallel the
stock should be evenly divided between them; this is assessed either
visually with open screens or from the relative pressure drop across en-
closed screens.
175
2C.3 1 SCREENS AND CLEANERS
In older systems, particularly those using a mixing box, it is customary

to turn water and fresh stuff on together so that the first flow fed to the
screen contains stock at near normal consistency. This can happen too,
of course, if there has been a temporary shut on a machine and the back-
water pit already contains stock. The danger here is that the initial rush
of stock is at a relatively high consistency and in this event immediate
clogging of the screen plate may well occur. To avoid this there should
always be a delay after starting the system up before the fresh stuff is
turned on. With enclosed screens provision is sometimes made to fill the
screen with water before the impeller is started and stock admitted. Both
these precautions should make clogging much less likely.
Secondary screens are of course started at the same time as the primary
and sprays, dilution water, and power consumption checked. With the
open fiat screen the vibration mechanism and, when used, the scrapers
should be examined. Once flow of stock from the primary reject starts
to come through to the secondary screen the operation requires more
attention. Generally, however, there is little point in setting dilution water
to the secondary screen and adjusting the sprays to give the reject desired
until flow at the wet-end has settled down. Secondary screens of the flat
type can often be tricky to set satisfactorily so this is probably best left
till the machine is under way.
A cleaning installation comprising several stages of cyclone or cylindrical
cleaners does not generally require the same care in starting as screens
do provided a standard sequence of operations is followed. Dilution
water is first turned on and, with the conventional cyclone set-up, troughs
allowed to fill with water. When appropriate, elutriation and 'eductor'
water is turned on and the pump on a reject vacuum line started. The
procedure then is to set the final stage going first, then the previous stage,
leaving the primary stage until last. The sequence must be carried out
reasonably quickly, as there may be an insufficient capacity of dilution
water available to keep the stages filled, so normally it is not commenced
until immediately before the stock is ready to be delivered from the
mixing pump.
Each time a stage is started pressures on the inlet and (when enclosed)
on the reject lines are adjusted if necessary. Note must also be taken of
the accept pressure because any restriction in this line causing an increase
in pressure will reduce efficiency considerably. When a vacuum pump is
used on the reject or accept line the gauge on this and the level in the
receiver are checked. Occasionally a re-circulation line is used to balance
out flow to the cleaner installation and operation of this to keep the
accept pressure at the usual level must also be correct. It is always prefer-
able to design the system so that only a single valve on the discharge side
of each pump has to be opened, leaving all other valve settings unaltered
from previous running. This is because it is easy particularly with a totally
enclosed system to upset the balancing of flows to different stages.
Once stock has approached equilibrium conditions the cleaners should
be examined carefully. Stock building up in closed reject chambers and
discharging from the final stage of a cyclone battery should be checked
176
SCREENS AND CLEANERS 2C.3 2
and the dirt content noted. Especially at start-up there can be a large
accumulation of impurities which necessitates early cleaning out of closed
reject chambers unless these are emptied automatically. The reject nozzles
of smaller diameter cyclone cleaners are particularly apt to plug at start-
up and when this occurs there is an immediate drop in efficiency of the
cyclone unit concerned. Unplugging cyclone reject orifices can be a very
demanding task for the first few hours after start-up.
2C. 3 2 Shut-down
Before shutting down screens and cleaners it is always best if backwater
or fresh water is pumped through to prevent thick stock lodging anywhere
in the system. The best method of shutting the wet-end is to close the fresh
stuff line and run the fibre out on the wire to the hog-pit, allowing the
backwater in the machine wire pit to be pumped round for a short time;
if this procedure is followed then it should suffice to clear the screens and
cleaners of fibre. Otherwise fresh water should be run through.
Open rotary screens should be rotated for some time with the sprays
on and then thoroughly hosed down with a high-pressure jet. Enclosed
screens sometimes have facilities for back flushing and this is effective
for clearing screen plates. Even so every enclosed screen unit should be
opened up and the impeller examined (taking special note of the gap
between the impeller heads and the plate) and the visible part cleaned out
with a jet. Unfortunately, it is not generally possible to give an enclosed
screen a thorough clean without actually removing the plates so periodi-
cally this should be done also. Apart from this attention screens need
little in the way of regular maintenance so far as the papermaker is con-
cerned. With inward-flowing rotary screens the sealing strips should be
carefully checked. The trash-collecting chambers of enclosed screens
have to be emptied and swilled out.
Cleaners of the rotary centrifugal type require a really thorough hosing
out after the main body of fibre held in the various compartments has
been dug out. Cylindrical and cyclone cleaners cannot be cleaned inside
so it is particularly important that water with low fibre content is pumped
through before the installation is shut. The first stage is then shut down,
followed by the second and so on. Open troughs are washed out and with
some models of cyclone cleaners the nozzles can be taken off and cleaned
if this is usually necessary. Apart from this no other action is needed.
2C. 3 3 Checking screens and cleaners during running
Normally screening equipment requires little attention while running.
With the open rotary type an occasional check must be made that the
rotating and vibrating mechanism is alright and that level in the vat
and flow from the reject are satisfactory. Most attention is probably
needed to the sprays which should always completely cover the length
of the cylinder. When whitewater is used a partially blocked spray is
more likely and this leaves a ring of uncleaned screen plate which eventu-
ally becomes made up, thus effectively reducing the screening area avail-
able. Sealing strips and spray troughs should also be inspected.
177
2C.3 3 SCREENS AND CLEANERS
With the enclosed screen there is little examination possible other than to
check that the inlet and outlet pressures are normal, the power consump-
tion at its usual level, and the reject flow steady. When a continuous air
bleed from the top of the screen is used this must be examined as the
valve will frequently tend to get made up due to the low flow. Compart-
ments catching large impurities must be emptied with reasonable regularity.
Flat screens can require more attention because they are generally
fairly sensitive to changes in flow and dirt content of the inlet stock.
Frequent inspection is advisable or it will be found that the reject has
suddenly increased to a high level and a considerable quantity of fibre
is being put down the drain. Flat screens, particularly when used in a
secondary position, should be cleaned at least twice a shift, the vibration
mechanism checked and other points such as the sprays and action of
scrapers examined. Unfortunately secondary screens of the flat type are
very often placed in out of the way positions and tend to be ignored by
machinemen. This does nothing to improve an already dismal performance.
Centrifugal cleaners of the cylindrical type with reject collection com-
partments have to be emptied at intervals. Sight glasses are usually pro-
vided to assist in deciding when this is needed, but it is probably best for
the machineman to get into a routine of emptying once or twice a shift
to avoid neglect. Cyclone cleaners of small diameter will need frequent
attention to the reject orifice tips to avoid plugging. Provided quick-action
valves or clamps for the inlet and accept pipes are available so that isola-
tion of the cyclone is simple, cleaning out a reject nozzle need not take
long. Occasionally it should be possible to clear out a nozzle simply by
temporarily closing the accept line thus putting the full pressure on to the
reject. Otherwise inspection of cylindrical and cyclone cleaners is confined
to checking over the various pressures through the system. When vacuum
is applied to accept or reject lines, the functioning of this system also
requires occasional examination.
There are always periods on any machine when impurities of one kind
or another appear in the paper or cause breaks. It is at such times that the
screening and cleaning equipment is the first to be scrutinized. Once it
has been thoroughly checked over and no faults can be seen, the machine-
man is naturally at a loss to know what to do except blame the pulp,
which he usually does. Occasionally, of course, this proves to be a correct
accusation, and then little can be done except either re-circulate more
stock flow round the screens and cleaners (when this is possible) or divert
secondary accept flow to the preparation system in the hope that some
impurities will be broken down to a more acceptable level.
A laboratory analysis of impurities appearing in the paper can be very
valuable to give a clue as to the source of trouble. It is after all always
preferable that impurities are eliminated at source rather than to depend
on the screens and cleaners to remove them. There are one or two other
actions that may be taken. Holes and slots in screens and cleaners can be
checked (particularly if this is not done on a routine basis). The fresh
water and backwater lines can be opened to see that they are not a source
of contamination from rust, scale, slime, etc. Additive systems, especially
178
SCREENS AND CLEANERS 2C.3 3
of loading when this is placed in slurry form into the mixing box or
pump, are also worth checking over. There are usually many places where
dirt and slime can lodge in open tanks, the breast box, vacuum receiver
tanks and manifolds, and so on. Such positions should be gone over with
a tooth comb. Finally, of course, there is always the possibility that the
impurities are getting into the paper at a later stage by falling from the
roof on to the wire, being picked up by felts at the press, or coming from
the drying cylinders and felts. If an analysis is made of the impurities in the
first place, this should serve to narrow the field and enable attention to
be confined to those parts of the system most likely to yield up the culprit.
179
REFERENCES
(1) Chester, C. A.: 'A Study of the Operation of Sand Tables.' Proc. Tech. Sec. 31,
3, 617. November, 1950.
(2) Minute, A.: 'Modern Methods of Cleaning Paper Stock.' Proc. Tech. Sec. 33,
l, 29. February, 1952.
(3) Rastatter, E. L., and Croup, A. H.: 'Hammermill Dirt Separator.' T APPi 35,
5, 223. May, 1952.
(4) Steenberg, B.: 'Principles of Screening System Design. Studies in Screening
Theory I.' Svensk Pap. 56, 20, 771. October 31, 1953.
(5) Almin, K. F., and Steenberg, B.: 'The Capacity Problem in Single Stage Screen
Cascades. Theory of Screening II.' Svensk Pap. 57, 2, 37. January 31, 1954.
(6) Anderson, D. D.: 'Centri-Cleaning of Sulphite Pulp.' TAPP! 38, 9, 183A.
September, 1955.
(7) Stockman, L., and Ruus, L.: 'Cleaning of Pulp with Cyclones.' Svensk Pap. 58,
15, 483. August 15, 1955. (In Swedish).
(8) Boadway, J. D., and Freeman, H.: 'Centrifugal Classifier for Cleaning Pulp and
Paper Stocks.' TAPPI 39, 11, 797. November, 1956.
(9) Gorham, A. B.: 'Reducing the Dirt Content of Unbleached Sulphite Pulp with
Centrifugal Cleaners.' TAPPl 39, 7, 199A. July, 1956.
(10) Smith, J. A., and Stewart, J. C.: 'The Deculator-Cleaner Process.' TAPPI 39,
3, 198A. March, 1956.
(11) Wen berg, H. B.: 'Nicolet's Experience with the Vorvac System.' Paper Trade
Journal 140, 5, 34. January 30, 1956.
(12) Brown, A. D.: 'Cleaning of Board Furnish with Vorject Cleaner and Vortrap
Classifier.' Canadian Tech. Sec. Proc. p. 381. 1957.
(13) Casper, R. R.: 'An Investigation of Centri-Cleaner Performance using Radio-
active Tracers.' TAPPI 40, 5, 175A. May, 1957.
(14) Dyson, A. 0.: 'The Vorvac, Vorject and Vortrap as applied to the Paper Machine.'
TAPPI 40, 1, 136A. January, 1957.
(15) Espenmiller, H.P., and Zuehlke, E. H.: 'The Shartle Liquid Cyclone-A Centri-
fugal Cleaner.' TAPPI 40, 1, 140A. January, 1957.
(16) Graham, W. M.: 'Centrifugal Cleaners on a Newsprint Machine.' Canadian
Tech. Sec. Proc. p. 387. 1957.
(17) McLean, J. A.: 'Use of Miami Selectifier Screens for cleaning Stock Supply to
Newsprint Machines at Great Northern.' Canadian Tech. Sec. Proc. p. 360. 1957.
(18) Nuttall, G. H., and Hendry, I. F.: 'Design and Coupling of Cyclone Installations.'
TAPP! 40, 12, 951. December, 1957.
(19) Brecht, W., and Weishaupt, K.: 'The Suitability of a Hydracyclone for Pulp
Screening.' Woch. fur Pap. 86, 7, 221; 8, 281. April 15, 30, 1958. (In German.)
(20) Corte, H.: 'Calculations of Screening Systems.' Zellstof und Pap. 7, 4, 116.
April, 1958. (In German.)
(21) Jacobsson, J. J.: 'De-aeration of Paper Stock with or without Cleaning.' TAPP!
41, 5, l 79A. May, 1958.
(22) Nuttall, G. H.: 'The Use of Cyclones for Cleaning Pulp and Paper Stock.' Paper
Maker 135, 3, 230; 4, 326. March and April, 1958.
(23) Salomon, S., and Lukianovitch, S.: 'The Application of the Bird Centriscreen
on Newsprint Furnish.' P. and P. Mag. Can. 59, 7, 111. July, 1958.
(24) Tirado, A.: 'Theory of Screening.' TAPP! 41, 5, 235A. May, 1958.
(25) Gavelin, G.: 'Shive Content Reduction of Groundwood Pulp by Centrifugal
Cleaning.' Svensk Pap. 62, 10, 355. May 31, 1959.
(26) Kemp, G., and Rance, H. F.: 'Hydraulic Cyclones in the Cleaning of Paper-
making Stock.' Proc. Tech. Sec. 40, l, 53. February, 1959.
(27) Klinga, I.: 'The Design of Hydrocyclones for the Paper and Pulp Industries.'
Svensk Pap. 62, 1, 11. January 15, 1959. (In Swedish.)
180
SCREENS AND CLEANERS
(28) Gavelin, G., and Sikstrom, L.: 'An Investigation of Two Hydrocyclones for
Groundwood Pulp.' Svensk Pap. 63, 11, 398. June 30, 1960.
(29) Robinson, W., and Kingsnorth, S. W.: 'The Cleaning of Paper Stocks with
Particular Reference to Development of a New Screen for Sisal Removal.' Paper
Technology 1, 6, 635. December, 1961.
(30) Buglass, J.: 'Cleaning Stock for Esparto Papers.' Paper Technology 3, 6, 565.
December, 1962.
(31) Downey, C., and Blake, M.: 'Newsprint Stock Cleaning at Bowater's, Newfound-
land.' P. and P. Mag. Can. 63, 12, T-587. December, 1962.
(32) Hopkins, R. M., MacPherson, R., and Morin, L. J.: 'Analysis of the Effects of
Centrifugal Pulp Cleaners and Pressure Screens on Newsprint Runability.'
P. and P. Mag. Can. 63, 12, T-563. December, 1962.
(33) Anon.: 'The Radiclone-A New Type of Centrifugal Cleaner.' Paper Trade
Journal 147, 52, 22. December 30, 1963.
(34) Lambert, J. E.: 'The Comparative Evaluation of Screens and Screening Effici-
encies.' P. and P. Mag. Can. 64, 9, T-406. September, 1963.
(35) MacMillan, F. A. et al.: 'Shives in Newsprint: Their Detection, Measurement
and Effects on Paper Quality.' P. and P. Mag. Can. 66, 7, T-361. July, 1965.
(36) Sears, G. R. et al.: 'Shives in Newsprint: The Role of Shives in Paper Web
Breaks.' P. and P. Mag. Can. 66, 7, T-351, July, 1965.
7 131
PART 3
THE WIRE SECTION

INTRODUCTION
31 In Part 1 the characteristics of the Fourdrinier paper machine were
considered up to the point where the stock passes through the slice.
General features of recirculation in the main backwater system, the reten-
tion of fibre and loading, and the use of whitewater from the suction boxes
and couch were also discussed. It is proposed now to cover the wire section
itself from the point where the stock impinges on the wire to where it is
couched off the wire. Certain subjects which have already been considered
in the context of the overall wet-end flow system, for example breast box
stock consistency and temperature, are now treated from the point of view
of drainage and formation of the sheet on the wire. Attention is confined
to the basic Fourdrinier design; sheet forming in cylinder vats, rotary
formers, combinations with Fourdrinier wires for duplex and other special
types of paper, the use of secondary headboxes, or such relatively new
processes as Inverform, Twinverform, or Verti-forma are not considered.
In the first few feet of the wire the most important structural features of
the web are determined and the main object of the machineman is always
to achieve as good a formation as possible, in the general sense of the
term. This subject is dear to the hearts of every papermaker and ideas
abound of what happens at this crucial stage of the papermaking process.
There is also a considerable difference between opinions of what constitutes
a good formation, particularly when the term is applied to newsprint and
tissue on the one hand and to fine papers made on slow machines on the
other. Some attempt is made to clarify this whole problem in the light of
observations obtained in recent research but it must be emphasized that
probably in this sphere more than any other it is not yet possible to talk
in terms of a theory as such; generalizations can only be made with extra
care.
Throughout the length of the wire drainage of water from the sheet is
taking place, and this comprises the second main purpose of the wire
section. Drainage and formation are closely connected in the sense that
satisfactory conditions for the one are no use at the expense of the other;
the machineman must constantly compromise between having adequate
dilution of the stock in the breast box to form the sheet and sufficient
capacity on the wire to drain the sheet adequately. Accordingly, close con-
sideration is given in the text to the manner in which drainage of the sheet
occurs, both in the table roll section and at the suction boxes.
When the sheet is sufficiently dry it must be transferred from the wire
to the press section where removal of water by pressing becomes more
efficient and economical than continued application of suction in the boxes
or couch. Transfer cannot occur efficiently before the sheet is strong enough
to withstand the stresses involved and, since the wet strength of the web
increases with dryness, this usually means that the sheet is run as dry as
possible into the presses to avoid breaks. This is generally beneficial due
185
31.l THE WIRE SECTION
to the fact that when the sheet is drier entering the presses, it will be drier
entering the drying section where removal of moisture is relatively costly.
But on the other hand additional dryness at the couch is usually obtained
during day-to-day operation either by reducing the consistency of stock in
the breast box, and thereby affecting formation, or by applying more
vacuum in the suction boxes, thereby affecting the power consumption
and the life of the wire. Each of these factors has to be carefully balanced
and, as in so many aspects of papermaking, some compromise in operation
has to be found.
The couch itself extracts water and so affords an additional means of
reducing the moisture in the sheet before it is couched off the wire. But
other factors are involved here and the couch is, or should be, primarily
designed to facilitate the actual removal of the web from the wire. In this
sense the extraction of water, though an extremely valuable corollary of
the couch operation, should not be allowed to prejudice the successful
transfer of the sheet.
These are the main aspects governing the general operation of any
wire section. In addition, the wet strength of the web at the couch and the
overall quality of the paper is affected in some degree by the compaction
to which the web is subjected on the wire. The vacuum applied to the suc-
tion boxes and the pressure and vacuum at the couch are significant in
this respect, but the most important influence comes from the operation
of a dandy. The compacting action of the dandy inhibits subsequent
drainage from the sheet (as is readily observed from movement of the
dry-line position when a dandy is raised from the wire) and thereby affects
both the stock consistency which may be run in the breast box and the
dryness at the couch.
These brief remarks serve to emphasize the interactions occurring
between the various factors which influence operation of the wire part on a
machine. They also serve to show the complexity and difficulty of assessing
the function and performance of each section of the wire part and illustrate
why the design of this part of a new paper machine remains above all a
purely empirical process relying almost entirely on analogies and extra-
polation from the performance data of existing machines. What follows
represents an attempt to set down the more important features governing
this section of the paper machine; as in other Parts of the book, the method
of presentation involves division of the material into three basic chapters,
though it must be admitted that in this case it has not always been easy to
distinguish between what, because of its fundamental and general nature,
comprises theory, and what should rightly be treated as one of the factors
influencing operation. The author hopes that the division actually chosen
does not appear too arbitrary.
186
CHAPTER 3A
3A .1 DRAINAGE IN THE TABLE ROLL SECTION

On slow machines running a relatively heavy substance, drainage in the
table roll section is largely a continuous process by gravity; the pressure
of water in the fibrous mat on the wire is sufficient to overcome the resis-
tance of surface tension forces within the wire meshes and natural drain-
age through the wire occurs between the table rolls. On very slow
machines this can result in a ff.ow down the leading side of the table rolls,
similar to that occurring in a plain press, as water carried on the underside
of the wire is compressed between the wire and table roll. Under conditions
of gravity drainage the mesh of the wire and the consistency, porosity and
thickness of the mat must be particularly important in determining the
rate of dewatering.
With increase of speed the time available for gravity drainage between
the table rolls diminishes and dewatering of the web by this means becomes
rapidly insignificant except in the region where the stock first meets the
wire. Instead it is readily observed that drainage occurs almost exclusively
on the trailing side of the table rolls. All experimental work on this subject
reported in recent years has concerned faster machines with this form of
drainage at the table rolls, and what follows is confined to this aspect.
3A. 1 1 Suction developed at a single table roll

At one time it was considered that surface tension forces were responsible
for pulling water out from the web on the trailing side of a table roll, but
in the early 1950's it became clear that this force was inadequate to explain
the quantities actually discharged and that separation of the roll and wire
surface in fact produced a suction force on the web. Several investigators
approached the problem of elucidating the nature of this suction force
both from the experimental and the theoretical angle and as a result of
their efforts though many details remain unexplained an adequate general
picture can now be given.
Direct measurement of the suction force has been attempted by Bennett
(9), Burkhard and Wrist (21), and Mardon and his colleagues (67). This is
an extremely difficult task though the results of Burkhard and Wrist,
obtained by attaching a specially designed piezo-electric pressure transducer
in the body of a table roll on an experimental machine, are sufficiently
reliable and consistent to give the general pattern. Typical curves for
different machine speeds obtained by these investigators are illustrated in
Fig. 3. 1 and show the pressure or suction exerted on the table roll surface
as it passes under the wire.
187
3A.l 1 THE WIRE SECTION
The interpretation of these curves by the authors, based additionally on

observations and photographs of the relevant part of the wire section, is as
follows. The first small rise in pressure as the gap between the roll and wire
is closing is caused by water carried into the nip; this has a relatively small
F.P.M.
2''
\!l 10 laoo f.P.M.
::i:
Ill
k s
~
0
UJ F=. P.lr'\.
rt.
:J ~s
I/)
If!
w -lo
et ·IS.
Q..
TOP DeAO
CENTRE
POSITION ALONG TASiLE SURF~CE.
Fig. 3. 1. Pressure exerted on table roll surface as it passes under the wire at different
machine speeds (after Burkhard and Wrist)
effect (though in this work the table roll was doctored) but causes the mat
on the wire to lift very slightly. From the point where the wire meets the roll,
just before top dead centre, round to where contact is broken, there is a
region of high pressure; this corresponds with an area where the stock level
on the wire is observed to rise and then dip as the wire wraps round the
table roll on the trailing side. The length of this region clearly increases
with speed due to the greater wire wrap which occurs with the development
of higher suction forces once the wire has left the roll. The length was
188
THE WIRE SECTION 3A.l 2
shown to depend, as expected, on the tension in the wire, decreasing with
increased tension; also to a lesser extent the length decreased with increasing
drainage resistance of the mat caused by, for example, a drop in tempera-
ture or rise in consistency or substance of the mat. The magnitude of the
pressure also appeared, up to a point, to depend on the speed, though the
development of the double peak is not explained.
,_ 4·o
:c
0
?:
i
~3-o
~
a;
\!J
..J
5
ot
l·o
0 11 000 2,000 ~.ooo

$PEED IN F.P.M.
Fig. 3. 2. Drainage at three different table roll positions for varying machine speed
(after Wrist)
After the pressure region a small plateau occurs where pressure is

approximately zero as the web begins to leave the roll; the suction region
then commences with pressure rapidly falling to a minimum followed by a
slow subsidence. The peak magnitude of the suction depends closely on the
speed and was shown to increase in value at a rate almost proportional to
the square of the speed. As the length of the suction region also increases
slightly with speed, this and the lower suction peak have a combined effect
in increasing the total quantity of water discharged.
3A . 1 2 Discharge from a table roll

The relationship between quantity discharged and speed is of great impor-
tance in papermaking, but unfortunately the precise relation is not easy to
189
determine experimentally due to the difficulty of keeping other conditions

constant, particularly that of the drainage resistance of the mat above the
table roll. Fig. 3. 2 shows a graph given by Wrist (84) and illustrates how
for a groundwood/kraft furnish the discharge at three different table rolls
down the wire increases with speed, the relationship in this case being such
that the discharge is proportional to the 1·2th power of the speed, i.e. the
discharge increases at a faster rate than the speed. Unfortunately this result
cannot be generalized for Tellvik and Brauns (62) in similar work obtained
a power of O· 3 for kraft stock, which implies a substantially reduced
increase in discharge for proportional increase in speed, although they
also obtained the same power of 1·2 for a newsprint furnish.
Although the mat conditions on the wire complicate the issue, it is
evident that the quantity of water removed by a table roll will be dependent
in some way on the total suction applied in the nip, i.e. the integrated value
of the suction over the length of its application. Thus, as a first step in
any theoretical approach determination of both the magnitude and shape
of the suction curve are important and a great deal of effort has been
directed towards establishing these, notably by Cowan, Wrist, Meyer,
Bergstrom, and particularly Taylor. An extensive summary and analysis
of the work of these authors has been given by Marden et al. (67) and it is
not proposed here to consider the subject in any detail, though a few general
comments on the methods used will be made.
3A. 1 3 Taylor's theories of table roll discharge

There are two main theoretical derivations which are usually taken as a
starting point for checking against experimental results and these are
based on radically different hypotheses as to the nature of flow in the table
roll trailing nip. In the first of these it is assumed that no mixing or tur-
bulence occurs and the flow is taken to be completely streamline; in the
second case complete turbulent mixing is assumed to take place in such a
way that the velocity at any point in the nip depends only on the gap
between the wire and the roll. Both these theories were first conceived and
applied by Taylor and each gives a suction curve similar in the general
sense to that determined experimentally by Burkhard and Wrist. There are
however, two main differences between them: firstly, the no-mixing hypo-
thesis predicts a lower value for the maximum suction in the nip (though
in both theories the peak suction is shown to be proportional to the square
of the speed and therefore agrees reasonably with experimental observa-
tions); secondly the suction region extends further along the table roll in
the case of the turbulent mixing hypothesis. The net effect of these differ-
ences is that the turbulent mixing hypothesis predicts a discharge from a
table roll equivalent to 2· 5 times that predicted by the no-mixing hypothesis.
This may appear such a great difference that there should be no difficulty
in deciding which theory agrees closest with observations. It has already
been mentioned, however, that the quantity of water discharged by a table
roll depends not only on the suction applied to the web but also on the
resistance presented by the mat itself to the passage of water through it.
190
THE WIRE SECTION 3A.l 4
The main difficulty is, therefore, that there is no initial absolute comparison
to act as a standard and it is only possible, in the absence of data directly
relating to the drainage resistance of the mat under the particular conditions
pertaining, to compare the discharge of one roll with another. Attempts
have been made actually to determine the drainage resistance of a fibre
mat under different conditions and these will be considered in the next
section.
It should also be possible to decide between the two theories by direct
measurement of the peak magnitude of the suction in the nip but according
to Mardon evidence in this direction is, at the moment, contradictory.
Thus, although the turbulent mixing theory is most favoured in that it is
considered likely to express closer the actual hydrodynamic conditions
in the nip, the choice remains open. Several attempts have been made to
modify and compromise between the two basic approaches; Bergstrom
for instance has produced corrections covering the degree of wrap by
the wire round the roll, but these need not be considered here. It may be
noted, however, that Taylor has been able, with the aid of some funda-
mental experimental work, to show that the curtain formation which is
such a common feature of drainage at a table roll is due basically to the
instability arising from the separation of two surfaces carrying a fluid.
The object of this work is partly, of course, to account for the total
drainage in a wire section and also for the variation in the quantity of
backwater drained at each table roll according to its position along the
wire. Accurate predictions of the effect of altering such things as the length
of the wire table and the number and diameter of the rolls may then be
possible. This has become more imperative as the speed of machines has
been increased to the point where considerable difficulty may be experienced
in giving the sheet reasonable time to form while removing the water in a
practicable length of wire at the same time. For this reason considerable
attention has been concentrated on examining changes in the drainage
properties of the mat as it progresses down the wire and in the discharge
from the individual rolls. This topic will now be considered in some
detail.
3A. 1 4 Variation in drainage resistance of the mat down the wire

One approach to the problem of determining how the resistance to drainage
of a mat varies with its changing condition down the wire is to attempt to
measure it in a simulation experiment. This has the advantage that the
effects of different properties of the fibres and water comprising the web,
the degree of beating, temperature, addition of fines and loading, etc.,
can be investigated at the same time; this in turn may help to establish
some fundamental conceptions of what determines drainage resistance
and explain why with different stocks on the same machine variations
occur between the quantity drained at one part of the table roll section
compared to another. Laboratory simulation can also indicate how
drainage resistance depends for different pulps on the weight of the mat
already formed on the wire and on the amount of suction applied (the
191
3A.1 5 THE WIRE SECTION
latter is particularly important in the application of drainage theories for

it enables the rate of movement of the water under a given suction to be
calculated).
In this field Hendry et al., Ivarrson and Johannson, Ingmansson,
Higgins and de Yong, and, for simulation of higher speeds, Meadley and
Anneus have done useful work; for details reference can again conveniently
be made to the survey by Marden et al. (67) in which the results of some
of their own work are also presented. More recently Wahlstrom and
O'Blenes (83) have reported results obtained with a specially modified
apparatus designed specifically for simulating drainage conditions close
to those appertaining on the wire; they found that the rate of mat formation
and drainage is very dependent on pulp type and the degree of beating
while neither the concept of specific drainage resistance, which has proved
useful for low consistency work, nor the freeness test is related to the drain-
age rate figure obtained in their work. Boadway and Gray (85) have also
reported work in which the drainage resistance of the fibre mat both
vertically and horizontally, and of the wire itself, have been estimated
separately; they suggest that the structure of the wire has an important
influence on the drainage resistance of the mat and does not act simply as
an extra resistance to drainage in addition to that of the mat itself.
It is evident from this that the factors affecting the flow through a
compacting mat of fibres on a wire mesh are extremely complex and this
is no doubt the reason why up to the present time little useful progress has
been made in predicting theoretically the variation down the wire of the
quantity of water extracted at the table rolls. Some typical curves of this
variation are shown in Fig. 3. 3 and were obtained for different speeds by
Tellvik and Brauns (62); in addition curves giving the change in consistency
of the table roll discharge are shown. These results were obtained on an
experimental machine and no changes were made in the pulp and making
conditions other than that of speed.
3A. 1 5 Predicting drainage conditions on a machine

Using data on the variation of discharge from individual table rolls on a
wire and making assumptions about the manner in which the average
drainage resistance varies down the wire, both with the suction applied
and the thickness of the formed mat, it is possible to determine which of
the two basic theories (mixing or non-mixing) mentioned above best fits
observations taken under different conditions; Mardon and his colleagues
used this approach amongst others but found fundamental differences
between the results for newsprint and kraft stock and once again obtained
no clear-cut indication as to whether either theory is likely in general to be
correct.
Tellvik and Brauns, on the other hand, used their data as a starting point
to determine what influences the drainage resistance through the settled
mat of fibres formed on the wire, and, by assuming this resistance to be
dependent only on the mat properties and independent of speed, deter-
mined how the flow from the table rolls would vary with speed if other
192
THE WIRE SECTION 3A.1 5
looo
t<.RAf:T PULP
E!>Rl!AST e,ox CoNSl~TENCY O·S %
Sub&TANci. <Do 9.s.M.
0
Sc.HoPPER REtGLER 4o &.R..
~00
i" M~c~INs SPe.ao
0
J
---.::::::::::::::::=:::::::::~<~75
410
M!tlRES/M1"4
It
u. ~,!,oa
o ........~-L-__...___._~...___.~_._~......_~~2_0_0~
S 7 ~ 11 I~ IS
TA.e,\,.E; ROl..I.. NUMe;,E~
\
~ \
)- O·O~ \
\ \
u \ \
'Z
,_w ,,~
' ' MAc.Hu'4E. 5Pe:e.o
Ill o·o4 <~~~4lo MlT~~s/M1N.
~.....-315 "
"'3oo 11
"-200 ,,
3 S. 7 9 II I?> IS
f>1usT ~"'-" TAE?>LE. Ro1..1... NuM.ei.i.R.
A FORMING
f>OAR.0
Fig. 3. 3. Flow and consistency of discharge from table rolls at different positions
down a wire and at different speeds (after Tellvik and Brauns)
193
factors remained constant. This proved radically different for two different
types of stock due probably, it is suggested, to a difference in compression
of the settled fibre mat on the wire; in fact, as mentioned earlier, the in-
fluence of speed for a kraft as opposed to a newsprint furnish was small
and the magnitude of the drainage for kraft was determined more by the
weight of the fibre mat and slowness of the furnish. Although these authors
claim that on one particular machine operated with a constant type of
stock it is possible to take certain measurements under operating conditions
which may be extrapolated to predict the effect of a change, insufficient is
yet known about the detailed mechanism of drainage for any generalizations
to be possible.
3A. 1 6 Bennett's observations

It is appropriate at this point to mention in more detail some results
reported by Bennett (9) in which the quantity and consistency of discharge
from the table rolls of a newsprint machine were, together with much
other data, measured on several occasions. The flow from the table rolls
declined down the wire in exactly the same manner as Tellvik and Brauns
have shown in Fig. 3. 3; also the consistency of the discharge was at a
maximum at the second table roll (though in this case the breast roll and
first table roll discharge were not sampled separately). The decrease in
consistency of the discharge (which incidentally applied mainly to fibre-
the ash content did not alter appreciably down the wire) is readily explicable
in terms of the filtering effect which occurs as the mat builds up in thickness;
the increase in consistency occurring in the early part of the wire may be
attributed to a greater reduction in the quantity of water as opposed to
fibre discharged, although the behaviour of the table roll flow in this
region before the mat has fully formed must be highly dependent on the
forming conditions.
Bennett and his colleagues also analysed statistically the effect of varying
several conditions on different regions of the table roll part of the wire; the
variations were such as occurred during normal production and so in all
cases the range is comparatively small. Increasing speed over the range
610 to 690 f.p.m. appeared to increase the flow proportionately greater in
the early positions (i.e. the speed had less effect down the wire), and the
net effect was that the consistency of the mat at the first suction box was
relatively unchanged. An exception to this was the combined breast and
first table roll discharge, which was apparently unaltered by the speed;
this may be due to the fact that at this part of the wire substantial gravity
drainage occurred and the proportionate reduction of this with increasing
speed offset the increased drainage produced by the rolls. But on the other
hand the suggestion has been made that the slice jet may have impinged
further down the wire at the higher speed and reduced the effect of the
breast roll discharge. Again it can only be said that conditions in this part
of the wire are highly critical and generalizations are probably very risky;
other aspects of this topic will receive consideration in later sections.
Other results obtained by Bennett indicate that an increase in temperature
194
THE WIRE SECTION 3A.21
of the stock, while generally increasing the flow from the table rolls, causes
a greater proportional increase further down the table. This has been
considered reasonable on the grounds that flow against capillary forces
(such as occurs through the thick mat more towards the end of the table
rolls) is related to surface tension, and this is affected by temperature
more than the purely hydraulic flow which predominates in the first few
feet of drainage area on the wire. The freeness, drainage time, and consis-
tency of the breast box stock did not appear to have any significant effect
on one part of the table roll section as opposed to another though there
were indications that an increase in consistency retarded drainage more
in the earlier part of the wire.
3A.2 FORMING THE SHEET ON THE WIRE

The term 'formation' is used in a variety of senses in the paper industry.
In a general context it may be applied to describe the overall quality of the
sheet as gauged by looking through it; in this respect it can amongst other
things cover freedom from pinholes, blotches, light patches produced by
splashes and bubbles on the wire, streaks, worm marks, and even absence
of dirt and slime particles. In fact, in this most general sense any aspect of
the sheet originating in the wire section, at least up to the point where the
sheet is fairly well compacted, could be included under the heading of
'formation'; the term then has a different emphasis of meaning according
to the common faults encountered in the particular mill where it is
used.
At the other end of the scale, 'formation' is used to describe the homo-
geneity of appearance when viewing the sheet in transmitted light. This
restricted sense is nowadays in more common use and will be adopted here.
It thus covers the terms 'wildness', 'cloudiness', etc., applied to describe
the presence, size and frequency of clumps of fibres or areas of thicker
concentration in the sheet; as such the formation is essentially dependent
on the manner in which the stock is lead through the siice, deposited on the
wire, and then drained, without reference to other defects which may
exist.
In what follows an attempt will be made to describe recent thoughts
and theory on how the sheet is deposited and formed on the wire and how
this influences the formation. It is convenient at the same time to include
two other aspects of this complex subject, two-sidedness and fibre orienta-
tion, since both of these properties are established at the same time as the
formation and governed by the same general conditions. The influence of
various factors such as the shake, wire mesh, relation between velocity of
the slice jet and wire speed, and dandy on these properties of the sheet will
be given individual attention in the appropriate place later.
3A. 2 1 The nature of formation

Except for certain specialized papers it is generally agreed that a sheet of
paper should be as homogeneous as possible. This implies a completely
195
3A.21 THE WIRE SECTION
even distribution of long and fine fibres and loading throughout the volume
of the sheet and a paper approximating to this ideal will have an excellent
formation. In practice, formation is assessed subjectively by the machine-
man from observation of the sheet held up to a light to compare the light
and dark patches; the information gained from this examination is
supplemented by following the sheet down the wire by eye to assess the
degree and manner of small scale movement occurring as the fibres settle
into a mat (at least on machines running at most 1,500 ft./min.). Altering
the formation on the basis of this information may justly be termed,
perhaps above all else, the art of papermaking, and the steps taken on
any particular machine to remedy or improve some aspect of formation
often seem to be completely individual to that machine and to depend on a
wealth of accumulated experience in running the particular paper and wire
section.
In an attempt to achieve some degree of objectivity many instruments
have been devised for the purpose of assessing formation in some measur-
able terms. An excellent summary of the investigations made in this field
has been given by Robertson and Mason (81), to which reference may be
made for further details. Most of the instruments are, of course, essentially
laboratory devices for examining and analysing variations in the light
transmitted from varying sizes of slit through a single sheet of paper.
An essential point of the design of these instruments is to check how results
derived from the light measurement correlate with subjective assessment
of the ranking of a variety of papers having different formation; in some
cases elaborate attempts have even been made to take account of the
behaviour of the human eye. Some models have now been placed on a
commercial basis and may be purchased.
There have also been some reports, notably by Robinson (23), Maclaurin
(58) and Eastwood (87), of the application of a formation measuring
device actually on the machine at the dry-end, and attempts have even been
made to use the readings obtained to regulate refiner settings. The merits
of this latter step are debatable since refining is only one of the factors
which have a bearing on the formation of a sheet, but nevertheless this
does represent a worthy move to place control of the property on a more
systematic basis.
Apart from optical methods of measuring formation use has been made
of alpha, beta, and gamma radiation from radioactive sources to measure
very small-scale substance variations in a sheet of paper, notably by
Brazington and Radvan (43), and Attwood and Parker (71). This is
doubtless a valid approach, since optical and substance variations must be
closely correlated, and in fact this could eventually prove to be a more
suitable and useful means of assessing formation.
At this point it is worth emphasizing that the line dividing formation
problems from those of relatively small-scale substance variation is not
·· clear-cut. Generally speaking, it is convenient to think of relatively small
substance variations as one of the causes of such faults as cockling: thus,
the areas of lighter substance dry faster than the heavier and shrink,
causing the damper and more plastic parts of the web to become stretched
196
and dried under tension; the lighter and less strained areas of the sheet will
then (as discussed in greater detail when dealing with drying) have a
greater stretch under a given tension and a greater moisture expansivity,
particularly if the paper is liable to shrinkage, and in reaching equilibrium
with the atmosphere by relaxation and re-gain of moisture cockling may
occur. Formation differences, on the other hand, may be considered to be
confined to areas so small that differences in drying do not have a significant
effect on structural properties of the sheet. Nevertheless in extreme cases
the difference in thickness over a small area between different parts of the
sheet can become noticeable not only as poor formation in the look-through
of the sheet but also in the surface characteristics, since the thicker portions
become more glazed in the calenders.
3A . 2 2 Factors contributing to the formation

In the strict definition of the word, good formation depends on achieving a
regularly ordered deposition of fibres on the wire. It is usually considered
that, in a general sense, this is best achieved by ensuring two things. Firstly,
that the discharge from the slice is as homogenous in distribution of fibres
as possible and in this context this implies, above all, no flocculation.
Secondly, that there is sufficient small-scale agitation of the stock when it
is on the wire to preserve this state until the fibres are set in the mat.
A third important aspect is that the impact of the jet and wire should not
be disruptive in character though a certain degree of turbulence in that
area may help to complete the even distribution of fibres; as this is largely
a negative effect it is more appropriately dealt with later in 3B .1.
While this theory is no doubt substantially correct (though not of much
direct value in overcoming poor formation in any particular circumstances)
it is important to remember that on all but very slow machines the sheet is
formed in the same way that the drainage is accomplished, in a series of
diminishing surges. The quantity of fibres fixed at each roll into the existing
fibre mat diminishes down the wire and at each successive roll the layer
deposited is thinner. Each layer is deposited in a predominantly horizontal
direction, and drainage is essentially one of gradual filtration through a
mat of fibres already thickened, except possibly in the later stages beyond
the table rolls (see, for example, references 116 and 117). It is now proposed
to examine this whole subject in detail with particular emphasis on some
recent work in this field which has contributed to understanding the
processes affecting formation.
With slow machines it is considered that screens of various design
perform the important function of deflocculating the stock, a task which
is especially vital with longer-fibred rag finishes, while the shake on the
wire ensures that the fibres are kept adequately distributed until they settle
into the mat. Whether or not the screens do other than keep tangled
strings of fibre out of the stock flowing to the breast box need not be
considered here; a great deal must depend on the nature of the flow
system and the furnish and, in particular, on the rate at which the fibres
reflocculate in the conditions pertaining. Regarding the shake there is
197
strong evidence to show that formation can benefit in comparison with an

unshaken condition; however, on the subject of shake more than any other
it is extremely difficult to generalize and a detailed consideration of this
topic is left till 3B. 3.
One observation which is of interest is reported by Judt (48). If drainage
of the pulp is delayed, by removing one or more table rolls, then on a fairly
slow machine without shake the formation deteriorates. Other workers
have noticed a similar effect and it has also been stated that the increased
cloudiness is predominantly on the top side of the sheet. This is confirmed
by the work of Luhde (112) who has demonstrated that floes are more
prominent on the top side of a sheet, though the effect diminishes at higher
speeds. The usual explanation of this is that as the sheet progresses down
the wire the stock still in suspension has more time to flocculate and is less
subject to the shear forces induced in the wire side by the action of table
rolls and other dewatering devices. In addition, on machines with shake
the influence of this is also gradually diminishing.
Recently the manner in which the stock first contacts the wire, particu-
larly with regard to the quantity drained by the breast roll, has been
investigated by Manson (90) using a formation tester. Quite small altera-
tions in the position, horizontally and vertically, of the straight slices on an
old machine running at 130 feet per minute were shown to have a sub-
stantial effect on the formation, a confirmation of the importance of the
stock condition as the mat first sets on the wire.
3A. 2 3 Formation on faster machines

Similar but more comprehensive work has been done in this direction on
faster experimental machines and there are indications that other aspects
become more important at higher speeds. For instance, Parker (68) has
reported that the substitution of low-vacuum suction boxes for table rolls
in the wire section considerably worsens formation. This is despite the
fact that under these conditions the flow on the wire is practically free from
the normal disturbances which plague high-speed operation, in fact almost
glassy in appearance, and small-scale substance fluctuations are con-
siderably reduced. In this case there is no question of delaying drainage as
in the work reported by Judt (though earlier drainage may well have been
slower than normally occurs) so the explanation for the deterioration in
formation cannot lie entirely in the increased time in which the sheet is
allowed to form on the wire. Though Parker agrees with the theory that the
formation is dependent on how the rate of drainage compares with the rate
at which flocculation occurs, he considers that the production of controlled
turbulence before the slice (usually with a perforated roll) to keep the fibres
adequately distributed is less important than ensuring there is sufficient
small-scale turbulence in the immediate drainage zone where the sheet is
formed. In particular, he considers on the basis of the experiment men-
tioned above that table rolls induce such a turbulence to the stock on the
wire and therefore are important for obtaining good formation as well as
draining the sheet.
198
Wrist (84) is in general agreement with the contention that inducing a
deftocculated condition in the slice jet is of less importance than has been
hitherto considered. His work on perforated rolls (though other workers
disagree with this, see lB. 5 3) has indicated that they do not effectively
deftocculate fibres in suspension; any micro-turbulence induced by the
roll decays so rapidly that reflocculation occurs almost immediately,
certainly by the time the jet contacts the wire. The production of some
small-scale turbulence on the wire, especially in the immediate drainage
zone where the sheet is largely formed, is therefore of great importance
though it must be kept to a suitable level; this turbulence, by inducing
relative motion between the undrained stock and the deposited fibres,
produces a combing and shearing action which inhibits flocculation. The
main problem is, however, that the same turbulence if excessive will cause
the overall fibre distribution to deteriorate (as happens, for example, if
the shake is too hard) and it is difficult to obtain the beneficial flocculation-
preventing turbulence without at the same time increasing the substance
variations.
In the discussion following this last paper further evidence of the
importance of table roll action in helping formation is presented. In the
same work by Parker already referred to, the experiment of reintroducing
some table rolls ahead of the suction boxes proper was tried; an improve-
ment in formation over that obtained with no table rolls was observed.
In addition table rolls appeared to improve formation compared to grooved
rolls which are known to disturb the stock on the wire less as well as
extracting less water from the sheet. The general conclusion is that some
agitation to the stock on the wire is necessary during the forming period
if adequate deftocculation, and consequently good formation, is to occur,
and this is conveniently provided by normal table rolls.
Also in the discussion of Wrist's paper it is pointed out that the differing
drainage conditions in the various experiments involving removal of table
rolls would produce different retention conditions which in turn would
affect the consistency of the breast box stock once equilibrium conditions
for the normal substance were achieved. The suggestion was made that
these variations in consistency may have been at least partly responsible
for the differences observed in formation. This is a valid criticism since
consistency of the breast box stock has an important effect on formation.
At the moment then there is some possibility that the observations des-
cribed are capable of a different explanation.
Before turning to consider in more detail the effect of breast box con-
sistency on the formation it is worth drawing attention to some other
evidence of the agitation induced in the fibre mat by table rolls. Several
workers have measured the consistency of the discharge from various
drainage devices under the wire and in all cases it has been observed that
if a table roll follows a foil or grooved or dandy-type roll then the consis-
tency of discharge rises at the table roll. Measurements quoted by Hendry
et al. (3) for a tissue machine generally show this increase, and Burkhard
and Wrist (21) give results for an experimental machine in which after
passing over 13 foils the consistency of discharge rose at the first table roll
199
and continued to rise up to the third before declining. These observations

suggest that solid table rolls have a substantial loosening effect on the under-
side of the fibre mat and as this action would generally aid deflocculation
it provides support to the theory that their operation is advantageous to the
formation.
3A. 2 4 Effect of breast box consistency on formation
It is well-known by machinemen that in running the wire section a com-
promise is always necessary between the rate of drainage and the formation.
With any particular furnish the quantity of backwater in circulation is set
to give a workable position of the dry-line; in other words as much back-
water is used to dilute the fresh stuff as can be comfortably drained away,
leaving the moisture in the sheet at the couch satisfactory for transfer from
the wire. The lower the volume of backwater in circulation, the higher the
consistency of the stock in the breast box; so if the stock on the wire works
wetter, i.e. drains slower, reduction must be made in the backwater volume
resulting in a higher consistency in the breast box.
A higher consistency of stock in the breast box produces more floccula-
tion and an increased network strength of the fibres deposited on the wire,
which results in a deterioration in formation; Wrist and Norman (115)
have confirmed this by demonstrating that the random variation in sub-
stance of paper over small areas is greater at higher consistencies, while
Schroder and Svensson (106) have presented a comparison of formation at
two different consistencies which also shows deterioration at the higher con-
sistency and in addition indicates a lower general strength and stretch.
At really high consistencies it becomes impossible to make a sheet that is
not like a snowstorm in appearance. Thus, if stock becomes wetter the
machineman has a choice between risking an excessive number of breaks
at the couch due to higher moisture, or shutting off backwater and allowing
the formation to deteriorate. The usual remedy is to reduce treatment of
the furnish so far as possible to make the stock run more free but this,
if taken to excess, inevitably affects the quality of the paper from the point
of view of strength, bulk, surface characteristics, and freedom from larger
shives.
With the gradual speeding up of machines the point is frequently reached
where the difficulties in draining the stock adequately to achieve the
required quality becomes a real problem. Increasing the vacuum on the
suction boxes or the pressure at the couch inevitably produces a reduced
wire life and can only be taken so far without becoming economically
ruinous. Alternatively, the drainage rate can be improved by such expedients
as increasing the number of table rolls, using deflectors touching the wire
between rolls, and increasing the flow of water removed by the breast roll;
the danger in some cases, particularly with the latter expedient, is that if
the drainage in the early part of the wire is too rapid flow instabilities
become less controllable and wire mark more prominent, resulting in a
generally poorer sheet and making the situation no better. This is the case
especially with fast draining newsprint and tissue stocks and will be con-
sidered in more detail later.
200
3A.2 5 Two-sidedness; Hansen's experimental work on loadings

All papers exhibit two-sidedness to a greater or lesser degree and in most
cases it is an undesirable characteristic. The difference between the two
surfaces is present even with glassine and persists after supercalendering,
as an excellent series of light micrographs obtained by Emerton and his
colleagues at the British Paper and Board Research Association shows (72).
Wire and felt marks contribute to two-sidedness but the basic cause, and
the one which is considered here, is due to variations in structure and com-
position through the thickness of the sheet which manifest themselves by
producing different surface properties on the two sides. These variations
can also affect the relative moisture expansivity through the sheet causing
one side to expand or contract more in the presence of a change in moisture
equilibrium than the other. This phenomenon is commonly exhibited after
drying in the form of curl, but as this particular subject is more closely
associated with differences in fibre orientation, consideration of it is left
till a later section, 3A. 2 11.
Several techniques have been used to investigate the variation in com-
position within a sheet of paper, the most common involving stripping off
successive layers of fibre with adhesive tape, grinding or scratching off
layers with fine carborundum paper or a razor blade, or using a microtome
to cut very thin sections. These methods have enabled several workers to
show that differences in average fibre length, loading, the proportion of
one type of fibre to another, and even the frequency of air bubbles (16)
can exist through the thickness of a sheet of paper. The average fibre
orientation can also vary (this will be dealt with in 3A. 2 10), as can the
frequency of floes or fibre clumps through the sheet (as described in 3A. 2 2).
With the aid of a light microscope it is possible to examine the structure
in microtomed cross-sections of paper with considerable detail, and studies
with this technique demonstrate the manner in which a sheet is built up
from successive layers of fibres (116).
The first experimental results of some value in elucidating the basic
causes of two-sidedness were obtained by Hansen (1) and are worth
considering in some detail. In this case different amounts were ground off
the surface of sheets of paper and ashed, enabling the distribution of china
clay through the thickness of the paper to be determined. Distribution
curves for paper made under different conditions on an experimental
machine, but with the same basic furnish, were obtained in this manner;
for example, Fig. 3. 4 shows how the shape of the curve alters as the
substance is increased (and the machine speed reduced approximately
pro rata). In a different experiment it appeared that adding Sveen glue or
·altering the speed, though affecting the overall retention as discussed in
describing the factors affecting the wet-end flow system, did not signifi-
cantly alter the shape of the distribution curve.
The curves in Fig. 3. 4 indicate that change in substance does not ap-
preciably affect the percentage of filler at the two surfaces, only the inner
part of the sheet changes and evens out to a steady value as the substance
increases. In some cases the percentage of filler on the top side is consider-
201
ably greater than in the breast box stock though this increase towards the
top side appears to be caused mainly by the dandy, a subject which will be
covered later.
Hansen also ran the machine with doctors on the table rolls to reduce
the volume of water carried into the nips; this had no effect at all on the
distribution curve. In another experiment the machine was stopped with
Sue.STANCE
~ ,3.o '° 0 9.~.ni. 3o 80. 9.s;.ni. 30 ......,__ _ _10_0_,..;9_.t>_._'"-1·
~
l
ls.I
~(} '2.o Zo
u
\l)
~ I o 1-----"t--1
0
_J
g
TOP WIRE. T W T
~~o
t-
z
._
II.I
d '2.oi--------T--~ Zo...__ _ _ _ _ _ __,.__~
"l0 101----------'t-+ \01-----------.+I

<!
0
J
0 .___ _ _ _ _ _ _..... 0 L------------'
T W T w.
Fig. 3. 4. Distribution of loading through thickness of paper made at different
substances. Dotted line indicates average content (after Hansen)
the stock on the wire and then left to allow the water to drain away; the
same distribution curve appeared even in samples taken just after the
slice before the mat had passed over a table roll. Other experiments showed
no change occurring in the shape of the curve at the suction boxes or presses
and as a result of all this evidence Hansen reached the conclusion that the
loading distribution is fixed when the stock has been run on the wire. The
low loading content on the wire side may in fact be attributed basically to
the natural process of drainage, particles of filler having less opportunity
to drain out with the water through the fibre mesh as the mat grows.
202
3A . 2 6 Other experimental work on loading distribution
This explanation has been contested by Underhay (15) who tried various
methods oflaboratory simulation of drainage to obtain a difference between
the top and wire side loading content similar to that commonly observed
in machine-made paper. In this work the main method was to split the
sheet into four thicknesses and make a straight comparison of the ash
percentages in the upper and lower portions. It was found that the only
way in which the composition and appearance of the sheet could be made
similar to machine-made papers was by applying a vertical oscillating
movement to the drainage machine wire. On the basis of these observations,
together with evidence that the difference in loading content between the
two sides is not present in paper made on very slow machines, Underhay
considered that only a 'washing' and disturbing action to the underside
of the sheet at the table rolls could account for the loss of loading and
fines which occurs there.
A similar approach, this time involving simply splitting of the sheet
and ashing the two portions, was used by Pritchard (29) on relatively thick
and heavy machine-made papers. This yielded some rather peculiar results
in that there appeared to be more loading on the top than the wire side
for 135 and 270 g.s.m. papers but less for 220 and 250 g.s.m. papers.
Also differences were found in the relative loading content of the two sides
from the couch down through the presses, and there were indications that
more loading is removed from the side in contact with the felt during pres-
sing. The order of accuracy and repeatability of the results quoted by this
author appear high despite the fact that they are based on quite small
differences in ash content, and with the aid of laboratory work on hand-
sheets some attempt is made to explain the various phenomena which
were found. For example, it was demonstrated that the relation of top to
wire side loading content depended on the interaction of the degree of
beating of the furnish and the substance and this is explained in terms of
the relative ease with which loading settles through the fibres during the
drainage process. Though highly interesting it is unfo.rtunate that most
results obtained in this work only permit the loading content of the two
sides to be contrasted and little information is available as to the relative
distribution through the sheet. It does appear, however, that for fine,
heavy papers different mechanisms may come into play and this should be
borne in mind in what follows.
3A. 2 7 Groen's observations

The results obtained by Hansen, Underhay, and Pritchard which have
now been outlined are in many cases contradictory and this largely re-
mained the situation until Groen reported some further work (75). This
involved stripping off thin layers of fibres using adhesive tape and with this
technique the loading distribution of numerous papers, both machine-
made and hand-sheets, was determined. In all cases the general characteris-
tics of the distribution curve for Fourdrinier papers follow those found by
203
Hansen; in some there is a gradual increase in filler content from the wire
to the top side of the sheet but more commonly when no dandy is used
the percentage increases from the wire side until a plateau of steady value
is reached, while close to the top side the percentage may show a small
decrease. The shape of the distribution curve does not appear dependent
on the fibre composition (for example, wood free as opposed to up to 70
per cent. groundwood) nor on the type of filler (china clay, talc, or coating
clay) except, possibly, in one case, annaline; increasing the total percentage
of filler appears to affect the top side layers rather more, while increasing
machine speed causes the plateau region of relatively steady filler content
to narrow, and the reduction in content as the wire side is approached
extends over a greater thickness of the sheet.
These results, together with those from other experiments, led Groen to
confirm Hansen's original findings and to conclude that the basic shape of
the filler distribution curve is explicable in terms of natural self-filtration
through a growing fibre mat. Further, in one direct comparison the substi-
tution of several open-type table rolls for solid rolls had only a slight
effect on the distribution curve and this is taken as confirmation that the
'washing-out' effect, which should be greatly reduced in open rolls due to
the lower quantity of water adhering to the surface, is not of great signifi-
cance. Wrist (84) concurs with this view and has stated that the substitution
of foils for table rolls, though affecting the overall filler retention, does not
alter the distribution within the sheet. In sum then the weight of evidence
favours the natural drainage theory though it is only fair to point out that
the 'washing' effect of the table rolls still has much support as an explana-
tion and considering the other disturbances which are known to occur in
the roll nip it could well be relevant at least in high speed operation.
3A . 2 8 Two-sidedness in fibre distribution

All the work referred to so far has concerned loading materials though it
has always been considered that differences in smoothness between the
two sides of a sheet are due equally to loss of fines from the wire side.
Close observation of the two surfaces, as in the light micrographs of
Emerton, shows that on the wire side longer fibres are exposed with more
gaps and concavities prominent between them, particularly when ground-
wood and fine fibre fragments are present in quantity in the fresh stuff.
Further evidence of the increased fines content of the top side of a sheet is
that it has been found to be composed of fibres having a slower freeness
(112).
This difference in fibre composition on the two sides is also considered
to be one of the causes of colour difference which can be so annoying in
fine papers. Hinton (76) has pointed out that surface reflectance differences
originating from the differences in smoothness and fibre orientation between
the two sides are partly responsible for this, but the main trouble is essen-
tially that various sizes and shapes of fibres and particles of loading are
liable to be dyed to a different intensity. Pigment dyestuffs should obviously
be avoided to reduce two-sidedness in colour because they behave in a
204
similar fashion to loadings, but little can be done to alter the fact that
soluble dyestuffs have different affinities to loadings and fibres; the intensity
of colouring is generally greater for fines than long fibres though it varies
with the origin of fines (whether resulting from beating or naturally
occurring) and depends on many other factors. If, as is usually the case,
the difference in intensity between the dyeing of long fibres and fines
presents the more important contrast, then the top side of the sheet will be
enriched in colour due to the greater quantity of fines present there; if,
however, the dye has a much stronger affinity to fibres than to the filler,
this may become the more important effect and it is then possible for the
wire side to be enriched due to its lower loading content.
There has not been much work done on how fibre composition varies
through the sheet though several investigators have tried adding dyed
fines to furnishes and examining the relative quantities appearing in the
two sides of the sheet. There has been a general confirmation that fines
appear in greater quantities on the top side of paper made on both slow
and fast machines. Forgacs and Atack (73) have investigated a special
though highly relevant aspect of this subject by determining the proportion
of chemical pulp to groundwood present through the thickness of sheets
of newsprint; a microtome was used to obtain the sections and the propor-
tion of each pulp was determined by a technique dependent on the differ-
ences in lignin content remaining in the fibres after their respective treat-
ments. Several sheets of newsprint made at different machine speeds and
with a relatively wide range of average chemical pulp percentage were
tested; in all cases the proportion of mechanical pulp was much lower than
the average on the extreme wire side, though within the sheet it increased
rapidly to reach a maximum and thereafter decreased slowly towards the
top side.
This is similar to the distribution of loading through the thickness of a
sheet of paper and both may be explained in general terms as follows. At
the start of drainage the longer fibres are held in the meshes of the wire
while the fines tend to pass through. As the mat builds up the longer fibres
act as a sieve of decreasing mesh size, so retaining more fines, and at the
same time the lower layers become more compacted. In the upper layers
some movement of fines downwards between the longer fibres may occur
(thus accounting for the reduction in fines and loading on the top side)
unti] the web becomes so compacted that relatively little movement between
the fibres is possible any longer.
3A . 2 9 Fibre orientation; general observations

In the vast majority of papers it would be desirable for the various strength
and other physical properties to be approximately isotropic, i.e. the same
in all directions. In practice strength and rigidity tests taken in the machine
and cross directions can differ appreciably and the basic cause of this is a
preferential alignment of fibres in the machine direction (see, for example,
reference 91). Differences between the machine- and cross-direction stresses
205
in the sheet, particularly during drying, also introduce anisotropy but this
subject will not be considered at this point.
Several methods of assessing fibre orientation based on the highlighting
and examination of a small proportion of fibres in the sheet have been
used by different investigators, but without exception they are rather
tedious to use. Usually in laboratory work a small proportion of fibres are
dyed and a silurian effect produced in the finished sheet; for work actually
on the paper machine the fibres may be dyed with a fluorescent dye which
is visible in ultra-violet light or in certain cases fibres may be mordanted
with tannic acid and made visible in the finished sheet by suitable dyeing.
Various devices have been designed to aid the process of determining and
recording the direction in which individual fibres lie, but in all cases a
great many fibres need to be counted to achieve any reasonable accuracy
and there are many difficulties, particularly with longer fibres that may
twist in several directions. Recent reported methods have generally been
adapted from the technique developed by Danielson and Steenberg (35, 45),
although Forgacs and Strelis have devised a relatively simpler method
based on counting the intersection of fibres with two lines in the machine
and cross directions (88). Attempts have been made to instrument this
process of counting by using transmitted light through suitably placed slits,
but little success has been achieved. A more promising line has been the
application of X-ray diffraction techniques to show up any preferential
alignment and reasonable correlation with anisotropy of strength measure-
ments has been achieved in this way (33, 41).
In earlier discussions of the flow of individual fibres in the slice jet,
evidence has been cited to show that there is a preferential alignment in the
machine direction. In the first layers deposited on the wire this preferential
orientation may be accentuated as differences in velocity between the jet
and the wire produce a drag on the fibres; layers deposited later when the
relative motion of stock and wire is practically zero may be expected to
retain only the original orientation effect and consequently fibre orientation
is generally greater on the wire than the top side, see Fig. 3 . 5. Shake also
has an effect on orientation of the fibres but discussion of this will be
left till later.
3A . 2 10 Changing fibre orientation on the wire

The manner in which the initial orientation of fibres in the jet stream is
affected before they are fixed in the sheet depends on the presence of a
relative motion between undrained stock and the existing mat, and can
be explained as follows. If the fibre were lying horizontally in the stock on
the wire and at one particular table roll settled completely into the mat
during drainage, then even if there were a substantial difference in velocity
between the stock carrying the fibre and the mat fixed on the wire, there
is no reason why the orientation of the deposited fibre should be affected;
in this case the initial orientation of the fibre in the slice jet is unaltered
and the whole fibre is, as it were, fixed simultaneously into the sheet.
But in a jet in which the fibres lie at random in all directions, most will
206
have some vertical component of orientation; subjection to the short,
impulsive downward forces produced by suction at the table rolls may then
fix one end of the fibre leaving the other relatively free to move in the
MACHINE: Duacnori
ToP SlOE
Fig. 3. 5. Fibre orientation distribution on top and wire sides of paper (after Glynn,
Jones and Gallay)
liquid stock. In this situation any relative velocity between the free end of
the fibre and the mat will affect the orientation to an extent dependent on
the time lapsing before the next roll produces another downward impulse
and fixes more or all of the fibre into the mat. In particular, a slice jet
meeting the wire at a lower relative speed will accentuate the small existing
machine-direction alignment of fibres especially in the slow, narrow bound-
ary layer of the jet which first contacts the wire.
This theory was first developed by Finger and Majewski (11) to explain
some observations on the effect of shake. Wrist (84) has since pointed out
207
that the conception of one end of a fibre being fixed and the other free
to move is too simple. In fact the boundary between a fixed and a free
portion of a part-vertically aligned fibre must be diffuse, and it is more
realistic to consider that one end of the fibre extends down into a zone of
higher consistency and stronger fibre-network structure which inhibits
movement to a greater extent than in the lower consistency of the upper
regions. Change in orientation due to relative motion of wire and stock
is thus still possible but the effect must become less pronounced further
down the wire as the overall consistency of the mat increases (which is
why attempts to apply shake at the dry-end of the wire have met with little
success). This, combined with the progressive decay in relative motion
between fibre in the liquid stock and the mat beneath as the sheet builds up,
explains the reduced fibre orientation in the top side of paper.
It is apparent from this explanation of fibre orientation that natural
forming conditions on the wire and the initial orientation of fibres in the
slice stock both contribute to alignment in the machine direction; only
the influence of shake produces forces which can counteract this tendency,
although it is also possible that the small-scale turbulence induced in the
stock as the wire passes over a table roll has a randomizing effect which
reduces alignment. For these reasons paper made on the Fourdrinier
machine inevitably has some degree of machine-direction orientation
producing anisotropy; as in the case of two-sidedness, although some
alleviation of the differences may be achieved by careful experimentation
in the light of the foregoing explanations of the phenomena, it is not
realistic to expect that the differences can be eliminated altogether.
3A.2 11 Curl
Because of the differential shrinkage of fibres during drying along and
perpendicular to their length, two-sidedness and fibre orientation effects
interact to produce the condition known as curl, a phenomenon which is
far from easy to explain. This subject will be dealt with more fully in
Part 5 and for the moment attention is confined to discussing the influence
of formation conditions on curl.
The difference in degree of orientation between top and wire sides, as
shown for example in Fig. 3. 5, is the basic cause of curl in paper, though
the issue is complicated because the average potential shrinkage of the two
sides is different (due to differences in composition of the fibres as explained
in the previous section) and because of the effects of different drying stresses
in the machine and cross direction. Work on this subject particularly by
Brecht and his colleagues (44), Glynn and his colleagues (45, 55, 74), and
Hendry and Newman (89), though producing some clarification has served
to underline the interactions between these various factors which occur in
practice, and so far only a general qualitative explanation of the direction
and degree of curl exhibited in any particular case is possible.
Briefly, it is important to realize that neither two-sidedness nor fibre
onentation on their own would produce curl along a particular axis, both
must be present. If the fibres were randomly arranged in the paper but
208
the usual two-sided composition were present, then in the absence of
drying restraints, or of drying taking place more on one side than the
other, the sheet would exhibit a general bending inwards in a circular
manner towards the side which, by virtue of its composition or density,
had contracted most; for example, handsheets dried in air show a tendency
to fold inwards towards the wire side because, presumably, the higher
density and presence of a greater percentage of fines on the wire side of
handsheets gives proportionally more contraction during drying. If, on
the other hand, fibre orientation were strong but uniform through the
thickness of the sheet and there were no two-sidedness, then on drying
without restraint the paper would contract more in a direction perpendicu-
lar to the preferred orientation of the fibres, but would remain fiat. Only
when fibre orientation differs one side of the sheet from the other will the
relative shrinkage in machine and cross directions differ also on the two
sides, and then the interaction of shrinkage forces will cause the paper to
bend along one particular axis (this axis is usually defined as the direction
of the line of paper remaining fiat on the support surface). This axis
appears to be determined largely by the side having stronger fibre orienta-
tion; this side has greater shrinkage perpendicular to the direction of
orientation which produces a tendency to curl towards that side along an
axis parallel to the direction of alignment. In his investigations Glynn
found that the difference in the degree of orientation between the wire and
top sides bears quite a close linear relation to the degree of curl exhibited
by papers dried under similar conditions (45).
Due to the greater machine-direction fibre orientation normally found
on the wire side the usual structure of machine-made paper produces on
drying a tendency to curl with axis in the machine direction and towards
the wire side. It must be emphasized, however, that in practice drying
restraints affect this and can even reverse the direction and axis of curl;
in addition differential drying or damping, for example in an M.G. paper
or when using a sweat roll, can affect curl irrespective of other factors.
Finding the solution to any particular curl problem requires study from
many aspects, including those of the drying conditions which are described
in SA. 3 8 when the subject is dealt with in more detail, but the approach
must always be made bearing in mind the fundamental considerations
outlined above.
3A. 3 DRAINAGE AT THE SUCTION BOXES

The water/fibre ratio of stock on the wire after the last table roll is between
50 and 30, i.e. the consistency, solids content, or dryness of the fibre mat
(by definition these are all the same) is roughly from 2 to 3 per cent. After
passing over the suction boxes on the average machine the water/fibre ratio
is down to between 9 and 5, i.e. the solids content between about 10 and
17 per cent. Bennett (9) estimated that on one machine the suction exerted
in the first box was similar to that developed in the last table roll, though
the box extracted six times as much water. The value of suction boxes is
thus clear and to a great extent their efficiency under given conditions
209
governs the dryness of the sheet at the couch. It is all the more surprising
then that so little has been done to investigate this particular part of the
Fourdrinier and until relatively recently there have been only rough ideas
current as to how the suction box works and what influences its efficiency.
In passing over a suction box each part of the web is subjected to vacuum
for a particular length of time and this (assuming the open length of the
box is the same for all points across the wire) should result in a uniform
increase in the solids content. One of the fundamental aspects to determine
is how the increase in solids content depends on the vacuum, the length
of time it is applied, and the initial solids content, i.e. the position of the
box. A certain quantity of air is pulled through the sheet into the suction
box and up to a point this affects the power used by the vacuum pump.
It is therefore important to know at the same time how this volume depends
on the three variables so that, if necessary, pumping costs can be related
to the increase in solids content. In the discussion following let it be clear
at the outset that reference to a higher, greater, or increased vacuum
implies a lower absolute pressure, i.e. a greater difference from atmospheric
pressure.
3A. 3 1 Machine experiments by Brauns and Oskarsson

Brauns and Oskarsson (5) were the first authors to report results relating
to the foregoing and their work remains valuable because it was undertaken
on an experimental machine rather than, as in most subsequent cases
mentioned below, on an apparatus designed to simulate suction box action.
All the results were taken on one particular suction box, one section of
which comprised a single compartment beneath a normal slotted cover,
while another section was divided parallel to the slots into three separate
compartments; the vacuum applied to each compartment could be varied
and the resulting drainage from the web measured. The solids content at
the leading edge of the suction box was approximately 4 per cent. .in all
cases.
Applying the same vacuum to each of the three successive compartments
gave the results shown in Fig. 3. 6; in this graph the quantity of water
extracted is expressed as a cumulative percentage of the total in the web
before passing over the box. It is apparent firstly that the quantity of water
removed in each successive compartment reduces and this may be expected
because there is less water remaining for removal. If the quantity extracted
is expressed as a percentage of that remaining to be removed, however, it
may be shown that this also reduces so that the effect of a particular
vacuum is less the drier the web. Finally it may be noted that as vacuum is
increased it has less effect on the water removed; thus, with a vacuum of
800 mm. H 20 or 2·3 in. Hg almost 50 per cent. of the water initially in the
web is extracted, but with a vacuum double this the percentage increases
to only about 65 per cent.
From this data two points may be deduced: to remove the same propor-
tion of water at each successive box a higher vacuum is required; and the
quantity of water extracted increases as the vacuum is raised, but with
210
THE WIRE SECTION 3A.3 I
diminishing effect. The second of these points is important when con-
sidering the relative value of extracting more water at the suction boxes
by means of a generally higher vacuum at the expense of greater power
consumption and shortened wire life. The first point leads to the notion of
gradually increasing the vacuum applied to successive boxes instead of
keeping it uniform, and this formed the second part of the experimental
work reported by Brauns and Oskarsson. In this case exactly the same
general conditions pertained but instead of applying the same vacuum in
each box the first one was reduced by 200 mm. H20 and the third increased
,.........,_
w
~
:J
,J
80
~
.J
< LiNIFO{;?M
;::: VACUUM
i IG;.oQ mm. Hz 0
~ Go
'-' 1'2.oo mm. 1-lz o
ll
w
:::..
0 800 n'lm. H2o
~ 4o
UJ
~
tl
w
!{
~ 2o
J
~
~
2
COMPARTMENT NUMBt;R
Fig. 3.6. Water removed in successive compartments of a suction box expressed as a
percentage of the initial volume of water present in the paper; shown for three different
vacua uniform in each box (after Brauns and Oskarsson)
by this amount. The result of this was to increase substantially the quantity
of water removed and in Fig. 3. 7 the effect on the ultimate solids content
of the sheet for different average vacua is shown. To achieve the same
solids content without staggering the vacuum applied, a higher uniform
vacuum would be required with consequently greater power consumption,
lower wire life, and other disadvantages.
Measurements of the volume of air passing through the box compart-
ment were also made and it was estimated that air flow commenced when
the sheet was between 6'5 and 7·5 per cent. dry, though this figure will
211
depend essentially on the degree of consolidation of the web and other

factors. The results show that under a given vacuum the air flow increases
as the sheet becomes drier, but this is not accompanied by a corresponding
rate of increase in water removed; Brauns and Oskarsson conclude that
the initial quantity of air flowing through is therefore of greater significance
and the reason for this will be examined shortly.
3A. 3 2 Suction box simulation experiments

Apart from a general confirmation by Tepelnev and Sokolov (30) that
staggering the vacuum in the boxes of a paper machine gives an improve-
ment in drainage over uniform vacuum (the average being the same in
1z \~ x
$
UNl FORM VACUUM
\NCltEA!»\NG VACUUM
0 IN SUC:C.ESSIVc
j::
u CoMP.:lt\~TMENT.:S
:I
l'Z
\I) (Ave:~AG! f'1..orre:~
at
UJ
....
""
~
ll
~
I-
z lo
,_
I.LI
z
0
v
~
e
0
:::;
() IN1T\AI... So1..1os. CoNT&NT: 4 %.
If)
8
200 4op ~00 800 looo 1200 1400
VACUUM, rwi ni. Hz·o.
Fig. 3. 7. Solids content of web drained at different vacua under uniform vacuum and
under progressively increasing vacuum (after Brauns and Oskarsson)
both cases), almost all other experimental work on suction boxes, notably
by Nordman (14), Millier-Rid and Pausch (28), and Attwood (53, 40),
has been done with apparatus designed to simulate the action of a suction
box. The first two workers used a stationary drainage device in which
there was no simulation of the scraping action between the wire and box
which occurs on the machine; Attwood has shown that this action is highly
important so that the value of this earlier work is limited though the general
conclusions probably hold good.
Nordman used a constant delivery vacuum pump in his work which
meant that the vacuum varied according to the porosity of the sheet; thus,
212
although he found that adding fines to the stock produced an increase
in the dryness after a given suction period, this may well have been the
result of a greater vacuum due to the increased density of the sheet.
Muller-Rid and Pausch controlled the vacuum and obtained an opposite
result, i.e. that free pulps drained easier under given vacuum conditions,
and also a greater volume of air passed through the sheet; this accords
with common observations that when stock on the wire becomes free either
the vacuum falls or, if it is regulated, the air flow from the boxes increases.
These workers also confirmed that the vacuum is the most important
single variable affecting the drainage rate and under given conditions there
appeared to be a substance of sheet at which the final solids content
was highest; the value of this optimum substance increased at higher vacua
and with greater beating of the stock.
Apart from these results the work of Milller-Rid and Pausch is interesting
in showing the effect of continual application of a steady vacuum. Fig. 3. 8a
shows some representative curves and it will be noted that prolonging
application of vacuum has a steadily diminishing effect on the ultimate
solids content of the sheet. This is particularly noticeable for lower vacua,
when it may also be observed that the curve flattens off earlier even though
the quantity of air passing through the sheet steadily rises, as shown in
Fig. 3. 8b. The implication of this is that the lower the vacuum at a box,
the less time it should be applied, because a point is soon reached where
little further dryness is achieved despite the fact that air drawn through the
sheet, and consequently the pump power consumption, continues to
increase. Under any particular machine conditions it cannot be easy to
find whether the total time of application of suction to the sheet represents
a good compromise between producing a sufficiently dry sheet and using
excessive power (both as a result of drawing through too much air and
having a larger friction area on wider boxes). It may be suspected that,
especially on slow machines, it is common to have an excess of suction
area; indeed one worker, Shamolin (61), has reported a considerable
power saving with negligible decrease in solids content of the web resulting
from reduction of the width of suction boxes to a third.
Certainly, it is worth experimenting with the available suction area on a
machine and it is quite possible to find that one or two boxes at the dry end
can be lowered, and the overall vacuum on the remaining boxes raised
slightly to compensate, but the effect is to decrease the total power con-
sumption and increase the life of the wire.
3A. 3 3 Attwood 's experimental work

B. W. Attwood (53, 70) began his investigations into the action of suction
boxes by designing a laboratory simulation apparatus similar in basic
principles to that of Muller-Rid and Pausch. However, it soon became
apparent that the apparatus, though producing similar results to those
obtained by Muller-Rid and Pausch, in fact removed far less \Vater under
the same vacuum conditions than occurred on a Fourdrinier machine
when using the same sort of pulp. Various differences in the characteristics
8 213
\~\TIA.\- SOLID& CO~TilNT; A.%

SU<::T'IOr-4
INS. l-l3
_ _ _ _ _ 2.4
... 22
z
.UI
f-
-z.
3 ls
II)
0
J
0 l4.
If)
TIME SUCTION A?PLIE:D, 6ECONO~
Fig. 3. Sa. Final solids content of sheet after application for varying time period of
different vacua (after Millier-Rid and Pausch)
...u 12.
~
Cl
:> lo
\ff
...
<. 8
":::>
«
I.I.I
IZ
0...
~
"'
~ 4
u.
()
\I.I
~ z _ _ _ _ _ _ _ _ _ <O
:J
_J
~
0
0 O·I 0·2 O·~ 0·4-
TtME SUCTION AFf>L.lt:O, SEC.ONOS.
Fig. 3. Sb. Volume of air passing through sheet after application for varying time
period of different vacua (after Muller-Rid and Pausch)
214
of formation of the sheet on the machine and in the laboratory apparatus
were ruled out as possible explanations of the discrepancy, and it was
decided that the scraping action of the wire on the suction boxes must have
an important effect.
Observation with a high-speed cine camera of the underside of the wire
on the laboratory apparatus and (by fitting a simple slotted suction box
with transparent plastic sides) on a pilot Fourdrinier machine confirmed
the difference in action occurring in the two cases. With the static labora-
tory apparatus, during the vacuum period water is disengaged from the
wire as drops and spray from bursting bubbles but a substantial quantity
of water remains attached to the underside of the wire by surface tension
forces. This water, particularly at higher solids content, is rapidly sucked
back into the sheet in a matter of a few hundredths of a second after the
l"'\TIA\... SOLl05 CONTENr: 3·S7o.

s S1...oiS x Ya'
l<O
~
,_ IZ
z
uJ
.....
~ s
u
II)
04
J
0
If)
0 4 8 1'2 I <P.
VAC.UUM IN.,,. H9.
Fig. 3.9. Relation between solids content and vacuum applied to different widths of
slots, the suction width totalling one inch in each case (after Attwood)
vacuum is shut off and this is attributed by Attwood to elastic expansion

of the layer of fibre next to the wire; under vacuum this layer is compacted
by the difference in pressure and on release of the vacuum water is pulled
from the underneath of the wire and from the mesh of the wire to fill the
intersticial volume created between fibres. On the Fourdrinier the water
pulled underneath the wire by the influence of vacuum could clearly be seen
being removed in considerable quantities by the scraping action of the
trailing edge of the suction box slot.
Accordingly, a new laboratory simulation apparatus was built comprising
a shallow container with a wire mesh bottom placed above a vacuum
chamber; the wire and vacuum chamber are separated by the outer part of
215
a large solid circular disc into which slots or other shapes of hole can be
cut at intervals. When the disc is set in motion revolving about its centre the
effect is to reproduce the dynamic conditions of drainage on a paper
machine with the only difference that the suction box slots move relative
to the wire instead of vice versa. This apparatus proved to simulate con-
ditions of drainage on the machine wire very closely.
Observations by Attwood with this modified laboratory simulation
apparatus showed that increased temperature and reduced consistency
both produced better dewatering. The importance of the scraping action
was confirmed by determining the solids content achieved using numbers
of different width slots adding together in each case to the same total length;
thus, in Fig. 3. 9 the resulting solids content from using 8 x l· in. slots,
4 x ! in., 2 x t in., and 1 x 1 in. slot can be compared for different
vacua (initial solids content was 3·5 per cent.). It is apparent that the
more slots are used, i.e. the more the vacuum area is subdivided into slots
of smaller width, the more effective is the application of vacuum (there
will, of course, be a lower limit to this effect, as discussed below). Looked
at another way, if for any particular slot the width is extended beyond a
certain point, the vacuum will cease to have any further significant effect
and it would be more beneficial to use up the available area by starting
another slot; this, therefore, confirms the conclusion reached earlier by
Muller-Rid and Pausch when they compared the effect of applying vacuum
for varying lengths of time.
Attwood also determined from his results that it is possible to reach
only a certain maximum solids content in a reasonable time with a given
vacuum, and to increase the solids content above this value requires a
higher vacuum. This again confirms the same result obtained by Milller-
Rid and Pausch. The advantages of increasing the vacuum as the web
gets drier, compared to using the same average uniform vacuum on all
boxes, was also demonstrated. Increased speed had the effect of reducing
the dewatering capacity with a given vacuum, as did increase in substance
of the sheet (contrary to the findings of Muller-Rid and Pausch with their
static simulation apparatus).
3A. 3 4 Summary of suction-box operation

From the foregoing it is possible to give a reasonably consistent explanation
of the course of water removal at the suction boxes and to some extent to
deduce how present suction box arrangements may be operated or modified
to improve their performance.
At the first few suction boxes the increased pressure differential acting
on the two sides of the sheet causes compression, and water is squeezed
out through the capillaries in the web; due to the elastic condition of the
mat the compression must initially occur mainly in the lower layers of the
sheet where the change in pressure is first felt. Some water sprays through
the wire into the suction box, but probably the majority is removed at the
trailing edges of the slots or holes of the box where it is scraped off from
underneath the wire. No air is sucked through the sheet at this point.
216
On release of vacuum the web expands and very rapidly draws any water
in the wire mesh or remaining underneath the wire back into the web.
As the web gets progressively drier water must be removed from smaller
capillaries, requiring a greater differential pressure. Thus, it is to be expected
that the higher the temperature of the water, and hence the lower the sur-
face tension and capillary resistance, the easier it should be to dewater the
sheet. In this phase, when the web is becoming relatively less saturated,
the compressibility and elasticity of the fibre mat will also presumably
be of importance.
The point is soon reached when the upper layers of fibres in the sheet
are freed of their overall film of water (visible as the dry-line) and air
enters the larger capillaries. Each release of vacuum will then cause some
water to migrate upwards back into these capillaries, but as soon as vacuum
is re-applied they will be cleared, together with more capillaries in the
lower layers, until eventually continuous air channels are formed through
the sheet (this occurs at 6·5 to 7·5 per cent. solids content according to
Brauns and Oskarrson). A small quantity of air passing into the suction
box is probably of some benefit in assisting the transportation of water
particles, but it is almost certain that evaporation from the web due to
passage of air through it is of no particular significance. Fibres which have
been more heavily beaten will generally compress easier and form a smaller
average capillary size, so that removal of water from the web will require
greater suction to achieve the same dryness while less air will be sucked
through the sheet.
Together with this general explanation the following points drawn from
the experimental results mentioned earlier are worth reiterating:
(a) A higher vacuum is required to remove the same proportion of
water when the sheet is drier.
(b) At any particular dryness of the web, increasing the vacuum
increases the quantity of water removed, but with diminishing effect.
(c) There is no value derived from applying vacuum beyond a certain
(relatively short) length of time because this only raises the quantity of
air drawn through the sheet for little increase in solids content; also
the lower the vacuum applied the less time is needed to reach the opti-
mum point.
(d) The scraping action of the trailing edges of suction box slots
provides a highly important contribution to the removal of water and,
up to a point, the greater the number of scraping edges, i.e. the narrower
the slots, the better.
From these observations the following conclusions may be justifiably
drawn:
(a) In the first two or three boxes the vacuum applied need only be
relatively small (sufficient to ensure an adequate film of water under the
wire) but should be progressively increased.
(b) The vacuum applied to the sheet when it is becoming relatively
dry, and particularly once air is being drawn through the sheet, still
217
needs to be progressively higher to have an adequate effect on the sheet,

but a limit must probably be reached in the last two or three boxes
above which little additional benefit is derived from continuing to increase
the vacuum to too high a value.
(c) As a consequence of (a) and (b) probably the best arrangement of
applying vacuum in the suction boxes is to have a progressive increase
(of perhaps tin. to 1 in. Hg.) up to the last two or three boxes beyond
the dry-line which then remain at a normal running maximum.
(d) With the qualification that the proportion of surface area of a
box to the vacuum applied has an influence on the drag exerted on the
wire (see 3B .41), the gap between different boxes and between slots and
holes in a box should be as narrow as possible to avoid reabsorption;
this point applies especially to the drier boxes but may be expected to
be of significance mainly on slower machines.
(e) A compromise must be reached in the machine-direction length
of each slot or hole between subjecting the web to vacuum for a sufficient
enough time for it to have effect, and providing as many scraping edges
as possible; for the first two or three boxes a further consideration may
be to ensure the slot is wide enough to conduct the water away adequately.
(f) The width of slots at present in common use may generally be too
wide (Attwood assessed that there is little advantage in having slots
greater than t in. width; however, his results in this respect, though
admittedly showing no dependence on speed, were for machine speeds
equivalent to under 400 feet per minute and are not necessarily capable
of extrapolation to higher speeds).
(g) Combining (d), (e) and (f) it seems reasonable to suggest that the
first two or three boxes should be equipped with the usual cross-direction
slots, narrowed to about t in. in width and possibly with edges angled
in the direction of flow to help scrape off the water under the wire and
conduct it away (though this presents a design problem because, unless
the whole box is constructed with each slot falling away at an angle
(see for example reference 101), account has to be taken of the surface
gradually wearing thus causing the land area to diminish and neces-
sitating replacement after a smaller depth than usual has been worn
away); also the gap between individual slots should be as narrow as
possible compatible with adequate structural rigidity of the box.
(h) Towards the drier boxes it is desirable to lengthen the time of
application of vacuum. Since the vacuum itself is greater in these boxes
a point must be reached in widening the slot design where wear on the
wire is accentuated by the tendency of the wire to dip into the slots;
thus, drier boxes are probably better equipped with the staggered hole
pattern, though the individual holes could perhaps usefully be elongated
in the machine-direction to 2 in to 3 in in length provided their arrange-
ment still permits each part across the web to be equally treated.
As an alternative to this one or other of the herring-bone designs
should achieve roughly the same effect.
218
3A. 4 CONDITIONS FOR COUCHING

In this section it is proposed to give only some brief details relating to
general conditions of couching the web from the paper machine wire; the
various methods by which couching takes place and the sheet is carried
over to the presses will be given individual treatment later. The open draw
commonly used on all but the faster machines is intimately associated
with the general problem of couching, so this will also be considered here.
3A. 4 1 Forces governing couching

The effort required to pull the web from a paper machine wire is governed
by the relation between two primary forces: the adhesive strength of the
web to the wire, and the effectiveness of tension in the web in overcoming
this adhesion. This relation has been found to depend closely on the angle
at which the sheet is pulled off, not simply because the effective pull
vertical to the wire obviously reduces as the angle becomes more acute to
the wire, but also because the adhesive resistance itself varies with the angle.
For the moment this aspect will be ignored and attention confined to
discussing what affects the adhesion of web to wire irrespective of the man-
ner in which the web is stripped off.
In fact, little work has been done on this subject though it is considered
likely that several factors can increase the adhesion of web to wire at the
couch, in particular coarser mesh wires and the greater fibre entanglement
in the wire which would be expected to occur under conditions of early,
rapid drainage or heavy suction on the boxes. Also the material and weave
of the wire is likely to affect the adhesion (it has been suggested that plastic
wires give less adhesion than metal) and the composition and degree of
beating of the furnish.
In practice, however, the most impoi;taqt factor affecting adhesion at the
couch appears to be the presence of a film of water on the underside of the
mat. This may be thought of as providing a loosening of the fibre layer in
immediate contact with the wire by reducing the strong surface tension
forces binding individual fibres into contact with the wire strands. All
methods of couching depend for their effectiveness on the provision of such
a film of water on the underside of the sheet. With the jacketed top couch
above a solid wire couch roll the pressure exerted on the wet web serves
to produce this film; with a suction couch the same effect is achieved by
the pressure differential resulting from the suction pull. In both cases, of
course, an important effect of the couching arrangement is the increase
produced in the solids content of the web, though this is really a separate,
and subsiduary, function from the actual operation of couching.
If the film of water, once produced, remained between the web and the
wire, then designing a couching arrangement would be relatively easy.
In practice the water is sucked back into the web as pressure on the sheet
is released in a manner identical to that occurring with suction box opera-
tion, as discussed above. Moreover, at the solids content met with at the
couch this reabsorption occurs extremely rapidly; Baggallay (17) and Hen-
219
dry (19) in their investigations of suction couch shadow-marking (see next

section) have both given examples of this. The first author quotes an increase
in water/fibre ratio to 6-9 from 5·2 on one machine, and 5·7 from 4·9 on
another, obtained by running a slack instead of a tight draw and thereby
leaving the web in contact with the wire a few inches further round a
suction couch roll.
Baggallay and Hendry also point out that the web can be pulled off the
wire immediately after the final suction box with little trouble, but further
down the wire this becomes more difficult and fibre is likely to be left
adhering to the wire. Baggallay also showed that, irrespective of the solids
content of the web, a simple suction box placed under the wire at the point
of couching considerably reduces the tension needed to pull off the sheet.
Originally in couching from a position between a suction couch and for-
ward drive roll on one machine, adhesion of the sheet was so high that
an appreciable fibre pattern was left on the wire and the tension required
to effect couching was considerable; applying the suction box removed this
completely and allowed a much slacker draw. These observations again
demonstrate the importance of a film of water between sheet and wire for
reducing the adhesion.
3A. 4 2 Shadow-marking at a suction couch

It is appropriate at this point to deal with the question of shadow-marking
caused by a suction couch; this subject was the original reason for the
investigations of Baggallay and Hendry mentioned above and what follows
is largely drawn from their reports. If the web is pulled off the wire close
to the trailing seal-strip of a suction couch then a pattern can be observed
in the paper, either by eye on slow machines or with the aid of high-speed
photography or a stroboscope on faster machines. This pattern shows up
in the form of darker areas which correspond to the holes in the suction
couch, and it has been proved that the sheet in fact comes away wetter
over the hole regions under these conditions. The pattern usually remains
visible in the paper at the reel-up due most likely to a difference in bonding
of the fibres at the presses that produces an optical effect in the finished
paper. This particular effect is described in more detail when considering
shadow-marking at a suction press in 4A. 2 4.
The difference in water content in the two regions is attributed to the
fact that immediately the vacuum is released the web absorbs water from
the wire opposite the holes due to the sudden change in pressure differential
and up-rush of air, while over the land areas surface tension forces on the
couch shell hold water more strongly in the wire and prevent reabsorption
by the web to the same degree. If couching takes place closely after the
trailing seal-strip terminates the wire is observed to be dry opposite the
holes and wet opposite the land areas, which confirms this explanation.
It might be expected that the suction force in the holes would have the
opposite effect, i.e. tend to remove more water from the web opposite the
hole areas so making them drier. This does not appear to occur to any
significant extent and it must be presumed that water drawn into the suction
220
couch holes comes fairly evenly from the web while at the same time the
wire itself is evenly flooded. When a presser roll is used, however, a slightly
greater pressure may occur opposite the land areas of the couch which
would produce a lateral movement of water towards the hole regions,
thus enhancing the eventual difference in water content.
If, instead of being drawn off the wire close to the suction box seal-strip,
the sheet is allowed to remain on the wire for a short distance round the
suction roll, then the pattern is observed to fade. This is due to the fact
that a greater reabsorption of water by the web occurs opposite the land
areas where the wire carries more water, thereby tending to equalize the
water content of the web. Redistribution of water laterally within the
sheet cannot explain this as shadow-marking when present after couching
shows no similar signs of fading in the sheet before reaching the presses.
There is, therefore, a choice between couching the web from a suction
couch roll close to the trailing seal-strip, when shadow-marking is likely
to be prominent but the average water/fibre ratio of the sheet is low, or
alternatively the sheet may be allowed to wrap round the roll a bit further
in which case the shadow-marking decreases but the water/fibre ratio
increases. In practice the most sensible approach with a straightforward
suction couch and open draw arrangement seems to be to draw off the
couch at a point where shadow-marking is just within tolerable limits.
3A.4 3 Angle of draw from the wire

With increasing machine speeds the frequency of breaks occurring at the
couch when an open draw is used invariably rises, and this can either make
faster running economically unsound or necessitate a more costly furnish
or greater treatment of the pulp to make the web stronger at the couch.
The greater likelihood of breaking is caused principally by two conditions
which affect the wet web increasingly at higher speeds. Firstly, a higher
tension becomes necessary at the draw to overcome the greater inertia
of the web and also to provide the extra effort which is known to be required
to effect separation from the wire at higher speeds. Secondly, on any par-
ticular machine higher speeds make it more difficult to effect a compromise
between having the breast box consistency sufficiently dilute to form the
sheet and having adequate drainage in the wire section without excessive
vacuum on the suction boxes; the usual result is that the solids content at
the couch drops and with it the wet strength of the web.
It is difficult to assess the relative importance of either of these factors;
the first will increase in importance with the substance of the paper and
the second depends on how sensitive the wet web strength is to alteration
in solids content. Lyne and Gallay (13) and Robertson (94) have demon-
strated that in the region of 20 per cent. to 30 per cent. solids content,
which is approximately that found at the couch on most machines, the
change with dryness of wet web strength (tensile breaking length) may be
considerably less compared to that at both higher and lower solids content,
though some decrease in strength with lower solids content always does
occur, see Fig. 3. 10; this phenomenon is attributed to a change
221
lo,ooo
&1000
SUL.-P ... l"rc
!Ji 1 000
1
lU
°'
t-
UI
~
Soo
~
:r
t-
\!J
z
uJ
..J GROUNOWOOO
I!) loo
z
i
.(
w
~ So
dl
lo
o 2o 4o ~o 80 \oo
501..1os CONTE.NT %
Fig. 3. 10. Change in breaking length of sulphite and ground wood wet web with
increasing solids content (after Lyne and Gallay)
occurring from surface tension to fibre bonding forces as the main

source of cohesion in the web.
Apart from these factors it is apparent that the strain on the sheet at an
open draw is very dependent on the draw applied, i.e. on the difference in
speed between the first press and the wire. A relatively large difference in
speed produces a greater tension, giving a tight draw with the sheet peeling
off the wire at an acute angle; a small difference in speed produces less
tension, giving a slack draw with a larger angle of take-off. In practice
222
the machineman uses his experience to set the draw in a suitable position.
If he runs the draw too tight the higher tension increases strain on the
relatively weak and plastic sheet and so the web is more likely to break
either at the couch or further down the machine; at the same time the
web width will shrink more and a greater anisotropy of strength and stretch
characteristics of the finished paper in the machine and cross directions
may be produced. Running with the draw too slack on the other hand
~
0
~
~
~
~
! 8
"Z
()
'°
w
'X
Q G
~
UI
§4
rt
lit
6
~ 2
z
0
'1)
z
w
t-
0 3o co~ ~o 1'2o t5o
ANGL..E. OF IA.KE-OFF IN OE.GRees.
Fig. 3. 11. Relation between tension necessary in web to overcome adhesion to a
twill wire and angle at which web is drawn off (after Radvan and O'Blenes)
makes the sheet liable to run into longitudinal creases and wrinkles,
especially at the edges, and these may be cut through in the calenders.
At first sight it would be thought that maximum effect is achieved if the
sheet is couched off the wire at an angle of 90 deg., since in this position
the tension has the greatest pull normal to the wire; in other words, if at
any time the adhesive force were too great for the tension applied in this
direction then the sheet must carry round the couch roll and eventually
break. This would be correct if the force required to overcome adhesion
were independent of the angle at which the sheet is pulled off, but in fact
this is not the case. Several workers have investigated this question both
experimentally and theoretically, notably Gavelin (34), Mardon and his
223
colleagues (37, 59), Howe (46), Radvan and his colleagues (51, 80), Camp-
bell (64), and Osterberg (79). There seems to be general agreement that the
effort required to overcome adhesion is closely dependent on the angle of
separation and the actual tension needed in the sheet to accomplish couch-
ing is at a minimum for an angle greater than 90 deg., usually in the region
of 120 deg. to 130 deg., see Fig. 3. 11. The implication of this is that under
given conditions least strain is put on the sheet if it is drawn off the wire
backwards at this angle.
Normally the machineman tends to set draws relatively tightly. This is
probably because the creasing likely to come from slack draws is readily
observable and to be avoided, whereas to a large extent frequent breaking
is not immediately associated with undue strain in the sheet resulting from
a tight draw. This is especially the case if the breaks occur further down the
machine when any weakness induced in the sheet at the couch is subjected
to greater strain; when this occurs the reason for such breaks may be
sought more readily in the formation of the sheet or preparation of the
stock and the true cause is not always easy to trace. Radvan, in particular,
has reported how the frequency of breaks on a paper machine at the open-
draw from a press (where a similar situation to that at the couch prevails)
steadily decreased as the draw was gradually slackened back from an angle
of take-off of about 40 deg. up to 70 deg.; other experiments also showed
that running the machine became easier at larger angles of take-off.
How slack the draw may be run in practice depends essentially on the
stability of the machine, in particular of the drive and uniformity of the
paper. If the angle of take-off from the wire is relatively acute then any
increase in adhesion is compensated for by the web remaining on the wire
until a larger angle of take-off is reached where there becomes sufficient
normal pull to provide the necessary extra force; in other words the opera-
tion is stable and self-correcting for any variation in adhesion. If on the
other hand the draw were run with the angle of take-off from the wire very
close to the minimum tension point, then any increase in adhesion of the
web is liable to cause the angle of take-off to exceed the position where the
tension has most effect; beyond this point the effective force pulling the
web from the wire decreases and a break is certain. In addition, as the
angle of take-off is increased flutter of the separation line grows due to
the greater change in angle necessary to overcome a given increase in
adhesion (this is reflected in the decreasing slope of Fig. 3. 11 at higher
angles of take-off); this in itself is likely to create stresses which weaken
the sheet. Because of the unknown influence of these various factors it is
hardly practicable to predict that a machine should run best at some
particular angle of take-off and over a long period the point where the
frequency of breaks is a minimum can only be found by trial and error.
Possibly on most machines once the operating conditions have settled down
fair compromise will be found with the draw slackened back to give an
angle of take-off approaching 90 deg. In this respect it is important to
realise that any relatively large change of draw during normal running may
have an effect on the substance. Elongation in the machine direction caused
by tension at any draw, but especially at the couch, is not accompanied by
224
a corresponding shrinkage in the cross-direction; the net effect of increased
tension is therefore to increase the area of the paper giving the effect of a
reduced substance.
3A. 4 4 Breaks at the couch

Apart from the question of the overall stability of the machine being an
essential factor governing the slackness with which the couch draw can be
run, in practice the presence of local variations in adhesion of the web to
the wire is probably equally if not more important. At any open draw it is
possible to see indentations in the line of take-off which may be permanent,
intermittent in one position, or simply occurring spasmodically across the
sheet. These show up the difference in adhesion caused by such factors as
uneven substance streaks, worn or dirty wire patches, plugged couch holes,
clogged dandy, and in particular fibre clumps and contraries in the sheet.
Some interesting experiments have been reported by Radvan and his
colleagues (51, 80), and by Mardon and his colleagues (59) involving an
examination of what happens when the local adhesion at some part
of the web becomes too great and causes a break. At normal angles of
take-off if a small spot sticks to the couch or a press roll it initiates two
tears which run outward on either side of the spot to the edge of the sheet
to produce a break. If the angle of take-off is increased by slackening the
draw then the tears separate outward more slowly and a point is finally
reached at an angle of take-off greater than 90 deg., i.e. when couching the
web off backwards, where the tears do not separate at all but run together.
In this case no break occurs (at least not immediately) and instead only a
small hole is pulled out of the web. Although this situation does not appear
to occur on all machines, the rate of separation of the tears always becomes
slower at greater angles of take-off and the phenomenon has been observed
on both slow and fast machines.
Other observations have shown that a sticky spot having greater ad-
hesion than the rest of the web may not actually start any tears but instead
appears to be plucked out of the web leaving only a thin spot in an other-
wise intact sheet. This has been termed 'skinning' and has been found to
occur more frequently at higher angles of take-off. A break is, of course,
less likely to be caused by skinning, so this provides a further reason apart
from the reduced tension needed in the web for running at higher take-off
angles.
225
CHAPTER 3B
OPERATING FACTORS AFFECTING THE WIRE SECTION
3B .1 EARLY DRAINAGE CONDITIONS

The conditions prevailing when the stock meets the wire and drainage
first takes place are highly critical for the structure and formation of paper.
Unfortunately, generalizations on this subject, except for more obvious
remarks, are particularly difficult to make since practice in the industry
varies appreciably not only between slow and fast machines but also between
practically similar machines making the same quality of paper.
Once reasonable running positions for the apron, slice jet, breast roll,
forming board, etc. have been determined over a period by intelligent
manipulation, the papermaker is generally most reluctant to make any
alteration. And not surprisingly so, for interaction of the numerous variables
in this part of the process makes it impossible to predict the outcome
except in the vaguest of terms; in addition a workable arrangement has to
be found for any machine which suits all the grades manufactured without
the need for a lengthy engineering alteration each time there is a major
change and this, by and large, is a purely empirical process. Even for those
machines making the same grade day in, day out, adjustments such as those
of the slice lips relative to the wire and to themselves, of the forming board
angle, breast roll discharge, slice jet to wire velocity, and many other
variables need to be very gradual in nature and carefully evaluated over a
long period of time. For this type of investigation a very convenient ap-
proach is to use an Evolutionary Operation type of experiment; this is
eminently suited to seeking over a long period, and without affecting
production, an optimum position of a number of variables which may
interact with one another and which affect test properties of the end-
product in different ways.
Despite these qualifications to the value of any detailed discussion of this
subject, there are several observations which have been reported in the
literature that are well worth noting. Most of these concern avoidance
of undue disturbances during early drainage of the web on faster machines,
though one or two remarks will also be made on how the formation of the
sheet (in the general sense of the term) is affected, especially with regard
to variation of the stock or slice jet velocity relative to the wire speed.
For slow machine operation the most important variable in this field is
that of shake but this topic is not considered until 3B. 3.
3B . 1 1 Impact of stock on the wire

The point where the stock first contacts the wire and the angle of the impact
are both highly important from the point of view of formation and the
prevention of defects in appearance caused by stock jump and other
226
THE WIRE SECTION 3B. l I
disturbances. There are generally two conflicting interests in the early
part of the wire section which have a close bearing on where impact of the
stock occurs: the necessity for drainage to commence as soon as possible
in order ultimately to increase the solids content at the couch, and the
requirement that this process of dewatering does not disturb or spoil the
general formation.
On most machines a compromise is usually necessary on the one hand
to avoid excessively rapid drainage in the first few feet of wire length which
is liable to disrupt the whole structure of the sheet, and on the other to
avoid delaying drainage to such an extent that the sheet leaves the couch
very wet and the breast box consistency has to be raised to offset this,
while in addition the formation may be spoilt on the top side of the sheet
by the results of flocculation. On faster machines an added disadvantage
of early rapid drainage can be a prominent wire mark. According to
Forgacs and Atack (73), this is due to short fibres being drawn end-on
into the wire meshes, wedged there between the predominant longer fibres
(which are more likely to span across the meshes), then flattened out in
pressing and calendering to form tiny mounds; the mark will, of course,
be greater with coarse wires and furnishes composed of a large proportion
of fines. Laboratory investigations reported by Dushnicki (110) also
confirm that early drainage rate is the key factor affecting wire mark.
This will be strong in conditions where the slice jet meets the wire with a
high vertical velocity component and early drainage at the breast and
succeeding table rolls is rapid. Later stages of drainage at the suction
boxes and couch can also affect wire mark but only if the initial drainage
conditions are not severe. If, on the other hand, early drainage is very slow
and in addition the consistency of the stock is low, then flow over the early
part of the wire becomes unstable and oscillating streams emanating from
small discrepancies in the slice jet appear on the surface ('skating') and
produce variable cross-web substance fluctuations.
The effective total suction in the nip of the large breast roll is considerably
greater than for a table roll revolving at the same speed, so if a machine
wire is short of drainage capacity it is usual to arrange for stock to contact
the wire close to top dead centre of the breast roll. This produces a con-
siderable discharge but nonetheless appears in most cases to give a satis-
factory sheet, possibly because the alternative of running without breast
roll discharge may necessitate an increase in breast box consistency which
destroys any potential benefit to the formation. When running in this way,
however, it is very important to ensure that the point of impact of the stock
is always close to the top dead centre; if the stock meets the wire beyond
this point on the breast roll, allowing a partial discharge, variations in the
drainage can occur as the head in the breast box and hence the stock speed
fluctuates, or if the wire tension and hence the angle of wrap round the
breast roll alters. In practice trouble from these causes is more likely to be
noticed in cross-web variations and any misalignment of the breast roll
and slice will readily produce a variation across the sheet.
On slow apron machines it is not generally advisable to have the apron
terminate over the top of the breast roll, particularly if there is the possibility
227
3B.l 2 THE WIRE SECTION
of a flow down the back of the roll, and in any case drainage on such a
machine will generally be more by gravity through the wire. Usually it is
more important on such a machine to delay drainage sufficiently for the
shake to have the opportunity to influence the mat; for this reason the
apron usually extends down the wire short of reaching a position where
trouble may be incurred from water under the apron running into channels
since this can disturb the mat and mark the wire side of the paper. A better
alternative to an excessively long apron would be to use a forming board.
3B. 1 2 Reducing disturbances on fast machines
On fast machines the suction exerted at the breast roll becomes considerable
and, except for tissue machines with specially designed pressure formation
breast boxes where suction may be applied in the breast roll, there are
strong arguments against having any breast roll discharge. For one thing
the dangers of operating with variable drainage are greater but perhaps the
main reason is that stock jump and other disturbances are much more
severe at a breast roll.
Mardon and Truman (50), amongst others, have made some interesting
observations and a theoretical analysis of the forces operating at the point
of impact of a slice jet which are highly relevant to this problem. When the
jet meets the wire a certain vertical velocity component is inevitable and
this produces a condition in which the upper surface of the jet may become
more disturbed on impact by destruction of this velocity. Any small ridge
in the jet produced by an imperfection in the slice flow meets the wire with
a slightly heavier impact and this may cause the ridge to split or become
accentuated. In the worst cases spouts form on the surface of the stock
which break away to produce the spray familiar on all fast machines.
With a breast roll discharge the additional downward acceleration
induced by the roll, coupled with an upward impulse at the point where the
wire wrap round the roll terminates, makes these disturbances more likely
to occur and for this reason it is more satisfactory to avoid this method
of operation. Higher wire tension, by reducing the wrap round the breast
roll, improves the situation, but as the breast roll is unfortunately at the
opposite end to the driving roll the wire tension there is comparatively low.
Disturbances can also be reduced by grooving the breast roll. Other
workers, including Miiller-Rid and Pausch (28), have confirmed that less
disturbance occurs when there is no breast roll discharge, and it may also
be noted that suction of air between the underside of the jet and the wire,
which can be very troublesome and affect the trajectory of the jet unevenly
and spasmodically, is more likely to occur when the jet meets the wire
over the top of the breast roll.
It is equally evident for the same reasons that the stock should meet the
wire with as little vertical component as possible; when an apron is not
used this requires that the jet leaves the slice near horizontally (setting the
slice lips to achieve this condition has already been discussed), and meets
the wire in as short a distance as possible (which is also advantageous for
reducing the growth of instabilities in the jet). Plainly when no breast
roll discharge is desired an engineering problem arises here because a short
228
THE WIRE SECTION 3B. I 3
jet length with impact beyond the breast roll can only be achieved with the
lower slice lip extending well over the roll; to make this sufficiently rigid
with adequate clearance above the wire will necessarily increase the
vertical drop of the jet, and hence the vertical velocity of impact. In
practice a compromise position, in which the jet impinges on the wire
close after top dead centre on the trailing side of the breast roll, is often
used. In this case the disadvantages possible due to variable discharge
have to be faced.
With the slice jet meeting the wire after the breast roll less disruption of
the sheet occurs at the point of impact but substantially the same difficulties
with stock jump and other disturbances may arise as soon as the mat
passes over the first table roll where the first drainage impulse occurs.
Grooved rolls and dandy-type rolls have been designed to help overcome
this problem and these are discussed in 3B. 2 3, but for the moment atten-
tion will be turned to the use of forming boards as these are more intimately
connected with the conditions of early drainage.
3B. I 3 Forming boards

Forming boards have become increasingly common on fast machines and
in their usual position between the breast roll and the first table or grooved
roll are thought to serve two main purposes. First, and most important,
they reduce drainage in the first few feet of wire length and thereby make
the process of deposition of the wire side fibre layer more gentle, and cer-
tainly less subject to disturbances than at the breast roll or a table roll.
Secondly, their construction enables them to be set closer to the breast roll
than a table roll and this has the advantage on wider and faster machines
with large diameter breast rolls of reducing the gap between supports,
which in turn reduces the angle of wrap of the wire round the breast roll.
Forming boards come in many designs and lengths depending on the
papermaker's conception of their function on a particular machine. Some
are completely solid but generally slots are cut out at intervals of 2 in. to
3 in. allowing some drainage to take place; occasionally a light suction
and drop leg may be applied to these slots. The leading edge of the boards
are normally pointed at an angle not greater than 30° to give a smooth cut-
off of water carried on the underside of the wire. On fast machines there
may be a long series of forming boards interspersed with grooved rolls and
foils (see 3B. 2 3), the object being to de-water the rolls gradually, but
sufficiently to reduce stock jump at the first solid table roll. In all cases,
however, correct setting of each board is most important and in this
respect it is essential to have adequately designed adjustments on the
support brackets to permit alteration when the machine is running.
It is usually advised that the leading edge of the first forming board is
initially set a fraction of an inch (kin.) below the line formed between the
tops of the breast roll and first table or grooved roll, while the trailing
edge is set on this line. This allows for the wire wrap round the breast roll
and ensures that the board is set somewhere near parallel to the run of the
wire. When the machine is in operation final adjustment is made and
229
3B.14 THE WIRE SECTION
it is generally considered best that the board and wire are in contact but
without pushing up the wire. If the board slopes down towards the breast
roll, the sheet may be marked in uneven streaks or worms in the same
way that can occur with table rolls when too much water enters the leading
side of the nip under the wire. On the other hand, if the board slopes
towards the couch it begins to act as a foil and although the suction
developed is less than with a table roll an appreciable quantity of water can
be discharged at the trailing end of the board which may produce precisely
the same sort of disturbances that the board is trying to minimize. Bad
alignment or too high a setting under the wire can have an appreciable
effect on the drag exerted on the wire so that careful observation of the
power used by the wire drive can be of assistance in setting the board
correctly.
When using forming boards the slice jet should ideally meet the wire in
the gap between the first board and the breast roll. If the jet impinges on
the wire too close to the leading edge of the board then considerable
discharge may occur at the edge due to the initial influence of the down-
ward component of the jet, and this may have an undesirable effect on the
wire side surface. If the jet meets the wire actually above the forming board
then disturbances similar to those described with a breast roll discharge
will occur. The length of the forming board section, particularly if solid,
should not be too great or the excessive delay in early drainage may induce
skating.
3B . 1 4 Relation between wire and stock velocities

Although relation between the wire speed and the velocity of the stock or
slice jet (the 'efflux ratio') is known to be of great importance in regard to
the degree of anisotropy of the paper, surprisingly little has been published
on this topic until recently. Andersson and Bergstrom (8) were in fact the
only authors to have made a specific investigation of this variable, and
their work was performed on an experimental machine with an apron slice
at speeds ranging from 300 to 600 ft./min. The relative velocity of the stock
to the wire was varied by opening up the slice and allowing stable conditions
to be reached with the same flow from the breast box but at a new level;
the machine speed and substance remained effectively steady. To eliminate
the effect of drying strains on the relative machine and cross direction
properties of the paper, the web was cut between the couch and first press,
pressed between blotters and then dried between polished plates.
Fig. 3. 12 shows results obtained at a single speed of 300 ft./min. though
these represent an average for various conditions which involved altering
the position of the apron and also the suction applied to a forming board
between the breast roll and first table roll; these latter variations had some
degree of influence on the strength ratio but not of the same order as the
relative wire and stock velocity, nor did their alteration appear to change
the optimum relative velocity. It is apparent that in the region when the
slice jet velocity is some 5 per cent. to 10 per cent. slow, anisotropy in the
sheet is at a minimum and other experiments indicated that this remained
230
THE WIRE SECTION 3B.l 4
substantially the same at higher speeds; this incidentally agreed fairly well
with determination of the orientation of a number of coloured fibres which
also showed least alignment in the machine direction with the stock
velocity about 10 per cent. slow. Judt (48) has also reported that the lowest
strength ratio occurs with the stock velocity between 10 per cent. and 15 per
cent. slower than the wire speed, though this author also states that look-
through and uniformity of fibre deposition is best with the velocities equal.
A fuller investigation has now been reported by Schroder and Svensson
(I 06) in which the effect of changing the effl.ux ratio in regard to various
paper properties was studied on the Swedish Central Laboratory experi-
mental machine and on commercial machines. Various graphs are presented
Fig. 3 .12. Influence of the relative velocity of stock to the wire speed (efflux ratio) on
the strength ratio of paper (after Andersson and Bergstrom)
and all show a turning point in the relation between the different paper
properties and the effl.ux ratio at the point where this is unity; in this work
the jet speed was calculated without allowing for any discharge coefficient
and if this were taken into account the turning points occur for the jet
speed about 4 per cent. to 5 per cent. slower than the wire (assuming
C = 0·95 or 0·96) which is in agreement with earlier work. Of those
properties tested, tensile and tear ratio, bulk, burst, and breaking length
in the machine direction had minima at this effl.ux ratio, while formation,
porosity, and cross-direction breaking length had maxima. Some of the
curves are not simple but assume a W shape with a sharp increase or
decrease in the region of the critical effl.ux ratio; this applies especially to
burst and rigidity properties, the sharpness apparently becoming greater
at higher speeds.
Andersson at;id Bergstrom also assessed that earlier and more intense
drainage resulting from alterations to the apron and forming board
conditions produced a greater fibre orientation, although the general
visual formation of the sheet appeared better due presumably to the more
efficient deflocculating shear forces operating with rapid drainage. How-
ever, if dewatering were delayed and the sheet allowed to form under
conditions of moderate drainage then, apart from a better isotropy, a
231
closer sheet was obtained as indicated by the lower air flow through the
suction boxes and higher air resistance of the paper.
From this work it is apparent that the relative velocity of stock to the
wire requires close control and this is particularly the case if the sheet is
being formed to give least anisotropy, since there are strong indications
that small variations on either side of the optimum can often have a
considerable influence on the degree of anisotropy obtained. Robertson and
Mason (81) have emphasized the importance of this point particularly
with regard to the possible influence of local variations across the web.
Using a technique for separation of chemical and groundwood fibres in
newsprint sheets they were able to demonstrate that a large proportion of
samples exhibited alternating regions of orientated and random chemical
fibre alignment on the wire side across the sheet; it is quite likely that
this was produced on machines operating close to the optimum relative
velocity point though in the absence of further details the influence of
other factors, in particular whether there was partial breast roll discharge,
cannot be discounted as a possible explanation.
3B.2 TABLE ROLLS AND OTHER DEWATERING DEVICES

The original reason for using table rolls on the Fourdrinier paper machine
was undoubtedly their convenience as a support for the wire. That they
prove to be highly efficient devices for removing water from the web and
also apparently have a useful role in assisting good formation of the sheet
could be regarded as fortuitous but highly beneficial. But table rolls are
not, of course, without their disadvantages and these become more apparent
at higher speeds.
3B. 2 1 Disturbances with table rolls

The suction forces acting in the nip of a table roll and their effect on the
wire and underside of the mat have already been discussed in detail. At
the point where· the vacuum is broken it is apparent that the wire experi-
ences sharp upward acceleration that will react directly on the fibre mat;
this produces an impulse that can have an appreciable effect on the stock
surface and produce various disturbances, in particular that known as
stock jump. Burkhard and Wrist (21) were the first to investigate the
phenomena associated with these disturbances and their observations will
now be described.
Using a high-speed cine-camera it is possible to follow the development
of disturbances as the stock passes over a table roll; Burkhard and Wrist
found by this means that any small depression existing in the stock surface
on approaching a roll increases rapidly in depth over the suction zone.
At the end of the suction zone the upward impulse on the mat produces a
pimple in the centre of the depression which grows in height to form a sort
of spout; this spout falls back relative to the stock surface as it meets
resistance from the air and this accelerates a process of breaking up into a
spray that consists mostly of fines at a consistency lower than that of the
232
THE WIRE SECTION 3B.22
breast box stock. This is observed on the machine as stock jump and,
though directed forwards, the spray of course lands on the wire behind the
point where it originated.
The type of depression in the stock surface which can produce kick-up
of this nature appears to be caused primarily by small bubbles and ir-
regularities already on the surface of the stock, and also as a result of
earlier spray landing on the wire immediately ahead of a table roll. Stock
jump can be observed at speeds well under 1,000 f.p.m. but its onset is
influenced by stock conditions and other factors. Consistency of the stock
is important and Burkhard and Wrist assessed that increasing consistency
from 0·6 per cent. to 0·85 per cent. on their experimental machine delayed
the start of stock jump by about 500 f.p.m.; a similar effect was observed
if initial drainage was made more gradual by the substitution of foils-the
mat consistency at the first table roll was then raised and this too delayed
onset of stock jump. Other effects observed in this connection were that a
temperature increase of 15 deg. F. reduced the critical speed for stock jump
by roughly 100 f.p.m., while increased aeration or a lower tension in the
wire also reduced the speed at which stock jump was first observed. The
probable explanation of these observations lies in the increase in surface
tension and viscosity associated with increased consistency and decreased
temperature; increasing wire tension will reduce deflection of the wire and
the consequent impulsive force on the mat at the point where the suction is
broken.
Mardon and Truman (50) have observed that individual ridges in the
stock surface approaching a table roll may branch into smaller ridges on
either side, while Wrist (84) has stated that even if no pronounced dis-
turbance exists in the slice jet the first table roll or foil will create a regularly
spaced set of ridges on the surface and these will be perpetuated in suc-
ceeding rolls until they eventually become smoothed out as the mat con-
sistency rises. Theoretical and experimental work carried out by Yih and
Lin (99) has shown that these phenomena are an inherent feature of the
hydrodynamic instability of a moving liquid surface subjected to vertical
forces and small changes in its direction of motion.
3B. 2 2 Other aspects of table rolls

Wrist also gives some interesting results of measurements of the power
transferred from the wire to a table roll (additional to that used to overcome
bearing friction). This power decreases down the wire, corresponding closely
to the decline in the quantity of backwater discharged, and increases with
higher speeds and with greater flexing of the wire (obtained by removing
deflectors). Wrist concluded from these and other observations that most
of the power consumed by a table roll goes in agitation of the stock.
Increased drainage is obtained from a table roll by application of a
doctor, but generally speaking this is not considered advisable due to the
probability of increased power consumption and excessive wear on the
wire which will occur if resistance to rotation causes slippage at a number
of rolls. In some cases, particularly on slow machines, a table roll may stop
233
3B.2 3 THE WIRE SECTION
completely and it must be considered poor practice not to remedy this

immediately because once a flat is worn on the roll it will become useless
for running under the wire; occasionally a roll may be stopped deliberately
with the idea of delaying drainage, but this action must obviously be toler-
ated only as a temporary expedient for it would be more satisfactory simply
to remove the roll, if nothing else.
To reduce the drag exerted by a table roll on the wire, anti-friction bear-
ings have come into general use. Rolls have also tended to grow in size
to maintain dynamic stability at high speeds; a certain critical speed
exists for any roll at which whipping and vibration occurs and this is raised
if the roll is of larger diameter and more rigid. A vibrating or eccentric
roll can be very troublesome especially close to the slice, and produces
barring in the paper.
3B. 2 3 Use of deflectors and foils

On faster machines a considerable quantity of water discharges from the
nip of a table roll and if this splashes on to the roll following it can produce
marks on the wire side of the paper where a stream of water has been
carried into the nip. Regular-spaced streaks from the curtains under the
wire carrying forward into the nip of the next roll can also affect the appear-
ance of paper.
For these reasons it is common to place a barrier between the table rolls
and this is termed a baffle, deflector, or scraper, though the latter descrip-
tion is, perhaps, rather unfortunate. Although the top surface of a baffle
must be very close to the wire to be effective, it is not good practice to have
the top so hard up on the underside of the wire that it actually scrapes the
surface. This can lead to a considerable increase in power consumption
at the wire drive, and uneven wear across the baffle will lead to a variable
carry-through of water under the wire which can produce streaks. Never-
theless, as a precaution against scraping, the surface of a baffle should be of
a low friction material similar to those used for suction box tops.
If the baffle has a comparatively wide surface and is adjusted to slope
down towards the couch with the leading edge touching the wire, then a
suction is developed in the nip. The device then becomes what is commonly
known as a 'foil' and the surface may be 1t in. to 3 in. long. Individual foils
generally extract less water at a lower consistency (loading as well as fibre)
than table rolls, except possibly when these are of comparatively small
diameter, and they also disturb the surface less; however, they require
less room than table rolls with baffles, and in an equivalent wire length a
bank of foils (usually arranged in units of three to seven) extracts a greater
quantity of water (86, 96, 97, 100, 103, 109, 111). Typical results have
been reported by Sisler and Maves (96) who substituted four foil units for
three table rolls immediately after a forming board and found, for various
substances of a letterpress printing grade, that in every case the total
drainage from the foils was greater. These workers also found by adjusting
the dip of the foils that an optimum angle (between 2 deg. and 3 deg.)
occurred at which the foils extracted most water, and this has been con-
234
firmed by Roecker (105) who also considers it preferable for the trailing
edge of a foil to be straight rather than convex.
However, Descary (103) has reported data which indicate that the
optimum angle should be nearer 4 deg. for maximum drainage consistent
with least disturbance to the fibre mat, while Hansen (111) recommends
the use of increasing angles down the wire, so the question of optimum
angle would seem to be in dispute.
One advantage of a foil provided it is not too long is the reduction in
stock jump and disturbance to the underside of the sheet at faster speeds;
over a foil any depression in the stock surface gradually grows in the same
way that occurs over the suction area of a table roll, but the impulse to the
wire at the point where suction is broken is much less. For these reasons
it has been considered likely that foils might well replace table rolls
provided wear on the wire is not excessive. The development of abrasion-
resistant plastic surfaces (especially high-density polyethylene) for foils
has reduced considerably the wear and drag on wires, and foils are now
used solely on quite a number of machines. However it is often thought, as
described in 3A. 2 3, that some table roll action is important to maintain the
formation of the sheet, and it has also been observed that the advantages of
foils diminish as they are substituted for more and more table rolls on a
particular wire section. Nevertheless, the application of foils is becoming
increasingly common in the wire section and they are now used success-
fully on many high-speed machines in the early part of the drainage area
where it is particularly desirable to reduce the amount of stock jump and
at the same time increase the overall rate of drainage. At medium speeds
the substitution of foils for table rolls at intervals in the wire section is often
particularly advantageous because the lower consistency of backwater
removed by foils results in a higher retention of loading and fines, and
some reduction in two sidedness and increases in burst and air resistance
have also been reported. The increased drainage rate and lower backwater
consistency both assist in lowering the consistency of the breast box, which
can be useful for forming the sheet more satisfactorily or retaining quality.
3B . 2 4 Other drainage devices

In previous discussions it has been emphasized that rapid drainage in the
first few feet of the wire section is generally undesirable; stock jump is
closely associated with rapid drainage and the need to minimize such disturb-
ances is a further reason for delaying drainage. Apart from the foil, two
main modifications to the table roll serve this purpose, the grooved roll
and the dandy-type roll. Of the two, the grooved roll is most popular and
is easiest to run provided the groove is spiralled and preferably terminates
sufficiently far from the roll end to avoid any possibility of catching the
wire edge. In varying degrees either device generally removes less water,
at a lower consistency, than a table roll, and their use in the early part of
the wire serves the dual purpose of making drainage more gentle, thereby
permitting the bottom layers of the mat to be formed without undue
235
38.24 THE WIRE SECTION
violence, and allowing the consistency of the mat to increase to a point

where stock jump is less likely above the first table roll.
The use of grooved rolls and the other devices is largely a matter of
individual experiment and unfortunately it is frequently difficult to draw
definite conclusions as to the effect of their use in different positions. Too
many foils or grooved rolls before the first table roll may prevent excessive
stock jump but at the expense of loss of formation. More frequently the
problem arises that the use of grooved rolls so reduces the dewatering rate
as to bring in complications similar to those encountered in trying to speed
up a machine with an inadequate wire drainage capacity. It is not possible
to cram rolls together too closely or there will be insufficient space for
water to escape between the roll and baffle. Larger diameter rolls extract
more water but only approximately in proportion to the space they need.
Experiments involving carefully controlled alterations to the various
factors which have been discussed above have been reported by Hendry
et al. (3), Duskin (18), Burkhard and Wrist (21), Parker (68), and Wrist
(84), to which reference can be made for further details.
The problem of dewatering on an existing wire part has lead recently to
the introduction of several special devices designed to increase water
removal by the application of vacuum. One arrangement (102) applies a
light vacuum of a few inches w.g. to a conventional foil unit and is claimed
to give a considerable increase in dewatering on slow stocks or high sub-
stance sheets run at relatively low machine speeds, conditions in which
foils without vacuum do not perform so well. 'Forming boxes' or 'wet
suction boxes' are a similar device except that the surface contacting the
underside of the wire is flat, as with forming boards, instead of angled on
the trailing side. Experimental work on these units (98, 107, 108) has
indicated that increased drainage occurs except when the web is very close
(greaseproof and other wet-beaten furnishes), and also an improvement is
seen in strength and other properties of the paper and there is a more even
distribution of fines and loading through the thickness of the sheet. How-
ever, earlier work reported by Parker (68) noted a deterioration in forma-
tion when this type of box was substituted for table rolls on a fast machine,
though flow on the wire was free from the normal disturbances encountered
at higher speeds with table rolls. A third device employs vacuum on a
number of conventional table rolls to which appropriately designed end-
seals are applied to retain the vacuum. This is also claimed to improve
drainage rate.
It is in fact likely that any arrangement allowing a vacuum under the
wire will increase drainage rate, but only up to the point where the vacuum
induced by the natural action of table rolls or foil units remains lower than
the vacuum applied. With fiat-surfaced forming boxes no natural vacuum
would normally be experienced, but here the time available for the imposed
vacuum to be felt can be expect to limit its effect. There is thus likely to be
an upper limit to the machine speed at which these devices will effect
useful improvement in drainage by provision of an economical and
practicable vacuum. However, when a useful improvement in drainage
can be obtained an added facility is that variation of the strength of
236
THE WIRE SECTION 38.3
vacuum can be used to adjust the drainage rate, thereby permitting reason-
able changes in wetness on the wire to be corrected without having to
alter the breast box consistency.
3B.3 SHAKE
Of all aspects of papermaking on the slower Fourdrinier machine none is
more individual to each machine than the operation of the shake. It is not
then surprising that a survey of the literature soon brings to light a con-
siderable diversity of views on the value and attributes of shake. For in-
stance, some consider that shake has most effect over the first few table
rolls, others that it is of relatively little importance until later on (hence,
incidentally, the reason for attempts to apply the shaking force to the wire
near the suction boxes instead of at the breast roll). Some authors state that
the application of shake causes water to be removed more rapidly, others
that the opposite effect occurs, while a third body of opinion holds that a
slow shake promotes drainage but a quick shake retards it. Yet again, some
consider that shake increases anisotropy of the sheet, others that it im-
proves evenness of fibre orientation and directional strength properties,
and similar differences of opinion no doubt also exist as to whether shake
serves any purpose at all.
When it comes to advice on how to operate the shake on a machine, in
particular what length of stroke to use and what frequency, it is certainly a
case of 'comment is free, but facts are expensive'. Laboratory simulations
have been attempted to throw some light on the subject but though they
lead to more consistent results within themselves their value as a guide to
what happens on the paper machine cannot be considered great. The
worker who has done most to throw a gleam of light on this subject is Judt
and what follows is largely a summary of his researches (47, 48).
3B . 3 I General effects of shake

Judt has made a direct investigation of the operation of shake in more than
one hundred experiments on six fine paper machines mostly running at
speeds within the range 100 to 300 f.p.m. The results obtained from the
mass of data collected in this work may be conveniently summarized as
follows:
(a) Evenness of fibre orientation and also, to some extent, the ratio of
strength tests in the machine and cross directions improve with
application of shake, though the effect diminishes at higher speeds;
improvement in fibre orientation is greater with wet stock and on
the top side of the sheet.
(b) Drainage of the mat with free stock is retarded by shake, particularly
in the early part of the wire, and the effect is increased by a faster
shake; for wet stock, shaking increases the drainage rate and again
the effect is greater with a faster shake.
237
(c) Formation improves with application of shake though the effect

diminishes at higher machine speeds; this improvement is greater
for wet stocks and is present up to greater speeds than for free stocks.
(d) Formation improves with increasing frequency and stroke of shake
though a point is always reached where further increase brings about
a deterioration; the shake is also more effective when the substance
is heavy and the breast box consistency high.
(e) With the stock velocity equal to the wire speed shake can cause a
deterioration in formation; with the stock velocity lower than that
of the wire by a few per cent. this does not occur.
(f) When application of shake has a positive effect on the formation,
and anisotropy of the sheet is reduced, then the density, burst, air
resistance, and average stength of the paper all increase; this effect
diminishes at higher speeds.
(g) Increasing the amplitude as opposed to frequency of the shake has
more influence in increasing the density and average strength pro-
perties of the sheet; increasing the frequency of shake has more effect
on the formation.
It is apparent from these results that there are many complications
involved in defining the effect of shake. The basic reason for this is un-
doubtedly that shake influences any particular sheet to an extent dependent
essentially on how much improvement it is possible to make; in other
words shake has the largely negative purpose of remedying the deficiencies
(particularly of flocculation) present in the stock as it flows on the wire,
and of inhibiting the development of flocculation during drainage. The
quality of the stock, in particular whether it is free or wet, and the design
of the wet-end of the machine itself must therefore have considerable
influence on how effective shake proves to be under any particular circum-
stance. At higher speeds improvements in the approach system and breast
box design, the use of shorter fibres, and increased agitation on the wire all
reduce the tendency for flocculation to occur; this, more than the mechani-
cal difficulties involved in increasing the shake speed for corresponding
increases in machine speed, has probably been responsible in the past for
shake being found unnecessary on faster machines.
Turning to one or two specific points, it is possible to give a general
explanation for most of the items listed above; attention will, however,
be confined to three of the more important effects. Firstly, for shake to be
effective in improving formation and producing a more evenly aligned
sheet it must act on the fibre mat for a sufficient time; hence, in general, the
reduced effect observed with faster draining free stocks and at higher speeds.
Secondly, when the sheet is difficult to form on the wire without floccula-
tion, e.g. when the substance is heavy or the breast box consistency is high,
then shake can be particularly effective; also improvement is likely to occur
particularly on the top side of the sheet which is normally the more floc-
culated due to the longer time the upper layers offibre remain in suspension.
Thirdly, as regards the different effect of shake on the drainage rate of
free and wet stocks, it is probable that for free stocks the rate is slower with
238
THE WIRE SECTION 38.3 2
shake due to the formation of a denser fibre mat on the wire, particularly
in the wire side layers; for wet stocks on the other hand the tendency to
compact down quickly on the wire side, producing a mat with low perme-
ability that retards drainage, is hindered by shake which tends to keep the
mat more mobile.
These points by no means explain all the varied observations presented
by Judt and, in particular, nothing has been said regarding the reason for
the different effect of shake amplitude and frequency although it is obvious
that these, by governing the relative velocity between the stock and wire at
different points, are highly important. Judt has evolved several ingenious
hypotheses to explain the varied effects he noted and the interested reader
may care to refer to the original papers for a fuller treatment of this
subject; for the present purpose, however, the author sees no value in
attempting to present explanations for observations that are not necessarily
general, nor of much practical significance.
3B . 3 2 Other observations on shake

Several of the results obtained by Judt have been observed by previous
workers and of these four investigations are selected for further considera-
tion because they contain some further points of interest.
Brauns, Bergstrom and Svenson (10) carried out an investigation into
the effect of shake on an experimental machine with an apron. A compari-
son with and without shake indicated that paper made with shake in general
had less fibre orientation and was less flocculated. Increasing the amplitude
of the shake beyond a certain point, however, produced an increased
fibre orientation in the machine direction. Both the amplitude and frequency
govern the accelerating force applied to the wire and if either is made too
great formation begins to deteriorate and a point is reached where waves
appear on the surface and the sheet is liable to become completely dis-
rupted.
Hitchings (26) is another worker who investigated the effect of varying
shake frequency and amplitude on a slow paper machine with particular
reference to the effect on drainage. He found that applying shake decreased
the drainage rate and that increasing the amplitude of the shake with
frequency constant caused first less, then more water to be removed from
the table rolls until a point was reached where the drainage rate became
almost equal to that observed without shake. Increasing the frequency of
shake with amplitude constant had a fairly similar effect. From his results
Hitchings claims that there exists a combined optimum level of frequency
and amplitude, which he terms the 'total shake speed', at which the drain-
age rate is a minimum; this condition then results in maximum strength
properties in the sheet.
An unusual investigation has been reported by Finger and Majewski (11)
who sought, by stripping off successive individual layers of fibres from a
sheet of paper, to compare the average fibre orientation at different points
within the thickness of the paper. By this means they succeeded in demon-
strating that the frequency of shake in relation to the distance between the
239
table rolls can produce an observable effect on the orientation in different

layers. If, for instance, the time taken by the wire to travel between two
consecutive table rolls is a multiple of the time for a single shake oscillation,
then the relative velocity of the stock and wire is always the same for any
particular position on the wire when it passes over a table roll. As the
average orientation of fibres is influenced by this relative motion it is
apparent that fibres will then tend to be deposited above one another in
the same general direction at each roll; this effect will produce a waviness
in fibre alignment along the length of the sheet. Other relationships
between the shake frequency and the time taken to travel between two
consecutive table rolls produce different effects on the sheet. In general a
cyclic variation in average fibre orientation exists through the thickness
of the sheet though it tends to become blurred towards the top side due
to the reduction in drainage rate, the smaller relative motion produced
by the shake, and the unevenness of the thin layers deposited.
Finger and Majewski conclude from this that certain conditions of
frequency and amplitude of stroke are critical in determining the basic
structure of the sheet both in the machine direction and through its
thickness. On any machine there will presumably be a range of speeds
where this effect becomes more prominent. At very slow speeds relatively
little extra drainage occurs at table rolls so the suction impulses necessary
to produce a predominant fibre orientation at particular positions will be
absent. At faster speeds, of course, the diminishing influence of the shake
as a whole will also reduce the significance of this particular phenome-
non.
Finally, Manson (90, 113) has presented some data on the effect of shake
on formation, as measured on a laboratory test instrument, which further
confirm the complexity and contradictions surrounding the use of this
device. In his first experiment the amplitude of the shake on a very slow
machine was reduced from y36 in. to s32 in., and finally the shake was stopped
altogether. The surprising result emerged that although overall formation
was worse without shake than at the original setting, halving the amplitude
to / 2 in. worsened the formation to an even greater extent.
Manson also investigated the effect of removing shake on a faster (900 ft.
per min.) machine. Even at this speed, at which shake is normally considered
to have relatively little effect, an observable deterioration in formation was
measured without shake. The effect was not, however, so great as was
caused by the removal of the dandy, see 3B.5 I.
A later analysis of the effect of shake on a variety of grades of paper
run at speeds ranging from 300 to 1,000 ft. per min. indicated that forma-
tion improvement is related to the product of amplitude and the square of
the frequency of shake, and is less at higher machine speeds; however, the
effect of shake was often found to be erratic and sometimes there was little
improvement in formation. The effectiveness of shake was reduced at
higher breast box consistencies or if the refining degree was increased
drastically. When improvement occurred it was generally found to give a
more even distribution of fibres and reduction of clumps in regions spaced
t in. to 2 in. apart.
240
3B .4 SUCTION BOXES
In this section it is proposed to deal primarily with two aspects of the use
of suction boxes which have a close bearing on their efficiency of operation.
Firstly, the frictional resistance and abrasive action of suction boxes on
the wire will be considered; as these affect the rate of wearing and the over-
all vacuum which can economically be applied to the boxes they are of
obvious importance to the working efficiency. Secondly, attention will be
given to the method of application of vacuum; this is of importance in
determining the flexibility of operation of the boxes although other ques-
tions such as that of ease of assessing performance and general stability
have also to be taken into account.
3B . 4 1 Frictional resistance at the suction boxes

The effect of abrasion between the wire and the suction box surfaces is
threefold: power consumption to rotate the wire is increased by the
frictional resistance, the wire is worn rapidly, and the suction box surfaces
are also worn. Generally speaking, each of these factors can be expected
to increase in degree with higher speeds, and on high-speed machines the
rate of wire wear particularly can become a difficult problem. It has been
found on faster machines that between 50 per cent. and 70 per cent. of the
total load required to revolve the wire section is due to frictional resistance
at the suction boxes. The importance of minimizing friction is therefore
clear; this topic will be considered first.
The resistance of the wire to passage over the suction boxes depends
essentially on two factors: the total load acting between the wire and box
surfaces, and the coefficient of friction. Apart from the relatively small load
due to the support given to the wire by the suction boxes, under any
particular running conditions the effective load is dependent entirely on
the vacuum applied to the boxes and so may be expected to increase almost
proportionally with the vacuum. The coefficient of friction, on the other
hand, depends on a number of variables including the materials of the
boxes and wire, the pressure between the two (i.e. the relation of the load
to the supporting or land area of the boxes) and any lubricating action
that may result from a film of water under the wire. The question of
materials is closely allied to the problem of wear and is conveniently left
till the next section, but apart from this aspect some other useful experi-
mental results relating to the friction coefficient have been reported.
Macdonald et al. (49) have carried out an extensive survey of the factors
affecting wire wear and this will be considered in greater detail in 3B. 7 1 ;
for the moment attention will be confined to that part of the work relating
to the suction boxes. In this, the load on each of the boxes of several paper
machines was determined by relating the vacuum to the open area, while
the drag on each box was found by observing the reduction in power
consumption which occurred when vacuum on the box was gradually
lowered to zero (the vacuum on the other boxes being readjusted to their
original value, if necessary). As a matter of interest no further observable
241
reduction in power consumption occurred if a box were lowered from the

wire, which confirms the negligible effect of drag caused solely by the
weight of the wire on the box surface. This technique, though simple,
gave good correspondence with a direct measurement of the drag ex-
perienced by a suction box when operating normally (this was obtained by
using strain gauges incorporated in the supports), and any change to
other running conditions brought about by shutting off one of the boxes
was not considered to be of significance.
Results from this work indicated that the coefficient of friction increases
as the percentage open area in a particular box is reduced, or if the vacuum
is reduced, i.e. in both cases if there is a lower pressure on the box. Put
another way, this implies that an increase in the load between the wire
and boxes does not give a proportional increase in the drag. The coefficient
of friction also increases as a wire ages; this is probably a direct result
of the lower pressure produced by spreading the normal load over the
greater area of wire exposed from wearing down of the knuckles. The
increase in power consumption observed as a wire ages is mainly caused
by this change in frictional drag at the boxes.
Macdonald found no difference in coefficient of friction according to
the position of the suction box, i.e. whether at the wet or dry end. Generally
wet boxes had higher friction coefficients than dry, but this was no more
than could be accounted for by the difference in pressure (the vacuum was
usually lower for one thing). It therefore appears that if there is any lubri-
cating action due to a water film, it is present fairly equally at all the boxes
irrespective of the solids content of the web or the quantity of water
removed.
All the results quoted above have also been found by Boadway (42) on
a small experimental machine. This worker observed in addition that speed
of itself did not appear to affect the coefficient of friction.
Macdonald also observed that a freshly dressed box had a relatively
high coefficient of friction until the surface became bedded in; because of
this it is prudent to take care during resurfacing to prevent any unevenness
(in particular to support the box at the ends, as on the machine, when
planing or grinding), and whenever possible to change only one box at a
time to prevent excessive drag on the wire.
From these results the useful conclusion can be drawn that the greater
the proportion of open area in a box the better, and this is so especially for
dry boxes which carry a higher vacuum; fortunately this agrees completely
with the requirement that the gap between consecutive suction zones should
be as small as possible. Thus in practice the individual slots or holes in a
box should be separated by a minimum of solid surface compatible with
the demands of rigidity. This applies especially to the cross-machine
edges of boxes where an unnecessarily wide land area represents a useless
increase in drag.
A further point that is clear is that the proportional drag experienced
by the first few boxes will be higher when these are run at a lower vacuum.
This does not mean that the vacuum may just as well be as high as in late
boxes, because the increase in total load resulting from this would more
242
than offset the slight decrease in friction coefficient at the higher vacuum
and produce an unnecessary increase in total drag on the wire; rather the
implication is that the vacuum in the first few boxes should not be reduced
too much, particularly if this leads to the need for greater vacuum in later
boxes. The staggering of vacuum should, in other words, be gentle.
3B. 4 2 Abrasion of the wire and suction boxes

Abrasion of the wire and the box surfaces causes wear to both, and in
selecting appropriate materials the rate of wear in each case must be con-
sidered. A suction box surface material that offers a low coefficient of
friction and low rate of wire wear is no use if the material itself wears away
rapidly and requires frequent dressing and replacement. The rate of wear
of both wire and boxes can be expected to increase with the vacuum applied
at the boxes and, in fact, with any other factor which increases the drag.
The quantity of abrasive material in the stock, especially loadings and grit
coming from the grinding stones used in mechanical pulp, also affects the
rate of wear; this applies especially if the material can become embedded
easily in the box surfaces and also, according to Boiteux (24), the abrasive
effect depends (at least for loadings) mainly on the particle size. If wear is
uneven due to the boxes being out of line or tilted, or particularly if
uneven wire tension is allowed to wear hollows into the box surfaces,
then frequent resurfacing will be necessary, the vacuum may be lost, and
it may also become difficult to keep the wire straight. It is now, of course,
common practice to avoid the possibility of grooving from individual
wire warp strands by oscillating the boxes laterally, though a more modern
alternative is to use a wire guide which gives the wire a continual positive
oscillation from side to side. Another device which is claimed to avoid the
necessity for frequent dressing of box surfaces and to ensure a good seal
involves the use of flexible hollow ribs which are pressurized with water
or air. These ribs form cross-machine slots and can be renewed by sliding
out of a retaining groove.
The most common material for suction box surfaces for many years has
been hardwood, particularly end-of-grain maple, which may be wax
impregnated. Under normal conditions with the usual phosphor-bronze
wires this has given a tolerably long life, but with increasing machine
speeds the relatively high coefficient of friction of this material causes the
frictional drag and rate of wire wear to become excessive. Much research
has been conducted into finding a more appropriate material and many
different covers are now on the market. Some reports of changes in cover
material on individual machines are available, but the most thorough
assessment has been given by Lawson and Lambert (78), who have carried
out extensive laboratory and machine comparisons. Rubber covers, for
instance, apparently wear less than hardwood but have too high frictional
drag; plastics like nylon and teflon (fluon) are more satisfactory, while
overhard boxes made from steel, ceramic, glass or aluminium are likely
to cause too much wear to a metal wire. The results of Lawson and Lam-
bert indicated that high-density silicon carbide covers give the best all
243
round performance; of all the materials they tested these apparently had
the lowest coefficient of friction and wearing effect on the wire, while
lasting longest and also having a good resistance to accidental damage.
However, this material is very costly, and an alternative which does not
require the same capital outlay is high-density polyethylene. But it would
be unwise to generalize and conditions are likely to vary substantially from
one machine to another, necessitating individual trials to determine the
most suitable from the many materials now available commercially.
In theory the most economic choice of material would minimize over a
period the combined cost of labour and replacement of both the wires and
suction box surfaces, together with the average power consumption re-
quired to overcome drag between the two. On fast machines particularly
this implies, more than anything else, the necessity of reducing the friction
coefficient and keeping wire wear and the frequency of box dressing to a
minimum. Plastic, chromium-plated or stainless steel wires will no doubt
require different suction box materials to those most suitable for phosphor
bronze wires, but at least for faster machines it seems likely that specially
developed surfaces will soon gradually replace the traditional materials.
Whatever material is used it is also, of course, essential that it can be
resurfaced satisfactorily and is sufficiently rigid to permit a low ratio of
land to open area.
38. 4 3 Rotary belts over the suction boxes

One method of overcoming the problem of abrasion between the wire and
suction boxes is to interpose between the two either a perforated or slotted
neoprene belt or a coarse plastic fabric which runs in contact with the wire;
this is supported separately and is designed and lubricated to minimize
friction and wear over the box surfaces. The belt types are the only ones
about which there have been any operational reports and these have
invariably stated that much less power is required to rotate the wire and
that wear on the wire is greatly reduced. Generally installations appear
to be placed over boxes at the dry end because it is in this region that the
higher vacuum used produces most drag.
Although applications at quite high speeds have been reported, it would
seem that difficulties can occur in trying to keep the belt rotating satis-
factorily without flying off. Also it is apparently not always an easy
matter to guide the belt and the wire (which usually drives the belt), while
a special tray may be needed to catch water thrown out of the belt perfora-
tions. Even so, Mills (60) has estimated that the capital cost of a Rotabelt
rotary belt unit on one high-speed machine was recovered in eighteen
months to two years as a result solely of the reduction in wire costs, and
Friese (54) has also given a satisfactory report for a slower machine.
Application of the Beloit Flo-Vac model has been described by Hill and
Clark (65) and Smith (82) and similar advantages are claimed.
Smith, in particular, gives a detailed account of the problems encountered
in substituting a rotary belt for three 'dry' suction boxes and making it
operational. The eventual reduction achieved in wire wear (40 to 50 per
244
cent.), and lower power consumption required to drive the wire section
(30 per cent.), made the device particularly beneficial in this case because
originally it was a difficult problem on the machine concerned to obtain
an adequate wire life. Experience with this particular model lead to
replacing the rollers originally holding the belt with stationary guides:
the belt itself was made from neoprene reinforced with nylon cord, while
teflon surfacing appeared appropriate for the suction box covers. Guiding
the belt, and the appearance of fine wet streaks on the wire, gave consider-
able trouble at first until the lubricating water used to float the belt over
the boxes was reduced in quantity and the application method modified to
give a more even spread of water across the belt. Life of the belts up to the
date of the report was only five months and replacement was generally
necessary due to wear and cracking at the edges; it was, however, expected
to improve on this.
One final point should be noted with regard to the rotary belt. Although
several general statements have been made that water removal is improved,
so far as the author is aware no figures have been quoted to illustrate this.
The general assumption is that since the use of a belt permits a much higher
vacuum to be applied (12 to 15 in. Hg.), then water must be removed from
the sheet more effectively. But it has been seen earlier that the scraping
action at the trailing end of the open areas in a suction box is responsible
for removing probably the majority of water, certainly on faster machines.
This scraping action on the underside of the wire is absent when a rotary
belt is used and is replaced by a scraping action between the belt perfora-
tions and the suction boxes which cannot be expected to produce the
same effect. It would, therefore, be interesting to learn by direct comparison
whether in practice the greater vacuum that can be applied through the
belt perforations offsets a probable reduction in the scraping action
effect.
A different method of overcoming abrasion between suction boxes and
wire is to remove the boxes altogether and substitute hard-rubber-covered
rolls underneath which run soft-rubber rolls in contact along the full
length, the 'Rolvac' system. Vacuum is then applied to the volume formed
by the three rolls and the wire, with an appropriate sealing deckle at the
edges. As with rotary belts, the absence of scraping action under the wire
will diminish the efficiency of this device unless high vacuums are used,
but there is no doubt that wire life is extended to a high degree.
3B. 4 4 Application of vacuum to the suction boxes

On older machines suction is applied directly to the upper section of the
box through a pipe connected to the underside in one or more places.
But nowadays it is more common for the box to be divided lengthways by a
plate in which there are several large connecting holes or slots, and suction
is then applied at one or both ends of the box. Alternatively the box may
be divided only at the ends to carry the deckle seals in the upper compart-
ment, and incorporate in the body of the box a number of sloping baffles.
Whatever the design the object is to ease removal of water from the box,
9 245
give a more even suction across the web, and permit individual boxes to be
taken out quickly and easily.
The method of applying suction to the boxes varies considerably but
particularly on older machines a popular arrangement involves connection
of each box to a manifold which leads to a large separator tank to which
suction is applied by the pump. Regulation of the vacuum in such a system
broadly follows one of three ways: manual operation with release valves
usually at the back side of the machine, together often with a fine adjust-
ment admitting air through cocks at the front; by a spring-loaded vacuum
release mechanism which in effect sets an upper limit to the vacuum in the
system determined by manual setting of the spring pressure; or by a
controller which maintains vacuum at the desired level, for example by
letting more or less air into the system. The first system is the crudest and
allows only a rough setting of the vacuum which furthermore tends to
become unevenly distributed across the web especially when air is admitted
at the front side. With the spring-loaded release a more sensitive setting
of vacuum is possible though in this case, as in the previous arrangement,
considerable quantities of air are likely to be sucked from atmosphere and
this increases pumping costs.
The spring-loaded mechanism is often used largely as a precaution to
put an upper limit to the vacuum that it is possible to apply. In this event
the vacuum will, to a certain extent, vary with the closeness and solids
content of the web, i.e. the less porous the web and less easily air is sucked
through, the greater will be the vacuum applied. In this case, therefore, a
compensating action exists in which stronger suction is applied to a web
which is damper or more resistant to dewatering, thereby reducing varia-
tions in solids content at the couch; this is probably the most convenient
system on slower machines running well beaten furnishes and heavier
substances. The fully controlled arrangement on the other hand keeps the
operation steadier, particularly with regard to drag on the wire and hence
draw at the couch, and permits the vacuum to be set and kept precisely at a
level which is most economical from the point of view of wire wear versus
drainage capacity; for these reasons this system is more suitable for faster
machines.
Drop-legs can be placed on a box to assist direct removal of the water
before it enters the separator tank manifold; this requires sufficient depth
available by the side of the wire for the height between the box and the
seal-pit level at the bottom of the drop-leg to be greater than the equivalent
water column of the highest vacuum applied, otherwise the boxes will
flood. Using a manifold and separator tank with assistance of one or more
drop-legs is a relatively trouble-free method of applying suction to the
boxes so long as adequate precautions are taken to maintain the water
level in the .tanksteady and, of course, below the level of the manifold
pipe inlet and the boxes. This requires either a simple level control device
in the separator tank, or a level sight glass which is used to adjust manually
a valve on the discharge side of the extraction pump. One disadvantage of
this system is that observation or measurement with any accuracy of the
flow from individual boxes is not possible. Also, since it necessitates throt-
246
tling the suction inlets, the vacuum on individual boxes cannot easily be
adjusted without risk of flooding the box. Uneven application of suction
across the machine is likely if the valve on any particular box is partially
closed, and oscillations in vacuum which give rise to instability of operation
are possible.
In more modern installations these disadvantages are overcome by
separate application of suction to each box. This requires a drop-leg on
each box (usually at the back side of the machine but occasionally if the
flow is large in the first one or two boxes a second leg may be placed at the
front side for assistance); the maximum suction available is fixed by the
height of the drop-leg but this restriction can, if necessary, be overcome by
evacuating the receiver tank at the bottom. A suction inlet is connected
vertically above the drop-leg, with the pipe from the suction box joining
at right angles, and the individual suction lines connect to a common
manifold joining the vacuum pump. Suction on each box is then individually
controllable either by a straight-forward valve in each suction line together
with a main vacuum controller in the pump, or, as in the Broughton
system, by a more elaborate control mechanism which serves to keep losses
in the vacuum system to a minimum (in some cases a separate suction box
vacuum pump can be dispensed with), and is claimed to have greater
stability.
The drop-legs require careful sizing: too narrow a diameter will cause
the leg to back up and flood the box, too large and it gives no assistance
to the vacuum pump which must then keep a greater volume evacuated
to maintain the required suction. Arrangement of the drop-legs can be in a
variety of ways but the most useful system involves dropping each leg
into an individual compartment; the flow from each leg can then be
arranged to pass over a weir into a common pit thereby giving a visual
indication and permitting, when necessary, easy measurement or even a
record of the flow from each suction box. A further advantage of this
arrangement is that separate use can be made of the water flow from the
first one or two boxes to maintain the main backwater pit level, leaving
flow from the remaining boxes to enter the whitewater system in the normal
way.
3B. 5 THE DANDY

The function of a dandy is twofold: to improve the formation of the sheet
and to give a watermark when required. Stated thus the matter appears
simple, but it is no exaggeration to say that used incorrectly or inappro-
priately the dandy can have both an adverse effect on the quality of the
sheet and cause a great deal of downtime. Consideration will be given
first to the effect that a dandy running satisfactorily has on the sheet, and
this is followed by a brief discussion on the best way of achieving this end
and on the special requirements for watermarking.
3B . 5 1 Effect of using a dandy

It is generally agreed that running a dandy in the correct position down the
247
wire brings about a considerable improvement in the appearance of the

top side of the sheet. This is particularly true with longer-fibred stock made
on slower machines where flocculation of the stock in the upper layers of the
fibre mat is more likely to spoil the sheet; the dandy helps to deflocculate
these upper layers and spread out any thick lumps that protrude on the
surface due to uneven deposition and drainage.
A further effect of the dandy is to consolidate or compress the sheet.
When a dandy is lowered on the wire the dry-line will generally recede
down the machine; this indicates that the sheet has been closed up making
dewatering at the suction boxes more difficult. One result of this should
be to improve the wet-web strength at the couch (provided operation is ad-
justed to attain the same moisture content) and to reduce porosity of the
paper, though it is doubtful whether a significant increase in final strength
also occurs.
Despite the common use of a dandy it is only relatively recently that any
quantitative data on its effect on the quality of paper have been published.
Manson (90) has compared the formation (using a laboratory instrument)
of paper made with and without a dandy; in each case samples at various
positions across the web were measured and the experiment was carried
out on different grades from five paper machines covering a speed range
from 130 to 1,300 f.p.m. In every case overall formation proved worse
without the dandy, though general uniformity across the web was un-
altered. The maximum effect appeared to occur for variations in formation
having an amplitude of between tin. and 1- in.; large-scale variability was
relatively unchanged by the use of the dandy. This work very satisfactorily
confirms the general impression that the dandy 'closes up the sheet'.
An extension of this investigation at higher speeds has since been
reported by Luhde (112). This involved assessing the effect of a dandy on
an experimental machine running a coating base grade at four speeds,
in steps from 1,200 to 1,800 ft. per min. Formation was measured with
the same instrument, and in addition the sheet was split into four sections
and photographed in transmitted light. Comparison of results obtained at
the different speeds indicates that the dandy caused a definite improvement
in overall formation, though the effect diminished at higher speeds and at
1,800 ft. per min. had virtually disappeared. Porosity of the sheet was
reduced with the dandy; vacuum on the suction boxes increased, as did
fines content in the suction box water. The top layer of the sheet showed a
marked change with the dandy down, the visual denseness increasing,
fibre clumps becoming less well defined, and a slight increase taking place
in the proportion of fines and loading (the latter being already some three
times greater than on the wire side). The presence of fibre clumps was
always more pronounced in the top side compared to the wire side, though
with increased machine speed the top side became as well dispersed as the
wire side. Luhde interprets these results as giving evidence of a lateral
squeezing action in both machine and cross directions, and the general
benefit of the dandy in compacting the sheet and improving appearance on
the top side of the sheet is confirmed by this work. However, a disadvantage
to using the dandy appeared in the development at higher speeds of a
248
THE WIRE SECTION 38.5 1
prominent wire mark. This in fact became so serious that at 1,800 ft. per
min. the mark even penetrated into the wire/middle section. Luhde assessed
that any improvement in formation the dandy gave when run at a speed
over about 1,400 ft. per min. may well be outweighed by deterioration in
appearance caused by the intensity of the mark.
It is highly important that the dandy is correctly positioned in relation
to the solids content of the web. It is normally placed on the web in front
of the dry-line to ensure that there is a water film on the surface. If the
web is too wet the dandy, particularly when heavy, will tend to press too
Fe:.t..T ---OA.~OY
WIR.C.
So -----No DANDY
-:::i
I
I
I
I
lo L
1.,
L.
o.____._~__._~__._~.....___,.___,_~_.._~_.L....~_.____.
o 5o too
% OF e,ASIS WE-IGMT
Fig. 3.13. Distribution of loading through the thickness of a paper made with and
without a dandy (after Groen)
deeply into the mat; this produces a bulge in the surface of the sheet on
the leading side of the dandy which disrupts the upper layers, spreads the
whole sheet outwards, and at higher speeds produces a heavy spray which,
especially with coloured papers, marks the sheet. On the other hand, if the
web becomes too dry the dandy will have less and less effect.
One aspect of the importance of correct solids content of the sheet
beneath a dandy has been illustrated by both Hansen (1) and Groen (75)
in their work on the distribution of loading through the sheet thickness.
With an appropriate film of water on the mat surface a typical change in
the distribution of loading produced by a dandy is shown in Fig. 3. 13; the
concentration in the uppermost layers is substantially increased at the
249
38.5 2 THE WIRE SECTION
expense of lower layers (even to the extent of becoming higher than the
percentage of loading in the breast box stock), an effect due presumably to
the results of compression. If, however, the dandy is run at a position
further towards the couch, then this change in distribution becomes much
less noticeable, showing that the dandy has less effect on a dry mat.
Fig. 3 .13 also shows, incidentally, that by increasing the loading concen-
tration on the top side (which has already a higher percentage than the wire
side) the dandy accentuates two-sidedness in composition; with some
papers the effects of this could represent a distinct disadvantage.
The weight of the dandy creates a slight sag in the wire which will
increase wear at the leading edge of the next suction box after the dandy.
For this reason a dandy is preferably supported either by a single or a
couple of special table rolls inserted between the suction boxes, or above
the solid portion of a suction box. The position or pressure of the dandy on
the wire particularly at higher speeds must be finely adjustable to give
just sufficient contact to drive the dandy. Excessive pressure creates an
unnecessary disturbance in the upper layers of the sheet, producing a similar
effect to running the dandy with the web too wet. Too light a pressure
results in a speed differential between the dandy and the wire, which will
scuff the surface and completely spoil the appearance of the paper
3B. 5 2 Running the dandy

The problems involved in running the dandy satisfactorily are nearly all
accentuated by speed but some, apart from the watermarking aspect, can
be present at all speeds. On all machines the dandy tends to pick up fibre,
especially at the coarse deckle edges, and this gradually builds up to clog
the mesh of the dandy (which is normally coarser than that of the wire).
Pitch can act in a similar manner though this is less easy to observe on the
dandy. A second problem is caused when froth is generated, creating
bubbles of air on the surface of the stock and the dandy itself. Either of
these effects produce familiar thin and transparent marks on the top side
of the sheet, either singly or in the form of long worms.
These defects are usually obviated by blowing one or more steam and
water sprays through the dandy, probably with extra force at the edges, and
draping a piece of felt as a wiper on the top of the dandy. A tray may also
be used to collect the fibre and froth dislodged from the dandy but it may
still be necessary to lift the dandy at intervals for cleaning with a suitable
solvent and jetting with steam and water. Because the action is more gentle
and the dandy mesh digs into the sheet less, picking and bubbles are not so
likely to occur with a larger diameter running with only a light pressure on
the mat. But if the dandy is run too dry or the mesh is too fine then,
especially with a laid dandy, picking will become very troublesome.
On faster machines the problem of running the dandy fast enough to
prevent scuffing of the top surface and the production of excessive spray
becomes acute. To bring the dandy up to speed before lowering may be
difficult enough in the first place but to keep the running speed close to
that of the wire may require such a heavy pressure as to disrupt the mat
250
THE WlRE SECTION 3B.5 3
anyway. The only adequate solution to this problem is to provide a separate
drive for the dandy, and for the larger shaftless type normally in use on
faster machines this is most conveniently arranged at the back side of the
machine. Either a timing belt running round an appropriate gear on the
end of the dandy journal is used, or a direct drive an to one of two revolving
rollers which also serve to support the dandy. The drive requires a flexible
connection to permit vertical movement of the dandy. Control of the
relative speed of wire and dandy should be reasonably fine and is accom-
plished with a variable speed drive controlled either from tacho-generator
references on the wire drive and the dandy motor, or by governing the
torque transmitted to the dandy. It may then be found that the best results
in formation and watermark are obtained with the speeds the same, but
it is possible to find an improvement with the dandy running faster or
slower than the wire. There will, of course, be an interaction effect between
the speed differential and the pressure of the dandy on the wire so it is
advisable at the same time to measure and control the latter. This is usually
done indirectly by using the relieving pressure of air in a cylinder or
Airide spring arrangement. Adjustable stops restrict the lower limit of
movement of the dandy to a position a few thou clear of the wire.
3B. 5 3 Watermarking
Under some circumstances it can be very difficult to obtain a good
watermark, though generally speaking all the factors mentioned above as
contributing to satisfactory running of the dandy have a similar effect on
the watermark quality. Thus, the mark has a watery appearance and is
distorted by the dandy sitting too heavily on the mat, running too slow, or
if the mat is too wet; on the other hand if the mat is too dry or the pressure
of the dandy too light then the mark is indistinct.
Certain other aspects of watermarking require special consideration.
In the first place a watermarking dandy is generally more likely to create
trouble with picking and bubbles at the places where the marks themselves
protrude, especially when using an intricate or heavy design. If this cannot
be overcome by raising the dandy or by more efficient steam spraying while
running, the quality of the mark deteriorates and frequent stoppage of the
dandy for cleaning will be necessary.
The type of furnish and beating also has an important bearing on the
quality of a watermark and in general shorter fibres such as those produced
in well-beaten esparto yield the best results. Loadings, to a certain extent,
are also beneficial. What seems to be required is a reasonable proportion
of fine fibres or particles which can be compressed through the upper
layers of the fibre mat by the watermark to a greater extent than by the
remainder of the dandy mesh, thereby creating an optical difference in
reflectivity through the sheet.
Watermarking to register requires precise matching of the dandy and
wire speeds coupled with careful control of shrinkage of the sheet in both
directions. The first requirement is achieved automatically on slower
machines provided adequate pressure of the dandy on the wire can be
251
obtained and the frictional resistance to rotation is kept low. On faster

machines it becomes more imperative to have a dandy drive and a measure-
ment of the differential speed.
Controlling shrinkage is a more difficult matter, depending as it does on
so many factors such as wetness of the stock, the tightness of the draws
and felts, and the general characteristics of the sheet. In the long run
better control all through the stock preparation and machine, and repeat-
ability of operation from one making to the next, are the only means of
obtaining a satisfactory register with reasonable speed and accuracy.
Otherwise if the register becomes too narrow in the machine direction it is
necessary to tighten draws, free the stock, or slow the dandy either by
altering the drive speed, lifting the dandy slightly, or, if neither of these are
possible, the edges of a dandy free-running on the wire can be built up.
If register becomes too narrow in the cross direction, tightening the dry
felts will usually reduce cross-direction shrinkage; slackening the draws
will also help though this, of course, is also liable to affect the machine-
direction register to a certain degree. Excessive fibre alignment in the
machine direction increases cross-direction shrinkage and it may be possible
to effect an alteration to this. Lack of uniformity of watermark across the
sheet is, however, more commonly associated with the differential shrinkage
caused by a general unevenness in cross-machine substance and moisture
at the reel-up, which requires a more extensive investigation to remedy.
3B.6 COUCHING
There are three principal methods for couching the web from the paper
machine wire: the jacketed top couch and plain bottom roll, the suction
couch usually with presser roll, and the plain or suction felt transfer.
Each method has its own variations and refinements and these will now be
described in turn.
3B . 6 1 Jacket couch
The jacketed top couch is the oldest type and derives directly from the
practice in hand-made paper of couching by pressure on to a piece of
dampened felt. The normal operation involves running the jacket roll
slightly set back above a solid or grooved brass or rubber covered wire roll,
which is generally also the drive roll. The top roll exerts a slight pressure
on the wire and this serves the dual purpose of removing some water from
the sheet and, more important from the point of view of couching, provides
the essential water film between the wire and the fibre mat. Some water
passes through the wire and escapes down the wire roll, similar to a plain
press nip, but probably the majority enters the jacket to be squeezed out
at the squeeze roll or guard board.
Once the sheet has passed under the jacket it is important for ease of
couching to remove it from the wire as soon as possible. The longer the
sheet remains on the wire the more water is reabsorbed from the wire
mesh and the greater becomes the adhesive force which it is necessary to
252
overcome. The usual configuration with this type of couch results in the
sheet being drawn from the wire well down the drive roll; the position of
the jacket roll prevents couching the sheet too close to the jacket nip
except with a very tight draw and this is undesirable for many reasons
discussed earlier. It is suggested that Fig. 3 .14 shows the most preferable
arrangement permitting early take-off at a reasonably high angle. In this,
both the lead roll and first press felt roll are higher than normal; adjustment
of the former serves the purpose of guiding the path of the web to enable
J'ACKETl!!D -
Couc~ Ro&.."-
_o______ ----
- l w'u~e,tt"
Fig. 3.14. Preferable arrangement of open draw with a jacketed top couch roll
the take-off point to be as close as possible to the jacket nip with the
draw slackened back during running as far as seems advisable. Transfer
of the tail at the speeds associated with this type of couch would normally
be by hand, possibly with the aid of a lump of wet broke.
The most difficult feature of this type of couch is the operation of the
jacketed top roll. The load exerted by the roll must be light, otherwise
crushing will occur and the wire, which is only partly supported underneath,
may be subjected to strain; for these reasons a large roll is preferable as
this spreads the load and reduces the pressure on the mat and wire.
Nevertheless the wire usually has to drive the top roll so sufficient pressure
is necessary for this purpose; to assist in this end roll bearings have to be
kept as free as possible but· even then at higher speeds running the roll in
this way without crushing becomes difficult. A grooved bottom couch roll,
which inakes removal of the expressed water easier, is one way of helping
to prevent crushing but this has other disadvantages. A table roll is always
necessary between the suction boxes and the top roll so that the wire does
not rub on the trailing edge of the last box.
A guard board is frequently used on top of the jacket to squeeze out
water and conduct it to the side of the machine, but this can cause undue
wear on the jacket and produce a heavy breaking force on the undriven
253
top roll; also it is not easy to adjust pressure on the ends to keep the jacket
level. For these reasons the board is commonly replaced, particularly
nowadays on faster machines, by a rubber squeeze roll. This roll serves
the same purpose but has a gentler action; also a greater pressure is
possible than with a guard board so that more water can be removed before
a point is reached where the jacket tends to loosen and ruckle up into
creases. The roll requires a certain amount of camber to offset the pressure
applied and a doctor to keep the surface clean. Occasionally an emergency
board is still used in front of the squeeze roll; it is placed just clear of the
jacket but may be quickly lowered if the sheet goes up the jacket-it is not
a guard board and only serves the function of saving the jacket from being
spoilt by the sheet being pressed into it at the squeeze roll.
Keeping the jacket in satisfactory condition can be a problem, especially
where it tends to pick up fibre from the deckle edges. A spray could be
necessary to prevent this but even so the jacket becomes plugged eventually
and the nap gets worn. When this occurs the sheet will tend to leave the
wire and follow the jacket, making it necessary to stop the machine to
clean the jacket and (though some do not recommend this) rub up the nap
with teezers. These difficulties become greater at higher speeds, of course,
and the substitution of a rubber presser roll may be required. But this
reduces appreciably the water removed and is normally associated with a
suction couch.
3B . 6 2. Suction couch roll

According to Hendry (19) the original invention of the suction couch was
intended as a replacement of suction boxes to minimize friction on the wire,
but for some unknown reason it was first substituted for the couch. This
apparently fortuitous application has had far-reaching effects for there is no
doubt that the suction couch has many advantages over the plain jacket
couch, especially on fast machines. There is good evidence to indicate that
it removes more water but possibly the main advantage is the reduction of
accidents to the wire and the avoidance of trouble with jackets. The suction
couch roll is frequently also the drive and turning roll for the wire which has
the added advantage compared to a solid roll that the suction helps to re-
duce slip of the wire. However, it is becoming more usual on faster machines
to have a forward roll to provide all or most of the drive requirements. On
machines with a felt transfer from the wire, the forward roll is in general use
anyway.
In operation the vacuum in the suction box, aided in most cases by a
presser roll, draws water into the couch holes; after passing the trailing
seal-strip, air entering the holes from the inside throws the water outwards,
to be largely contained by the wire then thrown out subsequently by centri-
fugal force at the point where the wire leaves the roll. This throw-out is
not usually troublesome though occasionally when the wire follows well
round the roll a deflector plate may be required to prevent the spray hitting
the underpart of the wire and carrying into the couch nip. A light rubber
doctor can also be helpful for reducing the volume of water carried round
254
on the roll surface into the nip. On most machines it is unlikely that much
water is drawn through the holes into the vacuum pump.
Once the sheet has passed over the suction area reabsorption of water
from the wire will begin immediately, reducing the solids content of the
web and increasing adhesion to the wire. As with the jacket couch, there-
fore, it is important to remove the sheet from the wire as early as possible,
though in this case a compromise must be reached to reduce the appearance
of shadow-marking which, as described in greater detail in 3A. 42, dimi-
nishes in intensity as water is reabsorbed. The rate of reabsorption and
reduction of shadow-mark must be primarily a function of the web
porosity; where shadow-marking cannot be removed satisfactorily by
0 \---~1R$r PR~ss
lE..oi..O \ FEl.."'T'
RO\.L \
\
\
- ----
\
\
\
Fig. 3.15. Preferable arrangement of suction couch and pressure roll with an open
draw
permitting the web to remain longer on the wire, removal of the presser roll
and reduction of the vacuum in the couch both help though at the expense
of reduced water removal at the suction couch.
When the shadow-marking problem is under control and it is possible
to effect earlier removal from the wire, then it is not necessary to alter the
usual couching geometry by raising the lead and felt rolls as in the case of
the jacket couch. Instead the suction box may simply be moved round the
roll until the angle between the trailing seal and the take-off point is brought
to the minimum tolerable for shadow-marking; even with a reasonably
high angle of take-off this angle should normally be sufficient not to cause
any problem in positioning the presser roll, nor in blowing the tail over
during feeding up, see Fig. 3.15. The position of the lead roll close to the
couch is important to prevent excessive movement of the take-off line; if
the draw were tightened sufficiently for this line to fall back to the suction
box zone, picking occurs due to the uneven pull on hole and land areas.
When driving the wire section the suction couch roll is made relatively
large in diameter to reduce the speed of rotation and the radius of tum of
the wire round the roll, both of which contribute to keeping down the
255
degree of slippage of the wire relative to the roll and affect wear on the
wire, as will be discussed in 38. 7 1. Holes are countersunk to help apply
the vacuum as evenly as possible to the sheet and pitched to maintain the
hole area over the suction box as constant as possible during rotation,
thereby reducing the possibility of vacuum oscillations. A fresh-water
spray inside the roll should be of the high-pressure, low-volume type,
preferably oscillating, and set carefully to give even application across the
length of the roll; the function and operation of the spray is essentially
similar to those used in suction presses and the precautions necessary for
efficient use are fully dealt with in the section devoted to that subject.
This applies also to setting of the edge deckle and maintenance of the
suction box seals.
Some suction couches are equipped with a double box; the first compart-
ment is wide and under a relatively low vacuum (10 in. Hg.), having the
object of removing air from the couch holes and sealing the sheet; the
second compartment is the same as the normal box though rather narrower.
No data has been made available on the relative benefits of this design
though prior evacuation of the holes, because of their high volume, may
well lead to power savings depending on the arrangement of vacuum
equipment on the machine.
The function of the presser or lump-breaker roll is to close up the sheet
at the suction couch and raise the vacuum to improve the dewatering
action. This increases the wet-web strength at the couch and is also generally
reported to reduce draw variations. Particularly when a dandy is not in
use the presser roll is also thought to reduce the frequency of breaks at the
couch by removing lumps in the web and reducing picking. The roll is
usually situated either over the leading seal of the suction box or, with a
double box, over the middle of the first compartment.
Although the presser roll is undoubtedly advantageous it must be
recognized that smooth operation may not always be easy to attain. The
rubber should be soft (200-250 P. & J.) and pressure of the roll on the
wire, allied to a suitable camber, requires careful setting; if the roll does
not extend over the trim (a precaution sometimes adopted to prevent build-
up of fibre on the roll at the deckle edges) then ridges in the wire may gradu-
ally appear opposite the ends of the roll. To keep the roll clean it is always
necessary to have a fine spray skimming the surface on the up-going side
and this is preferably of warm fresh water. If this spray is allowed to become
uneven in application moisture variations across the web are observable,
and if the volume of water becomes too great the whole value of the presser
roll in improving dewatering at the couch may be lost.
On faster machines it may be found necessary to equip the presser roll
with a separate drive either for lowering the roll at an adequate speed or for
satisfactory continuous running. Particularly under such circumstances, as
also on the many machines normally running with a presser roll which is
very troublesome to operate due to fibre picking or repeated damage to
the wire, it may well be found that lifting the roll only reduces the vacuum
on the couch by 2 in. or 3 in. Hg. and this may have little effect on the
moisture of the web, at least after the presses. Provided the frequency of
256
breaks at the couch were effectively unaltered there would in this case be a
strong argument for removing the roll altogether.
3B. 6 3 Operating the suction couch

The suction couch assists in the action of couching by providing a film of
water between the sheet and the wire, though the occurrence of shadow-
marking reduces its efficiency in this respect; before the sheet can be
removed some reabsorption of water back into the sheet must be permitted
in order to equalize the moisture. For this reason the whole value of the
suction couch has at times been questioned. In particular, Baggallay (17)
has reported that he placed a small suction box in front of the suction couch
of a medium-speed machine and, by moving forward the lead roll over the
wire, arranged to remove the web from the wire at that position; the solids
content of the web showed no reduction despite the fact that the suction
couch was then not in use. This finding has lead to development of the
'Baggallay box' as a cheap and simple couching device which can sometimes
with advantage replace a difficult jacket couch arrangement. The box
normally consists of a single compartment to which a vacuum of 5 to 10 in.
Hg. is applied, together with a blowing compartment or jet to assist in
getting the sheet up when feeding the tail through. In operation the web
leaves the wire just at the trailing side of the suction slot and the open draw
to the lead roll must be kept as short as possible by placing the roll just
above the box.
On the other hand Macdonald et al. (49) have reported data from one
machine where the suction couch appears to act as a stabilizing influence
on the solids content of the web at the couch; lowering the vacuum in the
suction boxes only resulted in the suction couch extracting a greater quan-
tity of water with no significant change in the solids content entering the
presses. Even if this applies only partially to other machines it could
well represent a valuable means of reducing power consumption and wire
wear, particularly since any small decrease in dryness at the couch is
partly compensated for in the presses.
The only systematic investigation of the dewatering action of the suction
couch appearing in the literature is due to Brauns and Oskarsson (5).
These workers used an experimental machine and a couch in which the
width of the suction box could be varied, as well as the vacuum applied.
In addition the air flowing through the box was varied, with other condi-
tions constant, by means of a plastic cloth which could be pressed on to
the web above a portion of the suction box. Unbleached kraft pulp at two
degrees of wetness was used and in all the experiments the solids content
of the sheet before the suction couch remained constant.
The results of this work indicate that for a given vacuum and air flow
per unit area, the width of the suction box (or the time vacuum applied
to the sheet) has little effect on the dryness, giving only a very slight
improvement with increasing width. The vacuum applied, as expected,
has a much greater influence on the dryness, see Fig. 3. 16a, and the relation-
ship appears to be approximately linear. The graph is transposed in Fig.
257
24 A.1R. FL..OW
LITRES/ s~. Meri<e .
..-)c:>o
7S
... 22 So
z ........25
"'...l
0
\J Zo
If)
0
.J
~ IS
16 ~~~~~~~~~~~~~~~~~-
IS .35 45 SS
VACUUM. cm. ~ 5 .
Fig. 3.16a. Relation between solids content of web and vacuum applied to a suction
couch with different air flows through the sheet (after Brauns and Oskarsson)
VACUUM
~ 24 Cm. 149.
,_:
z
Ii.I
.... 22.
z
0
\j 2o Za
Ill
a 22.
..J lB
0
~
1'1>
0 .So loo 1$o 200 2So
I
QUANTITY OF AIR, LITRE!'> SQ. N\ll!T"~E..
Fig. 3. 16b. The results of Fig. 16a transposed to illustrate the effect of air flow on
dryness of the sheet with vacuum constant.
3. l 6b and shows there that when a greater volume of air is passed through
the sheet for a given vacuum and suction area, i.e. when the velocity is
greater, then this also produces an improvement in the dryness; however,
the effect diminishes for higher volumes and beyond a certain quantity of
air the dryness is only imperceptibly improved.
258
The implication of these results is that dewatering at the suction couch
is closely dependent on the pull exerted within the sheet, presumably
because this governs the smallness of the capillaries which may be
evacuated. In contrast to the suction boxes, at least in the early 'wet'
positions, which depend for removal of water mainly on the scraping action
of the wire on the box, at the suction couch the passage of air through the
sheet contributes relatively more to the effectiveness of the operation. In
this respect dewatering is dependent on the use of air as a medium for
transportation of the water droplets extracted and, to a lesser extent, for
evaporation into vapour form.
From this work it is clear that suction boxes could profitably be narrower
in width on many machines. A smaller box has little effect in so far as it
diminishes the time of application of vacuum, but it will certainly decrease
the volume of air passing through the sheet; the effect of this on the dryness
will be of slight value unless the quantity of air becomes comparatively
small, but the power consumption will decrease approximately pro rata
with the air volume.
The vacuum has the most effect on the dryness of the sheet so that the
use of a presser roll to increase it by compacting the web should normally
be justified. With a fixed-performance vacuum pump a further advantage
to narrowing the width of the suction box is that, by reducing the volume of
air removed, a beneficial increase in vacuum will be brought about, though
in this connection it is important to realize that the holes themselves carry
round an appreciable volume of air that must be evacuated. This vacuum
increase alone can be expected to improve the dewatering efficiency so
there seems to be a good case for constructing the normal suction box to
make an alteration in width practicable during operation; to avoid inter-
ference with the take-off conditions it would be preferable to achieve this
by making the leading seal adjustable, not a difficult engineering operation
and one which could well permit an optimum working position to be
found for any particular set of conditions.
3B. 6 4. Felt transfer

With increasing machine speeds a point is reached where operation of the
open draw becomes uneconomical-the reasons for this have already been
considered in some detail. The avoidance of an open-draw by using a felt
to transfer the sheet from the wire, either directly ('lick-up') or with the aid
of a suction roll ('pick-up'), has therefore many advantages. The success
of this method of transfer to the presses depends on ensuring that the felt
presents a greater adhesion to the top of the sheet than the wire does to the
underside of the sheet. With a lick-up this necessitates arranging for the
face side of the felt to be extremely dense and also ensuring that a film
of water is present on the surface by using a full-width shower on the felt.
For suction pick-up a relatively open felt is necessary to allow the influence
of the suction to be exerted, though at the same time the face side of the
felt must be sufficiently closely woven to stick easily to the web. Both types
259
are generally used in conjunction with a forward-drive roll, the transfer line
lying between this roll and the couch.
Considering first the actual couching operation it is particularly impor-
tant, especially when operating this form of transfer close to the maximum
substance possible under the particular machine conditions, to ensure that
adhesion of the sheet to the wire is at a minimum. On many machines the
pick-up line is, for felt and wire changing convenience, placed much closer
to the forward drive roll than the suction couch roll; this gives the web a
much longer time to reabsorb water from the wire, decreasing the solids
content and increasing the adhesion, a process which is all the more rapid
since the wire separates from the couch roll and is, therefore, able to release
water much easier. Hendry (3) has confirmed that maximum water removal
occurs when the suction boxes in the pick-up roll and couch roll are practic-
ally aligned and, as Baggallay (17) observed, if couching is delayed too long
after the web has passed over the suction couch roll then the results can be
visible in the amount of fibre left adhering to the wire. If it is not practicable,
for engineering or shadow-marking considerations, to position the transfer
roll close up to the suction couch roll, then much benefit is gained by
using a small suction box immediately before the pick-up line.
A separate drive for a suction pick-up roll is necessary, though not usually
for a straightforward lick-up roll. Frequently with this sort of arrangement,
at least on faster machines, the couch roll is driven in addition to the for-
ward roll, and sometimes the first wire return roll may also be driven. The
trailing seal-strip in the suction couch roll should for preference be wide;
this increases the angle through which the roll turns before vacuum in the
holes is released to throw out a spray of water which must not, of course,
splash the wire or forward drive roll.
Pressure of the transfer roll on the wire requires careful setting to avoid
crushing the sheet or damaging the wire. When feeding up it is more usual
to raise the wire by moving the forward drive roll than to lower the transfer
roll, except possibly with a plain lick-up roll. A slight positive draw is
usually applied as this is found to prevent creasing or a tendency for the
paper to crepe in the nip. With a suction pick-up the deckles are set in the
roll to leave the trim on the wire; with a lick-up the trim is either blown
off the wire before the couch or taken up with the main part of the sheet
and removed from the lick-up felt by rollers or by some other means. The
edges of the felt in a pick-up arrangement are apt to get dirty and a shift
in deckle can create a problem in preventing the trim following the felt.
The special requirements of these two basic methods will now be discussed
in greater detail.
3B. 6 5 Couching with a lick-up arrangement

The type of transfer involving a straightforward lick-up felt is of limited
use and, in fact, is applied exclusively to extremely light-weight papers.
Any attempt to lick-up heavier papers involves having the felt so damp,
and the paper so dry, as to be uneconomical and operationally hazardous.
It is essential to keep adhesion to the wire to a minimum; the lick-up roll
260
THE WIRE SECTION 3B.6 6
is therefore situated above the couch roll which, for this grade of paper,
is either plain, or grooved to enable air to enter under the web immediately
after the nip formed by the lick-up roll.
A further difficulty with the lick-up arrangement is removal of the sheet
from the felt itself. It is impossible to transfer to another felt, because the
lick-up felt is so dense and damp, so the sheet must be taken through the
press section on the lick-up felt and eventually transferred on to either an
M.G. or a small-diameter pony cylinder. This requires the assistance of a
roll giving quite a heavy pressure on to the cylinder and makes the whole
pressing operation a relatively inefficient process. Further, the large
volume of water in the felt makes crushing difficult to avoid. Even so,
provided the felt can be kept with an adequate 'skin' on the surface, for
light papers this method offers some advantages.
Dixon (32) has reported comparative figures taken from two machines
making the same grade of tissue from the same furnish, one with a lick-up
arrangement and the other with a suction pick-up. He cites the advantages
of the former as follows: less felt marking of the sheet due to the closer
weave of the lick-up felt, longer felt life if it is kept properly cleaned (though
the felt is more expensive), and lower power requirements. Against this the
pick-up arrangement, being more efficient in the presses and in this case
extracting twice as much water for the same loading, yields a higher dryness
entering the drying section; also transfer from the pick-up felt is easier, as
is cleaning of the felt because it is much more open and runs comparatively
dry.
3B. 6 6 Suction pick-up

It is generally considered that use of a conventional suction pick-up arrange-
ment and transfer to a second felt at the first press nip has no advantages
from the point of view of water removal. Burnett (25), for example, found
that little water was extracted at either the pick-up or transfer nips on a
typical machine, conditions at both being governed primarily by the neces-
sity of transferring the sheet. The main advantage gained by this arrange-
ment is undoubtedly safe transport of the web.
Because of the relatively heavy quantitities of water sucked through the
pick-up felt it tends to get dirty quickly unless efficient cleaning devices are
used. This applies especially to the face side which, if allowed to get dirty,
marks the sheet and can create a problem by throwing fibre clumps on to the
wire as the felt drops down to the pick-up roll. Dixon (32) gives details of
using a variety of pick-up felts, many with synthetic content, and em-
phasizes the importance of matching this felt with the felt on to which the
paper is subsequently transferred. The outcome of many trials in this
particular case was adoption of a high terylene content pick-up felt which
does not reduce in thickness throughout its life and is usually removed
when marking becomes excessive.
Jordansson (27) has reported experimental work on a pick-up arrange-
ment and amongst other things demonstrated that an upper substance
of sheet exists beyond which transfer by this means becomes unreliable and
261
the paper is liable to drop off the felt. The point when this is reached will
depend on many factors, including the dryness and constitution (particu-
larly porosity) of the web, openness of the felt, etc.; in this work substances
of 150 g.s.m. never caused any trouble, while at 200 g.s.m. a critical stage
was being approached. Placing a felt suction box between the pick-up roll
and transfer press made the operation more reliable at this heavier sub-
stance, especially if the pick-up felt were dampened with a spray before
meeting the sheet. It was also found useful to make the sheet as dry as
possible at the pick-up line since this reduces the weight to be carried.
However, modern installations combine the pick-up roll and transfer
press, and with this arrangement the paper can be held to the felt simply
by applying suction right round the roll (see Part 4).
Another way of overcoming this problem has been proposed by Kitano
(36); this involves using air pressure to lift the sheet on to a 'post-up' roll,
which is a wire net cylinder of fairly small diameter, and thence to a felt.
A rather complicated set of air seals is needed but it is claimed that speeds
up to 2,000 f.p.m. have been achieved without transfer difficulties by this
means.
When feeding up the sheet it is useful if the vacuum in the pick-up roll
can be temporarily increased. Under normal operation the vacuum is kept
as low as is necessary for safe transfer (5 to 10 in. Hg.), otherwise water
drawn into the holes can be troublesome to remove without splashing.
Dixon (2) has discussed the problems involved in this with regard to a stack
press which presents precisely the same problem at the pick-up roll;
further discussion is conveniently left to section 4B. 3 6 et seq. on the
presses which deals with this.
3B. 7 THE WIRE AND SHOWERS

The wire is, of course, an essential part of the Fourdrinier machine process,
acting as a convenient means of transporting the web while drainage occurs
and the sheet is formed; however, the direct influence of the wire on the
paper is, by and large, confined to the size of the mesh. A fine mesh gives a
greater fibre and loading retention and less wire mark, but a slower drain-
age. In practice the best compromise is governed to a large extent by the
substance range run on the machine: coarse (55 x 50) for board and kraft
wrapper; average (64 x 52) for newsprint and printing papers; slightly
finer (70 x 60) for ordinary tissues and writing papers where loading
retention may be more important; and very fine (100 x 80) for thin
cigarette and condenser tissues. Apart from this aspect the main problem
with any wire is to prevent an excessive rate of wear and retain the same
drainage conditions through the life of the wire.
3B. 7 1 Wire wear

If run long enough and kept free from damage, the vast majority of
machine wires eventually have to be changed because of holes in the body of
the wire and cracks appearing at the edges. These may be caused directly
by mechanical faults, but they may equally represent the weakest link in a
262
wire which has been worn down to the point where the slightest damage
or misalignment will show itself. Much attention has been given to this
whole problem in recent years, with particular reference to faster machines
where wire life may become reduced to a few days, and this will now be
summarized. Paton (7), Boadway (42), Macdonald et al. (49), Friese (54),
Johnson and Gavelin (77), Lawson and Lambert (78), Redfern and Gavelin
(92), and Reynolds (93) are the main contributors from which this sum-
mary is drawn.
During the life of a wire the inside knuckles are worn and the thickness
of the wire diminishes. The rate of wear is greater at the beginning of the
wire's life because the smaller area of contact creates a higher pressure and,
therefore, a greater abrasion rate of the wire. Most wear occurs at the
suction boxes, as discussed in 38.4 1 et seq. in some detail; slip at the
driving roll and at other rolls, though probably small (Macdonald found
w
::i.
Qt
0
{}. ~Q()
rt
w
~
~
ZSo,__~--~....._~~~-'-~--~~...________,_
0 5 lo IS 2o
DAYS o""' MA.CHINE.
Fig. 3. 17. Increase in power required to drive wire during its life on the machine
no case where a difference greater than 0·5 ft./min. occurred on fast

machines exceeding 1,200 ft./min. in speed), can also be expected to
contribute to wear, particularly if bearings are stiff or doctors too hard on.
As the area of contact increases, the total drag on the wire is also raised
producing an increase in the power consumed by the wire drive, see Fig.
3 .17. Normal wear of the under-surface of the wire will cause the knuckles
to become smooth and polished; if they are scored or scratched this is due
to the action of abrasive particles and it has been shown that the rate of
wear is then on average much greater.
When the drive roll is a suction couch it has been found that the initial
rate of wear of the wire is greater. Some creep round the drive roll is
263
unavoidable because of the slight shrinkage of the wire which occurs with
decrease in tension from the pulling to the return side of the roll; this
wears the wire at a rate which is higher when the open area of the couch
is greater. Wear also appears to be influenced by the extent to which the
wire is driven by the vacuum in the suction box; if the greater part of the
change in tension of the wire round the roll occurs within the width of the
box, then the higher pressure between the wire and roll in that region
produces greater friction on the wire as slippage occurs. For these reasons,
on faster machines it appears more important to reduce or eliminate the
wire driving force exerted by a suction couch in favour of using a separate
solid driving roll; over this forward drive roll, wire wear will be less due
to the lower pressure of the wire on the solid surface and the avoidance of
additional abrasive action produced by the wire knuckles rubbing over the
couch holes. With this arrangement some drive to the suction couch is still
preferable to overcome friction at the box seals and bearings, otherwise
some slip will occur and the wire will be worn there anyway. To ensure
there is little wire creep in the suction zone it is advised that the wire laps
at least 10 deg. on either side of the box.
With increasing machine speeds the wire needs to be run at a higher
tension to prevent slipping at the drive roll; if this occurs it causes the
load to drop spasmodically and the draws oscillate. Greater tension may
also be necessary to reduce stock jump at the table rolls. When a suction
couch is used as the drive roll it has been shown that higher tensions create
greater wire wear due to the greater creep round the roll occasioned by the
bigger difference in tension at the pulling and return side. In this case a
forward drive roll becomes even more essential and automatic control of
the tension by one or other of the devices available is also desirable.
38. 7 2 Damage to the wire

Although steady wear is the most common cause for removal of wires, they
are, of course, particularly vulnerable to accidental damage, especially
when being put on the machine, and no machineman needs to be warned
that great care and close observation of the wire are needed during running.
The various parts of the section which contact the wire should be kept in
good mechanical condition to prevent the appearance of ridges particularly
when running tension is high: thus, rolls need to be well balanced; forming
board, baffle, and suction box surfaces smooth; presser roll pressure not
excessive; jets from the wire showers uniform (especially if the water is at a
different temperature to the wire at start-up); and so forth. Any unevenness
of tension across the machine may cause stretching early on in the life of
the wire which gives trouble if tension is slackened off later on. Also
stiffness in wire guide spades or undue tightness at the edges subject the
wire to greater stress in this region and increase the likelihood that cracks
appear; it has been suggested that the edges of new wires should be deliber-
ately stretched to reduce the possibility of cracks appearing, but a more
common palliative is to use plastic edges.
A small hard particle passing between the wire and a roll creates a pimple
264
which wears faster than surrounding portions of the wire, especially when
facing inwards, so that sooner or later a burst occurs. Inward indentations
are frequently caused at the first wire return roll and the incidence of
damage from this has been shown to be higher for faster machines (where
wire tension is usually greater); other factors likely to increase the frequency
of this type of damage are finer mesh wires, a wide wrap round the roll, and
a small diameter return roll. Outward illdentations are not so serious but a
bad pimple can eventually lead to a crack across the wire.
In some mills corrosion of wires presents a more serious problem than
either wear or damage due to other causes. Studies of this have been made
by analyzing stock and whitewater systems and it appears that in some
cases a greater loss of copper can take place from over the whole body of the
wire than is lost at the knuckles due to normal wear. Provided the loss is
fairly uniform over the whole surface area of the wire the strength is not
unduly affected but it is more usual for attack not to be uniform and
concentrations cause pits in the wire which severely weaken it. Corrosion
increases with a lower pH but also depends on many other factors;
addition to the stock or showers of an inhibitor, or putting a suitable
coating on the wire, are reported to help but economically the best solution
to the problem in the long-run is likely to be the use of plastic wire.
3B . 7 3 Cleaning the wire

It is most important to keep the wire scrupulously clean both to prevent the
mesh gradually becoming clogged and to reduce the possibility of damage.
As most damage to a wire occurs when clumps of fibre pass between the
wire and the first return roll the most important cleaning shower is situated
between the couch and this roll. To do an efficient job all trace of fibre
adhering to the wire on the return run should be eliminated and this task
will, of course, be made easier if the sheet has been cleanly couched.
With efficient cleaning and the use of trays to prevent backwater from table
rolls cascading through the wire underneath, it has become rare for wires
to give drainage difficulties due to plugging. But damage of the wire
stemming from an indentation caused at the first return roll is not so un-
common.
There is some difference of opinion regarding whether fresh water or
whitewater should be used for wire cleaning showers. The adherents of
fresh water claim that the use of whitewater damages the wire and is
particularly dangerous if the wire is stopped without showers being
immediately turned off. There is also doubt cast on the ability of whitewater
showers to do an efficient job of cleaning. But using fresh water, unless it
be kept separate from the rest of the backwater system and partially
recirculated, can be very detrimental to general running stability and it is
probably the case these days that most machines, certainly faster ones, are
changing over to using whitewater.
With whitewater cleaning showers the lower the consistency, the less
likely it is that the wire is damaged or plugged at an emergency stop. For
this reason on modern systems whitewater that has passed either through
265
two save-alls in series or is from the clear section of a disc filter, and is
consequently highly clarified, is generally used. This, coupled with the use
of high pressures and dual showers close to the couch, gives an efficient
means of thoroughly cleaning the wire. Even so, especially when fibre is in
the shower system it is important to ensure that plugging does not occur,
as nothing is more likely to spoil a wire and give uneven drainage than a
gradual fibre build-up in the mesh at one or more places across the machine.
For this reason oscillating and self-cleaning showers of various designs are
becoming more familiar on machines; some of these are of the automatic
purging type while others require manual operation of the cleaning action.
Regular inspection for scale and other deposits, together with periodic
measurement of the flow to the showers, are highly advisable.
Such showers alone are not usually sufficient to deal with the whole web
when this is carried round the couch during a break. In this case it is
essential to have a separate powerful knock-off shower that sprays out
comparatively large volumes of whitewater. The arrangement of whitewater
systems to provide this facility has already been touched upon and one of
the problems involved is arranging to start the shower sufficiently quickly
when the sheet breaks. With open-draw machines such showers when
needed usually have to be kept on all the time and provision made for
adequate recirculation. On pick-up machines more warning is available
and the turning on of a knock-off shower can be connected to work
automatically with movement of the forward-drive roll or other mechanism
which governs operation of the pick-up.
There appears to be no generally recognized system of positioning
showers for knocking off the sheet at a break. In most cases a first shower
inside the wire loosens the sheet and a second more powerful one blows it
off; care must be taken here that the trajectory of the sheet as it falls from
the wire does not foul the first return roll. On some machines a single
shower loosens the sheet which is then transferred to the first return roll
and doctored off; in this case with the sheet passing between the wire and
return roll there is greater danger of damaging the wire and an excellent
doctor on the roll is essential for success.
The only systematic investigation into the action of wire cleaning sprays
known to the author has been due to Brecht and Weitzel (63). The efficiency
of various models of circular and flat jet sprays was evaluated and the
conclusion reached that the flat jet gives the better cleaning action for a
given amount of water; however, flat jet sprays have a tendency to vary
considerably in cleaning efficiency across their width and careful design is
required. With all types the most efficient cleaning action occurs when
the jet impinges vertically on the wire.
38. 7 4 Wire material

The use of non-ferrous metal wires was universal up to a few years ago and
the development of suitable metals to give the necessary ductility and
strength, together with abrasion and fatigue resistance, has been described
by several authors (notably Paton (7)). Twill weave wires have been shown
266
THE WIRE SECTION 38.74
to have a greater tensile strength than plain wires for a given reduction in
their thickness due to wear, and this is of great importance to the wire life.
The wire seam rarely gives any trouble on the machine though the occa-
sional wire does give way first at the seam. At one time the seam was
especially prone to becoming plugged with fibre and other deposits but
modern techniques of welding have eliminated this problem. It is also
usual for the end of one wire to be joined to a different wire two or three
positions away; this gives added protection against the possibility of groov-
ing the suction boxes.
The primary disadvantage of metal wires is their relatively short life.
Plastic wires or fabrics, though costing substantially more than metal at the
present time, can give an appreciably longer running time particularly on
machines where corrosion limits the normal wire life. Other advantages
are greater ease of cleaning off pitch and other contaminants with steam
showers (pitch adheres less in the first place), better retention, and (though
opinions differ on this point) less tendency to get made up. Also, because
the fabric can be flexed much more than a wire, it is simpler and quicker
to put on the machine, particularly where space in the aisle is cramped.
Adhesion of the web to a fabric is considered less than to a wire; drainage
and wire mark appear to be similar. In comparing the merits of the two
materials a full economic appraisal is, of course, necessary and this should
take account of wire changing costs and downtime for cleaning and
repair; in general it is thought that plastic wires are particularly beneficial
on machines which have corrosion or pitch problems, or when the fre-
quency of wire changes represents a substantial total cost.
Plastic wires used to stretch 2 to 3 per cent. during the course of their
life and this always required modification of the run of the wire because
normally a stretch of only at most a few inches is possible; on slower
machines this did not present any real problem but faster machines
requiring greater tension needed considerable alteration. For example,
Macpherson (66) has described in some detail the experiences on one
machine where a change over to plastic wires involved installing three
expander rolls and more return rolls to give a larger stretch and an ade-
quate wrap round the expander rolls. However, the problem of excessive
stretch has now been overcome so this particular installation cannot be
regarded as typical. Some alterations are generally needed to take plastic
wires and these include a single expander roll to keep the wire tight across
the machine, plastic or stainless steel suction box covers, and the installa-
tion of one or two plastic or stainless steel foils to improve drainage where
this is likely to produce useful results. Disadvantages mentioned in the use
of plastic wires, apart from the greater stretch, are higher risk of being
punctured by hard objects, lower strength, and greater liability to wrinkle.
The use of a coarser wire to support the finer plastic wire over the suction
boxes has been successfully applied as a means of further reducing wear
(104).
Reports have also appeared describing the merits of stainless steel and
chromium-plated wires, and under certain conditions these appear to
improve on the life of ordinary phosphor-bronze wires.
267
3B. 7 5 Slope of the wire

The idea of sloping the wire downwards first arose due to the difficulty in
getting a fast enough flow from the old dam-type slice. Also, with the low
heads used with this simple type of slice a large difference in velocity can
exist between the upper and lower layers of the flow particularly with
heavier papers. Sloping the wire, by giving a slight acceleration to the upper
layers before they become fixed in the mat, may be expected to reduce this
difference.
Upward sloping of the wire is occasionally used on old machines when
it is desired to increase the drainage rate by the creation of a deeper stock
thickness in the first few feet of drainage; it is stated that faster speeds can
then be achieved and an additional advantage often claimed is that the
slight deceleration of the upper regions of the stock assists in deflocculation
of these layers. Very steeply sloping wires are used in association with
pressure formation on special tissue machines and also when strong machine
direction alignment of the fibres is desired.
It is sometimes claimed that downward slope on faster machines assists
in delaying drainage (by carrying the water further down the wire) and
thereby can benefit formation; another idea current is that slope will help
to overcome air friction on the upper surface. In the author's opinion
neither effect is likely to prove significant in practice. Firstly, drainage rate
is unlikely to be altered by the slope because on faster machines the
gravitational element is insignificant; effectively all drainage occurs at the
table rolls and foils and the forces involved in this are unaffected by the
slope. Secondly, any frictional effect of the air on the upper surface of the
stock cannot have much chance to affect the velocity significantly because
the increasing pull of lower layers as the mat thickens will keep the velocity
identical to that of the wire.
There have never been any controlled experiments on the effect of slope
and its influence at slower speeds can only be conjectured; however, it is
worth pointing out that as a way of altering the drainage it is certainly less
flexible compared to other means available because the engineering
arrangements effectively require that once decided upon the slope is fixed.
A simple calculation shows that moderate degrees of slope cannot have a
significant effect on the formation and drainage of the stock over about
1,000 ft./min., so for faster machines whether or not the wire slopes may
be settled primarily for convenience of design.
3B. 8 BREAST BOX STOCK CONSISTENCY AND TEMPERATURE

The influence of breast box consistency on formation has already been
discussed in some detail; in particular, the point has been made that in
general under any given machine conditions the lower the consistency
(to the point where the flow from the slice becomes so great that instabilities
occur), the easier it will be to obtain a satisfactory formation. Wrist (84),
for example, demonstrated that the random variation in substance over
small areas is significantly less at lower consistencies. In practice the breast
268
box consistency is governed by the relation between the drainage rate and
the position of the dry-line: if drainage rate decreases and the dry-line moves
towards the couch, the machineman circulates less backwater which
increases the consistency.
It is not possible to define an ideal position for the dry-line which in
practice is arrived at by reconciling the conflicting requirements of low
breast box stock consistency and high solids content at the couch. A
difference in moisture after the suction boxes becomes proportionally
lower after the couch, and this in turn becomes proportionately lower after
the presses. Thus there is no particular merit in prejudicing formation of
the sheet in an attempt to obtain too high a dryness at the couch. But at
the same time, of course, the wet strength of the sheet must always be
above the point where transfer from the wire becomes hazardous.
A frequent problem is shortage of drainage capacity; this necessitates
compromising formation by the need to run the breast box consistency
higher than is desirable in an endeavour to obtain a sheet sufficiently
dry to withstand couching. On any machine drainage rate can be increased
in several ways by modifying the wire table or increasing the water extracted
at the suction boxes or couch. These points have been dealt with earlier.
Otherwise under any particular running conditions the factor having the
greatest influence on drainage rate is the temperature of the stock; an
increase of temperature, by lowering the viscosity and surface tension,
reduces resistance to flow through the fibrous mat and wire mesh.
It is common practice to heat stock in the main backwater pit and
elsewhere purely to gain this advantage in quicker drainage. In integrated
mills the stock has already been heated during preparation of the pulp
but otherwise some consumption of steam is required for this purpose; in
this case it is desirable to make some assessment, however approximate, of
the added cost of the steam against the benefits obtained in the form of a
lower frequency of downtime or increased speed. Generally such a calcula-
tion may be expected to show that the addition of steam is well worthwhile
to quite high temperatures but a limit must be imposed by the necessity of
keeping working conditions at the wet-end tolerable. A disadvantage of
heating the stock, as shown by the work of Burkhard and Wrist mentioned
earlier, is that a rise in temperature lowers the machine speed at which the
onset of kick-up occurs.
Little data is available to demonstrate the change in drainage rate with
temperature under actual machine conditions. Bennett's work indicated
that a 4 deg. F. rise in temperature on 35 deg. F. increased the overall
drainage rate by approximately 2 per cent., the increase being proportion-
ately greater in the second half of the table.
3B.9 MACHINE SPEED, SUBSTANCE AND STUFF TREATMENT

The influence of machine speed on most aspects of the wire section has
already been treated in detail and for discussion of its effect on drainage
behaviour, couching, the dandy, and the wire itself, reference may be made
to the appropriate section. In-so-far as conditions in the wire section govern
269
the maximum speed a machine is capable of under given circumstances, the

total drainage capacity is the most important feature which is affected by
speed.
The problem posed by limited drainage capacity has been dealt with in
the preceding section in relation to the effect on breast box stock consis-
tency and the advantages of heating the stock. There are two further factors
which closely affect drainage rate; substance and stock treatment (in the
broad sense of changing wetness by beating, refining, addition of broke,
etc.), and these are now briefly discussed.
There is little information on the direct effect of substance on dewatering
in the wire section except that obtained as part of a larger experiment by
Jordannson (27) which is discussed in greater detail in the Part dealing
with the presses. This work clearly showed that increasing substance
reduces dryness at the couch with other factors constant. Common practice
indicates that within a fairly wide range drainage capacity is approximately
inversely proportional to the substance-in other words an increase in
substance generally requires a corresponding decrease in speed of the
same proportion, in such a way that the ultimate production remains
roughly the same. Because under normal operation conditions some change
in speed is customary when altering making substance (unless other factors
such as limitations in the drive have some bearing on the matter) it is
usually not possible, nor particularly useful, to try to separate the two
effects.
The effect of stuff composition and treatment on drainage rate is a
central factor in papermaking. Ideally the furnish should be chosen,
blended, and treated to produce a paper with the desired physical pro-
perties, but in practice a compromise is always necessary to enable a
machine to be run economically. The most important factor affecting this
compromise is the behaviour of the stock on the wire. Generally speaking
the effect of greater stuff treatment or the addition of a higher proportion
of a stronger fibre is to produce a sheet with greater strength properties;
this is important not only with respect to the final quality of the paper put
also because the wet-strength of the web at the couch for a particular
dryness is increased. But greater treatment also produces a slower draining
stock and ultimately to obtain a dry enough sheet at the couch either speed
must be sacrificed or the treatment reduced. This problem is familiar to all
papermakers and its characteristics vary from one machine to the next
and depend essentially on the qualities of paper manufactured. In particular
on machines making a wide range of substance it is not found possible as
the substance is increased simply to decrease the machine speed to keep
the production pretty well constant; a reduction of treatment gradually
assumes greater importance otherwise it becomes impracticable at the
heavier substances to dewater the sheet adequately.
The behaviour on the paper machine wire of different furnishes and the
effect of varying degrees and types of treatment is complicated by the great
change in drainage rate produced by backwater recirculation of fines and
loading, in other words by the question of retention. Laboratory simulation
of machine recirculation conditions has only been approached in a rudi-
270
mentary manner and what knowledge is available on the drainage charac-
teristics of different fibres under actual operating conditions is largely
derived from common observation. Some fibres, such as those from
mechanical pulp, or fines produced essentially by cutting rather than
fibrillating, appear to reduce the drainage rate at the table rolls but have,
if anything, the opposite effect at the suction boxes, (Bennett (9), for
example, found a change of breast box stock C.S.F. had a greater propor-
tional effect on drainage at the earlier table rolls). In a similar way, wet and
dry broke usually exhibits a different effect on the wire. Such phenomena
can be presumed to depend on the relative flexibility of fibres and compac-
tion of the mat at different stages down the wire, and explanations can only
be proferred in general terms.
Apart from its effect on drainage rate, stuff treatment also alters the
porosity of the sheet and this produces a change in the vacuum that is
sustained at suction boxes under the wire and in suction rolls. This change
is a familiar phenomenon and, so long as it can reasonably be isolated from
other factors that also influence porosity such as substance, sheet dryness,
general formation, dandy and presser roll operation, etc., is useful to the
machineman as a guide to changes in composition and treatment of the
fresh stuff. Again, however, such changes must be individual to the machine
and always require interpretation in the light of prevailing operating con-
ditions. J ordansson reported some observations on the effect of beating
an unbleached sulphate pulp with other conditions relatively steady;
dryness at the couch and particularly at the presses reduced (details are
given in section 4B. 7) and it is interesting to note that the effect of a change
from 19 deg. S.R. to 66 deg. S.R. was to increase the vacuum at the couch
from 10 in. Hg. to 20 in. Hg., and at the pick-up roll from 10 in. Hg. to
14 in. Hg.
3B.10 EQUIPMENT
To conclude this chapter dealing with operating factors affecting perform-
ance of the wire part, it is appropriate to make a few remarks on some
other pieces of equipment commonly in use that have not so far been
mentioned. These include return rolls and their doctors, the breast roll,
trays, steam sprays, wire stretch and tension devices, wire guides, deckles,
aprons, and cutters.
3B. 10 1. Rolls and doctors

All return rolls under a wire require as careful attention in regard to
balancing and positioning as table rolls; in the same way also they require
frequent attention to the journal bearings to ensure that drag on the wire is
kept to a minimum. Doctors are not generally considered necessary on
inside rolls except for machines which do not tray all the table roll back-
water; in the latter case the shower of water falling on to the return run
of the wire saturates the inside rolls and makes a doctor and collecting tray
essential. A doctor is always used on the first outside roll and in this position
271
is of critical importance in reducing the possibility of damage to the wire

caused by lumps of stock failing to be washed off and adhering to the roll.
The angle the doctor makes to the roll it contacts is important: too fiat
an angle reduces the effect of the doctor and makes it easier for lumps of
fibre to pass under when there is a small discrepancy in the evenness of
contact; too steep an angle lessens the scraping effect and in excess results
in blades chattering. The angle generally recommended is between 25 and
35 deg. and attention is required to ensure that as wear takes place the
angle does not exceed the upper figure. Important also is the pressure of the
doctor on the roll and to avoid the necessity of heavy pressures to overcome
unevenness of contact it is beneficial to construct with light material and
use flexible plastic blades. Air-loaded doctors are becoming increasingly
common because the facility of being able to vary pressure to suit conditions
has been found to be a useful one; the type which allows equal pressure to
be exerted all along the doctor, though more expensive than the type which
simply loads the ends, is obviously more likely to do an efficient job.
Setting the doctor load is largely a matter of trial and error but excessive
weight is unwise for this results in a greater drag on the wire particularly
when the angle of wrap is small; in some cases it may even then become
necessary to apply a separate drive to the roll. As for all doctors, oscillation
from side to side geared to the roll rotation is a simple but wise precaution
against scoring.
The breast roll doctor is also highly important; without it fibre lumps
passsing under the doctor would be squeezed under the wire during its
long and tight wrap round the breast roll and eventually mark the wire.
Fortunately the consistency of backwater discharged down the roll is low
so the risk of marking the wire from this source is much smaller than at the
first return roll. A more frequent source of trouble from a poorly-fitting
breast roll doctor occurs when an uneven quantity of water is carried
round the roll. This is often a source of aeration in the form of relatively
large bubbles of air visible on the mat surface and can be alleviated either
by a shower or compressed air jet, directed into the nip between the lower
slice lip and the breast roll, or a soft rubber wiper lightly pressed on to the
wire; these devices also reduce the possibility of streaks caused by water
channelling under a wire apron and are especially important when trays are
not used. Where forming boards are not possible for reducing the gap
between the breast roll and first table roll, a relatively small breast roll can
be used and a separate larger roll is then required to act as the turning roll;
an advantage of this arrangement is that the smaller breast roll permits a
more variable setting of the point of impact of the slice jet without re-
quiring the lower slice lip to protrude too far over the roll, thereby increasing
the vertical fall of the jet on to the wire (see discussion in 3B. 1 2).
3B. 10 2. Other equipment
Reduction of aeration achieved by using trays under the wire to catch
table roll backwater has been mentioned several times and there is no
doubt that when the return wire is showered with backwater a large volume
is carried through to the breast roll and can spoil the underside of the sheet.
272
Machines having the main backwater pit at the side of the machine always
have trays under the wire and it is only when the pit under the wire is used
to collect the backwater that the practice of not using trays is sometimes
found. Although on some occasions it may be found that leaving out
trays does not appear to influence the operation of the machine in any way,
if frothing or air-bells on the wire occur replacement of trays is certain to
help alleviate the problems caused. Trays are often avoided by the machine
crews when they have been badly designed or are too heavy for easy
removal and replacement during a wire change. In this case the remedy
is obvious, and consideration should be given to using plastic materials
for lightness.
Steam showers on the wire, though effective for bursting bubbles when
these cannot be avoided in any of the ways mentioned above, are an
unmitigated nuisance and make general working conditions and observa-
tion of the wire and dandy very difficult. Furthermore, badly designed
steam jets drop water on to the sheet and, if too strong, can produce a
secondary disturbance in the form of a spray as the bubbles collapse.
Mansell and Saunders (6) have suggested using an anti-foam compound
spray on to the web with the aid of compressed air. Their experiments in
this direction lead them to the conclusion that this was a more effective and
cheaper way of removing bubbles on the stock surface but probably the
relatively higher capital cost of the equipment compared to a simple steam
spray has prevented a more general use.
Stretch in the wire is usually taken up on one, or on longer wires two,
rolls with vertical adjustment on a screw; this serves at the same time to
set the tension in the wire. Although movement of the stretch roll may be
two or three feet, the difference between minimum to maximum lengths of
the wire will only be a matter of inches and great accuracy in supplying the
exact ordered length is imperative for wire manufacturers. On faster
machines the wire tension set by the stretch roll becomes a very important
factor affecting wire life and sheet quality so that some indication of
the tension is highly desirable both for setting and running the wire more
consistently. A load cell incorporated at each end of a suitable roll is on
way of obtaining a measure of the wire tension and this can also indicate
if there is any difference in tension across the wire. As a final development,
automatic tensioning devices enabling the wire to be set (preferably while
the machine is in operation) and maintained at any desired tension are
becoming more common on faster machines and a number of efficient
designs are now available.
For guiding the wire the spade-actuated mechanical movement has
proved a most efficient design when properly set up, but with increasing
speeds the response for stable correction of a drift in the wire becomes too
sluggish and the greater pressure of the paddle on the edge of the wire
needed to effect movement of the guide roll more easily causes trouble with
cracking. For these reasons air and water jets or photocells are now more
frequently used to detect movement of the edge of the wire and initiate
appropriate adjustment of the guide roll. Alternatively, the self-acting
servo-roll type of guide may be used.
273
The older type of movable deckle straps were never very satisfactory and
if allowed to become too slack or dirty on machines with no trim the
resulting rough deckle edge could cause frequent press breaks. The
stationary deckle strips now in general use except on slow apron machines
can also be troublesome if lumps of fibre collect between the strip and the
wire, and their setting is critical on faster machines if edge waves are to be
avoided. The stock jet should leave the slice parallel to the wire and the
deckle can then be pushed in until it is just contacting the edge of the jet.
Some machinemen consider that the vertical line of the strip is best angled
outwards as this reduces the tendency for cross-waves at the edges of the
jet to bounce off and affect the level further in; on the other hand it is also
frequently recommended that the deckle strip is bowed slightly outwards
down the wire and unless the strip is vertical it is not possible to do this
without leaving a gap somewhere between the strip and the wire. The best
setting must in fact be found empirically and the ease with which this
proves possible will certainly depend on how smooth the flow from the
slice can be made. Flexibility of adjustment with adequate rigidity is most
important in the design of the holding brackets but inevitably after some
time trouble is likely to occur with the deckles when they have become
twisted in all directions in an endeavour to remedy an edge fault; it is then
simplest to start from scratch again with a perfectly straight, upright strip.
There is also a method of obtaining a deckle using compressed air jets but
little information on the advantages of this device has yet become available.
The rubber or leather apron still used on very slow machines with dam-
type breast boxes can be a trouble-maker if neglected. It requires careful
fixing to avoid leaving any projecting points on to which longer fibres
easily collect and should be replaced when cracks or wrinkles appear, or if
the end becomes frayed. It is generally advisable not to let the apron
become too dry and on some machines an old apron is used for support
to reduce wear on the underside from the wire.
Finally, cutters on the wire used for the web edge-trim and feeding-up
tail need to be kept in first-class condition. For this reason they always use
fresh, filtered water, and require frequent wiping round the edge where
splashed fibre collects. A cutter that does not produce a clean straight,
vertical jet with no feathering can create an enormous amount of trouble
by dragging the trim at the couch and producing a ragged edge that sticks
at the presses. To avoid breaking the sheet for cleaning or changing cutter
nozzles it may be prudent to have two sets available on the machine.
274
CHAPTER 3C
RUNNING THE WIRE SECTION
3C.1 DAILY OPERATION

The general problem of operating the wire section efficiently is now con-
sidered from the point of view first of day-to-day running and then of
general long-term maintenance. Discussion of daily operation will follow
the procedure adopted in the other Parts of this book and classify measure-
ments that can be made to assist the machineman into essential, important,
and useful categories. In the section dealing with general maintenance and
long-term records, frequency of wire changing is given prominence because
this is the single most important piece of regular maintenance necessary.
Finally, there is a section dealing with the more practical aspects of running
the wire section.
3C . 1 1 Essential measurements
Measurements which can be classed as essential for running the wire part
are those which show the performance of the vacuum systems on the suction
boxes and also, when applicable, on a suction couch and pick-up roll.
In addition the draw at the couch and, where the drive is from individual
electric motors, the power taken by the wire may be included in this
category.
With regard to the suction boxes a straightforward vacuum gauge in the
main manifold is the simplest indication available. With the type of suction
arrangement involving a single manifold carrying water to a separator
tank, it is also essential as a minimum to have a sight glass for indicating the
level in the tank and a pressure gauge on the extraction pump; these enable
performance of the separating system to be easily checked if operation is
sluggish or :flooding of the boxes is suspected, though it is preferable in
addition to have a level controller in the tank operating a throttling valve
on the discharge side of the pump.
Apart from showing that the system is functioning as required and that
air leaks have not become excessive, the use of the main vacuum gauge
depends to some extent on the suction arrangement, and in particular on
whether an automatic spring-loaded air relief valve or similar type of
regulator is used. In this case the vacuum either varies below a maximum
value determined by the regulator setting or, if the relief is continuous,
keeps relatively steady. A simple regulator of this type is probably the most
common system in use on older machines and the primary function of the
gauge may then be regarded as being to set the vacuum in the boxes to the
desired level by whatever means of adjustment is available. When there is
no vacuum regulator the gauge serves to indicate if release cocks on
275
3C.11 THE WIRE SECTION
individual suction boxes at the front side or in some other position need to
be opened to prevent the vacuum becoming too high. Also, since the vacuum
can then vary somewhat depending on changing air resistance in the sheet,
some indication of whether the stock is running wetter or freer on the wire
is available, though for this particular purpose to achieve much accuracy a
measure of the air flow is preferable as discussed in the next section.
The use of vacuum gauges on suction boxes in a suction couch or pick-up
roll is essentially similar to their use in a suction press. They serve to
indicate malfunctioning of the system due to inefficiency in the pump itself
or to unusually great air leakage at the seals. During a particular making
the suction couch vacuum fluctuates due to changes in porosity and
consolidation of the sheet, and can be used as an indication of varying
making conditions, particularly when similar changes in vacuum occur at
the suction boxes. Continuous recording of the vacuum in both suction
couch and (where appropriate) suction boxes, with presentation of the two
traces side by side, can for this reason provide very useful information on
many machines, though some precaution is needed when interpreting
relative changes of the two readings because, in comparison with the
suction box vacuum, the couch vacuum is in practice affected by changes in
solids content of the web (and hence by substance and freeness changes)
to only a limited extent. Normally, recording the couch vacuum would not
be considered essential, but logging of average vacuum readings over a
period is valuable for indicating gradual make-up of the holes in the roll
(vacuum increasing as less air is carried round in the holes) and other long-
term changes in operation. However, when suction box vacuum is con-
trolled, the suction couch vacuum is the only indication readily available
of the stability of making conditions and would then be worth recording
if the better alternative of using air flow through the suction boxes is not
available.
The draw applied at the couch, whether of the open or felt transfer type,
is the most critical of all draws down the machine. The importance of the
position of the take-off line and the angle of draw from the wire has been
stressed earlier and the effect that the web tension can have on paper
quality has already been emphasized. A continuous measurement of draw in
this position is a most important guide to the machineman, enabling
repeatable conditions to be attained at the couch and, from observation of
the apparent tension in the sheet at the draw (i.e. movement of the take-off
line), variations in adhesion of the web to the wire can be readily detected.
Without a draw measurement, when the web carries further round the
couch roll and the draw therefore appears slacker, the machineman cannot
know whether this is due to a change in adhesion to the wire or to the draw
itself.
A particularly sensitive method of draw measurement has been des-
cribed by Schroder et al. (95) and can be used for studying the relatively
rapid fluctuations (or 'fluttering') of the draw that are observable on many
machines. These originate mainly from drive variations caused by changes
in power transmission or in load on the drive, and apart from the resultant
operational instability they induce, have been shown to affect directly the
276
THE WIRE SECTION 3C.12
substance and strength and stretch properties of the sheet. However, for
regular machine use measurement of the draw is simpler at the couch
than in other places down the paper machine because it is normally at least
2 per cent. of the machine speed even for pick-up transfer, and can be
as high as 8 or 10 per cent. on lightweight papers. Various methods can be
used though perhaps for comparative purposes the most suitable con-
tinuously monitors the difference between pulse counts from special
tachometers mounted on two appropriate rolls.
When practicable a measure of the power consumed by the wire section
is invaluable for preventing excessive drag of the wire, and should be
associated with a limit warning device to ensure that the machineman
notices when consumption exceeds a reasonable level; to give early indica-
tion of adverse trends a log should be kept of the ammeter readings. The
effect of altering the tension, vacuum on the suction boxes, roll doctor
pressure, and other important variables can all be studied by observation
of the drive ammeter. Further, provided the normal increase in power
consumed over a wire life has been noted for several wires and averaged,
a valuable means is available after changing a wire or replacing suction
boxes, forming boards, deflectors, etc. to indicate if alignment is poor; in
fact often the power demand can be used to assist in obtaining correct
setting. Also the drive ammeter often detects cyclic variations more
sensitively than is possible by observing fluctuations either of the sheet at a
pick-up roll or of the suction box vacuum; these variations may be due to
such things as poorly set deckle seals in oscillating suction boxes, excessive
slippage of the wire, eccentric breast or couch rolls, or bad cases of uneven
stock discharge from the slice.
3C. 1 2 Important measurements

Amongst data which are all too rarely available on paper machines but
which can provide considerable help in checking and tracing abnormalities
of wire operation may be listed an indication of wire tension, air flow from
the suction boxes, the vacuum on each individual suction box and, when
the system design permits, the flow from each box.
The value of having some indication of wire tension is particularly
important for faster machines where it can have a close bearing on wire
life and stock disturbances. In such cases the familiar method of setting
tension by pressing down with the hand on the wire is too subjective and
anyway gives no idea of running tension which is varied by the drag
exerted on the wire. Several methods of indicating tension are available
and depend on whether or not an automatic tensioning device is used.
When automatic tensioning is not used the position of a roll free riding in a
vertical slot under the wire provides perhaps the crudest indication of
running tension, but a far more accurate method involves measuring the
force exerted on an appropriate roll by means of strain gauges. With
automatic tensioning devices it is always possible to select the tension
precisely at the desired value either hydraulically or by means of weights.
Each of these devices will permit repeatable setting of running tension and,
10 277
if necessary, systematic alteration during the life of a wire to prevent an

excessive increase in the power consumed by the drive.
When vacuum in the suction boxes is largely kept steady by means of a
relief valve or some other control system, where practicable a measure of
the total air flow from one or all the boxes can be a useful guide to stability
of operation. Such a measurement need not be absolute and is obtained
very simply by means of a suitable orifice in an individual line or the main
air manifold with a differential pressure cell coupled to tappings either side
of the orifice; this generally provides a more sensitive measure of the
condition of the web in the later stages of drainage than the suction couch
vacuum which is sometimes used for this purpose (see previous section).
Over relatively long periods the reading obtained in this way is affected by
several variables, in particular the condition and setting of the suction box
surfaces and deckle seals, and to some extent also the age and cleanliness
of the wire; but for shorter periods of a few days or less a measurement of
the air flow can prove very useful for several purposes.
Comparison with a continuous substance measurement obtained from
a beta ray gauge at the dry-end shows that a large proportion of the
variation which occurs in air flow from suction boxes over periods of
several minutes is due to fluctuations in the substance. But over longer
periods than this any trends which occur can often be accurate indications
of alterations in composition of the fresh stuff or retention on the wire.
To detect these trends adequately requires a continuous record of the air
flow which should preferably be situated by the side of, or on, the same
recorder as the substance record, an essential elaboration if full use is to
be obtained of this measurement.
Deliberate alterations to the quantity of backwater in circulation, the
operating vacuum, or formation conditions also affect this reading and its
comparative nature should, therefore, be thoroughly understood by the
machineman. But such alterations apart, air flow through the web at the
suction boxes will depend on the closeness of the mat and its solids content
above the boxes, both of which are functions of drainage and retention
conditions and hence primarily of treatment of the fresh stuff. For this
reason perhaps the most useful potential which measurement of suction
box air flow presents is when it is used in conjunction with a last refiner in
the fresh stuff line during intervals between obtaining paper tests or stock
tests in the preparation system upon which the overall control of beating
and refining will normally be dependent; provided the response to small
alterations in the load of such a refiner is reasonably soon apparent on
the wire it may be regulated quite effectively to keep the air flow steady.
Whether or not this keeps the qualities most desirable in the paper more
consistent is, however, a matter for experimental determination.
In operating suction boxes the benefits obtainable from grading the
vacuum on individual boxes have been detailed in 3A. 31. Considering that
these benefits are fairly well-known it is not easy to understand the general
lack on paper machines of any indication of the vacuum on each box.
The expense of a vacuum gauge on each box is small but alternatively it is
not difficult to mount a series of small mercury tubes on the front side of
278
THE WIRE SECTION 3C.1 3
the machine and connect these to the boxes. In fact the latter arrangement
has some merit because the grading of vacuum is more readily apparent,
though if the vacuum gauges are also mounted close together almost the
same visual effect is achieved and maintenance is less. It has been recom-
mended that vacuum should be gradually increased from the first box
approximately up to the box under the dry-line, and thereafter be constant;
with vacuum measurement on each box, changing the rate of increase in
the earlier boxes and the overall vacuum during normal production can be
studied in relation to the influence on average moisture content of the web
at the couch and performance of the wire.
As a useful adjunct to measurement of the vacuum in each box, on
suitable systems the flow of backwater extracted by each box can be
measured. This requires individual drop legs and suction lines on each box,
an arrangement which is in any case superior to a single manifold and
separator tank but which requires adequate depth below the wire to func-
tion; the discharge from each drop leg can be placed into a compartment
with overflow at one side across a suitably sized weir into a collecting pit.
The level over each weir then serves to give a reasonably accurate indication
of the flow from each box, certainly on a comparative basis. With this
additional information it should be possible to study directly the effect on
dewatering efficiency of varying the vacuum applied in each box, a very
valuable facility for any machine especially when drainage capacity limits
production.
3C. l 3 Useful measurements

Several other measurements may usefully be made in the wire part with the
object of providing further data on general drainage conditions and of
regulating more closely the use of a dandy, upper couch or presser roll,
pick-up roll, etc. With regard to the drainage conditions, measurements of
consistencies, flows, and so forth in the machine backwater circuit and
suction boxes are most important and their observation and control does,
of course, contribute considerably to keeping the general drainage con-
ditions steady, but this aspect has already been dealt with in Part 1.
It has frequently been proposed that the flow from one or more table rolls
either over the whole width of the machine or a selected width would give
an extremely valuable record of conditions on the wire. Certainly when
applied to a breast roll discharge such a measurement even on a relatively
crude basis could well be valuable for indicating changes in running
conditions produced by the many factors which affect drainage when the
stock first meets the wire, in particular on faster machines the wire tension
and slice geometry. A drainage rate measurement applied on a continuous
basis to one or more table rolls, may also prove interesting from the experi-
mental point of view, but the author doubts if such a reading can with the
present state of our knowledge provide information which could usefully
augment other indications of changes in drainage which are more easy to
obtain. Certainly it can be expected that for any particular making at a
single speed, changes in the drainage rate will occur due to temperature,
279
stock conditions and many other factors, but at the present time it is almost
certainly more useful to set about keeping these steady by means of other
more direct measurements and control applications than to attempt to use
an indication of table roll flow as a basis for tracing sources of variability
during operation.
A simple measurement of the load of a dandy on the wire and, especially
for a driven dandy, any indication of the relative speed with the wire would
both be very useful. The load indication can be quite rudimentary, the
position of springs at back and front sides or of some relieving weight,
and provided the mechanism is kept free from resistance this is better than
nothing; for more precise measurement a load cell or statimeter type of
instrument could be used, or alternatively with pneumatic relieving the
air pressure is sufficient.
Similar remarks apply to the pressure exerted by a presser roll, upper
couch roll, or pick-up roll which, if consistently too high, can cause damage
to the wire in the form of machine direction cracks. In addition in each case
though probably within fairly wide limits, variation of pressure could have
a significant effect on the ultimate sheet properties and for this reason also
it is a useful facility to have some indication of the load applied, however
indirect.
Regarding the relative speeds of dandy and wire, a separate dandy drive
and differential speed measurement has several advantages for controlling
the running with precision and independent of the pressure of the dandy
on the web. By this means an optimum range for the relative speed can be
found and kept to without difficulty, and for watermarking to register
a ready and closely controllable adjustment is available to take care of
changes in shrinkage and other factors.
3C. 2 MAINTENANCE OF THE WIRE SECTION

3C. 2 1 Wire changing
Changing the wire represents the most frequent and important piece of
maintenance, in the broadest sense of the word, necessary to keep the wire
section running efficiently. The life of a wire can vary from a few days on
very fast machines to many weeks; the reason for removal depends, of
course, very much on the machine conditions, though straightforward
wear is almost certainly the most common.
As with other clothing used in the paper machine the policy adopted with
regard to frequency of changing varies appreciably from mill to mill, but in
a majority it is safe to say that the tendency of the papermaker is always
to err on the side of obtaining as long a life as possible. If a hole or crack
appears the machine is stopped and, unless the damage has become really
extensive before it is noticed, a routine repair is effected. This repair may
last through to the normal life expected of the wire, but then again it might
not and a long series of stoppages for further patching may ensue with
the crews anxiously hoping that the wire will last out till the next shut
period. Avoidance of such a state of affairs is, of course, the mark of an
280
experienced papermaker but in making a judgment in any particular case
whether to repair or remove a wire it is always helpful if some simple but
soundly-based economic criterion is available instead of the normal vague
feeling that the wire should be kept on if at all possible.
One approach to this problem is the use of past records to determine
an optimum life for a wire which gives, on average, the maximum return.
This involves analysis of production and time lost due to repair at different
periods through the life of as many wires as possible, and comparison of
the cost of a new wire with that of downtime for repairs together with
labour costs in carrying out the actual change. The methods and refinements
which may be adopted for determining an optimum life in this way are
essentially the same as those described for achieving the same object with
press felts, to which reference may be made for further details. Such calcula-
tions will illustrate the false economy of trying to obtain a few extra days
of use from a wire after it has reached a certain age, and it follows that
adoption of a standard life based on this will lead to a reduction in the
frequency with which risky repairs are necessary.
Nevertheless, accidental damage or poorly regulated running will
always necessitate some repairs at one time or another, occasionally quite
early on in a wire's life. The. machineman or supervisor is then faced with
assessing the likelihood that the patch he can put on will last, or whether
further trouble may be expected and possibly a reduction in machine speed
advisable to carry the wire through to the next scheduled shut period.
Such decisions are made easier if two standard costs are obtained: the
average cost per hour of the wire, and the average cost per minute of
downtime. With the first figure a simple calculation will give the additional
cost incurred by changing the wire before the next shut period when it
would have been changed anyway; the second figure enables this to be
converted to the equivalent of downtime which, when added to the time
that would be needed to change the wire, can then be balanced against the
time for repair plus the likelihood of further stoppages.
For example, suppose for simplicity that the average cost for the wire is
10s. per hour and the average cost of downtime is 15s. per minute. The
machine is stopped to inspect a hole halfway through the week and another
60 hours running will be needed to carry the wire on to the week-end or the
next scheduled shut period. Replacing the wire incurs additional wire
costs equal to 60 x 10s. = £30, or the equivalent of 40 minutes down-
time. Add to this the time needed to change the wire which, including
preparation time, may be 150 minutes, and the total cost of changing the
wire becomes equivalent to 190 minutes downtime. If the hole is small,
seeming unlikely to give further trouble and taking perhaps only 10 to 15
minutes to patch and get the machine under way again, it is clear that the
wire should be repaired. If on the other hand the hole is already large,
experience may indicate that half-an-hour would be needed to complete the
job and that even then the patch would need examining and probably
replacing at least once a day and possibly once a shift; in this case it is
very likely that downtime could exceed 190 minutes and the wire should be
changed. The effect of another alternative which may be chosen, running
281
the machine slower to nurse the wire, can easily be assessed on the same
terms by estimating the resulting loss of production and converting to
running time at the normal rate.
3C. 2 2 Wire records

More than with any other clothing it is essential to keep careful records of
the makes and types of different wires used on the machine and their
general performance in regard to ease of running. It is customary to
include data concerning the life of the wire and this can be very important
for the purposes of comparison and for establishing an optimum life. To
this end the criterion of performance for each wire may be in terms either
of revolutions turned, tonnage made, days or weeks run or some other
suitable measure. Of these the first is probably to be recommended (even
though to account for varying speeds and downtime it may require a
certain amount of calculation if convenient data such as yardage of paper
produced is not available); this is because the number of revolutions of the
wire is connected closest with wear, the primary cause of deterioration.
Cost per unit (per thousand revolutions, ton of paper, or hour) is a con-
venient means of easy comparison of different types of wire provided it is
remembered that any difference in average loss of time for repairing must
also be taken into consideration.
It is, in fact, a worthwhile addition to the wire records if the total loss of
time required for attention (directly to the wire and not, for example, to
removing pitch) is available. The value of this depends on the circumstances
but it can be particularly useful not only as a refinement for comparing cost
of different wire but also, if the causes of downtime are categorized
carefully, as a means of keeping check on the more familiar troubles-
accidental damage, edge cracking, holes and so forth. In addition, as a
further check on these factors the reason for eventual removal should be
recorded if this takes place before the allotted span.
On any particular machine the wire life will be limited by one or more
factors. These may be fairly obvious, in which case a course of action to
lengthen the life can be defined: better cleaning sprays to overcome
making-up, additives to combat corrosion, a change of suction box material
or running tension to reduce wear, and so forth. If the factors responsible
are not clear from experience then careful wire records may help. More often
often, however, a special programme of examining the wires will be
necessary. This could involve identifying and finding the cause of faults,
and detailed testing of samples of each wire when it is removed from the
machine to determine changes that have occurred due to wear, corrosion,
make-up and so forth. But collecting such data is more in the nature of a
special project in which the results require careful planning and analysis;
the tests are not really appropriate as a routine form of checking for the
papermaker's purposes and their establishment on this basis may give a
false illusion of thoroughness at the expense of a great deal of work.
Even so, certain tests of this type are worth considering as a form of
282
long-term control and as such may be undertaken periodically on typical
used wires with the object of indicating trends and helping to assess the
performance of new types of wire. Included in these tests might be: deter-
mination of reduction in thickness (related to revolutions turned) to indi-
cate deterioration or improvement in the wear-resistant properties of the
wire material, or more likely a change in the abrasion rate at the wire part
due to alteration of suction box covers, forming boards, operating vacuum,
etc.; micro-photographs of the knuckle surfaces as an indication of rough-
ness caused by grit and other hard particles; assessment of the change in
drainage properties of water through the wire (difficult to accomplish
with any accuracy and requiring a specially-designed apparatus) to assess
changes in the degree of making-up or overall corrosion; and measurement
of the tensile and stretch properties of the wire to determine reduction
in strength in relation to that of thickness. Such data when available for
typical wires also assist in detecting the reason for a wire failing earlier than
expected. For example, excessive abrasion of the knuckles might indicate a
deficiency in the cleaning system or unusually dirty pulp; a drop in tensile
strength even though the thickness reduction is less than usual could point
to a wire manufactured from poor metal.
3C. 2 3 Other maintenance

Many items in the wire section need frequent attention and the responsibility
for this and the way in which the maintenance is organized depends essen-
tially on the attitude and progressiveness of each mill. At one end of the
scale little or nothing is renewed until the evidence of neglect becomes
apparent in the quality of the paper or in general smoothness of running
the wire part. At the other end of the scale each item of equipment is
either renewed at regular intervals or examined and renewed when a
specified deterioration has taken place. The wire part is the area of the
paper machine most sensitive to neglect in this respect but it is not appro-
priate here to discuss the techniques and economics of long-term main-
tenance organization, and attention is confined to detailing the main
items of equipment requiring regular inspection.
On machines using couch jackets and pick-up felts these, of course, need
regular renewal. The requirements for adequate records in dealing with
this are essentially similar to those for press felts, and as these are con-
sidered in detail in Part 4, further discussion is unnecessary. The same
applies to the maintenance of suction box seals, internal sprays, hole con-
dition, etc., of a suction couch, which is basically similar to the suction
press. Also the camber on an upper couch roll, squeeze roll, rubber presser
roll, or on any roll inside the wire, though normally very small, nevertheless
should be checked at fairly frequent intervals and the results, possibly
expressed in the form of wear curves, may then be treated in the same
manner as described for press rolls. Neglect of these items can eventually
affect running of the wire as it becomes increasingly difficult to guide and
frequent ridging is likely; also evenness of cross-web solids content at the
couch deteriorates, culminating in an increased frequency of breaks at
283
that point and a worsening of the evenness across the sheet at the
reel-up.
The condition of all doctors in the wire section is especially important
and the blades on each doctor need regular attention; the angle of the
blade to the roll should also be checked at intervals as particularly on
smaller diameter rolls relatively little blade wear has an appreciable effect
on the angle. Beside the doctors, other items of equipment such as deckle
straps or strips, wire aprons, the guide and stretch mechanism, sprays,
forming board and deflector surfaces, cutter nozzles, guard boards, etc.
should all preferably be placed on a preventive maintenance schedule and
a proper record kept of the types in use, their life, condition when removed
and so forth. Cracks or perishing of deckles and aprons, and nicks in the
surface of a forming board may appear insignificant but can cause a
surprising amount of trouble.
The breast roll needs to be checked periodically for alignment, especially
when shake is used. Observation of the wire during running will indicate
any unevenness in rotation of either breast roll or table rolls and this
should never be allowed to continue for long as tendencies to eccentricity
or flats on a roll rapidly become accentuated. Grooved rolls need more
careful cleaning than solid rolls to prevent an accumulation of fibre, scale,
and other debris which are liable to break away unevenly causing the roll
to extract more water at one part than another. Dandy rolls require
thorough off-machine cleaning and examination each shut-down and
records of performance for each dandy should be similar to those des-
cribed for wires.
Suction box covers should always be attended to before the surface has
worn into hollows and loss of vacuum or difficulties guiding the wire occur.
Dressing of covers off the machine should always be done with the suction
box supported exactly as in the wire frame to allow for the sag in the
middle; if water in the box under normal running conditions makes a
significant contribution to the sag, the box should also be part-filled with
water before dressing in this way. It is useful to plane down each box in
rotation over a suitable period, though whether in fact this is possible
depends on the relative rate of wear of the wire and boxes; on faster
machines it may be necessary to attend to almost every box at each wire
change, though when this state of affairs exists on slow machines it may
well be that bad alignment of the wire is creating unequal wear.
It is an advantage to start up with no more than one or two freshly
dressed boxes at the same time to avoid a condition which may produce
too much drag on the wire and make it difficult to obtain the normal
suction. On replacement, careful alignment of the box top to be just parallel
and in even contact with the wire surface, and at right angles to the run of
the wire, is essential to prevent guiding trouble or excessive drag. For fixed
deckle machines, setting the position of the box deckles is also important,
especially when the boxes are oscillated, to ensure that the suction area
does not extend outside the deckle edge of the sheet. A record of the time
spent dressing boxes from each position under the wire, together with the
frequency with which covers have to be replaced, provides important data
284
as conditions of speed, suction-box vacuum, type of wire, and so forth
change over the years. It also, of course, provides essential information
with which to compare the performance of new cover materials.

For the purpose of long-term records tests at the wire section are preferably
obtained after the wire has bedded down, i.e. when the power consumed by
the drive has approached the normal maximum. This will enable compari-
sons to be made without requiring allowance for any effect that the age of
the wire may have in requiring operation at a different tension or suction
box vacuum.
Special measurements which should be taken in a wire section test will
include firstly any of those considered earlier for which instruments do not
already provide an indication. These are: vacuum on each suction box
and the extraction rate, total air flow from the boxes, wire tension, and
whenever possible the load applied by an upper couch or presser roll.
Vacuum on each box is easily obtained from a tapping specially made for
the purpose in the ends of each box. In order to measure the extraction
rate when the water is not all discharged down a drop-leg, probably the
simplest method is that involving addition of salt solution at a fixed rate
through a tapping at the front-side of the box and determination of the
equilibrium salt concentration in samples drawn off from the extraction
pipe before it meets the main manifold. Measurement of air flow through
the boxes can be done with a pitot tube and some idea of the wire tension
can be obtained for comparative purposes when the wire is stationary by
noting the depression caused by lowering a roll of known weight in the
middle of a convenient run of the wire (see reference 49).
The flow from a number of table rolls and the breast roll could prove
useful for long-term records though this is hardly practicable unless
troughs are specially designed to catch the water and channel it to a con-
venient measuring pipe at the side of the machine. Such equipment may
not be convenient to leave permanently in position under the wire. Alter-
natively, in the drier section of the wire it is possible to measure the solids
content of the web (this is best expressed in terms of water-to-fibre ratio);
this may be achieved by using a slit-shaped air nozzle to blow off lumps of
stock from underneath the wire into a receiving pan. Though the accuracy
and repeatability of the results obtained by this technique leave something
to be desired, they can be sufficiently reliable to provide useful data. In this
way determinations of solids content along the table roll section, imme-
diately before the suction boxes, in between the boxes and before the
couch can be made. With practice this does not require too much loss of
production and is conveniently organized in association with press and
drying section tests for which the sheet also needs to be broken down.
This data permits the dewatering efficiency of the couch to be calculated
and provides a means of cross-checking the extraction rates of the suction
boxes when these are measured separately; too close agreement is not to be
expected especially for faster machines and those making lightweight paper
285
because splashing and breaking up of the web on the wire make it difficult
to obtain a representative sample. The position of the dry-line and power
consumed by the wire section should also be noted.
3C. 3 PRACTICAL POINTS

From the practical point of view, operating the wire part of a paper machine
requires more training and experience than any other job in a paper mill.
It also requires a detailed knowledge of the individual characteristics of the
machine. What follows, therefore, can only be of limited value although it
is hoped that it contains some information and points of use to the person
with little or no experience of actually running a wire part.
3C. 3 1 Start-up
The factors in starting-up the wire part which have the most direct influence
on how quickly saleable paper is produced are those affecting formation,
in so far as it is governed by the flow on to the wire, and substance; these
particular aspects have already been dealt with when describing the wet-end
flow system. With the exception of machines where a dandy or top couch
roll is used, the wire part itself probably affects the start-up directly only
when something is overlooked which prevents the normal procedure
taking place smoothly. In other words the machineman needs to be con-
cerned primarily in checking that all the devices comprising the wire part
are functioning satisfactorily before and after the sheet comes on the wire,
as opposed to actually setting the conditions to those desired (as with the
stock flow system) or ensuring that conditions are as near as possible to
those eventually pertaining when the sheet is up (as with the press and drying
sections). This emphasis requires painstaking attention to detail particu-
larly with regard to starting the wire itself, a task in which damage is easily
caused as a result of a small oversight. To assist the machineman to do a
thorough job it is very useful to provide a comprehensive check-list that
he can refer to.
Before the wire is first set in motion it should be thoroughly inspected
and jetted inside and out to ensure so far as is practicable that no objects
have become lodged where they can cause damage in the nips between the
wire and rolls. The wire is then very slowly crawled round, a foot or two at
a time, while the jetting is continued and any doubtful areas are cleaned
with a grease solvent, with the aid perhaps of a wire brush or steam jet.
When the wire is considered to be thoroughly clean it may be stopped for a
while until the stuff is ready, though it is preferable when water is being
pumped round the system and through the slice that the wire is kept running
slowly round. This applies particularly when the water is being heated in
the backwater pit as the wire will then become gradually warmed with the
rest of the system, an added assurance that rapid temperature changes
when the stock comes on the wire do not cause the sudden appearance of
ridges.
When the wire is ready to be put in motion after cleaning, the action of
286
the sprays should be checked to ensure that the pressure is adequate and
the coverage even. The normal tension is then set and the operation of the
guide mechanism carefully checked with somebody posted ready to make
an immediate switch over to manual operation or work the hand guide if
the wire looks like going over the edge of the table rolls. All undriven rolls
should be glanced at to ensure that they are in motion and the scraping
action of the doctors should appear smooth and continuous. This is the
time also for a first inspection of other items of the equipment which are
easily overlooked: the smoothness of shake operation, the jets from cutter
nozzles, the deckle straps or strips, the wire apron, breast roll wire wiper,
tray positions, etc.
Finally the vacuum pumps on the suction boxes and, when applicable,
on the suction couch and pick-up rolls are started. The presser roll or
jacketed couch roll normally have no separate drive and are lowered on to
the wire; where a new jacket needs to be well run in the roll may be lowered
while the wire is being jetted. Pressure on the wire should be light until the
sheet passes through to avoid damage to the wire and lessen the possibility
of the sheet being picked up as it first passes underneath (an occurrence
which is, of course, particularly troublesome with a jacketed couch roll as
it might ruin both wire and jacket).
Water from the machine pit should be circulated first while head at the
slice and other flow box conditions are set, then when the stuff is turned on
the substance of the sheet will be low at first and gradually creep up to the
desired value as the backwater consistency approaches equilibrium. There
is little risk with this procedure that the vacuum at the suction boxes
becomes excessive due to high air resistance of the mat and causes the
wire to freeze, a most undesirable accident. But as a double insurance the
vacuum regulator may initially be set at a lower value than normal and
then re-adjusted shortly after flow on the wire bas been established.
With the sheet on the wire the vacuum in individual suet.ion boxes
requires setting and where a separator tank is used a check is immediately
made that this is functioning correctly. The suction box deckles should also
be examined to ensure that the edges of the sheet are not too wet (leading
to trim troubles and crushing at the couch) and that air is not being sucked
in (this may give an audible indication but otherwise reliance must be
placed on the suction box vacuum gauge or air-flow indication). If one or
more boxes are freshly dressed an especially careful examination is required
to make certain that they have been replaced level and are not rubbing hard
on the wire and creating an excessive load on the wire drive, or causing the
suction to gulp or oscillate. To give the sheet time to approach final running
substance it is then appropriate to check again the action of the shake,
cutters, sprays, dandy wiper, etc. and note any disturbances to the flow
at the edges, adjusting if necessary the deckle straps and positioning the
cutters to give the required trim width.
On older machines without a hog-pit it will be generally considered
preferable to pass the sheet over to the presses as soon as possible. But
before this is done it is far better, when applicable, that the dandy is
lowered. On faster machines to get up speed fast enough the dandy is
287
3C.3 1 THE WIRE SECTION
occasionally lowered before the stock passes on to the wire; this is an

unsatisfactory practice unnecessarily endangering the dandy, and it is
much more sensible to provide the dandy with a helper drive. The dandy
should have been thoroughly sprayed before lowering, as should the presser
roll when used.
For a jacket couch, preparation should have been made with as much
care as for a press felt. The ends require inspection and tightening if
necessary, and a thorough checkover is made for grease marks. Care
must be taken to wet out the jacket evenly, possibly with a hot water
high-pressure hose, but even then occasionally trouble may occur as the
sheet tends to follow up the jacket; this is more likely to happen when the
sheet is too wet or the jacket dirty, but a mixture of resin size and china clay
poured into the guard board or squeeze roll nip is generally thought to
help avoid picking. Weights on the jacket couch or presser roll should
immediately be adjusted to normal, and if sprays are used on the squeeze
roll these are also checked.
If this procedure is carried through efficiently, by the time the substance
of the sheet is approaching its ultimate equilibrium value other properties
of the web at the couch should also be comparable to normal. In particular
the solids content and wet-strength derived from adequate consolidation at
the dandy and couch should be sufficiently close to their eventual values
to make the action of transfer to the presses little different from any
other occasion.
For feeding across to the presses the main cutter is positioned to give a
tail four to six inches in width. With open draws the tail may be picked by
hand from the wire (aided by a lump of broke), or more commonly on
faster machines it is blown over by means of a special air jet in the suction
couch. To assist passing over the tail a light aluminium lead roll may be
initially raised by hydraulic pressure at the front side, and then immediately
lowered; such a roll near the wire is very useful for restricting movement of
the web take-off line but needs to be kept clear of fibre clots or the sheet is
marked. With a suction pick-up roll either the roll itself is lowered or a
forward drive roll moves the wire up into contact with the roll; transfer
may be assisted by diverting extra suction capacity to the pick-up box
but normally little difficulty occurs with this method of couching and it is
common to feed up the whole sheet through the first press at once without
first passing a tail.
Whenever practicable it is preferable from the point of view of the press
that the tail is widened straightaway after being successfully fed through.
For this an automatic widening device carrying the cutter across operates
more smoothly and safely than movement of a connecting hose by hand,
and permits the system to be tied in with one or more break detectors,
normally of the photoelectric or ultrasonic type. When installed at inter-
vals down the machine these detectors can actuate the cutter to narrow the
sheet to feeding-up width and reduce broke gathering in the drying cylin-
ders. Also on some machines it may be possible to detect tears starting at
the edge of the web at the couch or presses with sufficient speed to enable a
cutter positioned centrally above the wire to be turned on; this stops the
288
tear spreading across the full width of the sheet and causing a break, so the
cutter may then be made to traverse automatically to the edge to restore
the sheet, returning to the centre after the water has been turned off.
Difficulty can be experienced at first as the tail end tends to follow round the
wire; a tighter draw may help to overcome this if the sheet is reasonably
strong but it should be possible to return close to the normal operating
draw as soon as the cutter is taken across.
Once the sheet is through the presses the machineman may return his
attention to the various parts of the wire. Sprays are turned over to white-
water if this does not function automatically by means of a head tank feed
or a non-return valve when the pressure is high enough; a check is made
that the trim is passing smoothly to the hog-pit without tending to tear at
the point of separation or follow the main sheet, and the couch or pick-up
deckle positions adjusted if necessary; the dandy is examined for picking,
the jacket operation is checked; the appearance of air-bells on the stock
surface is noted, and so forth. These points will be given a little more
attention when describing what to look for when the machine is running.
There is one adjustment which the machineman is likely to make as soon
as possible, that of setting the dry-line across the machine. The use of the
dry-line for giving an early indication of the evenness of substance distribu-
tion across the web has been mentioned previously but it is appropriate here
to stress that ideally the line should only be taken as an indication which is
useful at this stage of the start-up before proper substance tests can be taken
on paper at the reel-up. The reason for this is that the dry-line profile is
influenced not only by the substance at any particular point across the
sheet but also by any factor which affects the drainage rate unevenly: a
partly clogged grooved-roll, dandy, or suction box; uneven dandy pressure;
sagging deflectors or forming boards giving unequal action; a breast roll
slightly offset extracting more water one side than the other; the wire
partially made-up due to uneven spray or couch action. All these and
probably many other small differences could produce a bulge in the dry-line
even with a perfect substance profile. The dry-line position is highly impor-
tant as an indication to the machineman that the overall conditions with
regard to drainage rate and solids content at the couch are satisfactory, and
also relative movement of the line and the frequency of tongues spurting out
give some indication of the stability of the stock and backwater system;
but variations across the machine, except gross ones which occur, for
example, when the deckles are badly set, need more careful interpretation.
3C. 3 2 Shut-down
The procedure for stopping the wire section either for a scheduled shut or
for an emergency inspection involves relatively little trouble. Once circula-
tion of stock through the slice ceases with stoppage of the pumps or
opening of a dump valve, the wire can be brought to a standstill. Until
the wire is stopped it is advisable to leave the suction box vacuum pump
on or the wire may start wandering about. Sprays are either turned off or,
if the wire is subsequently cleaned by crawling or inching round with the
sprays on, then when the action is not automatic a change must be made
289
immediately from whitewater to fresh water to prevent damaging the wire

or getting it clogged with fibre while almost stationary.
For a scheduled shut-down, the upper couch roll or presser roll should be
raised and, particularly in the case of the former, given a thorough clean.
This applies not only to the jacket but also, when applicable, to the guard
board and squeeze roll both of which should be raised off the jacket.
A dandy needs to be removed immediately from the machine and cleaned
because if fibre is allowed to dry in the delicate mesh permanent damage
may be caused when trying to remove it.
Particularly when stock has been running at a high temperature and a
cold fresh-water shower is turned on, the wire should be slackened back
to permit easy shrinkage as cooling takes place. A strong jet of water,
preferably warm, is generally sufficient to remove debris from the wire, but
if slime or pitch is present additional treatment as described in 3C. 3 5 is
necessary. Rolls also require jetting, particularly when grooved, and the
pressure of all doctors should be relieved. Finally items like the deckles,
apron, forming board surface, doctors, showers, cutter nozzles, and so
forth should be given a routine inspection, and arrangements made when
necessary for renewal. The wire itself is also checked over for holes and
cracks. With hardwood suction box covers and wooden doctors water may
be left playing on the surfaces during a longer shut period to prevent them
drying out and cracking.
3C. 3 3 Changing and running the wire

When changing the wire the first step is to remove the old one; this is
usually done by cutting across the wire at the couch and bundling together
the folds as the roll is slowly crawled round. With the old wire off, before
hosing down the opportunity should be taken to inspect various parts of the
wire section which are normally difficult to see: the middle parts of table
rolls, doctor blades, apron, grooved rolls, etc. For fixed deckle machines the
suction-box edge seals may also be checked and re-set, particularly if
trouble with the edges of the sheet has necessitated a temporary adjustment
at some time which may not have been corrected satisfactorily.
In addition, the positions of the table rolls, forming boards, deflectors,
and suction boxes may be checked for alignment, particularly if any rolls
have been changed or surfaces planed. This can be carried out in several
ways but perhaps the most satisfactory is to stretch a thin steel wire at both
sides of the machine at a gap of i in. or so above the normal run of the
machine wire, then check or set each clearance to this. Though ensuring
correct vertical alignment this check would not detect if one end of a roll or
suction box were displaced horizontally compared to the other; this can
be done only by comparing diagonal measurements to the opposite ends of
a roll known to be square to the wire (preferably the breast or couch roll)
but this task should not be necessary very often.
The routine for actually putting on a new wire depends, of course, on the
basic method designed for the machine and it is of no value here to discuss
this in detail. So much depends on the experience and training of the
290
changing crew and whether or not a soundly thought out procedure is
followed. In this respect there is probably much merit, particularly for
mid-week changes where saving of time is highly important, in having a
single crew specifically trained for this task and having the whole job
method studied with particular reference to the danger-areas where greater
precaution is needed. Wires are easily damaged and the attention given by
a good crew to such little things as unpacking and lifting carefully from
the crate and removing all obstacles, hoses, tools, etc. from the machine-
room floor before starting the operation can save much expense. All
changing equipment, particularly the surfaces of the poles, should be
kept scrupulously clean and adequately protected from warping when
not in use.
With the wire on, the procedure for checking over follows that detailed
in the section on starting-up. A new wire will often require more careful
inspection and cleaning with a grease solvent due to the handling, and the
seam or any odd kinks which have appeared may need smoothing down
by pressing (or in the case of plastic wires lightly ironing at a set tempera-
ture) with a fiat board underneath which is preferably kept specially for
the purpose. Greater attention is also necessary to check the alignment of
the wire and watch for signs of a ridge appearing; when first crawling
round a few turns should be made with the wire fairly slack before it is
gradually tightened up to near eventual running tension.
During running, provided the guiding mechanism functions satisfactorily
the wire itself should not require much attention except an occasional
touch to the hand guide to bring the automatic guide roll to a more central
working position. In the absence of an automatic tensioning device the
stretch roll will be adjusted at intervals to take care of changes in the
length of the wire as it expands under tension and for this purpose an
indication of running tension is particularly useful. Observation is always
necessary to detect the early appearance of any faults or of deposits of
pitch and slime which affect the drainage in patches, and this the machine-
man does continually in his constant watch on how the sheet forms on the
wire.
3C. 3 4 Changing and running a jacket

Although jacket couches are now obsolete there are still a large number of
slower speed machines which use this arrangement. Perhaps more than
any other piece of clothing, selection and running of a jacket is individual
to one machine and in adverse conditions a poor jacket can create an
enormous amount of trouble.
When changing a jacket the old one is cut off lengthwise and gathered up
as the wire is inched round. After jetting out the roll with hot water the
new jacket is opened out and edged over the roll, care being taken to
ensure that the nap runs in the correct direction. This operation should
normally not take long though a great deal of struggling is necessary if the
jacket is slightly undersize and needs to be stretched; for this reason the
inside length of the jacket should correspond to within a fraction of an
291
inch to the ordered size. After sewing or tying up the ends, shrinking of the
jacket must be done with special care and it is preferable to use a spray
which can apply hot water evenly across the surface. As the jacket becomes
thoroughly wet the guard board or squeeze roll is put down to help con-
dition the felt and flatten the nap.
A final examination of the surface of the jacket is advisable to ensure
that there are no hard lumps caused by the roll or a sliver of wood under
the jacket; this would rapidly lead to a ridge in the wire. The tightness
should also be sufficient to prevent the jacket wandering or ruckling up
with the squeeze roll or guard board at normal pressure, and to ensure this
the ends may need to be undone and pulled tighter.
Controlling a jacket during running is straightforward so long as it has
been adequately shrunk and tightened. The seam normally has a slight
lead in the middle from the small amount of camber on the roll, but if
this gets excessive greater load on the ends of the couch roll is required. If
the seam leads at one side it is likely that the guard board or squeeze roll
pressure is unequal at the ends and a little extra load on the leading side
is required. Likewise a tendency to run over to one side may be overcome
by slightly relieving the load on that side. Alterations to the load on the
board or squeeze roll in this way should not be allowed to become excessive
or the pressure of the couch roll on the sheet will become uneven, causing
the jacket to receive unequal wear which will ultimately accentuate the
difficulty of running. For the same reason every effort should be made to
avoid applying unequal weight to the ends of the couch roll to correct a
fault either of drainage earlier up the wire or of the jacket itself.
Jackets become dirty very quickly and frequently need to be given a
thorough clean by jetting with hot water. The nap can also be brushed up
with felt teezers though the long-term benefit of this in an older jacket which
is well clogged and worn is doubtful.
3C. 3 5 Checking the wire section during running

The machineman spends most of his time watching the wire section and a
list of the points he looks for, or should look for, would fill many pages;
even then it would be impracticable to attempt to detail them all for so
much is essentially individual to each machine. All instrument and recorder
readings should be regularly checked and attempts made to explain
movements which occur beyond the normal expected limits of variation;
when required a log is completed. The working of all the ancillary pieces of
equipment should also be systematically inspected-a simple check list
can be very useful for this purpose.
At the same time the most important criterion of successful operation,
the sheet quality, is carefully checked each time a sample is available;
apart from the general appearance of the sheet, any blemishes or faults
which are present are noted and explanations for them sought. These may
be identified as coming from fairly obvious sources such as air bells, or they
may originate in many more obscure ways due to such things as water
passing unevenly over a deflector surface. So far as possible, sources of
292
trouble have been mentioned earlier when dealing specifically with faults
originating at each item of equipment, but connecting a particular fault
with its cause is all too frequently extremely difficult and it is in this realm
that the experience of a good machineman is invaluable.
The machineman takes continual note of the appearance of the sheet
on the wire and the general forming conditions, with particular emphasis
on any sign of unevenness which recurs as the wire comes round each
revolution. In this way a small crack or hole in the wire can be detected
very quickly and rectified before becoming serious. The accumulation of
slime or pitch in patches on the wire can also be noted by observing uneven-
ness of drainage, though often the first indications of this sort of trouble,
particularly on faster machines, occur when holes are noticed in the paper
and breaks occur.
Early detection of slime and pitch on the wire may prevent the necessity
for a thorough clean-up of all parts of the wire involving removal of forming
boards, doctors, and suction boxes. Both deposits can be difficult to
dislodge and the decision to shut a machine for cleaning is not made
easier by knowing that once started there is a tendency for both slime and
pitch to accumulate more rapidly, particularly on those areas already
affected which may not get completely cleaned. Sometimes, especially in
the case of pitch trouble, time can be saved whenever practicable by
removing suction boxes one at a time before stopping the wire; this has
the advantage of preventing any pitch on the boxes from being pushed into
the wire mesh when the machine is shut. Removal of slime from the wire
mesh is generally easiest with a steam jet, but to dislodge bad patches may
need a suitable solvent together with scrubbing with a brush with hard
bristles (but not so hard as to scratch and wear the wire). Pitch is tackled in
a similar way and often chalk or clay may be rubbed over the affected areas
first to help push it out of the mesh. In both cases it is useful to place a
special tray under the wire to collect the particles blown through, otherwise
when re-starting the backwater system may be full of small specks which
give further trouble.
Some useful indications can also be obtained from watching the surface
of the stock mat for other types of irregularities. The line on the surface
made as the mat passes over each table roll should be perfectly straight and
stable; wavering indicates variations in the mat conditions and a regular
oscillation may be traced to unevenness in the roll itself, indicating the
desirability for changing and re-covering, or to a vibrational disturbance
in the stock flow system. The couch line or take-off should also be even
without fluttering, and a permanent depression in the line could indicate a
blocked spray causing the wire to become made-up or some other fault
affecting drainage rate at that particular position across the mchine.
Changes in position of the take-off line which occur under unaltered
conditions of the draw tension can also give a pointer to changes in the
general substance level and the stock drainage conditions: if the take-off
line moves back up the wire (i.e. the draw appears tighter) and at the same
time the dry-line has moved forward on the wire, it is likely that the sheet
has become heavier; but if on the other hand the take-off line moves back
293
but the dry-line has also moved back, then on many machines this is a
clear sign that the stock is freer. Opposite indications are given when the
take-off line moves further forward on the wire (i.e. the draw appears
slacker): in this case if the dry-line has also moved forward on the wire it
is quite likely that the stock has come wetter; on the other hand, if the
dry-line has moved back the sheet is probably lighter. To make this clear,
changes in the position of take-off line and couch line are produced as
follows:
Dry-line Change in
movement Take-off line movement property
Forward Back (appears tighter) Sheet heavy
Back Forward (appears slacker) Sheet light
Forward Forward (appears slacker) Stock wet
Back Back (appears tighter) Stock free
The validity of these interpretations depends, for any particular machine,

on the relative movement of the dry-line that occurs as a result of fluctua-
tions in sheet substance or stock wetness; on some machines the degree of
alteration in either substance or stock wetness which is normally encoun-
tered during operation may affect the drainage rate, and hence the dry-line
position, much more than the other, and so becomes the dominant factor
determining the position of the dry-line. Further, the validity depends on
the extent to which changes in adhesion of web to wire at the couch, and
also (since it alters the effective tension) in the stretch of the sheet in the
open draw, are produced on the one hand by alterations in the solids
content at the couch, and on the other by changes in composition of the
sheet as a result of differences in stuff treatment. Normally, for example, a
drop in solids content occasioned by a substance increase will reduce ad-
hesion to some extent (causing the take-off line to move back up the wire);
but if the stock works wetter on the wire, though this results in a similar for-
ward movement of the dry-line, the take-off line tends to move in the oppo-
site direction (forward down the wire) due to a combination of greater
adhesion to the wire and greater stretch occasioned by the change in stock
composition (the latter because greater stretch lessens effective tension for a
given draw). Despite these qualifications, the relationships listed above
certainly appear to occur to a significant extent on many machines and the
machineman finds that careful observation of such signs, together with
similar indications that may be obtained at press draws or from the vacuum
on the suction boxes or a suction couch, enables him to detect changes and
make corrections far sooner than if he waits for confirmation from test
results on the paper.
Breaks at the couch are often the first sign of trouble and require some
ingenuity to remedy because of the numerous sources that are possible.
Any weakness in the sheet caused by uneven drainage at some point of the
wire due to pitch, slime, or other causes may produce a break which is
more likely to occur at the couch than anywhere else; in addition lumps of
fibre breaking away after accumulating in the breast box, slice lip, wire
apron or other places, and poor edges caused by the deckle, dandy or
294
unevenness of contact at the couch, these and a host of other minor ailments
can all first show themselves at the couch.
Trouble can also result indirectly from quite a small alteration made by
the machineman in the course of his inspection of the wire part; for
instance, slackening back the tension in the wire as it ages might increase
the wrap round the breast or table rolls to just the extent which causes an
appreciable increase in drag on the forming board or deflectors, and this in
turn affects the stability of draw at the couch making the presence of
occasional faults in the sheet which hitherto passed the couch successfully
suddenly sufficient to create a break. Much of papermaking is affected by
the occasional appearance for no known reason of trouble of this nature
which suddenly exhibits its presence by causing breaks at the couch or
press, or by making blemishes in the paper, and then as suddenly disappears
followed by a sigh of relief from the operators. More instrumentation and
control, better maintenance procedures, and the keeping of comprehensive
long-term records should reduce such occurrences considerably, but it
would be an optimist who predicted their eventual elimination.
295
REFERENCES
(1) Hansen, E.: 'The Distribution of Filler in Paper.' TAPPI 34, 4, 180. April, 1951.
(2) Dixon, W. B.: 'The Operation of a Stacked Press With Vacuum Transfer on an
English Newsprint Machine.' P. and P. Mag. Can. 53, 11, 115. October, 1952.
(3) Hendry, I. F., Mardon, J., Scruton, H., and Wilkinson, D. M.: 'Drainage on a
Tissue Machine Wire Table.' Proc. Tech. Sec. 33, 3, 603. December, 1952.
(4) Rance, H. F.: 'The Influence of Papermaking Variables upon Sheet Charac-
teristics.' Proc. Tech. Sec. 33, 1, 173. February, 1952.
(5) Brauns, 0. and Oskarsson, R.: 'Drainage at the Flat Boxes and Suction Couch
Roll.' Svensk Pap. 56, 18, 691. September 30, 1953. (In Swedish). Also in English
in Paper Maker 127, 3, 190. March, 1954.
(6) Mansell, G. F. J., and Saunders, K. C.: 'A Novel System of eliminating Bubbles
on the Fourdrinier Wire.' Proc. Tech. Sec. 36, 3, 535. October, 1953.
(7) Paton, A. W.: 'Machine Wire Maintenance.' Proc. Tech. Sec. 34, 3, 437.
December, 1953.
(8) Andersson, 0., and Bergstrom, J.: 'Influence of Differential Speeds Between
Stock and Wire on Sheet Properties.' TAPPI 37, 11, 542. November, 1954.
(9) Bennett, W. E.: 'Quantitative Studies of Water Removal by Table Rolls.'
TAPPI 37, 11, 534. November, 1954. Also P. and P. Mag. Can. 55, 13, 131.
December, 1954.
(10) Brauns, 0., Bergstrom, J., and Svenson, A.: 'Formation of the Web on a Paper
Machine Wire. The Influence of Shake.' Svensk Pap. 57, 19, 693. 1954. (In
Swedish).
(11) Finger, E. R., and Majewski, Z. J.: 'Sheet Formation on the Fourdrinier
Machine.' TAPPI 37, 5, 216. May, 1954.
(12) Holden, H.: 'Some Observations on Suction Couch Rolls, Suction Press Rolls
and Vacuum Transfer.' Tech. Bull. Tech. Sec. Brit. Paper & Board Makers'
Assoc. 31, No. 1 : 7-12. February; No. 2 : 41-6. April, 1954.
(13) Lyne, L. M., and Gallay, W.: 'Studies in the Fundamentals of Wet-Web
Strength.' P. and P. Mag. Can. 55, 11, 128. October, 1954.
(14) Nordman, L.: 'Laboratory Investigation of Water Removal by a Dynamic
Suction Box.' TAPPI 37, 11, 553. November, 1954.
(15) Underhay, G. F.: 'Drainage Effects and Two Sidedness.' TAPPI 37, 11, 547.
November, 1954.
(16) Alston, M. P., Goodhew, I. E., and Chapman, J.: 'The Study of Formation on
the Fourdrinier Wire.' Proc. Tech. Sec. 36, 3, 535. December, 1955.
(17) Baggallay, M. B.: 'Some Theories and Experiments in Couching.' P. and P.
Mag. Can. 56, 9, 131. August, 1955; also Canadian Tech. Sec. Proc. p. 517, 1956.
(18) Duskin, A. W., and McCaughan, A. C.: 'What Crosset learned about Grooved
Table Rolls on a Light Weight Machine.' Paper Trade Journal 139, 51, 48.
December 15, 1955.
(19) Hendry, I. F.: 'Suction Couch Rolls-Some Observations on their Performance
and Effect on the Paper.' TAPPI 38, 9, 163. September, 1955.
(20) Brecht, W.: 'The Effects of Wire Shake.' Woch. fur Pap. 84, 11/12, 446. 1956.
(In German).
(21) Burkhard, G., and Wrist, P. E.: 'Investigation of High Speed Paper Machine
Drainage Phenomena.' P. and P. Mag. Can. 57, 4, 100. March, 1956.
(22) Mardon, J. and Shoumatoff, N ..: 'Flocculation, Formation and Headbox
Design.' Part IV. P. and P. Mag. Can. 57, 3, 305. Convention Issue, 1956.
(23) Robinson, W. F. E.: 'The Instrumentation of a Paper Machine and the Results
Obtained.' Proc. Tech. Sec. 37, 2, 217. June, 1956.
(24) Boiteux, P.: 'The Effect of Drainage Characteristics of the Machine Wire and of
the Machine Speed on the Wear of Suction-box Covers.' Assoc. Tech. Ind.
Papetiere, Bull. No. 5 : 192-8. 1957. (In French).
296
THE WIRE SECTION
(25) Burnett, R. W.: 'Vacuum Transfer from Wire to Press Section.' P. and P. Mag.
Can. 54, 11, 132. November, 1957.
(26) Hitchings, R. G.: 'The Effects of Shake in Slow Speed Fourdrinier Operation.'
(27) Jordansson, L.: 'The Press Section with Vacuum Transfer. Part I. General
Operating Conditions and Water Removal.' Svensk Pap. 60, 5, 59. 1957. (In
Swedish).
(28) Muller-Rid, W., and Pausch, G.: 'Difficulties in Sheet Formation on the Wire
of a High Speed Paper-machine.' Zellstofund Pap. 6, 10, 307. 1957. (In German).
(29) Pritchard, E. J.: 'Two Techniques for Studying the Distribution of Loading in
Paper.' Proc. Tech. Sec. 38, 2, 425. June, 1957.
(30) Telepnev, A. N., and Sokolov, N. N.: 'Sheet Drainage at the Suction Boxes.'
Bumazh. Prom. 3, 5. 1957. (In Russian).
(31) Cole, E. J.: 'How the Sheet is Laid on the Wire.' Paper Trade Journal 142,
49, 44. December 8, 1958.
(32) Dixon, P.H.: 'A Comparison of the Operation of Two Wet Felts.' Also Discus-
sion. Proc. Tech. Sec. 39, 2, 173. June, 1958.
(33) Friedlander, P. H.: 'The Measurement of Fibre Orientation in Newsprint with
respect to the Machine Direction by X-Ray Diffraction.' P. and P. Mag. Can.
59, 1, 102. January, 1958.
(34) Gavelin, G.: 'The Case of the Open Draw.' Svensk. Pap. 61, 9, 282. May 15, 1958.
(35) Judt, M.: 'Fibre Orientation in Paper.' Das Papier 12, 21/22, 568. 1958. (In
German), Summary in Tech. Bull. Tech. Sec. Brit. Paper & Board Makers'
Assoc. 36, 4/5, 128. August/October, 1959.
(36) Kitano, A.: 'Post-up transfer from Wire to Press Section.' Paper Trade Journal
142, 14, 27. April 7, 1958.
(37) Mardon, J., Truman, A. B., O'Blenes, G., and Meadley, K.: 'A Consideration
of the Factors Involved at the Open Draws at Couch and Presses of Fourdrinier
Machines.' P. and P. Mag. Can. 59, 9, 135. September, 1958.
(38) Moen, C. J.: 'A Formation Standard.' TAPPI 41, 4, 212. April, 1958.
(39) Millier-Rid, W., and Pausch, G.: 'Contribution to the Problem of Water
Removal from Fibrous Webs by Means of Vacuum.' Woch. for Pap. 86, 8,
298. End April, 1958. (In German).
(40) Rodman, J.: 'A Survey of the Operating Characteristics of the Original and
Latest Designs of a Suction Pick-Up.' Paper Maker International Number,
p. 48. 1958.
(41) Ruch, H., and Krassig, H.: 'The Determination of Fibre Orientation in Paper.'
P. and P. Mag. Can. 59, 6, 183. June, 1958.
(42) Boadway, J. D., Friese, J., and Husband, R. M.: 'Some Studies of Research
and Development Department, Consolidated Paper Corporation Ltd.' P. and P.
Mag. Can. 60, 8, T23 l. August, 1959.
(43) Brazington, E. S., and Rad van, B.: 'A Method of Determining the Substance over
Small Areas of Paper.' TAPP! 42, 7, 545. July, 1959.
(44) Brecht, W., Raderer, P., and Weitzel, W.: 'Tendency of Paper to Curling.'
Das Papier 13, 11/12, 237. June, 1959. (In German).
(45) Glynn, P., Jones, H. W. H., and Gallay, W.: 'The Fundamentals of Curl in
Paper.' P. and P. Mag. Can. 60, 10, T316. October, 1959.
(46) Howe, B. I. et al.: 'The Characterization of Couch Draw Stability.' P. and P.
Mag. Can. 60, 4, T99. April, 1959.
(47) Judt, M.: 'The Effect of Shake of Paper Machines on Sheet Formation and
Fibre Orientation.' Das Papier 13, (3, 4), 46. February, 1959. (In German).
(48) Judt, M.: 'The Influence of Wire Shake on the Technical Properties of Paper.'
Das Papier 13, 19/20, 483. October, 1959. (In German).
(49) MacDonald, C. E. et al.: 'Study of Newsprint Machine Wire Life.' P. and P.
Mag. Can. 60, C. 185. Convention Issue, 1959.
(50) Mardon, J., and Truman, A. B.: 'The Wake Effect, Ridge Formation and Spout
Development on the Wire of a Fourdrinier Machine.' Paperi ja Puu, 4i, 9,
391 and 41, 10, 457. 1959.
(51) Radvan, B. and O'Blenes, G.: 'Principles of Adhesion on a Paper Machine.'
TAPPI 42, 12, 921. December, 1959.
297
THE WIRE SECTION
(52) Ure, R. A., Nuttall, G. H., and 'Ad-Lignum.': 'Fourdrinier Wire Slope.' Paper
Maker 137, 4, 41. April, 1959.
(53) Attwood, B. W.: 'A Laboratory Investigation of Dynamic Drainage at Vacuum
Boxes.' P. and P. Mag. Can. 61, C, T-96. Convention Issue, 1960.
(54) Friese, J.: 'The Improvement of Paper Machine Wire Life.' P. and P. Mag. Can.
61, 10, T-467. October, 1960.
(55) Glynn, P., Jones, H. W. H., and Gallay, W.: 'Drying Stresses and Curl in Paper.'
Canadian Tech. Sec. Proc., p. 378. 1960.
(56) Howarth, J., and Banks, W. A.: 'Machinery for Automation.' Paper Technology.
1, 3, Tl 12. June, 1960.
(57) Ives, A. A.: 'Suction Couch Operation on Newsprint Machines.' Paper Tech-
nology 1, l, 43. February, 1960.
(58) MacLaurin, M. I.: 'New Instruments for Control of the Paper Machine.'
Paper Technology 1, 4, 381. August, 1960.
(59) Mardon, J., Meadley, C. K., Duchesne, J.P., and Howe, B. I.: 'An Experimental
Investigation of the Problem of Breaking and Tearing of the Wet Web on the
Fourdrinier Paper Machine.' P. and P. Mag. Can. 61, 7, T-341. July, 1960.
(60) Mills, R.: 'Development of the World's Fastest and Widest Rotabelt Suction
Unit.' Paper Trade Journal 144, 18, 36. May 2, 1960.
(61) Shamolin, I. S.: 'Factors Affecting Drainage on the Wire Part.' Bumazh. Prom.
35, 9, 6. 1960. (In Russian).
(62) Tellvik, A., and Brauns, 0.: 'Studies of Table Roll Drainage.' Svensk Pap.
63, 22, 803. November 30, 1960.
(63) Brecht, W., and Weitzel, W.: 'The Specific Cleaning action of different types of
Wire Spray Pipes.' Das Papier 15, lOa, 635. October, 1961. (In German).
(64) Campbell, J. G.: 'Simple Machine Design Changes Could Reduce Couch
Breaks.' Paper Trade Journal 145, 13, 20. March 27, 1961.
(65) Hill, V. C., and Clark, C. L.: 'Function of Flo-Vac Rotary Suction Box on
Fourdrinier Machines.' TAPP! 44, 6, 218A. June, 1961.
(66) MacPherson, R.: 'Interim Report on the Performance of Formex Fabrics on
No. 2 P.M. at Laurentide Division of Consolidated Paper Corp. Ltd.' P. and P.
Mag. Can. 62, 7, T-351. July, 1961.
(67) Mardon, J., Truman, A. B., Howe, B. I., and Meadley, C. K.: 'A Review of the
present state of knowledge of the Mechanism of Water Removal on the Four-
drinier Paper Machine.' Paperi ja Puu 43, 8, 457; 9, 499; 10, 575; 11, 581.
August-November, 1961.
(68) Parker, J. D.: 'Recent Theoretical Work in the Area of Sheet Formation.'
TAPPI 44, 4, 162A. April, 1961.
(69) Vercoe, J. E.: 'Fourdrinier Maintenance-Key to Machine Efficiency and Wire
Life.' Paper Trade Journal 145, 14, 36. April 3, 1961.
(70) Attwood, B. W.: 'A Study of Vacuum Box Operation.' Paper Technology 3,
5, Tl44. October, 1962.
(71) Attwood, D., and Parker, J. R.: 'Basis Weight Variations Over Small Areas
of Paper.' Paper Technology 3, 5, Tl33. October, 1962.
(72) Emerton, H. W., Page, D. H., and Hale, W. H.: 'Structure of Papers as seen in
their Surfaces.' Formation and Structure of Paper. Transactions of the September
1961 Oxford Symposium, p. 53. B.P. & B.M.A. Tech. Sec. publication, 1962.
(73) Forgacs, 0. L. and Atack, D.: 'Distribution of Chemical Wood pulp and Ground-
wood through the thickness of Newsprint.' Formation and Structure of Paper.
Transactions of the September 1961 Oxford Symposium, p. 721. B.P. & B.M.A.
Tech. Sec. publication, 1962.
(74) Glynn, P. and Gallay, W.: 'Further studies on the Mechanism of Curl in Paper.'
P. and P. Mag. Can. 63, 8, T-418. August, 1962.
(75) Groen, L. J.: 'Fundamental Aspects of Filler Distribution in Paper.' Formation
and Structure of Paper. Transactions of the September 1961 Oxford Symposium,
p. 697. B.P. & B.M.A. Tech. Sec. publication, 1962.
(76) Hinton, A. J.: 'The problem of Two-sidedness in coloured papers.' Paper
Technology 3, 3, 229. June, 1962.
(77) Johnson, B., and Gavelin, G.: 'Investigation of Wire Wear on Swedish Paper
Machines.' Svensk Pap. 65, 22, 897. November 30, 1962. (In Swedish).
298
THE WIRE SECTION
(78) Lawson, J. J. A., and Lambert, J. E.: 'Newsprint Machine Wire Life and Silicon
Carbide Flatbox Covers.' P. and P. Mag. Can. 63, 11, T544. November, 1962.
(79) Osterberg, L.: 'Transfer of Paper Webs from Press Rolls.' Svensk Pap. 65, 6, 222.
March 31, 1962. (In Swedish).
(80) Rad van, B., Dodson, C., and Thompson, A.: 'Wet End Breaks and Related
Problems in operating Open Draws.' Paper Technology 3, 3, 215. June, 1962.
(81) Robertson, A. A., and Mason, S. G.: 'Wet-end Factors affecting the Uniformity
of Paper.' Formation and Structure of Paper. Transactions of the September
1961 Oxford Symposium, p. 791 B.P. & B.M.A. Tech. Sec. publication, 1962.
(82) Smith, R. T.: 'Application of the Roll-less Flo-Vac on a Fourdrinier Machine.'
TAPP! 45, 10, 830. October, 1962.
(83) Wahlstrom, B., and O'Blenes, G.: 'The Drainage of Pulps at Papermaking Rates
and Consistencies Using a new Drainage Tester.' P. and P. Mag. Can. 63, 8,
T-405. August, 1962.
(84) Wrist, P. E.: 'Dynamics of Sheet Formation on the Fourdrinier Machine.'
Formation and Structure of Paper. Transactions of the September 1961 Oxford
(85) Boadway, J. D., and Gray, J. T.: 'The Role of the Supporting Medium in Drain-
age Resistance.' P. and P. Mag. Can. 64, l, Tl4. January, 1963.
(86) Dyer, H.: 'Foils and Fabrics mean more and better Paper for Domtar.' Paper
Trade Journal 147, 15, 39. April 15, 1963.
(87) Eastwood, G. G. M.: 'A Continuous Two-Spot Formation Recorder.' TAPP!
46, 1, 32. January, 1963.
(88) Forgacs, 0. L., and Strelis, I.: 'The Measurement of the Quantity and Orienta-
tion of Chemical Pulp Fibres in the Surfaces of Newsprint.' P. and P. Mag. Can.
64, 1, T3. January, 1963.
(89) Hendry, I. F., and Newman, J. A. S.: 'Effect of Machine Variables on the Curl
of Paper.' Paper Technology 4, 4, 381. August, 1963.
(90) Manson, D. W.: 'The Effect of Selected Paper Machine Variables on Formation.'
TAPPI 46, 8, 472. August, 1963.
(91) Prusas, Z. C.: 'Laboratory Study of the Effects of Fiber Orientation on Sheet
Anisotropy.' TAPPI 46, 5, 325. May, 1963.
(92) Redfern, A. P., and Gavelin, G.: 'A Study of Conditions affecting Wire Life.'
P. and P. Mag. Can. 64, 3, T-165. March, 1963.
(95) Schroder, S., Johnson, B., and Gavelin, G.: 'Drive Studies on Paper Machines.'
P. and P. Mag. Can. 64, I I, T-478. November, I963.
(96) Sisler, G. C., and Maves, K. L.: 'Comparing Table Rolls and Ladding Wire
Doctors.' Paper Trade Journal 147, 8, 30. February, 1963.
(97) Lettenberger, r.: 'A Comparison of Hydrofoils with Table Rolls on a Four-
drinier Machine.' TAPP! 47, 10, 192A. October, 1964.
(98) Svensson, 0., and Brauns, 0.: 'Function of Wet Suction Boxes.' Svensk Pap.
Part 1: 67, 19, 772. October I5; Part 2: 67, 21, 860. November 15, 1964. (In
Swedish).
(99) Yih, C-S., and Lin, S. P.: 'Effect of Variation of Acceleration on Free-Surface
Instability.' TAPPI 47, 2, 88. February, 1964.
(100) Abbott, E. C.: 'Practical Considerations and Results obtained from the Applica-
tion of Hydrofoils to a Fine Paper Machine.' P. and P. Mag. Can. 66, 10, T-538.
October, 1965.
(101) Anon: 'New Suction Box Design and Cover Material.' Paper Trade Journal
149, 3, 38. January 18, 1965.
(102) Clark, R. C.: 'New "Vacu-foils" provide Adjustability, High Dewatering Power.'
Paper Trade Journal 149, 25, 52. June 21, 1965.
(103) Descary, I. G.: 'Dominion Adjustable Foils.' P. and P. Mag. Can. 66, 12,
T-611. December, 1965.
(104) Doane, F. P.: 'The Use of Synthetic Forming Fabrics on Fine Papers.' P. and P.
Mag. Can. 66, 6, T-313. June, 1965.
(105) Roecker, A. C.: 'Evaluation of Fourdrinier Foil Designs.' TAPP! 48, 7, 109A.
July, 1965.
299
THE WIRE SECTION
(106) Schroder, S., and Svensson, 0.: 'The Influence of Efflux Ratio on the Properties
of Paper.' Svensk Pap. 68, 2, 25. January 21, 1965. (In Swedish).
(107) Svensson, 0., and Brauns, 0.: 'Forming Boxes increase Production and improve
Quality.' Paper Trade Journal 149, 14, 38. April 5, 1965.
(108) Svensson, 0., and Osterberg, L.: 'Function of Wet Suction Boxes.' Part 3.
Svensk Pap. 68, 11, 403. June 15, 1965. (In Swedish).
(109) Thorp, B. A.: 'Hydrofoil Design and Application is an Evolving Science.'
Paper Trade Journal 149, 2, 37. January 11, 1965.
(110) Dushnicki, T. S.: 'A Laboratory Investigation of the Effect of the Drainage
Forces on Wire Mark.' P. and P. Mag. Can. 67, 8, T-356. August, 1966.
(111) Hanson, V. E.: 'Control of Wire Table Drainage-Foils show Versatility.'
Paper Trade Journal 150, 52, 34. December 26, 1966.
(112) Luhde, F.: 'The Effect of Wire Speed and Dandy upon Sheet Formation and
Wire Mark.' P. and P. Mag. Can. 67, 9, T-381. September, 1966.
(l 13) Manson, D. W., and Gillis, R. C.: 'Effect of the Fourdrinier Shake on Forma-
tion.' P. and P. Mag. Can. 49, 10, 2A. October, 1966.
(114) Mardon, J. et al.: 'Free Surface Flow for Papermaking.' P. and P. Mag. Can. 67,
7, T-314. July, 1966.
(115) Norman, R. J.: 'Dependence of Sheet Properties on Formation and Forming
Variables.' Consolidation of the Paper Web. Transactions of the September 1965
Cambridge Symposium, p. 269. B.P. & B.M.A. Tech. Sec. publication, 1966.
(116) Page, D. H., Sargent, J. W., and Nelson, R.: 'Structure of Paper in Cross-Section.'
Consolidation of the Paper Web. Transactions of the September 1965 Cambridge
(117) Rad van, B., Dodson, C., and Skold, C. G.: 'Detection and Cause of the layered
Structure of Paper.' Consolidation of the Paper Web. Transactions of the Sep-
tember 1965 Cambridge Symposium, p. 189. B.P. & B.M.A. Tech. Sec. publica-
tion, 1966.
300
PART 4
THE PRESS SECTION

INTRODUCTION
41 In one of the early classical text-books on papermaking the press
section is dismissed in just under one page, and part of that is covered by a
photograph of a typical old plain straight-through press. It is only in
comparatively recent times that this neglect of the presses has been remedied
and during the last few years a considerable amount of research has been
devoted to this essential part of the Fourdrinier. Normally the press section
provides little trouble to the machine operators, at least once the sheet has
been successfully threaded through, and apart from attending to the con-
dition of the wet felts and seeking to remedy the occasional wet streak
which can be attributed to the press operation very little attention is
required or given. The pressure applied in the presses is frequently kept the
same for long periods once a new wet felt has been run in, and alterations
are then more frequently to relieve pressure than to apply more.
Except in the rare case where some restriction to the quantity of steam
available for the drying cylinders puts a limit on production and highlights
the importance of running the paper leaving the presses as dry as possible,
the desirability of this objective is rarely in the forefront of the operator's
mind. One of the reasons for this is undoubtedly the difficulty of assessing
at all accurately whether the press is functioning as efficiently as it might.
An excessive demand for steam is one indication that the moisture content
entering the cylinders is too high but, except on some machines with hoods
and an efficient controlled air recirculation system, there are many other
factors which can alter steam requirements; in particular, the moisture
content of the paper at the reel-up is important though usually it is not
easy to keep a close check on this property. Changes in press performance
which are not due to definite malfunctioning are all fairly gradual in nature
and this increases the difficulty of keeping a high efficiency of water removal
by visual observation. Alterations over a period of the pattern of water
throw-out from the shell of a suction press or the flow of water from the
press trays are not of a nature to cause the operator to take any action until
they become very obvious.
Yet the efficiency with which a press removes water is by far the most
important criterion for assessing performance. The properties of paper can
alter very slightly as a result of pressing under different conditions, but
not to an extent which is usually of any great importance in the manu-
facture of most grades, at least within the normal variations in running
found on presses. Indeed the precise effect of wet pressing is the subject of
some difference of opinion. Generally it has been thought that as a result of
the promotion of fibre bonding under pressure, bulk is decreased and
strength and smoothness increased in pressing. But recent work (57, 58)
indicates that although consolidation of paper structure takes place it may
303
41. l THE PRESS SECTION
in some circumstances be due to the collapse of individual fibres which

could damage bonds already formed and produce a reduction in strength
properties. The response to press loading appears to depend on the
strength of bonds already formed, the relative intensity of local hydraulic
pressure in the nip, and on the deformation characteristics of the individual
fibres involved. Other work (54) has indicated that reduction of bulk is
greater when the same solids content entering the dryers is achieved as a
result of three separate presses working at a lower linear pressure instead of
two with a high linear pressure. Also, it was observed that although
smoothness improved on the top side of the sheet it was impaired on the
wire side due to felt mark.
In practice the only noticeable effect of the presses on quality is when
excessive pressure causes structural deformity in the sheet and leads to
crushing. At loads slightly under the crushing point deterioration of quality
occurs as fibres are washed out and re-aligned in the press nip, but when
the crushing point itself is reached the web stretches and eventually folds
over into creases and breaks up. Up to the point where these adverse
effects are apparent, increasing the load in the presses has most influence
in improving the efficiency of water removal. But there are, of course,
many other factors affecting press efficiency and it is proposed to discuss
these in some detail so that their relative importance to any particular
application can be assessed more easily.
41. l Importance of dryness leaving the presses

It cannot, however, be stressed too much that obtaining as dry a sheet as
possible leaving the presses consistent with keeping up the quality of the
paper is one of the machine operator's most important tasks. This applies
not only to the average moisture content of the sheet but also the evenness
of the moisture profile; a section across the sheet which is wetter than the
rest of the web entering the cylinders will almost always show up damper
at the reel-up and the dryerman will then control his reel-up moisture
content to that part of the sheet and dry down the rest of the sheet more
than is necessary.
As an approximate guide, the cost of drying an equal weight of water in
the cylinder section of a paper machine is usually quoted at about seven to
ten times the cost in the presses, though this of course depends a great deal
on the relative steam and power costs and the efficiency of heat usage.
In several text-books graphs and tables show that if, for example, the water
to fibre ratio of the paper leaving the presses increases by one third from
l ·5 to 2·0, then the approximate steam consumption changes from 2·0 to
2·7 lb. steam per lb. paper, which is also an increase of almost one third,
the slight discrepancy in the calculation being due to allowance for water
left in the sheet at reel-up. In fact these figures are derived from average
conditions and the degree of change in steam consumption is too pessimistic
because it does not take account of two important considerations. Firstly,
a proportion of the steam used on any machine serves only to balance heat
losses in the system and is not used to evaporate water from the sheet;
304
THE PRESS SECTION 41.2
secondly, less heat is required to remove water in the sheet when it is
relatively damp than when it is almost dry, so that moisture added due to
press inefficiency does not need an exactly proportional increase in heat to
remove it in the drying section. A more realistic picture is obtained if it is
considered that additional steam usage is one half to two thirds that of the
theoretical figure derived from a simple proportional calculation based on
the percentage increase in water/fibre ratio at the press in the manner
mentioned above. Nonetheless when the relative costs of the water removal
are taken into account it is still at least five times as expensive to dry
additional moisture in the drying section than to remove it in the presses.
41. 2 Representation of moisture figures

Before commencing a description of the theory of pressing it is worth
noting the importance of quoting moisture figures at the press, as on the
wire, in terms of water to fibre ratio. Too often results are quoted in the
form of percentage moisture or dryness figures, although these do not
convey the degree of change in the actual water content of the sheet
adequately. To illustrate this point consider the table below which covers
the usual range of moisture values applicable to the press section:
Moisture content % Dryness% Water/fibre ratio

85·7 14·3 6
83 ·3 16·7 5
80·0 20·0 4
75·0 25·0 3
71 ·4 28·6 2·5
66·7 33·3 2
60·0 40·0 1 ·5
50·0 50·0 1
For a unit change in the water/fibre ratio, and hence in the quantity of
water remaining to be removed from the sheet, the moisture content
percentage alters appreciably less the higher the water/fibre ratio, i.e. the
wetter the sheet, and this confuses assessment of press performance.
Moisture content figures are preferably confined to the reel-up and drying
section where they are more useful and representative of the properties of
the paper.
305
CHAPTER 4A
GENERAL THEORY OF PRESSING
4A.1 EARLY THEORY
An interest in discovering precisely what happens in the nip of a press has,
in common with interest in other parts of the machine, been stimulated in
post-war years by the general increase in machine speeds. Press performance
has been observed to fall off as higher speeds have been attained and
research has been directed to discover why. In addition the felt manu-
facturers, anxious to keep up and improve their products in the face of
these increasing speeds, have devoted time and money in an endeavour to
determine the precise role of the felt in the press. The culmination of this
interest came in 1960 when the publication of a paper by Wahlstrom (21)
completely overturned the theory that was generally accepted as describing
the operation of a press. Since then a great deal of confirmatory evidence
has been published and Wahlstrom's theory must be accepted as approxi-
mating closest to experimental observations. The comparative newness of
the theory and the controversy that preceded its appearance makes it
desirable to give a fair, if brief, recapitulation of the old theory first.
4A.1 1 Nissan's theory of pressing

The theory that held favour up to recent years was propounded most fully
by Nissan (5). Within the ingoing side of the press nip the compression was
assumed to close up and saturate the felt causing water to be expressed
from it either back out of the nip in the case of a plain press or into the
holes in the case of a suction roll press. The main function of this part of
the nip was to remove the water carried round by the felt; apart from a
small compression of the paper, little or no water was imagined to leave
the sheet which was almost as wet at the centre of the nip as when it
entered the press. Within the outgoing side of the nip the gradual relaxation
of pressure allowed the felt to expand and this action was envisaged as
providing a suction force drawing water from the paper to the felt and to be
responsible for removing the greater part of the water from the web.
The consequences of this theory were that it appeared desirable to have
a felt which was as compressible and elastic as possible, as well as being
permeable through the thickness of the felt and along the length. The felt
should, in fact, be bulky and possess large interstices compatible with not
marking the paper. As the water was removed from the web essentially by
being sucked into the felt in the outgoing side of the nip it would be prefer-
able for this to take place over as long a time as possible, which implied a
wide nip width and allowing the paper and felt to remain in contact as long
as possible.
306
THE PRESS SECTION 4A.12
4A. 1 2 Observations of Osborn and Wicker
There were certain aspects of this theory which did not seem quite to fit
observations of the press section. As early as 1956 Osborn and Wicker (7)
had published the results of an interesting investigation in which a beta ray
gauge was used to determine the moisture in a first press felt while it was
running on a paper machine. When increasing the total load on the press
they discovered, as expected, that a greater quantity of water was removed
from the press roll tray. But also it appeared from the beta ray gauge
measurements that the felt carried round less water as the nip pressure
increased. The relevant graphs showing the variation between the water
removed and the percentage of water in the felt as the nip pressure altered
are shown in Figs. 4. 1 and 4. 2.
IS.____.........,.--~~.--~~--~~---~~~-
loo IZS I So 175 '200 225.
PRESS NIP PRessuRe, Le,s PER LINEAR OJCH.
Fig. 4. 1. Relation between water removal and nip pressure (after Osborn and Wicker)
Although Osborn and Wicker made no comment in their paper, the

theory current at that time could not satisfactorily explain the decrease in
water carried round by the felt as the nip pressure increased. If suction of
water from the paper to the felt in the outgoing side of the nip were the
primary cause of water removal it would be expected that when more water
was extracted from the web into the felt due to the increased pressure this
additional water should either have no effect or lead to an increase, not a
decrease, in the water content of the felt. It is possible to elaborate the
theory to explain this discrepancy, but a straightforward application of the
theory undoubtedly leads to a contradiction.
Another observation made in the same experiment was also unexpected.
When a break occurred and paper stopped passing through the press nip,
the moisture content of the felt was observed suddenly to increase by about
307
4A. l 3 THE PRESS SECTION
2·5 per cent. If the felt were extracting water from the web in the outgoing
side of the nip and the water was suddenly not there to be extracted, it is
evident that the moisture content of the felt should decrease. Osborn and
Wicker sought an explanation of this phenomenon in terms of the reduction
of suction press vacuum (from 10 in. Hg. to 5 in. Hg.), which occurred at
the same time due to a greater air flow through the felt when the paper was
no longer present to seal it. This, they argued, would result in less effective
retention of the water by the suction roll and consequent partial rewetting
z 38·5
c:t
w
~
~ .!>8·o
Cl
'"
\!)
~ 37-5
UI
u
rt
~ 37·o ,__~~~~~-..,.~~~,-~~.....,.~~..:l-.....-
\oo \Z.S lSo 175 Zoo 225
PRi5S NIP PRESSURE:, L~S. PER. L.lNEA~ INCH.
Fig. 4. 2. Relation between moisture content of felt and nip pressure (after Osborn
and Wicker)
of the felt as it emerged from the nip. This explanation is now known to be
incorrect mainly because, as will be seen later, alteration of the vacuum of
a suction press has relatively little effect on water movement in the holes
of the press roll.
4A. 1 3 Experimental work by Sweet

One further set of experimental observations although not published until
after Wahlstrom's paper appeared were known to him and provided
evidence completely contradictory to the old theory of pressing. This work
has been described by Sweet (26) and involved the use of an experimental
press. Sheets of paper were made up from a varying number of layers of
thin wet sheets of tissue and these were passed through the press with
different felts. The reduction in moisture content of the paper resulting
from the pressing was determined. Under any particular set of press
conditions and with the same felt, increasing the number of layers which
comprised the sheet, i.e. the thickness and substance of the sheet, always up
to a certain point resulted in the pressed sheet having a lower moisture
308
THE PRESS SECTION 4A. l 3
· content. Figure 4. 3 illustrates a typical result. In other words the thicker
the sheet, despite the greater quantity of water entering the nip in the paper
the more water proportionally was pressed out. If water were removed from
the sheet by suction in the outgoing side of the nip it is impossible to explain
why the presence of a thicker sheet should cause so much more water to
be removed that the sheet emerged at a lower moisture content than it
would were it thinner.
The only explanation of this unexpected result is that water is removed
from the paper in the ingoing side of the nip and reabsorbed in the out-
going side. The volume of water removed when entering the nip would
ri I Go
w
0..
~
~ 140
....u..
\.IJ
...J
I.Lt \1o
. C!
:::J
I-
~
()
~ loo
Ii.
0
u.I
\!)
80
~
z
LU
u
Oil
w
0.. Go
0 lo '20 3o
NUMe>E:R OF .SHEETS OF PAPER.
Fig. 4. 3. Relation between moisture content and number of sheets of tissue passed
through an experimental press (after Sweet)
appear, given the same nip conditions, to be fairly dependent on the

thickness of the paper, so that the moisture content at mid-nip is always
the same, whereas the volume of water absorbed on leaving the nip
remains fairly constant irrespective of the thickness. The water reabsorbed
thus influences the ultimate moisture content of the sheet to a more promi-
nent degree when the sheet is thinner because it then represents a greater
quantity in relation to the volume of water carried by the sheet.
As the reabsorption in the outgoing side of the nip appears to be rela-
tively independent of the paper thickness it may be deduced that it is a
surface phenomenon involving transfer of water from the felt surface into
the sheet. One way of checking this hypothesis is to see what is the effect
11 309
4A.21 THE PRESS SECTION
of using two wet felts, one on either side of the paper, since this would
allow reabsorption to take place on both sides of the paper. An experiment
for this purpose is reported by Sweet and involved comparing the final
moisture content of the sheet, made up of varying layers of tissue as
before, under two conditions; in the first case the paper was run through
the press sandwiched between two wet felts and in the second case the same
two wet felts were placed together on the underside of the paper. In this
way the press nip conditions were kept the same but the paper was con-
tacted by a felt first on both sides and then on only one side. The result of
this elegant experiment confirmed that water is transferred from the felt to
the paper when leaving the nip. In every case, whatever the thickness of
the sheet of paper, the moisture content after pressing was significantly
higher when the paper was sandwiched between the felts.
Other experiments are reported by Sweet and in every case evidence of the
transfer of water was observed. A similar result was found by White and
Berdux (22) using a static press when it was observed that migration of
water from felt to paper occurred when pressure was released. With the
evidence that has been described, coupled with considerable work of his
own, Wahlstrom put forward a new explanation of pressing which will
now be described.
4A.2 MODERN THEORY OF PRESSING

As paper and felt enter the nip of a press a gradual compression between
the press rolls occurs. The paper, being more saturated with water than the
felt, can compress only to a relatively small extent. Most of the compression
which takes place up to the nip centre occurs in the felt as it compacts with
the removal of air from its structure. From the point where compression
commences up to nip centre a gradually increasing pressure is exerted by
the press rolls. First, consider the situation at a point within the nip where
the gap between the two rolls has narrowed sufficiently to exert a com-
pressive pressure on the felt and paper.
4A. 2 1 Transfer of water from paper to felt within ingoing side of press nip
At any particular point within the nip if an imaginary line is taken across
the machine parallel to the rolls then the thickness of the nip, i.e. the gap
between the rolls, must be the same all along the line and the paper and
felt together must be compressed to the same extent. The pressure applied
by the rolls must be identical at each point along this line and must be
balanced by an equal and opposite pressure in the felt and paper. A section·
through the felt and paper along this line would be subject to the same
pressure all over it, in exactly the same way that a horizontal section through
a tank of water will also have a uniform pressure acting on it. Passing
further into the nip is equivalent to going lower down in the tank though
in the case of the press rolls the rate of change of pressure with change of
position through the nip is not uniform, but is greater at the point of entry
where the distance between the rolls is narrowing most rapidly.
310
THE PRESS SECTION 4A.2 I
Returning to the imaginary section, it has been stated that the roll
pressure is balanced by an equivalent pressure within the felt and paper.
This pressure comes from two main sources: the force required to put the
felt into compression (analogous to the force needed to push down a spring)
and the hydraulic force which comes from subjecting the water in the felt
and paper to pressure. Throughout the section the sum of the compres-
sional force and the hydraulic force must be the same and be equal to the
pressure exerted by the rolls. Although the sum of the two forces must be
the same this does not, of course, mean that the individual compressional
and hydraulic forces in the felt and the paper should be the same; it is
precisely the difference between the relation of hydraulic to compressional
forces in the paper and felt that is important.
The paper enters the nip fairly close to saturation. As pressure through
the thickness of the paper increases a small amount will be used to over-
come compression resistance, but as the paper passes further into the nip
by far the majority of the pressure increase will cause the hydraulic pressure
of the water in the paper to build up. By comparison the felt is far from
saturation on entering the nip; the pressure it encounters will be used up
much more in overcoming the compressional resistance required to close
up and saturate it and only towards the centre of the nip will the hydraulic
pressure of the water in the felt begin to grow. It is evident then that shortly
after entering the nip, when the paper approaches saturation, there will be
an increasing hydraulic pressure difference between the paper and felt.
This pressure difference will, against capilliary resistance in both fabrics,
produce a flow of water from the paper to the felt. Further the magnitude
of the pressure difference must grow at least up to the point where the felt
also becomes saturated. After this point is reached it may remain fairly
steady or alter slightly depending on the small additional pressure absorbed
by further compression of both fabrics.
The precise relationship governing the difference between the hydraulic
forces in the felt and paper and the resulting water movement between
them are difficult to characterize in any detail. They must depend on many
factors beside the relative water content of the two fabrics entering the
press and the extent to which they may be compressed; in particular the
change in porosity that occurs with compression and the movement of
water itself will have an important influence. It is reasonable to deduce that,
other things being equal, the final magnitude of the hydraulic gradient
between the paper and felt in the region of the nip centre is a crucial factor
and this implies that the greater the pressure at that point the more water
will be expressed into the felt. Thus, for a given load on the presses it
should be preferable to have a short nip width in order to distribute the
load over a small area and achieve maximum pressure at nip centre. Where
the same pressure at the centre of the nip can be achieved with a wider nip
by greater loading, this would be expected to be better because the hydraulic
gradient will have operated for a longer time and therefore will have been
able to pass through a greater quantity of water. These aspects of the theory
will be discussed in more detail later when experimental evidence is
produced, but it may be noted that the traditional representation of
311
pressure in the presses by relating the total load to the length of the nip is
not sufficient to define the influence of load on the performance of the
press.
4A. 2 2 Removal of water from felt within ingoing side of press nip
So far discussion of the forces operating at the ingoing side of the nip has
been confined to those acting through the thickness of the felt and paper.
These forces lead to the movement of water from paper to felt. But a
pressure gradient also exists along the width of the nip directed from nip
centre back to the position where compression starts at the entry to the
nip. It is this gradient which is responsible for removal of the water from
both fabrics. The flow of water back out of the nip must be almost entirely
through the felt and under a given pressure gradient must be governed
entirely by the resistance encountered. The velocity of the flow relative to
the felt must be greater than the machine speed in order that water can
move backwards relative to the press rolls, so it is immediately obvious
that increasing machine speed, with all other factors constant, will lead to a
reduction in the water expressed at the nip.
With a plain press water must flow through the felt over the whole
distance from nip centre to the point of entry before it can escape and
therefore has to overcome a high total resistance to flow through the felt.
The resistance to flow in a lateral direction is known to be smaller than
through the thickness of a felt so that water expressed from the paper into
the felt should not meet a build-up of resistance on this account. But it is
clearly desirable to use a felt which, even under compression, has a high
lateral porosity along its length. In addition, other factors being equal, a
short nip width would reduce the total resistance to water flow through
the felt.
The essential advantage of a suction press at the ingoing side of the nip
is to reduce the distance water must travel through the felt before it can
escape. With increasing machine speed this advantage is extremely impor-
tant because on a plain press the felt reaches the stage of being completely
clogged with water and a pond forms in the nip. The movement of water
through the felt in a suction press is not easy to define because the pressure
gradient along the nip will be modified by the presence of the holes. Over
the holes there must be an area of lower pressure than at a corresponding
position parallel to the roll axes over a land area. Hence some movement
of water will occur across the machine. At all events it can be seen that the
distance between holes along the circumference of the suction roll is very
important and it is desirable to keep this distance as short as possible. It is
also apparent that the lower pressure over the hole areas and the flow of
water into them from surrounding land areas, particularly from the side
of the hole facing towards nip centre, will cause the felt to be wetter over
the holes. The paper will also be wetter over the hole areas at nip centre
because the pressure at that point, which has been seen to be important in
determining the water flow from paper to felt, will be lower than in the
312
surrounding land areas. These points will be reconsidered when shadow-
marking is discussed later. ·
4A. 2 3 Summary of desirable conditions within ingoing side of press nip

Before considering the second phase in the outgoing side of the nip it will
be useful to summarize the characteristics which, on the basis of the theory
propounded, appear to be desirable. Two objects have to be achieved:
(i) as much water as possible must be forced from the paper to the felt by
having as high a hydraulic gradient as possible acting for as long a time as
possible, (ii) the total resistance to flow of water through the felt to a
region where it can escape must be as low as possible. This requires:
1. High pressure at nip centre.
2. Felt dry entering nip.
3. Felt porous under pressure, particularly along its length.
4. Felt relatively incompressible.
5. Hydraulic gradient from paper to felt acting for a long time.
6a. In plain press, nip width small.
6b. In suction press, distance between holes along circumference of roll
small.
Some of these requirements are incompatible. The load that can be
applied to any press is limited and, if this limit is approached in operation,
1 can only be achieved with a shorter nip width which affects 5 adversely
(although in fact the time factor may not be so important). In a plain press
a short nip width is desirable for removing the water on another score, 6a,
so considerations in this case point to the preference for features contri-
buting to a short nip such as hard rubber-covered rolls which are small in
diameter compatible with the requirements of structural rigidity. For a
suction press the same features are also desirable though not to quite the
same extent. It is difficult to make a felt which satisfies both 3 and 4 at the
same time. What seems to be ideal is an open felt which requires a sub-
stantial force to compress it. This would contribute to keeping a short nip
width and, by absorbing the roll pressure in compression, would allow the
hydraulic gradients to build up while at the same time offering only a
small resistance to movement of water in the felt. It may also be noted that
in a second press where the paper is more dry and compressible each of
the requirements detailed above will be relatively more important, especially
2 and 4.
4A. 2 4 Conditions within outgoing side of press nip; shadow-marking

Passing through the centre of the nip the pressure on the felt and paper is
gradually relieved. In this phase of the nip, water is drawn back into the
paper from the felt largely by capillary attraction. Tests on a variety of felts
and papers indicate that the average capillary size of the paper is always
smaller than that of the felt so that there is ample evidence that a force would
operate across the felt/paper surface drawing water out of the felt. At the
same time, as the fabrics expand air/water intersurfaces will be formed,
313
4A.2 5 THE PRESS SECTION
mainly in the felt, and water will redistribute into the smaller capillaries.
This will also draw a certain amount of the water adhering to the surface
of the felt over the holes of a suction press back into the felt.
In a suction press the amount of lateral as opposed to vertical redistri-
bution of water which takes place within the felt and paper is difficult to
determine. One school of thought inclines to the view that very little
redistribution occurs; the paper and felt becomes wetter over the hole
areas largely because the pressure in the nip is lower in those regions and
also, in the case of the felt, because of the flow of water through the felt
to the holes. A second school believes that the felt, and in turn the paper,
becomes wetter over the hole areas mainly because of the water reabsorbed
from the holes in the outgoing side of the nip; redistribution of water from
hole to land areas in the body of the felt prevents the difference in moisture
from becoming excessive.
Questions concerning the extent of lateral redistribution of water from
hole to land areas and the amount of reabsorption of water from holes
in the outgoing side of the nip are important principally in connection with
shadow-marking. It has been thought for some time that the basic cause
of shadow-marking is attributable to differences in moisture content of the
web over land and hole areas as it leaves the roll. But the reason this
difference persists through to the finished paper has not been clear.
Work reported by Redfern and Gavelin (32) has now indicated that
shadow-marking is due basically to the difference in pressure experienced
by the paper over the hole and land areas of the shell. They detected a
significant difference between the structure and density of the sheet in the
two areas and demonstrated the existence of a slight plastic flow of fibre
from the land to the hole areas. It appears that a difference in the degree of
fibre bonding between the two regions of the sheet occurs due to differ-
ences in compression in the nip and it is this, rather than the difference in
moisture content of the sheet over the hole and land areas, which is
considered responsible for producing shadow marks in the finished paper.
The hole areas in fact appear lighter than the land areas due to the differ-
ence in light reflection caused by the lower number of bonds.
Shadow-marking often becomes troublesome mainly when a felt is
drawing near to the end of its life, and it is precisely at this time, when the
felt has become hard and plugged, that the difference in pressure over the
hole and land areas will be accentuated. Relieving the load on the press or
using softer rubber-covered rolls will help to reduce the difference in
pressure but both must be considered retrograde steps since they will
reduce press efficiency. Redfern and Gavelin consider that the ·only
feasible measure against shadow-marking on an existing press, which need
not at the same time reduce efficiency, is to use a different type of press
felt that is stiff enough on the underside not to be pressed into the holes.
4A. 2 5 Reducing rewetting of paper in the outgoing side of press nip

It will be appreciated that while extraction of water in the ingoing side of
the nip is largely independent of paper thickness, reabsorption in the
314
THE PRESS SECTION 4A.2 5
outgoing side of the nip, being a surface phenomena, will affect both
thinner and drier paper to a greater extent. If the quantity of water flowing
over the felt/paper surface is relatively constant it will raise the moisture
content to a proportionally greater extent when there is less water already
in the paper. Thus, prevention of rewetting will be of greater importance
in a second press. The evidence for this transfer of water already quoted
in the work of Osborn and Wicker and of Sweet was confirmed by
Wahlstrom when he observed that the water carried round by the press felt
increased with paper no longer passing through the nip. In addition he
observed that without paper passing through the nip the top roll was
wetter and this supports the view that capillary attraction is less in the
felt than the paper.
The transfer of water from felt to paper in the outgoing side of the nip
will be kept to a minimum by having:
1. Short nip width to reduce rewetting time.
2. Quick separation of paper and felt.
3. Small felt capillaries, particularly at surface contacting paper.
The third requirement is partially contradictory to the desirability of
having an open felt which has been shown to be preferable for the ingoing
nip phase. This emphasizes the difficulty of obtaining an ideal felt and
the necessity for compromise in felt properties to produce the most
effective type for assisting water removal.
These considerations have led to the idea that a more efficient press
arrangement would come from separating the two functions of absorbing
water from the paper and then removing it from the felt. For the first
function a felt can then be used which is thick and heavy enough to absorb
all the water expressed in a plain press nip without attempting to pass any
water out of the nip for removal. The felt, carrying all the expressed water
with it, then enters a separate high-load press or some other method
designed to remove most of the water it contains. The means adopted for
water removal from the felt can be much more drastic than is possible
where paper is present, and the felt should enter the main press nip
relatively dry.
The effectiveness of such an arrangement must depend on the degree
of rewetting which occurs on the outgoing side of the nip compared to the
ability of the felt to absorb all the water removed from the paper. That
rewetting has a fairly minor influence in comparison with the ability to
remove water from the web has been suggested by the work of Robinson
(43) and Swanberg (45) who have both demonstrated that the use of two
felts, with the web sandwiched in the middle in a 'double-divided' press,
gives superior operation.
Robinson, on a machine making an extremely wet-beaten transparent
paper and equipped with five plain presses, made a modification in such a
way that one of the presses could be missed and used instead for dewatering
a second felt added to an adjacent press. Two different arrangements like
this were tried and in each case there was an appreciable improvement in
the overall efficiency of the press section. Swanberg adapted a conventional
315
suction press and bottom felt arrangement to incorporate a top felt and
wringer rolls; in a later experiment a fabric felt was also added between the
top felt and top roll. Under practically all operating conditions {speed
above 300 f.p.m., substance above 50 g.s.m., freeness above 20° S.R.) the
first of these arrangements gave improved water removal at all linear nip
pressures compared to the conventional set-up, while addition of the
fabric felt improved removal even further. It would appear from this work
that under conditions where the web is dewatered from both surfaces the
resultant increase in pressing time and lower resistance to water flow from
the web can be more important than the maximum hydraulic pressure
achieved in the centre of the nip (since this will be lower for two felts
compared to a single one). Secondary advantages of the double-felt
arrangement are the absence of any adhesion to the top roll, which can be
particularly troublesome with some grades, and reduction of shadow-
marking for the same load on the press.
4A. 2 6 Recent developments

Coincident with and as a consequence of the findings of Wahlstrom there
has been a considerable development of press section design culminating
in recent years in several new arrangements appearing on the market.
Each has been developed in an endeavour to overcome the shortcomings
of the normal plain press design without incurring the high cost of suction
press rolls and pumps. They have generally been installed in second press
positions where there is more scope to improve efficiency, and represent in
different ways attempts to reconcile the need to attain higher nip pressures
(in order to express more water from the web into the felt) with the ability
to allow the water expressed to escape from the felt without building up a
high hydraulic pressure leading to crushing. Not all have yet proved fully
operational, though several arrangements are now used in sufficient
numbers to warrant mention.
The first of these is the fabric press which incorporates a specially
designed, incompressible, fine open mesh, plastic fabric running between
the felt and the press roll (33, 34, 35, 46). Water absorbed by the felt is
pressed into the fabric and subsequently thrown out as the fabric separates
from the roll, or is sucked out at a suction box. The fabric requires its own
guide and stretch rolls, and a camber bar or expander roll to keep it taut
across the width of the machine. An adaptation of this is the sleeve press
in which a fabric equal in length to the circumference of the roll is shrunk
onto the roll (48, 49). Reports on these two arrangements have indicated
improvements in dryness leaving the press compared to both plain and
suction presses of conventional design. Each can be used in conjunction
with an existing suction press and this gives an even greater improvement
in overall efficiency and can be useful to overcome a severe shadow-
marking problem. There are also indications that the use of a fabric press
with high load can improve strength properties at the expense of a reduction
in bulk and air resistance {54).
316
A second arrangement, known as the Venta-nip press, utilizes a 10 to
30 thou wide spiral grooving in a plain bottom roll. This assists removal
of water from the felt and allows greater nip pressures than could be used
with a plain roll; water is removed from the grooves by a type of foil
followed by a doctor (a low vacuum may be applied between the two for
speeds under about 700 feet per minute), assisted at higher speeds by
centrifugal force (when only a doctor is necessary) (39). Reports on this
type of press are fewer but several are known to be working successfully,
giving improved dryness even compared to a suction press. Problems in
their use appear to be mainly concerned with obtaining a satisfactory
design and life for the clothing and also, in some applications, with
keeping the grooves clean.
A third innovation is the Hi-I press which involves the use of a small
diameter, stainless steel, grooved roll which runs in the nip between the
two principal press rolls in such a way that it lies between the felt and a roll.
It is claimed (56) that nip pressures three times greater than normal are
possible with this arrangement and this gives an exceptionally high improve-
ment in press efficiency. Several installations are now in operation. Other
modifications in the process of development include use of an impervious
belt running from the wire into the first press nip, and of a special type of
porous nylon material covering a perforated steel press roll which allows
water expressed from the web (possibly without any felt in the nip at all) to
pass into the body of the roll. Preliminary laboratory tests have confirmed
the advantage of using this latter material (42), but there have been no
reports of a full-scale operating application.
317
CHAPTER 48
OPERATING FACTORS AFFECTING PRESS
PERFORMANCE
48. 1 PRESS LOAD AND NIP WIDTH
The influence of nip pressure on water removal and on the moisture content
of the felt has already been mentioned in 4A. 2 1. Experimental work
shows that the felt has less moisture in its body the higher the nip pressure,
and it was seen that this is one of the pieces of evidence favouring the new
press theory. The general relation between water removal and nip pressure
obtained by Osborn and Wicker and shown in Fig. 4. 1 has been confirmed
under different conditions by several other workers (2, 10, 21) and is
familiar to all papermakers. Alteration of the press load is probably the
most important variable on a press and all the experimental curves show
that the moisture content of paper leaving a press is closely dependent on
this. There is evidence that greater nip pressure affects bulk and strength
properties, see 41. A greater nip pressure also appears to reduce the life of
a felt (17) and will increase the power required to drive the press.
4B. 1 1 Importance of nip width

The pressure in a press nip is customarily quoted in lb. per linear inch
because of the difficulty of obtaining an accurate measurement of the
width of the nip especially under operating conditions. Even if the force
exerted were related to the area of the nip to obtain a more conventional
pressure figure it is evident that the result obtained would give only an
average value since the pressure is not evenly distributed over the nip area
but is, of course, much greater at the centre of the nip. Nevertheless
Wahlstrom has found that the figure for pressure obtained by relating the
force applied to the area of the nip (and termed by him 'specific nip
pressure') bears a closer relation to water removal than pressure measured
at lb. per linear inch. This is in accord with the theory since the specific nip
pressure, by taking into account the width of the nip, gives a more accurate
measure of the magnitude of the pressure at the centre of the nip and this
maximum pressure is the most important factor governing the expression
of water from the paper to the felt.
As water removal is dependent on the specific nip pressure it is evident
that for a given load on the press the nip width is very important and the
narrower the nip the more efficient will be the water removal. The upper
limit for water removal on any press is when crushing occurs and shear
forces in the plane of the sheet create disruption by stretching which leads
eventually to disintegration of the web as it folds over in small creases in
the nip (40, 47). If a press is run close to maximum load for the bearings
318
THE PRESS SECTION 4B. l 3
and there is no trouble with crushing or felt rubs then it would be desirable
to attempt to narrow the nip. This could be done most easily by using a
harder and more incompressible felt but it will be realized that large press
rolls and softer rubber covers both contribute to producing a wide nip.
When a press is run with plenty of load to spare then, up to the point of
crushing, a wide nip width does not matter to the same extent, especially
on a suction press.
With increasing applied load, the nip width increases slower in propor-
tion than the load so that the specific nip pressure also increases and
improved water removal takes place. A wide nip will be more noticeably
disadvantageous when the paper is run between two felts as on a tissue
machine or a transfer press; in this case not only will the double cushioning
of the felts widen the nip but rewetting of the paper within the outgoing
side of the nip will take place for a longer time in a wider nip and will occur
over both surfaces of the paper. In such presses the necessity to effect
transfer limits the type of felts that can be used and it becomes particularly
important to be able to apply a heavy load and use relatively hard rolls if
the press is to function efficiently.
48. 1 2 Experimental evidence of the influence of nip width

Wahlstrom obtained figures confirming the influence of nip width on
water removal by assessing the effect of felt compressibility and also of the
hardness of rubber on a bottom roll. In both cases with the same general
press conditions the narrower nip gave greater water removal and resulted
in lower steam consumption. As a felt ages and becomes less elastic and
compressible the nip width decreases and this was also shown to be
advantageous; this will be discussed more fully in 4B. 2.
The moisture content of paper entering the first press is considerably
greater than for the second press. This implies that in the second press a
greater pressure is required to compress the paper while the build-up of the
hydraulic pressure providing the force for water removal will be slower.
For this reason greater specific nip pressure can be attained at the second
press without crushing although the effect of increasing load is not so great;
Wahlstrom quotes a reduction of 0·04 in the water/fibre ratio through the
first press as against 0·025 through the second press for an increase of
10 p.s.i. specific nip pressure on 100 p.s.i. Narrow nips are therefore of
greater importance in a second press and in this respect the type of felt
used is of great significance.
48. 1 3 The need to measure total press load

On most presses it is very difficult to assess nip width with any accuracy
when the press is in operation. Carbon impression and other techniques
such as the use of specially embossed aluminium foil permit static measure-
ments of the nip width with and without felts but there is considerable
doubt about the relation of such results to dynamic conditions. Methods
involving the use of strain gauges and piezo-electric crystals have been
described but are not really practicable for normal production presses.
319
4B.21 THE PRESS SECTION
The machine crews must rely for day-to-day operation of the press
mainly on knowing the total load applied to each side of the press, yet it is
surprising how many machines still have no direct measurement of this
important value. Over a period the force applied by weights and levers in
the older type of press can alter appreciably due to large friction losses
which can build up in the system at the contact points between tension rods
and lever arms and at the pivots; having the weights at the same position
on each side of the press need not mean an even pressure application and
machinemen often do not appreciate the difference in load produced by
quite small movements of the weights along the lever arm. Likewise where
springs are used their characteristics change with ageing. No press should
remain unequipped with some form of load indication and on older presses
devices like the Statimeter or load cells using strain gauges are extremely
valuable.
Presses using air or hydraulic loading invariably have a fluid pressure
gauge which is perfectly satisfactory-it is less important to have an
absolute load measurement than a means of measuring the variation in
load during normal operation. Whatever load measuring device is used
should of course be checked and calibrated against a dead-weight mechan-
ism at regular intervals by the instrument p€rsonnel. The relationship
between the instrument reading and the absolute load applied should be
known in order that the theoretical camber of the press rolls can be
determined for normal operating conditions.
4B.2 FELTS AND FELT CLEANING

The function of the felt in a press, apart from conveying the paper and
substituting a less objectionable felt mark for a wire mark, is to act as a
conducting medium for removing water. It used to be considered that the
best type of felt was one which allowed considerable compression and was
sufficiently elastic to have immediate recovery; the expansion of the felt
taking place in the outgoing side of the nip was believed to draw water out
of the paper so that the greater the expansion, the better the water removal.
The more recent press theory predicts that the best type of felt, for the first
press at least, will have little compression in the nip and will be sufficiently
open to retain good porosity under pressure.
These requirements have been generally verified by measuring the
relative press water removal for a variety of different types of felt of a
similar weight (21, 25, 30). The openness of a felt is largely determined by
the weave pattern and in two separate investigations it was found that the
more open reverse broken twill or chain weave was more effective in water
removal than the closer plain weave. Felts which were less compressible
and rather denser also gave improved water removal.
4B. 2 1 First and second press felts

In the first press the paper becomes saturated after a very small amount of
compression (one estimation gives the figure of 10 per cent.) while in the
second press the paper will not be saturated until compression has reduced
320
THE PRESS SECTION 4B.22
its thickness by about half. Also, in a second press the quantity of water
extracted is normally between 10 per cent. and 40 per cent. that of the first
press. The requirements of a second press felt are therefore somewhat
different and it would be expected that openness of the felt could be
sacrificed to achieve lower compressibility. It has in fact been confirmed
that denser felts with lower compressibility give better water removal in a
second press than the more open felts suitable for a first press. The finer
capillaries of a denser felt would also be expected to reduce the rewetting
of paper in the outgoing side of the nip and this is of much greater impor-
tance than in the first press.
The felt normally carries round a quantity of water of the order of ten
times that brought to the press in the paper. The hydraulic conditions in
the nips are thus governed largely by the moisture content of the felt
entering the nip and, especially in the second press, it is very important to
keep the felt as dry as possible for efficient water removal. From this point
of view it is essential when a felt is leaving the nip of a suction press to
prevent any rewetting by water thrown out of the holes. Any form of
cleaning device on the felt should preferably extract more water than it
puts in and soft doctors on the bottom press roll preventing water being
carried round to the nip have also been found effective (21). A form of
suction box on the felt would help to reduce the water content where this is
excessive, but the effectiveness of this must be balanced against the
possibility of increased wear of the felt.
4B . 2 2 Change in felt characteristics with age

When a new felt is put on a press as much load as possible is usually applied
in an attempt to extract sufficient water. As the felt gets run in and becomes
mechanically conditioned the load is relieved, but the water extracted will
still generally be greater than initially. The process of running a felt in
brings about a compaction of the felt fibres making the felt thinner and less
compressible and this in itself, since it is known to improve water removal,
is a good confirmation of the validity of Wahlstrom's theory of pressing.
Also during the preliminary wetting and running-in periods there can be
considerable dimensional changes in the length and width of the felt and
it is one of the concerns of a felt manufacturer to assess and minimize
these changes in advance and allow for them when supplying the felt.
The changes in characteristics and performance of a felt after being put
on has been the subject of a certain amount of experimental work;
Wahlstrom in his extensive paper dealt most thoroughly with this aspect
and it is from his work that subsequent remarks are largely drawn. He
found that throughout the life of a felt the thickness decreases and the
density increases, especially in the earlier stages. At the same time the
compressibility of the felt decreases and it carries round a smaller volume
of water. Reduction of the compressibility corresponded to reduction of
the width of the nip and to this is largely attributed the increased water
removal efficiency. Fig. 4 .4 illustrates some typical results on a second
press felt. The increase in moisture content of the paper for a short period
321
48.2 3 THE PRESS SECTION
immediately after start-up is attributed to the loss of water repellent

material from the surface of the wool, making the felt take up more water.
The thickness of first press felts reduced and levelled off to a greater
extent than second press felts and this may be due partly to differences in
the type of felt and partly to differences in the quantity of water removed
and carried round by the felt. The running-in period, when thickness is
changing relatively rapidly, can vary from a few hours on a high-speed
machine using high nip pressures to several days for lower speeds and
pressures. The weave of the felt is also important and it has been found that
0 f> to lS 2o 2S
AGS OF- F'EL.T, O~Ys.
Fig. 4 .4. Variation of felt thickness, nip width and moisture content of paper with age
of second press felt, (after Wahlstrom)
the more open reverse broken twill felt reached equilibrium faster than a
plain weave. With slower machines, when the running-in period is longer,
the difference in time required to reach equilibrium is of much greater
significance and the reduced water removal efficiency during this time
becomes an important consideration.
4B . 2 3 Effect of plugging in a felt

As a felt ages a certain amount of material is worn away and the dry
weight of the felt is reduced. This leads to a gradual deterioration in
performance and may be the determining factor governing removal of the
felt from the machine. On probably a majority of machines, however, a
felt is eventually taken off because it has become so hard and plugged with
fibres, pitch, loading and other deposits that press efficiency has become
seriously reduced; this applies particularly to first press positions. The
extent of this problem varies considerably from one machine to the next.
322
In one investigation Peters (8) reported that he found no particular change
in water removal efficiency or steam consumption during the life of felts
used on a kraft machine-material filling the felt after it was removed
amounted to only 1 per cent. to 3 per cent. of the total weight of the felt.
By contrast on another machine making newsprint the amount of filling
in the felt when it was taken off amounted to between 11 per cent. and
35 per cent. of the felt weight and in this case over the life of the felt a
reduction in water removal and increase in steam consumption was
observed.
Although the detrimental effect of felt plugging has long been recognized
from practical experience, until recently very little headway has been made
in assessing the effect quantitatively because of the difficulty of obtaining
some measure of the result of plugging on felt properties while the felt is
in use on the machine. Recently some success in this direction has been
achieved by the use of a simple device known as the Huyck-Smith porosi-
meter (41). This instrument consists of a perforated disc covering a small
air chamber which is held against the felt while it is running; the air
chamber is kept at a constant regulated vacuum which falls as air leaks
through the disc and thereby gives a measure of the porosity of the felt.
The device can be effectively used for assessing differences in plugging
across the width of a felt as the reading should not be unduly affected
during a single traverse of the felt by calibration difficulties, variations in
the pressure applied on the felt and other changing conditions.
Howe (25, 30), Wicker (36) and Delisle et al. (37) have also used the
instrument to assess the variation of felt porosity with age and report some
interesting results from this work. The measurements obtained by Howe
show a steadily decreasing porosity over the whole life of a first press felt,
which may be attributed at least in part to a growing quantity of material
plugging the felt, while the water removal efficiency of the press gave a
corresponding reduction during the same period. First press felts showed a
much more prominent reduction in porosity and water removal efficiency
with age than second press felts. This may be attributed to the higher flow
rates in the first press and the greater removal of fines from the paper into
the felts. Also it was observed that the more open the first press felt the
easier it could be kept clear of filling materials; open reverse-broken twill
weave felts filled to a lesser extent than denser plain weave felts. The
choice of an open felt for the first press position on the grounds of obtain-
ing high water removal efficiency will thus also help in minimizing the
problem of plugging.
4B . 2 4 Cleaning of felts
Recognition of the high cost in excess steam consumption due to a plugged
wet felt reducing press efficiency has brought about recently a renewed
interest in the question and economics of felt cleaning. Many machines do
not possess any cleaning device at all on the wet felt and in this case it may
be the practice to remove the felt periodically for cleaning off the machine,
using a second felt while this is being done. On some machines the wet felts
323
may be washed and occasionally turned over at scheduled shut periods,

with the felt still on the machine, or this may even be done specially in
mid-week and involve stopping the machine. The relative merits of any
of these methods require careful economic consideration and a time-
hallowed system in use on any machine may well be completely un-
satisfactory under present-day conditions. This subject will be discussed in
4C. 2 1 and 4C. 2 2 when attention is given to methods of supervising
wet felts on a machine.
On faster machines allowing a felt to become plugged can result in a
serious increase in production costs and additionally can create a great
deal of trouble due to difficulties in producing a level reel. For these
reasons some form of felt cleaning is in general use on faster machines.
The Vickery felt conditioner has sustained a great deal of popularity and
appears to be reasonably efficient even though it cleans a restricted area at
any one time and also, with the exception of more recent models, gives an
uneven time distribution between consecutive cleaning applications due to
the design of the forward and backward traversing system. With this
cleaner it is important to ensure that about the same quantity of water is
removed through the suction orifice as is forced into the felt at the pressure
orifice, otherwise a wet streak in the felt and paper will be observed
opposite the cleaner shoe and eventually crushing will occur; this applies
particularly if the felt is already well plugged.
On many machines devices cleaning the whole width of the felt are used,
sometimes in conjunction with squeeze or wringer rolls; the simplest of
these involves a shower across the machine followed by a full-width suction
box. The width of the suction box and the vacuum applied appears to vary
considerably from one mill to another, but should in practice be arranged
so that as much water is extracted as is put on by the shower. An alternative
to the use of a suction box, which may curtail life of a felt, is the Rolvac
unit which applies vacuum to the volume formed by three rolls, two of
which run in contact with the felt (the first acting as one of the squeeze
rolls), the third closing the gap between the other rolls.
There have been several reports recently of the effect of using small
quantities of detergent with the cleaning water, and the value of this in
improving press water removal, particularly for badly plugged felts, has
been amply demonstrated (8, 30, 37). The use of hot rather than cold
water also keeps felts noticeably cleaner. Intermittent applications of a low
concentration of detergent certainly help to prevent excessive build-up of
filling material, but this procedure has been shown to have only a tem-
porary effect and not to be as valuable as continuous washing with
detergent (37).
To illustrate the importance of felt cleaning, the effect of starting a
conditioner after leaving it off for over an hour has been examined by
measuring the combined moisture and weight of a felt continuously with a
beta-ray gauge (7). The weight of the moisture and filling material reduced
from about 61 ·5 per cent. of the felt weight to 59 per cent. over a period
of less than an hour, a change which would be expected to improve water
removal.
324
THE PRESS SECTION 4B.3 1
When considering any new method of felt cleaning it is important to
bear in mind not only the improvement in felt life and general running
which may be obtained but also the effect on the water removal efficiency
of the press. It is no use gaining an extra week's life for a felt if as a result
the paper is running appreciably wetter to the drying section all the time.
It is unfortunately very difficult to assess long-term changes in moisture
content of the paper leaving the press so that a complete economic assess-
ment is very often impossible and a decision on the merits of a change of
this nature becomes more of an inspired guess.
48.3 FELT AND PAPER RUNS

Felt runs are designed in the first place with the necessity of changing the
felt as easily as possible very much in mind. Even then it is surprising how
complicated some runs become and how the demands of the machine frame
introduce elaborate twists and turns. For every pass round the press the
amount of wear on a felt will be greater if there are more felt rolls and also
the felt will be more difficult to keep square. Except on the stretch roll it is
undesirable to have too severe a turn over a felt roll; although a certain
degree of flexing will help to keep the felt open, an excessive amount will
subject the felt to undue strain. One very important facility which is needed
on a press is to be able to alter within reasonably wide limits the angles
with which the felt approaches and leaves the press. This is particularly
necessary with suction and transfer presses where the relative angles at
which the felts and paper leave the nip can be quite critical.
48. 3 1 Conditions for a plain press

On any plain press there are two important considerations which should
govern the directions of the felt and paper runs in the vicinity of the press
rolls. Entering the nip it is important that bubbles of air are not trapped
between the paper and felt because these will cause the paper to stretch
and produce creases on either side of the bubble. Leaving the nip it is
desirable to separate the felt and paper immediately to reduce the re-
absorption of water from the felt into the paper.
The felt and paper will be carried together into the first press and should
lap the top roll to ease water removal down the bottom roll. It is usual to
couch the top roll backwards to aid the water removal. It may be necessary
to have a suction box under the felt to prevent blowing, though passing the
paper and felt fairly tightly over a roll at the point where they meet is
usually sufficient. A suction box can be water lubricated if its operation
appears to cause undue wear of the felt.
The felt should leave the first press nip lapping the bottom roll and the
paper will be peeled off the top press roll. On many machines the paper is
allowed to drop back off the top press roll onto the felt before transferring
over to the second press felt. This should be avoided because there may be
some transfer of water from felt to paper, particularly with a heavier sheet,
and also creasing and rubbing can occur because the draw is not applied
directly to the sheet as it leaves the first press.
325
4B.3 2 THE PRESS SECTION
The paper should be taken off the first press top roll over a lead roll
(which is occasionally driven at a slower speed to reduce web slackness
when the draw is long) then enter the second press lapping the top
roll. The felt should also lap the top roll at a point a little nearer the
nip. By pressing the paper and felt onto the top roll a certain amount of
water may be transferred to the felt before entering the nip. This arrange-
ment should prevent blowing and thereby remove the need for a suction
box as in the first press. When it is necessary to support the sheet as much
as possible between the presses the paper should be transferred straight
FELT.
Fig. 4. 5. Preferable arrangement of runs for paper and felts through two straight-
through plain presses
Fig. 4. 6. Runs for paper and felts through two straight-through plain presses when
sheet requires maximum support between presses
from the press to the second press felt; a suction box may then be needed
and then similar considerations apply as for the first press. For a reverse
press the felt will carry the paper into the nip and in this case also condi-
tions are similar to a first press. Leaving the second press nip the same
considerations apply as in the first press nip. These points are all illustrated
in Figs. 4.5 and 4.6.
4B . 3 2 Conditions for a normal suction press

When considering the runs for a normal straight-through suction press
there are a number of different requirements to take into account; in
326
particular, the position of the suction box in the bottom roll is of great
importance and must be considered at the same time. The leading edge of
the suction box must be sealed and it would be impracticable and inefficient
to ensure it is positioned actually within the width of the nip. Hence the
felt and paper must both lap the bottom roll sufficiently to seal the suction
box and the effect of the vacuum then serves effectively to remove air
entrained between the web and felt.
In the air-bleed principle an extra wide suction box is used to allow air
to be drawn through the felt before the paper contacts the bottom roll and
this introduces different considerations. However, discussion of the merits
of air-bleed will be left till 4B .4 3.
On leaving the nip it is equally important, as with a plain press, that the
felt and paper should be separated immediately, the paper passing up the
top roll and the felt lapping the bottom roll. This usually necessitates the
suction box terminating at the point where the nip finishes otherwise the
suction will under certain conditions tend to hold the paper to the felt
instead of allowing it to pass up the top roll. To facilitate this the suction
box position must be movable.
Further, the precise angle which the felt laps the bottom roll is very
important in relation to the position of the trailing edge of the suction box
because this determines the manner in which water in the holes of the
suction press is removed. Dixon (1) first drew attention to this as a result
of observations he made on a stack press and his 'pop' theory will now
be described.
4B. 3 3 Movement of water in suction press holes
Leaving the nip the press holes can be thought to contain a plug of
water drawn from the felt by suction and to be under vacuum sealed at
their outer end by the felt. The felt must not leave the roll before the
leading edge of the sealing strip is reached otherwise there will be a direct
leakage of air to the suction box. This is easily avoided so long as the
leading edge is just within the press nip.
As the roll rotates, immediately the leading edge of the sealing strip is
reached the holes in the shell are effectively sealed at both ends and still
under vacuum. If now the trailing edge of the sealing strip is reached before
the felt leaves the roll, air will rush in from the inside through the un-
covered end of the holes and the plug of water will be driven out into the
felt. This is obviously undesirable since the felt will then be rewetted. Even
if the felt is removed from the roll just after the trailing edge of the sealing
strip has been passed and the air pressure is equalized on both sides of the
hole, the kinetic energy of the water already produced by the air pressure
on the inside of the hole will still carry the water outwards, aided by
centrifugal force, and into the felt.
On the other hand, suppose that the felt is removed from the roll while
the inside of the hole is still sealed by the sealing strip. Air will now rush
into the uncovered outside of the hole forcing the water inwards. When the
inside of the hole is unsealed the air in the hole will already be effectively at
atmospheric pressure and there will not be any substantial force acting to
327
drive the water outwards. The water will in fact leave the hole later by
being thrown out by centrifugal force.
There is one remaining possibility and it is this one which was observed
by Dixon and led him to consider these points in detail. If the trailing edge
of the sealing strip is reached very shortly after the felt leaves the roll, the
air rushing in through the uncovered outer end of the hole and pushing the
water inwards will develop sufficient kinetic energy to force it to the inside
of the hole just as it is uncovered. This will result in a spray directed
inwards into the suction roll which is normally undesirable because it is
not possible to place a tray actually inside the roll for the purpose of
catching the water. This phenomenon was observed on the middle roll
of a stack press where it was discovered that in a certain setting water was
actually thrown upwards from the lower nip into the middle roll, a rather
surprising effect.
4B. 3 4 Position of felt take-off and suction box seal
It has since been proved fairly conclusively that water does not collect
in the holes in the form of a plug but is in fact atomized; if the quantity of
water expressed into each hole is calculated it is easily shown that it would
occupy too small a volume of the hole to act as a seal across the hole. This
does not, however, affect the substance of Dixon's theory since the flow
of air into the holes under a pressure difference will have the same effect
on water droplets as does pressure on a plug of water sealing the hole.
From the theory it is evidently desirable that the felt should leave the
bottom roll at an angle slightly further round the roll than the leading edge
of the sealing strip. The holes are then unsealed first on the outside where
the felt leaves the roll, and air enters to carry the water inwards; by this
means the eventual throw-out of water by centrifugal force is delayed as
long as possible making collection of the water in a tray easier. As the
sealing strip leading edge should be just within the nip, the felt will leave
the nip almost directly outwards, i.e. at right angles to the plane joining the
roll axes. The lead roll directing the felt off the bottom roll should be
adjustable in position so that the desired angle can be set; though not too
critical in most presses, the angle is obviously dependent on machine speed
and, to a lesser extent, on the vacuum in the suction box.
The sealing strip should preferably be wide thereby allowing the air to
reach near atmospheric pressure in the hole before it is uncovered on the
inside. On some suction presses, in particular in a stack press where it
may be impossible to place a tray in a suitable position, the normal suction
box can be followed by a narrower box under a small vacuum designed to
draw air at high velocity from the atmosphere and collect any water from
the holes before it is thrown out. On most presses however the tray is
positioned as close to the nip on the underside of the felt as possible to
prevent rewetting of the felt from water throw-out.
48. 3 5 Ideal arrangement of suction press felt and paper runs

The normal suction press is couched forwards to allow the felt to carry the
paper into the press almost horizontally and to seal the suction box in the
328
bottom roll at the same time. Figure 4. 7 embodies the points raised above
and illustrates the preferred arrangement for two straight-through suction
presses. The position of the leading edge of the suction box trailing sealing
strip is just within the nip and it will also be noted that the felt lead rolls
after the press nip are adjustable to help in setting the correct angle at
which the felt leaves the bottom roll.
On most suction presses the ideal arrangement analyzed above will be a
counsel of perfection. In practice the leading edge of the trailing sealing
Fig. 4. 7. Preferable arrangement of runs for paper and felts through two straight-
through suction presses; the positions and width of sealing strips for the suction boxes
are exaggerated
strip may have to be set well into the press nip otherwise variations in the
width of the nip may lead to difficulties in preventing the sheet from
following the felt. Especially as the felt gets dirtier and closer the paper
tends to stick to the felt at the edges if the suction extends beyond the nip.
It is doubtful if this will have any substantial effect on press efficiency
because the nature of the capillary transfer of water from felt to paper in
the outgoing side of the nip should not be affected to any great extent by
continuing to subject the felt to suction. It is, though, obviously important
that the suction should cover the full extent of the ingoing side of the nip
since it is in this region that the water is expressed.
Likewise if the felt laps the bottom roll further round on the trailing side,
provided it does not seal the roll when the suction box sealing strip termi-
nates severe rewetting of the felt will be avoided, especially if the press tray
extends as far into the nip as possible. It would appear, however, that in
many cases an extension in the width of normal suction box sealing strip
on the trailing side could be advantageous.
4B. 3 6 Conditions for a press nip where paper transfer occurs

With increasing machine speeds and the desire for more compact press
sections there have been in recent years a growing variety of designs
embodying suction pick-up and transfer rolls, the main purpose of these
being to eliminate open draws between the presses. The stack press early
took this trend to its logical conclusion but it has been found to be rather
329
too compact and to possess several difficulties from the point of view of
clothing replacement and general maintenance.
It is not proposed to discuss the merits of different press designs in any
detail except with regard to the common practice of transferring from one
felt to another in a nip formed by two suction rolls. Elimination of the
open draw together with automatic feeding up and convenience in press
arrangement are the main advantages to this arrangement although
without considerably increasing the load press efficiency is reduced by the
presence of two felts in the nip. Also the requirement of transferring the
paper from one felt to the other tends to control the types of felt used in the
different positions in preference to selecting the felts primarily for their
water removing properties.
Consider the press shown in Fig. 4. 8 in which the paper is transferred
from the bottom carrying felt to the transfer felt on the top roll. The bottom
Fig. 4. 8. Illustration of the preferable arrangement of paper and felt runs and position
of suction boxes at a transfer press
roll would often also be the pick-up roll with a second suction box at the
bottom where the roll contacts the wire. Entering the nip of the press the
paper and carrying felt meet with similar conditions to those in an ordinary
suction press and must be in contact with the bottom roll before the
leading sealing-strip of the suction box. The transfer felt should also lap
the bottom roll and cover the suction box to assist in sealing the box; further,
the possibility of blowing from air trapped between the paper and the
transfer felt will then be reduced because the pressure of the felt on the
paper will help the air to escape upwards and out through the transfer felt.
The position and size of the bottom roll trailing suction box seal are
governed essentially by the same factors that have been discussed for the
case of an ordinary suction press. However, the necessity for transfer
makes it particularly important that the suction area terminates well
within the nip. Also in this case the bottom felt should just lap the top roll
so that when it separates from the paper the underside is at atmospheric
pressure and there is no possibility of a suction being applied to the paper
which would tend to make it stick to the bottom felt.
330
The suction box in the top roll is primarily responsible for ensuring that
the sheet follows the top felt and commences at about the centre of the nip.
The suction area could extend into the ingoing side of the nip in which
case some water from the paper would be drawn into the transfer felt;
although this might reduce the moisture content of the paper slightly at
mid-nip, this method of operation is not likely to yield an improvement
because the transfer felt should be as dry as possible to obtain the best
results in the next press nip.
The transfer felt and the paper must lap the top roll beyond the suction
box area and, especially in the cramped situation of a stack press, they
could possibly be arranged to leave the top roll at a point in relation to the
sealing strip which encourages water in the holes to shoot inwards into the
top roll. In this way, according to Dixon (1), the water can be collected
easier than by trying to fit a tray in a very awkward position. The condition
best suiting this method of operation (see 4B. 3 3) will be achieved when
the vacuum in the holes of the top roll is broken first on the outside by the
felt leaving the top roll and immediately afterwards on the inside as the
trailing side of the sealing strip is passed. The precise position where this
takes place can only be found by experiment and for this it is generally
necessary to be able to observe what is happening in a roll with an open-
ended construction. The lead roll over which the transfer felt and paper
passes must be adjustable within reasonable limits to effect the desired felt
direction.
In practice, the use of a subsidiary air chamber using low vacuum, high
velocity air is necessary to ensure that water is not thrown out. This
chamber is situated adjacent to and on the trailing side of the main suction
box, and is very effective though difficult to engineer in the space available.
The alternative, of course, is to try to delay the throw-out of water in the
usual way (possibly by using an extra wide trailing seal-strip or an air-
bleed arrangement) until a point round the roll is reached where it is
possible to construct a tray between the top roll and the felt which catches
and retains water thrown out from the holes. In addition to this a nozzle
can be built to project into the nip with a high-velocity low vacuum air
connection to catch and draw up the spray.
The paper will always stick to the smoother surface it contacts in a press
nip so that transfer is usually eased by selecting a denser and closer transfer
felt and this fortunately approximates to the desirable characteristics of a
second paper felt. The suction applied to the top roll box can then be
reduced to a minimum which considerably lessens the water taken into
that box.
The same general principles apply in whatever form the transfer press
operates though it will be noted that where the paper is on the underside
of the carrying felt, as it would be for example if coming straight from a
suction pick-up roll, transfer is from the top felt to the bottom felt. In this
case it would not generally be practicable to use the top roll suction box to
extract water and normally the top roll would be plain and the bottom roll
suction box would undertake the dual function of aiding the water removal
and effecting the transfer.
331
The double-felted transfer nip has now lost its popularity in favour of
transfer from a pick-up felt direct to a granite plain roll. With this arrange-
ment it is simpler both to feed up the sheet and avoid the complications of
water throw-out which have just been described. Except for very lightweight
papers the open draw necessary from the plain roll to a second press felt is
not disadvantageous and avoidance of two felts in the transfer nip enhances
water removal. The pick-up roll may have two separate suction compart-
ments, one to effect the pick-up and the other to aid water removal in the
second nip where the web transfers to the plain roll, or in some modern
designs a single wide suction box accomplishes both functions. The latter
arrangement ensures that the felt and paper adhere to the suction roll
between the point of pick-up to transfer, but is perhaps not so flexible for
adjusting water throw-out at the transfer nip in cases where this is critical.
A further advantage in transfer to a plain roll is that broke disposal is
simpler. The full width of web can be doctored into an extension of the
hog-pit underneath the plain roll, whereas with transfer to a second felt it
is necessary to make provision to dispose of the full web at the second press,
which with faster machines is often inconvenient.
4B. 4 SUCTION ROLL CONDITIONS

The preceding section has dealt in detail with the position of the suction
box in a suction press roll and with the critical setting of the trailing
sealing strip. These factors have been seen to play an important role in
controlling conditions particularly in the outgoing side of the press nip.
In the present section attention is directed towards examining the influence
of the vacuum exerted in the suction box and in addition some reference
will be made to the air-bleed principle. A greater or higher vacuum is
understood to mean a lower absolute pressure, i.e. a higher amount of
vacuum relative to atmospheric pressure.
4B. 4 1 Factors affecting the suction box vacuum

It is common practice with a suction press to use a constant-volume
vacuum pump, usually of the liquid-ring type. A vacuum breaker or limit-
ing device of some other kind may be used to prevent the vacuum rising
beyond a point which would cause trouble on the press; as the vacuum in
the press suction box increases, added friction at the box seals usually has
a noticeable effect on the power demand which, if allowed to carry too far,
would lead eventually to overload of the press drive motor or belt drive.
This is particularly important at start-up when the press load will initially
be very heavy until the bearings and other equipment approach equili-
brium temperature.
During normal operation the vacuum attained in the suction box
becomes a useful indication of performance and condition of the press. The
greater the volume of air drawn by the pump the lower will be the vacuum
achieved; the running vacuum thus gives a measure of closeness of the seal
in the press roll and can be expected to be higher when the paper is thicker
or less porous and when a felt is getting older and so is less porous. However,
332
evacuation of air from the holes in the press shell can also be shown to
have an important bearing on the normal running vacuum because the
holes entering the influence of the suction box carry air which is initially at
atmospheric pressure; the greater the vacuum in the box the higher will
be the effecitve volume of air (reckoned at normal atmospheric pressure
and temperature) which the pump will have to cope with to reduce the
air pressure in the press holes to that in the box. As the holes in a press
roll become plugged they carry a lower volume of air and thus, with
other conditions the same, a greater vacuum would then be expected.
A.lR THR.OIJGH
FEL-r' ANO SEA.I..:'>.
0 co lZ 18 24
VACUIJM 1 INCHES Hs·
Fig. 4. 9. Characteristics of a vacuum pump compared with equivalent volume of air
required to evacuate holes in press shell (after Wahlstrom)
There are two reinforcing considerations governing the vacuum achieved.

A higher vacuum is associated at the pump with a lower volume of air but
it also means that more of the air that is drawn through the pump will
simply have come from evacuation of the holes and less will have been
drawn through the sheet (or equivalently through a poor seal). Likewise a
lower vacuum is associated at the pump with a greater volume of air but
also less of the air drawn through the pump comes from evacuation of the
holes and more must have been drawn through the sheet (or through a poor
seal). It appears then that the vacuum is a sensitive indication of press
condition because quite small changes in sealing conditions resulting in
alteration in the volume of air drawn through the paper and felt will give
correspondingly large changes in vacuum. These points are illustrated in
Fig. 4. 9 which shows the characteristics of a vacuum pump as given in
reference 21.
333
4B . 4 2 Effect of varying the suction box vacuum

These comments do not resolve the question of which is the more desirable
in order to extract water from the sheet, a high vacuum or a high flow of
air through the sheet? It is clearly not possible to have both so it is interest-
ing to turn to reports of experimental work to see what indications there
are as to which is preferable.
There are only two reasonably comprehensive reports of experimental
work in the literature in which the effect of deliberately altering the suction
box vacuum with other conditions constant has been investigated.
Jordansson (10) on the experimental machine at Stockholm decreased
vacuum from 14 in. Hg to 4 in. Hg and noted only a small decrease in
dryness of the sheet leaving the press. The quantity of water collected
through the pump depended to a much greater extent on the machine
speed and above a certain speed (which depended on other operating
conditions) no water at all was extracted by the pump, leaving a greater
quantity to be ejected from the holes further round the roll.
Wahlstrom (21), in his work, varied the vacuum on a second press
between 19 in. and 9 in. Hg and found no significant change in paper
moisture except at very low nip pressures when decreasing the vacuum
increased the moisture content of the paper.
Dixon (1) and Sulatycki (20) confirm that little water is extracted in the
suction pump so the concensus of opinion is that at reasonably high speeds
and nip pressures the vacuum in the suction box is not critical. It is probable
that to a great extent the suction in the press holes is important only
insofar as it helps to direct air flow in the critical stage when the felt
leaves the roll and the suction box is unsealed.
4B. 4 3 The air-bleed arrangement

If the actual vacuum in a suction box is not of great importance can it be
said that increasing air flow by allowing air to leak directly through the felt
and thereby running at a lower vacuum improves the performance of a
press? This is the principle of the air-bleed as described by Molsberry (6)
in which an extra-wide suction box allows air to be drawn through the felt
on either side of the nip; the paper follows close to the top roll and does
not provide any sealing effect entering or leaving the nip.
There are some practical difficulties to this arrangement, especially on
entering the nip, where Pollard (15) has reported trouble with blowing and
blistering which could only be overcome by wrapping the sheet tightly
round the top roll. In addition when leaving the nip, as the felt is still under
vacuum when the paper separates from it, there is a natural tendency for
the paper to follow the felt which can be very troublesome when threading
through.
These difficulties apart, the evidence of whether the flow of air through
the felt has a significant effect on press performance is not sufficient to
draw definite conclusions. Jordansson (10) simulated an air-bleed arrange-
ment on the outgoing side of the nip and found a slight increase in dryness
but only at slower speeds where water was still being drawn through the
334
suction pump. Wahlstrom (21) experimented with moving the position of
the suction box in a second press and with increasing air flow at a constant
vacuum by allowing air to be drawn through the felt in the outgoing side of
the nip. Apart from encountering similar practical difficulties to those
reported by Pollard (the paper tended to stick to the felt particularly at the
edges when the felt was dirty) no particular change in sheet dryness
was noted.
Wahlstrom concludes that the position of the trailing sealing strip of the
vacuum box is more important with regard to keeping the felt clean and
dry and the small increase in dryness of the felt achieved with the air-bleed
can have had only a marginal effect on the efficiency. It may be noted
particularly that the influence of reducing the felt moisture content by
air-bleed in the ingoing side of the nip would be expected to be greater for a
second than a first press where hydraulic pressure components within the
felt have less influence on the efficiency of water removal. However, both
these investigations applied only to bleeding air through the felt on the
outgoing side of the nip; it is possible that an air-bleed arrangement for the
ingoing side of the nip would have more effect if the practical difficulties
can be overcome satisfactorily.
The weight of evidence indicates that within wide limits neither the
vacuum nor the air flow has much influence on press performance. It
would therefore appear prudent to keep the vacuum fairly low, thereby
minimizing frictional losses and power requirements. The use of a Sulzer
centrifugal exhauster and blower in place of a normal type of suction pump
would seem to commend itself as a means of satisfying this condition
economically and it may be noted that a recent report (23) has commented
on the lower power consumption and economy of hot air usage in the
dryer section which accrues from using this type of equipment.
48. 5 MACHINE SPEED AND DRAW CONTROL

The effect of machine speed on press performance has already been
mentioned in various contexts. Increasing speed hinders the efficiency of
water removal in two direct ways. First, and most important, the decreased
time of passage through the nip means that expressed water has to travel
faster in relation to the felt to escape out of the nip or into suction-press
holes within the ingoing side of the nip. To overcome this it is important
to reduce the resistance to flow in the felt so a more open felt is needed.
Second, the greater quantity of water handled in unit time means that the
felt has less time to absorb water from the paper under the influence of the
hydraulic pressure difference operating across the surface of contact. To
overcome this a greater specific pressure and hence a greater load on the
press is required. Both these effects are a result of the decreased time in the
nip. But increasing speed has another effect on performance of a press
which has been termed the 'wedge effect'. The hydraulic pressure of water
in the felt rises with increased speed due to the greater velocity backwards
through the felt which is needed to permit removal; for a constant applied
load on the press, this greater hydraulic pressure has the dynamic effect of
335
wedging the rolls apart and allowing a greater percentage of water in the
felt to pass through the nip. The result of this increased separation of press
rolls in practical terms is still disputed.
This direct influence of machine speed has been investigated by
Jordansson (10) and Fig. 4. 10 reproduces some of the curves he obtained
for dryness of the web leaving the press under different press loads. All
other conditions were constant except that above 300 metres/min. there
was a steady increase in the moisture content of the web entering the press
which would contribute to the decreased dryness after the press.
3S
~
II)
U\
<I.I
~ 3o
"~
1l.
.....J
II)
zs
In
"'z
>-
0
cc
2o -----......
0 loo 200 .Soo 400 Soo
h\ACHINE SP!!l D. h\ETRcs/MIN.
Fig. 4 .10. Influence of machine speed on suction press performance at different press
loads (after Jordansson)
Machine speed has other effects on press operation. Felts are run in
faster but because of the greater quantity of water they handle the detri-
mental effects of plugging will appear earlier. Increasing speed will h~ve
an effect on the throw-out of water from a suction press shell and it was
mentioned in the previous section that a speed is reached when little or no
water is taken out through the vacuum pump. The conditions governing
the direction and force of the throw-out of water in relation to the relative
position of the end of the trailing sealing strip and the point where the felt
leaves the roll are also influenced by the time element; a large change in
machine speed might require an alteration in the direction of the felt run
leaving the nip of a suction press if rewetting of the felt is to be avoided.
4B. 5 1 Draw from press rolls

It is not proposed here to discuss the question of draw control at the
presses in any detail as the subject has already been considered with regard
to an open draw from the couch where the influence of the various factors
is more pronounced and readily observed. Setting the speed difference
336
between the presses and between the last press and the drying section to
produce a suitable draw tension is nonetheless an important part of the
machineman's activities, even though the situation becomes less critical
as the sheet gets drier. In one sense the conditions at press draws are
simpler than at the couch, as there is no question of adhesion varying with
the reabsorption of water from the wire. But the same general conclusions
apply for press draws as the couch, that in practice they are almost always
run too tight; once the machine is settled down, in most cases draws
should be slackened back to a point where the angle of take-off is close
to 90 deg.
The draws at the presses in most cases involve a speed differential of
about 1 to 2 per cent. and on many machines, especially fast ones, it
would be a useful guide to the machineman if he had an accurate indication
of this. Such an indication would be partly useful in pre-setting the draw
at start-up, but the main advantage would be to give a measure of the
adhesive force between the sheet and the roll and prevent it becoming
excessive. This can be done to some extent by observing belt positions, or
some mechanical or electrical setting, but all of these are influenced by
other factors.
Recent experimental work on handsheets by Radvan and Karpati (31)
has shown that adhesion increases with speed and beating (of sulphite
pulp), while other variables such as fibre composition, temperature,
ingoing moisture content, and roughness of the roll all have some effect
on adhesion. Varying the pH of the water in which the handsheets were
made between 7·0 and 4·5 surprisingly showed no effect on the adhesion
measurement. On many machines the pH of the backwater is treated as a
sensitive value to watch and if allowed to rise too high from an acid
condition the press doctor on the top roll will soon exhibit a growing
collection of fibre stripped from the sheet.
48. 6 MOISTURE CONTENT OF PAPER ENTERING PRESS

Two investigators, Jordansson (10) and Wahlstrom (21), have produced
some figures on the effect on press efficiency of differences in the moisture
content of the paper entering the press. It would be expected from the
theory that variations in ingoing moisture content would have a more
pronounced effect in the second press than the first because the point within
the nip where saturation of the paper is reached is more critical in the
second press. Wahlstrom's work, which covered both a first and second
press, confirmed this; in Fig. 4. 11 it can be seen that the slope of the graph
relating ingoing to outgoing moisture is much steeper for the second press.
It also appears from both sets of observations that the action of the
press has a damping effect on moisture variations in the sheet; an alteration
of couch moisture content from 80 per cent. to 88 per cent. in Jordansson's
experiment led to only a O· 5 per cent. increase in the moisture content
leaving the press. In Wahlstrom's graphs a reduction in the couch moisture
content from 4·0 to 3·5 lb. water per lb. paper (12·5 per cent. decrease) led
to a change from 2·22 to 2·08 lb. water per lb. paper after the first press
337
(a 6·3 per cent. decrease). Similarly this led in turn to a change from
1·97 to 1·91 lb. water per lb. paper after the second press, which repre~
sents only a 3 per cent. decrease. Thus a 12·5 per cent. change at the
couch produced on average only a 3 per cent. change after the presses.
F\RST ?Ress.
~·G ~·8 4·o 4·2.

INGOING WATER/ FIBRE RA.TlO.
secoN D PRESS.
2·o 2·2 z.4 2·G Z·B

INGOING WATER/F12>RE: RATIO.
Fig. 4. 11. Effect of changes in ingoing moisture content on outgoing at a first and
second press (after Wahlstrom)
These results are important because they indicate that the actual quantity
of water removed by a press is very dependent on the moisture content of
the paper entering it. Further, the shapes of the curves in Fig. 4. 11 show
that the more water there is to remove the more proportionally the press
will take out. It is not, therefore, so critical that the paper should leave the
couch as dry as possible.
Certainly it appears that the drier the paper entering the press section
the drier it will be entering the drying section, other things being equal,
but the shape of the curves in Fig. 4. 11 shows this to be very much a case
338
of diminishing returns. It would hardly be worth while prejudicing forma-
tion by running insufficient water on the wire, or shortening wire life by
having excessive vacuum on the suction boxes, if the motive were only to
achieve a low moisture content leaving the couch. On the other hand if the
0 So loo loo
sue:isTANCE, G.5.M·
Fig. 4. 12. Variation in moisture content of sheet leaving press with sheet substance
(after Jordansson)
sheet is allowed to become much damper entering the press section it is

evident that it will begin eventually to have an increasingly important effect
on the moisture content leaving the press section.
4B.7 PAPER PROPERTIES

The influence of paper properties on press performance has already been
mentioned in two contexts. Variations in the degree of beating of the stock
influence the adhesion of paper to the press roll. Also to certain extent
the thickness of the sheet has an effect on the moisture content leaving a
press nip.
Jordansson (10) found that with other conditions constant there existed
a certain sheet substance at which the dryness leaving the press was highest.
For lower substances the reabsorption effect would be expected, following
the experimental work of Sweet (26), to cause the dryness to decrease with
increasing sheet substance. Presumably also, increasing the substance
beyond a certain point so increases the quantity of water entering the nip
that there is increasing resistance to the greater flow within the felt and this
has the effect of reducing the final dryness of the sheet. Fig. 4. 12 reproduces
the graph obtained by Jordansson, showing the effect of substance on the
moisture content of the sheet leaving the press. The moisture content
leaving the couch also varied with the sheet substance, but not sufficiently
to account for the shape of the curve as shown.
339
Both Jordansson and Wahlstrom (21) state that increased freeness gives
an increase of dryness of the sheet leaving the press and this is attributed
to the effect of the fines content and fibrillation on the size of capillaries
in the paper. A greater degree of beating and the presence of more fines
produce a smaller average capillary size; this will both retard the passage
of water into the felt under the hydraulic pressure gradient within the
ingoing side of the nip and increase reabsorption from the felt within the
outgoing side of the nip, both of which will reduce the water removal.
A graph given by Jordansson is shown in Fig. 4. 13 and illustrates the
35
~
\Jl
Ill .30
uJ
z
>
Cl
0
"25
lo 3o So 7o
ORAINAG~
0
RES!5TA.N.CE, 6. R.
Fig. 4 .13. Variation in moisture content leaving press with drainage resistance (after
Jordansson)
effect of beating on the moisture content of the sheet leaving the press.
(There was in this case hardly any variation in moisture of the paper
leaving the couch.) It was also noted that the vacuum in the suction press
increased from 11 in. to 14 in. Hg over the range of 19 deg. to 66 deg. S.R.,
which further illustrated the effect of closing up the sheet.
A similar point has been made by Wrist (47), who considers that next
to loading on the press the furnish has the most influence on water removal.
A sheet made from a pulp high in fines and fibre debris was found relatively
difficult to dewater by pressing in an experimental set-up, but removing the
fines fraction considerably increased the water removed under similar
conditions and also lessened the tendency to crush.
4B. 8 TE1\1PERATURE
By lowing the viscosity of water and hence reducing the flow resistance, an
increase in temperature would be expected to assist water removal at the
presses in the same way that it promotes drainage on the wire. There are
340
one or two reports evaluating the effect of temperature change produced by
applying heat at the press in various ways and these vary somewhat in
their estimates.
Wahlstrom (21) carefully tabulated the effect of temperature change
during start-up periods and calculated that a 10 deg. F. increase in tem-
perature of stock in the breastbox reduced the water/fibre ratio of the paper
by 0·15, 0·08 and 0·04 lb. water/lb. paper at the couch, first press and second
press respectively. The effect of temperature rise in the breast box is
considerably reduced by the time the presses have been reached so that
these figures in fact refer to a lower effective temperature rise at the
presses; an actual rise of 10 deg. F. in temperature at the presses was
calculated to yield a reduction of 0· 12 lb. water/lb. paper leaving the
second press. This reduction represented approximately 6 per cent. of the
water in the paper entering the drying section and would thus give a very
considerable decrease in steam consumption.
Increasing temperature at the presses has been recognized for some time
as an effective means of improving press performance and there are several
publications and patents which deal with this. Heating the stock itself in
the breast box or pit, or by steam directly on the wire, could hardly be
justified solely for the improvement in dewatering it would produce at the
presses. These methods are adopted on many machines where a quicker
drainage on the wire is necessary to increase production, but the improve-
ment in press efficiency would then be regarded as a marginal benefit.
Other methods depend on applying heat directly at the presses. The use
of infra-red radiation produced from oil, electricity or propane gas and
directed on the paper has been reported and various claims have been made
that such applications are economically worthwhile (3, 17). It is important
to have the predominant wave-length of the radiation chosen to suit the
particular application since the useful wave-length band for heating is
restricted and depends on the thickness of water film.
Preheating the felt or paper with steam or hot-air is another approach
that has received attention while another technique involves impinging
open flames from gas burners directly onto the sheet. Yet another system
is to pass a high voltage alternating current through the felt and paper
either in the press itself, when a special metal wire can form one of the
conductors, or separate from the press (12, 16, 29). The possible hazards
involved and the economic validity of each of these methods require very
careful consideration and to provide an accurate assessment it would be
necessary to obtain figures relating reduction in steam consumption with
increased dryness of the paper entering the drying section.
It is reasonable to assert that the most promising technique both
practically and financially would very likely involve heating of the felt in
the second press. Heating of the felt itself is important because it carries
round by far the greater quantity of water entering the press nip and it is
through the felt that water from the paper is expressed; also of the two
presses improvement at the second press is more likely to be useful because
it handles much lower quantities of water and the effect of a reduction in
moisture content of the paper carries straight to the drying section.
12 341
48.9 THE PRESS SECTION
48. 9 EQUIPMENT
It is not proposed in this section to enter into any detail regarding the
relative efficiency of various types of presses even if enough comparative
information were available to be worth quoting. The general characteristics
of the basic types of press are well known and the present trend in design
seems to be towards a compromise between the ease of maintenance and
layout of an ordinary pick-up and suction press followed by one of several
types of second press compared to the lower overall power requirement and
use of machine space of the stack press.
48 . 9 1 Press rolls
In the discussion on the influence of nip width it was mentioned that,
especially where load is limited, it is important to keep the width of the
nip narrow. One way of achieving this is to use harder rubber covers and
comparative results on an experimental plain press with rubber covers of
different hardness have confirmed that this has a very important influence
on water removal efficiency (20). There have also been other reports of the
beneficial results obtained on machines by using harder roll covers; in one
case increasing hardness from 45 deg. to 17 deg. P & J decreased the
moisture content by as high as 4 per cent. (21). In the same way, smaller
diameter rolls and thinner rubber covers will reduce the nip width and
hence can be expected also to have a similar effect in improving the efficiency
of the press.
The main disadvantage to running with very hard rubber covers, or
even no covers at all, is that small discrepancies in the camber shape
become relatively more troublesome. Any small bruise in the roll causes
a wet spot in the paper and produces a blemish at the calenders. Opposed
to this softer rolls, apart from reducing the press dewatering efficiency, are
more difficult to grind and the rubber distorts more within the nip. This
distortion could, if the roll were very soft, be sufficient to reduce the
diameter of the holes in a suction press to a point where they ceased to be
effective. But apart from this the rubber bulges on either side of the nip,
particularly on the ingoing side because of the effect of the bottom roll
drive, and the relative motion of the rolls in this region will subject the felt
to greater wear (20).
The holes in a suction press should be as small as possible. Water
collected in the holes of a typical press occupies only a small fraction of the
volume of the hole so there is no fear of the hole becoming waterlogged.
The smaller the diameter of the holes, the more holes can be put in the shell
and the smaller can be made the distance between holes round the circum-
ference of the roll. The advantage of this has already been stressed and
would be especially valuable in the first press. Further, the volume of a
hole is proportional to the square of the diameter, so a greater number of
smaller holes will have a proportionately lower volume of air to evacuate
and this will reduce the demand on the vacuum pump.
Whether the practical limits of reducing hole diameter have not already
been reached in an endeavour to minimize the possibility of shadow mark-
342
ing is difficult to say. In mills which experience trouble with the holes
getting made up it would probably not be feasible to reduce their diameter
further. Also small lateral distortions of the rubber cover in relation to the
metal surface of the main shell of the roll onto which it is bonded would be
more troublesome in throwing the holes in the rubber and shell out of
alignment.
Finally with regard to suction roll holes it may be mentioned that it is
preferable to pitch the holes slightly across the length of the roll. This evens
out the area of holes coming under the influence of the suction box at any
one time as the roll moves across the box, and reduces the chance of
oscillation of the air flow and vacuum.
An investigation has been made (9) in an attempt to reduce adhesion of
the paper to the top roll which becomes more troublesome at faster speeds;
it appears that the adhesive force depends partially on chemical character-
istics and partly on the micro-geometrical nature of the roll surface.
Ground but unpolished granite has lower adhesion than polished granite,
but some hard synthetic rubbers give the best results. These general
findings are in broad agreement with recent experimental work by Radvan
and Karpati (31).
48. 9 2 Ancillary equipment

It is becoming more common now to see the doctors on top press rolls
air-loaded and oscillating slightly across the machine. These features have
a definite advantage in minimizing wear in localized positions across the
roll and reduce considerably the possibility of trouble eventually occurring
with fibre lumps passing under the doctor and into the press nip. Air-
loaded suction box seals are also proving beneficial for the flexibility they
offer.
On faster machines the simple tray behind the doctor becomes in-
adequate to collect the sheet as it runs up the press and is being superseded
by a belt conveyor which removes the broke from the machine as it piles
up behind the doctor. Alternatively a special rotating screw pulper to
which water is added can be used to break up the sheet and dilute it for
pumping to the broke chest. It is possible to have both these systems
started manually by the operator although, as they would usually be
installed on faster machines, automatic operation is desirable to avoid
jams. In this case the mechanism could be actuated by signals from
photo-electric or ultrasonic break indicators but an alternative is to use a
light aluminium feeler attached to a microswitch which senses when the
sheet begins to build up on the doctor and then sets the conveyor or
pulper into motion. Modern designs for the first press overcome the
problem of broke disposal by arranging for the web to be doctored off
straight into an extension of the hog-pit.
On some presses water throw-out from the holes in the suction roll can
be delayed sufficiently to cause wetting of the felt at the ingoing side of the
nip and this would be expected to have an adverse effect on the water
removal efficiency of a press. This was confirmed by Wahlstrom (21), who
343
devised a rubber doctor suitable for contacting the bottom suction roll of
a second press. The goctor was shaped to deflect water thrown against it
away from the roll and proved to yield a significant improvement in
performance. A subsequent modification of the original design has been
described by Delisle et al. (37).
It is usual to have a water shower inside a suction roll, and stretching
across the full length of the roll, the purpose of which is to lubricate the
inside of the shell and minimize friction with the suction box seals. Such a
shower is particularly useful during start-up periods when it is important
to prevent the press using excessive power, but later on it is rarely necessary
to use a high pressure on the shower and may even be possible to run
without it at all. If the press uses no more power nor sounds any different
with a reduced shower pressure it is preferable to retain the lower pressure
in order to avoid the possibility that the splashing of water into the suction
roll holes may adversely affect water removal in the press nip (the author
has known even a clean, well-adjusted spray to increase the average
moisture leaving a press by 2 per cent.). Further, in mills using hard water
it is easy for uneven wear of the suction box seals to occur and uneven
deposits in the holes across the roll to build up due to the shower producing
ajet stronger in some places than in others. This localized plugging of press
holes and wear of the seals is difficult to observe and correct and eventually
produces uneven water removal causing both the paper and felt to run
wetter over the area opposite the badly made up holes. Using as little
water in the shower as possible helps to reduce these difficulties, but in
some cases it may be advisable in addition to oscillate the shower. In some
types of suction roll the spray is dispensed with in favour of water purges
on either side of each sealing strip; this should, given a satisfactory design,
be less liable to cause unevenness across the length of the roll.
344
CHAPTER 4C
RUNNING THE PRESS SECTION

4C. l DAILY OPERATION
As for other parts of the paper machine the problem of keeping a control
on the efficiency with which the press section operates is now considered
from the standpoint first of day-to-day working, and second of longer-term
running. In the latter case most attention will be given to the question of
felt changing because this is the most important single maintenance prob-
lem. But other aspects of maintenance will be considered, in particular that
of suction roll cleaning; also the purpose and value of having periodic
tests of the press section efficiency will be discussed.
Another problem of press operation is that of keeping the press action
as near as possible the same the full width across the machine. Some features
of this problem are daily and some long-term, but it will be more convenient
to treat this subject separately.
4C. 1 1 Essential measurements

In the preceding chapter the influence of various factors on press perform-
ance was considered in detail and three measurements were frequently
referred to: the load on the press, the vacuum in the suction box (for a
suction press) and the power used by the press (for electrical drives).
Each of these measurements can give a great deal of useful information on
the state of the press and are vitally important.
An indication of the load on the press at both sides of the machine
obtained by one or other of the methods discussed in 4B. 1 3 is important
primarily for accurate setting of the load applied. At start-up and when
running-in a felt load has to be varied, perhaps over a fairly short interval,
and this can be done more accurately and consistently when a proper
indication is available. Although once a felt is run-in the tendency will be to
adopt the same setting all the time, it may often be possible to put on a
heavier load without crushing and an accurate load indicator is then useful
for giving a figure to the higher load applied so that it may be tried again
under similar conditions. When a press is run just under crushing load
almost all the time, observation of how this load varies under different
conditions should give a guide to the state of the felt and suction roll holes.
Unless the press is faulty the load should always be exactly the same on
both sides, which is why it is essential to measure and set the load applied
at both the back and front of the machine. Deliberate application of
uneven loading as a means of correcting a consistent moisture difference
caused by a fault elsewhere on the machine is often resorted to, but should
be avoided because it becomes over a period of time less and less effective
and gradually ruins the press; this point will be fully dealt with in the
discussion on cross-web variation in 4C. 3.
345
4C.12 THE PRESS SECTION
Both the vacuum in the box of a suction press roll and the power
consumed by the press are of obvious importance for showing if something
is radically wrong. But apart from that, careful observation and logging of
the appropriate figures can give a useful indication of the state of the felt,
suction box seals and holes in the press roll. If the vacuum is lower than
usual it does not mean that the press will be working less efficiently but it
may indicate that one of the seals is leaking or that the pump requires
attention; greater speed will also reduce the vacuum because more air will
be drawn from the press holes in a given time. More often the vacuum will
progressively increase as the felt or press holes get made up and it should be
realized that there is no benefit in letting it get too high.
The power consumption depends mainly on the frictional forces involved
and these are increased by greater load on the press, greater speed and
higher vacuum in the suction box. The draws would be expected to have a
small influence on the power but probably the condition of the felt roll
bearings and the tension in the felt are of more importance. After the sheet
is threaded through at start-up the power consumption should gradually
drop as the press warms to equilibrium running conditions; nevertheless,
it is useful to have a warning device linked to the ammeter to ensure that
the attention of the machineman is quickly drawn should the load become
excessive.
4C. 1 2 Useful measurements

There are two variables of press operation that, especially on faster
machines, could beneficially be measured: the draws and the felt tension.
The value of measuring the draws at the presses depends on the accuracy
of the method; after the first press the draw may be under 1 per cent. of
the machine speed and to obtain useful information an accuracy of
probably at least 5 per cent. or so of this difference would be needed. This
is a very exacting requirement but such accuracy must be attained if a
draw measurement is to be of more use than the rough indications which
are available anyway to the machineman.
The main value attached to measuring the draw is similar to measure-
ment at the couch which has been discussed in detail in Part 3. Indirectly
it enables a figure to be given to the tension in the sheet when the run of
the paper web is in the usual operating position. When a draw is normally
run fairly tight (in the sense that it is adjusted so that the paper peels off
the top roll at a pretty acute angle) it is not easy by inspection to see that
the tension is greater than usual, as it would be if adhesion of the paper to
the press roll were higher. Measuring and keeping the draw to modest
proportions, apart from helping to reduce the possibility of frequent breaks
due to overstretching, will be of particular importance when the paper is
required to have equal strength and moisture expansion properties in both
directions (moisture expansivity is closely dependent on permitted shrink-
age though this effect is of rather more significance in the drying section).
The tension of the felts in a press section is very frequently much too
high. Apart from the probability of using more power to pull the felt
round due to the increased friction on the felt roll bearings, the felt will be
346
THE PRESS SECTION 4C.l 3
strained by excessive tension and will almost certainly wear faster. High
tension can be justified, for example, when the felt is too wide at the
beginning of its life and tightening enables it to be brought to the necessary
width for comfortable running. But normally the stretch roll should be
slackened well back within the limit of the felt threatening to become
uncontrollable.
Huyck Felt Co. have designed a simple device for measuring the tension
of a felt when running. This instrument consists of two smoothly bevelled
supports joined on a frame about one foot apart; in between is a weight
which can fall below the level of the supports by a small amount which can
be measured by a micrometer suitably attached to the frame of the
instrument. In use the instrument is held by a handle attached to the frame
and allowed to rest on the running felt; the weight pushes the felt down
below the supports by an amount which depends on the tension of the felt
and this can be determined from previous calibration of the micrometer
reading.
Using this instrument it is possible to take periodic checks of felt tension
and this may help to keep the tension down to a reasonable level. Alterna-
tively it is not difficult to attach a strain gauge to the stretch roll
mechanism to give a continuous measurement of the tension. Modern
stretch mechanisms which are pneumatically and hydraulically loaded
have the advantage of giving a direct indication of felt tension.
4C .1 3 Measuring moisture content leaving the presses

Each of the measurements discussed above provides essential or useful
information to assist the operator in running the presses efficiently and
help him recognize and diagnose the cause of unusual conditions. But he
still has no direct assessment of how well the press is performing its primary
function-to remove water.
Ideally a continuous record of the average moisture content of paper
entering the drying section would be invaluable. The beta-ray gauge has
been used for this purpose but there are several difficulties which, though
they can be overcome in a research investigation, need careful thought if
the instrument is to be useful for unsupervised daily operation. Two of
these difficulties are worth mentioning. On most machines there is likely
to be substantial variation in moisture content across the web and un-
representative readings would be produced if, for instance, the head of the
beta-ray gauge coincided in position with a damp streak in the paper. This
could only be overcome satisfactorily with either a traversing mechanism
or using more than one head and both of these methods are expensive.
The other difficulty relates to the fact that the beta-ray gauge measures the
total weight of the sheet, fibre plus moisture, so that variations in the dry
weight of the sheet will affect the moisture content figure. The machine
direction substance fluctuations may not be sufficiently great for this to
give a serious error when averaged over a reasonable period of time, but
cross-web variations would on all but the most stable machines add com-
plications in ensuring a representative position. A beta-ray gauge reading
from the reel-up subtracted from a press gauge in the same cross-web
347
position provides one solution to this problem; but it does not even then
take account of variations in the moisture in the paper at the reel-up
though this presents no particular difficulty when a moisture meter is used
at the dry-end.
Despite these problems some continuous indication of the average and
also in some cases the profile moisture content entering the cylinders is
likely in the future to become standard equipment and at least one type of
moisture meter has been developed which is claimed to have a high
accuracy at this position. The advantages for the operator would be very
important because he would be able to observe immediately the effect of
making alterations to the press operating conditions. Adjustment of load
on individual presses and the balancing of total load between the different
presses would make a change to the moisture content of the sheet leaving
the presses which could be assessed straight away, and this would en-
courage the operator to experiment with changing the load much more
than is customarily done at present on most machines. Further, when
alterations are made to the wire conditions, in particular to the volume of
backwater circulating round the wire pit and mixing pump and to the
vacuum on the suction boxes and couch, any change in the moisture con-
tent of the sheet at the couch which became transmitted to the drying
section would be immediately shown up. Over longer periods changes in
the average level of dryness of the paper leaving the press section would
permit the felt condition and the usefulness of cleaning the felt in various
ways to be assessed. Conditions in the drying section could also be
watched more easily if the dryness level entering the section were com-
pared with the steam flow.
4C. 1 4 Measuring water extraction at the presses

It is possible to obtain information from the presses yielding most of the
benefits mentioned in the preceding paragraph without attempting the
difficult task of measuring the actual moisture content of the web. This can
be done by obtaining a measure of the volume of water extracted by each
press. In some respects treating each press individually has additional
merits from the point of view of trouble-shooting and considering the
relative ease and simplicity with which it is possible to measure water ft.ow
from presses it is surprising that this is not done on more machines.
In the case of a simple plain press, water collected in the trays is generally
channelled to one side of the machine and led to a drain. It is not difficult
to arrange the ft.ow so that it passes over a suitably-sized weir or flume
giving a direct visual measurement of the flow. By using a simple float
and cable element or the back-pressure in an air bubble-tube system placed
below the level of the weir, this can be elaborated without much cost to
give a continuous record; the flow from two presses can be displayed
simultaneously on a two-pen recorder.
The reliability of this measurement depends to a large extent on whether
other flows of water to and from the press represent a significant part of the
total flow; this particularly concerns water added to the felt in cleaning
348
THE PRESS SECTION 4C.l 4
and water removed by suction in the felt cleaner and the suction box prior
to the press nip. If these flows amount to 10 per cent. or more of the
main flow from the press tray, and especially in the case of water added in
cleaning they may quite easily, then it would be preferable to ensure at least
that the water added in felt cleaning is kept constant; this is likely to apply
particularly to a second press where the main flow is smaller. This step is
advisable anyway and can easily be accomplished by incorporating a small
rotameter in the line to allow the flow to be set; alternatively, if the flow
with a fixed entry valve position hardly ever varies, it may be sufficient to
set the valve to a standard opening and have the flow checked and adjusted
periodically by the laboratory staff. With a suction press, water extracted
by the suction pump presents an added difficulty. However, it has already
been mentioned that in most cases the suction pump appears to extract
little water so that any variations in this flow may be regarded as negligible.
A simple check can be made by temporarily shutting off priming and sealing
water to the pump and measuring the remaining volume, which should
then be the flow from the suction press box; provided this flow stabilizes
before vacuum begins to alter as the pump is starved of water, the figure
obtained would be reasonably accurate. If such a test proves the flow to be
substantial it can be added to the main flow from the press tray for
measurement. In this case it would be important to keep the pump sealing
water constant in a similar way to the felt cleaning water. Finally, the
water added in the suction roll spray would also need to be regulated as
this usually represents an appreciable fraction of the total water removed
in the tray.
Although the points detailed above may seem to involve many compli-
cations it must be remembered that the preliminary estimations of the
various auxiliary flows require little preparation beyond arranging suitable
tappings. If any were found to be variable there would be some advantage
anyway in arranging for the flows to be regulated from the point of view
of consistent press operation, apart from increasing the accuracy of the
main press water flow measurement. The flow of water in a suction press
internal spray has been known to give a great deal of trouble by causing
cross-web moisture variations until someone discovered that a valve that
used to be just cocked open had somehow gradually got turned fully open
and was admitting a huge quantity of water to the spray.
To assess the potential value of measuring the flow from individual
presses (assuming the necessary care with accuracy were taken in the
manner described), it is worth considering the relative influence of various
factors which will affect the flow. These may be divided broadly into
variations in the paper itself in the form of moisture content entering the
first press, substance, freeness, temperature, etc., and variations in the
performance of the press itself due to load, suction box vacuum, felt
condition, etc. Variations in the press water flow caused by the paper
affect a first and second press in the same way, though to different extents,
so that on a combined double pen recorder showing both flows the traces
would follow roughly parallel. By contrast, variations due to the press
itself (load, felt condition, cleaner operation, etc.) will show up as a
349
4C.2 l THE PRESS SECTION
separate movement in the relevant trace and the effect of making altera-
tions can be readily observed. Further it should be very easy to see what
happens to the water removal of a second press when an alteration is made
to the first press, and hence to adjust the total load appropriately. In
day-to-day running the operator should be able to distinguish without much
difficulty between effects caused by the paper and those attributable to the
individual presses and this will assist him in no small measure to keep the
performance of the whole press section at a high level.
4C.2 MAINTENANCE OF THE PRESS

4C.2 1 Felt changing
The most frequent and important aspect of running a press that comes
under the heading of maintenance in its broadest sense is clothing renewal.
The life of a wet felt can vary from a few days on a fast machine to several
weeks. On most machines felts are generally taken off the press when it
becomes fairly obvious that their performance has deteriorated; the felt
may have worn severely in some place or more likely it may have become
so hard and plugged that the load has to be constantly relieved from the
press to prevent crushing while a noticeable increase in drying steam
pressure become necessary.
Policies with regard to felts vary considerably but as with other clothing
it is fairly safe to say that in a majority of mills the papermaker regards it
as one of his obligations to get as long a life as possible from the felt. He
may do this by stipulating careful washing, and in some cases possibly also
turning, of the felt at every opportunity when the machine is stopped, or
even require that the machine is stopped occasionally for that purpose alone.
Some form of record will be kept in which the various makes and types
of felts used on the press are listed, together with their cost and a means of
assessing their performance. The latter is usually in the form of the number
of days or weeks each felt lasted on the machine, the total distance run, or
the number of tons of paper it made and perhaps also the cost of the felt
per ton of paper. Such details together with general comments on the
running of the individual felts are, of course, an essential means of ensuring
that only suitable felts are used, and they help considerably to decide the
physical make-up and characteristics of the felt best suited to each particular
application.
The main weakness in this simple record system is that it emphasizes the
importance of getting felts which run satisfactorily, in the sense of being
dimensionally stable throughout their life, withstanding accidental damage,
not having weak places, etc., and which last longest, produce the most tons
or generally do best in whatever measure is taken of their performance.
These aspects are certainly important but there is no virtue in squeezing
the last ounce of life out of a felt at the expense of appreciable loss of
production in downtime. It is fairly easy to use most felt records to
determine in the first instance whether felts are being kept on too long, and
if so to estimate roughly how long felts of similar types may be kept on
the press before they become uneconomic.
350
THE PRESS SECTION 4C.22
Suppose, for the sake of example, that felts are normally used on a press
from four to six weeks and that a straightforward record of the number of
tons made each week by each felt is available. On machines which make a
variety of different types of paper a simple total tonnage figure for each
felt may not be considered representative of the work done by the felt but
there is no reason why the individual figures for each making should not
be weighted in some way to take account of this; this adds complications
to the business of keeping convenient records, as would allowance for
variable deckles and paper lost in the dryers but included in the tonnage
figures, but these points must be decided on their merits and can probably
in most cases be ignored without unduly affecting the accuracy of the
figures.
During the first two weeks of the four to six week life, production can
reasonably be regarded as uninfluenced in any way by the state of the felt
(this assumes that the running-in period is not so long and difficult as to
involve loss of production-only greater steam consumption). An average
for these first two weeks taken over many felts will produce a figure which
may be regarded as normal when the felt is satisfactory, and this may be
compared with the average production in succeeding weeks of the felt's life.
By a simple calculation it is easy to determine whether the average loss of
production by retaining a felt for an extra week of life is uneconomic.
This procedure is, of course, a standard maximization problem and a
person with a little mathematical sophistication would use the tonnage
figures in a more rigorous manner to estimate how accurate the results
are and to check that the conclusions do not give a false impression. Also,
if a reasonably comprehensive machine log is kept, it is possible to check
from recorded downtime caused by attention to a felt that this becomes
greater later on in its life. But even a straightforward estimate along the
lines first suggested is likely to show up some alarming facts and it is often
surprising how little downtime on a machine immediately wipes out the
small difference in average running cost of the felt achieved by gaining an
extra week on the five, four, or less weeks already run.
In the author's experience it is well worth using this approach to deter-
mine whether, as is nearly always the case; felts are being left on the
presses too long, and then to decide on a standard life for the felt in each
position. This has the additional merit of permitting easier organization of
clothing changes and provides the papermaker with a sensible economic
basis for taking decisions to leave or remove a felt which is giving trouble.
It will also, incidentally, be appreciated by the cost accountant working to
standard cost figures and will allow clothing to be more easily purchased in
a systematic manner (even, as an application of operational research
techniques, with a frequency which reduces stocking costs to a minimum).
4C. 2 2 Felt performance throughout its life

In the previous section optimization of felt life was considered from the
point of view of preventing uneconomic loss in production as the felt ages.
But experimental work has been quoted which shows that after a certain
351
4C.2 2 THE PRESS SECTION
length of time on the press the porosity of a felt is frequently reduced by

materials plugging the interstices and then the efficiency of the press begins
to drop. On fast newsprint machines this consideration may outweigh all
others in deciding the life of a felt and as soon as steam pressure in the
drying section begins to require raising the felt may be taken off. Though
admittedly an extreme case, this illustrates that on most machines, par-
ticularly those using loadings and mechanical pulp which might cause
rapid plugging of felts, this factor warrants investigation.
Most of the work involved in an inquiry into this aspect of felt life
depends on obtaining figures which show up changes in the moisture
content of the paper leaving the press. The complexity and variety of the
experimental results obtained and analyzed for this purpose by Wahlstrom
makes any investigation of this subject appear at first sight to be a formid-
able undertaking. Wahlstrom in fact emphasized that normal variations
in ingoing and outgoing moisture content, disregarding extreme conditions
at start-up or with a new felt, could be very high and made it impossible
to characterize a press by means of occasional spot tests. This difficulty was
well brought out by the completely negative results of the analysis of the
1958 questionnaire circulated by the B.P. & B.M.A. Technical Section
which was designed to discover, amongst other things, how water removal
and felt life are affected by felt type and general press conditions on
different machines.
The main problem lies broadly in isolating changes in the water removal
efficiency of the press due to the press itself, and in particular to the felt as
it ages, from those caused by the paper, of which ingoing moisture content
is the most important variable. By careful analysis of the moisture content
of the paper before and after pressing Wahlstrom was able to isolate and
allow for the more important variations caused by the paper itself and this
enabled him to consider the influence of the press and in particular to
discover how serious was the effect of the felt ageing on press efficiency.
But this required press tests involving breaking the sheet for samples at
fairly frequent intervals through the life of a felt. It is unlikely, as the
author has in fact found, that analysis of spot test results, such as many
mills take at intervals of a few weeks primarily for examining cross-web
variations, would yield any useful information on this point.
An alternative approach, which has much to commend itself in a mill
with restricted facilities for testing, is simply to observe the variations in
water removed at the press and measured in the manner suggested in 4C.13.
This preferably requires frequent makings at similar speeds of the same
grade of paper, but over the life of a number of felts any decrease with age
should become apparent. Similar observations of variation in steam usage
in the drying section should enable an assessment to be made of the relative
saving by keeping the felt on after a certain point in its life compared with
the cost of using more steam for drying. The interaction of the presses on
one another, in the sense that the less water is taken out at the first press,
the more will be removed by the second, must of course be taken into
account in a calculation of this sort and allowance must be made for
variations under different loads, the other main press variable.
352
A further point worth mentioning is that felts which cause reduction in
press efficiency by becoming plugged will show a considerable change in
their physical properties before and after use. Unfortunately such changes
occur anyway due to mechanical conditioning and the same difficulties in
interpretation arise as were mentioned in connection with the measurement
of porosity while the felt is still on the machine by using the Huyck porosi-
meter. It is not, therefore, likely that measurement of thickness, weight,
density, tensile strength, chemical composition and other properties can
help much to indicate whether the felt is being left on too long, though they
may be useful in connection with the analysis of cross-web variation and
for long-term checking and comparison of the performance of different
types of felt. As such, these comprehensive tests on used felts would be
undertaken at intervals on typical samples.
Another advantage to be gained by making an investigation of this
nature is that it will give an idea whether or not the existing method of felt
cleaning is adequate. If it appears that a felt should be removed from the
press because it is badly plugged, yet otherwise it is in sound condition,
then some new form of cleaning on or off the machine is obviously indi-
cated. If figures for the increased steam costs due to plugging are available
it will allow a sensible economic evaluation of alternative cleaning methods
to be made.
Finally no papermaker will be content to determine a satisfactory type
of felt and standardize its life, and then regard the job as complete. Felt
manufacturers are constantly introducing new varieties of felt by altering
the material, weave, and finishing processes and it is encumbent on the
papermaker to recognize their efforts in this direction by giving a new type
of felt a fair trial. In fact, of course, if reliance is placed purely on sub-
jective comments received from the machine crews, then the papermaker
deserves to lose the opportunity of improving his presses. It is certainly
important to know that the felt runs satisfactorily, can be controlled on the
press, and does not mark the paper, but how well it performs its primary
function of removing water can only be answered satisfactorily by careful
measurement. Here, again, appears one of the advantages of having to
hand records of the average water removal by felts already in use so that
an immediate comparison is possible.
4C. 2 3 Other maintenance

Apart from the felt there is little on a plain press in the way of long-term
maintenance that can be regarded as directly the obligation of the paper-
makers, as opposed to the engineering and instrument staff. The question
of cambering of rolls is dealt with separately.
On a suction press it is necessary to clean out the holes at intervals. This
should preferably be done on a routine basis for the same reasons that
apply to felt changing, and observation of water removed by the press
should, in this case also, help to lay down the frequency with which
cleaning is needed. It is likely that the holes can get quite small in diameter
without unduly affecting water removal and generally trouble will eventually
353
begin to appear at one point of the roll, rather than over the whole roll at
once, causing cross-web variations.
Formerly the usual method adopted for clearing the holes was to punch
or drill out the deposits from each hole individually, but most mills
nowadays find that it is perfectly satisfactory to use strong alkaline or acid
solutions specially formulated for this purpose by chemical manufacturers.
These are best used at the moderate temperatures recommended and can
easily be applied to press sections from a special tank equipped with a
heater and pump, which circulates the chemical into the press roll and back
through a suitably adapted collecting tray, while the roll itself is slowly
revolved. This form of periodic cleaning can be assisted by the use of high-
pressure water cleaning units which traverse slowly across the roll. In
hard-water areas, where deposits can be very hard to punch out, loosening
the adhesion of deposited solids by this means and then chemical cleaning,
though it may require doing at more frequent intervals than punching, is
then very much cheaper.
The suction boxes, seals, deckles, and sprays inside press rolls will
require periodic inspection and cleaning and are conveniently examined all
at the same time. The sprays whenever practicable should be tested off the
machine where they can be observed and, if of the fan-tail type which are
the best for this purpose, the distance should be checked to prevent
overlapping of the jets in the press roll. If this is avoided deposits from the
water are less likely to cause build-up in press holes at one place across the
roll in preference to another. The load in use and oscillating mechanism
for doctors require frequent examination, and the condition and angle of
the blades to the roll surface also should be regularly checked. Records of
the doctor changes giving details of the type used, condition when removed,
life, etc., are of obvious use if a systematic check is to be kept.

It has already been emphasized that any equipment on the paper machine
should be made the subject of thorough checks at intervals and this applies
especially to the presses. Press testing involving breaking down the sheet
for samples to test for moisture content across the machine is often
adopted on a routine basis for the purpose of controlling and spotting
cross-web variations; this will be considered in 4C. 3 4 but is, in the
author's opinion, of limited practical value. It is, though, important to
obtain this type of data periodically, when one of the major grades of
paper is being run on the machine, for the purpose of long-term records
of performance. The occasions when this is done are preferably chosen to
avoid very new or old felts which may confuse interpretation of the results.
A thorough press test that is well organized need not require the sheet
to be down for longer than five minutes. During this time several samples
are taken at fixed positions across the sheet from all the relevant points of
the press section, placed in suitable air-tight containers, and then tested
for moisture content and the results expressed in the form of water /fibre
ratios. Calculations from this data of water being removed from the sheet
354
THE PRESS SECTION 4C.3 1
in each nip should tally to within 10 per cent. or less with the quantities of
water as measured leaving each press in the tray (sprays, felt cleaning water,
etc., can be temporarily shut off to facilitate obtaining a quick and simple
measurement). Other measurements that may be useful apart from the
obvious ones of press load, power consumption and suction box vacuum
are temperature of the paper, air flow to the suction pump, and draw.
4C. 3 CROSS-WEB VARIATIONS

The problem of variation across the sheet in the press section is essentially
one of keeping the moisture content profile as even as possible. If the
moisture content were even across the sheet leaving the press section then
the presses would have done a good job. Unfortunately this is rarely the
case and the machine crews spend a great deal of their time trying to
achieve this ideal. The main difficulty is, of course, that in the normal course
of events the operator can only check the finished reel for evenness across
the sheet so that the cause of any variation may be anywhere down the
machine and location of it often represents the hardest task.
Cross-web variations attributable directly to the presses may be divided
into three groups: those due to inaccurate cambering of the rolls, those due
to differences in the felt dewatering characteristics, and those due to other
isolated faults such as spasmodic plugging of press holes or uneven cleaning
shower pressure. Transient fluctuations in moisture content across the
web coming from the couch will be damped out in an efficient press section
but permanent differences will gradually begin to have an adverse effect on
press performance so that even an initially perfect press will eventually
be thrown out of balance. In addition, each of these individual causes of
unevenness reacts with the others and further complications arise. These
problems have received a great deal of study recently, notably by Chinn
(18), Sulatycki (20), Wahlstrom (21), Grant (24), and Delisle et al. (37)
and the following sections draw heavily on their findings.
4C. 3 1 Camber of the rolls
The subject of cambering rolls to ensure even pressure in the nip all the
way across the roll is very complicated and it is intended here only to bring
out the main points. There are three main aspects to the problem as it
affects press rolls: to decide the magnitude of the camber required, the
shape of the camber curve, and the distribution of camber between the
two rolls forming the nip.
The magnitude of camber can be calculated with a fair degree of accuracy
once the total load on the roll and certain physical characteristics of the roll
itself are known. Apart from standard dimensions and weight the important
physical constants which are needed are the Young's modulus of the roll
and the point of balance in the journal bearings. Both these values can
only be obtained approximately and the accuracy of the calculation is
further affected by the fact that the rolls are always of different com-
position, in hardness if not in material, through their body. Further, since
the camber depends on the total load on the roll it can be suited only to
one running condition; if more load is applied to the press the nip pressure
355
will be greater at the edges of the roll, if less load the pressure will be
greater in the centre.
Because of these difficulties it is usual to find the best camber by experi-
ence, although calculation can be a very useful guide if a new or different
set of rolls are put into use. Any error in the camber will show up greatest
when the roll is first put on the press; as the roll is run-in the places where
pressure is hardest gradually wear more than the other places and the
initial errors are evened out. There are one or two methods of assessing
whether the camber is correct for normal operating pressure, but they are
very approximate and with one exception can only be utilized in static
conditions; most frequently used are the embossed foil or carbon impres-
sion techniques which depend on measurement of the width of a fairly
ill-defined band so that only relatively gross errors of camber can be
determined with any accuracy. A different technique has been developed by
Sussman and Grimwood (44) which permits a dynamic check of nip
conditions to be taken; in this method full-width strips of a suitably
embossed aluminium foil are fed through the press while it is moving as
fast as is practicable, and variations in the resultant thickness of the foil
can, with calibration, be used to assess differences in nip pressure. However,
all these methods are primarily of value to check that loading of a press is
even on both sides and to avoid any excessive error in a new roll.
The best method of controlling the camber is, in the author's opinion,
to determine the amount of wear across the roll periodically through its
life and use that as a basis for fixing the camber for the next grinding.
By this means the magnitude of the camber and also the camber curve
giving the best compromise under average conditions can be found. The
success of this approach depends, however, on there being no other
persistent variations across the machine and this is usually where the
difficulty arises. If the roll has, for instance, been loaded heavier on one
side than the other or has gradually plugged in one place, then the wear
curve will give a false impression. For this reason it is important to check
the camber at intervals when the roll is still on the machine and determine
from these figures an average wear curve. In theory once this has been done
a few times it should become possible to lay down a definite camber figure,
but in practice conditions are constantly changing on the machine (in
particular, normal operating loads may be altered) and the procedure has
to be continually repeated. Rolls should always be ground shortly before
being put on the press to avoid the appearance of flats in storage.
On the machine, wear of the roll is influenced by many factors and the
better the treatment of the roll the longer it may be run without reaching
a point where grinding becomes necessary. If the drying is different on one
side of the machine compared with the other and the machine crews
habitually correct this by weighting the press roll heavier on the damper
side, then the roll will wear more and more on that side until the load has
gradually become almost even again across the roll; hence a greater and
greater load difference will be needed as time goes on until the press be-
comes practically impossible to run and one or both rolls must be removed
for grinding. Again, there may be overpressing at the edges because the
356
sheet is always run heavier there to avoid cracking and to even up the
caliper of the finished roll; over a period the same unevenness of wear and
trouble on the press will occur. Under such conditions as these a frequent
grinding cycle will be needed and because of the trouble they give on the
press a better solution to avoid the necessity of running in that way should
be sought. Softer rolls will also wear more and be more sensitive to
differences in load, so that they too will require more frequent grinding.
Apart from the actual magnitude of the camber the shape of the camber
curve is very important and presents similar difficulties when it comes to
calculation. The original analysis of this problem by S. F. Smith in the
Technical Section Proceedings of the British Paper Makers' Association
for November, 1936, remains the best source of information on this subject
and his formula for the shape of the camber curve is probably the closest
that can readily be calculated. Even then the effect of shear stresses within
the roll, temperature gradients, and horizontal deflection under the driving
stress are ignored, but nevertheless will all influence the shape of the ideal
camber to an unknown degree.
The shape of the camber normally imparted by a grinder comes from a
circular cam with a slightly eccentric centre of rotation. Rotation of the
cam is usually geared to travel with the carriage along the grinder bed and
the position of the cam giving the high point is set at the centre of the
roll; the precise shape of the curve then depends on the amount of rotation
of the cam and hence on the roll length. According to Chatwin (53)
rotation of the cam is best set at 70°, though for rubber rolls 90° is better
due to movement of the rubber after grinding. Instead of using a circular
camber, occasionally what is known as a parabolic camber is put on the
roll; however, neither shape corresponds exactly to the curve determined
by Smith and the greatest discrepancies usually occur at a distance of
between 10 per cent. and 20 per cent. of the roll length in from the edges.
It is in these regions that variations in pressure are most frequently noticed
in press rolls.
Due to these various difficulties it is common practice to adjust the
camber shape imparted with the circular cam by re-grinding selected regions
of the roll. This can be a time-consuming task which could only be elimi-
nated by the introduction of special gearing into the grinder mechanism
to permit a variety of slightly varied shapes to be available for a given roll
length instead of just one shape, so that a better compromise may be
achieved. Wear curves can be used to assist in determining the best initial
camber shape and the same procedure as for determining the magnitude of
the camber, i.e. frequent on-machine checks and the plotting of wear
curves, can be used. Greater care needs to be exercised in using wear
curves for this purpose, however, and any alterations required to the shape
of the curve should be apparent in symmetrical wear of the roll. Localized
high degrees of wear should be interpreted with care and only very small
changes in shape of the crown made at any one time.
Ideally it would be desirable to avoid having to place any camber on
press rolls at all. Apart from all the difficulties of putting on a suitable
camber, when a roll has a greater diameter in the middle then the speed of
357
the surface of the roll is higher there than at the ends. This means that some
slippage must occur between the press rolls and the paper will be slightly
stretched in the middle. Normally the total camber required is distributed
between top and bottom press rolls to give a nip as near horizontal as
possible; the precise division of the camber depends mainly on the load
applied and can be calculated. It usually amounts to around one third of
the total camber for the top roll and the remainder for the bottom. Possibly
the best conditions in the nip are not achieved when it is horizontal and a
different compromise would allow more satisfactorily for the effect on the
paper of the speed difference of the rolls, but such considerations as these
are still a matter for controversy. The introduction of such devices as the
'Accra-nip' anti-deflection roll and the 'swimming' roll, which largely avoid
the necessity for camber, should solve a great many of these problems
once they are perfected for presses; the design and operation of these
rolls is discussed briefly in connection with calender camber-compensation.
4C. 3 2 Variations in the felt
It is unlikely that faults originating in the felt itself are the cause of many
frequent cross-web variations in performance. In most cases these are
caused by other parts of the press and are transmitted to the felt. Prob-
ably the most common variation from the felt itself comes from uneven-
ness of the seam; in places where the seam takes up the greatest angle to
the cross-machine direction the felt weave is most distorted and this has
been observed to reduce the porosity and hence the dewatering properties
of the felt. If the felt seam can be kept as level as possible across the
machine most trouble from distortion of this nature should be avoided.
Any variation in pressure applied in the press nip is eventually trans-
mitted to the felt. If the pressure is higher at one point across the machine
the felt will start its life by responding to the higher pressure and taking
more water out of the sheet; in time the felt will wear faster at this position
and due to the greater quantities of water passing through may become
plugged to a greater extent than the rest of the felt. This will have the effect
of reducing the ability of the felt to carry water and gradually the greater
pressure will have less and less effect until the dewatering efficiency of the
press pretty well equalizes all the way across. For this reason any errors in
cambering of a roll or in load application should be more noticeable when
a felt is first put on the press and will gradually diminish as the felt ages.
A similar effect occurs if one position across the web is consistently
different in moisture content or substance when entering the press (this
may be accidental due to an uneven condition in the wire part, or it may
be deliberate, as for example when the sheet is run heavier in one part to
counteract a condition of over-drying which produces a dry streak at the
reel-up). If the sheet were continually wetter in one part of the web, the
felt opposite that position initially extracts more water and evens up mois-
ture content, but then gradually becomes more plugged than the remainder
of the felt due to the greater quantities of water passing through. So
the dewatering efficiency of the felt in that position gradually diminishes
and the wet streak reappears in the web to a more prominent extent after
358
the press section. Thus the felt tends to return conditions to their original
state: if the cause of the wet streak were accidental in the wire section
then it begins to give more trouble again as the felt ages; if the wet streak
were deliberate then the fault it was designed to remedy will be more
apparent with the new felt and the substance difference would need to be
much greater at first.
The moral of all this, as with loading the press unevenly to correct a
drying profile error, is that the use of the press to remedy faults elsewhere
in the machine is likely to lead to more trouble than it overcomes. This
applies especially when deliberate alterations are made to remedy faulty
drying conditions and such measures should never be other than temporary
until the trouble can be found and corrected.
4C. 3 3 Other sources of cross-web variation

There are several other things which can affect the cross-web moisture
profile through a press section and most of them are likely to occur in
random positions anywhere across the roll. The most common originate
in faulty felt cleaning equipment and, in the case of suction rolls, in
sealing spray pressure differences across the roll.
Poor cleaning in one or more positions across the felt will obviously
allow the felt to become dirtier and more plugged with material so that as
it ages a wet streak will become more noticeable. If an uneven roll shower
causes more deposits to build up in the holes in one position, then removal
of water from the sheet eventually must become inhibited in that region and
a similar result occurs over a longer period. In a similar way a fairly perma-
nent difference in substance or moisture originating in the wire section may
cause the holes opposite to become more made up purely from carrying
the greater quantities of water removed in that position; hence over a
long period the press would cease to cloak the fault and the uneven condi-
tion of the sheet at the couch would gradually become more noticeable.
Some of these differences in moisture content may become so accentuated
as to cause breaks at the press because of the difference in adhesion to
the top roll, or to cause creasing in the drying section.
If a small patch of holes in the press roll becomes made up, or if the roll
itself becomes damaged at one spot due perhaps to something passing
through the nip, then the problem is mainly one of reducing the resulting
blemish to manageable proportions. If a patch of the sheet is damper
entering the drying section than the rest it will shrink slower as the sheet
dries; this will cause the patch to become stretched and visible in the
finished sheet due to a variation in optical refraction properties. Larger
patches will also exhibit cockling. If the patch is markedly damper at the
calenders it can become blackened and distorted into wrinkles and possibly
stamped out completely.
4C . 3 4 General considerations
Enough has been said in the previous sections to illustrate the complexity
of the problem of controlling cross-web moisture variations at the presses.
359
One or two people, notably Chinn (18), consider that frequent and com-
prehensive testing of the press section involving sampling across the web
for moisture tests can help considerably in keeping variations to the
minimum. Chinn cites several examples of the effects of various alterations
to the press section and mentions one case, for example, where analysis of a
press test eventually led to the discovery of a seized dry felt stretch roll!
Unfortunately, analysis of the results of a press test is by no means
simple and the interactions between rolls, felt and paper complicate
deductions enormously. Certain variations, such as those due to poor
cambering, are relatively long-term in their effect and may be discovered
from careful comparison of press test data; but other faults some of which
show up more at the beginning of a felt life, some towards the end, may
completely resist detection. The author considers that, apart from periodic
sampling and testing to keep check on the general press performance as a
whole and provide suitable standards, comprehensive press tests involving
taking many samples from across the web are best kept as an aid to
trouble-shooting to help unearth the cause of some disturbing unevenness
in profile.
In general, cross-web variations will be kept to a minimum if three ideals
are aimed at: keep the substance profile at the reel-up as level as possible
(as discussed in Part 1); minimize camber discrepancies by careful analysis
of wear curves taken during the life of and before regrinding a roll; and
keep the felt seam straight. If despite this variations occur at the reel-up, as
they inevitably will, then a detailed press test involving the various tech-
niques that have been discussed may pin down the trouble. But the cross-
machine substance must be evened up first and so long as the wire section
is reasonably in order variations in press performance due to differences in
substance and ingoing moisture will then be minimized and the moisture
content figures should be sufficiently representative for the purpose of
providing comparison across the web.
4C.4 PRACTICAL POINTS

4C. 4 1 Start-up
When starting-up the press section the aim should be to get the general
conditions of both press and felt as close as possible to normal before the
sheet is actually fed through. This implies that the press as a whole should
be well warmed up and approaching the equilibrium temperature attained
when running, while the felt moisture content should also be close to its
eventual value with the sheet up.
The press section is very often one of the last to be started on the
machine. After a check-over to remove any oil or grease and see that the
felt is clean and undamaged both inside and outside, the top roll is lowered
slightly from the raised position it was left in when the machine was shut
and the felt tightened sufficiently to allow it to be driven slowly round.
Water (preferably warm) is sprayed onto the felt and, especially if this is
done by hose, it is important to maintain an even wetness over the whole
360
width otherwise the felt may crease; pools of water should definitely be
avoided as slack areas may develop later. As the felt shrinks the stretch roll
is steadily slackened back to keep the felt tension just sufficient to drive the
felt rolls. In some of the newer types of felt, such as the needled or batt-on-
base type, the felt may behave rather differently when being wetted out and
the shrinkage commonly encountered with woollen felts may not occur.
Shortly before the web is ready for feeding up from the couch the top roll
is lowered to near normal running load, the felt is tightened to the usual
tension, and the sprays are shut off. By this means each press is brought up
to near normal running conditions and the felt is not likely to be so wet
that the sheet may tend to adhere to it. The tail is either passed to the first
press by hand (perhaps with the aid of a small felt-covered wooden roller
to lift the tail off the wire), or it is blown over from the couch onto the
first press felt; in the case of a lick-up or suction pick-up, the forward drive
roll is raised or the pick-up roll lowered so that the felt touches the paper
lightly and carries either a tail or the full sheet over to the first press. The
passage of a narrow tail 4 in. to 6 in. wide through one end of the press
throws a greater pressure on to it than on the full width sheet so if the tail
crushes it may be necessary briefly to relieve the load on the front side. As
soon as the tail is running satisfactorily up the top roll, whenever practicable
it is preferable before feeding through to the next press or the drying section
to widen the sheet using the movable jet cutter on the wire and to even up
the press load. If this is done, the vacuum on a suction press and on a felt
suction box will come up to the usual level and conditions generally in the
press will approach equilibrium quicker so that the paper is less liable to
give trouble.
Transfer off a top roll on a slow machine is effected by hand, sometimes
with the aid of a lump of wet broke. On faster machines a pair of com-
pressed air jets shaped like the horns of a cow and directed inwards towards
each other will draw a tail off the roll while another suitably placed air jet
will help to carry the tail over to the next press; these jets require careful
setting and should be checked periodically because if they get knocked out
of line a great deal of frustration and delay can be caused. For passage
through the drying section a slightly wider tail may be cut and this is
thrown or blown between the carrying ropes when these are used. A narrow
strip of paper often will not stand up to the normal running tension and it
may be necessary to slacken back the draws when feeding up the tail to
prevent the sheet snapping.
At the earliest opportunity the draws and other press conditions should
be checked and reset. If there is difficulty getting the tail off the press roll
some dilute vitriol may be dropped onto the roll to reduce adhesion but
this is essentially a stop-gap. The most effective way of preventing sticking
at the presses is to ensure that start-up conditions in respect of composition
of the paper, temperature, and ingoing moisture content are as close as
possible to those pertaining during normal running. If the sheet tends to
follow the felt it is probably caused by the felt being too dirty and plugged;
if the felt is clean a lighter load may temporarily help to reduce this
tendency.
361
4C. 4 2 Shut-down
When stopping the press section, even for a short time, the top roll pressure
should be relieved slightly and the felt examined for plugging and run
round with sprays on for cleaning. An acid washing compound may be
used on the felt and the addition of small quantities of a neutral detergent
can help to remove pitch, rosin and wax deposits. When cleaning equip-
ment rubs on the felt, as with a Vickery conditioner, the shoes should be
cleaned of pitch, felt hairs, etc.; it is good practice to do this anyway at
least once a shift when the machine is running.
For a complete shut the top roll is lifted completely off to avoid putting
a flat on the rolls and the felt tension is relieved. The doctor blades should
be cleaned and checked for passing stuff and changed if necessary; the
doctor should preferably be left raised off the top roll, particularly if it is
heavy. Other parts of the press, such as the suction boxes on the felt, will
require cleaning and should be checked for wear.
If the felt seam is uneven it can be squared up, but this is unlikely to
prevent it going out of true once the press is started again, particularly if
a different nip pressure is used. In the case of a suction press, the holes
should be examined and any tendency to make up especially at the ends of
the roll should be noted.
4C. 4 3 Changing and running wet felts

There are as many ways of changing a wet felt as there are machines.
Generally the old felt is cut across after the press rolls and collected behind
the top roll doctor or in the pit as the rolls are crawled round. While the
new felt is prepared the frame and rolls should be well cleaned down and
the stretch roll slackened right back to give a maximum length of felt to
play with. The top roll is lifted and locked in position while the bottom
roll is cantilevered or slung up allowing the front end support to be
removed so that the felt can be slipped over the bottom roll.
Clean paper should be laid on the machine house floor before un-
wrapping the felt, and the direction of the nap should be checked to see
that the arrow on the felt shows the correct direction of running. With
two-sided felts a check should also be made that the right side will face the
paper. It is useful to arrange for the seam to be near the press roll when the
felt is being put over the roll to make it easier to check that everything
is right at least after this most difficult part of the operation is complete.
The felt is handled easier if the two laps are well separated and plenty of
slack is available at each roll.
Finally the seam is straightened out, the stretch roll tightened slightly
and the top roll lightly lowered. After a thorough inspection to ensure there
are no wrinkles, creases, or dirty patches, the felt is wet out in the manner
described in 4C. 4 1. In some mills the felt is kept very slack and two
members of the crew pull the felt out as it passes into the press nip; this
may be advisable when the felt is first wetted but it should not need to be
continued for very long if sufficient care is taken to wet the felt evenly. A
wetting detergent may be used on a new felt to assist wetting out.
362
Occasionally the felt will be too wide even when it is thoroughly wetted
and even tightening the stretch roll as hard as possible will not bring the
width down to a manageable level. In this case, if experience of the parti-
cular make of felt indicates that it is likely to shrink appreciably soon after
the machine is started, then as a temporary expedient the stretch roll can
be slackened back at one side causing the seam to run at an angle to the
cross direction and the felt width to reduce. This procedure affects the
drainage properties of the felt which will plug much faster if not corrected
reasonably quickly. As a final expedient, of course, the felt can be trimmed
at the edges.
Once the machine is running it is very important that the seam should be
kept as straight as possible. If the seam leads on one side of the machine
the indication is that the felt run is shorter there and the stretch roll
should be tightened at that side to equalize the length of run. The most
common occurrence is for the seam to lead in the middle; this can usually
be attributed to the fact that camber on the rolls causes the felt to travel
slightly faster in the middle, though when tension is high and felt rolls
deflect inwards this also causes the seam to lead in the middle due to the
felt having a slightly shorter run in that position. Normally this tendency
will be taken care of by having one or two felt rolls with a slightly hollow
camber or having a worm roll with the worm only on the edges of the roll;
the surface of these rolls then travels at a slightly faster speed at the ends
compared to the middle which counteracts the opposite motion caused by
the press rolls. Expanding 'spreader' rolls, usually of the fixed-bow type,
have also proved helpful for keeping the felt stretched wide and the seam
straighter. For further discussion of this topic reference can be made to an
article by Woodside and MacMillan (52).
If the camber on the press rolls is higher than usual this will be shown by
the felt seam leading slightly more than usual in the middle. When a new
roll shows this tendency it may be sufficient as a temporary correction to
slacken the felt if it is more tight than usual. Alternatively, on some
machines it is possible to put on a string worm at the edges of a convenient
felt roll and provided this is done by an experienced person it should give
a satisfactory temporary remedy (the string should never become a
permanent feature, as often happens). In some mills lumps of wet broke
are thrown onto the felt rolls at the edges for the same purpose; this has
recently been roundly condemned because it has been shown to cause
filling up of the felt and press rolls, leading eventually to reduced de-
watering. If the camber on the press rolls is less than usual the opposite
tendency will occur, the seam will lag behind in the middle and one of the
felt rolls will need building up slightly in the middle. A seam which is
uneven but symmetrical across the machine indicates that the camber
shape is likely to be incorrect with the camber too great in regions where
the seam leads. Associated with wear tests on the press rolls this can give
a useful clue to grinding faults.
Felt guides normally require little attention. If the stretch roll is tightened
on one side to correct the seam then the felt will tend to run over to the
opposite side, which is then slacker. For this reason, whenever the $tn~tch
363
roll is altered in this way the felt guide should be watched and if necessary
the hand guide roll should be adjusted to even up the felt.
4C . 4 4 Checking the press during running

Periodically during each shift the machineman will check over the presses.
This means that he examines all instrument and recorder readings and when
required completes a log. He will have an idea of the normal degree of
variation affecting each reading and his interpretations of the instruments
will take account of this. Any sign that an instrument reading is out of the
normal expected running limits will occasion an inquiry to find the reason;
many aspects of this were dealt with in 4C. 1 to illustrate the use of different
measurements in trouble-shooting.
Apart from this, examination of the press section depends a great deal
on the individual design. Attention to the felt has already been mentioned.
The press doctor should be checked and if fibre normally builds up slowly
on the blade it should show an even rate of movement all the way across
the machine. If the build-up is excessive the tray will require frequent
cleaning out and the acidity of the stock may require adjusting within the
limits normally accepted.
It is particularly important to check that the draws are satisfactory; if
the sheet has tightened or slackened, and the mechanics of the press are
satisfactory (no belts slipping, etc.), this can often give a useful indication,
when interpreted with the position of the dry line on the wire, of changes in
substance or freeness. On a plain press the flow of water down the back of
the press roll should issue from the nip fairly evenly all the way across the
roll, although it is normal for the flow lower down the roll to break into a
series of regular streaks. Likewise on a suction press roll sometimes the
throw-out from the holes can be seen and this can be a useful indication of
the state of the press though it is never possible to see very far along the
roll to compare the throw-out all the way across.
Most breaks at the press originate in the wire section. Poor formation,
lumps, slime, poor edges, air bells, dandy picking, and numerous other
defects can cause trouble at the press where the weakness is shown up
either by crushing in the nip or in excessive adhesion to the top roll. Poor
substance control, with the sheet coming spasmodically damper to the
press all across or in streaks, can cause the same sort of trouble. Likewise,
variation in freeness or vacuum on the suction boxes and couch can show
up in the same way.
The causes of breaks attributable directly to the press section include the
following: unevenness in the draw caused by variation in shrinkage in the
sheet itself or by mechanical slippage in the draw control causing the sheet
tension to become excessive; felt or suction roll becoming badly plugged so
that rubbing and crushing occurs locally or the sheet adheres too strongly
to the top roll; felt badly guided and worn at edges or in body so that
dewatering is very uneven in places in the nip: press doctor worn unevenly
or pitch building up under the blade causing fibre to pass round roll; and
air between the paper and felt causing blow-creasing.
364
In addition probably a never-ending list of minor ailments can be
thought of, including: water and lumps passing down the edge of the
· doctor onto the roll and being splashed into the sheet; loose threads
from the edges of the felt getting entangled into the paper; condensation
water gathering on the back of the doctor and dropping on the sheet; and,
in a comparatively dirty felt, too much water carried round from a Vickery
conditioner causing excessive crushing and a long ragged hole in the sheet.
All these faults are avoided only by a careful supervision and anticipation.
365
REFERENCES
(1) Dixon, W. B.: 'The Operation of a Stacked Press with Vacuum Transfer on an
English Newsprint Machine.' P. and P. Mag. Can. 53, 11, 115. October, 1952.
(2) Hanson, K. F.: 'The Press Felt and its Water Content.' Svensk Pap. 56, 16, 525.
1953. (In Swedish.)
(3) Anon: 'Application of Infra Red Irradiation to the Pre-drying of Paper.' Papeterie
76, 9, 559. 1954. (In French.)
(4) Holden, H.: 'Some observations on suction couch rolls, suction press rolls and
vacuum transfer.' Tech. Bull. Tech Sec. Brit. Paper & Board Makers' Assoc.
31, No. 1 :7-12; No. 2:41-6. February, April, 1954.
(5) Nissan, A. H.: 'Functions of the Felt in Water Removed on the Papermaking
Machine.' TAPPI, 37, 12, 597. December, 1954.
(6) Molsberry, M. V.: 'Modernising a Press System with an Air Bleed Press at
Wisconsin Rapids.' Paper Trade Journal 140, 8, 42. February 19, 1956.
(7) Osborn, R. V. and Wicker, D. B.: 'Beta Ray Measurements of Moisture in Run-
ning Felts.' TAPPI 39, 7, 480. July, 1956.
(8) Peters, J. H.: 'Effect of Felt Design on Paper Machine Operation and Felt Life.'
TAPP! 39, 8, 575. August, 1956.
(9) Fahlin, B.: 'Adhesion of Paper to Press Roll Covers.' Svensk Pap. 60, 13, 484.
1957. (In Swedish.)
(10) Jordansson, L.: 'The Press Section with Vacuum Transfer.' !.-General Operating
Conditions and Water Removal. Svensk Pap. 60, 5, 59. 1957. II-Suction Roll
Operation. Svensk Pap. 60, 12, 447. 1957. (In Swedish.)
O 1) Dixon, P.H.: 'A Comparison of the Operation of Two Wet Felts.' Also Discus-
sion. Proc. Tech. Sec. 39, 2, 173. June, 1958.
(12) Engel: 'Improved Water Removal from the Paper by Resistance Heating.'
Woch. filr Pap. 86, 22, 953. 1958. (In German.) Summary in Tech. Bull. Tech. Sec.
Brit. Paper & Board Makers' Assoc. 36, 4/5, 130. August/October, 1959.
(13) Gavelin, G.: 'Thoughts on Pressing.' Svensk Pap. 6, 11, 345. June 15, 1958.
(14) Grant, J.: 'Summary of Literature on Papermakers' Felts.' Part I-Wet Felts.
Also Discussion. Proc. Tech. Sec. 39, 2, 151. June, 1958.
(15) Pollard, D. F.: 'An Operational Report on Sheet Pressing at East Millinocket.'
P. and P. Mag. Can. 59, 7, 133. July, 1958.
(16) Anon: 'A New Process for Better Dewatering at the Wet End.' Paper Maker 137,
3, 65. March, 1959. (Translation from Allg. Pap. Rund, November 5, 1957,
p. 1077.)
(17) Pallavicini, M. R.: 'How Infra-red Heating is Used to Increase Production at
A.P.M.' Paper Trade Journal 142, 26, 36. June 29, 1959.
(18) Chinn, G. P.: 'The Practical Application of Press Testing for Improved Paper
Machine Performance.' P. and P. Mag. Can. 61, 5, T-287. May, 1960.
(19) Ford, T.: 'Needled Felts and particularly "Needle Reinforced Felts".' Paper
Maker. Annual Review Number, 1960.
(20) Sulatycki, R. W.: 'A Few Thoughts on Paper Machine Pressing.' P. and P. Mag.
Can. 61, 3, T-109. Convention Issue, 1960.
(21) Wahlstrom, B.: 'A Long Term Study of Water Removal and Moisture Distribu-
tion on a Newsprint Machine Press Section.' I.-P. and P. Mag. Can. 61, 8,
T-379. August, 1960; II-61, 9, T-418. September, 1960.
(22) White, R. and Berdux, F.: 'Static Pressing of Wet Sheets and Felts.' TAPPI 43,
6, 580. June, 1960.
(23) Anon: 'Sulzer Vacuum Installation Proves Highly Successful on Paper Machine.'
Paper Trade Journal 145, 6, 41. February 6, 1961.
(24) Grant, D. M.: 'A Preliminary Report on Felt Air Porosity Measurements.'
P. and P. Mag. Can. 62, 5, T-289. May, 1961.
366
THE PRESS SECTION
(25) Howe, B. I.: 'Towards a better understanding of the Press Section of the Paper
Machine.' Paper Technology 2, 4, T-156. August, 1961.
(26) Sweet, J. S.: 'A Basic Study of Water Removal at the Press.' P. and P. Mag. Can.
62, 7, T-367. July, 1961.
(27) Wicker, D. B.: 'Press Felts and Water Removal.' Paper Trade Journal. Part I:
145, 42, 40, October 16; Part II: 145, 43, 33, October 23, 1961.
(28) Anon: 'Newsprint Men Discuss Felts at Conference.' P. and P. Mag. Can. 63,
2, 46. February, 1962.
(29) Gavelin, G.: 'New Approach to Paper Machine Press Sections.' Paper Trade
Journal 146, 15, 38. April 9, 1962.
(30) Howe, B. J. and Cosgrove, C.: 'Factors affecting Pressing Efficiency on the
High-Speed Suction Press.' P. and P. Mag. Can. 63, 11, T-515. November,
1962.
(31) Radvan, B. and Karpati, D.: 'Adhesion of Wet Paper to Solid Surfaces.' Paper
(32) Redfern, A. P. and Gavelin, G.: 'The Causes of Shadowmarking in Paper.'
Paper Technology 3, 5, T-161. October, 1962.
(33) Brandberg, M., Brauns, 0., and Swanberg, 0.: 'The Wire-Felt Press.' Svensk Pap.
66, 24, 1031. December 31, 1963.
(34) Dyer, H. : 'Plastic Fabrics iron out many Wet Pressing Problems for Mead.'
Paper Trade Journal 147, 3, 26. January 21, 1963.
(35) Hamilton, H. D. and Wrist, P. E.: 'The S.P.F.I.-Mead Fabric Press: A Novel
Approach to Wet Pressing.' P. and P. Mag. Can. 64, 5, T-219. May, 1963.
(36) Wicker, D. B.: 'Progress Report on the Use of the Felt Porosity Tester.' Paper
Trade Journal 147, 49, 28. December 9, 1963.
(37) Delisle, J. C., Garofalo, J. H., and Cooper, S. R.: 'An Investigation into Wet-
Press Operation. Part I: Improving Press Operation. Part II: The Effect of
Continuous Felt Cleaning on the Papermaking Operation.' P. and P. Mag. Can.
65, 8, T-315 and T-331. August, 1964.
(38) Howe, B. J. and Cosgrove, J.C.: 'A Unified Concept of the Suction Press.' P. and
P. Mag. Can. 65, 10, T-424. October, 1964.
(39) Justus, E. J. and Cronin, D.: 'The Vented-Nip Press.' TAPP! 47, 8, 493. August,
1964.
(40) Lyall, J. D.: 'Observed Crushing in Press Roll Nips.' TAPPI 47, 2, 119. February,
1964.
(41) McNamara, S. J.: 'Water Removal Capacity of Running Press Felts.' P. and P.
Mag. Can. 65, 12, T-529. December, 1964.
(42) Perry, F. G., Hervey, L. R. B. and Emslie, A. G.: 'Water Removal from Wet
Webs by a Porous Nylon Press Roll.' TAPP! 47, 11, 683. November, 1964.
(43) Robinson, W. F. E.: 'The Double Divided Press.' P. and P. Mag. Can. 65, 12,
T-550. December, 1964.
(44) Sussman, P. G. and Grimwood, K.: 'Nip Pressure and Gap Tests under Dynamic
Conditions using Commercially Embossed Aluminium Foil.' Paper Technology
5, 3, T-59. June, 1964.
(45) Swanberg, 0.: 'New Studies point up Potentials of Double Felted Presses.' Paper
Trade Journal 140, 44, 35. November 2, 1964.
(46) Tenfiilt, E. and Wahlstrom, B.: 'First Report on Scandinavian Experience with
Fabric Press.' Paper Trade Journal 148, 8, 38. February 24, 1964.
(47) Wrist, P. E.: 'The Present State of our Knowledge of the Fundamentals of Wet
Pressing.' P. and P. Mag. Can. 65, 7, T-284. July, 1964.
(48) Eiderbrant, A. and Swanberg, 0.: 'New Fabric Press Development greatly
simplifies Operation.' Paper Trade Journal 149, 26, 32. June 28, 1965.
(49) Haller, H. C. and Curtis, D.: 'Shrinkable Sleeve improves Press Section Water
Removal.' P. and P. Mag. America 39, 12, 23. March 22, 1965.
(50) Halas, K. K.: 'Finnish Mill Experience with the Divided Press.' Paper Trade
Journal 149, 16, 42. April 19, 1965.
(51) McNamee, J.P.: 'A Study of Rubber Covered Press Roll Nip Dynamics.' Part I
TAPPI 48, 12, 673. December, 1965.
(52) Woodside, L. M. and MacMillan, H. J.: 'Guiding Felts-which way the right
way?' P. and P. Mag. America 39, 50, 35. December 13, 1965.
367
THE PRESS SECTION
(53) Chatwin, M. A.: 'Reel Moisture Improvement by Crown Curve Selection.'

(54) Eiderbrant, A. and Andersson, W.: 'The Fabric Press and Paper Quality.' Svensk
Pap. 69, 24, 847. December 31, 1966.
(55) Klass, C. P.: 'Garden State gets 47 % dryness from new Swedish Press Section.'
Paper Trade Journal 150, 52, 30. December 26, 1966.
(56) McNarnara, S. J. and Baxter, M. K.: 'Water Removal with the High-Intensity
Nip Press.' Paper Trade Journal 150, 8, 70. February 21, 1966.
(57) Nordman, L. et al.: 'Relationships between Mechanical and Optical Properties
of Paper affected by Web Consolidation.' Consolidation of the Paper Web.
Transactions of the September 1965 Cambridge Symposium, p. 909. B.P. &
B.M.A. Tech. Sec. publication, 1966.
(58) Pye, I. T. et al.: 'Structural Changes in Paper on Pressing and Drying.' Consolida-
tion of the Paper Web. Transactions of the September 1965 Cambridge Sym-
posium, p. 353. B.P. & B.M.A. Tech. Sec. publication, 1966.
368
PART 5
THE DRYING SECTION

AND CALENDERS
INTRODUCTION
51 In this Part of the book the final stages in the operation of the Four-
drinier paper machine are considered. The drying section has the single
object of removing from the web of paper water which is left after pressing.
The use of steam drying cylinders is a relatively cumbersome method of
achieving this end in comparison with pressing, and in terms of steam and
power usage, overall maintenance, and floor space, conventional dryers are
by far the most expensive part of the process. For this reason a good
understanding and regular checking of the drying section are of paramount
importance, and it is such aspects that will form the main concern of what
follows.
Despite the cost of drying paper by this traditional method, it remains
one which is handsomely economic in comparison with other drying
techniques. Fluid-bed drying, straight-pass dryers using high-velocity air
nozzles, flash drying as used with pulp, dielectric or high-frequency drying
in a strong electro-magnetic field, ultrasonic drying, and several other
comparatively new ideas involving basically different concepts of drying
are attracting increasing attention as possible successors to the common
banks of drying cylinders, but for the present there is little indication of a
real break-through. These particular techniques and their potentialities
will not be considered, (the reader interested in further details may consult
articles by Burgess (57) and Luckins (94)). Certain other techniques, in
particular high-velocity-air hoods (otherwise known as air caps, forced-
convection hoods, and accelerator hoods, but these terms are less expres-
sive of the essential difference between this type of hood and the ordinary
type with an extractor fan), infra-red drying, and drying under vacuum,
may be used to augment and speed-up the normal process and also, in the
case of the first two, have some potential value as methods of regulating the
moisture profile across the sheet. These will receive due attention in the
appropriate section.
It is commonplace to incorporate in the drying section various other
pieces of equipment. Certain of these may be regarded as essential to
achieving the desired finish of the paper on machines in which they are
installed, and include the M.G. or Yankee cylinder, smoothing press,
breaker stack, and sweat roll. Each of these items, particularly the first,
will be considered in detail. Other pieces of equipment found in the dryers
for a specific purpose are, strictly speaking, not an integral part of the
Fourdrinier machine and will not be dealt with. These include the size
press, coating press, and various other devices such as the double-roll unit
used in the drying section to manufacture extensible paper. Similarly the
use of water doctors on the calenders for the purpose of applying colours,
starch solutions, wax emulsions, and so forth, for improving surface
smoothness and other characteristics of the paper such as oil and scuffing
resistance will not be mentioned.
371
51.1 THE DRYING SECTION AND CALENDERS
51. 1 Moisture in the paper

The moisture content of paper entering the drying section depends very
largely on the efficiency of the presses; the economic necessity of having
these functioning as close as possible to the point where disruption of the
sheet becomes likely has already been stressed. Depending on this efficiency,
and on the type of paper being produced, the moisture content is frequently
as low as 60 per cent. (l ·5 water/fibre ratio) but may be as high as 70 per
cent. (2·3) or even 75 per cent. (3·0). At this point it is worth noting that
for representing moisture content throughout the drying section the 'wet'
basis, i.e. water as a percentage of total weight, is in common use. This is
certainly satisfactory for production calculations and as a description of
the state of the paper towards the end of the dryers and at the reel-up, but
for fundamental work it is frequently better to use the 'dry' absolute basis
in which water is taken as a percentage of dry fibre. This latter form of
representation permits a better indication of the state and movement of
water within the paper during drying. But moisture content percentage
values on the 'dry' basis can be a little confusing in that they are able to
exceed one hundred; they are, in fact, exactly the same as water/fibre
ratios expressed as a percentage. So unless specifically mentioned, all
moisture figures will be expressed on the 'wet' basis. Subtraction from 100
then gives another form of representation occasionally used, the dryness or
solids content.
At the reel-up the paper may be dried down to as little as 2 or 3 per cent.
moisture content though it is more usual, especially on faster machines
producing newsprint and kraft grades, to achieve much higher values of
around 5 to 7 per cent. When allowed to humidify under standard atmo-
spheric conditions the equilibrium moisture content is invariably higher
than that at which the paper is produced, varying from 5 per cent., for
some grades of thin tissue up to as high as 10 per cent., and with common
values in the 8 or 9 per cent. region. This difference between moisture con-
tent at which the paper is produced and at which it stabilizes in air is import-
ant from many angles. When sold in reel form on a tonnage basis, the closer
the paper is to its equilibrium value the cheaper it is to produce because
for every 1 per cent. of moisture lacking in the paper, 1 per cent. of fibre
has to be substituted. When required in sheet form the costly separate
conditioning process will be necessary for the same reason and also because
it reduces disagreeable manifestations of moisture pick-up in the form of
expansion and curl of the sheets. Apart from these points the moisture
content at which the paper is reeled alters the effect of the calenders, the
presence of static, and the apparent strength and other test results on the
paper; also, according to the extent by which moisture is picked up so
certain qualities, smoothness in particular, can alter appreciably. The
difficulties of getting sufficient moisture into the paper at the reel-up are
well appreciated yet the consequences of not achieving a reasonable and
consistent level are seldom fully realised; this particular topic and the
difficulties attendant on its solution will, because of its tremendous
importance, be subjected to a particularly close analysis.
372
CHAPTER 5A
5A.l REMOVAL OF WATER IN DRYING
The manner in which water is removed from the paper web during drying
has been the subject of much interest and speculation over a great many
years. This applies especially to the precise function of the dryer felts, a
subject that has created a great deal of dispute which on occasions has
(somewhat appropriately) become very heated. Although several details
remain unsettled it is now possible to give a fairly comprehensive descrip-
tion of the process of drying, at least in a qualitative fashion. The modern
theory to be described is the outcome of work by many distinguished
investigators, though a really detailed understanding of how water is
removed from the web may be said to date largely from the fundamental
work reported in a paper by Dreshfield and Han (22).
5A. 1 1 Evaporation from the paper surface
When any surface is damp, a continuous interchange of water particles
takes place between the surface and the air; as a preliminary to more
detailed discussion of the drying of paper it is important first to clarify the
mechanism involved in this. The rate of evaporation of water from a damp
surface depends basically on the difference between the molecular vapour
pressure of the water in liquid form and the pressure of water vapour in the
air in the immediate vicinity of the surface. The vapour pressure of liquid
water at a free surface is produced by the continual escape of molecules
from the surface and is greater the higher the temperature due to the
increased speed of movement of the molecules. The pressure of water
vapour in air is the partial pressure any vapour exerts in a gas and is
dependent primarily on the absolute quantity of water held in vapour form
in the air. Evaporation occurs so long as the molecular pressure of the
water in liquid form exceeds that of the vapour pressure of air in the
vicinity of the liquid, and ceases only when this air is saturated.
A unit volume of air at a particular temperature can only hold a certain
quantity of water vapour, at which point it is saturated and any excess of
vapour will immediately condense; the vapour pressure at this point is
termed the saturation pressure. Below saturation point the vapour pressure
is almost linearly dependent on the quantity of water vapour held in the air
and decreases steadily to zero in completely dry air. In other words at a
given temperature the vapour pressure of air below saturation point is
virtually the same percentage of the vapour pressure at the saturation
point as the relative humidity; air at 50 per cent. relative humidity has half
the vapour pressure of air at 100 per cent. humidity at the same temperature.
The effect of increasing the temperature of air is to increase the quantity
of water vapour it may hold before becoming saturated. If air is cooled the
13 373
5A.11 THE DRYING SECTION AND CALENDERS
opposite effect occurs and a point is reached, the dew point, where the
quantity of water vapour actually present in the air becomes sufficient to
make the air saturated at the lower temperature and further cooling brings
about condensation. This principle is sometimes applied to reducing the
humidity of air: incoming air is cooled well below dew point, water
vapour condenses out, the air is then re-heated and, as it contains a lower
quantity of vapour, will then be drier than formerly.
Thus, even though the actual quantity of water in a given volume of air
is unchanged, increasing the temperature of the air reduces the relative
humidity and increases the capacity of the air to pick up and hold water.
Also the vapour pressure of the water in liquid form at a free surface rises
with increased temperature, so the evaporation rate increases. These are
the reasons for heating air blown onto the paper in Grewin systems and
hoods: to increase the capacity of the air to retain water, and (by
transferring additional heat to the paper) to increase the rate of evapor-
ation.
It may have been noted that stress has been placed on the rate of evapora-
tion being dependent on the state of the air in the immediate vicinity of the
surface. This is highly important because even a fast moving body carries a
thin layer of stagnant air next to its surface and natural evaporation only
occurs so long as this layer is not saturated. Once the vapour pressure of
the air in this layer has built up to saturation point no further evaporation
occurs. For this reason it is vitally necessary to keep air moving over the
surface at an adequate rate for evaporation to be continuous. Generally
speaking, the faster the rate of air flow the lower is the relative humidity of
the air in the layer adjacent to the surface and the greater the rate of
evaporation, although a limit is reached when further increase in air flow
produces little effect because it makes only a negligible difference to the
relative humidity of the layer of air next to the surface. This principle is
used to good effect in high-velocity-air hoods.
To summarize, evaporation from an exposed, damp web of paper or felt
is improved the higher the temperature of the surface, the lower the
humidity of the air in the immediate vicinity of the surface, and the
greater the flow of air across the surface of the paper or felt. One further
point which will be considered in more detail in the next main section is
that evaporation also involves a transfer of heat energy. To vaporize water
requires a quantity of heat equivalent to the latent heat of vaporization at
the particular temperature involved. This heat must either be provided
from an external source (the heat of the drying cylinder or the surrounding
air) or from the paper and felt itself, and in the latter event cooling of the
surface would occur and this will then reduce the rate of evaporation. In
practice under any given conditions a state of dynamic equilibrium is
reached in which heat required for evaporation at the surface is balanced
by the inflow of heat and the temperatures stabilize; the evaporation rate
achieved is thus the result of a number of fairly complex interactions.
Figures 5. 1 and 5. 2 illustrate the effects described above and refer to
some experimental work on evaporation from a damp felt reported by
Preston-Thomas and Dauphinee (29).
374
THE DRYING SECTION AND CALENDERS 5A.l 1
0 lo 2o ~o 4o So IOo 7o Bo 9o
DEW POINT. oc.
Fig. 5. I. Effect on evaporation rate, from surface of a damp felt at different tempera-
tures, of air temperature and dew point (after Preston-Thomas and Dauphinee)
._: ~~
w ::i
4-o u. r
~
o<
l::l
CL z It
(~
s
I.II ~~
~-() 0 -< ~
ZZ"c. 'Bo ° C
Z·o ss"c. 9o"c.
Zo"c. 7o 0 c
57°'-. 7o"c.
o 2So Soo 7So 1,ooo l,'25o l,Soo l,75o

AlR VGLOC\TY. lN i:e&T/N'.1N\JT!l;.
Fig. 5. 2. Effect on evaporation rate of air velocity for two air and felt temperatures
and two dew points (after Preston-Thomas and Dauphinee)
375
SA.12 THE DR YING SECTION AND CALENDERS
SA. 1 2 Drying on unfelted cylinders

It is proposed first to consider the process of drying on unfelted cylinders.
Although few Fourdrinier machines in fact operate entirely without dry
felts it is quite common for at least one or two of the early dryers to be left
unfelted, so it is appropriate to deal with this relatively simple condition by
way of introduction to the subject.
From the moment when the damp web contacts the first heated cylinder
(this being often a smaller diameter pony cylinder) a process of heating up
begins. Heat is transferred by conduction from the cylinder surface to the
contacting surface of the paper web and thence, also by conduction in the
initial stage, into the body of the web. In this way a gradient is established
between the hotter contacting surface and cooler exposed surface, and the
whole sheet rapidly rises in temperature from the region of 70 to 95 deg. F.
(22 to 35 deg. C.) up to 140 to 190 deg. F. (60 to 90 deg. C.). As the exposed
surface increases in temperature there will be a corresponding increase in
natural evaporation to an extent dependent on the condition of the air in
the proximity of the sheet, as discussed in the previous section. As evapora-
tion occurs a gradual capillary movement of water within the web will
maintain the exposed surface wet.
Even if the temperature of the sheet surface in contact with the cylinder
is still below vaporization point when the wrap is completed, in the free
draw to the next cylinder there will be rapid evaporation from that surface
because it is at a relatively high temperature. This will be accompanied by a
drop in temperature of the web as heat is extracted from it to provide
latent heat for the evaporated water. On making contact with the next
cylinder the new outer surface will still be slightly hotter, but as the web
heats up again a reversed temperature gradient, again from the hotter
contacting surface to the cooler exposed surface, will be created.
At some position, which is usually very early in the drying section, the
temperature of the web is raised to the point where vaporization com-
mences. This will occur at the hottest place within the thickness of the
paper, the region in immediate contact with the surface of one of the
cylinders. Vapour created in this manner will be under a certain pressure
and will pass through the thickness of the web from the contacting surface
towards the exposed surface. This takes the vapour into cooler regions of
the web, causing a certain amount of condensation and release of heat into
the web, while the remaining vapour that is formed at the cylinder surface
finally escapes from the outer surface to the air. The heat released in the
web by condensation of some of the vapour will augment conduction of
heat from the cylinder surface, thereby helping to raise the temperature in
layers nearer to the outer surface and thus increase the rate of natural
evaporation.
If the web were in contact with the cylinder for a sufficiently long time, as
for example with an M.G. cylinder, an equilibrium state is finally reached
where the rate of vaporization at the cylinder surface balances the propor-
tion of heat available from the cylinder that is not used to maintain the
temperature of the web constant against the loss of heat from natural
evaporation and various wasted heat losses mainly in radiation. Numerous
376
THE DRYING SECTION AND CALENDERS SA.12
factors such as the air porosity and thickness of the web, the evaporation
conditions, and the resistance of heat transfer from the cylinder surface
thus affect the drying rate actually achieved.
Water that is vaporized at the cylinder surface must be replaced by
capillary movement of water within the body of the web, leading to a
gradual migration of water towards the cylinder. Two movements are thus
in operation: vapour passing out from the cylinder to the exposed surface,
partially condensing on the way, and water moving in the opposite direc-
tion within the sheet to fill the pores in the web vacated by vaporized water
thereby maintaining the cylinder surface wet. Dreshfield and Han provided
~
z
w
~
Q
\J
UJ
>-
Q
Kor
SURFACE.
Fig. 5. 3. Dye content through the thickness of a sheet dried with one side in contact
with a hot surface, illustrating migration of water towards the hot surface (after
Dreshfield and Han)
elegant proof of this movement of water within the web during drying by
making up laminated sheets with water containing a non-volatile dye so
that wherever vaporization occurred evidence of this was left in the form of
dyed fibres. Examination of the intensity of dyeing of the different
laminates indicated that there was always more movement of water
towards the hot cylinder surface than the exposed surface (Fig. 5. 3),
whereas only movement towards the latter would occur in natural evapora-
tion. Confirmation of this was also obtained with another technique
involving the use of beta-ray radiation to determine the moisture content of
the different laminates comprising the sheet. Sheets dried completely on
377
SA.I 3 THE DR YING SECTION AND CALENDERS
one side had a maximum moisture content at a depth of 20 per cent. to

30 per cent. of the thickness below the exposed surface; a condition similar
to this may be expected after paper is dried on an M.G. cylinder.
When the paper separates from the cylinder vapour trapped in the inner
surface of the web will be immediately released, while at the same time
rapid evaporation will occur from the hot surface, accompanied by a drop
in temperature of the web as latent heat is provided. Evaporation from the
outer, cooler surface of the web will also continue in the draw. Finally the
web contacts the next cylinder and the whole process begins again as the
new contacting surface of the web is rapidly heated up to the point where
vaporization again commences at the cylinder surface.
It is interesting to observe the effect on the movement of water within the
web that is produced by the conventional arrangement of alternating
cylinders. If the rate of vaporization at both top and bottom cylinders
were reasonably balanced, one would expect that there would be a net
movement of water from the centre to the surfaces of the web which was
approximately the same in both directions. In fact this is unlikely in
practice because the cylinder temperatures and heat transfer-rate are
frequently different in the top as opposed to the bottom cylinders. Also the
resistance to the movement of water at the wire side of the paper is different
from that at the top side due to the basic two-sidedness of the sheet; this
will produce a difference in the relative rate of migration of water to the
two sides of the sheet irrespective of any difference between the rate of
drying on top and bottom cylinders.
The difference in migration rate is the reason why when using certain
types of dye a colour difference can occur in the finished paper irrespective
of any colour two-sidedness due either to preferential dyeing of different
constituents of the furnish or to movement within the sheet of pigment
dyes. Where sheet structure and substance is varied, differences in the water
migration rate in different regions of the web produce a patchy, mottled
effect on the colour. Conditions which accentuate the difference in migra~
tion rate, such as larger moisture gradients resulting from more rapid
drying, may be expected to increase colour two-sidedness originating in
this manner.
SA . 1 3 Later stages of drying
The rate of drying achieved in the conditions described in the previous
section tends to be relatively constant down the drying section. If the
cylinder temperatures and draw conditions remain relatively constant, then
it is reasonable to expect this. But at some position in the later stages of
drying a point is always reached, sharper with some papers than others,
where the rate of reduction of moisture in the web begins to fall off. Figure
5 .4 (which actually applies to a felted machine but, as seen in the next
section, the overall pattern of drying is not affected by the presence of felts)
shows a typical drying curve illustrating the very short heating up period,
the constant rate of drying, and finally the falling-rate region; the curve is
a typical one from several given by Montgomery (11 ). It should be noted
that the moisture results are reported on the 'dry' basis in order to make the
378
rate of drying equivalent to the slope of the curve (a graph of this type with
moisture plotted on the 'wet' basis would be impossible to interpret
because the reference point, the total weight of the paper, is constantly
altering).
The generally accepted explanation of the falling rate of drying is that it
is associated with growing resistance to water movement within the web.
As long as the web is sufficiently damp for there to be steady migration of
water to the hot surface to replace the vaporized water, this surface will
continue to be saturated and conditions of heat transfer and vaporization
7o ~
"'
<
d>
.....
w
GS~
I
~ ~
I So <Oo
.... t-
z z
tll w
>-r. t-
-z:.
3 loo so 3
dJ
et.
40 :J
1-
lf)
~o o
2o ~
lo
0 '--~--~~-i-~~_,.-~_:.:::;::::::::::;;;;;;~o
o 2o 4-o Go Bo loo
% OF DRYING :5URFACf PA S:!i E-D Ov'ER
f:>Y WE:\?.
Fig. S. 4. Change in moisture content of a writing paper passing through the drying
section (after Montgomery)
remain relatively unaltered. But when a point is reached where the growing
dryness of the inner layers of the web acts as a resistance to movement to
the surface, i.e. when only finer capillaries within the web remain saturated
and thus require a stronger pull to overcome capillary attraction, then the
hot surface will no longer remain saturated. The zone where vaporization
continues must begin to recede gradually from the cylinder surface into
the thickness of the web, leaving paper in the outer layers comparatively
dry. As this occurs heat from the cylinder surface has to be transferred
through a widening barrier of comparatively dry fibre, so the rate of
drying gradually falls off and the overall web temperature increases. In
addition to this, water which is held more closely within the pores of
fibres is less easy to vaporize than 'free' water which may move along
379
SA.14 THE DRYING SECTION AND CALENDERS
capillaries, and this will contribute to lessening the rate of drying in the
final stages. Also, the vapour formed will find it more difficult to reach the
surface of the sheet as the continuous network of water capillaries is broken
and air intrudes into the web; the vapour must then escape by a process of
diffusion rather than by pressure difference through the water capillaries.
The critical moisture content at which the relatively constant drying rate
ceases is usually in the 30 per cent. to 45 per cent. region. The precise value
must be governed by several factors, but particularly by those which affect
the moisture gradient through the web as any increase in this will also
cause an increase in the average sheet moisture content at which the hot
surface is no longer saturated; Dreshfield and Han obtained some data on
this and found that the moisture gradient (and hence the critical moisture
content) increases when cylinder temperatures are hotter and also when the
sheet is thicker. The lower rate of drying occurring after the critical
moisture content has been passed implies that in general the longer this
point is delayed, the less total drying effort will be required; as the falling-
rate period may begin as early as half way down the dryers, the actual
value of the critical moisture content can thus be very important in
governing the overall drying rate.
In the falling-rate region of drying, the sheet properties (especially bulk
and porosity) are comparatively more important, particularly in their effect
on the transfer of heat as the zone of vaporization recedes from the surface
of the sheet. External conditions such as the velocity of the air stream and
the cylinder temperature are relatively much less critical because there is a
lower quantity of moisture removed from the web; the rate of removal of
the remaining moisture can only be increased significantly by raising
appreciably the surface temperature of the paper.
SA . 1 4 Drying on felted cylinders
The essential advantage gained by using a felt is that resistance to
transfer of heat from the cylinder to the web is substantially reduced. This
much is not in dispute (evidence will be presented later), and with other
things equal better heat transfer may be expected to permit faster drying
and reduce general heat losses. But with a felt covering the outer surface of
the paper the conditions of removal of vapour and of evaporation from the
web are appreciably altered; as a direct result of this the net effect of using
a felt may not be quite so beneficial.
In considering the process of drying on a felt-covered cylinder, it is
convenient to think of four distinct phases. The first phase is the short
period during which only the paper touches the cylinder surface, the second
phase covers the period when the felt contacts the paper and extends round
a majority of the cylinder surface, the third phase is when once again for a
short time only the paper remains in touch with the cylinder after the felt
wrap is completed, and the final phase covers the period when the web is in
open draw between cylinders.
During the first phase the situation is the same as that for an unfelted
cylinder. It is of such short duration that no vaporization may be expected
in this region, only a commencement of heating-up of the web.
380
In the second phase the heating-up period is more rapid due to the
increased heat transfer; once the process of vaporization at the cylinder
surface begins, the movement of vapour and water within the web will
proceed exactly as detailed for the case of drying without felts, but with
the difference that at the outer surface of the web the vapour and evaporat-
ed water must penetrate the felt. The mechanism of this transfer into the
felt has been the subject of much discussion and a detailed consideration
will be deferred for the moment. Suffice it to say at this point that move-
ment of water into the dry felt may be envisaged in two basic ways: by
direct transfer of liquid water either by capillary attraction under the
pressure of contact or from migration due to the force of vapour pressure
built up within the paper; or by diffusion of vapour into the felt together
with partial condensation when the felt temperature is lower than that of
the web at the outer surface (the heat released from condensation raising
the temperature of the felt and reducing further condensation). During
passage round the cylinder, a small proportion of vapour may diffuse right
through the dry felt and also some natural evaporation may occur from
the exposed surface of the felt if this becomes relatively damp.
During phase three when the felt leaves the web a release of vapour
occurs from the separating surfaces of both materials. Evaporation from
the surfaces of both web and felt will also be rapid due to their saturation
at high temperature. This will be accompanied by a rapid drop in tempera-
ture of the inner felt surface and, to a lesser extent, of the paper web which
is still in contact with the cylinder. During the short period the paper
remains on the web vaporization continues though probably at a lesser
rate due to some loss of close contact with the cylinder as the felt pressure
is relieved.
In the final phase, the open draw, more vapour is released from the
contacting surface of the web as it separates from the cylinder, and the
process of evaporation and cooling in air occurs for both paper and felt in
the manner already described. As it turns round the felt roll to contact the
next cylinder, the felt may approach an equilibrium condition similar to
the one existing immediately prior to contacting the first cylinder, but it is
more likely that as the felt travels down the drying section its state changes,
at least at the surface contacting the paper, and it both heats up slightly
and gets progressively damper. The effect of getting damper would be to
inhibit further transfer of water from the web and if this were the case the
advantages of having a means of drying the felt before it completed its run
along the cylinders would be potentially advantageous and might be a
strong argument for the use of intervening felt dryers, hot-air drying felt
rolls, and other devices. These will be discussed in a later section but it is
worth stressing that at the moment little evidence is available to indicate
that such progressive dampening of a dry felt does have practical signi-
ficance; it is more likely to occur early in the drying section with the first
top or bottom felts, though obviously it depends on many other conditions
in the drying section. One would expect, were drying to deteriorate
appreciably as a particular dry felt continued from one cylinder to the
next, that there would be a noticeable increase in the rate of dryin~ at th<;
381
SA.1 5 THE DRYING SECTION AND CALENDERS
point where the next felt section commenced; obtaining a complete drying
curve through the dryers could therefore be a useful method of detecting
this and assessing its importance.
SA. 1 5 The experimental work of Janson and Nordgren

The behaviour of paper and felt through the various phases as described
above has been illustrated in detail by Janson and Nordgren (35) and some
of their results will now be quoted. This work was undertaken on the
Swedish Central Laboratory experimental machine and involved studying
the drying of paper over a small felted M.G. cylinder with an associated
felt dryer. The moisture content of the paper entering the M.G. cylinder
could be altered by varying conditions in some pre-dryers; after-dryers
enabled the final reel-up moisture to be kept constant. The most remark-
able feature of the work was devising a means of continuous measurement
of the temperature both of the dryer at different distances from its surface.
of the felt, also at different depths through its thickness, and at the surface
between the felt and the paper web. This was achieved with thermocouples
embedded in the dryer and felt and placing ingoing leads to the junctions
in such a way that each in tum could be connected to a recorder from
which a curve showing temperature variation through an entire cycle
could be obtained.
Figure 5. 5 shows the temperature variation of the paper, and of the felt
at depths of 1, 3·5 and 6-5 mm. from the contacting surface (the felt was
approximately 8 mm. thick). As each of the phases is passed the behaviour
of temperature can readily be interpreted in the light of the description in
the previous section. For instance although the web temperature will begin
to rise on contact with the cylinder, the rate of rise is increased enormously
once the felt touches the cylinder at the beginning of phase 2; this illustrates
the important benefit of the felt in improving heat transfer from the
cylinder. The web temperature flattens off towards the end of the cylinder
wrap as equilibrium conditions of vaporization and water transfer to the
felt are approached, but then falls with reduced contact and increased
evaporation when the felt leaves the cylinder in phase 3. The drop in
temperature when the paper itselfleaves the cylinder in phase 4 is extremely
rapid.
The temperature of the felt closest to the paper surface varies appreci-
ably, showing a substantial rise while wrapping the cylinder, a fall as
evaporation occurs in the open draw, followed by a rise in temperature on
the felt dryer and a further fall afterwards. The same pattern can hardly be
distinguished at all further into the felt towards the middle, while the outer
surface of the felt shows little alteration in temperature at all. The implica-
tion here is that there is little or no evaporation and diffusion of water
vapour through the dry felt; it is only at the contacting surface that the felt
is affected significantly. A further point of interest is that temperature rise
on the felt dryer is at a much lower rate than on the M.G. cylinder, despite
the fact that the surface temperature of the felt dryer was higher than that
of the M.G. cylinder; this is due essentially to the difference in heat
382
THE DRYING SECTION AND CALENDERS SA. I 5
transmission when this is purely by conduction, as on the felt dryer, and
when augmented by the transfer of water vapour from the wet web.
While the temperature curves shown in Fig. 5. 5 are stated to be fairly
characteristic of a number of runs in that their general appearance was
always retained, observations taken over a range of various conditions
showed up some important and interesting variations. The effect of such
factors as the type of felt used, felt tension, machine speed, and steam
pressure were all investigated and reported either by Janson and Nordgren
I Ft1..T
k\.G. ('y1...1NOER DftYER
S.U~FACll! TfMP.
loo
So
FELT.
(0&PTH e:,ELOri
80 <::ONTACTING Sui:tl"ACc)
ou
u.I
3·5 rnrn.
et.
:J 14-----t+---..l,----1------'r---.;:,.,~<O·S rnm
r 7o I M<'I'"\.
<
Q!
.....
~
~
Go
So
0 5 10 15 2o
TtM.E IN SECON0.5.
Fig. 5 .5. Variation in temperature of paper and of felt (at different depths) round an
M.G. cylinder and felt dryer (after Janson and Nordgren)
or by their colleagues in later papers; these will be dealt with later. For
the present it is worth examining the results obtained when ingoing
moisture content of the paper web and the amount of felt wrap round the
M.G. cylinder were varied as these serve to illustrate further the process of
drying as detailed above.
Figure 5. 6 shows the effect of ingoing paper moisture content on the
pattern of temperature rise of the paper on the M.G. cylinder. At high
ingoing moisture content there is a tendency for the temperature during
the heating up period to increase to a higher level before the curve begins
to drop away as equilibrium conditions of vaporization and water transfer
are approached. This is primarily because the moisture content and
383
temperature of the felt would be initially higher under these conditions

due to the overall higher rate of transfer of water from the paper, and this
would force the paper web temperature to adopt a higher value before
transfer of water (and thereby heat) from the web commenced. But the rate
at which the rise in temperature then drops away is much greater at higher
llo DRYING PHASE.

II 1m1
I I
I I
1 I
I
loo
INGOING MOISTURE:
CoNTiNT.
15%
9o
Qu I
\lJ
I
a I
::J
t:;:
c:t. 80
!/
~
0.
~
.__
\ll
7o
<Do I
0 2 3 4 5 c. 7
TIA",E IN SEC.<.:>ND5.
Fig. 5. 6. Dependence of temperature of paper over drying cylinder on ingoing moisture
content (after Janson and Nordgren)
ingoing moisture contents and the ultimate web temperature in fact

approaches closer to the cylinder surface temperature when the web is
initially drier.
Using the same experimental set-up Brauns and Ponton (40) investigated
the effect of varying the run of the felt and arranged that it left the M.G.
cylinder at three different positions giving wraps of 3·8, 2·5 and 0·8 metres,
compared to the paper wrap of 4·4 metres. In this instance ingoing
moisture content of the paper was kept steady at about 40 per cent. The
results indicated an evaporation rate (including the free draw of the paper)
of 83, 73 and 70 kilograms per hour respectively for the three degrees of
wrap, showing conclusively that it was preferable to have the felt wrap as
large as possible. But it may be noted that the difference is not so great as
384
THE DRYING SECTION AND CALENDERS SA.1 6
might be expected, though there is no evidence that the rate of drying was
significantly decreased by the growing moisture content and temperature
of the felt as it progressed round the cylinder.
SA. 1 6 Transfer of water from paper web to felt
In 1954 and 1955 Nissan and Kaye (12, 19) presented some theoretical
calculations based on heat transfer considerations that have sparked off a
controversy that is still not satisfactorily settled. However, there is now
sufficient evidence to put the matter in perspective and show that in
practice the issue is not likely to be of any real importance anyway.
What Nissan attempted in the first place was to calculate how much
water would be removed from the paper web on the assumption that the
felt does not absorb any water and in fact retards drying by restricting
evaporation from the web to 10 per cent. of what would occur if it were
not present. Calculation of the water that would be removed on this basis
produced a result that was too small to account for the actual rate of
drying, certainly on slower machines, and Nissan therefore concluded that
a substantial transfer of water from the paper web to the felt must occur.
As the actual water removal in this phase is greater than it would be in free
air it appeared that a percentage of the transferred water must be in liquid
form, i.e. by direct migration of water from the web to the felt, rather than
by means of transfer in vapour form followed by subsequent condensation.
Nissan concluded from these results that liquid absorption of water into
the dry felts was an all important function of the felt (hitherto the general
feeling was that the felt served only to improve heat transfer from the
cylinder to the paper web), and that this form of transfer was more effective
for some reason on slower machines. Following from this he considered
that dry felts (especially the first pair) should be constructed on the paper
side to encourage liquid absorption, i.e. be soft, bulky, with a good nap,
and made of wool, while retaining adequate strength and dimensional
rigidity; also the felt should not be allowed to get too damp (or this would
inhibit liquid transfer) so that intervening felt dryers or even the use of a
drying cylinder for the felt rather than the paper should be considered.
The conclusion that some reduction of moisture content of the paper
occurs actually on the cylinder in addition to evaporation in the free draw
is not disputed nowadays and several workers have confirmed this, e.g.
Soinenen (25), and Janson and Nordgren in the experimental work
described above. The actual proportion of the drying occurring in the two
phases is known to vary with the machine speed and, as will be discussed
in more detail in SB. 6 1, increasing the speed reduces the actual proportion
of drying on the cylinder in favour of loss of moisture in the- open draw.
But the second conclusion reached by Nissan, that a significant proportion
of drying on the cylinder occurs by direct migration of liquid water, has
been hotly disputed.
Nissan and his colleagues have elaborated this theory (24, 54, 79) and
reported confirmatory evidence obtained from measurements on a slow
paper machine (27, 32) and in laboratory experiments (on drying muslin,
not paper(54)). In the machine work, samples of the paper were drawn
385
5A.1 6 THE DRYING SECTION AND CALENDERS
from between the cylinders and the general conditions of drying over the
whole section were determined; being a very slow machine (120 f.p.m.) in
this case most of the water was shown to leave the paper web when in
contact with the cylinders and consideration of the evaporation conditions
indicated that some of this must have transferred to the felts in liquid form.
Brauns and Ponton in their work on the Swedish experimental machine
referred to earlier, used the comprehensive data available for cylinder,
sheet and felt temperatures, steam and air conditions, and steam consump-
tion to perform heat energy and mass transfer calculations in three different
ways; each of the three methods gave a good correspondence and indicated
that the proportion of water transferring from the paper web to the felt in
vapour form was approximately twice that in liquid form. These authors
considered that this liquid migration was probably partially by capillary
effect (only when the felt is comparatively dry, transfer to a damp felt is
hardly possible in view of the relative capillary sizes being so much larger
in the felt than the paper) but mainly the liquid transfer was a result of the
pressure within the sheet created by the expansion of air and steam.
In a later paper, Ponton (77) also points out that, despite the com-
paratively low temperature of the web entering the dryers, appreciable
evaporation often begins immediately on the first cylinder and the rate of
drying is hardly less than in subsequent dryers, even though the web is still
largely being heated up at this stage. Exceptions to this are when no felts or
felt dryers are used. These considerations support the occurrence of liquid
transfer to the felt at the beginning of the dryers.
Aligned against this formidable weight of evidence are the results
reported by Kirk and his colleagues (92). This work was designed specific-
ally to obtain experimental evidence as to the nature of the water transfer
without relying (as all the results reported above do) on calculations which
are dependent on various assumptions, particularly with regard to heat
transfer coefficients. Kirk used an experimental apparatus designed to
simulate drying (similar to that used by Smith and Attwood (1, 7) to be
described later) with three different types of felt: wool, terylene, and
reinforced cotton. Handsheets were made from water containing a 1 per
cent. solution of potassium chloride which is non-volatile, i.e. water
removed from the sheet by vaporization would leave the chloride content
of the sheet unchanged. By using a carefully standardized extraction
technique Kirk was able to determine the chloride remaining in the sheet
after drying and compare this with the known water content of the undried
sheet; any drop in the total chloride recovery would indicate liquid
transfer.
Over a wide range of conditions (including different felt pressures,
substances of paper, dryer temperature, etc.) in no single case was -~my
evidence found that the least liquid transfer had occurred. Composite
sheets showed a net movement of water between the laminates towards the
hot surface, as found by Dreshfield and Han, but again no liquid transfer
to the felt. On a small experimental paper machine similar work indicated
that on the first cylinder no tnore than 0·2 per cent. of water passed into
the felt in liquid phase, a result hardly significant. The implications of this
386
THE DRYING SECTION AND CALENDERS 5A.2
work are that to all intents and purposes water is removed on the drying
cylinder only by the vapour transfer mechanism. The nature of the felt
itself is only of secondary importance (Kirk draws attention in this respect
to the small difference in performance found between a wide range of felts
tested by Smith and Attwood) and should preferably be relatively porous
to allow as little restriction to the passage of vapour as possible.
What happens on operational paper machines remains undecided. But a
reasonable assessment of the situation is that when liquid transfer does
occur (as it is known to on occasion because dry felts can become coloured
from the paper dyes) it will only be likely to any significant extent over the
first and possibly second cylinder, and then only on relatively slow machines
where the moisture content of the web entering the dryers is very high. With
regard to the desirable properties of dry felts, the possibility of some liquid
transfer in the first top and bottom positions must be kept in mind, but by
and large other considerations weigh more important.
5A.2 PERFORMANCE OF THE DRYING SECTION

The previous section has described the manner in which water is removed
from the paper in the dryers. Attention is now turned to another important
aspect of the section, the overall rate at which the paper is dried and the
general performance of the dryers with regard to the efficiency of utiliza-
tion of the heat provided in the steam.
SA . 2 1 Transfer of heat to the paper
The amount of heat transferred from a drying cylinder to the paper at any
given position is dependent on two variables, the temperature and con-
dition of the steam in the cylinder and the temperature of the paper con-
tacting the surface of the cylinder. For the purposes of the present discus-
sion it is assumed that unpolluted dry saturated steam is used (use of
superheated steam is mentioned in SB. 1 5), so that temperature and
pressure of the steam are precisely related.
Suppose, for example, that saturated steam in the cylinder is at a
pressure of 6 p.s.i.g. or 230 deg. F., while temperature of the paper at the
surface of the cylinder is 150 deg. F. Then the rate of heat transfer will be
determined by the difference between these temperatures, i.e. 80 deg. F., in
relation to the overall resistance to transfer of heat between the inside of
the cylinder and the paper. This resistance may be considered to be the
sum of the resistance of each intervening layer comprising: (i) the con-
densate film on the inside of the drying cylinder; (ii) scale or rust on the
inside of the cylinder surface; (iii) the metal of the cylinder; (iv) scale and
fuzz on the outside of the cylinder; and (v) air between the outer face of
the cylinder and paper. Not a great deal is known about the resistances to
heat transfer offered individually by these different layers but it is generally
considered that (i) is small unless rimming of the condensate occurs, (ii) is
relatively small, (iii) is also relatively small except for thicker-walled M.G.
cylinders, while (iv) and (v) together account for a considerable portion of
the total resistance. In the example, resistance in each layer might be (i)
387
0·0005, (ii) 0·0013, (iii) 0·003, and (iv) and (v) together 0·02, each figure
being expressed as the reciprocal of B.t.u. transferred per hour per sq. ft.
per 1 deg. F. difference. Total resistance would then be the sum of these,
i.e. 0·024, and heat transfer equal to 80 x l/0·024 = 3,300 B.t.u. per sq. ft.
per hour. The temperature at the junction of the condensate film and
inside of the cylinder would thus be only a fraction under the cylinder
steam temperature at 230 deg. F., the temperature of the inner and outer
metal surfaces of the cylinder about 226 deg. F. and 216 deg. F. respective-
ly, and the biggest temperature drop would be between the outer cylinder
surface and the paper at 150 deg. F.
If this supply of heat were just sufficient to balance the rate of evapora-
tion and heat losses, then the temperature of the sheet would remain in
equilibrium at 150 deg. F. In fact, of course, such a state of affairs is only
likely to be attained at the point on a drying cylinder where the paper has
been heated to a maximum stable temperature, if such a point is ever
reached. At most of the positions round the surface the flow of heat from
the cylinder (though governed always by the gradient pertaining at each
instant) will exceed that used for evaporation and heat loss, so that the
sheet and felt will be heated up in the manner described earlier. As the
temperature of the sheet next to the cylinder rises, so the heat transfer will
diminish.
SA. 2 2 Heat balance for the whole drying section
Under steady conditions, the steam pressure used for drying the paper at
any instant may be said to give a rate of heat transfer through each cylinder
which is just sufficient under prevailing conditions to evaporate water from
the paper down to the desired reel-up moisture content; in other words, the
supply of heat meets the demand. What is the situation if, for some reason,
the paper enters the drying section containing more water? In this event
more heat is required both to heat up the sheet and evaporate the additional
water.
Even if the steam pressure in the cylinders is unaltered, it is interesting
first to note that more heat is in fact provided automatically to the drying
section. Thus, because the sheet contains more water entering the dryers it
will not in the early stages of the section heat up quite so quickly; the sheet
temperature will be lower at any position round each cylinder than pre-
viously in the same position, so it follows that the rate of heat flow will,
overall, be greater until the sheet reaches the normal drying temperature
associated with the 'constant' rate period. Again, although other things
being equal the drying rate in the 'constant' period must be the same in
both cases, the point along the dryers where the zone of vaporization
recedes into the sheet and the drying rate falls off must be reached later
when initially there is more water to remove; in other words, the relatively
higher heat flow associated with the 'constant' rate period will extend over
a greater number of drying cylinders. It is for these reasons that in any
drying system there is an important built-in automatic correction for
fluctuations in the moisture content of the paper entering the dryers. Varia-
tions in heat demand in the drying section are to a great extent taken care
388
of without the need for any instrumentation. This is perhaps as well,
otherwise it is doubtful whether many machines could run at all.
Further consideration indicates, however, that only a limited automatic
compensation for variations in sheet condition entering the dryers may be
expected. It can be shown that the extra heat transferred will only partially
provide the heat required to remove the additional moisture. The inevitable
result is that the moisture content of paper at the reel-up rises, even though
only slightly, whenever the sheet entering the dryers contains more water.
A further and more important effect in an uncontrolled drying system is
that with greater heat demand the steam pressure does not, as assumed
above, remain constant. It will in fact gradually diminish and this, by
reducing the temperature inside the cylinders, inevitably reduces the heat
transferred to the paper and brings about an even greater increase in
moisture content at the reel-up. Such a situation is very easily overcome by
provision of a simple control system in which measurement of the steam
pressure or temperature in the main section of cylinders is used to regulate
the opening of the main supply valve. More will be said about this later but
it may be noted at this point that such a regulator is superior to the un-
controlled system simply because it provides a straightforward method of
preventing the steam pressure from falling in the wake of an increase in
demand for heat; the supply valve is automatically opened to maintain the
pressure and provide more steam instead of remaining in a fixed position.
A final example of the relation between supply and demand of heat
which may be cited is the situation when a break occurs at the wet-end of
the machine. As there is suddenly no web on the dryers, the dry felts
contact the cylinder surface directly and, despite the greater resistance to
heat transfer across the cylinder/felt interface, immediately begin to heat
up and dry out. An equilibrium state would be reached only when the felt
temperature (and even more so the cylinder surface) approached that of
the steam temperature to an extent which reduced the temperature gradient
and hence the flow of heat to the relatively small amount needed to
counteract radiation and convection losses. Even though the overall
resistance to heat transfer is greater without the paper, the result would be
that the dry felt became extremely hot with the risk of scorching. This is
the reason why it is always necessary during a break to reduce the steam
temperature and on modern steam control systems this is provided for
automatically. The extent to which the steam temperature is reduced
depends on several factors which will be discussed later; apart from keeping
down the cylinder surface and dry felt temperature to prevent damage to
the dry felts, the situation when it comes to feeding up the sheet is also of
considerable importance.
5A. 2 3 Relationship between steam pressure and drying rate
The steam pressure in the cylinders necessary to dry the paper to the
desired moisture content is widely regarded as an indication of how well
the section is operating. If when running a familiar grade it is found that a
much higher pressure than usual is needed to dry the paper, then it is
suspected either that the paper is too damp entering the dryers or that
389
waterlogging or some other trouble in the section is occurring. This, taken

by and large, is perfectly sound reasoning, but care needs to be taken in
relying too much on steam pressure as an arbiter for whether the drying
section is working efficiently. The reasons for this will now be discussed.
At any particular time the average steam pressure provides a temperature
inside the cylinders giving an appropriate heat transfer rate to dry the
paper. Should the heat demand increase appreciably (such as when speed
is increased or a heavier substance is run at the same speed, i.e. well beyond
the bounds where automatic compensation of supply takes place to meet
variations in demand, as discussed in the previous section) then maintain-
ing the same steam temperature in the cylinders will not give an adequate
heat flow and the paper will reel too damp. To overcome this the remedy
is, of course, to increase the steam pressure: for the same or possibly
slightly higher average paper temperature a greater temperature gradient
is established and the overall heat flow increases, providing the additional
energy needed for drying. In the same way, increased steam pressure is
effective in compensating for a substantial increase in moisture content
entering the drying section, or for an increase in resistance to heat transfer
occasioned by waterlogging of cylinders; hence the dryerman's usual
interpretation when he finds a higher steam pressure necessary is in
general perfectly sound.
But there is an important distinction that should be made between high
steam pressure necessitated by excessive moisture entering the dryer section
or by waterlogging of cylinders, and this shows up in the difference in flow
of steam for the two situations. Where the presence of more water to be
evaporated from the paper is the reason for a higher drying pressure being
necessary, then the effect of this higher pressure is to produce an increase
in the average temperature gradient between the inside of the cylinders and
the paper and hence in the flow of heat. The additional heat used in this
way must be provided by the steam. Within a fairly normal working range
the heat given out by condensation of unit weight of saturated steam is
practically independent of the pressure (actually, being equivalent to the
latent heat, the heat given out decreases slightly with increasing pressure).
So effectively the additional pressure used for drying will in this case be
reflected almost exactly by a corresponding increase in the flow of steam to
the cylinder section.
But in the case of waterlogged cylinders, the increase in steam pressure
is required purely to overcome the greater resistance to heat transfer within
the cylinders. In other words, to achieve the same rate of drying a greater
temperature gradient is needed and this is achieved by increasing the
temperature within the cylinders. No increase in flow of heat occurs, and
hence the steam flow remains unchanged (actually there would be a very
slight increase to offset greater radiation losses from the feed-pipes and
sides of the drying cylinders).
Increase in steam pressure is thus not of itself an indication of greater
usage of steam. However, as a corollary to this, if knowledge of the steam
flow is available in addition to the pressure then the general source of any
change in demand (i.e. the paper or the dryers) can be more easily identified.
390
THE DR YING SECTION AND CALENDERS SA.24
A steam flow measurement is thus very useful for indicating whether or not
the drying rate has altered with the steam pressure.
A further point worth noting is that apart from small changes in ex-
traneous heat losses, there is no reason to expect that the performance of
the drying section, i.e. the quantity of steam used compared to the water
evaporated from the paper, would alter in either of the two situations just
considered. This brings in a further aspect which deserves some considera-
tion.
When cylinders are waterlogged it will be said that the dryers are in-
efficient. So they are, but not in the sense that steam utilization is poorer.
The meaning here is essentially that greater pressure is needed to accom-
plish the same rate of drying. On many machines the pressure available is
limited due to the necessity to prevent the back-pressure of power-
generating turbines exceeding a fixed value. In such cases reduction of the
heat transfer rate by waterlogging may limit production because the
machine is already run with practically the maximum steam pressure
available and this is effectively the governing factor determining the
machine speed. This is highly important, more so than the relative efficiency
of steam utilization, and it is justifiable in such cases to think of the per-
formance of the drying section as dependent on the steam pressure needed
to keep up normal production. But it must be realized that this assessment
of performance applies essentially to changes in the heat transfer resistance
between the cylinder and paper. The correct measure of performance of the
dryers refers to how efficiently steam is utilized, and this aspect will now be
discussed.
SA. 2 4 Efficiency of steam utilization
The performance of the drying section of a paper machine at any particular
time is assessed by determining the weight of steam which is used to remove
a unit weight of moisture from the web. To calculate this figure it is
necessary to measure the flow of steam to the machine, including any
supplied for felt dryers and air heating (though the latter may be treated
separately); strictly speaking any steam used in calenders, where the
purpose and effect is not to remove moisture, should be excluded. Also the
average moisture content of the web entering and leaving the dryers must
be determined. For practical purposes the moisture content at the reel-up
is usually taken as equivalent to that leaving the dryers; if a sweat roll is in
use moisture at the reel-up may be taken only if the moisture added to the
paper after the dryers is determined or allowed for separately.
The water removed from the web is calculated (in British units) as
(M m) / (100 - M) = (d - p) / (1 + p) lbs. per lb. finished paper,
where M, mare the moisture content (wet basis) entering and leaving the
dryers, and d, p are the water/fibre ratios entering and leaving the dryers,
respectively. This figure multiplied by the production, determined for the
relevant speed, deckle, and substance in the usual way, gives the total
water evaporated per unit time (a value which is, incidentally, of use in
assessing ventilation conditions) and this is then compared with the lbs. of
steam used in the same unit of time.
391
Performance calculated in this way gives a useful measure of the

efficiency of steam utilization. The figure obtained depends on the type and
substance of the paper and its moisture content entering and leaving the
dryers, but is relatively independent of the steam pressure used. Depend-
ence on the paper itself occurs because the heat energy required to dry the
web is a function of the structure and thickness of the paper. Dependence
on the moisture contents occurs because it is easier to remove additional
water in the web entering the dryers than to remove the same quantity of
water at the end of the dryers. Some dependence on the steam pressure
arises due primarily to the lower latent heat associated with higher
pressures, and also because greater heat losses must be expected at higher
operating temperature, but unless the range of pressure is wide this should
not be very significant compared to the accuracy it is possible to achieve
when determining the performance figure. Despite the various factors
affecting the result, for any particular grade made on a paper machine the
straightforward determination of steam utilization efficiency in the way
detailed above is of use for evaluating alterations to the drying or ventila-
tion conditions of the machine and for detecting long-term trends in
performance.
With regard to the evaluation of alterations to the drying section it is
appropriate at this point to emphasize the importance of using a perform-
ance figure relating the water evaporated to the steam used rather than, as
is frequently the case, relying solely on the steam pressure to indicate
whether an improvement has occurred as a result of making the alteration.
The argument that if a drop in steam pressure occurs for similar running
conditions then this indicates an improvement is not always true. Certainly,
on a machine with limited drying capacity, an increase in production
would be likely to follow and under such conditions this will probably be
an all-important consideration. But with regard to the performance or
efficiency of the drying section it is quite possible for an alteration to bring
about a drop in the steam pressure and at the same time an increase in the
weight of steam needed to evaporate unit weight of water from the paper.
This may happen (to take a hypothetical example) if a new type of felt
were used which allowed a greater tension to be applied by virtue of con-
structing the felt to be very dense and non-porous. The greater tension
could reduce heat transfer resistance at the cylinder/paper interface to such
an extent that even with a lower temperature gradient between the inside of
the cylinder and paper there is a greater heat flow. Thus, a situation might
arise where the denseness of the felt considerably inhibited transfer of
vapour from the web to the felt on the drying cylinder, necessitating an
overall higher temperature of the paper to achieve the same drying rate,
but with the result that the greater heat flow demanded for this could be
supplied at a lower steam pressure. Performance in the sense of steam
utilization efficiency would therefore be poorer despite the drop in steam
pressure.
A second and simpler example of a situation where lower steam pressure
resulting from an alteration to the drying section may be less efficient could
arise when steam is newly used for heating of air supplied to the section. A
392
greater rate of drying is achieved because of improved evaporation in the
open draws, so for the same production there would be a lower steam
pressure in the cylinders. But the overall consumption of steam related to
the water evaporated may, as a result of the extra steam used for air
heating, be significantly greater.
SA. 2 5 Comparing performance with other machines
Apart from the use of performance or steam utilization efficiency figures to
keep a long-term check on the drying section of a particular machine, or to
evaluate alterations to the machine, it is also instructive to compare the
figures with those obtained on other machines making similar grades of
paper. Unfortunately the difficulty here is that little reliable information is
readily available. A detailed analysis of data provided in answer to a
carefully standardized questionnaire circulated to a large number of
Canadian mills making newsprint has been reported by Snider (30), and in
this close agreement was found over a wide range of conditions between
the lbs. of steam used on a machine and the lbs. of water evaporated from
the paper, the average slope of the curve, i.e. the ratio between the two,
being 1·76. Variations from this figure on individual machines would
indicate a relatively higher or lower steam utilization efficiency than
average. Snider also pointed out that extrapolation of the accumulated
results indicated that a significant evaporation would occur at zero steam
consumption, so that relatively higher utilization efficiency figures may be
expected on lower speed machines.
There is little similar data relating to other grades of paper and this is
partly due no doubt to the relative paucity of steam flow instruments on
individual machines. But another form of comparison between the per-
formance of different machines making the same grade of paper which can
be useful is the overall rate of drying (commonly known as the evaporation
rate); this figure when related to standard data for the grade of paper in
question may be particularly useful for giving a clue as to the likelihood of
improving the capacity of a machine where the production is limited by the
drying section. The evaporation rate is usually determined by relating the
total water removed from the sheet in unit time (determined as above) to
the area of dryer capacity in use. The latter figure for convenience includes
the whole area of each drying cylinder (together with felt dryers) within the
width of the sheet, not just the area wrapped by the paper which is assumed
in practice to vary relatively little from one machine to another.
The evaporation rate depends very closely on the steam pressure (see
Fig. 5. 7 which is from the same report by Snider relating to newsprint;
similar curves for other grades of paper may be found in the T APPI Data
Sheets). This dependence is, of course, due basically to the fact that, other
things being equal, the rate of drying depends on the rate of heat transfer
and this is greater for a higher steam temperature. Also the evaporation
rate is dependent on the moisture content entering the drying section; the
greater this is the higher will be the evaporation rate, and under such a
condition the capacity of the dryer part would appear better than it
actually was.
393
SA. 3 THE DRYING SECTION AND CALENDERS
Another form of comparison is to use the straightforward production

figure instead of the water evaporated and to calculate steam utilization
efficiency or the drying rate on this basis. The spread of results for different
machines then tends to be wider because variations in moisture content
~
Ill
rt
-(
"l.
}:
~ Cl!
~ a
Ci ...:
II.
z
0 d
111
~rt. d.
() ""'
t'l.
0.. et
<( ')
'.:> 0
UJ I
c:c
Q'.
"'
~
210 2~o 250 '27o 290

STE:AM TE:MPERATUR.E- °F.
o 4 8 12. !Go '20 ,Oo 4o

STEAM PRES5URE. p.6. i .•5'
Fig. 5. 7. Evaporation rate for newsprint related to steam pressure in the drying
cylinders (after Snider)
entering the dryers exert a relatively greater influence. Information on

drying rates based on this method can also be found in the T APPI Data
Sheets for different grades of paper.
5A.3 EFFECT OF DRYING ON PAPER PROPERTIES

Although the purpose of the drying section on a Fourdrinier machine is
essentially to remove water from the paper web, it has long been recognized
that the manner in which this is accomplished has an important effect both
on the behaviour of the web during the drying and on the characteristics of
the finished paper. The key to these phenomena lies in the relationship
between two fundamental properties: the amount of shrinkage in the web
394
THE DRYING SECTION AND CALENDERS 5A.3 1
which would occur naturally as the paper dried, and the amount of shrink-
age which in fact is allowed to take place on the machine.
It is proposed in this section to discuss first the nature of shrinkage in
the drying section. The effect on the final properties of the paper of this
shrinkage, and of its restriction during drying, are then described together
with other phenomena associated with the drying process, such as cockling
and curl. A brief discussion is also included of the factors affecting the
moisture content of paper in equilibrium in air, and the influence this has
on the properties of the paper.
SA. 3 1 The nature of shrinkage
The basic source of shrinkage of the web is the shrinkage of individual
fibres comprising the web. Surface tension forces are responsible for some
of the overall contraction in the earlier stages of drying, but these alone
could not account for the manner and magnitude of the shrinkage com-
monly observed. But even though the origins of shrinkage are thus clear,
the manner in which the contractions of individual fibres affect the whole
web has been argued about for many years.
Intimately connected with shrinkage is a second characteristic of paper
that requires explanation: this is its dimensional instability in the presence
of alteration in humidity (moisture expansivity) and when wetted (hygro-
expansivity). Any theory of shrinkage on the paper machine itself must be
able to explain these dimensional changes and account for the varying
degrees of irreversibility which are known to occur when paper is dampen-
ed and re-dried.
To consider first the nature of shrinkage itself, in recent years the theories
of Page and Tydeman and their colleagues at the British Paper and Board
Research Association have become generally accepted as giving, in broad
outline, an adequate explanation of this phenomenon. The more important
results reported by these workers (see particularly references 80 and 128)
will now be briefly described.
During drying, individual fibres may shrink as much as 20 per cent to
30 per cent. in the transverse direction, but only 1 per cent. to 2 per cent.
longitudinally. This difference in shrinkage, related to the preferential
orientation of fibres in the machine direction, has for a long time been
regarded as responsible in some way for the common observation that
paper made on a Fourdrinier machine always exhibits greater shrinkage
in the cross direction than the machine direction.
However, Page and Tydeman have pointed out that a straightforward
relationship between individual fibre shrinkage and contraction of the
whole web is not compatible with the..conventional idea of paper consisting
of fibres bonded together at their cross-over points; such an assemblage
of fibres would contract only by an amount equivalent to the longitudinal
shrinkage of individual fibres and would in fact be independent of any
transverse shrinkage of the fibres. Figure 5. 8 makes the reason for this
clear by illustrating that, in the absence of longitudinal shrinkage and with
the fibres restricted only at the cross-over points, transverse shrinkage is
possible without the overall area of the paper being affected.
395
This anomaly implies that during shrinkage either a relative sliding

movement occurs at the cross-over points, or that the longitudinal shrink-
age of fibres is actually greater in the sheet than when measured on
individual fibres in isolation, possibly due to a form of enforced miniature
creping along the fibre length. By dyeing a few fibres in a handsheet and
comparing microscope photographs covering the same field before and
after drying the sheet with complete freedom to shrink (by floating in a
Fig. 5. 8. Diagram to illustrate that transverse shrinkage of fibres does not produce
sheet shrinkage if only the centres of the bonded crossings are fixed (after Page and
Tydeman)
mercury bath), Page and Tydeman found that individual fibres did in fact
shrink in length to a much greater extent than had hitherto been measured
on isolated fibres not forming part of a sheet. By increasing the beating
degree of the fibres comprising the handsheet, fibre shrinkages as high as
12 per cent. were obtained; furthermore, it was shown that the contraction
of the whole sheet was always the same as the average longitudinal shrink-
a~e of the fibres. Careful examination of the positions of the individtJal
~96,
THE DR YING SECTION AND CALENDERS SA.3 2
dyed fibres showed no relative movement at the points where they crossed
other fibres.
The relatively large longitudinal shrinkages observed in this work appear
at first sight to confirm that some form of micro-creping of individual
fibres between bonds must occur during drying. But there is no reason why
such micro-creping should not be observable when isolated fibres are dried,
and also the largest sheet shrinkages occur in papers that are highly beaten
and therefore have a lower portion of their length free from contact with
other fibres. It is apparent, therefore, that the shrinkage of a fibre is
promoted not between bonds in the unattached portions of the fibre, but at
the bond sites themselves. To account for this it must be assumed that
fibre-to-fibre bonding is strong enough during shrinkage for the transverse
shrinkage of one fibre at a crossing to produce a compressional force
causing a contraction in length of the other. Although the unbonded areas
of the fibre shrink in length only by a very small degree, at the bond sites
the high transverse shrinkage of other fibres in contact induces a much
greater total reduction in length. This in turn is transmitted to the sheet as
a whole and results in an overall shrinkage of the same order. The greater
the degree of beating with corresponding production of bonds between
fibres, the greater the bonded length of individual fibres and the greater the
overall shrinkage.
This explanation of the nature of shrinkage in the web has been con-
firmed by Page and Tydeman in a number of different experiments in
which, amongst other results, longitudinal shortening from micro-com-
pression in fibres at bond sites has actually been observed and photo-
graphed. Apart from contraction at the bond sites, kinks (especially in
longer, unbeaten fibres) and micro-creping can also be observed in the
unbonded portions of the fibres; this is a direct result of the compressional
forces which must be set up between the fixed bond sites in individual
fibres if the sheet is to shrink and at the same time retain its general shape
and structure. Thus, the total longitudinal shrinkage of individual fibres
in the sheet is compounded of the shrinkage at the bond sites themselves,
together with a certain amount of enforced compression in the free areas.
It may therefore be expected that the magnitude of shrinkage in a sheet of
paper depends primarily on three factors: the intrinsic potential shrinkage
of the fibres primarily in the transverse direction; the extent of bonding
which occurs in the sheet; and the resistance of fibres to micro-creping and
to bending and kinking in the free regions of fibres, i.e. the general rigidity
of the fibres towards compressional forces acting along their axes. Where
orientation occurs in the sheet, the transverse shrinkage of fibres in the
preferred direction acts to a greater proportional extent on fibres aligned
at right angles, hence more contraction of the sheet will also occur at right
angles to the preferred direction.
SA. 3 2 Expansion of paper on re-wetting
In the presence of a higher humidity, or when immersed in water, a
dimensional expansion takes place in a sheet of paper. The relationship
between this expansion and the shrinkage which occurs during drying is of
397
both fundamental and practical interest, and much investigational work

has been done on this subject. Page and Tydeman, in their experiment
using dyed fibres in handsheets dried without restriction to shrinkage,
found that re-wetting of the paper produced a longitudinal expansion of
individual fibres which was the same as that of the sheet as a whole. This
confirmed that changes in sheet dimensions depend essentially on the
changes in length actually taking place in individual fibres in the sheet.
Expansion of a sheet on immersion in water is not completely reversible;
normally only about 40 per cent. to 70 per cent. of the shrinkage is
recovered, and reversibility tends to diminish when fibres are more highly
beaten. This can be attributed basically to incomplete recovery in length of
individual fibres due to physical changes produced by the kinking and
micro-creping. It is probable too that the original transverse shrinkage of
the fibres may not be completely reversible and also surface tension con-
tractions associated with the earlier stages of drying may result in irrever-
sible changes in position relative to contiguous fibres within the framework
of linkages set up by the bonds.
Dimensional changes in the presence of a higher humidity have a
similar effect, though of a lower magnitude.
SA. 3 3 Effect of restricting shrinkage
On a paper machine the web is always subjected during drying to forces
which modify the amount of shrinkage that can take place: obvious
examples are the draw, which affects machine-direction shrinkage, and the
felt tension, which affects both machine- and cross-direction shrinkage.
S. F. Smith (2) was the first to recognize the importance of these restraining
forces and to investigate the effect of what he termed the 'dried-in strain',
the difference between the percentage shrinkage that would occur without
restraint and the percentage shrinkage in the same direction that actually
does take place on the machine. He measured first the percentage shrinkage
in the cross direction which occurred up to various positions in the drying
section of an M.G. board machine with pre-dryers and after dryers, and
compared this with the shrinkage that took place when samples taken
before the drying section were allowed to dry to the same moisture content
but with freedom to shrink; the extent to which shrinkage in the cross
direction had been prevented (or 'dried-in strain' developed) was thus
determined. Samples drawn from the same positions in the section were
then immersed in water and re-dried without restraint, the change in length
caused by doing this being measured. In every case wetting and re-drying
caused the initial cross-direction dimensions to shrink by an amount which
corresponded (but with lower magnitude) to the 'dried-in strain' that had
been induced up to tI:iat position in the dryers.
The importance of this work lay in the recognition that the restraint
conditions of drying affect the final condition of paper and that the process
is partially reversible by wetting out and re-drying the paper under different
conditions of tension. Other workers have since shown that restraint in
drying also affects the moisture and hygroexpansivity of the paper: in
general the less shrinkage is allowed to occur during drying, the lower are
398
the dimensional changes of the paper when afterwards immersed in water
or placed in a higher humidity.
These phenomena are readily explicable by Page and Tydeman's theory.
Restraint during drying will hold individual fibres in tension and prevent
the development of kinks in the free area of the fibres and of micro-
creping at the bond sites. This will cause the longitudinal shrinkage of
individual fibres, and hence of the sheet as a whole, to be less than would
occur under conditions of complete freedom. If the sheet is allowed to pick
up moisture or is immersed in water, the length of fibres will increase but
the magnitude of the increase must be smaller because there is less length
to recover. The greater the restraint during drying, the more this will apply.
However, once the sheet is re-wetted, apart from physical changes induced
in the fibres which prevent complete recovery of their original length, the
general sheet conditions are returned to their previous state. Drying
without restraint will then result in similar behaviour to that of a sheet
dried free to shrink which has not first been dried under restraint. If a sheet
dried under restraint is put through a number of humidity or wetting
cycles involving pick-up and loss of moisture with freedom to shrink, a
further, but gradually diminishing, amount of the shrinkage lost by the
initial drying restraint is commonly recovered at each cycle. Thus, some of
the physical changes induced by restraint during drying must also be
reversible given suitable conditions.
SA . 3 4 The effect of drying restraints on strength properties
Restraint during drying affects the strength properties of paper as well as
the dimensional stability. Most of the work on this subject has been done
in the laboratory and a useful summary of this, together with a report of
some recent detailed work of their own, is given by Gates and Kenworthy
(85). The only detailed report on the practical effects of changing restraint
conditions on an actual paper machine is described in the next section.
If tension is applied in any particular direction during drying, and
shrinkage in that direction thereby restrained, then it appears that the
tensile properties of the paper in the same direction are also affected in
such a way that the extensibility is reduced. In general, the less the shrink-
age that is allowed to occur, the lower the stretch of the paper and the
total energy required to bring about rupture, but the greater the tensile
strength.
According to Gates and Kenworthy the relationship between shrinkage
on the one hand and stretch and tensile strength at rupture on the other is
approximately linear; as regards stretch the relationship applies irrespec-
tive of the degree of fibre orientation in the sheet (which governs the
relative magnitude of directional tensile properties in the first place),
though the relationship between shrinkage and tensile strength is more
prominent as orientation increases. Anisotropy of stretch may thus be
attributed mainly to differences in drying restraints (primarily because of
the close association between stretch and the shrinkage that has occurred
in drying), though if the sheet has a high degree of fibre orientation, some
anisotropy would still exist even though tension during drying were even in
399
both directions. By contrast, anisotropy of tensile strength, though also

dependent on differences in the drying restraint applied, is mainly con-
trolled by the fibre orientation. The following table of ratios between
results in the machine and cross directions quoted by Gates and Kenworthy
illustrates this:
Draw Degree of anisotropy
Extension to
Tensile strength rupture
Random Orientation None 1·09 1·27
High 1-15 2·58
High Orientation None 2·05 1·89
High 2·34 2·99
These results emphasize the difficulties of making a sheet of paper that
is free from directional properties with regard to strength. As long as fibre
orientation exists, it is hardly possible to obtain tensile isotropy, though by
using very slack draws together with devices for restricting shrinkage in the
cross direction it appears possible that isotropy of stretch properties may
be obtainable.
A general explanation of these phenomena has been presented by Page
and Tydeman. In the first place, they consider that under the gradually
increasing load of the tensile test the kinks and the micro-compressions at
bond sites of individual fibres are gradually pulled out until a point is
reached where little further stretch is possible; greater load then comes to
bear on the bond sites themselves until the shear strength of the bonds is
exceeded and slippage follows with final rupture at the weakest spot. Thus,
the strong correlation between the shrinkage occurring during drying and
extensibility of the paper is accounted for by the fact that greater shrinkage
implies a higher proportion of kinking and micro-creping in the fibres
which are therefore able to stretch more. When shrinkage is inhibited, this
is at the expense of a weakening of bond strength; hence the energy required
to rupture paper.. is lower when less shrinkage has occurred during drying.
If dry paper is subjected to tension and then released, above a certain
tension a permanent extension or 'set' is found to have taken place. This
may be attributed to disruption of micro-compressed regions at bond sites
and of kinks that have been straightened out under the tension and do not
return to the same position. Over a long period some recovery of the lost
shrinkage occurs, i.e. the permanent set gradually diminishes, and also
according to Rance (4) 40 per cent. or so of the permanent set can be
removed by wetting and drying; these are further indications that physical
changes induced by tension in the sheet are partially reversible under
suitable conditions. In the draw at the calenders tension is applied in the
machine direction to paper which is dry, hence on release from the web
some permanent set may be expected to be already an inherent part of the
paper and this may well modify the dimensional stability and strength
characteristics. Rance argues from this that apart from shrinkage restraint
during drying, draw at the calenders must also affect the final properties of
the paper, though no direct confirmatory work has in fact been reported
on this.
400
SA. 3 5 Drying restraints on a paper machine
Arlov and Ivarsson (3) have reported some interesting results obtained by
varying the draws and felt tension of the Swedish Central Laboratory
experimental machine, and it is proposed now to describe these in some
detail and discuss their significance. In this work for each condition of
draw and felt tension the dryness of the web from cylinder to cylinder was
measured, together with dimensional changes in the machine and cross
directions. All the work was undertaken with constant wet-end conditions
and a machine speed between 160 and 220 f.p.m. Dimensional changes
were measured using a special device consisting of two cogged wheels set
at a fixed distance apart; these could be used to mark the web at any point
in the dryers, and the distance at the reel-up between the lines formed by
the two wheels, and between adjacent cogs on each wheel, then gave the
shrinkage that had occurred from that point in the dryers in the cross and
machine direction, respectively.
It was found in the first place that changing the draw and felt tension
had little effect on the drying curve due, it was presumed, to the good
ventilation on this particular machine. Normally it would be expected that
slack felt tension would reduce heat transfer, necessitating a higher steam
pressure and a different drying rate. The drying curve common to all the
tests is shown in Fig. 5. 9, together with the dimensional changes measured
down the machine in machine and cross directions.
These results show that a relatively high tension in the sheet, i.e. tight
draws, not only decreases shrinkage in the machine direction, but also
causes increased shrinkage in the cross direction (this has also been found
in several laboratory experiments). Tight felts reduce cross direction
shrinkage but appear to have an opposite effect on the machine direction.
Shrinkage of the sheet is small down to a moisture content of about 40 per
cent., and up to this position in the dryers extension of the sheet in the
machine direction is possible by tighter draws; this produces a correspond-
ing contraction of the sheet in the cross direction, which is also seen to be
particularly influenced by the draw in this region. Below 40 per cent. down
to about 15 per cent. moisture content, shrinkage of the sheet increases
sharply and in this region, although the influence is still present, different
draws and felt tensions have a decreasing effect possibly because the forces
of contraction against which they must act become relatively great.
General confirmation of this has come from laboratory simulation
experiments reported by Kenworthy (103), who found that the influence of
tension in drying is rapidly reduced as moisture content decreases from
about 55 per cent. down to 30 per cent. or so.
Many other workers have found in laboratory experiments that the
shrinkage rate increases below 30 per cent. to 50 per cent. moisture content.
It has also been reported that greater beating of the paper furnish, apart
from increasing the rate of shrinkage throughout the drying section,
causes the shrinkage rate to start increasing when the web is damper
though the transition is then less sharp between the two rates of shrinkage.
A more highly beaten paper shrinks more in the early part of the drying
section presumably because surface tension forces act on a more closely
401
connected assemblage of fibres which are, in turn, more flexible and

yielding.
The change in shrinkage rate appears to take place at about the position
in the dryers where the rate of drying begins to fall off (in Fig. 5. 9 this is
at the inflexion point of the drying curve, i.e. about 40 per cent. moisture
DR.YING se.<:.TIOM.
OP..A.W. F 1:1.:T.
l. TIC.I-IT. 51..ACK.
2. TIGH'T'. TtGHT
3.
.4.
ORA.W: HLT.
I. SLACK. TIGHT.
2. SLAC.K. SLACK.
3. T1'1Ki. TIGHT.
4. TIGHT. 5.LAC.K
,
2.
~·
""4·
Fig. 5. 9. Change in moisture content and dimensional changes in the cross and
machine directions of a web passing along the drying section under different conditions
(after Arlov and Ivarsson)
content). It is reasonable to suppose that the change is closely associated

with completion of the removal of 'free' water from the web and the com-
mencement of evaporation of water within the fibres themselves. The basic
cause of shrinkage thus changes at this point from the relatively small
contractive forces produced by surface tension to the much more substantial
forces resulting from the reduction in length of individual fibres.
It will be noted particularly that a large decrease in machine-direction
shrinkage occurs at the calenders, to such an extent that in every case the
402
overall dimensional change in this direction amounts to a stretch rather
than a contraction in the sheet length. This bears out Ranee's contention
that the magnitude of the calender draw can be very important in deciding
the final properties of the paper.
The effect of drying on the mechanical properties of the paper was also
investigated by Arlov and Ivarsson under conditions when the drying rate
on the machine was adjusted to be as linear as possible all the way down
the section; samples were extracted at various positions down the dryers
s
LEA.\l\t-1 <:i
O~YE~S 1"1°18.
i.N'tiRlNGr
ORYllRS.
~ < 10
0/
/O·
Fig. 5 .10. Machine direction stress-strain curves of paper extracted from different
parts of the drying section (after Arlov and Ivarsson)
and were then allowed to complete their drying with freedom to shrink.
Figure 5 .10 shows the change produced in machine direction stress-strain
curves obtained with a special tensile tester; the results bear out the
observations that the progressive effect of tension in the sheet during
drying is to reduce shrinkage and subsequent stretch, but increase the
tensile strength required to break the paper.
SA. 3 6 Controlling shrinkage on the machine
For many types of paper it is important to be able to obtain some desired
degree of dimensional stability, strength isotropy, or other special
properties such as maximum extensibility in one or other direction. The
role of fibre orientation in the sheet has an important bearing in such cases,
for example it has already been seen that tensile strength isotropy is largely
dependent on the degree of fibre orientation, but in the present context it is
proposed to discuss only the influence of drying. In particular, it is most
useful to be able to control separately the shrinkage taking place in
machine and cross directions in the dryers.
Taking first machine-direction shrinkage, the results obtained by Arlov
and Ivarsson indicate that this can be regulated to some extent by the felt
tension, but the most important and controllable influence is the sheet
403
SA.3 6 THE DR YING SECTION AND CALENDERS
tension in the draws. Increasing the sheet tension on a machine within the
practical limits demanded by smooth running may be expected to produce
a decreased extensibility and less dimensional instability in the machine
direction, though it also appears to induce the opposite effect in cross-
direction properties. This applies particularly to the early part of the
drying section although often there is only one position at this stage where
a draw can be applied so the scope for manipulation is limited. When the
expense is justified, greater flexibility could no doubt be obtained by
dividing up the early cylinders into a number of individual drives each, of
course, with its own pair of dry felts.
Shrinkage is also affected on some machines by the practice of gradually
reducing the diameter of the last few cylinders in an endeavour to maintain
as near as possible the same tension in the sheet in the region where most of
the shrinkage takes place. But this is obviously a permanent arrangement
and is only suitable where there are no substantial differences in the
shrinkage pattern, such as occur between different grades of paper and
under different beating conditions, for papers produced on the machine in
question.
Shrinkage in the cross direction, which may be as little as 2 per cent. but
with heavy, well-beaten papers can amount to 9 or 10 per cent., is less easy
to regulate on most machines; but it is particularly important to keep a
check on this, if only to prevent shrinkage becoming excessive and affecting
production by narrowing too much the deckle at the reel-up. It has been
seen that tight felts help to reduce cross-direction shrinkage; also slackening
the machine-direction sheet tension, though possibly not restricting
shrinkage in the cross direction, will prevent its increase at the expense of
tight draws. But the use of felt tension to regulate cross-machine shrinkage
of the paper carries with it the objection that the drying conditions with
respect to heat transfer and steam consumption may thereby be affected,
and anyway on many machines it is not practicable from the point of view
of running the felt to alter the tension to any significant degree.
A device known as the 'textile stenter' has been adapted to grip the edge
of the paper web and restrict excessive web shrinkage when the web is
unsupported in a free draw, and details have also been given of equipment
consisting of a rubber belt which is pneumatically pressed and drawn across
the paper as it passes over a drying cylinder, thereby spreading the web
across the machine as it follows the belt (120). But in recent years a more
flexible method of regulating cross-direction shrinkage has appeared, the
expander roll with fixed or variable bow. Gallahue (73) and Fahey and
Chilson (84) have given details of the use of this device, and claim that
shrinkage in the cross direction is reduced and also (as would be expected)
there is an improvement in cross-direction tensile strength and dimensional
stability. Expander rolls have been used in several positions and appear
most effective in the early part of the drying section (where, it may be
noted, draw has also been found more effective for altering machine
direction shrinkage). No 'noticeable' effect on machine-direction shrinkage
is reported, nor any 'appreciable' effect on tensile strength in the machine
direction.
404
It is also possible with expander rolls to bring shrinkage at the edge of
the sheet under better control by adjusting relative web tension. This is of
particular importance because shrinkage of the paper up to a few inches
from either edge is generally affected by differences in tension during
drying (due to the greater freedom to shrink cross-wise, to differences in
substance, and to differences in drying rate which affect the eventual
shrinkage, as discussed below); in extreme cases very prominent differences
in the dimensional stability and strength characteristics of the web at the
edges can occur.
SA. 3 7 Cockling
Cockling of paper when cut into sheet form, and certain forms of distortion
in reel form, are also due basically to differences in tension set up during
drying whenever the drying rate is different in one portion of the web to
adjacent positions. Suppose, for example, that drying is slower for some
reason in one region of the web. The surrounding area, because it dries
quicker, will begin to shrink earlier and will consequently exert a tension
on the damper region. This will cause the damper region to dry with a
lower amount of shrinkage and when the web is finally dried to the same
moisture content this region has become stretched to a slightly greater
area.
Distortions of the machine reel often emanate from differential shrinkage
and this may originate in cross-machine substance variations: a persistent
narrow heavy streak will produce a wrinkle running round the reel the
prominence of which will depend on the winding tension. Normally
though, the causes of reel distortions are complicated by associated
moisture and calendering differences, a topic discussed further in SC. 5 4
and elsewhere.
When cut into sheets, a region of the paper which has been stretched
will exhibit itself as a bulge; the bulge may at first be hardly noticeable,
particularly if the sheet comes from a tightly wound reel, but once the
paper relaxes in sheet form the familiar appearance of cockling becomes
more apparent. It is in this form that distortions due to differences in
tension set up during drying are most commonly encountered. In certain
cases, as the sheet rapidly picks up moisture from the atmosphere the
lower moisture expansivity of the stretched area of the sheet may gradually
reduce the prominence of the cockle.
Cockling produced by cross-machine substance unevenness shows up
very prominently in sheets stacked off the cutter and in this respect, as in
the case of similar faults in a reel, the slightly heavier substance of the
stretched portion is generally associated with slightly greater thickness
which therefore accentuates the trouble. Differences in moisture content
across the web entering the dryers, caused for example by a ridge in the
wire or a plugged wet felt, can produce this same effect when they are
severe and confined to a relatively narrow width. Likewise differences in
tension of the dry felt as a result of being stretched on some occasion and
very bad unevenness in cylinder surfaces produced by a build-up of scale
and dirt can both have a similar effect on the paper. A poor breast box and
14 405
5A.3 8 THE DR YING SECTION AND CALENDERS
slice can deposit relatively large areas of stock which are over- or under-
weight and this also can produce cockling, though in a less systematic
manner.
SA. 3 8 Curl; the fundamental mechanism
The subject of curl, in so far as it is associated with two-sidedness and
fibre orientation during formation on the wire, has already been con-
sidered in some detail. In this section the added complexities of the
influence of drying on curl are dealt with. An understanding of how the
conditions of drying come to affect curl in the finished paper has gradually
emerged over the years as a development from observations of the effect
of drying restraints on paper, as described in the previous section. The
foremost workers in this field, from whose reports the following is derived,
are Brecht et al. (14, 21, 42, 70), Glynn et al. (52, 74), and Newman et al.
(28, 86).
The development of curl in a sheet of paper depends in the first instance
on differences in the unstressed dimensions of the two surfaces. For
instance, if the top side would take up when completely relaxed a larger
area than the wire side, then the sheet will tend to curl towards the wire
side. The axis of curl depends primarily on the relative difference in
relaxed length of the two sides in the machine and cross direction; the
magnitude of curl depends on the relation between the extent of these
differences and the rigidity of the sheet, i.e. on the balance between the
stress towards bending set up by the differences in dimension of the two
sides of the sheet related to the resistance to bending inherent in the
structure of the sheet. Both axis and magnitude of curl thus depend on an
interaction of stresses in the sheet; a 20 in. x 30 in. sheet may curl in one
direction while two 20 in. x 15 in. sheets cut from it may curl in the
opposite direction.
This fundamental conception of the mechanism of curl has been directly
demonstrated by Glynn and Gallay (74). In this work a laminated paper
prepared by couching together two standard handsheets was subjected to
non-uniform drying on the two surfaces and the curl developed was
measured. The two sections were then separated and their equilibrium
lengths measured: a close correlation was shown to exist between the
differences in unstressed length of the two sides and the magnitude of curl.
Other work which illustrates the dependence of curl on differences in
the unstressed dimension of the two sides is reported by Brecht et al. (42).
In these experiments one surface of a sheet of paper was insulated from the
other by a specially devised container, and the relative moisture contents
(and, hence, the natural dimensions) of the two surfaces were then altered
either by blowing dry air on one side or maintaining the air in contact with
two surfaces at different relative humidites. Whenever the difference (and
hence unstressed area) between the equilibrium moisture content of the
two sides was increased, e.g. by increasing the difference in relative humid-
ity between the two sides, then the curl observed was greater. The rate at
which the sheet curled also appeared to depend primarily on the magnitude
of the moisture gradient that was set up through the sheet: increasing the
406
velocity with which dry air was blown onto the sheet, or the difference in
humidity between the two sides, increased the rate of development of curl.
Also, thin paper curled faster than thick paper in comparable conditions,
presumably due to the lower resistance to bending, although the final
degree of curl once equilibrium was established was little different in the
two cases (this may be attributed to the lower difference for the thin sheet
in final moisture content of the two sides, associated with the lower
rigidity). Sheets made from beaten pulps always gave a greater degree of
curl in comparable conditions, a result that is expected due to the overall
higher moisture expansivity.
SA. 3 9 Curl; the effect of changes in atmospheric conditions

These experiments illustrate the basic mechanism of curl and the inter-
action of forces that comes into play. Under normal conditions, of course,
both sides of a sheet of paper are subjected to the same atmospheric
conditions, but even so changes in humidity will affect the degree of curl.
For this to happen it is only necessary for the moisture expansivity of one
side of the sheet to differ from the other, as is the case whenever there is
appropriate two-sidedness in the constituents of the sheet. Change in curl
then takes place towards the side having lower moisture expansivity
when the humidity is increased, and away from that side when humidity is
decreased. A perfectly homogeneous sheet does not curl in the presence of
humidity changes; Brecht illustrated this by gluing strips of paper together
to form a sheet having two identical surfaces.
From this it follows that anything which affects the relative moisture
expansivity of the two sides of a sheet will alter the behaviour of the paper
under a change of atmospheric conditions. The axis of the curl, though
mainly determined by the relative degrees of fibre orientation in the two
principal directions, will also be affected by the restraint imposed on drying.
If shrinkage is prevented in one direction during drying, this will reduce
the moisture expansivity in that direction and thereby create a tendency
for curl also to take place in the same direction because the greater
expansivity is at right angles. Such an effect would be independent of fibre
orientation, though of course any orientation in the sheet would affect the
relative degree of expansivity in the first place.
To summarize, although a sheet of paper may lie fiat in one particular
atmosphere, curl can occur if the sheet is placed in a more moist or dry
atmosphere due essentially to differences in moisture expansivity of the
two sides of the sheet. The magnitude of the curl depends on the internal
stress caused by the differences in expansivity in relation to the rigidity of
the sheet; the axis of curl depends on the relative differences in expansivity
in the two principal directions which in turn are governed by fibre orienta-
tion and the restraints exerted in drying. To this extent, it may be noted,
altering draws and felt tension can affect curl by changing the potential
expansivity in one or other direction. In practice, however, it is not easy to
decide how to alter curl by this means because the effect depends not on
the magnitude of shrinkage occurring in any particular direction, but on
407
the difference in expansivity on the two sides of the sheet in the one
direction compared to the other.
5A. 3 10 Curl; direct changes induced in drying
Attention has so far been focused on the curl which may be induced by
atmospheric changes. It will be readily appreciated that the same factors
are responsible for any inherent curl of the paper as produced on the
machine and in practice it is usually this source of curl which is the more
prominent in Fourdrinier papers. By 'inherent curl' is meant simply the
curl that a sheet of the paper would assume immediately it were released
from the web but before any change occurred in the moisture content.
Although in practice it is not easy to distinguish between such inherent
curl and the curl induced by changes in moisture content of the sheet as it
reaches equilibrium in the atmosphere, it is important nonetheless to be
clear that there are two sources.
Inherent curl is a product of the internal stresses set up through the
thickness of the sheet due to differences between the two sides in shrinkage
potential during drying. For example, if by virtue of its composition the
top side of the sheet has a natural overall shrinkage during drying which
is greater than the wire side, then the top side of the sheet will try to assume
a smaller area. The structure of the sheet will prevent this actually occur-
ring and instead a degree of stress is set up within the sheet in which the
top side is restrained from shrinking and is in tension while the wire side
undergoes a certain amount of compression. On release from the web
these stresses relax to create, in this case, a curling tendency towards the
top side.
It is possible during the drying process to alter the stresses set up through
the thickness of the sheet and, thereby, the inherent curl in the paper.
Glynn et al. (52) have reported several interesting experiments which
demonstrate this. Firstly, sheets dried free to shrink under an infra-ray
lamp always curled away from the heat source irrespective of whether the
top or wire side faced the heat. When dried with one side in contact with a
metal plate, the sheet curled towards the contact surface irrespective of
which side heat was applied to, i.e. through the plate or on the free surface
of the sheet. Finally, with handsheets dried on a curved dryer and the
sheet in contact in various combinations with the cylinder, a felt or wire
screen, or sandwiched between felts and/or screen, the direction of curl
was shown always to be towards the side from which moisture was
removed last. The same occurred with various types of machine-made
paper which were re-wetted and then dried under similar conditions. The
hotter the surface of the curved dryer, the greater the curl induced;
increased beating increased the curl but size, alum and loading had little
effect.
These results are explicable only in terms of the internal stresses set up
in the sheet during drying. In all cases the curl is towards the surface
which dries last: the side away from the heat source in the first experiment,
the side facing the metal surface (which would prevent moisture removal
from that side) in the second experiment. The side drying first will tend to
408
THE DR YING SECTION AND CALENDERS SA.3 11
shrink earlier and will therefore be continually subjected to more restraint
than the opposite side; only when drying of the first side is completed
and shrinkage ceases is the opposite side prevented from shrinkage any
more and an increasing restraint put upon it. With drying completed
the side of the sheet dried first will have tended to shrink less, i.e. have
a greater natural area, than the opposite side from which moisture
is removed last, so that the sheet will be subjected to a stress which on
relaxation will produce a curl away from the side dried first. The greater
the tendency for one side to shrink earlier than the other, by quicker drying
or by the sheet having greater potential shrinkage, the stronger is the curl
induced by this means. In any particular case, the magnitude of curl will
depend on the moisture difference set up during drying between the two
sides of the sheet in relation to the thickness, i.e. to the moisture gradient
through the sheet.
SA. 3 11 Correction of curl
It is evident from this explanation that a means is available for correcting
curl induced by the two-sided structure of the paper. To return to the
example of the sheet which by virtue of its composition has a greater
shrinkage potential on the top side, it was shown that under unrestrained
drying this would tend to curl towards the top side. If now drying is
performed in such a way that the top side is dfied quicker than the wire
side, a counteracting curling force will be produced. The tendency for the
top side to assume a smaller area by virtue of its desire to shrink more is
counteracted by removing moisture earlier from the top side and thereby
subjecting it to greater restraint during drying and increasing the natural
area.
This effect has been demonstrated on a paper machine by both Brecht
ef al. (42) and Hendry and Newman (86). However, difficulties arise in its
use as a means of correction because of a difference in behaviour in the two
principal directions. Suitable correction by differential drying of a greater
potential shrinkage on one side of the paper in the machine direction is
unlikely to be exactly suitable for correcting the same difference in the
cross direction.
Hendry and Newman describe running a machine with the top cylinders
hotter (200 deg. F. as against 145 deg. F.) than the bottom, then vice
versa. Curl in strips cut with axis parallel to the machine direction (a
measure of curl with axis in the cross direction) was unaltered, while curl
in cross-direction strips altered considerably in such a way that the curl
was always away from the side contacting the hotter cylinders, i.e. the side
dried earlier.
For all the knowledge of the sources and behaviour of curl that is now
available, it will be evident from the sheer complexity of the situation that
a solution to any individual curl problem presents considerable difficulties.
A thorough analysis of the causes of curl in one particular paper on a
machine can be extremely time-consuming and it is difficult to obtain
consistent results. Certainly differential top and bottom drying in the last
bank of cylinders provides an invaluable aid to the machine crew provided
409
a systematic and standardized procedure is laid down for assessing the

effects of changing the differential in terms of the resulting change in curl.
A sweat roll can also have some effect on curl because moisture is added to
one side of the sheet before calendering, but this is not so controllable. In
practice differential drying together with careful manipulation of draws
and felt tension can enable curling tendencies to be kept under control, but
considerable care and experience is required.
SA. 3 12 Alteration of paper properties with atmospheric changes

To complete this section, brief mention wiii now be made of the effect that
changes in atmospheric conditions have on the properties of paper. The
moisture content after conditioning in the standard atmosphere of 65 per
cent. relative humidity and 68 deg. F. varies considerably from one type of
paper to the next, being as low as 5 per cent. for bible tissue and over 10
per cent. for mechanical printing. According to Brecht (41) the equilibrium
moisture content depends on the hemicellulose and lignin content of the
paper, the presence of filler (which decreases the moisture content as its
proportion is increased), but not so much on beating.
The moisture content of paper in any atmospheric condition is closely
dependent on the relative humidity and actually depends on whether
equilibrium is approached from a higher or lower initial moisture content.
Fig. 5. 11 is a typical curve illustrating the behaviour of moisture content
as relative humidity is varied; much detailed work on this subject dealing
with a whole range of papers has been published by Crook and Bennett
(71). The time required to achieve final equilibrium can be long, a matter
of many hours, though blowing air on the paper hastens this appreciably.
However, when paper is first exposed to the atmosphere on extraction
from a reel the initial rate of change is rapid; newsprint taken off the reel
at 6 per cent. moisture content picks up a further 1 per cent. in under a
minute. More will be said about this, and the problems it presents with
regard to testing the moisture content of paper in a reel, in later sections.
Various properties of paper are closely affected by the moisture content.
Fold, stretch, tear, and thickness increase in value at higher moisture
contents, while tensile strength, air resistance, and stiffness decrease; burst
shows a maximum and begins to drop off sharply above 6 to 7 per cent.
moisture content. Fold is most affected by changes in moisture content,
followed by stretch, tear, stiffness, tensile strength, burst, air resistance,
and thickness. Comprehensive data on this subject can be found in the
same report by Crook and Bennett.
Other results have been reported by Brecht (41) who found that smooth-
ness decreases with increasing moisture pick-up. This work will be referred
to in greater detail when discussing the effects of calendering. A further
observation by this author is that static electric charge drops rapidly on a
damp reel that is placed in dry air, but a dry reel in damp air keeps its
charge indefinitely. Thus, the higher the moisture content of the paper the
less static it is likely to carry. Brecht also states that less dust is present in a
reel when moisture content is higher.
Temperature, irrespective of humidity differences, has only a small effect
410
on the properties of paper and is hardly significant in relation to the
normal changes in indoor temperature which occur in this country,
certainly in comparison with ambient humidity changes. Fold is most
affected by temperature changes and to a lesser extent strength tests; a
14
12
le
i 8
I-
z
UJ
I-
z
0
tJ
6
w
et
:::l
~
~
0
~ 4-
RE.LAT l\lE 1-lUMiOiTY ~-~.
Fig. 5. 11. Typical curve showing the change in moisture content of paper caused by
increasing and then decreasing relative humidity
15 deg. F. rise in temperature is reported to reduce tensile strength by about

5 per cent. and burst strength by 21 per cent. The effect of temperature on
moisture content, thickness, air resistance, and tear is negligible.
SA. 4 CALENDERING
At a really fundamental level relatively little is known of the way in which
calendering affects paper. Some consider that damage is done to fibre
411
bonds on the surface of the paper and this reduces the resistance of the
outer layers of fibres to vertical stresses such as are applied during printing;
fibres are thus more easily lifted up from the surface of the paper or
loosened completely from the sheet to create dust and fuzz. The predomin-
ant effect of the calenders is on the paper surface and a comparison before
and after calendering (such as in the light micrographs on supercalendered
paper presented by Emerton (72)) shows up well the change that occurs.
Individual fibres which are initially sticking out proud from the surface
are seen to be flattened into close contact with the sheet, while local con-
centrations offines and filler (mainly on the top side) become merged in the
calendered sheet into a smooth highly-glazed area. Photographs of micro-
tome cross-sections taken through paper before and after calendering
reveal that a reduction in the inter-fibre void space occurs and that the
lumens of mechanical wood fibres collapse (107, 127).
The effect of the calenders is produced by the heavy compressive force to
which the paper is subjected, combined with a certain amount of sliding or
rubbing action of the calendar rolls on the paper surface. The process
appears to some extent to be reversible: Farebrother, in the discussion on
the above paper by Emerton, presented photomicrographs of the same
area of paper which show that the general appearance of the sheet after
calendering (on a laboratory scale) is largely restored to the uncalendered
state by soaking in water and re-drying. On the other hand, bulk and
smoothness showed appreciable residual effects due to the calendering
and these are the properties most affected in the first place, alteration of
bulk and smoothness and reduction of variations in thickness across the
web being, of course, the principal purposes for which calenders are used.
What follows represents a summary of the present state of knowledge of
calendering. It is confined largely to describing the results of investigations
into the effect of calendering on the general properties of paper.
SA. 4 1 Change in paper characteristics down a calender stack
Experimental work designed to find out how the properties of paper change
as it progresses down a stack of calenders have been reported by Howe and
Lambert (61), Wultsch (68), Blanchard et al. (82), and Mardon et al.
(106, 126). There are many practical difficulties involved in obtaining
suitable samples for testing as the paper emerges from the different nips
under normal running conditions, and the most convenient method
appears to be to cut a strip at one edge and catch or blow it off at the
positions required, working progressively back up the stack. The rest of the
sheet runs through the stack while the strip is sampled and to prevent a
break it may then be necessary to widen the sheet back to full width very
quickly.
In the work reported by Howe and Lambert, sampling from an 18 in.
strip passing alone down the calenders was also tried and gave similar
results for thickness and smoothness changes to those obtained with the
full sheet. It is presumed that the rolls deflected sufficiently to prevent the
full weight of the stack falling on the strip and, if this is generally the case,
using only a strip could present an alternative means of experimenting,
412
although the possibility of a variable bias being introduced makes it
preferable whenever practicable in studies of this nature to extract samples
with the full sheet running in the stack.
It is interesting to remark at this stage that the change in properties
down the stack with only the strip running was more pronounced when
passing through at a crawl than at full speed. Under crawl conditions it is
to be expected that less distortion of the rolls within the nips occurs, and so
there is less likelihood of any relative motion between rolls and paper which
could contribute a smoothing effect; but the main difference from operation
at full-speed is that pressure is exerted on the paper at each nip for a longer
time interval. This is one piece of evidence which indicates the importance
of a straightforward compressive action in calendering.
Some of the results obtained by Howe and Lambert on a high-speed
newsprint stack are given in Fig. 5 .12 which illustrates the changes that
occurred in thickness, porosity, and smoothness (top and wire side using a
modified Bendtsen apparatus-note that lower readings indicate smoother
paper, the Bendtsen test strictly speaking measures roughness). The
general shape of each of these curves, the declining effect of successive nips
on thickness and the almost linear relationship (certainly as regards
percentage change at each nip) for smoothness and porosity, has been
confirmed by the results of both Blanchard (on five different stacks) and
Wultsch (on a machine making printing paper), though with fine papers
Mardon found that reduction of thickness in successive nips appeared on a
number of machines to be linear; similar changes in paper properties have
also been noted in work on supercalenders. Differences between top and
wire side smoothness remain unchanged through the stack.
While thickness, smoothness and porosity are the properties most
affected by the calender stack, other observations have shown that gloss
increases, and compressibility, stretch, and tear all decrease at a fairly
uniform rate from nip to nip. Printability as assessed on a proof press
increases fairly steadily (from which it appears that a measure of smooth-
ness can be used to give an indication of changes in potential printability).
Blanchard remarked that woodfree papers are much less affected by the
calenders than newsprint and considered that furnish has an important
effect. This has been generally borne out in the later work of Mardon, with
the additional observation that wood-free papers are relatively more
affected in later nips than the earlier nips of a stack.
SA . 4 2 Changes in smoothness and thickness on removal from the reel

It is unfortunate that few of the reports quoted above give any details of
moisture in the paper. Moisture content is important not only for its direct
influence on the effect of the calenders (see next section), but because the
time elapsing before testing, and the change in moisiure content as the
paper approaches equilibrium in the atmosphere, both have an appreciable
effect on its measured properties, and particularly on thickness and
smoothness. There are obvious difficulties in ensuring that samples are
tested in the condition they leave the calender, but the apparent absence of
specific precautions to minimize variation from these sources detracts
413
SA.42 THE DR YING SECTION AND CALENDERS
G
j
0
::t
t-
I/)
- 5
II)
tU
~
:J.
~ 4
:t
....
I-
.u
uJ
:r
lfl .3
~
't:
II.I
"1
....
0
r
tU
d)
'--'
150
"'
111
UJ
z
:a::
t-
()
0
~
Ill
5o
,,........
~
v
\)
0
3J
~
~
>- 20
t:
tll
0
ex
0 lo
0.
ENi~RlNG. 2 4 5 co.
N1PS •·
Fig. 5. 12. Change in thickness, smoothness and porosity of newsprint passing down a
stack of calenders (after Howe and Lambert)
414
somewhat from the value of the results (there is no mention of protecting
samples from the air as much as possible before testing, though Mardon
conditioned them for a fixed period to a standard atmosphere which
eliminates one source of variation).
Although changes in paper properties after extraction from the reel
(especially in moisture content, thickness and smoothness) are well
recognized, in fact there is not a great deal of information available
showing how rapidly and to what extent such changes take place. The only
z
~ 200
0
z
175
-!
1/1
\/\ 150
~
;::Cl 125
0
<: loo
ll'l
:;j ?>·4
Q
:!t'
I- 3·2
o11'
I/I
....% 3·o
:::L
u z..g
x
t- '2·~
lo·o
x
~ 9·o
w
Oi'. 8·0
:::J
I-
dl
(i
~
7·o
<D ·0 I
,____._ ___.._ __.__-.d-_ __.__ __..__;\r--·~
0 lo zo .;o 4o So e.o ~ OAY-5.

TIME- IN MINUTES.
Fig. 5. 13. Changes occurring in a sheet of newsprint after extraction from within a reel
really detailed report dealing with the magnitude of change under different
conditions is due to Brecht and Heyn (33), but before examining this work
it is interesting to consider the manner in which the characteristics of paper
can alter in the common case where extraction from the reel brings about
an increase in moisture content to equilibrium in atmosphere.
Fig. 5. 13 illustrates typical changes occurring in newsprint immediately
after it is removed from inside a reel and freely allowed to approach
equilibrium in standard atmospheric conditions. Over a period of three
days the smoothness decreases from 100 to 185 Bendtsen, and the thickness
415
increases from 2·75 thou. to 3·25 thou. The moisture curve is also shown
and attention is drawn particularly to the rapid change in the first minute.
These curves show that the moisture content of the paper is within
0·5 per cent. moisture of equilibrium after a little over 10 minutes, while
after one hour it has virtually stabilized and after three days little further
increase has occurred. In contrast both smoothness and thickness show a
substantial change between their values after one hour and after three days.
It may be strongly suspected from these results that the initial change in
smoothness and thickness on extraction from a reel is due primarily to
pick-up of moisture, but that some further alteration continues even after
the moisture content has virtually reached equilibrium. This later and
more gradual change may be attributed to a gradual visco-elastic relaxation
of the sheet from the compressional forces of calendering, a process which
can be expected to produce an increase in the thickness and a greater
unevenness of the surface as the stresses in individual fibres are relieved
and micro-movement occurs.
Though dependent in this way on rheological factors, the magnitude of
the change occurring in smoothness and thickness is nevertheless closely
associated with the degree of moisture change occurring in the sheet.
Brecht and Heyn have obtained substantial experimental data on this by
measuring the change after calendering in moisture content and smooth-
ness under a variety of conditions. Their work was done on a laboratory
apparatus using two smooth steel discs in a hydraulic press, but 'meticulous
studies' indicated that by employing corresponding pressures, the results
obtained with this static method are similar both with regard to calendering
effect and to change in properties afterwards to those obtained with
normal calenders. Three different papers were examined and in each case
the effect of calendering at different moisture contents was first determined
and then the paper was allowed to condition in three different relative
humidities, the change in smoothness and moisture content being measured
at appropriate intervals up to 24 hours later.
Considering paper made with a standard newsprint furnish, the following
table adapted from the graphs and data available illustrates the changes
which occurred in the smoothness values on conditioning to 65 per cent.
relative humidity:
Moisture Content per cent. Smoothness (Bendtsen)
After After 24 hrs. After After 24 hrs.
calendering conditioning calendering conditioning
5·3 10·2 180 350
9·3 10·3 100 160
13·4 11 ·3 56 76
23-3 11-8 86 185
These figures combined with similar ones obtained when conditioning to

35 per cent. and to 90 per cent. relative humidity (giving a lower and a
higher equilibrium moisture content respectively) as well as for the other
two types of paper, indicate that the bigger the change in moisture content
the greater is the drop in smoothness. This occurs whether a rise takes
place in the moisture content (as normally occurs), or whether the paper
416
loses moisture on conditioning, although the overall results show that a
much lower proportional drop in smoothness takes place when the moisture
content reduces. As an example, paper calendered at 13·4 per cent. and
conditioned to 90 per cent. R.H. showed a rise in moisture content by
2·1 per cent. and a drop in Bendtsen smoothness from 56 to 120; this may
be compared with the drop from 56 Bendtsen to only 76 for a decrease of
2·1 per cent. in moisture content, as shown in the table above.
When the moisture content after conditioning closely matches the
moisture content after calendering and only a very small change in
moisture content occurs then the loss in smoothness is undoubtedly at a
minimum. But even in such cases the results of Brecht and Heyn show
that an appreciable reduction in smoothness still takes place and this is a
further indication of the existence of a visco-elastic relaxation of com-
pressional stresses.
One final point worth mentioning is an observation made by Wultsch
(31). This worker reported a large drop in smoothness two hours after
normal calendering of a woodfree paper. Even when the same paper
received a second moistening and calendering, though this naturally gave
greater smoothness off the machine, it did not affect the change in smooth-
ness afterwards, and a similar drop occurred. But if the calender rolls in
the second stack were heated, practically no change in smoothness
occurred afterwards, although there was at least the same increase in
moisture content taking place during the two hour interval.
SA. 4 3 Effect of moisture content on calendering
Apart from the dependence of the final smoothness of paper on the change
in moisture content taking place after calendering, the initial effect of the
calenders is also very closely dependent on the moisture content of the
paper. There is unfortunately not a great deal of data to present on this
aspect of calendering, despite the fact that the calenders are well known to
be sensitive to changes in moisture content.
Laboratory work reported by Jackson and Ekstrom (104, 123) has
confirmed that up to the 12 per cent. moisture region paper is more com-
pressible at higher moisture contents and so would achieve a greater
smoothness after calendering. For newsprint, results have been obtained
which show that improved surface smoothness, gloss, strength and
printability are all achieved by calendering at a higher moisture content.
Similar improvements appear when the temperature of the paper web is
higher and there are also indications from their work that a temperature of
70 deg. to 75 deg. C. is best for obtaining a high smoothness in calendering;
normal operating temperatures of calenders are 10 deg. to 15 deg. C.
below this level.
'Blackening' of paper in the calenders, which is due to excessive com-
pression damaging the structure of the paper and creating an optical
distortion giving the appearance of greater transparency, is a familiar fault
that is caused by having the moisture content too high.
The work of Brecht and Heyn has provided clear evidence of the
existence of a moisture content at which calenders have a maximum
417
smoothing effect. Figure 5 .14 is adapted from their data and the lowest
curve illustrates the effect on smoothness of varying the moisture content
during calendering; the paper was made from a newsprint furnish and
some of the results have already appeared in the table on page 416. A
distinct maximum for the initial smoothness immediately after calendering
occurs at about 15 per cent. moisture.
The effects of conditioning to 35 per cent., 65 per cent. and 90 per cent.
relative humidity are shown in the upper curves which represent the values
to which the initial smoothness changes for the three different equilibrium
- - - ~MOOTtiNE.SS O:R~<:Tl..Y A!!ff'.R CA;.SND.,RlNG.

---&- SMOOill"1ESS t>.Hi;F!. 24 HOURS CONOIT10N!Nl3 /..."I
$00 ~5%!<.r<.
- X - - $MOOT~l'(E:;s .\F-if:R.24 HOUR~ C.ONOITIO!'liNG. Ai
c::.S%R.H
~ $t.\O!!TIHlcS$ AFHR 241-\oU;(S CCNOITIONINa. A.I
510% R,.ft
l'IDlf.; THE; POi~i MA.RV..<;;, DO l.iOT l't&Fli,,_ TO
E>tl?ERIN\S!>ITAL RESUL"i'S AND ~ERl/e ONLY
TO CH~TINGUISH IHE: LINlJ.5.
0 '-----'----~----'="'--~----=---------\....
S lo 15 Zo 2S 3.o 35
Fig. 5 .14. Variation of smoothness directly after calendering at different moisture

contents, and the corresponding smoothness after conditioning for 24 hours at three
different relative humidities (after Brecht and Heyn)
conditions (the moisture content scale does not apply to these curves). The
maximum point is still as evident in each curve but it has shifted position
due to the different changes in moisture content occurring before equili-
brium is reached in each case: for maximum smoothness in a 35 per cent.
humidity atmosphere calendering is best done in this case at about 13 per
cent. moisture content, for 65 per cent. humidity at about 15 per cent.
moisture, and for 90 per cent. humidity at about 18 per cent. moisture
content.
Though the general level of smoothness was different, similar results
were obtained for two other types of paper made from 100 per cent.
bleached poplar sulphite pulp and 100 per cent. bleached spruce sulphite
pulp. Maximum initial smoothness also occurred in each case at about
15 per cent. moisture content, but the maximum smoothness points after
418
conditioning to the different humidity levels, though following the same
sequence, corresponded to a lower initial moisture content in the case of
the first pulp and higher in the case of the second.
In the normal course of operation on the vast majority of machines there
is no possibility of calendering at moisture contents above 6 or 7 per cent.
So the main point of interest in these results from a practical aspect is that
higher moisture contents at the calenders can generally be expected to
produce a greater smoothing effect. But there is another aspect to this
which must be taken into account.
During any particular making the natural variation in smoothness of
paper at the reel-up that occurs from one moment to the next due to
MOIST4.l~i RSGIOI'{.
150 S·S % To co·o%
140
?
"'Ill
bl!
:D l~o
'-"
"'
U)
~ 120
:r
t-
<l
()
:ii;
II) Ito
k10 ,____ ___.__ _ _...___ ___.__ _ _.....__ _~--
2·7o 2·75 2·8o 2·85 2·9o 2·95

THICl'NE&S. 11-<0IJ.
Fig. 5.15. Regression lines showing the dependence of smoothness on the thickness of
newsprint reeled at different moisture contents
differences in substance, fibre composition, formation, and so on is closely

associated for given calender conditions with the thickness: in other words,
when for some reason the smoothness level alters, the thickness will usually
also be found to have altered. This association stems from the close relation
between the compressive effect on the paper at any time and the resulting
smoothness.
The precise relation between smoothness and thickness occurring in
calendered paper also appears to depend on the moisture content. This is
shown in Fig. 5 .15, which depicts four regression lines calculated from
quality control test results obtained during normal production of news-
print. Each line relates thickness (corrected proportionately for changes in
substance and therefore, strictly speaking, bulk) to Bendtsen smoothness
419
for results obtained when the moisture content fell within a particular
range. (Moisture content was measured as quickly as possible with a
portable meter but would be 1 per cent. to 1·5 per cent. higher than the
moisture content at the calenders.)
It will be observed from this graph that for a given moisture region an
increase in smoothness will on average only occur at the expense of a loss
of thickness. This relationship is well known and emphasises the need for
compromise when making printing papers for which both smoothness and
thickness (or bulk) are in general preferred as high as possible at the same
time. However, the point to note is that the relationship between smooth-
ness and thickness at higher moisture contents is such as to cause the thick-
ness to be lower for a given level of smoothness. It appears, therefore, that
running at higher moisture content levels, though giving a greater smooth-
ing effect at the calenders, does so only at the expense of a greater loss of
thickness than would occur were some quality of the paper other than
moisture altered to achieve the same change in smoothness. This, of course,
is particularly unfortunate from a papermaking point of view when it is
desired to produce a printing paper which is as smooth and bulky as
possible.
SA. 4 4 Other observations
Some other observations on calender operation which must be mentioned
have been reported by Howe and Lambert (61) and by Blanchard et al.
(82). In both these papers an analysis along similar lines is made of the
components of power usage in a stack of calenders. By measuring power
input, with and without paper, and making various assumptions and
deductions about power losses in the bearings and resulting from stress of
the rolls in the nips, it is concluded that only 30 per cent. to 50 per cent. of
the total power consumption can be attributed to straightforward work of
calendering, i.e. energy absorbed by the paper.
Howe and Lambert interpret their power analysis results as indicating
that the fundamental effect of the calenders is produced by 'rolling friction':
this is the elastic geometrical compression and distortion undergone by
rolls under heavy pressure in the nip regions which creates a small differ-
ential in speed between each pair of rolls that will have a shearing effect on
the paper. As further evidence in support of this theory the authors cite
some results they obtained on roll to roll slippage.
In earlier papermaking books the effect of the calenders was invariably
attributed to slippage between the rolls and it was assumed that this
created a rubbing or glazing action on the surface of the paper which
brought about the increased smoothness. Figures have even been published
showing the percentage of slip occurring on calender stacks for different
grades of paper. Howe and Lambert set out to assess the importance of
roll-to-roll slip by simultaneous differential measurement of the surface
speeds of top and bottom rolls in a high-speed stack. The method chosen
for this was basically similar to some of the more sensitive and advanced
means of draw measurement: it involved the use of discs with projecting
teeth fixed to the journal of each roll and serving to generate pulses which
420
were fed to suitable counters. Many comparisons were made under different
conditions with this equipment and in no case when paper ran through the
stack was there any significant difference (less than 0· 1 per cent.) between
the peripheral speeds of the top and bottom rolls observed. With no paper
in the stack a difference of a little over 3 per cent. was observed, the bottom
roll of course being the faster.
It is probable that older stacks of calenders with greater friction at the
bearings could possess a measurable amount of roll slippage with paper
running through the stack, but the importance of Howe and Lambert's
observation is that this cannot be put forward as the basic source of the
calender smoothing effect. To what extent the 'rolling friction' caused by
roll distortion is in fact responsible for producing increased smoothness, as
these writers go on to assert, is somewhat harder to judge. In the author's
opinion, the lack of roll-to-roll slippage, the greater calendering effect
obtained under crawl conditions, the observations by Brecht that use of a
static 'calendering' apparatus under suitable conditions results in the same
smoothness effect (both immediately and also after conditioning) as
machine calenders, and finally the visco-elastic relaxation that is known to
occur after calendering, all indicate that the change in paper properties is
due primarily to straightforward compression. Further confirmation that
roll slippage contributes little to calendering comes from an experiment
reported by Parker (129) in which it appears that hardly any effect is
observed on the smoothness and bulk of paper slipped over a polished
steel surface. Also experimental work by Mardon (126) indicates that
changes in paper compressibility are a function of the pressure and dura-
tion of pressure in the calender nip so that optimum smoothness is best
obtained from a stack of small-diameter rolls loaded externally. These
points make it seem certain that slip is relatively unimportant except to
increase gloss, particularly in supercalendering where fibre rolls are far
more flexible than steel and it is thus likely that 'rolling friction' becomes
more important.
421
CHAPTER SB
OPERATING FACTORS AFFECTING DRYING
AND CALENDERING
SB.1 STEAM HEATING AND CONDENSATE SYSTEM

The condition of steam supplied to the drying cylinders of a paper machine
is intimately related to the overall steam and power system of the mill. This
in turn is dependent on local conditions, on whether cheap power is
available or whether low-pressure steam can be obtained from an external
source, such as a power station. Some mills generate their own electricity
and steam in separate units, but a common and highly economical
arrangement is to feed high-pressure superheated steam into a turbine or
engine and use the low-pressure pass-out steam in the machine house. The
main disadvantage of this system is the limit it imposes on the pressure of
drying steam available for the machine.
In the machine house itself, steam may be required for a variety of
purposes from roof heating to wet felt cleaners. It is the job of the heating
and ventilation engineer to determine the most economical method of
supplying the various needs according to the temperature and quantity of
heat each demands; the installation of heat exchangers on air extracted
from the dryer section and on condensate water, the need for make-up
steam at high pressure, the optimum temperature of condensate return in
terms of the efficiency of running the boilers: these are the sort of questions
which can only be decided by a thorough examination of the entire steam
and ventilation system.
A detailed discussion of this topic is beyond the scope of this work and
only certain aspects are touched on in appropriate places. In this particular
section attention is confined to the use of steam on the paper machine
in so far as it directly affects the efficiency of production. Stearn usage in
the ventilation system and in ancillary drying devices will be dealt with in
later sections from the same point of view. This is not to say, however, that
the economical use of steam is not the concern of the paperrnaker; in fact
the drying section is the main part of any machine where substantial
reduction in running costs are possible. A close watch on the overall steam
consumption is most important and improvements to the system should be
continually sought. Areas of the drying section that repay examination will
be given particular attention in the following sections.
SB. I 1 Steam pressure requirements
The steam pressure required for drying depends on a multitude of factors
not least of which are the grade and substance of the paper concerned.
This question has already been considered in the theoretical section when
the influence of such variables as the resistance to transfer of heat from
422
THE DR YING SECTION AND CALENDERS SB.11
cylinder to paper and the ventilation conditions were discussed. Thus, an
M.G. cylinder is necessarily constructed from appreciably thicker metal
than an ordinary drying cylinder and this increases the resistance to heat
transfer and so necessitates a higher steam pressure within the cylinder.
On the other hand, the use of hoods improves ventilation conditions and
reduces the heat needed from the cylinders to achieve a given evaporation
rate; steam pressure is consequently reduced.
Increases in speed require greater evaporation rates. When other
economical means of increasing evaporation have been exhausted, it
becomes necessary ultimately either to add to the number of drying
cylinders or to increase the steam pressure in the main supply line. The
choice between these two courses will depend on whether a new or existing
machine is being considered, but the general tendency these days is
towards using a higher steam pressure. This is mainly for reasons of the
lower capital cost in terms of equipment and space, and partly the result of
a shift in the overall steam and power balance towards proportionately
lower power demand compared to steam, thus permitting the use of
higher back pressures on the turbine.
Higher steam pressure has several disadvantages: radiation heat losses
from pipes and cylinders are greater at the higher temperature, the latent
heat of vaporization is lower so that a greater weight of steam is needed to
provide the same quantity of heat on condensation, the paper temperature
may get too high, and there is more likelihood of steam escaping from
nozzles and joints. Further, the balance of steam and power generation
suffers considerably from using higher pass-out pressures: Robey and
· Webzell (97) have stated that an increase in pressure from 20 to 30 p.s.i.g.
reduces the steam flow for a given heat demand by 2 per cent. (this assumes
a drop to the same final temperature), but reduces the power generated in a
typical turbo-alternator set-up by over 10 per cent. Against this, apart
from its higher temperature the only real advantage of high steam pressure
is that smaller steam mains and valves are needed to carry the same weight
of steam.
For some purposes a higher steam pressure may be necessary on a
machine regardless of the pressure in the main bank of cylinders. Felt
drying cylinders are generally run at a higher pressure because the resist-
ance to transfer of heat from the cylinder surface to the felt is much greater
than between a drying cylinder and the paper. Calender heating, to be
effective, also requires higher pressures because of the temperatures
involved and the high resistance to heat transfer through the thickness of
the rolls. In cases where low-pressure steam is used in the drying cylinders,
separate supplies are necessary for these purposes.
On the other hand, some parts of the drying section require steam at a
pressure lower than that in the main bank of cylinders. The temperature of
the first few cylinders which the sheet contacts cannot normally be as high
as later cylinders, while at the end of the drying section also it is often
considered necessary to reduce the temperature of the paper before it is
calendered and reeled; in both cases the pressure of steam in the cylinders
must be reduced accordingly. For certain special purposes one set of
423
SB.12 THE DRYING SECTION AND CALENDERS
cylinders may need to be at a pressure quite different from another set, for
example, in curl control where a differential between top and bottom
cylinders is required. Such differences can be obtained simply by throttling
the main steam pressure at appropriate points (though the question of
pressure differential across the cylinders enters here, see SB. 1 4).
A more permanent arrangement for obtaining lower steam pressure,
especially for the first few cylinders, is the cascade system involving re-use
of flash steam. This system, though it has altered little over several decades,
is a simple and effective way of re-claiming steam blown through the
cylinders and extracting as much heat as possible from the steam supplied,
thereby reducing the overall weight of steam needed in the dryer. It will
now be discussed in some detail.
SB. 1 2 Cascade drying system
In the cascade system of drying there are usually two separate flash tanks,
though occasionally only one or as many as three tanks are used. Con-
densate from the main bank of cylinders drains to the first flash tank, and
steam from this tank is flashed off at a lower pressure and passes into the
second bank; similarly, condensate from this bank of cylinders drains to
the second flash tank, and steam from this tank then passes to the third
bank. Condensate from the two flash tanks (which are both level-controlled)
and from the third bank (which may go through a separator to vent gases)
passes to a common receiving tank from where it is pumped back to the
boiler under level control. With an M.G. machine the M.G. cylinder itself
takes the place of the main bank, and flash steam from the condensate,
usually augmented by lower pressure make-up steam, feeds pre- or after-
dryers, or is used for heating elsewhere.
To function smoothly the cascade system depends on there being an
adequate difference available between the steam pressure in the main inlet
manifold and the pressure in the condensate receiver. Where low-pressure
steam supplies the dryers, a vacuum pump is needed on the receiver to give
sufficient pressure drop.
The actual volume of steam flashed off at both tanks is very small com-
pared to the main flow; for a pressure drop even as high as 10 p.s.i.
between the inside of the cylinders (which determines the temperature of
the condensate) and the flash tank, the steam flashed off amounts to only a
little over 1 per cent. of that entering the cylinders, and this ignores any
loss of temperature in the condensate lines. This quantity would normally
be insufficient to supply the relatively large number of cylinders that are
usually fed by flash steam (the number of cylinders on the first flash tank
may be 15 per cent. or more of the main bank, while the number on the
second flash tank may be 40 per cent. of the second bank). Without an
additional supply of steam the condensing rate in cylinders fed solely by
flash steam would be extremely low and the surface temperature could not
be expected to be anything like as high as is normally found. In practice,
valves on the flash tanks are adjusted to give the temperature required and a
certain amount of steam is blown through the main bank of cylinders to
augment that available as flash-off.
424
THE DRYING SECTION AND CALENDERS SB.12
The more cylinders are placed on flash steam the lower will be their
average surface temperature under comparable conditions of pressure drop
through each bank of dryers. Too many flash-steam cylinders at the
beginning of the drying section can cause the surface temperature to drop
so low that sweating occurs even with the flash tank valves open as much
as possible. Resort must then be had to using make-up steam from the
main bank to increase their pressure, a practice which if permanent requires
certain precautions to be taken otherwise the resulting reduction in
pressure differential across the main section could lead to waterlogging.
This is only avoided by using differential pressure regulators which either
prevent the difference in pressure falling below a minimum value, or keep
it constant by automatic regulation of a make-up valve between the steam
headers of the two sections or by more elaborate arrangements.
Too few flash-steam cylinders, on the other hand, will cause rapid
heating-up of the web on the first few cylinders; there is then some danger
of the sheet sticking to the cylinders, and of excessive early vaporization
which causes the formation of blisters in the sheet. The paper also tends to
have a rougher surface when drying is initially rapid and sizing becomes
uneven. Despite these disadvantages, when the drying section steam
pressure limits production on the machine, it is preferable that as few
drying cylinders as possible are operated on flash steam in order to reduce
heating-up time of the web and reach the region ·of constant rate of
evaporation early in the drying section.
The most satisfactory compromise is one which suits all the different
conditions arising on a machine; once found this is in practice rarely
altered, though it is advantageous to have a relatively easy means of
switching one or more cylinders onto the main inlet or flash should the
need arise. Generally it is preferable to err on the side of having too many
cylinders on flash steam provided there is an adequate pressure drop
through the system as this seems to give more stable conditions and
provide greater latitude in adjusting the pressure drop through each
separate bank of cylinders.
The temperature of condensate returned to the boilers is high even when
a vacuum system is used, and this can create difficulties at the feed pump to
the boilers. It is generally found to be economical to extract as much heat
as possible from the condensate before it passes from the machine house to
the hot-well, so that in some cases it may be necessary to spray cold water
into the condensate receiver to condense any remaining steam and reduce
the temperature; this also helps to ensure where main supply pressures are
low that the maximum vacuum needed to operate the cascade system can
be obtained. Alternatively, heat exchangers are commonly located in the
main condensate line, and either water is heated for use in felt cleaning,
vacuum pump sealing, fresh-water sprays, etc., or air is heated as it enters
the machine house or supplies the make-up requirements of hoods. Flash
steam, instead of being used in drying cylinders, may also be utilized in a
heat exchanger (this is a popular arrangement on M.G. machines), and
many other systems are possible to suit the particular conditions on any
machine.
425
58.1 3 THE DRYlNG SECTION AND CALENDERS
An entirely different drying system which avoids the separate circulation

of flash steam involves the use of 'thermocompressors' or ejector-type
devices (99). These may either be used singly on each cylinder, or on a
complete section; an application of the latter type has been described in
some detail by McCaffrey (77) and will be briefly mentioned. With this
arrangement, before entering the drying cylinders the main steam supply
passes through an ejector which serves to provide the suction to draw out
flash steam from a tank into which condensate drains in the normal
manner. Three or more individual sections of dryers and a separate one for
felt dryers or an M.G. cylinder may be provided, each with its separate
main steam supply, ejector, level-controlled condensate tank, and integrat-
ed flash system. Each section is set independently to give the pressure
desired.
The main advantage claimed for this system is that increased differential
pressures can be run across the cylinders, thus aiding evacuation of con-
densate at high speeds without resorting to an excessive blow-through and
without the risk met in a cascade system that sectional pressures are
incorrectly adjusted and condensate builds up in some cylinders. Also the
system is claimed to be more versatile in that each section is independently
adjustable with its own inlet pressure or temperature controller and
differential pressure controller; the inlet pressure controllers can, if
necessary, be interlocked so that the pressures in different sections are kept
constant relative to each other. A strong air-bleed is needed to prevent a
build-up of air in the closed eductor circuits, and periodic blow-down of
dryers and separators may be necessary.
SB. 1 3 Operating a cascade system
It is essential for the smooth operation of the dryers that condensate is
removed from each of the cylinders as fast as it is formed; this can only be
done by ensuring that there is an adequate differential available between
the steam pressure at the inlet joint of each cylinder and the outlet pressure
in the condensate line. When operating a cascade system certain precau-
tions are necessary in order to maintain this minimum pressure differential
on each cylinder under varied drying conditions; this aspect will now be
considered in some detail. The actual magnitude of the pressure differential
required, its dependence on the machine speed, the method of condensate
removal, and so on, and also the difficulties occurring if condensate is not
removed efficiently, are discussed in SB. 3.
It is important first to appreciate the inter-relation between the heat
requirements in the drying part of a machine at any particular time, and
the conditions of steam pressure and flow which serve to provide this heat.
In the main bank of cylinders a certain pressure is found necessary to
achieve the correct drying rate. At the end of the cascade the condensate
receiving tank is at an absolute pressure (be this greater or lower than
atmospheric) which is set to a value found suitable for normal running
conditions. The difference between these two pressures is the total drop
available through the system, and must be accounted for in terms of the
following: the pressure loss incurred in removing condensate from the
426
THE DRYING SECTION AND CALENDERS 58.13
cylinders, the loss in the siphon and condensate pipelines and in the
cylinder joints, plus additional losses imposed by valves or orifices in the
steam and condensate lines. Each of these pressure losses is in turn affected
by the actual flow of steam and water through the condensate system,
becoming greater as the flow of either increases, so that in any particular
condition both the main steam pressure and the steam flowing to the drying
section are mutually dependent in providing the amount of heat required
to dry the paper. Further, if the absolute pressure in the condensate
receiver is permitted to alter according to conditions, this too will be
affected by the pressure loss through the system.
To illustrate this, consider the conditions at start-up when the paper has
been fed to the reel and the inlet manifold pressure is raised to a particular
value which it is hoped will dry the paper to the desired moisture content.
The flow of steam increases until it stabilizes at a level sufficient to cause
the total pressure loss through the system to become equal to that available.
Suppose that this flow of steam has a heat content in excess of that required
to dry the paper under the conditions applying; more steam then enters the
main bank of cylinders than can be turned to condensate (the rate of heat
transfer being fixed by the pressure in the cylinder) and so an excess is
blown through in a condensate/steam mixture, causing the temperature of
the flash cylinders to increase and the paper to become too dry. Conse-
quently the inlet manifold pressure would be decreased by throttling the
main steam control valve, giving both a lower flow .of steam and a reduc-
tion in steam pressure (and hence in heat transfer) in all the cylinders.
Eventually a pressure can be reached where the heat supplied by the steam
ft.ow equals that required to dry the paper.
There are two important points to note in this. Firstly, the lower the
overall pressure losses through the cascade system for a particular steam
flow, the lower is the drying pressure required to produce this flow. Hence,
where a machine is limited in production by the drying capacity, i.e. near-
maximum available steam pressure is always being demanded, it is useful
to eliminate all unnecessary restrictions in the cylinders, due for example to
scale build-up, and of course to have the valves on the condensate lines and
flash tanks as far open as possible. This allows as high a steam flow as is
practicable for a given pressure in the inlet manifold, thus ensuring that the
condensing rate in the cylinders (particularly those on flash) is large
enough to satisfy as much as possible of the heat demand from the paper.
The second point is that it is possible for the required flow of steam
through the cylinders to be obtained with a low overall pressure drop; in
other words, it may be found that the inlet pressure needed to dry the
paper is very low. Under such conditions the danger arises that the pressure
differential across the cylinders is insufficient to prevent condensate from
building up. To avoid this happening it is necessary to increase the total
pressure drop through the cascade system by increasing the amount of
vacuum on the condensate receiver, throttling the valves on the flash tanks
if the temperature of the first few cylinders then becomes too high.
Occasionally the heat demand for a particular grade may be so low that
the supply manifold pressure drops below atmospheric. Running under
427
such conditions is always difficult and drying is apt to become erratic due
to air leaking into the cylinders. One remedy is to cut off some of the later
dryers (thus increasing the pressure needed in the remainder) but this
requires efficient isolating cocks to prevent condensate being drawn back
into the cylinders and may cause them to sweat even when only alternate
dryers are isolated. Alternatively any supplies of hot air to the dryers could
be shut off at the risk of creating problems with the moisture level.
When the drying cylinders are being kept warm in preparation for
feeding-up the sheet, as just before start-up or during a wet-end break, the
temperature and flow of steam required is very much less than when the
machine is running. Consequently steam pressure is set lower. The danger
of condensate slowly building up thus becomes greater at this time,
especially where the vacuum on the condensate receiver tank is not
maintained but is allowed to fall as a consequence of the diminished
pressure drop through the system. Though this does not matter particularly
as long as the sheet is down, it creates difficulties in feeding-up. When the
steam pressure is put back to its normal operating position some cylinders
will be slower to clear than others (and, in fact, unless other precautions
mentioned below are taken, some may never clear) and the paper will
become too damp. The dryerman may then be faced with considerable
difficulty as the steam pressure is first increased further to correct this, then
decreased as more cylinders are cleared, with the continual danger that
conditions are so unstable at some point in the dryers that a break occurs
and the whole process has to be repeated.
For this reason it is a valuable provision to ensure that adequate
pressure differential is available to keep cylinders clear not only when
running, but also when the sheet is not actually on the dryers. If when
running normally the conditions are such that sufficient pressure differ-
ential is only just available, then almost certainly trouble will occur when
feeding up after a break. The remedy is to set the condensate receiver
vacuum pump so that adequate pressure differential is always available
even when the cylinders are just being kept warm. If it is suspected that the
maximum vacuum obtainable is still insufficient to clear cylinders when
keeping the dryers warm, the steam pressure must be reduced as little as
possible from the normal running value (in other words the outer surface
of the cylinders is allowed to get as hot as possible without overheating
especially the top bank).
SB . 1 4 Maintaining adequate pressure differential across the cylinders
In the previous section, some general comments have been made on
operating a cascade system to ensure there is a pressure differential across
each bank of drying cylinders adequate enough to prevent condensate
building up. To achieve this in practice is far from easy simply because it is
by no means obvious when waterlogging is occurring in any particular
dryer. It is safe to say that many more machines are run with some cylinders
waterlogged than even the most pessimistic papermaker would suggest.
In the first place a substantial difficulty is created by the fact that the
evaporation rate, and hence the heat extracted from each cylinder, varies
428
down the machine. Figure 5. 16 illustrates in a typical case the effect of this
on the flow of condensate and shows that cylinders at opposite ends of the
main bank may condense very different quantities of steam. Yet with group
drainage of the condensate lines, each cylinder is subjected to the same
overall pressure differential.
For a high rate of condensation, the pressure differential needed to keep
the cylinder clear will also be high due to the greater pressure loss in the
CDOO
MAIN e:>ANK. OF C.YL.IN DE-RS

1
No'!!:i. 8 TO 3 Ci:i.
Soo
al
'J 400
0
:!
tt
II.I
a.
Iii
d)
_J
300
~~
0
j
u.
w
~ 'Zoo
tfl
l
tlJ
a
z
0
v
loo
o'--~~-'-~-'-~..____.~_,_~_.___,.____....
8 1'2 I~ Zo 24 28 32 ~.
ORYING CYLlNDE:R NUi\\&E;R.
Fig. 5. 16. Flow of condensate from successive cylinders on the main steam supply
manifold of a paper machine
condensate lines. If the existing pressure differential is inadequate then

condensate builds up in the cylinder until the rate of condensation
diminishes to a level which can be extracted under the prevailing conditions
-in other words the cylinder becomes waterlogged.
On the other hand, for a cylinder with a low condensing rate the existing
pressure differential may be too great. In this case it may be expected that a
sufficient volume of steam blows straight through the cylinder until the
pressure loss of the condensate/steam mixture rises to the level of the
429
5B.14 THE DRYING SECTION AND CALENDERS
available pressure differential. The disadvantage of this is that the large

blow-through of steam could tend to raise the pressure in the flash tank,
diminishing the total pressure drop available across the bank of cylinders
and causing those cylinders with a much higher rate of condensation to
become waterlogged.
This state of affairs is avoided by grading the area of siphon tips or,
more conveniently, by the use of such devices as orifice plates in the
condensate lines from each cylinder. These are sized in such a way that
cylinders with a high condensing rate have a relatively open orifice, thus
leaving a higher pressure differential across the cylinder, while those with a
low condensing rate have a relatively narrow orifice. The dimensions of the
condensate lines from each cylinder, and to the flash tank, must of course
be ample to cope with the highest rate of condensate flow so that each
individual line can be restricted to the extent desired. Provided the relative
condensing rate of the different cylinders is approximately known, the
orifice plate dimensions can be suitably arranged to reduce considerably
the likelihood of condensate building up preferentially in one or more of
the cylinders.
Prevention of steam blowing through cylinders with a low condensing
rate can also be achieved by the use of steam traps and these are often used
on the first few dryers. But, in general, except possibly in the case of scoop-
removal of condensate on slow machines, the use of traps on every cylinder
is not favoured partly due to their cost and the maintenance needed to keep
them working properly, and partly because a certain quantity of steam
blowing through is desirable for smooth operation of a cascade system.
Further, when using traps on cylinder condensate lines the possibility of
steam locking is increased; this is a condition in which steam enters the
condensate pipe and trap and effectively blocks off further flow until it has
cooled sufficiently to condense. Steam locking can be avoided only by
having a release valve on each trap.
The proportion of steam blowing through each bank of cylinders is
determined by the overall pressure differential and is generally estimated to
be between 5 per cent. and 20 per cent. This amount is thought suitable
to ensure that cylinders are kept free from air accumulation and also assists
in stabilizing conditions as heat demand from the paper fluctuates during
normal running. But a high blow-through increases steam consumption
and creates a large drop in pressure in condensate pipe lines. Figure 5. 17,
adapted from a report by Wahlstrom and Larsson (109), shows how under
various conditions the pressure drop in a 15 ft. long, 2 in. diameter, vertical
pipe varies with the percentage of steam added to water flowing down-
wards. The rapid rise in pressure drop accompanying an increased propor-
tion of steam is of particular significance and indicates that loss of pressure
in individual cylinder condensate lines can be of overriding importance
when operating with a high blow-through. This point will be discussed in
more detail in connection with condensate removal systems. For the
moment, it will be noted that the absolute steam pressure is equally as
important in determining the magnitude of pressure drop as is the quantity
of water involved, and as the steam pressure reduces, the pressure drop
430
increases appreciably. The implication of this is most important: under
comparable conditions of blow-through and condensate flow the pressure
loss in the siphon and condensate lines of drying cylinders operating at low
steam pressures, in particular of those dryers working on flash steam, can
be substantially greater. To reduce the possibility of waterlogging, a
greater pressure differential must be made available at lower pressures;
thus, the practice of adjusting main valve positions in a cascade system to
14 W..'t'"t.R f"~OW'.
,3,?,go L&!>/HK
___ ..
----- _.,.
Zooi;>
G05o
L&&/HR.
LbS/ftR.
12
I
I
lo .I
~I
ci./
t-- / I
I
·-
~
e I I
I •I
cL ~I
Q..'
Q.-
0 ,.._I
~
CL I
0
I.II
1--:di I
I I
C:J.
::I I a..,
II\ 4 t 0'
Ill
w I
et
Q.
..... I I
2 10·
I '"
ui.
<V I
,'>J/
/ / / .,.,<i../
.
0
--- . ....----"' ./
'/
o·25 o·S l·o 2·S 5 lo 25 So ,0.

PERC!::NTAGE STEAM \~ WAiE:R.
Fig. 5. 17. Pressure drop in a downward vertical pipe for various flows of water
containing different percentages of steam at different pressures (after Wahlstrom and
Larsson)
give equal pressure drops through each bank of cylinders can be improved
by arranging instead for the available pressure differential to increase 2 or 3
p.s.i. in successive banks.
But to be certain that no cylinder is waterlogged it is necessary either to
have a differential pressure gauge on each dryer, or to have suitable tapping
points for using a portable gauge. Even then, particularly as the exact
pressure differential required to evacuate each cylinder is not known and is
dependent anyway on the quantity of steam blown through at any time,
431
58. l 5 THE DRYING SECTION AND CALENDERS
there must be some element of doubt that all cylinders are clear unless the
system is run in such a way as to have available fairly high overall pressure
differentials, and this may be neither convenient nor economical.
Probably the best approach to this problem is to undertake frequent
checks of the cylinder surface temperatures using a pyrometer or a non-
contacting infra-red radiometer which measures thermal radiation.
Though these instruments cannot always be relied on to give a correct
absolute measure of surface temperature, they are always useful for com-
paring the temperature of cylinders one against the other; those cylinders
in which condensate builds up have a lower heat transfer, and hence a
lower surface temperature, so that comparison with the normal running
pattern enables such cylinders to be detected with reasonable certainty.
Confirmation of the need for action can then be obtained by comparing
the pressure differential across suspected cylinders with the differential
across adjacent cylinders off the same bank. Alternatively, the flow of
condensate from suspect cylinders can be compared with those from
adjacent cylinders, though measurement of this (as discussed later)
requires some caution even for comparative purposes because the act of
measuring the condensate from individual cylinders is liable to interfere
with the natural flow.
Normally if a cylinder is suspected of being waterlogged there is little
that can be done until the machine is shut. It may then be noticeable that
the cylinder cools slower than the others, showing that is is holding more
water, which will be a good indication that removal during running was
not satisfactory. The cylinder and condensate removal system would then
be opened up for examination, and if these were in sound condition, with
no signs of pipelines being rusted up, siphon tips broken, etc., then a
change may be made in the size of orifice plate in the condensate line.
Possibly it would be useful to have valves in the condensate lines instead
of orifice plates to give greater flexibility and permit alteration of the
pressure differential during running. But most papermakers would feel
uncomfortable about this arrangement unless there were regular, possibly
weekly, checks through the whole system.
Sight glasses on condensate lines can also be useful to give an indication
as to whether the flow is stagnant, but these are probably too subjective
and difficult to interpret to do other than show up really bad cases of
waterlogging.
SB . 1 5 Condition of the steam
The performance of the dryers is dependent on the condition of the steam
and this can vary in three important respects: the degree of superheat, the
air content, and the extent of contamination. Practically all machines use
steam which is saturated, or very close to saturation. An ideal often aimed
at is to pass the steam out with just sufficient superheat to ensure that by
the time it enters the dryer supply manifold it is saturated but dry as this
minimizes condensation in the pipeline. But for a long time there have been
many who advocate the use of super-heated steam mainly, it would seem,
because in most integrated steam and power systems it is necessary to
432
THE DRYING SECTION AND CALENDERS SB. I 5
de-superheat the steam before passing it to the machine house, and this is
to some extent uneconomical.
Though this may be the case, the weight of opinion is firmly against
using steam with any but a moderate amount of superheat. The superheat
itself adds little in heat value over and above the latent heat (a superheat of
10 deg. F. gives out approximately 5 B.t.u. on falling to saturation tempera-
ture, compared to around 970 B.t.u. latent heat). But the main trouble
appears to be that drying becomes erratic with superheated steam, especially
in an M.G. cylinder and it proves difficult to control the steam without
the occurrence of wide swings in inlet temperature for only small changes in
the ft.ow. The reason advanced for the erratic drying is that condensation of
the steam inside cylinders is not even: small differences in drying rate across
the cylinder cause the temperature at some points to be slightly above
saturation and the steam does not then condense in those places. In
contrast only a small difference in temperature will condense saturated
steam, releasing the latent heat and providing a drop in pressure which helps
to maintain circulation in the cylinder.
There is no doubt at all about the detrimental effects of air in the steam.
Although feed water passing to boilers is de-aerated, there can still be a
small percentage of non-condensible gases diffused in the steam passing
into drying cylinders, and if this is allowed to accumulate its presence soon
has an effect on drying efficiency. Air can leak in through poorly sealed
steam glands in the parts of a dryer subject to a partial vacuum, and also
through manhole covers.
For pressures in the region of 10 p.s.i.g. each 1 per cent. of air in the
steam reduces the temperature approximately 0·5 deg. F.; this would
necessitate an increase of roughly 0·25 p.s.i. in pressure to maintain the
original temperature. Thus, 10 per cent. of air would reduce the tempera-
ture of 10 p.s.i.g. steam from 239 deg. F. to 233 deg. F., necessitating an
increase to almost 13 p.s.i.g. to maintain the original temperature of
239 deg. F.
For this reason it is essential to ensure that air and other non-condensible
gases are removed from the whole steam and condensate system. At
start-up, cylinders should not be allowed to remain stationary right up to
the time when the sheet is ready for feeding up. They should preferably be
rotated slowly right from the start, and the steam allowed to blow right
through the system and out at suitable vents in the condensate tanks for
sufficient time to clear all the cylinders (often such vents contain a thermo-
static element designed to shut off automatically when their temperature
reaches a suitable pre-set level). During running some form of continuous
venting can also be beneficial and, in cases where the main pressure is less
than atmospheric, absolutely vital. In systems based on the use of ejectors
it is quite common to have automatic venting devices which allow steam
and air to blow off when the difference between the actual temperature and
the temperature corresponding to the measured pressure exceeds a pre-set
value.
Apart from the loss in effective temperzfur.e occasioned by the presence
of non-condensible gases, the greater density of air (approximately double
433
58.2 THE DRYING SECTION AND CALENDERS
that of steam) will create a tendency for it to collect on the cylinder walls
where its effect on condensing conditions will be more noticeable. Also
there will be a substantial increase in resistance to heat transfer if the air is
trapped in an insulating layer between the cylinder surface and the con-
densed water. To avoid this, Hoyle (88) has suggested that it is essential to
have adequate turbulence in the cylinder, and this would also have the
benefit of giving a more even distribution of steam within the cylinder. For
high-speed machines, in which the condensate is rimming and there is,
therefore, little turbulence caused by water movement, it may well be
advantageous to follow Hoyle's suggestion of injecting steam into the
cylinders on to special baffles or some other device which will assist its
dispersion.
The growth of scale and rust in the drying cylinders and steam and
condensate lines can create difficulties over a period due both to the
unevenness and overall reduction which occurs in heat transfer through the
cylinder, and to the increase in pressure drop occasioned by roughening
and narrowing of the pipe-lines and fittings. The use of filming amines in
steam is reported to prevent this in drying cylinders by forming a non-
wettable film on the cylinder surface which protects against the attack of
oxygen and carbon dioxide in the steam. A further benefit stated to follow
from using filming amines is that at lower speeds of rotation condensation
in a dropwise fashion is promoted and this, by making it more difficult for
an even film of water of high heat-transfer resistance to be formed, helps
to create better conditions of heat transfer over the cylinder surface.
Schoonen (55) has given a summary of the results of using filming amines
on various machines and mentions that increases in production of from 5
per cent. to 15 per cent. have been claimed as a result of improvements in
heat transfer. But even though an increase in heat transfer was confirmed
in a laboratory experiment, a mill trial over four weeks on two slow
machines was not successful, and no improvement was observed which was
significant compared to normal operating variations.
The amine also acts as an inhibitor to corrosion and when applied to a
surface already corroded it penetrates under and loosens existing corrosion
products. If too large a dose is introduced into the system at first, there is a
danger that steam traps and other constrictions in the condensate lines
become clogged. Addition of filming amines is generally into boiler feed
water and must be continuous or former conditions soon return. As with
any other additive the economic value of its use should be carefully
assessed, though this requires accurate knowledge of the evaporation rate
and efficiency of the drying section.
SB.2 FELTS AND FELT DRYING

The purpose of felts in the drying section is mainly to increase the evapora-
tion rate and reduce the overall cost of drying. Which type of felt in each
section of a particular machine best satisfies this objective, whether indeed
a felt makes any improvement at all, are questions to which the papermaker
434
THE DRYING SECTION AND CALENDERS SB.2 1
should require an answer. Unfortunately, it is generally the case that felts
are selected for their performance in every other respect but these.
The reason for this is that on the vast majority of paper machines there
is a basic lack of information on the performance of the dryers. Commonly,
a pressure gauge on the main steam manifold is all that the papermaker
has to guide him and this, though useful as a rough indication of general
drying conditions particularly when the available pressure limits produc-
tion on the machine, is of little value on its own in assessing whether the
substitution of a new type of felt over one bank of cylinders has brought
about a significant improvement. The papermaker cannot, therefore, be
blamed for choosing his felts primarily for their running qualities:
resistance to wearing, lack of shedding fibres, ease of guiding, absence of
mark on the paper, good dimensional stability, and so on.
The difficulty of obtaining useful comparative data is apparent when the
literature is searched on the subject of dry felts. One of the few attempts to
get operational data about performance, the B.P.B.M.A. Technical
Section questionnaire in 1958, produced absolutely nothing. The benefit of
felt dryers is an even more obscure subject. The section that follows gives a
brief summary of the available published information.
SB . 2 1 General comments on different types of dry felt
Dry felts used to be made mainly of wool, frequently woven endless like
wet felts. Nowadays the length of dry felts usually necessitates seaming,
and wool is not so favoured except under extremely acid conditions because
of its tendency to shed dust as it ages. Difficulties in keeping an even
tension also frequently arise due to the varying degrees of shrinkage that
occur at different places across the machine, particularly at the edges. On
some machines dry felts containing asbestos are used where very high
temperatures are involved, but until comparatively recent years pure
cotton in a variety of weaves was probably the most popular basic material,
at least on faster machines. The main advantages of cotton have always
been its cheapness, strength, dimensional stability, and good resistance to
abrasion, though it gives a coarser surface.
The appearance of synthetic materials has radically changed the situa-
tion. Dry felts are now reinforced with nylon and terylene in various ways
in the body of the yarn and in separate yarns, or occasionally felts are
made exclusively from synthetic materials in various blends; their high
strength, flexibility, dimensional stability, and resistance to heat and other
contaminants have brought about many beneficial extensions in life. Race
(49) has shown that synthetic materials absorb water more rapidly and
with less fibre swelling than cotton (thereby retaining their permeability to
moisture better). Also he has demonstrated under a variety of experimental
conditions that synthetic felts dry comparatively quicker. This has been
confirmed by Bowden (32) who observed a reduction in steam usage in a
hot-air felt-drying system when a terylene felt was substituted for a wool
one on an M.G. machine.
The use of felts containing synthetic materials is nonetheless highly
individual to machine conditions, and in some cases they may at best
435
SB.22 THE DRY ING SECTION AND CALENDERS
serve only to give an extended life. In this respect Race states that felt life
is increased on average 2 to 3 times when replacing wool by synthetic felts
on slow fine-paper machines (giving a similar cost per ton), but replacement
of cotton felts on faster machines usually shows substantial economic
improvement. An improved evenness of drying across the sheet, less dust,
and lack of trouble with contamination of oil on the edges of felts (because
synthetic materials absorb oil less easily) are other advantages claimed.
Failure of terylene felts is often due to hydrolysis followed by mechanical
degradation, so that when used over early dryers under damp conditions,
felt dryers, or the use of one cylinder for drying the felt instead of paper,
are essential.
In recent years the advent of plastic fabric dry felts or dry 'wires' has
brought about a further development which has been very beneficial,
particularly on faster machines. Here the advantage is generally less one of
giving extended economic life, so much as improving the drying rate and
efficiency. Especially when used in early positions in the drying section,
fabric felts may be found to give a big improvement and, because they do
not hold water, less trouble is likely from damp streaks emanating from the
felts. Also they do not shrink or expand in use, and can be cleaned with
water or low-pressure steam. However, in some installations troubles with
wrinkling of the fabric, especially at the seam, has been reported. Marking
of the sheet does not appear to be a problem.
The type of felt which best suits a particular situation on a machine
obviously depends on many factors. Careful comparison of the performance
and the conditions prevailing with different dry felt types and materials
in each position of a drying section is most essential, though of necessity
each trial is extended over a long period. This subject is dealt with in
greater detail in SC. 2 2, and attention is now turned to examine what
pointers there are indicating which type of dry felt is best in different
conditions from the point of view of drying rate and efficiency-the most
important criteria, yet on most machines the most difficult to assess.
SB. 2 2 Specific comparisons between different felt materials

There have been only two comprehensive reports of investigations in
which direct comparisons of different types of felt have been made, one
using an experimental set-up and the other an experimental machine. The
results obtained will now be considered.
B. W. Attwood and S. F. Smith (1, 7) devised an apparatus for experi-
mental simulation of drying conditions which represents the first attempt to
obtain some fundamental data. In this apparatus samples taken from the
press of a board machine were dried in contact with a heated, curved metal
surface in a manner designed to simulate as closely as possible the sequence
of events in the drying section of the machine. The sheet was successively
pressed into contact with the heated surface by a felt, released and reversed
through 180 deg., exposed to moving air, then placed with the opposite
side in contact with the metal surface, the whole cycle being repeated as
often as necessary. The tension at which the sheet and felt were held, the
temperature of the heated surface, the condition of the air blown onto the
436
sample, the pressure of contact onto the drying surface, the length of time
in each cycle that the sheet was in contact with the surface and expoeds
to the air, and the number of drying cycles could all be varied. Conditions
of drying were in the first place selected to correspond so far as possible
with those observed on the board machine; the air velocity was set equal
to the machine speed.
By weighing the samples after a given number of drying cycles (corre-
sponding to passage of the web over so many pairs of dryers) a drying
curve could be constructed to show reduction in moisture content from the
initial value of around 65 per cent. In general, excellent agreement was
found with the equivalent machine drying conditions in terms of lbs. of
water removed per square foot of dryer surface per hour. The more obvious
differences in detail between the experimental and the actual methods of
drying were that the felt itself was not dried and only one was used, and
also the temperature of the heated surface remained constant.
Amongst their various experiments, Attwood and Smith made a direct
comparison (with other conditions constant) of the rate of drying with
felts made from several different materials, viz. wool, cotton, cotton/
asbestos, linen, wire cloth, and linen faced with tinfoil to make it imperme-
able to water and vapour. The drying rates with each of these types of felt
varied surprisingly little from each other, considering the wide difference
in characteristics between them. But of those tested the permeable
materials, wire cloth and linen, showed the highest drying rates, while wool
and cotton were practically identical, cotton/asbestos slightly worse, and
the tinfoil-faced felt by far the worst. It may also be noted, although the
differences concerned are small, that of the two best materials the wire
cloth gave slightly the better drying rate in the early drying cycles when the
board samples were damp, but that in the later cycles the drying rate was
faster with the linen felt.
The other investigation in which different felts have been compared was
part of the extensive work referred to earlier which was carried out on the
Swedish Central Laboratory experimental machine; this has been reported
by Ponton (47). In this experiment the moisture control of paper entering
an M.G. cylinder was varied, and the water evaporated over the cylinder
for each condition was determined. A direct comparison of the perfor-
mance with different types of felt (run for this purpose at the same tension)
was thus possible.
Figure 5. 18 shows the quantity of water removed from the web after
passing over the M.G. cylinder with various types of felt, viz. wool, cotton,
terylene, takryl, perlon (nylon) wire, and without any felt, in each case for
different moisture contents entering the M.G. cylinder. The line passing
through the origin represents the quantity of water in the web entering the
cylinder, so that the vertical distance below this line represents the water
remaining in the web under the various conditions. In Fig. 5. 19 the total
steam consumption (in both the M.G. cylinder and a felt dryer) is related
to the water removed from the web.
These results show first that when the web is comparatively damp
entering the cylinder, the type of felt does have a pronounced effect on the
15 437
water removed. In this case wool and cotton felts gave the best water
removal with the perlon wire only slightly worse for the dampest conditions.
When a considerable quantity of water has to be transported away the
ability of the wool and cotton felts to take up and retain this water is thus
advantageous; once the quantity of water requiring removal has fallen,
with adequate ventilation the wire-type felt appears to give a similar
performance. Furthermore, reference to Fig. 5. 19 shows that the wire-type
felt required a lower steam consumption, i.e. gave a higher efficiency of
steam utilization, despite the fact that in the experiment steam in the felt
dryer was kept at a controlled temperature although this was not required
(
I.II
IX
<(
v
z
:I
Cl:
WOOL. "~o COTTON F£.L.T$.
~ l'ZS P'i::.Rl...ON W1R&.
TIERYl..l::Mli: ANO TA~RYL FiLT5.
;J: loo W•n<our ~•L.T.
ri'
~ 75
.s: 5o
~
~ 25
~ 0
3 o O·S l·o l·S
<5 WATER/Fie.RE RATiO !:NTER.IN~ C.'r'UNDf\;l.
Fig. 5.18. Comparison of water removed on M.G. cylinder with different types of felt
(after Ponton)
for drying the wire. Synthetic felts removed a significantly lower quantity
of water, possibly because they are less able to take up water than cotton
and wool felts, yet offer similar resistance to the passage of water vapour.
Without a felt the drying capacity was severely reduced, though overall
less steam was required to dry the paper.
When the paper was drier entering the cylinder, the difference between
the various types of felt diminished until at about 0·6 water/fibre ratio
(37 per cent. moisture) it ceased to exist altogether. This confirms the
reduced importance of the felt in governing the rate of water removal in
later stages of drying, though the results also indicate that the drying
capacity will nevertheless be greater than without any felt, presumably due
to the better heat transfer condition across the cylinder/paper interface.
Further support of the diminishing role of dry felts as the paper becomes
drier may be adduced from the results obtained by Jepson (90) when he
observed the effects of removing dry felts from a number of machines: in
all cases reduction in drying rate was much greater when felts in the early
positions were taken off.
In a further experiment, Ponton compared the performance of two wool
438
WOOL. F.:.\.T.
ci 200
FCi\..TS.
':J
0
'::t
IX
~
~
~ \$0 P.l!H..,...
{
0
~
0..
~ loo
:J
Ill
1.
0
\J
~
<( So
llJ
...
I/I
0
0 So loo I So
WA."TE:R REN\O\/EO FROM. We & , t<.5. Pi<R l40Uft.
Fig. 5 .19. Steam consumption related to water removed from web on an M.G. cylinder
with different types of felt (after Ponton)
.'1_
al E.~iE.R1NG CYi...INOi;.Q. IN ~\IE:rt'.>.
'l t<.EN\OVEO FROM WEI) USIN.G.:
I-
z LIQ~T WOOi.. fE:.1...T.
:J HEAvY WOOL f-£.1...\.
Cl
l'2.$
.;.,
a.
loo
;!>
ci 75
*f(
~ So
I!.
() 25
>-
...tz 0
-(
0 0·5 l·o l·S.
::J WATeR / F-1 bRE R.ATIO ENTfRING C::Yl....l~Oci(,
G
Fig. 5.20. Comparison of water removed on M.G. cylinder using a light and a heavy
wool felt (after Ponton)
439
felts, one of basis weight 3,500 grams per sq. metre, the other 2,100 grams
per sq. metre. The results, represented as in the previous comparison, are
shown in Figs. 5. 20 and 5. 21. The thinner felt not only gave a higher rate
of water removal, but did this at a higher efficiency of steam utilization.
The difference in performance with the web entering the cylinder damp is
clearly significant, and is probably due to the greater drying effect of the
felt dryer on the thinner felt combined with a lower resistance to the
passage of water vapour. It might be thought that a thinner felt could get
~EAvV FEl..T.
/
LIG'°'T FEL..I.
/
2o()
/
/
/
ri /
::J
0
:r
150 /
IX
.... /
a. /
~
)i. /
r
r loo
Q
j:
0.
~
J
"'
5
\.) 5o
z<!.
w
ti
O·
0 So loo I So
WATER. REMOVED fR.OM. WEE>, K.j. \'ER ~OUR..
Fig. 5 .21. Steam consumption related to water removed from web on an M.G. cylinder
with a light and heavy wool felt (after Ponton)
saturated easier and cease to remove any more water, but in practice if this
were likely to occur a greater felt drying capacity would be called for.
Finally, an isolated but careful comparison by Bowden (32) of the
performance of two dry felts on an M.G. cylinder is worth mentioning.
In this work it was clearly shown that a wool felt removed more water
than a terylene one under quite a wide range of running conditions and
despite the fact that the cylinder surface temperature was invariably greater
with the terylene felt.
To summarize, over the first few cylinders, and particularly when the
440
THE DR.YING SECTION AND CALENDERS SB.2 3
paper is very damp entering the dryers, the fastest drying rate is likely to be
achieved with natural-fibre felts. This is provided the felt is kept sufficiently
dry, which means provision of adequate felt dryers and the use of lighter
weight felts. If the web is not unduly damp entering the dryers (which
should be the case if the presses are used properly), and the ventilation
conditions are good, then a plastic fabric dry felt may be expected to give
as good a drying rate as felts made from natural fibres, and possibly better
if there is no felt drying equipment on the machine. The choice can be seen
to depend essentially on which type of felt is best fitted in the prevailing
operating conditions to accept and remove the water vapour driven off the
web-with good felt drying provisions and when the web is so damp that
liquid transfer to the felt is possible, then natural fibres will be better; with
a well-ventilated hood the plastic fabric should undoubtedly be best.
Furthermore, plastic fabric dry felts may well have a slightly higher steam
utilization efficiency and so be more economical to use.
Towards the end of the dryers the felt itself has to cope with much
smaller quantities of water vapour and it is more important to keep the
temperature of the web high. Provided the resistance to heat transfer
between cylinder and paper is kept as small as possible by applying
adequate tension to the felt, the actual type does not appear greatly to
affect drying rate and efficiency. In this region the other factors for com-
parison of one felt with another, such as economic life on the machine and
dimensional stability, become more important criteria.
On machines making very heavy substances such as paperboard,
adequate tension can often be applied to the web to achieve the close heat
contact with the cylinders that it is the primary function of the felt to
assist; in such cases felts can be dispensed with altogether, and provided
the drying rate is adequate this could well be highly economical both in
eliminating the whole cost of felts and in giving a better steam utilization
efficiency. Normally sufficient drying capacity is only obtained by felting
all dryers except the pony cylinder; in cases where a comparatively gradual
heating-up period is thought necessary the first one or two full-size
cylinders may also be unfelted.
SB. 2 3 Felt tension and arrangement on the dryers
While the material and construction of dry felts have by far the most
important influence on performance of the dryers, there are some other
aspects relating to the felts which require consideration, in particular the
tension at which they are run and the manner in which they are arranged
over the dryers.
It is commonly believed that increasing tension in a dry felt, at least up
to a certain point, will improve the drying rate; this is a consequence of the
closer contact reducing resistance to transfer of heat from the cylinder to
the web. Evidence that this is so has been reported by Smith and Attwood
(7) from work on their experimental apparatus, and by Bowden (32) who
found that tightening a terylene felt over an M.G. cylinder from a very
slack tension increased water removal by about a quarter.
However, Ponton (47), in the experiment on the Swedish Central
441
5B.2 3 THE DRYING SECTION AND CALENDERS
Laboratory machine, could detect no change in water removal occasioned

by altering the tension. Although Ponton's work covered a wide range of
ingoing moisture conditions, from water/fibre ratio of 0·4 to l ·3, Smith and
Attwood's board samples were initially at a ratio of 1·7, and the ingoing
water/fibre ratio in Bowden's work averaged 2·4. The differences in
moisture content of the web are thus the most probable reason for the
discrepancy in the results, and within limits tension may only have a really
significant effect on drying rate in the early part of the drying section when
the web is quite damp. Possibly (as suggested by Nissan) the effect occurs
only when there is some transfer in a liquid state and tighter tension induces
a sustained pressing action on the web.
Except for the dry felts covering the first few cylinders, there thus seems
to be no reason to apply a tension which is at all heavy. The greater the
tension, the more the felt is likely to distort or crease under adverse con-
ditions of moisture change such as occur opposite a wet streak or at a
break. It is also probable that a felt run slacker will have a longer life.
Occasionally, other considerations may require heavier tension in a dry
felt, for example when it is desired to keep shrinkage of the paper to a
minimum, or when some cylinders are not geared and are dependent on
the felt to drive them. Otherwise it is probably more important to ensure
that tension is even across the width of the felt in order to reduce the
chance of creating an uneven drying rate in different parts of the web.
The wrap of a felt over individual cylinders should be as great as is
practicable, a requirement that is always met with in the building of a
dryer section. Apart from this the question arises, how many felts to use
and where to break up the different sections?
With the usual double bank of dryers, for convenience of arranging the
drive gears it is obvious in the first place that different sections are best
divided along the top and bottom in the same way. The two dry felts over
the first few dryers are the most critical and must cope with most water
removal. These should therefore be comparatively short and cover no
more than four or five cylinders each, preferably less. If a drying section
has been arranged with many more cylinders than this in the first section,
it is unfortunately a complicated matter to reduce the number to see if this
brings an improvement; the only alternative is to ensure the provision of
adequate felt drying facilities and in the absence of these an improvement
is very likely if the dryers can be made to run with one of the middle
cylinders of the section used solely as a felt dryer.
How many sections there are following the first depends on the total
number of dryers and the need to regulate machine direction shrinkage.
Often there is only one further section, making four dry felts in all, but on
longer and faster machines further sections may be used to keep the dry
felts to a reasonable length. When this is the only consideration, the length
of each individual section should be made progressively greater towards
the dry end of the machine.
On certain speciality machines, and machines with an M.G. cylinder in
the middle of the drying section, it is common to divide the dryers into as
many as six or seven different top and bottom sections. This is necessary,
442
for instance, where extensive shrinkage occurs in a paper during drying, or
when it is essential to have an accurate control on the amount of shrinkage
that can take place. Each individual section will have a separate drive with
relative speed control to allow the draw between the sections to be set.
Where Sheehan ropes are used to feed up the tail, these also have to be
split into the same sections.
SB . 2 4 Felt drying equipment
The importance of keeping felts dry, particularly when they are made from
natural fibres and pass over the first few dryers, implies that in certain
·circumstances some form of felt drying equipment is essential to maintain
a good rate of drying and efficiency. The small-diameter heated cylinder is
one popular method of drying felts, and to be effective the steam pressure
in this cylinder should be greater than in the main drying cylinders to allow
for the higher heat transfer resistance. The face of the felt contacting the
paper, being the one into which moisture is driven, is made to contact the
surface of the felt dryer.
Other types of felt drying equipment utilize hot air blown through
nozzles directed straight on to the felt or through a rotating perforated
roll over which the felt passes. Alternatively the felt may pass over a
perforated or wire-covered roll to which suction is applied drawing the
warm moist air of the drying section through the felt. Equipment of this type
is usually more suitable when plastic fabric or relatively porous dry felts
are used, especially synthetic which hold water mainly by adsorption.
Various claims are made regarding the advantages of using each of these
types of felt drying equipment, but unfortunately generalization from
individual applications must be highly suspect. Apart from improving the
evaporation rate and drying efficiency, an increase in the felt life and more
uniform drying across the web have also been observed when felt drying
equipment is first introduced. On the first top and bottom felts, two or even
more individual felt dryers may be used, one immediately before the dry
felt meets the first cylinder, the other halfway or less along the felt run. To
introduce a felt dryer in the middle of a felt run between cylinders can
involve a slight increase in the spacing to allow two felt rolls to be used,
and occasionally difficulties occur in keeping the felt running smoothly and
preventing a slack draw to the sheet.
The usefulness of any felt drying equipment requires checking and
perhaps the easiest way of doing this on a comparative basis is simply to
turn off the steam supplied to the dryer and observe what happens. It is,
however, not always easy to assess the significance of any changes that
occur, and it is preferable whenever practicable to divert the felt dryer
steam into the main drying cylinder manifold keeping the position of the
main flow control valve fixed. Changes in moisture content of the paper at
the reel-up should then indicate, other things being equal, whether the
extra steam in the felt drying equipment is being used to the best advan-
tage; note, however, that it may take up to an hour before complete
equilibrium under the new conditions is established.
Little work even of a simple comparative nature has been reported in the
443
5B.3 THE DRYING SECTION AND CALENDERS
literature. Jordan (10) mentions that he found two felt dryers at the dry-end
of a machine (also well provided with a flow of hot air to the felts) were not
necessary; in fact, it appeared that moisture uniformity across the web was
better without them. Although no details are given it is probable from
other data in the article that a measure of the steam consumption was used
as a basis for this comparison.
Rogers and Webster (20) have also reported some work on felt dryers in
which they measured the moisture content of a dry felt by using a beta-ray
gauge. They found that reducing the steam pressure in the felt dryers from
35 to 19 p.s.i.g. caused the dry felt moisture content to increase, although
only from 2·9 per cent. to 3·1 per cent. moisture. In a second experiment,
removal of one of six felt dryers caused an increase from 4· 3 per cent. to
4·5 per cent. in the felt moisture content, but removal of two of the dryers
caused a much greater increase to 5·4 per cent. moisture. It is difficult to
assess the importance of these figures but some spot comparisons indicated
that the modifications had no significant effect on the moisture content of
paper at the reel-up. It may be remarked that the magnitude of changes
measured in the dry felt moisture content appear relatively small, and on a
felt with the normal one or two felt dryers the effect of making similar
alterations will probably be much greater.
The use of hot-air blowing felt rolls in association with plastic fabric
dry felts has been reported by Jender and Gavelin (124) and Race et al.
(130). Improvements in moisture profile were noted in both cases, and in
the latter work some success was obtained in controlling the profile by
altering the position of air flow through the fabric. This approach depends
more on changing the ventilating conditions within cylinder pockets than
on actual felt drying, and as such is not so suitable when used with con-
ventional dry felts.
SB. 3 CONDITIONS INSIDE THE DRYERS

· ·Drying cylinders provide a simple and relatively trouble-free means of
transmitting heat from steam to paper. But to work effectively, steam must
condense evenly over the inside surface of the cylinder, and the heat
released should meet as little resistance as possible as it is conducted out-
wards. Fulfilment of these requirements is considered in this section.
SB. 3 1 The behaviour of condensate in the dryers

It is most important for satisfactory operation of any drying cylinder that
the condensed water should be removed quickly and steadily. Water
allowed to collect in a dryer has several disadvantages depending on the
speed of rotation. At relatively slow machine speeds up to 400 or 500 feet
per minute, condensed water simply forms into a pool at the bottom of the
conventional size of drying cylinder, and the greater volume this pool
occupies, the less is the effective area where condensation can occur on the
inside surface. This is likely to be especially detrimental in bottom cylinders
where the web wraps round the region where the pool forms.
At higher machine-speeds the pool of water begins to climb up the side
of the cylinder producing a growing amount of churning and cascading.
444
This introduces an unevenness in operation which affects the smoothness
of rotation (and hence the draw between cylinders), and is particularly
noticeable for the fluctuations it can create in the power required to drive
the cylinder.
At even higher speeds, above 1,000 feet per minute, the phenomenon
known as 'rimming' occurs when the pool at the bottom of the dryer
vanishes and water spreads in an even layer round the whole inner circum-
ference. This condition is characterized by a much steadier rotation and
power demand, but the layer of water between the steam and the cylinder
surface represents a substantial resistance to the transfer of heat so that a
greater steam pressure is needed to achieve the same drying rate.
2 I
A \
I
\ \
'-
0\~C
0
0 Soo looo 1500 Zooo
SURFACE: 5PE-E.D, FT, PE:R MIN.
Fig. 5. 22. Relation between power required to rotate drying cylinder and the speed of
rotation, for a fixed volume of water in the dryer (after Cooke)
The behaviour of dryers as the speed of rotation is increased through

these different phases has been illustrated in some interesting data present-
ed by Cooke (9) and White (26). Both workers used a single experimental
cylinder in which they could observe the motion of a known volume of
water and measure the power consumed at different speeds of rotation.
Figure 5. 22 from Cooke's paper shows the change in power found
necessary to rotate a 5 foot diameter cylinder at increasing speed when it
contained a fixed volume of water (in this experiment there was no siphon
action and frictional losses at the bearings were allowed for). The line A B
represents the growth in power consumed as the water became increasingly
turbulent; at Brimming occurred, and the power dropped rapidly until at
C it was practically zero. Reducing the speed gave the effect shown between
points C and D; the rimming condition collapsed at a lower speed than it
formed at, the power consumed jumping to its previous high value.
Increasing the volume of water in the cylinder made no difference to the
general shape of the curve shown in Fig. 5. 22 except that the rimming
445
velocity increased, and the power consumed at any particular speed in the
turbulent region was always greater. The increased power required when
the volume of water is greater, coupled with delay in the onset of rimming,
causes the power consumed at its maximum point (immediately prior to
reaching the rimming condition) to increase substantially. For example,
maximum power demand for 100 lb. of water in the cylinder was under 2
h.p.; but in Fig. 5. 22, which related to 175 lb. of water, the maximum
power demand increased to 5 h.p. Thus, speeding up a machine through
this critical region may well become hazardous to achieve smoothly if many
cylinders contain comparatively high quantities of water.
l,'200
tU
r
:::i
z
~
cJ Q.1MMlNG.
~ 1,oo
...
Ul
UJ
u..
t.
0 800
~
'o
Qt
u.
G ~00
0
uJ
I.LI
0.
I.I)
4001.--~~~-A-~~~---ii.-...~~~-"-~
0 o·o2
CONDENSATE.r c.u .fT, PER Sct-F'T: SuRrAC.C.
Fig. 5. 23. Speed of rotation of a 1 ft. cylinder at which cascading, rimming, and
collapse of rimming occur for different volumes of water in the cylinder (after White)
In a further investigation Cooke kept the speed of rotation of the
cylinder at a high value while increasing the volume of water to the point
where rimming broke down. When this occurred there was the same violent
jump in the power demand. Also the greater the speed used, the more water
could be held in the rimming condition before this broke down, e.g. 0· 14 in.
depth at 1,010 feet per minute, 1· 1 in. at 1,310 feet per minute.
White followed a similar experimental procedure, though on a 1 foot
diameter cylinder, and his observations agree closely with those of Cooke.
Power demand in the cascading region was found to be erratic, and the
same abrupt drop occurred when rimming commenced. On reducing
446
speed, the transition from a rimming to a cascading condition was also
sudden, but smoother. Figure 5.23 shows for varying volumes of water
the speed of the drum at which cascading first began, and also when
rimming and (on slowing down) collapse of the rimming condition
occurred. Though these curves refer, of course, to the 1 foot diameter
cylinder, other data indicate that so far as the rimming curve is concerned
the corresponding values for a 5 foot diameter cylinder would be approxi-
mately 1·5 times as high.
White reports that even at very low speeds of rotation there was a thin
film of water on the inner surface of the cylinder the thickness of which
grew with increasing speed. When cascading occurred, it did so in bands
across the cylinder (similar in appearance and origin to the curtains
observable under a table roll) and involved only a relatively small volume
of water. The rimming condition started at one or other end of the
cylinder, spreading quickly across the cylinder.
It is apparent then that under all conditions it is beneficial to keep the
volume of condensate in a dryer to a minimum: at low and medium speeds
primarily because of the effect of a greater volume on the power consump-
tion and evenness of operation, at high speeds primarily because of the
greater resistance to heat transfer through the cylinder offered by a thicker
layer of water. In the speed region where rimming may occur, the situation
is rather more complicated.
In the first place, the critical speeds for the commencement and collapse
of rimming depend on the volume of condensate in any particular cylinder,
and hence vary from cylinder to cylinder down a drying section due partly
to differences in the rate of condensate formation and removal, and partly
to other factors such as roughness of the inner surface. Alterations in speed
through the critical region will thus produce a series of surges in power
demand as individual dryers change their condition; these surges will in
turn increase the likelihood of a break in the drying section as momentary
impulses are transmitted to the tension of the sheet in the open draw
between cylinders.
Where a machine operates in the speed range between roughly 1,000 and
1,600 feet per minute, these difficulties are very likely to occur at one or
other point of the dryers. For speeds at the higher end of this region
fluctuations are likely to be particularly severe, and it would be prudent to
try to avoid the cascading condition altogether by increasing the steam
pressure immediately prior to feeding up the sheet, and temporarily
speeding up the drying section to well above that required. By this means
water in all cylinders is likely to be brought to the relatively stable rimming
state characterized by points C to D on the curve in Fig. 5. 22. Whether
this is satisfactorily achieved in practice could to some extent be judged from
the overall power demand of the dryers, which should become relatively low.
The growing thickness of the water film in a dryer as machine speed
increases in the lower range implies a progressive increase in resistance to
the transfer of heat through the cylinder, necessitating a higher steam
pressure (with its attendant disadvantages) to achieve the same drying rate.
There is some dispute about the extent to which the heat transfer resistance
447
does increase, some contending that convection currents are set up in the
water layer which act as a means of transferring the heat far more effectively
than occurs with pure conduction through water. It has even been suggested
that a reasonably rough inner surface such as occurs with a thin layer of
rust may be beneficial because it is likely to encourage turbulence within
the water film, thereby inducing such convection currents. In any event,
there is little doubt that the thicker the water layer against the cylinder
surface, the higher is the overall resistance to heat transfer, and this will be
particularly noticeable in the transition to a rimming condition where the
layer will increase in thickness and become more stagnant in relation to the
surface.
For these reasons, on some machines running in the critical speed
region it may be preferable to avoid the rimming condition so far as
possible, rather than encourage it to keep the power demand down. Such
a decision must be individual to the machine and though some latitude is
available, in that for instance the setting of siphon tips affects the volume of
condensate held in a dryer and hence the rimming speed, the whole business
of assessing whether and in which cylinders rimming is likely to occur
becomes rather a complex matter, inevitably involving much guess work.
SB . 3 2 Extraction of condensate from the dryers
In old, slow machines the common method of removing water from the·
drying cylinders is by scooping it out in a specially-shaped spiral groove
fixed to the dryer at the back side. No pressure differential is required to
remove water by this means, but of necessity a pool of water must
accumulate in the bottom of the cylinder before the rate of extraction can
equal the rate of condensation. The efficiency of removal has been found to
depend greatly on the design of the scoop (50) and as speeds increase the
size of the pool grows until a point is reached where this method of con-
densate removal severely limits the drying rate. A further difficulty is that
if the cylinders need to be stopped no water can be removed from each
dryer until the pool has crept up to the level of the journal; on re-starting a
heavy and erratic load is placed on the drive and a long time may be
required to get the surface temperature of the cylinders up to their normal
operating value.
These disadvantages become very evident at speeds around 400 to 500
feet per minute, and above this the scoop is generally replaced by the
stationary siphon. With this device a pressure differential is required to lift
the water up to the journal, to which losses along the siphon and conden-
sate piping, and the drop in pressure in the cylinder caused by condensation
of the steam must be added. From estimations of these losses, it used to be
considered that an available pressure differential of from 2 to 3 p.s.i.
between the steam inlet and condensate outlet pipe should be adequate to
ensure evacuation. However, the results obtained by Wahlstrom and
Larsson (109) referred to earlier have clearly shown that with a normal
amount of blow-through steam pressure losses in the siphon and conden-
sate pipes are of much greater significance than had been hitherto suspect-
ed; in fact, 2 to 3 p.s.i. represents the minimum pressure differential
448
suitable under common operating conditions, and especially at low
pressures the available drop required to maintain a reasonable flow of
blow-through steam and reduce the possibility of waterlogging should
probably be at least 5 or 6 p.s.i. across each cylinder.
The siphon itself is frequently just an open-ended pipe dipping to
within a fraction of an inch of the inner surface of the cylinder. The pool of
water collecting at the bottom tends to be pulled up the side of the cylinder,
so it is useful to offset the siphon a few degrees from the vertical in the
direction of rotation; this places the tip more centrally within the pool and
thus reduces the volume of water allowed to collect. As machine speed
increases the impact of cascading water on the stationary siphon tips
becomes much greater until with a rimming condition the velocity of impact
equals the machine speed. This affects the power consumption and in the
rimming condition could also create uneven heat transmission opposite the
siphon tip due to disturbance of the water layer. To reduce the impact and
take advantage of the velocity generated in the water, specially shaped
siphon tips are made which open in a wide horizontal slot towards the
direction of rotation and are designed to guide the water into the pipe as
smoothly as possible. A further improvement in this direction has been
achieved in one type of stationary siphon (Cowie) by attaching a piece of
neoprene to the bottom of the tip which scrapes against the cylinder surface.
The design of siphon tips to provide efficient condensate removal is a task
requiring the specialized knowledge of a steam engineer; a good survey of
the problems involved will be found in reference 131.
When water in the dryer is rimming, the setting of a stationary siphon
tip in relation to the cylinder is highly critical because this determines the
thickness of water held against the cylinder surface. It is difficult to set the
tip really close to the cylinder because allowance must be made for uneven
thermal expansion and dryer eccentricity; one way to overcome this, it is
claimed, is to cut a groove into the inner surface of the cylinder opposite
the siphon tip and allow the tip to run in the groove, but this necessitates
thicker cylinders and is therefore expensive. For these reasons, on many
modern high-speed machines a rotating siphon with the tip clamped to the
surface of the cylinder and joined to a special rotary joint has been
introduced. The advantage of this type of siphon, apart from being
mechanically simpler to design, is that the layer of rimming water can be
kept much thinner: Cooke (9) found that a rotating siphon could keep the
layer less than -b: in. thick, and this was much better than several models
of stationary siphon (including the Cowie scraper type) for speeds between
1,000 and 2,000 feet per minute. It was also observed that a rotating siphon
gave a much more stable load on the drying cylinder drive. Calkins (122)
compared the performance of a rotary siphon set at t in. and at -rt, in. from
the inner surface and found the latter superior at speeds from 1,500 to
3,000 feet per minute and over a range of condensing rates.
It was formerly thought that under comparable conditions the rotating
siphon required a greater pressure differential to ensure removal of the
condensate than with the stationary type; this is because centrifugal force
acting on the water must be overcome to draw it from the cylinder surface
449
SB.3 2 THE DRYING SECTION AND CALENDERS
to the axis, and this force increases with the speed of rotation (see for
example, reference 100). However, Wahlstrom and Larsson (109) have
shown that under normal high-speed operating conditions there is little to
choose in this respect between the two types of siphon. The reason for this
is that above 5 per cent. to 10 per cent. blow-through steam the frictional
pressure loss in the siphon and condensate piping predominates over other
losses of pressure across a drying cylinder, and is actually higher for a
stationary siphon. Only for a very low blow-through does the reverse
apply, i.e. the pressure differential required for evacuation under similar
t!l
V\ 0·2
uJ
z.
~
u
:t
I-
2,ooo L'b5/ l{ou~.
~ O•I
J
u:
0
0
PRESSURE DIFFE.RE:NTIAL..~ P.~.i.

Fig. 5. 24. Dependence of water film thickness on pressure differential in a cylinder
revolving at 2,000 feet per minute, with a t in. rotating siphon, for various condensing
rates and amounts of steam blow-through (after Wahlstrom and Larsson)
conditions is greater for a rotating siphon. For such conditions of low

blow-through the idea of having a suitably-sized hole in the side of the
rotating siphon pipe to add slightly to the amount of steam blowing
through, as reported by Daane (44), will aid efficient evacuation and
reduce the chance of flooding because the siphon can still function with the
tip covered by condensate. Nevertheless with a rotating siphon a consider-
able amount of blow-through must still be maintained during break
conditions: this as a rule requires separate condensers that are connected in
at such times and represent a significant addition to steam costs.
The actual thickness of the water layer when a cylinder with a rotating
siphon is in rimming condition has been found by Wahlstrom and Larsson
to depend on many factors. For instance, Fig. 5. 24 depicting their results
for a t in. rotating siphon illustrates how the water layer thickness varies
according to the condensing rate and the proportion of blow-through
450
THE DRYING SECTION AND CALENDERS 58.3 3
steam under different pressure differentials. Above a certain amount of
blow-through steam the thickness of the water layer is relatively independ-
ent of external conditions; but at lower levels of blow-through the layer
thickness can rise rapidly and unstable conditions will ensue possibly
causing the cylinder to flood. Thus, for efficient operation of a rotating
siphon, particularly in conditions where the condensing rate can change
relatively quickly, it is most important to keep the pressure differential well
above the minimum theoretically required in order to ensure sufficient
blow-through to retain the water layer at a stable thickness. It is for this
reason that rotating siphons need to be operated at a higher pressure
differential than stationary siphons, and the latter may be preferable even
at quite high speeds well within the rimming region when limitations to the
available steam pressure for drying are of over-riding importance.
The performance of rotating siphons also depends closely on the design,
in particular the size of the tip and diameter of the siphon pipe; if too large
the blow-through steam will in practice be reduced by using an orifice or
valve in the condensate line, and this could bring the operating condition
closer to the danger point if fairly wide variations in the condensing rate
occur. To size correctly the dimensions of rotating siphons, the character-
istics of both the siphon and condensate piping, and the condensing rate of
the cylinder, need to be known in some detail; stationary siphons, on the
other hand, always remove more water if the layer thickness increases, and
so do not suffer from the same sort of instability-a wider margin for error
is thus available.
A further disadvantage of the rotating siphon is that should cascading
occur and a pool of water form in any particular cylinder due to a reduc-
tion in speed or to the volume of condensate present, then removal can
become very erratic (26). Thus, apart from other considerations it appears
that rotating siphons are only really advisable when machine speeds are
high enough to make the rimming condition a reasonable certainty, and
also when sufficient steam pressure is available to permit relatively high
pressure differentials across each bank of cylinders. Future developments
may lead to the adoption of a combination of stationary siphon and a
scoop or rotating siphon arrangement, since according to Simmons (131)
this possesses the advantages of both types and requires a negligible
pressure differential to operate even at very high speeds. Another proposal
by this worker is to use a drying cylinder with a concentric inner cylinder
from which condensate is sucked by rotating siphons into the inner cylinder
where it is removed by a stationary siphon. Rimming conditions are
maintained in the outer annulus while the presence of a pool in the inner -
cylinder does not affect the efficiency of heat transmission.
SB . 3 3 Evenness of heat transmission across the cylinders
It is obviously important that both the flow of heat through the thickness
of a drying cylinder and the resultant temperature at the surface are as even
as possible across the machine. This can only be achieved in the first place
if condensation occurs evenly across the inner surface of the cylinder,
which implies a thorough diffusion of steam within the cylinder and an even
451
58.3 3 THE DR YING SECTION AND CALENDERS
distribution of water across the machine at each point on the circumference.

Little investigation has been carried out to check how quickly and evenly
steam actually does spread along the surface of a drying cylinder when
injection takes place through the journal (always nowadays at the back
side). On very wide machines, where difference in heat transmission at
different positions across the cylinders becomes more prominent, some
advantage has been claimed (89) in removing condensate at the opposite
side of the dryer to the one where the steam is injected (in a modern version
of the method still in use on some very old machines). For M.G. cylinders,
spraying the steam from a pipe across the dryer axis, removing condensate
from siphons at both sides, and cutting grooves in the inner surface to
collect the condensate are all reputed to be helpful as well as in some cases
giving an increase in the overall rate of heat transfer.
With the old scoop arrangement it is evident that water has to flow from
the front side to the back (where the scoop is located), which means that
the pool will be slightly deeper at the front side and drying on average
slower at that side. This also applies to the stationary and rotating siphons,
and Calkins (122) has demonstrated that location of the tip at one end of a
cylinder causes a slight but definite difference in heat transfer across the
width of the dryer which could be significant when cumulated through the
whole drying section. With rotary siphons it is easier, especially at higher
speeds, to arrange for the tip to be located towards the middle of the
cylinder, which must give an improvement in this respect. But in this case
there may be a tendency for the condensate to build up rather more near
the edges of the cylinder, giving a slower rate of drying than the middle;
however, this could be advantageous on most machines because, as seen in
SB .4 1, ventilation conditions are generally such as to make it very
difficult to prevent the edges of the sheet evaporating water at a faster rate
than the middle.
These conjectures leave aside the important fact that however even the
transmission of heat may be across the majority of the face of a cylinder,
conditions at the extreme edges must always be different both by virtue of
there being no paper to absorb the heat and also because some heat must
be conducted through the sides of the cylinder. Loss of heat for the latter
reason is small on fully-hooded machines and otherwise is minimized by
such steps as covering the cylinder sides with aluminium paint to reduce
radiation, but overcoming the change in conditions caused by the dis-
continuity at the edge of the web is a more difficult matter.
The presence of uneven heat distribution has been demonstrated by
Campbell and Hughson (43) who measured the surface temperature across
the dryers of an 1,800 feet per minute machine. They reported that this
became higher towards the edges of the web, particularly the front, at a
distance extending between one and two feet from each edge; outside the
edge of the paper the temperature was even higher. The reason for this
higher surface temperature at the edges of the paper was considered to be
either that the condensate rim does not extend to the side of the dryer, or
that steam scours out the sides giving better heat transmission. In the
author's opinion, it is not necessary to postulate that conditions near the
452
THE DRYING SECTION AND CALENDERS 5B.4
sides inside the dryer are appreciably different. At the extreme edge,
where there is no paper to receive heat, the flow must be much lower and
for this reason alone the outer surface temperature may be expected to rise,
in just the same way that the whole dryer temperature rises at a break if the
same steam pressure is maintained inside the dryer. The higher temperature
at the extreme edges outside the web will in turn raise the temperature in the
adjacent surface regions of the cylinder; this, together with the lower heat
demand at the edges of the paper due to more efficient ventilation, is
probably sufficient to cause the surface temperature also to be higher for a
short distance within the edge of the paper.
The outer edges of the paper made on the machine investigated by
Campbell and Hughson were normally very dry at the reel-up, and if the
flow of heat through the cylinders could be reduced near the sides it
appeared likely that this condition would be remedied. It was therefore
decided to study the effect of insulating the inside surface and a 9 inch wide
brick lining was built in at both sides of one of the dryers. This reduced
surface temperature at the sides successfully, and when seven drying
cylinders had been treated in the same way the moisture profile at the dry-
end was completely changed and the dry edges actually became damper
than the rest of the web.
The result of this and other attempts to level up the rate of drying across
the machine are valuable as an illustration of what can be done, though
except for correcting gross differences this method, apart from practical
difficulties, is probably too inflexible for general application particularly on
machines making a variety of grades because alterations can obviously
only be done with the machine shut. Certainly it must involve very careful
experiment over a long period, together with an efficient means of assessing
the changes in average moisture profile before a satisfactory compromise
can be achieved.
Other attempts to overcome the edge condition have included inserting
condensate-retaining rings round the circumference to increase the layer of
water at the edges, and insulating the cylinder with asbestos at the edges.
These appear to give similar results to those obtained with the brick lining,
but such methods are still suitable only for a rough, general correction of
gross differences at the edges.
Research into steam injection and condensate removal systems should
eventually help to even out the conduction of heat across drying cylinders,
but the edge problem is likely to remain less tractable. On existing machines,
obtaining surface temperature profiles of drying cylinders is a useful
exercise; though requiring some care to execute, it is probable that a
sustained programme of checking in this way could show up defects in
individual cylinders due, for example, to such things as the uneven deposi-
tion of rust which would produce cooler spots at different points across the
cylinder.
SB .4 VENTILATION
Adequate ventilation of the drying section is important for the effect it has
both on the overall drying rate and efficiency, and on the evenness of
453
drying across the web. In a larger context ventilation of the machine house
as a whole is also important, particularly from the point of view of giving
comfortable working conditions and preventing condensation.
SB . 4 1 Ventilation of the cylinder pockets
Evaporation of water from the paper web occurs in the free draw between
each cylinder; in fact, as will be seen later, there is considerable evidence
to show that at higher machine speeds a greater quantity of water is
removed in this way than by vaporization on the drying cylinders. In the
constant-rate region of drying where the web is saturated at its surfaces in
the open draw, the rate of evaporation (as discussed earlier) depends to a
close extent on the condition of the air in the immediate vicinity of the web.
In particular, the drier the air and the greater its velocity close to the web,
the faster will be the evaporation. In the falling-rate region evaporation
still occurs in the free draw, though conditions are less critical.
If the drying section were operated without any ventilation of the
cylinder pockets whatsoever, the air in these regions would become largely
stagnant and quickly rise in humidity until little moisture could be absorb-
ed from the paper passing through in the draw. The only circulation would
come from the natural rise in temperature of the air as a result of radiation
and convection of heat from the cylinders; this would induce saturated air
to leak out of the sides of the machine and be replaced by air from the
machine house entering the pockets at a lower level. Apart from the fact
that under such conditions the upper part of the drying section would be
shrouded in steam and condensation would become a serious problem, the
drying rate would be prohibitively slow and completely uncontrollable.
An early improvement on this situation was brought about by the
introduction of various types of air extraction equipment. A familiar form
involves a perforated tube extending across the width of the cylinders and
connected to a manifold at the back side of the machine to which vacuum
is applied. Condensation drains are provided on each extractor and
sufficient insuction of air is created to keep the pockets as free from steam
as possible. As with all pocket ventilation systems, a greater air flow is
usually required in the early dryers where more moisture has to be
removed.
This system has two important disadvantages. Firstly, it is cumbersome,
taking up a lot of room in pockets which are already cluttered with felt
rolls and cylinder doctors, and making removal of broke jammed in the
cylinders more awkward. Secondly, the saturated air removed by each
extractor is made up by air drawn into the pockets from the machine
house; any fluctuation in ambient air conditions created by opening doors,
altering drying on a second machine in the same building, and so forth is
therefore liable to alter the drying rate. Furthermore, air entering pockets
in the early dryers, which needs to be as dry as possible in order to remove
the greater quantities of moisture evaporated from the damp web, is more
likely to have come from the wet-end part of the machine house and
consequently be already very damp.
A further aspect of this type of ventilation is that air entering at the side
454
of the machine picks up moisture as it travels inwards; thus, air in the
pockets at the centre line of the machine tends to be at a higher humidity
and therefore absorbs less moisture from the web (though this will in fact
be partially corrected as the web progresses down the dryers because the
paper temperature will rise in regions where evaporation is least, thereby
increasing the drying rate and tending to even up moisture content across
the web). This lack of uniformity in evaporation rate due to inflow of air
from the sides of the machine is in fact common to most ventilation
systems and is probably the main cause of the tendency affecting all
machines to a greater or lesser extent for the moisture profile at the reel-up
to be drier towards the reel edges. At the extreme edges of the web, the
increased heat flow through the cylinders increases the dryness even fur-
ther, as discussed in the previous section, though that particular effect is
confined to a much narrower region than the hump in moisture content
across the reel resulting from ventilation deficiencies.
The tendency to dry edges is frequently kept within tolerable limits only
by increasing substance slightly from the middle towards the edges and by
over-cambering the presses (or, more likely, by running a lighter load on
the presses than the camber is designed for); both these actions make the
paper damper at the edges when entering the dryers and thus compensate
to some extent for the faster drying rate associated with that part of the
web, but differences in characterisitcs of the paper across the machine at the
reel-up and difficulties in running the press smoothly are inevitable
consequences.
Some allowance for the higher drying at the edges can be made by
altering the design of the extractor tubes so that more air is drawn from the
pockets opposite the centre of the dryers. Likewise, some allowance can be
made on individual machines for the fact that drying tends to be faster on
one edge compared to the other, a situation frequently encountered which
is caused by the masking effect on air flow of gears and other obstructions
at the back side of the machine, and also by air being at a different tempera-
ture and humidity at one side of the machine compared to the other as a
result of the layout and air flow in the machine house. The difficulty with
improving the moisture profile at the reel-up by this form of compensation
is that it requires an extremely lengthy process to achieve anything but a
rough improvement; also any change made to the drying profile in this way
is limited in extent and must be suitable for all grades of paper run on the
machine.
A different approach to ventilating cylinder pockets is to induce
turbulence and promote evaporation of water from web and felts by
blowing in dry, hot air; the equipment takes the form of a perforated tube
similar to that discussed above except that instead of sucking out the
saturated air from the cylinder pockets fresh air is blown in, (this system
should be distinguished from the use of special felt rolls to blow through
hot air-these are dealt with under the heading of felt-drying equipment
in SB .2 4). This permits closer control of the drying conditions than
is possible when air is extracted and replaced from the machine house
because the temperature and pressure of the supply air is readily varied.
455
Also, by supplying heat to the air in the cylinder pockets, less heat is
drawn from the paper as the water evaporates; this is responsible for the
appreciable rise in the drying rate that is usually found when this form of
ventilation system is first installed. Many different designs have been
devised but the most popular involve tubes stretching across the machine
from which air is blown through holes or through jets sunk at intervals in
the body of the tube to avoid protrusions on the surface. In common with
the extraction system, this also has the disadvantage that the tubes are a
nuisance in the crowded cylinder pockets, and as speeds have increased it
has become more important to keep the pockets clear making it easier to
remove broke and prevent it jamming in the cylinders.
This particular difficulty has been overcome in the Grewin system where
air is blown across the machine from nozzles situated at both sides.
Flexible connections to the supply manifold are comparatively simple to
fix along the machine frame and a normal arrangement involves blowing
air alternately from opposite directions into successive pockets. A greater
volume of air is required early in the drying section and two nozzles may
be fixed in each pocket, one below and one above the felt roll, blowing in
opposite directions. The volume of air blown into different pockets may be
regulated by the nozzle opening (adjustment of these may also be used for
compensation when drying is greater at one side compared to the other),
and the supply air is controllable with respect to pressure and temperature.
The pressure is normally set high enough to be able to feel air being pushed
out at the opposite side, but not so high as to set up :flapping at the sheet;
since air issues from the nozzles in a narrow jet which gradually widens out,
at normal supply pressures the effect of any individual nozzle is usually
greater at the opposite side of the machine.
Temperature of the air is set at a level which is economically sound and
must be determined by experimentation and measurement of the total heat
consumption under different conditions. If the temperature is too low, the
web is allowed to cool too much in the free draws as water is evaporated
from it and the only improvement in the drying rate comes from the
improved circulation of air in the pockets; if too high the sheet could
actually heat up in the free draw, reducing the heat flow from the cylinders,
and this is liable to be inefficient because losses will be much greater-also
the paper surface is likely to be affected if its temperature rises too high.
Normally the temperature is in the region of 200 deg. F. to 300 deg. F.
The Grewin system has been widely adopted because of the ease of its
installation on existing machines and the obvious improvement in drying
rate that it generally creates, particularly of course when ventilation had
previously been rudimentary. Especially where drying is limited by the
steam pressure available, such an installation is valuable as it can bring
about a reduction by several p.s.i. in the pressure necessary to sustain
production. Some improvement in the moisture profile at the reel-up is also
usual, though from the nature of the air flow set up by the nozzles it
cannot be expected that dry edges are overcome.
A more recent ventilation system which can be used with particular
advantage in the early dryers incorporates a method of blowing hot air
456
through the cylinder doctors. This avoids the awkwardness of having a
special tube and, especially if means of adjustment of the flow near the
edges of the sheet is available, this method of supplying air to cylinder
pockets appears very promising. Ideally, to obtain true cross-web uniform-
ity of ventilation it will obviously be desirable not only to inject hot air
evenly across the width of the cylinder pockets, but also to extract it
evenly. The use of such doctors together with suitable full-width extraction
equipment, possibly utilizing suction in specially-designed felt rolls, could
achieve this end especially where open-weave fabric dry felts allow easy
passage of air through the felt.
58.4 2 Removing the saturated air; hoods

Whatever the system of ventilation in cylinder pockets, a large volume of
saturated air is invariably expelled out of the sides of the machine and
rises by virtue of its high temperature. Some means of evacuating this
moisture-laden air is essential. Older machines may be covered by no more
than the natural roof of the building, with two or three vents to allow the
steam to escape; in many cases a false roof is constructed over the machine.
Natural ventilation of the machine in this way is uneconomical and too
uncontrollable for modern machines and the usual method of improve-
ment lies in provision of a hood with appropriate air exhaust and supply
fans.
A variety of hoods are available and depend for their construction on
the lay-out of the drying section, the drive, existence of a basement, and so
on. It is normally possible to walk alongside the dryers within the hood and
large access doors are available at intervals for entering and for removing
broke. An essential feature of all types of hood which cover the sides of the
machine is quick and generally automatic raising of the panels to permit
rapid attention in the event of a break. This applies also to covers placed
over the ends of the hood where paper is fed into and out of the dryers.
Material is commonly either asbestos board or aluminium panels which
are light, reasonably cheap, and give a good insulation. Air is withdrawn
from the hood at two or three points along the dryers by means of exhaust
fans and there is frequently provision of some heat-exchange arrangement
for warming fresh air and water.
The simpler 'open' type of hood may take the form of a straightforward
canopy over the top of the machine with side panels extending only about
as far down as the journals of the top cylinders; retraction gear is not
usually necessary in this case and normally there would be provision in the
hood only for exhaust fans (which could be speed-controlled from the
humidity of the exhaust air using a dew-point measurement), the air being
supplied partly by a Grewin or other pocket ventilation system and partly
from machine house fans.
'Closed' hoods may have covers extending to the machine-room floor
and into the basement; with this type it is usual to balance exhaust and
supply air within the hood itself, leaving only a small volume to be drawn
in from the machine house. The supply air can go to Grewin nozzles, to air
457
blowers directed on to felts (especially bottom), to slots discharging low-

velocity air along the lower edge of the hood ('air curtain'), and to other
places within the hood; the air is usually temperature and humidity
controlled.
Elaborate instrumentation is frequently associated with closed hoods
with the object of reducing heat consumption to a minimum and keeping
drying as consistent as possible irrespective of ambient air conditions both
inside and outside the building. The controls adopted depend essentiaJly on
the design of the hood and position of the fans and recirculation system.
As an example: the speed of the exhaust fans can be regulated to keep the
humidity of air drawn from the hood constant, the speed of the supply fans
being adjusted simultaneously to maintain the overall balance; the
quantity of air recirculated to the supply fan can be controlled to keep the
humidity of the supply air at a suitable value; and the dry-bulb temperature
of the supply air can be controlled (by admitting more or less steam
additional to the heat exchanger) at a set-point which can be altered to suit
conditions. With such a comprehensive system it is possible to maintain
drying steady for long periods though simpler systems, for example using
only a damper to control manually the recirculation of exhaust air
according to a measurement of its humidity, are probably sufficient in
many cases.
Where hoods have been installed on existing machines substantial
improvements have been claimed (for example, see references 5, 6, 8 and
10). Drying is often found to be more uniform across the sheet, though this
may be attributed more to the fact that when a hood is installed some
system of supplying hot air to the dryers is generally put in at the same
time. The life of felts over the early dryers is likely to improve due to the
overall drier conditions in the pockets and to lower running moisture in the
felts, while working conditions in the machine house will also be much
improved. But above all the specific steam consumption (as shown by
relating the total steam consumption to the water removed from the paper)
is certain to improve even with a simple open-type hood; it is frequently the
case, however, that no noticeable change occurs in the actual evaporation
rate. When the hood is closed further economies are effected because the
supply air can be made much drier than air in the machine house, thus
reducing the volume of air which must be handled for a given rate of
moisture pick-up.
With any hood heat-exchange equipment in the exhaust air is very
important if the full potential of the hood is to be realized. For instance,
Chalmers (8) has compared steam consumption on three machines making
the same grade of paper as follows: 6,200 lbs. steam/ton paper on a
machine with an open-type canopy hood and 3 exhaust fans but no heat-
exchange equipment; 5,780 lbs. steam/ton on a machine with similar hood
but equipped with four single-pass economizer units; and 5,450 lbs. steam/
ton on a machine with totally enclosed hood and double-pass economizer
units. Even with heat-exchange equipment the steam used for heating the
supply air may represent over a quarter of the total consumed in the
drying section so it is important to experiment with different settings of
458
the supply air humidity and temperature, the exhaust humidity, etc., in
order to get some idea of optimum operating conditions. Fortunately it
appears often to be the case that quite large variations in running condi-
tions can be made, for example by varying the volume of recirculated air
or the temperature of the supply air, without affecting markedly either the
overall steam consumption to dry the paper to a given moisture content or
the moisture profile at the reel-up (5, 6).
Hoods over M.G. cylinders are basically similar to those over multi-
cylinder drying sections though design of the hood to prevent stagnant
regions and stratification of air in different places across the machine is
much more critical. According to Knowles et al. (17), it is necessary to keep
the supply air volume much lower than the exhaust in order to give a large
leakage of air into the hood. The purpose of this is to prevent a layer of air
next to the paper (measured at up to 3 in. thick) from moving with very low
velocity relative to the paper and thus rapidly becoming saturated; the
disadvantage of having a large insuction is that air tends to be drawn in
mainly from the wet-end of the hood where it is already at a high humidity.
When exhaust air is recirculated to the supply line, it is taken preferentially
from the dry-end of the hood. With most ordinary types of hood on M.G.
cylinders it appears frequently to be difficult to get an even distribution of
drying across the sheet and recent years have seen the high-velocity-air
hoods, described in SB. 5, becoming increasingly popular for this particular
application.
SB .4 3 Ventilation of the machine house
On machines without any form of hood, the ventilation system of the
machine house has an important influence on the drying because it governs
the condition of air circulating into the cylinder pockets. Even when some
system of forced ventilation of heated air into cylinder pockets is used, a
large proportion of the volume of saturated air leaving the pockets must
normally be replaced by machine house air. This also applies to open-type
hoods and only a totally-enclosed hood with a balanced exhaust and
supply air system can be expected to function independently of the machine
house air circulation.
It is important for these reasons that the machine house should be
supplied with air by fans at a rate approximately equal to that at which it is
extracted. Otherwise, air will enter or leave the building wherever it can,
making the condition of the air and the drying sensitive to open doors and
windows. Most mills seem to operate with the volume of supply air some-
what lower than the exhaust air volume, but unless the machine is well
hooded the possibility of variable draughts affecting drying cannot be dis-
counted. On machines making grades of paper which require a high degree
of cleanliness, such as photographic base, ingoing air is well filtered and
the volume kept slightly greater than the exhaust to maintain a slight
pressure in the building.
Apart from equalling up the volume of air entering and leaving the
building, it is necessary to regulate the condition of the air blown into the
machine house if the effect of varying atmospheric conditions is to be
459
SB.4 3 THE DR YING SECTION AND CALENDERS
reduced. In this country climatic variations are relatively mild and normal
procedure is to heat air entering the building in winter and accept its
natural condition in summer. But this is done primarily to avoid con-
densation and is often not subject to any close control. In colder climates
with more extreme conditions the effect of variations in temperature of the
air entering the building become more evident and it is more common to
apply a strict temperature control through heaters at the supply fans. In
addition thermal insulation of the building is more carefully considered in
order to reduce natural heat losses. Where the climate is humid (and this
really applies to this country for a lot of the year), and particularly where it
is also hot, it may be economically worthwhile to cool air entering the
machine house in order to condense out moisture; by increasing the
quantity of water that can be comfortably picked up by the air it then
becomes possible to reduce the volume handled, and hence the power costs.
Unless a machine is fully hooded the exhaust-air volume may have to be
i o creased in summer to keep working conditions tolerable.
The choice of positions at which air is blown into the machine house is
equally as important as the condition of the air. This requires study by a
ventilation engineer if adequate circulation through the whole building is
to be achieved, particularly at the wet-end where working conditions can
easily become unpleasant when stock is heated. Currents of air need to be
steady, and planned to clear the hottest places; such things as changes in
humidity round the dryers caused by surges of damp air from the wet-end
into the aisle or down the back side of the machine must be avoided.
Condensation of hot, humid air contacting cold surfaces can be a problem,
especially where drips fall from the roof and girders on to the web and
cause slugs in the paper. With the adoption of closed hoods and the trend
towards reducing the amount of exposed surface at the wet-end, this
nuisance is becoming less troublesome nowadays; nevertheless it is now
common practice to counteract condensation by heating the false roof.
The whole arrangement and siting of fans, and control of the incoming air
condition, needs comprehensive design to cover the extremes of atmospher-
ic temperature and humidity which occur throughout the year. It may even
be desirable to vary the position of fans used at different times of the year,
for example by blowing down to machine-floor level in summer and up to
the roof in winter.
The necessity of heating air entering the machine house at various places
raises the question of how this is to be done. More and more mills are
taking steps to recover as much as possible of the heat leaving the building
in the exhaust air, and in closed-hood installations this is generally con-
sidered as an essential step for the full potential of the hood to be gained.
As much as 40 per cent. of excess heat in the exhaust air from a hood can be
recovered by this means. The simplest type of heat-exchange equipment
utilizes the exhaust air to heat air entering the hood and machine house.
More elaborate equipment like the Ross-Briner Economizer involves a
double-pass system which first heats air up to 150 deg. to 200 deg. F. for
passing to felt-blowing nozzles and other parts of a hood, then in a second
exchan~er heats air entering the machine house. Another arrangement
460
THE DRYING SECTION AND CALENDERS 5B.5
which is becoming increasingly common is to spray fresh or whitewater
into the exhaust air and in that way to recover heat for use in water supplies
to felt sprays, vacuum pump seals, flash steam condensation, and to other
parts of the mill. Supply air to Grewin nozzles is normally drawn in the
first instance from the machine house or the false roof and heated separate-
ly by steam under temperature control.
Recovery of heat in exchangers naturally varies with the season; in
summer, heating of supply air to the machine house would be turned off,
while in winter additional heating may be required as well. But over the
whole year the economic benefit of heat-exchange equipment is obvious
and installation of one or other type may be expected soon to be regarded
as an integral and indispensable part of the paper machine.
SB. 5 AUXILIARY DRYING METHODS

The use of steam-drying cylinders is the basic method of removing water
from the paper web, but apart from this there are other means of drying
which are frequently utilized in an auxiliary capacity. These are usually
adopted to increase evaporation rate on a machine where the dryers are
limiting production, or to shorten the drying section of a new machine;
also in some cases the auxiliary system permits differential drying across
the web and can thus be used as a means of correcting the reel-up moisture
profile. The high-velocity-air hood is the most important device falling into
this category; in fact, the use of this type of hood may now be considered
almost an integral part of an M.G. cylinder arrangement, though applica-
tion has spread extensively to multi-cylinder drying sections.
SB. 5 1 High-velocity-air hoods; general comments
High-velocity-air hoods, or H.V. hoods, were first installed towards the
end of the 1950s. Despite their relative newness there have been a tremen-
dous volume of reports put out about their operation and performance,
and hard competition to enter a new field has led to an unusual degree of
acrimonious discussion, claims and counter-claims. An attempt will be
made in what follows to assess the situation with as much care as possible.
Mention has already been made of the unevenness in drying which
frequently occurs when using an ordinary type of hood over an M.G.
cylinder. This is due largely to stagnant regions in the hood and to the
tendency of the paper to carry a moisture-laden layer of air along with it
over the M.G. cylinder. H.V. hoods were first developed to overcome these
disadvantages by projecting hot air at high pressure directly on to the
paper over the whole of the hood. Nozzles carrying the air are placed very
close to the paper, and after taking up moisture the air is extracted as
evenly as possible over the whole hood area from in between the nozzles;
fans supply air through a heating system at the appropriate velocity,
temperature and humidity, and also extract it. A H.V. hood over an M.G.
cylinder is usually divided round the circumference of the cylinder into a
number of self-contained compartments, though the exhaust from one may
be fed to the supply of the next compartment round the cylinder. It is
461
largely in the design field-the arrangement and shape of nozzles and

extraction area, the size of heaters for the supply air, the velocity-range of
the fans, and so on-that there has been much discussion, but also there
are differences of opinion as to where H.V. hoods should be located on a
multi-cylinder machine. Before going into these points in more detail, it is
appropriate first to make a few general comments on H. V. hoods.
The main purpose of projecting air at high velocity on to the paper
surface is to keep the layer adjacent to the paper as dry as possible, thus
promoting the evaporation rate; the volume of air passing through the
nozzles must be sufficient to create even conditions across the machine and
remove all the moisture picked up. Heating the air serves the twofold
purpose of reducing its humidity and keeping the high heat losses in the
paper caused by rapid evaporation to a minimum.
For multi-cylinder drying the advantages compared to using conven-
tional dry felts are primarily reduced resistance to transfer of vapour from
the web and lowered heat loss from the outer surface. The main disadvan-
tage is that the closeness of contact between the web and cylinder which is
promoted by a felt is lost (although the pressure of the air jets wi11 to some
extent reduce the deficiency); this increases heat transfer resistance from
cylinder to paper and in some applications the consequent drop in heat
flow may outweigh any gain in evaporation rate brought about by the
hood. Gardner (101) has proposed the use of a soft pressure roll, similar to
that used on M.G. cylinders but employing relatively low nip pressures, in
order to improve contact between web and cylinder before passing under
the H.V. hood; this would also have the advantage of reducing trouble
caused by air becoming entrained between the web and cylinder, a nuisance
which becomes more common as speeds increase until apparently a point
is reached when there is a complete air film that effectively insulates the
web from the cylinder and makes it sensitive to small pressure variations on
the surface.
Poor contact between web and cylinder may also be overcome by the use
of plastic fabric felts in conjunction with H.V. hoods, a combination
which largely retains the individual advantages of both provided the felt is
sufficiently open not to interfere with the air jets, yet does not mark the
sheet. Another advantage of fabric felts is that their use overcomes guiding
difficulties which usually occur when several H.V. hoods are used over
successive cylinders. Unfortunately operational reports of this technique
are still scanty.
In the normal H.V. hood application certain practical difficulties must be
overcome. The clearance between the nozzles and paper is very small and
if a break occurs the hood must be rapidly retractable to between one and
two feet from the cylinder. This applies particularly to hoods on bottom
cylinders where broke will easily jam up in the hood and foul the fans.
Retraction is generally automatic at a break, and at the same time the fans
and heat supply may either be automatically stopped, or reduced in output
to keep air in the hood at a constant temperature. Any tendency for paper
to be sucked into the exhaust fan at a break must be reduced by careful
design of the extraction unit; a useful arrangement in one type of H.V.
462
THE DRYING SECTION AND CALENDERS 5B.5 2
hood (Spooner), which makes it easier to clean and to clear broke should
any get sucked into the hood, is to have a reversible axial-flow extraction
fan. Ropes are used for feeding up and the hood is normally lowered only
when the full width of sheet is over the cylinder.
Installing a H.V. hood on an existing machine is generally a complex
task due to the bulk of the hood itself and the crowded conditions which
characterize conventional drying sections. Felt runs round adjacent dryers
have to be modified and to ensure a reasonable wrap of hood it is usually
preferable to alter cylinder spacing. The volume of steam condensed in a
dryer under a new H.V. hood can be substantially higher than formerly,
with a consequent risk of flooding; for this reason it is advisable also to
install individual pressure gauges and valves on any drying cylinder
covered by a hood.
There have been several reports of damage to the paper surface caused
by H.V. hoods. Small slugholes, in which a flap of paper is folded back
from the hole, are caused by the movement under the high-pressure air jets
of pieces of heavy dirt embedded in the paper; dirt carried in the air itself
can produce the same effect, which makes it advisable to filter fresh air
drawn into the hood to a reasonably fine degree. Another fault which is
sometimes observed on M.G. applications, particularly at the edges,
appears in the form of damp patches in the paper which if large enough
may show tendencies to cockle. These are caused when certain regions of
the paper are sucked off the cylinder surface (a result basically of poor
design of the hood); in extreme cases there may be repeated breaks due to
this and the trouble can be overcome only by increasing tension in the
paper or decreasing the air velocity.
Apart from these operational defects, the quality of paper appears to be
little changed when passing under H.V. hoods. Brauns and Larsson (69)
have reported indications that some loss of smoothness can occur with
H.V. hoods applied to a multi-cylinder drying section, especially where the
entering moisture content is between about 40 per cent. to 20 per cent.;
also some increase in cross-machine shrinkage of a greaseproof paper was
observed when moisture content entering a H.V. hood was lower than
35 per cent., though for other types of paper no change in shrinkage was
detected. Jepson (90) considers that H.V. hoods over early dryers can
reduce tendencies towards cockling (this is likely if the original source of
cockling was primarily due to felt unevenness, though even if cockling
originates largely from substance variation the drying rate with high-
velocity-air may be less sensitive to the varying quantities of water to be
evaporated and reduce cockling in this way). On the other hand, where
hoods have been applied to only one side of the web, and especially when
this has been towards the end of the drying section, it has been noticed
that curl is induced in the paper; this could, however, be turned to advan-
tage on some machines where differential H.V. hood drying on top and
bottom cylinders could offer a means of curl control less difficult to arrange
than applying differential steam pressure to the cylinders.
Several designs of H.V. hoods are compartmented at intervals of 12 in.
to 15 in. in order to permit the nozzle air velocity to be varied, usually by
463
adjustment of a damper. Harrison (60), Kerler (91), and Mercer (95) have
described installations of H.V. hoods with this facility (in each case the
hood was sited about three-quarters down the drying section), and there is
no doubt from the results presented that adjustment of the dampers enables
an important improvement in the reel-up moisture profile to be achieved.
It is not possible to control narrow moisture streaks very effectively and
adjustment of moisture by alteration of one damper can often affect the
moisture level in adjacent positions, but with careful use a much more even
moisture profile can evidently be obtained. Some authors claim, however,
that the moisture profile improves anyway when H.V. hoods are installed
and imply that the refinement of being able to adjust operation across the
width of the machine is not necessary or helpful; although the existing
designs of damper arrangement are certainly not perfect, it would be
surprising if the facility were not found economically beneficial on at least
some machines.
A useful summary of the design and operational characteristics of ten
different models of H.V. hood suitable for use on multi-cylinder machines
has been given by Larsson (76).
SB.5 2 Design and operating conditions of H.V. hoods
There are many differences in design of the various makes of H.V. hood
available on the market, and also in the conditions under which it is
recommended that they operate. The primary design feature, and the one
about which there has been most contention, is the shape and arrangement
of the nozzles through which the air blows on to the web; in particular,
should these be circular or slotted to give the best distribution and per-
formance? Allied to this is the question: to be effective how near to the
paper should the nozzles be placed? The optimum gap between nozzle and
paper is dependent on the velocity of emission through the nozzles, and the
choice of this is in turn affected by the condition of the air supplied to the
nozzles.
One of the reasons for the differences of opinion existing amongst H.V.
hood designers is that the mechanism of drying by this means is not
clearly understood. Several workers have formulated theories, notably
Jepson, Gardner, Allander, Hurm, and Daane and Han, but each differs
in various respects from the others; a useful summary of their salient points
has been given by Larsson (76).
Many of these authors have also reported the results of experimental
work designed to simulate H.V. hood operation in various ways, usually
with the object of determining conditions giving rise to the best transfer of
heat from the air to the surface it impinges upon. These too have in many
cases given contradictory results. For example Gardner (58), guided by his
theoretical calculations, found the best performance of several different
arrangements occurred when the nozzles were in the form of slots 0·75 mm.
wide at 1 in. spacings. But Allander and Eneroth (81) compared the
performance of both slots and round holes and arrived at the conclusion
that holes are capable of giving 25 to 40 per cent. higher heat transfer and
should also not be so sensitive to the gap between nozzle and paper.
464
Apart from the question of the actual shape of the nozzles, most authors
are agreed that the greater the number of nozzles in a given area, the better
is likely to be the uniformity of drying; however, the overall percentage
open-area affects the volume of air needed to achieve a given velocity (and
hence power costs), and must be kept low to avoid interference between
jets. The minimum nozzle size has to be consistent with there being a
negligible risk of blockage from dirt and fibre dust. The nozzle size is also
governed by the width of the gap between the nozzles and paper: in general
the narrower this gap can be set at within practical limitations (usually
between-!- in. and 2 in.), the smaller can be the nozzle size without reducing
heat transfer efficiency, though different designs vary in the sensitivity of
their performance to relative changes in these dimensions. A practical
point of importance in this respect is that the gap can alter significantly as
the system heats up and allowance may have to be made for this in the
initial setting-up.
Most authors consider that heat transfer increases with increasing
velocity of air through the nozzles, other conditions being constant, and
that the jets should be directed perpendicular to the paper surface. But
higher velocities demand increases in fan power proportional to the cube of
the velocity, while according to Allander and Eneroth the heat transfer
increases only proportional to speed raised to the power of 0·75. This
implies that to double heat transfer by increasing air velocity would
demand a 16 times increase in the power, yet it is calculated that this will
yield only a 20 per cent. higher evaporation rate. In addition to this,
increased velocities bring more risk of damage to the paper, particularly
for low substances, and there are likely to be greater losses around the hood.
In practice, the range of air velocities used in H.V. hoods is from 5,000 to
15,000 feet per minute.
The condition of the air supplied to the nozzles is primarily a matter of
finding the most suitable temperature. If this is too low, the high velocity
of the air will have a cooling effect on the paper surface which will inhibit
evaporation. On the other hand, very high air temperature may bring
about removal of moisture at a rate greater than can be diffused to the
paper surface; the web then becomes so hot that there is a risk of surface
hardening, and even blistering and burning of the upper layers may occur.
In between these extremes there is agreement that increasing temperature
produces better heat transfer, though higher temperatures demand greater
steam consumption so economic considerations must be taken into account;
on H.V. hoods with sectional cross-web control, it has also been reported
that operation of the dampers becomes increasingly sensitive at high
temperatures (95). Normal operating temperature is from 250 deg. F. to
450 deg. F., though some hoods employ temperatures up to 600 deg. F.
Steam-heating is common, especially for hoods working at lower tempera-
tures, though oil, gas burners and electric heating are also used.
There is normally some form of recirculation of air, with only a relatively
low proportion of fresh make-up air admitted at a rate which can be
controlled automatically to keep the humidity of the supply air at a desired
level. Because this level closely affects running costs, it must be carefully
465
chosen to give the best compromise between the evaporation rate achieved
and the overall heating costs (increasing the volume of make-up air will
lower humidity, thus improving the evaporation rate, but at the expense of
a higher heat demand). As an example of the effect of changing the per-
centage of make-up air, curves given by Jepson (36) are shown in Fig. 5. 25.
These indicate that for this particular hood increasing the proportion of
make-up air above about 20 per cent. yields little improvement in evapora-
tion rate, though the overall specific steam consumption of the H.V. hood
and M.G. cylinder concerned shows a steady increase.
In some cases separate heaters are used, one to keep up the temperature
of the recirculating air, the other to heat the make-up air. This is claimed
lo \oo .
"t:.
0
j::
...tt.i~ ~,o
~ n..
ex: ~Q So~
~v. Vl
z o~ z
0 c; 0
~ ~ IJ
4 '."\~!>-
ri. E>o ? ~o ~
ic.C..\~\"'
0 d:
ILi
0.
<t: t:;>'V ....
> 1,..\.. IA
UJ o'1~Q..P..
~
~ 4o 4o ::J
:::l ~
~ )(
;;( 4:
~ ~
~
u. 2o 2o ~
0
~
~
0 0
0 2o 4o <Oo Bo (00
FRE:$\-I M.A.K.E-UP A.IR A.$ % oF TOTAL. C!~CUL..A.TED.
Fig. 5. 25. Relation of evaporation rate and overall specific steam consumption of an
M.G. cylinder and H.V. hood for varying proportions of fresh make-up air in the hood
(after Jepson)
to allow more careful control under varying exhaust conditions. Exhaust

air passing to atmosphere can be usefully passed through some form of
heat exchanger, similar to those discussed in the section dealing with
machine hoods.
With all these variables in the operation of a H.V. hood, it is extremely
difficult to select between various designs and to know even approximately
what the most suitable operating conditions are like to to be. Each
installation requires careful individual study, with particular emphasis on
the economic aspects and taking into account reductions in operating costs
466
which will occur if any felt dryers can be dispensed with, or if air heating
associated with ordinary drying-section hoods is reduced. The capital cost
of a H.V. hood may well be slightly higher than the number of ordinary
drying cylinders it is equivalent to in terms of evaporation rate, while steam
consumption may be fairly similar. Ideally, once installed a H.V. hood
requires extensive testing under comparable conditions to determine the
optimum running levels of air velocity, temperature and humidity in
relation to overall heat and power consumption-a rather formidable
proposition especially as the best conditions may well vary with the grade
of paper being dried. This, of course, presupposes adequate instrumenta-
tion to control these air conditions, an important feature of any H.V. hood.
One final point must be mentioned. An important difference in operation
distinguishes one particular make of hood (Greenbank) which simply heats
up the saturated exhaust air and uses this for supplying the nozzles (67).
Very little saturated air leaves the hood and replacement is from normal
leakage into the hood at the sides; in effect, this hood uses superheated
steam as the drying medium instead of air. This has allowed a more com-
pact design to be achieved, with fans and heaters inside the hood and no
bulky air ducts. At temperatures above 250 deg. F. it is claimed that the use
of superheated steam gives a greater evaporation rate than air, and yet is
cheaper to use under comparable conditions. This has been hotly contested
by manufacturers of the conventional air-heated H.V. hoods, and an
objective comparison of the two systems is still awaited.
5B. 5 3 Installing H. V. hoods on multi-cylinder drying sections
When H. V. hoods are put on to existing multi-cylinder drying sections,
their performance is apt to vary appreciably according to the conditions of
installation, and the observed increase in drying capacity may be equivalent
to as few as two, or as great as six or more individual dryers in the corre-
sponding part of the drying section (69, 78, 90). Two or more hoods install-
ed in series over top or bottom cylinders are not additive in effect, the
second one has a much lower efficiency than the first while a third may have
hardly any observable effect at all. This phenomenon is generally attributed
to there being insufficient time available under extremely fast drying rates
for moisture to migrate through the thickness of the web to the surface in
sufficient quantity to maintain the same drying rate.
The most important question when installing a H.V. hood is: in what
part of the drying section should it be placed? Many case histories of
individual H.V. hood installations in all parts of the drying section describe
improvements in performance which have occurred, but there are only two
detailed accounts giving direct comparative data.
In the first of these, Brauns and Larsson (69) compared the performance
of two H.V. hoods on the Swedish Central Laboratory experimental
machine; one of these hoods was placed over a lower cylinder about two-
thirds the way along the drying section and had hole-nozzles, the other was
placed over the next upper cylinder and was equipped with nozzles in the
form of slits. Various types of paper were run on the machine and the
moisture content entering the first H.V. hood was varied from between
467
60 per cent. to 10 per cent.; temperature of air to the hoods was also varied
systematically, but otherwise experimental conditions were kept as steady
as possible. Sampling paper before and after each hood and testing for
moisture content enabled the evaporation capacity of each hood to be
determined.
Results obtained for kraft paper are shown in Fig. 5. 26. These illustrate
that the water evaporated diminished in both hoods as the moisture content
entering the respective hoods decreased. This is to be expected since with
decrease in moisture content there is less water remaining to be evaporated.
Also shown in Fig. 5. 26 is the corresponding evaporation for a felt-covered
Ao UPPER 't-\ooo.
la.5 ·c.
LOWER. HOQO.
LG.$"'"'."" 145 °C
t2S "c.
loo •c.
CPo So 4o 3o '20 lo
M'n:;.Tu~e-. CoNTe.I"\ ~.:;) fNT'ER1"'4G \.\ooo.
Fig. 5. 26. Evaporation rate of two H. V. hoods at different air temperatures for varying
moisture content entering the hood ~after Brauns and Larsson)
cylinder. If the water evaporated using the H.V. hoods is compared with
that for the felt-covered cylinder, the ratio between the two varies with
moisture content entering the hoods as in Fig. 5. 27. This illustrates that
the hoods are comparatively more effective for a lower entering moisture
content, in other words they should improve drying rate proportionally
more towards the end of the drying section; similar results were obtained
with newsprint. With greaseproof paper, although the evaporating
capacity of the hoods declined with decreasing moisture content, as for
kraft paper and newsprint, the water evaporated over a normal felt-
covered cylinder declined much more slowly (characteristics of the paper
having more influence on drying rate towards the end of the dryers).
Because of this, the ratio between the H.V. hood and felt-covered cylinder
keeps pretty much the same irrespective of entering moisture content, and
for this grade of paper the position of the H.V. hood in the drying section
would be less important from the point of view of obtaining maximum
increase in drying rate.
The difference between the performance of the two hoods as shown in
Fig. 5. 26 deserves some comment. The lower hood gave a better per-
formance than the upper, except at high entering moisture content when
the reverse was the case. The reason for this is attributed to the fact that at
468
THE DRYING SECTlON AND CALENDERS SB.5 3
high moisture contents the lower hood acted partly as a means of heating
up the web to the advantage of the upper hood which then followed; at
lower moisture contents the more common effect of having two hoods in
series explains the lower performance of the upper hood. Apart from this
point it is not possible to draw any firm conclusions as to comparative
performance of holes or slits because the hoods were not interchanged,
though there is perhaps an indication that the performance of hole-nozzles
is less dependent on the entering moisture content.
One further point about these results which should be noted is the effect
of varying air temperature in the hood: this becomes of increasing im-
portance at higher moisture contents. It may be expected therefore that
Lo"'1Ei:t l-looo.
I IO S °C.
/~ loo
0
C..
:/:°/ I '4 s °C.
,, I oo 0 c.
UPPEl< t\ooo.
0
7o So
,Mo1;:.TuR~ CON1'EN\ % .SNTE.~INC. H.ooo.
Fig. 5.27. Ratio between evaporation rate of two H.V. hoods and that of a felt-
covered cylinder for varying moisture content entering the hood (after Brauns and
Larsson)
towards the end of the dryers no advantage will accrue from having the air
temperature higher than normal.
The other report presenting comparative results on using H. V. hoods in
different parts of the drying section is due to Jepson (90), who cites
observations made on a paper machine to which ten H.V. hoods were
fitted over the top cylinders of a 21-cylinder drying section. The hoods were
arranged in three separately controlled banks comprising 4, 3, and 3 hoods
over individual cylinders. The speed which the machine could be run at
(without felts and presumably at maximum available steam pressure)
increased 15 per cent. with the first bank of hoods operating, 24 per cent.
with the second bank alone, and 34 per cent. with the third bank alone;
with all three hoods working the increase in speed was 55 per cent. In other
words the hoods at the dry end of the dryers had a greater effect, other
conditions being equal. It will be noted, as observed above, that the effect
of using all the hoods together is not additive.
These two reports thus show quite strongly that H.V. hoods are likely,
in general, to be more efficient in the later stages of drying. This need not
always be the case, as for example with greaseproof paper, and the position
16 469
chosen for any particular application needs to be related to the drying

curve of the paper; where this is comparatively slow over the early dryers
a greater improvement may occur with a H.V. hood in that position. The
reason for the greater efficiency of H.V. hoods when the sheet is compara-
tively dry is not absolutely clear, but it must be remembered that the basis
of comparison is with felt-covered cylinders. Removal of a felt in the early
stages of drying gives a bigger reduction in overall evaporation rate than
removal in the falling-rate zone of drying where the conventional multi-
cylinder section is less efficient. Also, installing a H.V. hood at the begin-
ning of the drying section leaves the remaining dryers to operate in the
less-efficient region. On both these counts, when substituted for felt-
covered cylinders H.V. hoods are likely on a purely comparative basis to
show greater improvement if installed towards the end of the dryers.
SB . 5 4 Other auxiliary drying techniques
There are a number of other methods of assisting the drying process, of
which the more important worth noting are infra-red drying and drying
under vacuum.
The use of infra-red heating from oil or gas burners or electricity has
been strongly advocated from time to time but does not seem to have
caught on to any extent. Undoubtedly the main reasons for this are that the
cost can be prohibitive, and there is often a fire hazard involved; further,
there is no general agreement as to how efficient particular installations are
likely to be nor where is the best position to put them in the drying section.
Bhargava and Robertson (39) have given a comprehensive review of the
problems involved in using infra-red radiation, in particular of the need to
select carefully the predominant wave-length of emission (which depends
basically on the temperature of the source) in order to ensure maximum
absorption by the web. These authors state that even with high-efficiency
energy usage the cost of evaporating water from a paper web can be
approximately twice that of steam (though this, of course, depends on
relative fuel costs) though capital expenditure for an equivalent capacity is
lower. Experimental work showed that heat is absorbed, and water
evaporated, primarily in the surface layers facing the heater; the rate of
evaporation is fairly constant irrespective of moisture content of the web,
in comparison with drying cylinders where there is a steady decline. This
suggests that infra-red radiation is more likely to be successful when used
to augment the later stages of drying.
Other authors (for example, 75) agree that infra-red drying is best applied
when moisture content of the paper is low, though some, including
Burgess (57), consider that it could be applied more effectively to augment
drying in the constant-rate period when moisture is present on the surface
of the paper. Eisele (45) has conducted laboratory experiments which
showed a decline in the effect of infra-red radiation at moisture contents
lower than 25 per cent. to 30 per cent., thus contradicting those of Bhargava
and Robertson; this worker also found that the results obtained using
infra-red depended on several other factors such as thickness of the paper
and exposure time. Prince et al. (64) is another author who believes that
470
THE DRYING SEC TI 0 N AND CALENDERS 5B. 6
infra-red should be applied before the drying section to obtain the best
performance; he considers that the heating effect from an early infra-red
application enables temperatures used in the first few dryers to be higher
without the risk of picking or burning the web. It is evident from all this
that at the present time the value of infra-red radiation as an auxiliary
means of drying, and the best method of applying it to a paper machine,
are by no means clear.
One possible application of infra-red drying is to allow correction of
cross-web moisture irregularities by dividing the heater across the width of
the machine into a number of individually controllable banks. For this
purpose it is more likely that an installation will work efficiently if placed
at or close to the end of the dryers. Incorporation of such an arrangement
in a H.V. hood has also been considered.
Drying under vacuum is theoretically a sound idea because evaporation
then occurs at lower temperatures. This either makes an appreciable
increase in drying rate possible on any particular machine, or can approx-
imately halve the number of cylinders needed to achieve a given rate of
drying. The Minton Vacuum Dryer operates on this principle and involves
placing all the dryers in a sealed hood to which vacuum of up to 28 inch
Hg is applied; several successful applications have been described by Hill
(46). Brightness of the paper is claimed to improve though there are
difficulties in obtaining satisfactory sizing due to the prevailing low
temperature of the web which may not exceed 100 deg. F. The main
problems with this device lie in the need for complex seal arrangements,
especially where the sheet enters and leaves the dryers, and the trouble
which may occur clearing jams if there is a break in the drying section.
5B.6 MACHINE SPEED AND DRAW CONTROL

If the speed of a machine is increased while making a particular grade of
paper, for moisture at the reel-up to remain the same the rate of evaporation
in the drying section must increase pro rata. This is achieved in practice
by some increase in the steam pressure in the cylinders, but it has long
been realized that the actual increase needed is much less than would
be required were the higher pressure solely responsible for producing
the greater evaporation rate. The reasons for this will now be discussed.
5B. 6 I Changes in drying conditions with increased machine speed

To understand what happens in the drying section when machine speed is
increased, it is necessary first to consider drying on an individual cylinder.
Reference to Fig. 5. 5 shows that a certain proportion of the surface of a
cylinder is utilized for raising the temperature of the web, and it is only in
the later portion of the dryer surface that temperature levels off and a
reasonable evaporation rate can be achieved. With an increase in speed a
greater rotation of the cylinder occurs in a given time, so with other con-
ditions constant the same heat can be transferred through the cylinder only
after the paper has travelled further round the dryer. In other words, the
proportion of the cylinder needed for heating up the web increases, leaving
471
5B.6 1 THE DR YING SECTlON AND CALENDERS
less area available for evaporation at near-maximum rate. Consequently a

rise takes place in the moisture content of the web leaving the dryer.
Figure 5.28 shows experimental results reported by Jansen and Nordgren
(35) which illustrate this.
Although with an increase in speed the reduction in moisture content of
the web achieved by an individual dryer is thus diminished, the capacity of
the dryer, i.e. the total water evaporated in unit time, actually increases. In
other words, doubling speed with all other conditions the same does not
lead to a reduction by half in the difference between the moisture content of
di 0·3
Ii.I
~
u.
0
Ill o.(p
"'z
OJ
>-
«.
0 0·4
~
~
"' 0•2
"~
x
\.) C>
2 1·5 l·o o·5 o
WATER/FIBRE IN WEe. e,eFoR.i' CVLlMOE-R.
Fig. 5. 28. The increase in dryness of a web passing over a drying cylinder at two
different speeds for varying moisture content before the cylinder (after Jansen and
Nordgren)
the web entering and leaving the cylinder. This is primarily due to the
increased heat transfer that must occur (for the same internal temperature)
when the average temperature on the cylinder surface decreases as a result
of the longer heating-up period; thus, at any point round the cylinder
circumference the outer temperature will be lower at the faster speed, hence
the heat conducted through the cylinder in unit time will be greater. In
addition to this effect, felt covering a given area is required to receive a
lower quantity of evaporated water, and thus may provide less resistance
(if any, that is) to transfer of further water into the layers adjacent to the
sheet.
To compensate for the lower quantity of water removed from a given
area of the web, the steam pressure of the cylinder has to be increased. It
does not follow from this that the efficiency of the drying section is
drastically altered; in fact, except insofar as a greater steam pressure
implies higher operating temperature and therefore leads to slightly
greater heat losses, the weight of steam needed to evaporate unit weight of
water in the web should theoretically be unaffected by the speed of the
472
THE DRYING SECTION AND CALENDERS 58.61
machine. The usual difficulty that occurs is caused simply by the steam
pressure rising to a point where it reaches the limit of what can be supplied.
In practice though, as the machine speed increases a change takes place
in the amount of evaporation occurring in the different parts of the drying
cycle for each cylinder, and a greater proportion of the total evaporation
occurs in the open draw between cylinders as opposed to on the cylinder
itself (partly this is a consequence of quicker evaporation due to increased
too
ri
IJJ
0
't.
J
).
u
z
Q
80
~
z
fi
rt:
:J
\)
\.I Go
()
z
a
~
ff.
0
0.. 4o
~
uJ
u.
Q
llJ
~
(
..... 2o
z:
llJ
0
tt
w
0..
0 loo 200
MACHINE;
Fig. 5. 29. Relationship between the percentage of total evaporation occurring on
drying cylinders with the machine speed (after Jansen and Nordgren)
air velocity relative in the sheet). This change in character of the drying
may be expected to have a more direct effect on the efficiency of steam
utilization. Figure S. 29 (though based on only a small amount of data from
different sources) gives an idea how much the amount of evaporation
occurring while the web is in contact with drying cylinders, as opposed to
the open draw, alters with increasing machine speed. Due to this effect, the
473
58.6 2 THE DRYING SECTION AND CALENDERS
increase in cylinder steam pressure necessary to sustain higher machine

speeds is kept within manageable limits, and it is evident that ventilation
conditions in the cylinder pockets become more and more important as
speed increases.
It has been suggested that the draw between the cylinders of higher speed
machines should be lengthened to assist evaporation in this region. But
most water evaporating from the web does so by flashing off almost
immediately after leaving the cylinder (this is demonstrated by the rapid
fall in web temperature shown in Fig. 5. 5); also longer draws become more
difficult to manage. It is therefore unlikely that much improvement would
follow from deliberately extending draws beyond the length demanded by
the geometry of the cylinder spacing.
To improve the evaporation taking place actually on the cylinder
surface, some increase in drying cylinder diameter appears desirable for
machines running at high speed. This is usually needed anyway as the
cylinder must withstand greater steam pressures and remain in balance
dynamically at a higher speed of rotation. Theoretically, if no other effect
were present the diameter would need to increase proportionally with the
speed in order to maintain the same rate of drying round its circumference,
But at high speeds this would soon require cylinders of inconvenient size.
and also the effect of felts may well diminish because (for the same tension)
pressure on the cylinder reduces as the diameter is increased.
58. 6 2 The influence of draw
In the drying section, the draws between different sections and between the
last bank of dryers, the calenders, and the reel-up are generally very small
compared to customary draws in the press section. In most cases they do
not exceed at most 1 per cent. at any single draw. But it is equally as
important to have a satisfactory control on these draws because relatively
quite small changes can bring about a significant alteration from the point
of view of shrinkage of the web. As the paper dries, its extensibility under
a given tension reduces; this means that a particular degree of change in the
draw has a much greater effect on the tension of the web in dryers com-
pared to the presses-in other words, the draw is more sensitive in the
dryers.
Changing the draw between different parts of the drying section and
calenders affects the amount of web shrinkage in the machine and cross
direction which is permitted to take place. This not only has a direct
effect on the reduction in deckle of the paper and hence on production, but
also changes occur in the strength properties of the paper to a degree
dependent on the magnitude of natural shrinkage that would occur were
the paper able to dry free from tension. It also affects anisotropy of the
paper. These and related topi9s have been fully discussed in the theoretical
section.
As it is important in this part of the machine that draw is adjusted to
keep the tension in the web as steady as possible, a strong case can be made
out for having an indication of the tension at each draw; in fact, it is
probably more important to have this than a draw measurement, though
474
THE DRYING SECTION AND CALENDERS 5B. 7
the latter is also valuable as an indication of changes occurring in behaviour
of the sheet. The simplest method of indicating tension is from a full-width
or narrower free-riding roll, the journals of which act against restraining
springs; the position of the roll then gives an indication of the tension in
the web and the movement may be magnified to give a suitable means of
display in any convenient manner. Alternatively, strain gauges (load cells)
in the roll supports may be used to give a direct measure of the force
exerted on the roll by the web, and in this case movement of the roll itself
need only be very small.
Changes in stock composition, particularly those associated with free-
ness, affect the natural shrinkage occurring during drying quite closely, and
in a poorly controlled stock preparation system the tension of the web
between draws may fluctuate continually; this necessitates repeated
alterations to the draw to maintain a steady tension. On machines making
grades of paper in which shrinkage properties have to be closely controlled
and repeatable, it would be useful to control the tension by automatic
changes to the draw. The feasibility of this depends entirely on the method
of obtaining the draw and also on the type of drive.
58. 7 PAPER PROPERTIES

The performance of the drying section is directly affected by certain
characteristics of the paper. This is because, particularly in the later stages
of drying, the fibre network structure comprising the web influences the
rate of migration of water through the sheet and hence the ease of drying.
In this respect, the quantity of steam required to evaporate a unit weight of
water and the general shape of the drying curve, i.e. the value of moisture
content as the sheet progresses down the dryers, differ considerably for
different grades of paper; several examples for various types of paper have
been reported by Montgomery (11) and it is interesting to note that the
shape of the drying curve for each grade is generally similar for different
machines, thus confirming the relative importance of the paper structure.
Apart from this, the main effect on drying of changing paper properties,
certainly during any one making, is indirect and depends essentially on
changes in the moisture content of the web entering the dryers. If this
moisture content increases, then the quantity of water requiring evapora-
tion in the cylinder section also increases and a greater volume of steam is
necessary (though not a pro rata increase-see 4I. 1). Moisture content
leaving the presses is dependent to some extent on moisture content at the
couch, and this in turn is affected by drainage conditions on the wire.
Further, the performance of the presses themselves depend on the charac-
teristics of the paper; for example, other things being equal an increased
freeness in the stock gives an increase in dryness of the sheet leaving the
press section.
Due to these effects, changes in the structural properties and composition
of the paper running on a particular machine have a definite influence on
steam demand. For this reason it can be particularly valuable to have a
measure of the flow of steam to the drying section for comparison with
475
SB. 8 THE DRYING SECTION AND CALENDERS
various other measurements indicating, for example, drainage conditions

on the wire and performance of the presses; over relatively long periods
changes in steam pressure and flow are likely to be a direct consequence of
differences in demand caused by an alteration of the moisture content of
the web entering the drying section (particularly this is so when moisture
content at the machine reel-up is well controlled), and the reasons for these
changes can more readily be traced when a record of steam usage is
available.
SB. 8 THE M.G. CYLINDER

The M.G. cylinder (Yankee cylinder in the U.S.A.) was first developed to
impart a one-sided, highly smooth finish as an integral part of the making
machine, hence the term 'machine-glazed'. This is still the primary purpose
of the cylinder when used for such grades as kraft and sulphite wrappings,
envelope and bag papers, etc., though the M.G. is also used extensively on
tissue machines. This is primarily because of the high rate of drying that
can be achieved, but also because it enables open draws to be avoided,
at least until the web is almost dry and sufficiently strong to withstand the
tension developed at the high speeds of manufacture associated with
this product. On some machines a more specialized function of the M.G.
cylinder is to introduce a crepe effect.
In this section it is proposed to discuss aspects particular to the M.G.
cylinder, as opposed to the ordinary drying cylinder. Arrangements for
steam supply, hoods, and condensate removal, in so far as they differ from
multi-cylinder drying, have already been mentioned in the appropriate
sections.
SB. 8 1 General comments on M.G. cylinders
The M.G. cylinder can be the only form of drying on a machine, but it is
increasingly common nowadays for a number of conventional drying
cylinders to be incorporated before or after the M.G., occasionally both.
This enables an increased production to be attained, and also permits more
flexibility in determining the degree of gloss imparted to the paper. After-
dryers are particularly common on machines making creped paper because
it is easier to produce the crepe effect when the paper is slightly damp; in
such cases a negative draw is needed between the M.G. and the after-
dryers (continuous indication of this draw in some form is an essential
prerequisite to successful production of this grade). Some machines are
constructed so that the position of the reel-up can be quickly adapted to
leave out after-dryers when required; this facility permits a useful extension
of the range of papers that can be made on the machine.
The size of the M.G. dryer makes it easy to lead the web on to the
cylinder quite close to the bottom, thereby utilizing as much of the surface
as possible. However, this does mean that the wire side of the web contacts
the M.G. surface, and with Fourdrinier papers this is always initially the
rougher of the two sides so inevitably the contrast obtained is not quite so
great. To attain the maximum two-sided effect, the top side has to contact
476
the M.G.; this either involves rather clumsy felt runs and draws at both
sides of the cylinder, or the M.G. has to be sited low down in a basement,
both rather unsatisfactory arrangements. Earlier M.G. cylinders were
covered with a felt but this practice, never very easy to operate nor parti-
cularly efficient, is probably now obsolete and has been replaced by hoods
of the ordinary and H.V. type.
Care of the cylinder surface is one of the most vital features of M.G.
operation because any blemishes on the surface are readily transmitted to
the paper. Doctors, wire scrubbers, emery cloth, and other materials are
used continuously to keep the surface burnished as the cylinder rotates,
while other devices such as the electro-doctor which prevents corrosion by
applying an electric charge to the cylinder surface are claimed to be of value
in certain operating conditions. A more drastic operation involving buffing
first with coarse then with finer-grit wheels, followed by other methods of
polishing with lubricants, is used periodically to brighten up a surface
dulled by constant use. The treatment most suited to any particular M.G.
cylinder must be found by experiment as deterioration of the surface
depends very much on the grade and contacting moisture content of the
papers and on characteristics, particularly hardness, of the water used at
the mill. The ideal degree of burnishing keeps the cylinder adequately
smooth and bright without too drastic an action; excessive polishing of the
surface is not necessary and beyond a certain point does not enhance the
gloss imparted to the paper.
The size of an M.G. cylinder makes it necessary to be manufactured
with a thicker wall than ordinary dryers to an extent dependent on the
dryer diameter and the strength of the metal. This produces an increase in
the resistance to transfer of heat through the cylinder wall, and to overcome
this and obtain a satisfactory temperature at the outer surface, higher
steam pressures are necessary. But higher internal pressures in turn produce
greater stresses which the cylinder must withstand, and mean that the wall
thickness has to be increased still further. The best compromise in thickness
and operational steam pressure to overcome this vicious circle has been the
subject of much calculation and discussion which is too specialized to
consider here; an article by Chapman (51) may be referred to for a sum-
mary though no straightforward solution appears to be available yet. The
size and thickness of the M.G. cylinder also make it essential to prevent
rapid changes in temperature in order to avoid setting up thermal stresses;
for this reason M.G. dryers are usually kept warm over a shut week-end
and heating up in preparation for start-up has to be at a very carefully
controlled rate (nowadays frequently using an automatic device which
controls the rise in temperature to a pre-set pattern).
To overcome the natural deflection of the cylinder on its journals, a
camber is put on the surface which as closely as possible corrects for this at
normal operating temperature. Uneven action of doctors can produce
changes to this camber which ultimately lead to running problems; the
camber must therefore be checked at intervals and corrected when neces-
sary by grinding. Larger M.G. cylinders are slowly rotated when the
machine is shut to prevent distortion under their own weight.
477
SC.8 2 THE DRYING SECTION AND CALENDERS
There are normally two or three doctors on the cylinder serving different
purposes. Apart from helping to keep the cylinder clean and polished,
doctors have a variety of functions: to hold emery cloth against the surface,
to hold a damp felt for easier removal of fluff and dust (a steam jet
immediately in front of an ordinary doctor sometimes serves the same
purpose), for creping (with a special doctor adapted for rapid changing
because the high angle of contact produces rapid wear), and for cutting off
the paper above the creping doctor when the latter is being changed. These
doctors require much closer attention than do doctors on ordinary dryers
and merit a careful record giving details of their life and the frequency of
changing so that the effect of alteration in hardness, angle of contact, load,
etc. can be examined. Oscillation is absolutely essential to prevent the
possibility of scoring the cylinder surface. The load on the doctors should
also be easily adjustable and examined frequently.
5B . 8 2 The pressure roll
Intimate contact between the web and the M.G. surface is obtained by
pressing the web hard on to the cylinder with a 'pressure' roll. This is an
essential and critical part in the operation of any M.G. cylinder, and it is
the close contact achieved in this way which is responsible for the high
evaporation rates associated with M.G. dryers-generally about double
the equivalent rate over the same area of ordinary felted drying cylinder.
The closeness of contact also governs the glazing effect of the cylinder.
Pressure rolls are made of rubber and are cambered to suit the pressure
normally applied. Systematic attention to the camber is extremely im-
portant for obtaining an even gloss across the web and keeping the felt
properly guided; records for pressure rolls must be kept with as much care
as for press rolls. Occasionally a suction pressure roll is used, mainly to
help in preventing the web from being thrown off as it rotates round the
roll, but this practice appears to be confined to machines making crepe
paper because shadow-marking may be noticeable in a flat sheet. Some
M.G. cylinders operate with two pressure rolls, the first to transfer the web
to the cylinder and the second (using the same or a different felt) to
improve intimacy of contact, but the value of this is not certain and trouble
can occur with air-blowing and distortion of the sheet at the second pressure
roll.
The pressure applied at a pressure roll closely affects the resulting gloss
of the paper. Chapman (51) devised a simple laboratory experiment giving
a static simulation of M.G. operation and obtained the results shown in
Fig. 5. 30; gloss was measured on an Ingersoll glarimeter. He also reports
that a programme of increasing M.G. nip pressures on a number of
machines lead to a substantial increase in average gloss. The curves in
Fig. 5. 30 indicate that improvement in gloss with increase in pressure
approaches an asymptotic value so there is obviously a limit to the degree
of improvement possible in any particular case.
For a given load on the roll the pressure varies with nip width, and
hence with the felt condition and hardness of the rubber used in the roll.
Normally a relatively hard nip is desirable to attain maximum gloss,
478
though on some grades where this not important, e.g. crepe, a softer
pressure roll and more compressible felt help to secure more uniform
adhesion across the machine at the creping doctor and can therefore be
preferable. The load applied to the front and back sides of the roll can
easily be measured, directly or otherwise as with press rolls, and it is
general practice to have an indication of this.
The type of felt used in a pressure roll nip is partly controlled by the press
conditions (the same felt passing through the final press and the M.G. nip);
on some grades the felt is also used to vary the top side finish of the paper,
9o
~ STUPF FA.eENES&.
aII.! 4o5
....
uJ
~ eo 292
a
-d'.
J
~
'-...I
"0
t:.
<
w
d. 70
Ill
!ii
0
_J
\!)
PRESSURE P.01..1.. PRfSSURT:, P.s.i.

Fig. 5. 30. Gloss variation of an M.G. paper at various pressure roll pressures and
stuff freenesses (after Chapman)
the appropriate face of the felt being woven in a special pattern, e.g. ribbed
or rough. Ample cleaning equipment is needed for this felt to avoid
adhesion difficulties as gradual plugging occurs throughout its life (the
sheet then tending to follow the felt on the trailing side of the pressure roll
nip); also the variation in gloss that results from narrowing of the nip
width as the felt becomes less compressible with age is minimized by
keeping the felt uniformly clean. Continuously operating full-width
479
58.8 3 THE DRYING SECTION AND CALENDERS
showers with the addition of a small quantity of detergent (as for ordinary
press felts) are likely to be the best means of achieving this, especially on
machines where the web is couched and carried on the same felt through
the press section and on to the M.G. cylinder.
58. 8 3 Factors affecting the gloss imparted by an M.G. cylinder

The gloss imparted by an M.G. cylinder is frequently the most important
reason for its use. Two factors affecting the gloss have already been men-
tioned: the smoothness of the cylinder surface and the pressure applied at
the pressure roll. Other variables will now be discussed.
The moisture content of the web when first contacting the M.G. cylinder
is very important. Chapman cites an example of the effect on a plain kraft
paper of decreasing initial moisture content over the range 50 per cent. to
47 per cent.; the gloss diminished appreciably as the moisture content was
reduced, particularly towards the drier end of the range, and the general
appearance of the sheet also deteriorated and exhibited more cockling.
Other reports indicate that low initial moisture content in the web also
leads to poorer contact (with consequent lower drying rate, though as less
drying is needed anyway this may not be a disadvantage); in addition, a
more uneven glazing effect is observed (presumably due to less uniform
contact and greater sensitivity to moisture content variations across the
machine). To avoid these disadvantages the presses may in some cases
have to be raised or not used at all in order to increase moisture content at
the M.G.; more commonly, on faster machines the temperature of a
number of pre-dryers is varied to suit conditions, though too many pre-
dryers can become a nuisance because occasionally it may be found difficult
to reduce their drying effect sufficiently.
The moisture content of the sheet when leaving the M.G. cylinder also
affects the gloss, this being greater when the paper is taken drier off the
cylinder. Thus the overall drop in moisture content over the M.G. has an
important effect on the gloss attained: if the moisture drop is reduced,
either by lowering initial moisture content or raising the leaving moisture,
then the level of gloss will to some extent fall. The drying rate of an M.G.
cylinder is therefore a critical factor in determining the production possible
for a given quality: a deficiency in the drying rate of the M.G. though it can
be made up by using more steam in either or both pre- and after-dryers, is
overcome only at the expense of a loss of gloss which may be unacceptable.
All that can be done in such circumstances is to pull back the speed to
permit the M.G. cylinder itself to reduce the moisture content of the web
by an amount sufficient to give the gloss required.
After-dryers are generally used only for papers requiring a low glaze. But
apart from this, if the moisture content at which the sheet is pulled off the
M.G. is too high, problems with adhesion to the cylinder occur. Adhesion
has been found to depend closely on moisture content, reaching a maxi-
mum at around 30 per cent. moisture depending on the grade of paper and
condition of the metal surface; thus if the machine is run in such a way as
to cause the sheet to be too damp when it is removed from the M.G., th~
480
THE DRYING SECTION AND CALENDERS 58 .9
high tension necessary to overcome sticking can cause rupture problems
and the paper surface may be disturbed.
Other factors affect the gloss to greater or lesser extent. According to
Chapman, if the fresh stuff becomes freer or if the proportion of moist pulp
is increased, in both cases average gloss is raised. Occasionally spraying the
pressure roll felt with steam or water is found essential to obtaining a
satisfactory gloss, but this presumably is simply a result of increasing
moisture content of the web at the pressure roll nip.
Oil or glue is sometimes sprayed on to the M.G. surface to improve
adhesion and this also gives an improvement in gloss, possibly by virtue of
the improved rate of drying which allows the initial moisture content of the
web to be raised. This practice is also used for creped papers, it being
claimed that the better adhesion gives an improved crepe effect.
SB. 9 CALENDERS
fn this section some general points concerning calender operation will be
discussed. Calenders are used primarily to improve the overall smoothness
of a sheet of paper and in any particular stack two principal factors affect
performance in this respect: the number of nips and the moisture content
of the paper. Both these have been dealt with earlier in the theoretical
section. A secondary function of calenders is to assist, by use of the air
blowers, the building up of an even machine roll, i.e. a roll which bulks
uniformly over its full width; this aspect will be touched on below, though
a full consideration of the use of calender air blowers is left till SC. 4.
On some machines calenders are used for applying colours, starch
solutions, wax emulsions, and so forth to the web surface from water
doctors. The purpose of this is to improve surface smoothness and appear-
ance, and other characteristics such as oil penetration or scuff resistance;
these are essentially finishing processes which it is found more convenient
to undertake on the machine rather than as a separate operation-as such
they fall beyond the scope of this work.
SB. 9 1 General comments on calenders

The size and number of stacks on a paper machine is governed entirely by
the maximum degree of smoothing that is required. Some older machines
are equipped with as many as five sets though it is unusual nowadays for
there to be more than two on new machines. There are several reasons for
this, most important of which are the difficulty of running several stacks
together at faster speeds, the development of supercalendering as a more
flexible means of obtaining a high final finish, and the growing size of
machine calenders which enables a higher finish to be achieved in any
individual stack.
Another reason likely for the reduction in the number of calender stacks
is that nowadays as a result of improved overall machine control paper
tends to be run with more moisture at the reel-up, and this increases the
smoothing effect of calenders. On machines equipped with several sets of
calenders paper may be run through each stack at an average moisture
481
content of 3 per cent. or lower, and the increase in smoothness occurring at

any one stack will then be quite small. On the other hand, when calendering
at such a low moisture content, the loss of bulk for a given increase in
smoothness will not be so great; thus, any required final smoothness can be
achieved for less loss of bulk than would occur were the moisture content
higher and fewer stacks used, a distinct advantage in almost all categories
of paper.
Any load applied to a calender stack adds to the weight of the rolls
themselves, increasing the pressure at each nip and hence improving the
resulting smoothness. This gives some measure of control on the final
finish though it is not good practice to change the calender loading except
within relatively narrow limits otherwise difficulties inevitably occur due to
unevenness across the web (the problem of cross-web control of thickness
and smoothness is discussed in SC .4). Instead the general level of finish
required for any one making is achieved by missing out nips or whole
stacks of calenders, and final trimming only is done by altering the load on
the calenders. Where a variety of grades are made on a machine, several
sets of calenders can thus give much greater flexibility. For the maximum
finish the load on successive stacks of calenders must be increased in
steps to offset the reduced effect of a given load when the paper is smoother
entering the stack (from this it follows that later stacks should be heavier);
Mardon (106) has even reported that a second stack with a similar loading
to the first has practically no effect, and considers that further development
is achieved only when nip pressure in the second stack exceeds that in the
bottom nip of the first stack. The first set of calenders normally has the
greater influence on the reduction of thickness or bulk, just as the first nip
in any particular stack has the greater influence on this property. Later
stacks affect surface properties more and should, according to Mardon, be
equipped with smaller rolls that can be adequately loaded; this gives the
higher specific nip pressure which appears to produce more rapid increase
in smoothness.
With increasing machine speeds and deckles, calender rolls have become
gradually larger and heavier. The open-sided stack is now replacing the
A-frame design typical of older machines and has distinct advantages in
making it easier to change rolls and ensure correct alignment one above the
other. Bearing lubrication has advanced from ring oiling to gravity and
forced-feed oil. Bottom roll bearings and occasionally others are generally
water-cooled.
Some calenders are equipped with two or more hollow rolls through
which steam is passed; normally these are arranged to contact each side of
the sheet in equal number. A sight-glass and steam trap are usually
provided and the system is commonly a simple blow-through arrangement
with the rolls in parallel or series and pressure gauges at either side. Con-
densate is returned to the main system, though when a higher pressure than
the main steam supply is used (to raise the useful temperature inside the
rolls) the exhausted steam may be passed into the main dryer header;
however, this practice can lead to difficulties when back pressure varies, and
it is preferable to use the low-pressure exhaust steam for some other purpose
482
THE DRY ING SECTION AND CALENDERS SB.9 2
such as heating broke pulpers or breakers at the dry end to assist dis-
integration of the broke. It is also possible, though unusual, to heat
calender rolls by gas or electricity.
However, the usefulness of steam-heated rolls is open to a certain degree
of doubt. Certainly the use of steam at start-up can help to heat up the rolls
to nearer normal operating temperature and thereby cut down on the time
needed to run the stack beforehand when there is no paper in the nips (a
practice which is harmful to the surface of the rolls). But it is also fairly
common procedure to keep some steam flowing through the rolls all the
time, on the supposition that this gives an improved smoothness and gloss.
By contrast, on other calenders cold water is run through one or more rolls,
and this is reputed also to give a smoother sheet, though without added
gloss. So far as the author is aware, experimental evidence has been pub-
Hshed to confirm the validity of only the first of these practices, see SA. 43.
SB . 9 2 Calender roll cooling
It is common practice to use low-pressure air for cooling one or more rolls
in a calender stack at selected positions across the machine. Where a jet of
air is directed on to a roll, the natural heat generated in the roll by frictional
resistance and deformation is dissipated more quickly, resulting in a
lowering of temperature in that region. The effective diameter of a roll at
any particular position is dependent on the temperature through the cross-
section and any reduction in this temperature, even if only near the surface,
will cause the diameter to be fractionally, but proportionally, reduced also.
Calculation of the precise effect which occurs in practice is difficult because
it is only possible to measure the surface temperature of calender rolls
where air will have the greatest effect; but as an example, a 10 deg. F
temperature reduction in a 14 inch roll would, if distributed through the
full area of the roll, reduce the diameter by the order of one thou., which is
appreciable compared to paper thickness. The effect of this diameter
reduction is to increase the gap between rolls adjacent to the one being
cooled, so wherever air is directed on to a roll a slight decrease in calender-
ing pressure occurs, causing less reduction in thickness of the paper passing
through in this position.
The cooling effect of a jet of air depends on the velocity and the tempera-
ture difference between air and roll (according to Howe and Lambert (61)
the temperature difference is not so important relatively). By varying the
velocity of the cooling air at different positions across a machine it is thus
possible to alter the thickness profile of the paper. In practice, of course,
the most desirable situation at the reel-up is where thickness is as even as
possible, since this makes it easy to build a machine roll of uniform tight-
ness all the way across; so the role of the calender cooling equipment is
essentially to correct unevenness in the thickness profile. Such unevenness
can originate either in the paper itself or because the calender rolls are a
different diameter in one place than another: in the former case correction
of thickness may lead to deterioration in the cross-web profile of smooth-
ness and other qualities; only when air is used to remedy uneven thickness
resulting from differential diameter of the rolls (poor cambers, lack of
483
5B.9 2 THE DR YING SECTION A ND CALENDERS
uniformity of heating-up due to cooling effects, etc.) is there likely to be an

all-round improvement in cross-web paper uniformity. This will be
discussed in greater detail in SC. 4.
On slower machines making heavier grades, several large air blowers at
intervals of one foot or less may be sufficiently flexible to give adequate
control; with this arrangement a relatively crude valve on each individual
blower enables the flow of air from the main header pipe to be regulated,
and in addition the nozzle can usually be tilted sideways so that three or
four may be concentrated on one spot if need be. For high-speed machines
it is advantageous to use a greater number of nozzles which are smaller in
area and closely spaced at intervals of 2 inch to 3 inch making tilting
unnecessary; in modern designs the nozzles are incorporated for con-
venience with an oscillating doctor using the frame as an air duct. To assist
the dryerman in keeping a more systematic watch on the amount of air he
is using at different positions across the machine, the air pressure in each
nozzle can be measured on simple gauges or water manometers which are
displayed where they can be easily seen as the roll is checked for uniform
hardness of reeling.
Normally there are two sets of cooling nozzles on the entering side of the
calenders, one of which will be on the third roll up and the other usually on
the fifth roll. In some designs the nozzles are on the leaving side of the stack
and then one set may be directed on to the second or queen roll. Choice of
the actual rolls on which air nozzles are used does not seem to affect the
result materially provided the air blows direct on to the roll itself (paper
of course has an insulating effect). Blanchard (82) considers that air should
be blown on to higher rolls in a stack since it is in the early nips that the
greatest caliper change occurs. But it may be that thickness differences
introduced so soon in the stack are nullified in later nips. Possibly the best
arrangement from a theoretical standpoint is to have one set of nozzles on
the roll next to the top where differences in thickness of the paper itself are
remedied, and one set in the middle of the stack to remedy diameter
differences of the calender rolls. But how the two functions could in
practice be separated is another matter!
Pressure of air used for calender cooling is usually about 1-2 p.s.i. on
the modern narrow nozzles, though on larger blowers pressure is not
usually greater than-} p.s.i. The air should be well filtered to avoid abrasion
and marking from grit particles entering the calender nips. In some cases
the air is cooled, though this is hardly necessary in this country; it may
even be a disadvantage to cool the air for there are indications (104) that
maximum smoothing effect is achieved in the 70 deg. to 75 deg. C region
and this is hotter than the normal operating temperature for calenders.
Temperature control of the air has been advocated but this degree of
elaboration cannot be justified unless it is apparent that normal changes in
ambient air temperature have a significant effect. This seems unlikely
because such changes would have to be quite large to alter the difference
in cooling effect as between the various nozzles across the machine; in
other words, alteration of cooling air temperature can be expected to have
only a second-order effect. Some heating may be found desirable when the
484
cooling air is very cold to avoid the nozzle valves becoming too sensitive,
but this can be easily achieved by drawing a proportion of the intake-air
from the machine room.
SB . 9 3 Barring caused by calender stacks

With increasing machine speeds the difficulties associated with barring at
calender stacks have become more and more pronounced and many
groups have carried out research to throw light on the causes of this.
Barring appears on calender rolls as alternate light and dark patches
stretching the full width of the roll at intervals of about one inch or more;
it generally starts either down the whole stack at once, or on one or two
intermediate rolls from which it appears eventually to spread over the
whole stack. Connected with the presence of these bars is the occurrence of
cyclic variations in thickness and surface properties of the paper in the
machine direction; these variations are more pronounced the greater the
moisture content of the web (due presumably to the greater effect of
calenders), and in several cases blackening occurs in the areas of low
thickness and bars are observed very prominently in the paper itself.
The difficulty in tying down the cause of barring is that the severity and
frequency of the cyclic variations produced in the paper can vary over quite
a short space of time: they are sometimes quite regular but at other times
are apparently irregular in appearance with periods of large fluctuations in
thickness interposed by periods of relatively little variation. Also the
amplitude of bars on the rolls themselves may, according to Pye (96), differ
from one calender roll to the next and can fhtctuate in prominence or even
in amplitude over a period of time; further, the amplitude of bar marks in
the paper can be quite different again from those on the rolls. Despite this
elusive behaviour, it is now firmly established that there is an association
between bars that appear on calender rolls, and barring in the paper (in the
sense of cyclic thickness variations, not substance barring such as is
produced at the wet end): both in fact originate in small vertical oscillations
or 'chatter' of the rolls in the calender stack and the questions to be
answered are how do these vibrations originate and how best can they be
overcome?
Analysis of thickness profiles in the machine direction to assess barring
in the paper web is confused by cyclic fluctuations in substance produced
at the wet-end because these may also be associated with thickness varia-
tions. This has led some workers, notably Howe and Cosgrove (87) and
Pye, to suggest that cyclic variations in substance can be responsible for
creating barring; this would occur when a resonant vibration in the stack
is excited due to the amplitude of successive low substance regions being
some simple fraction of the calender roll circumference. On the other hand,
Cuffey (83) failed to find any connection between calender stack vibration
and cyclic substance variations in the uncalendered sheet. There is more
general agreement that stack vibration can be set up by all kinds of external
oscillations similar to those produced by reciprocating and rotating
machinery at the wet-end of the machine.
485
Wahlstrom (98) considers that vertical vibration in calenders is essen-

tially a characteristic of the stack itself; it is not necessary to postulate any
continuous outside disturbances such as periodic variations in the un-
calendered paper or mechanical vibrations transmitted from other
machinery to explain the oscillations which are observed. The calender
rolls float, as it were, supported by the paper which acts in each nip as a
spring possessing some normal pressure-compression relationship; analysis
of the dynamic behaviour of such a system using measured compression
characteristics of paper and the weight of rolls in a particular stack showed
that stable oscillations could be produced which agreed well with those
observed in practice. Once excitation has been started in a nip (substance
variations or external vibration may have a r6le in this), the oscillation can
be transmitted through the whole stack which begins to act as a mechanical
oscillator using the paper itself to transmit the forces and act as a feed-back
medium. The frequency of stable vibration of such a system will vary
according to many local conditions, including the machine speed.
Persistent oscillations at specific frequencies will eventually mark rolls,
and may become self-perpetuating because of this. According to Howe and
Cosgrove, the marks themselves are due to differential work-hardening,
though Pye is of the opinion that they are produced largely by the effect of
the vibrations on doctors acting on the rolls. The optical appearance of
bars on the rolls depends on the angle at which they are viewed, and in fact
they comprise smooth and rough strips, the latter exhibiting fine, short
scratches and minute pitting. Parker (117) has detected definite corruga-
tions with severe pitting centred usually in the troughs on the up going side.
So much for the sources and nature of barring, but how can it be pre-
vented or reduced to tolerable limits on machines which are plagued with
this trouble? Numerous remedies have been suggested. Staggering rolls
alternately a fraction of an inch off-centre up the stack appears to help
delay the onset of barring difficulties (and also, incidentally, gives a better
fit if bearings are slack). A modern development of this technique (using a
'Torque Compensator') is in fact claimed to present a most efficient pre-
vention of barring (118). Also, according to Cuffey, changing the number
of rolls may disturb the resonant frequencies sufficiently to be helpful,
though he reports that adding or removing a roll can have the opposite
effect on different stacks. Alteration of the relieving or applied load on the
stack may also help temporarily to disturb a persistent, severe chatter (as
may making the stack jump by passing through a thick wad of paper), but
normally changing nip pressure in this way succeeds only in shifting the
frequency of vibration. Removal of grit from the paper stock reduces
barring by preventing rapid wear of calender rolls, and grinding should be
carried out on perfectly circular journals or initial corrugations can be
imposed (117).
The most direct remedy for calender barring is to prevent or damp the
vertical vibration movement of the rolls which is recognized as the basic
cause. To this end, Wahlstrom has constructed a device for replacing the
top roll by a light hollow roll (one-tenth the weight of a normal calender
roll) which is pressed upon by a number of other small rolls forced down by
486
air cushions. The nip pressure at the top then becomes relatively independ-
ent of vertical movements in the stack, and the feed-back mechanism
essential for setting up a regular oscillation down the whole stack is largely
eliminated. Use of this device is reported to have been very successful in
reducing barring though further applications have not yet been reported.
SB. 10 EQUIPMENT
To complete discussion of operating factors which affect performance of the
final section of the Fourdrinier, it is now proposed to bring together a few
miscellaneous points regarding equipment found on all machines. In
addition to this, the use of devices of a more specialized nature including
such equipment as the smoothing press, breaker stack, sweat roll, and so
forth is briefly described.
SB . 10 1 Drying cylinders and drive
Various details relating to the drying cylinders have already been given in
various contexts, in particular with regard to their size and number on any
machine and the factors governing their construction. Normally all
cylinders on a machine are of the same size, with the possible exception of
the last few where on machines running a small range of similar grades the
diameter may be gradually reduced a total of 0· 1 in. or more over the last
8 or 10 cylinders in order to offset the effect of shrinkage. Also the first or
pony cylinder can be of smaller diameter, being then more suitable to give
a first gentle application of heat to the web; this arrangement is generally
considered superior to having an unheated small-diameter turning roll and
a normal-sized first cylinder because wrinkling of the web and surface
picking is less likely to occur. Cylinder material is cast-iron, except for
machines operating at very high steam pressures (above about 80 p.s.i.)
when fabricated steel may be employed. This material is also used in the
Lukenwald dryer which consists of a jacket formed by an outer and inner
shell; steam is introduced into the jacket through a number of ribs con-
necting with the journal and condensate is removed through a fixed pipe
attached to the inner shell. This construction allows the use of relatively
thin steel even for high steam pressures, thereby giving a faster evaporation
rate than with an equivalent cast iron dryer, and the greater velocity of
steam in the jacket is claimed to produce a more uniform temperature
across the face of the roll. It is important that all cylinders are accurately
balanced for smooth rotation and aligned parallel with one another.
Sectionalization of the cylinders has also been discussed earlier; its use is
generally to permit some control on shrinkage and reduce the possibility of
the web creasing and breaking. On smaller machines splitting the dryers
presents no special problems except for the cost of providing separate
drives for each section; a greater variety of felt lengths may be necessary
but the total cost of storing and using should not be appreciably greater
than if fewer sections were used. Nevertheless there is no advantage to be
gained from having an unnecessary number of sections because control of
the draw at each is equally as important as when fewer sections are used :
487
thus, more draws in the drying section necessitate either a more expensive
instrumentation and control system or place greater demands on the
dryerman. On faster machines the governing factor in deciding the number
of sections is often the maximum number of cylinders that can be con-
trolled in any one section. Too many cylinders ganged together makes the
drive unwieldy and places a strain on the gears. This may be overcome by
removing gears and restricting the drive to only a proportion of cylinders
in each section, relying on the felt to pull round the remaining cylinders.
To prevent any possibility of slippage the felt tension will then probably
have to be higher than otherwise, and some difficulty, especially with ropes,
may be experienced when stopping the section quickly due to the tendency
for the undriven cylinders to lose speed more slowly. On the other hand,
creasing of the web between cylinders due to small unevennesses in trans-
mission through individual gears can often be substantially reduced by this
modification.
The drive is an especially critical aspect of the drying section, smooth
operation being closely dependent on the sensitive and uniform control of
draws produced by very small speed differences between the sections.
Ideally it should be possible to alter the draw at any section without
affecting any others, though on most older machines this is not possible and
adjustment of one draw generally means that others will need re-setting
also; it is frequently the case that one of the presses is taken for the master
speed-setting and in this event alteration of, say, the draw between the last
press and first section of cylinders means that each succeeding draw down
the dryers to the reel-up will also require attention. Facilities for crawling
the dryers are essential for putting on new felts and for the purposes of
inspecting the felts and helping to heat up cylinders evenly at start-up. It is
also very useful to be able to reverse the drive as this facilitates removal of
broke jammed in the cylinders. Modern electric drives can have other
valuable features such as a controlled acceleration up to the speed desired,
controlled braking, and means of temporarily overriding a pre-set draw
during feeding up in order to remove the slack quickly and thereby lessen
the risk of breaking through excessive flapping. One or more emergency
stop-buttons in prominent positions by the dryers are an important safety
measure, and on faster machines help to prevent jams becoming too
serious.
SB . 10 2 Doctors in the dryers and calenders

Doctors are important on the first few dryers to keep down build up of
scale, pitch, fibre, dust, etc., on the cylinder surface and prevent the sheet
wrapping round and jamming or bulging the felt in the event of a break;
frequently it is necessary to equip every dryer with a doctor for these
reasons. Suitable trays, particularly for top dryers, are also needed on the
first few doctors in order to collect the debris scraped off the cylinder and
so prevent it from contaminating the sheet and causing breaks especially in
the calenders; such trays need attention and must be regularly cleaned.
More modern applications include a suction arrangement incorporated in
488
the doctor to remove the debris as soon as it is formed by use of a low-
pressure air flow through a tube behind the doctor which carries dirt away
to an extractor fan; this type of doctor, though more expensive than the
simpler type, has much to commend it for the first few dryers. Another
innovation is the electro-doctor which prevents corrosion of the cylinder
surface by inducing an electric charge on to the surface; this is, however,
more likely to be of value on M.G. cylinders where retention of a polished
surface is much more critical.
A doctor on the final cylinder of the drying section or any section pre-
ceding a gap in the dryers for a breaker stack, size press, etc., is usually
essential to ensure that the sheet does not wrap round at a break or start-
up; it is often of a heavier design than on other dryer doctors. Where the
final cylinder is used as a sweat roll and a film of moisture on the surface is
necessary for its operation, it is not possible to have a doctor applied
continually; in this case some mechanism to detect a break and immediately
apply the doctor is needed.
It is normally possible to alter the load and horizontal position of a
dryer doctor and this is done when necessary to prevent rings of hard scale
and dirt forming round the cylinder surface; the trouble with this arrange-
ment is that it is often found that doctors tend to ride on top of rings once
they have formed so it can prove very difficult to remove them and at the
same time prevent others forming. Any contamination on a cylinder
surface affects appreciably the resistance to heat transfer through the
surface and is not, especially in the case of the uneven patterns associated
with rings, conducive to uniform and efficient drying. Doctors should be
adequate to prevent build-up in the first place and their construction,
operating load, and the provision of oscillation gear should be carefully
considered with a view to finding the best method of keeping all cylinders
continually shining without creating an excessive drag on rotation and
thereby pushing up the driving load.
Doctors on calenders are of a more elaborate construction designed to
ensure an even and relatively light pressure over the full length of the roll,
an essential consideration if wear is to be kept uniform. They are manu-
factured from a variety of materials including carbon and stainless steel,
phosphor-bronze, and fibre, the choice being made to compromise rate of
wear of the roll against wear of the blade; generally it is preferable to use
the softest blade that does an effective job because changing is much
cheaper than frequent re-grinding of a calender roll. Dust-extracting
equipment combined with the doctors is more common on calender rolls
where much fuzz and dust is frequently generated and it is highly important
to prevent abrasive materials entering the nips. A quick method of raising
each doctor, manually or pneumatically (in which case a single operation
can be arranged to raise all doctors simultaneously), is useful to permit
cleaning at every opportunity; any pieces of grit wedged between the
calender roll and doctor can rapidly lead to wear and scoring. Pneumatic
loading is beneficial in preference to spring tension as it permits a readily-
controlled pressure of contact, usually of between 1 and 2 lbs. per linear
inch. The angle of each doctor also requires careful setting and should
489
remain relatively unaffected by wear; also oscillation is particularly

important for calender doctors. The bulk of the doctor frame frequently
takes up enough room to act as an effective nip guard- a valuable asset.
SB .10 3 Felt and other equipment
Each dry felt requires a stretch mechanism (generally allowing a movement
equivalent to about 10 per cent. of the length of the felt) and this serves also
to apply the required running tension to the felt. On very old machines the
necessary stretch and tension may be achieved simply by positioning the
roll according to the dryerman's judgement, but this system is far from
satisfactory because of the rapid dimensional changes which can occur
particularly in cotton and wool dry felts at a break. A more common
method of applying an even tension under all conditions involves the use
of weights which can move up or down, on a frame situated at the back
side of the machine, in accordance with movement of the stretch roll.
However, on modern fast machines such weighting gear is cumbersome and
it is becoming more usual for the desired tension to be applied by means of
hydraulic pressure acting through pistons on both journals of the felt
stretch roll. This permits the tension to be accurately indicated and set to
any desired value, and it is also possible to arrange for tension to be auto-
matically relieved at a break to prevent straining a felt when it is liable to
shrink.
Various types of guide for felts are in use and nowadays these are always
automatic, though the old, purely manual adjustment is still retained on
one roll of each felt partly as a precaution against failure of the automatic
guide but also to allow some adjustment which permits the guide to be set
working centrally. Automatic guides in general use for many years include
the spade-actuated and drum or cone-roll type, while more modern
devices detect the edge of the felt by means of air jets or a photo-electric
system. Correction of the position of the guide roll requires (except for the
cone-roll system) some motive power, and air or fluid pressure on an
appropriate diaphragm or piston is commonly used for this purpose. The
self-actuating servo-roll is also coming into more general use for dry felts
which do not alter greatly in dimension, particularly the plastic-wire type
and synthetic felts.
Expanding spreader rolls with a fixed or variable bow are sometimes
used on the web in the drying section to regulate cross-direction shrinkage,
but normal application is between the dryers and calenders, and between
the calenders and reel-up. In these positions they help keep the web
spread out over what is frequently a long pull, avoiding (especially at the
edges) wrinkles and creases in the calenders (which produce cuts in the
paper) and in the machine roll. An alternative is to have a simple bowed bar
or 'spreader' bar: the advantage of this is that construction and support of
the bar can be readily designed to allow slight alterations to the bow any-
where along its length according to conditions; the disadvantage is that the
hard friction between web and bar generates dust and can in some cases
damage the surface of the paper.
In the position between the dryers and calenders it is good practice to use
490
a spring or dancing roll which moves gently up and down on springs
according to the tension in the web; this effectively damps fluctuations in
draw between these two sections where, because of the varying load that
may be experienced on dryers and calenders during running, small varia-
tions are common.
Disposal of broke at the dry-end is made much simpler when a broke
breaker or pulper is located in a convenient position; this is arranged so
that the full sheet can fall into the pulper from both the last cylinder and
the calenders. Operation of the pulper is generally automatic in that one or
more diluting deluge showers (normally whitewater) are immediately
turned on when there is a dry-end break (where there are two showers, one
for broke from the calenders and one from a size press, a more elaborate
break detection system can be used to select the shower needed). Some-
times the shower can also be turned on manually when other broke torn off
the machine roll or trimmed at the winder requires disposal, but it is
preferable if a consistency regulator is incorporated either in a small
recirculation line or direct in the main line leading from the pulper, and the
signal from this is used to control water dilution. Level control is essential
and is designed either to regulate a valve opening or to switch on the
emptying pump intermittently; in addition rotation of the pulper motor
can be arranged to idle when outward flow reaches some pre-set minimum,
thus avoiding over-disintegration.
Finally, in the drying section it is often useful, and on faster machines
imperative, to have some form of break detector system. Normally this is
either photo-electric or ultrasonic in operation (the latter is claimed to have
advantages in that it is less affected by dust or dirt and there is no lens
system to keep clean); there can be several individual detectors situated
down the drying section to ensure as rapid response as possible, and.each
detector can be arranged to actuate different parts of the system. Detection
of a break in this way can be used for a variety of purposes, depending in
some cases on the location of the break; these include giving an audible
warning, starting a broke conveyor or screw on the presses, sending across
the cutter on the wire, reducing steam pressure in the main supply, and
starting a deluge shower either on the wire or in a dry-end pulper. In the
latter case a selective detector system can be used to switch on one or other
shower (normally it would not be desirable to have both on together
particularly when the supply is from high-pressure white-water) and a
further provision would ensure that the wire deluge shower switch could be
manually overridden in an emergency to prevent damage to the wire.
Numerous other devices can be actuated by a break-detector system and
have been mentioned in the relevant context; many are now considered
essential safeguards on any machine.
SB . 10 4 The reel-up
Except when a size or coating press together with more dryers are arranged
to follow the calenders (or part of the drying section), the sheet passes
straight from the calenders to the reel-up. With the old direct spindle drive
reel-up it is difficult to apply an even tension as the reel builds up and
491
SB .10 5 THE DRYING SECTION AND CALENDERS
pressure has to be altered on a belt or clutch to allow for the growing

rotational momentum. Also this type of reel-up does not lend itself to easy
changing and would be quite impossible to use satisfactorily at machine
speeds much above 500 feet per minute or so. At one time it was not
uncommon for the same stand to be used for unwinding on the slitter but
this practice, never satisfactory because of the congestion it created, is now
obsolete.
Various types of drum reel-up are now in general use and, though they
differ in design and ease of operation, they invariably permit much more
evenly-wound and straighter rolls to be built up. Partly this is because
tensions are higher than with the direct-drive reel-up, but also it is easier
with this method to keep reeling tightness (governed by the pressure of
contact between roll and drum) relatively steady as the roll builds up. The
most common type of drum reel-up achieves this by reeling not at the top
of the drum but at an angle to it; the full weight of the roll does not then
fall on the drum, creating a proportional rise in reeling pressure, but is
offset by a gradually increasing angle to the vertical which causes the
greater part of the increase in weight to be borne by the reel-up arms. On
modern fast machines it is becoming increasingly common to reel horizon-
tally on the drum, keeping the pressure at the reeling line constant by
applying an appropriate pneumatic or hydraulic loading to the spindle
journals; larger sizes of rolls are possible with this technique. Water is
commonly used to keep the drum reeling temperature of the paper cooler,
and this also appears to reduce tendencies towards the development of
static. A doctor on the drum is needed to prevent the sheet wrapping
round in the event of a break.
Change-over of rolls is more or less automatic on most machines with a
second spindle ready to be accelerated up to speed on the drum when the
existing roll is ready to be removed. The completed roll is edged away from
the drum and a blast of air, aided sometimes by a squirt of water, ensures
that the paper begins to wrap round the new spindle; it is more satisfactory
when this can be done without having to lift the completed roll clear with
the crane. A braking mechanism on the old spindle journal to ensure that
rotation of the completed roll ceases rapidly is a valuable safety measure.
The Pope-type reel-up allowing the spindle to be passed from one set of
arms to the other is always preferable to the type which necessitates alter-
nate use of two individual pairs, because one change-over is then clumsier
than the other and when reeling on the back arms it can be awkward for
the dryerman to check the rolls.
SB.10 5 Special equipment

Equipment of a more special~zeq nature in frequent use in the drying
section ca1{ be
divided into two broad categories: in one the purpose is
primarily to achieve greater smoothness of the paper surface (the breaker
stack and smoothing press); in the other it is to increase the moisture
content at the machine roll (the sweat roll, spray damper, calender water
boxes, etc.). Of course, other specialized devices, such as size and coating
presses, are also found in the drying section but these, strictly speaking, are
492
THE DRYING SECTION AND CALENDERS 58.10 5
employed only to add some new specific property to the paper, not to
enhance the existing properties; discussion of the use of such equipment is
beyond the scope of this work.
The breaker stack is used in the drying section either immediately after
the presses or more commonly well down the dryer section. It consists of
two highly polished steel or chrome-plated rolls, loaded to give pressure
and sometimes steam or hot-water heated, and is thought to serve two
main purposes. In the first place the breaker stack effectively smooths out
fibre clumps and felt mark on the web surface, giving a more uniform finish
and reducing two-sidedness; at the same time the stack seems to help bind
surface fibres closer together reducing the extent to which fibre dust and
pick-off occurs at the calenders (of particular benefit for offset-litho papers
where surface dust readily contaminates the printing blanket). Secondly, it
allows more control over the eventual thickness of the sheet (the breaker
stack is especially popular for paperboard machines); the overall effect, as
may be expected, is to reduce bulk and increase smoothness though it is
claimed that the same smoothness can be achieved for less loss of bulk, a big
advantage for ordinary printing papers. Because the web is still appreciably
damp, the pressure applied at a breaker stack is much lower than at the
calenders; but small changes in the load have a relatively greater effect on
the finished properties and an accurate indication of this load is absolutely
essential to successful and consistent operation. Efficient doctors are
needed on both rolls as any grit or crumbs mark the sheet much more
readily than at the calenders: feed ropes pass through just outside the edge
of the nip.
On any machine equipped with a breaker stack it is important to
determine how much of the smoothness eventually required is contributed
by the breaker stack and how much by the calenders. A high load on the
breaker stack may compact the web to a degree which affects subsequent
drying, raising steam costs, while spare loading capacity at the calenders is
unused; generally speaking, provided the breaker stack is run at a pressure
sufficient to serve its primary function of achieving greater surface uniform-
ity and removing felt marks, there seems little point in raising the load any
higher.
The smoothing press is similar to the breaker stack in purpose with
particular emphasis on the removal of wire and felt marks in better
quality papers. The term is generally confined to the use of a pair of rolls
immediately after the main press section and before the dryers though
some applications have also been reported in the middle of the drying
section. As the name implies, the smoothing press is essentially the same as
an ordinary plain press as regards loading arrangements (a lighter pressure
is of course necessary), doctors, drive, etc.; the only difference is that no
felt is used in the nip and no water is expressed from the web. One of the
rolls (generally the top) is rubber-covered, the other usually bronze. If
required an embossed pattern of a relatively crude nature can be trans-
ferred from the rubber roll to the paper surface at this position, though
this technique is usually applied with a separate rubber roll running on the
paper web between the last press and dryers. Apart from this the smoothing
493
press serves only to reduce small scale unevenness in the paper surface on
both sides.
The most common method of increasing moisture in the finished paper
is by means of the sweat roll, which is always the last roll in the drying
section. A supply of cold water is sprayed into the roll and removed at the
back journal by a syphon and ejector; this promotes the continuous forma-
tion of condensation on the cylinder surface which is then transferred to
the paper. In addition, steam may be sprayed on to the outer surface of the
roll to increase the condensation rate, and this affords a rough means of
varying the effect across the machine. The diameter of the sweat roll must
be 50 or more thou. greater at ambient temperature than that of the
drying cylinder it follows in order to allow for the expansion of the latter
during operation. The need to build up a film of water on the roll surface
makes it impossible to have any doctor permanently in contact, so to
ensure that paper does not wrap round the cylinder in the event of a break
a doctor has to be held ready for immediate application to the surface; the
doctor is released when a signal is received from a break detector or (and
this is probably more effective if it is made sufficiently rapid) the operation
is made automatic by means of a mechanical connection from a spring roll
situated between the sweat roll and the calenders (this roll falling to a new
position when the web snaps).
The purpose of using a sweat roll should be clearly borne in mind: it is to
add moisture into the paper, both to permit reeling closer to the eventual
equilibrium condition in atmosphere and to attain a better finish in the
calenders. The reason that this is not so easy to accomplish direct off the
machine is largely tied up with the difficulty of attaining simultaneously a
moisture profile which is both high and uniform, a topic fully discussed in
SC. 3; but it is self-evident that the greater the moisture content at which it
is possible to reel without the sweat roll, the less steam will be needed to dry
the paper out. It is a common failing when operating a sweat roll that it
becomes used on a machine purely as a means of cloaking irregularities in
profile that ought to be corrected. If the cost of doing this were better
appreciated, then perhaps more effort would be made to produce a reel at a
reasonably higher moisture content in the first place, and only later start
using the sweat roll when a high average proves impossible to achieve
without it. Turning off a sweat roll, but at the same time keeping the
average moisture level at the reel-up unaltered by adjusting the steam
controller according to a moisture meter reading, can sometimes enable
the steam pressure needed for drying to be reduced by 2 to 3 p.s.i. with no
noticeable increase in the effort required to build up an even roll. There is
also evidence that the smoothness of paper increases by a small amount
when moisture is retained in the body of the sheet instead of being added at
the sweat roll. In this respect it may be noted that laboratory studies
reported by Jackson and Ekstrom (104) have indicated that it is inadvisable
to overdry the sheet surface, even if it is re-moistened at a later stage,
otherwise the fibres became hard and less compressible in the calenders.
Other devices have been developed with the same intention as the sweat
roll of adding moisture to the web. These include: felt-covered rolls dipping
494
into a water trough and transferring water direct to the paper surface;
various types of spray dampers in which a large number of very small water
jets impinge direct on to the web, or at an angle on to a plate thereby
creating a mist either close to the web or on a roll which the web contacts;
rotating brush dampers which also pick up water from a trough, then strike
a doctor which scatters a mist on to the paper surface; carefully designed
steam sprays positioned close to the nips of calender rolls; and water
doctors on the calenders. No objective assessment of their comparative
performance is available, though it is obvious that operation will be better
the more control there is on the actual moisture picked up both in total and
at different points across the web. From this point of view spray dampers
and certain types of calender water doctors are probably most adaptable.
This applies especially to the former because the flow to individual nozzles
can be controlled, or careful choice of the nozzle sizes at different positions
across the machine can allow relatively permanent discrepancies in the
moisture profile to be corrected, e.g. slightly larger nozzles towards the
edges of the sheet compensate for the common tendency towards being
dry. Much development has taken place in the construction of non-
clogging nozzles of very fine aperture which are designed to give a mist for
direct use, and those working on the air-injector principle with compressed
air and water lines regulated by separate valves are claimed to give a very
uniform dispersion. Water for spray dampers must be well filtered, other-
wise the fine sprays clog, while a better mist is possible when the water is
hot; increased efficiency has also been claimed for a device which applies a
very high-tension d.c. voltage between the spray nozzles and a backing roll,
though this naturally involves careful caging-in of the equipment and is
unlikely to find general acceptance. Calender water doctors on the other
hand need to be of rigid construction to give even application; they
should have an easily adjusted overflow head and be mounted perfectly
level.
Several of these devices are used after the calenders, for example a spray
damper to apply water to the reel-up drum. The main purpose then is to
raise the moisture level to improve subsequent supercalendering. But no
matter where or how moisture is added, the same general comments on the
practice apply as when using a sweat roll, and ample caution is needed to
prevent addition of water exceeding the level that is absolutely necessary.
In all cases of direct application some measure of the water being applied
is highly desirable and should be used to check excesses. Whenever possible
it is also desirable to measure the web moisture content before the applica-
tion of water, especially of course when the reading is used to regulate
drying. Neither of these controls is available in the case of a sweat roll and
in fact it is not possible, other than by using an elaborate moisture meter
set-up or regularly breaking down the sheet to obtain moisture samples, to
keep much check at all on how the roll is functioning.
A more recent addition to devices available for adding moisture is an
on-machine conditioner using high-velocity, high-humidity air blown on to
the paper from nozzles (121). The web is supported on wire-wound drums
and the whole device is hooded to give closer control of operating
495
conditions. It is claimed that in addition to allowing moisture to be added

to the web, the cross-machine profile is also improved.
Finally, to complete this brief description of more specialized equipment
a short note on static eliminators is called for. The need to reduce static at
the reel-up when this reaches unpleasant proportions has called forth a
wonderful variety of contraptions most of which, it is certain, do anything
but what they are intended for. The most popular comprise chains or strips
or tufts or bristles of copper wire suspended at judicious intervals across
the machine immediately after the calenders. These wires are carried on a
bar, lowered to a position where they just touch the web, and are connected
to a cable sufficiently thick to conduct a stroke of lightning away to some
mysterious and inaccessible point carefully selected by the electricians. In
theory static electricity is seduced from the web by this attractive array of
whiskers and led away to an obscure and anonymous fate in the bowels of
the earth. Gazing at a typical static eliminator one can easily visualize that
electricity is flowing along the wire, and the occasional spark re-inforces
the certainty that static is indeed being eliminated. It is then hardly com-
prehensible that, having recklessly abandoned the precaution of wearing
rubber-soled boots, a gentle touch of the roll with moist fingers indicates by
the strength of the kick that some static has surprisingly slipped through
the net.
All that can be said on this topic is that as yet no certain cure exists for
static. One of the difficulties is to get a satisfactory measure of how much
static is present, though a convenient hand instrument is available which
is calculated, by its shape, to cause concern to any dryerman and is well
worth using for this reason alone. Static is less alarming on the machine
roll the damper the web is reeled, and once present in paper it disperses
easier the drier the atmosphere; a water-cooled drum reel-up quite
definitely helps. Static can be extremely troublesome when cutting into
sheets and laying down. But the greatest difficulty is the complaints the
consumer makes, particularly the printer using sheets. Especially for his
sake it is best to retain your own personal brand of eliminator in a
prominent place on the machine.
496
CHAPTER 5C
RUNNING THE DRYERS AND CALENDERS
SC .1 DAILY OPERATION
In dealing with various aspects of running the drying section and calenders,
the same procedure is now used as for earlier sections of the Fourdrinier.
Measurements required by the machine crews for day-to-day operation are
discussed first, and this is followed by a detailed consideration of the
longer-term maintenance necessary to keep up efficiency. Two features of
the dryerman's work that command especial attention, those of keeping
the moisture content and thickness of the web at the reel-up steady and
also uniform across the width of the machine, will, because of their
importance, be dealt with separately. Finally, there is a general discussion
covering practical points.
SC . 1 1 Essential measurements
During normal everyday running, conditions in the drying cylinders are
closely characterized by the temperature or pressure of the steam supplied
to the main bank; if anything goes seriously wrong in the operation of the
dryers it is almost certain to be reflected one way or another in the steam
required for drying. Fluctuations in demand produced by varying moisture
content of the web entering the cylinder section are also reflected in the
steam pressure, as are variations in the boiler pass-out supply in an
uncontrolled system. On many machines drying pressure is close to the
maximum available and indication of this is vital. For these reasons, and
many others discussed earlier, a measure of the temperature or pressure of
the steam supply is absolutely essential for the dryerman, and only a
record rather than a straightforward indication can ensure that full use is
made of this. An alternative is to measure steam condition in a selected
cylinder or temperature of the condensate, but either of these methods
introduces condensing variables and is unlikely to be so effective parti-
cularly when used for control purposes.
Steam supplied to the main bank and throughout the cylinder section is
normally saturated, so either temperature or pressure may be taken as an
indication of its condition. This, however, applies only provided adequate
precautions are taken to avoid air in the cylinders; with a significant
proportion of air present, pressure does not indicate the relevant property
of the steam which controls the drying rate and strictly speaking only
temperature is appropriate for this purpose. A temperature measurement
using a bulb in the main supply pipeline and a filled thermal system to the
recorder is simplest, but this method is relatively insensitive to the small
changes in condition that it is important to be able to observe. This is
particularly the case at higher steam pressures such as are used on M.G.
497
THE DRYING SECTION AND CALENDERS SC.11
and very fast machines; in these cases pressure is almost always the
characteristic measured, because the temperature change corresponding to
a given increase in pressure becomes gradually smaller the higher the
pressure. Pressure measurement using a direct connection-tube from the
pipeline is not satisfactory because the pipe is normally situated at the
back side of the machine and so either the recorder must also be sited at
the back where it cannot readily be seen, or precautions have to be taken to
ensure that varying condensation in a long connection-tube over to the
front side of the machine does not affect the reading. A slack diaphragm
with enclosed air or liquid connecting-tube to the recorder is a better
arrangement, but the most suitable and sensitive (and the most expensive)
system is to install a pressure transmitting device using a standard air-
pressure signal to the most convenient position to display the record; this
arrangement is readily adapted for control purposes, a facility which is
particularly advantageous.
In addition to the main supply steam pressure, numerous other lines
must have at least a gauge to display the relevant steam pressure. The
positions where such measurements are taken depends entirely on the
lay-out of the system, but obviously include the steam supply to such
parts as the felt dryers, calenders, and air-heaters; separate pressure
measurements are necessary for each bank of cylinders on a flash system or
when sections are taken separately off the main steam supply, as for
example when using differential drying for top and bottom cylinders, or an
M.G. cylinder and pre- or after-dryers, or a thermocompressor system. In
addition, pressure measurements upstream of the main steam control
valve, in condensate lines or flash tanks, and on the final condenser are
required; when any individual cylinder is throttled separate gauges on inlet
and condensate lines or a differential pressure gauge are advisable. In
more modern installations differential pressure measurements giving the
pressure drop over each section in which steam is used are common, though
these are normally associated with control systems which are discussed
in SC. 14. In certain cases it is well worthwhile to use recorders rather than
simple gauges. For instance, a useful arrangement is to use the same
recorder for both main steam pressure and the pressure of the first flash-
steam line; this enables the condition of the most important part of the
flash system to be watched, together with the total differential held over the
main bank plus any controlling valves in the main condensate and flash
lines.
There are a number of other measurements which are essential to
efficient operation of the dryers and calenders. The load applied at the
calenders, breaker stack, smoothing press or, in the case of an M.G., at the
pressure roll must be displayed on a double-gauge system, one showing the
load at the front and one at the back; in general, the same type of instru-
ments as were described when dealing with the presses are suitable to
measure the load in each of these cases. The use of the gauges is also very
similar to their use in the press section: in particular, they are essential to
set the load to the desired level and examine how variations in the load
affect performance, though for these applications of course the effect of a
498
THE DRYING SECTION ANO CALENDERS SC.11
given load is assessed essentially in regard to the smoothness of the paper
rather than the water removed. Further, the importance of the measure-
ment in its relationship to the camber of rolls and cross-web uniformity is ,
also exactly as for the presses and as the same principles apply in each case,
the value of load measurement in this respect has also been covered in the
earlier discussion of the subject; the only major difference lies in the
operation and maintenance of the calenders to give a uniform smoothness
across the machine, and this is dealt with later in SC. 4.
Whenever possible, the power consumed by each individual section in
the dryers, and by the calenders and reel-up, should be displayed on
ammeters. These should preferably be equipped with a limit warning
which immediately draws the dryerman's attention to excessive demand at
any part of the machine under his supervision. This warning device is of
particular value in indicating instability and disturbances in the cylinders
due to varying condensate volumes, but also allows early attention to be
given to any conditions (heavy doctoring, excessive draws, tight felts, stiff
bearings, etc.) which may affect the power usage adversely. For the same
reason, and to give indications of adverse trends, a log of the power used
by each section should be completed at specific intervals; the same applies
of course to the power consumed by the whole machine as this information
is required for accounting purposes.
Another essential measurement is of the water flowing into any damping
device before the calenders or reel-up. Especially when using sprays, a
pressure gauge is not really adequate in this position because any change in
the spray orifices, either by design or due to gradual erosion or scaling up,
affects the flow even when upstream pressure is kept constant; it is prefer-
able in addition to have a simple indication of flow rate which is used for
setting water to the damper in conjunction with the pressure gauge. A table
showing the rise in moisture which is roughly equivalent to different flow
rates is also useful to have on hand in order to prevent excessive drying
down of the sheet before damping.
Certain measurements of air state (primarily temperature and humidity
but in some cases also the pressure or flow rate) are essential if ventilation
conditions are to be kept under control. This applies particularly to hooded
machines where the full value of the hood cannot be realized unless
systematic regulation is possible. Usually measurements are taken of the
condition of supply air to Grewin systems, felt blowers, high-velocity air
hoods, etc. and also of the exhaust air in appropriate places. In many cases,
of course, such measurements are part and parcel of the control systems
which are a normal feature of any well-designed ventilation system. Each
installation is individual to the machine but it is evident that the more
elaborate arrangements permitting close control and fine manipulation of
air conditions in different parts of the dryers, as indeed of the machine
house, are generally worthwhile economically because they show how the
heat supply can be reduced and heat wastage minimized. Even the oldest
machine may well benefit from placing one or two wet and dry bulb
thermometers in key positions in the machine house and noting how their
reading changes under different operating conditions.
499
SC.12 THE DR YING SECTION AND CALENDERS
Finally, for production control purposes it is absolutely essential for

each machine to be equipped with some form of recorder which can be
used for analysing downtime. This should preferably give a complete
record of when paper was being made on the machine, and so requires a
device (feeler arm, photoelectric cell or ultrasonic signal, tension bar, etc.)
which senses when the sheet is passing through to the reel-up but is not
affected by normal reel changes. The record may be a straightforward one
on an appropriate time scale showing, for example, the start-up time,
various breaks which occurred during the week or between scheduled
shuts, and shut-down time. Or, and this is the author's own preference,
such a breaks record can be combined with a speed record in order to allow
ready analysis of machine performance at appropriate time intervals.
Together with an arrangement of this type can be associated integrating
clocks to show time lost during specific periods, and more elaborate
arrangements are possible which distinguish between dry and wet-end
breaks and even integrate time lost due to different causes. The use of such
data in conjunction with machine reports is of great importance to the
production personnel, the accounting department, and management: this
will be fully discussed in Part 6.
SC. 1 2 Useful measurements
Other measurements which can with value be made available to the dryer-
man include flows of steam and condensate, the draws and sheet tension,
felt tension, and (depending on the reel-up design) the pressure of the
machine roll on the reel-up drum.
To obtain full value from their use, steam flow meters are necessary in
the supply lines to each separate part of the drying system; this means
having individual meters on the main supply to the drying cylinders, on an
M.G. cylinder, felt dryers, hoods (especially H.V.), and the machine house
ventilation system. In each case a straightforward orifice system of
measurement is adequate, certainly for comparative purposes, provided
upstream supply pressure is reasonably controlled; it is beneficial always to
install recorders and integrators at each position though the same recorder
can frequently be used to chart two or more different pressure readings.
Such equipment serves two main purposes. In the first place it provides
useful data for cost accounting without which in a multi-machine mill only
very crude estimates of operating steam costs are possible. Secondly, the
effect of changes in clothing and other operating conditions can be studied
with a closer knowledge of the steam costs involved. Attention has already
been drawn in this respect to the fact that the main dryer or M.G. steam
pressure found necessary to dry the paper to a given moisture content does
not of itself necessarily alter in accordance with the actual quantity of
steam being used: it is quite possible for a rise in the main steam pressure
used for drying to be associated with a drop in the quantity of steam used.
To assess the full effect of any change to the drying section both pressure
and flow measurements are necessary. The same applies to the steam used
in felt dryers, hoods, or the machine house ventilation system, though for
these positions the use of steam flow measurement is essentially to compare
500
THE DR YING SECTION AND CALENDERS 5C.1 2
the effect of changes in operating conditions, e.g. in the felt dryer pressure
or the temperature of a Grewin or H.V. hood air-system, on the com-
parative steam flow rate to the equipment concerned as opposed to the
drying cylinders; on modern machines the steam used to heat air in hoods
can represent a third of the total heat consumed by the drying section and
so obviously requires careful setting. Without flow meters on the various
steam lines to each separate part of the system, it is hardly possible to
assess the effects of changing operating conditions sufficiently accurately; a
single flow meter measuring the total steam used on a machine is a cheaper
substitute, but is rarely satisfactory because some steam is likely to be
supplied at a higher pressure than the main dryer steam and also because a
single measurement will permit only relatively large changes in total flow
to be detected.
Measuring the flow of condensate from a machine is also important.
This reading, especially when integrated over a suitable period, can be
compared to a similar reading from a steam flow meter and thereby
indicate losses occurring in the system (important for reducing steam
wastage and because after de-oiling condensate water is more suitable and
cheaper than fresh water for re-use in the boilers); it also provides a useful
check on accuracy of the two instruments. This function of providing a
check on the accuracy of flow measurement also applies from the point of
view of the boiler-house unit, where water entering from condensate lines
together with that from returned low-pressure steam and from fresh water
sources can be balanced against a measurement of the total water entering
the boilers, a highly important piece of information for assessing boiler
operation. Finally, when used in conjunction with a recorder, the overall
stability of condensate flow can be examined and this provides useful
information on how well the machine drying section, especially in a cascade
system, is functioning: surges in condensate flow can be surprisingly large
and may be a direct consequence of poor operation (water-logging of
cylinders, temporary blockage, faulty traps, etc.) that is difficult to detect
in other ways.
The purpose of draw indicators between each section of the dryers,
calenders, and reel-up is basically the same as that already discussed with
reference to the couch and presses. They are valuable for pre-setting draws
(especially when creping) and thereby remove a frequent source of trouble
in feeding-up; also alterations to the draw found necessary to operate
satisfactorily give an indication of the composition and structure of the
sheet. In the dryers, draw indicators are of most use when associated with
sheet tension measurements between the sections. This is because it is
essentially by the tension in the web that the draw is set. Tension can be
measured in several ways and the most popular is to use a full-width ro11
acting either against springs, the positions of which at any time give a
measure of tension, or against appropriate strain gauges in the journal
supports; the latter arrangement is often used in conjunction with draw
controllers (see 5C .1 4). Setting web tension between the different sections
is especially important when manufacturing a paper with closely-specified
stretch properties, but on any machine facilities for keeping the tension at a
17 501
5C. l 3 THE DR YING SECTION AND CALENDERS
low value are valuable for avoiding creasing, quality variation, and breaks
caused by overstraining the web.
Felt tension devices which rely solely on positioning a stretch roll are far
from satisfactory because there is no means of judging (except crudely) at
what tension a felt is running. With such equipment some method of
measuring felt tension (similar to those used for measuring web tension)
provides a most valuable indication of running conditions; for this purpose
a simple spring-supported roll with a lever to magnify movement is an
inexpensive though adequate means of giving a qualitative indication.
The tensioning arrangement which involves the use of weights sliding on a
vertical carriage improves considerably on the simple stretch roll, and the
weights in use on any felt are normally rarely altered so that in effect felts
are always kept at the same tension. Measurement of felt tension is useful
(and readily available) in association with a pneumatic or hydraulic loading
system on the stretch roll and this permits running tension to be set and
altered as desired; in particular, automatic slackening of felt tension at a
break is possible.
The pressure applied by the machine roll on to the winding roll of a
drum reel-up can usually be varied either mechanically by weighting the
lever arms or by means of special pneumatic or hydraulic loading cylinders.
Whatever the system an indication of this load should enable the dryerman
to produce a roll to a more uniform hardness especially when, as is gener-
ally the case, the load has to be altered anyway as the roll grows to take
care of the changing pressure on the drum.
SC. 1 3 Dry-end paper quality instrumentation
Several instruments have been designed for the continuous measurement
at the dry-end of various properties of the finished sheet. These include
substance (dealt with in Part 1), moisture and thickness (discussed separate-
ly in SC. 3 and SC .4), formation (see also Part 3), and gloss, opacity,
colour, air permeability, and edge-of-web curl; in most cases the instru-
ments are available commercially and can where appropriate be arranged
to traverse across the web. There are also devices which scan the sheet for
dirt specks (using an optical transmission system) and for the presence of
holes and wrinkles (this can be done either with electrical feelers, or using
optical methods or ultra-violet light). This whole group of what may be
termed 'quality' instruments will now be considered in broad terms; the
value and operation of individual instruments will not be discussed because
their importance and possibilities depend closely on the particular paper
being manufactured.
To begin with it is worth stressing that no advantage is gained from
measuring any property of the sheet continuously for its own sake.
Substance, moisture and thickness are important basic properties of any
paper and can readily be adjusted or automatically controlled (along if not
across the machine) to enable a more consistent product to be turned out;
for these reasons a strong case can be made out for the continuous measure-
ment of these particular properties at the dry-end. In addition, other
characteristics of a paper almost always depend in some way on substance,
502
THE DRYING SECTION AND CALENDERS SC.1 3
moisture and thickness, so these should first be under reasonable control
before venturing into further quality instrumentation. Once this is achieved
it may well be discovered that any changes in other properties, formation,
opacity and so forth, are relatively slow to occur; in this event it must
seriously be questioned whether the alternative of a quality (or process)
control system involving the testing of samples taken from each machine
roll may not provide adequate information and in the long run be simpler
to operate.
Generally, whenever a quality instrument is installed at the dry-end and
found to function reasonably well one of the earliest tasks is to try to set up
some sort of control system. The pitfalls inherent in this have been well
documented by MacLaurin (63) who has illustrated the important but
neglected point that no control system is adequate unless it controls the
right variable: he cites an instance where control of opacity by addition of
titanium dioxide proved extremely expensive compared to making other
less easily controlled alterations of preparation conditions which could
achieve the same end. All qualities such as air permeability, curl, forma-
tion, gloss, and so on are affected by numerous variables on the machine.
The alterations which are practicably possible and which lend themselves
to being coupled up to some form of control system are often not the ones
which have most effect on the property in question. Difficulties with
excessive time-lags in the system also occur. Hence control can be difficult,
or at worst damped to such a degree as to be virtually ineffective.
A better approach to the problem of general quality variation may well
be to improve overall machine instrumentation and control a greater
number of characteristics right through the system. With carefully
documented data showing as comprehensively as possible the operating
conditions during the making of each grade, and with a well-run quality
control system checking the performance, it should be possible to keep
quality consistent both within and between different makings. The main
shortcoming of relying solely on this general method of achieving greater
product consistency is the difficulty of experimenting to improve or alter
the paper quality in some way. Sometimes a relatively long-term statistical
method such as Evolutionary Operation can be resorted to because this
does not require changes in operation which are immediately evident.
Otherwise production time has to be sacrificed in order to make it possible
to use the classic techniques of experimentation which, to yield useful
results, require adjustments to the process of an order that result in readily
detectable alterations and so almost always involve the making of some
paper that is out of quality specification. Continuously-measuring quality
instruments can provide a valuable function by allowing the effect of small
operational alterations to be observed more readily, thus giving informa-
tion which must otherwise be obtained more laboriously. But for this
purpose it can be argued that only an experimental unit is necessary, a
topic beyond the scope of the present discussion.
It is also worth pointing out that even the use of continuously-measuring
instruments to enlarge knowledge of the system may in time be superseded
by the advent of 'on-line' data-logging computers which incorporate
503
5C. l 3 THE DR YING SECTION AND CALENDERS
quality control results with routine data from machine instruments. One of
the prime values of such computer installations is likely to be their facility
for mathematical model building which will enable the effect of all the
measured operational variables on each aspect of quality to be effortlessly
isolated, leading to a clearer recognition of what is involved in keeping
each property uniform; following on this it is possible to envisage auto-
matic optimization of each important property (taking care even of
uncontrollable environmental changes) by use of a systematic procedure
for altering the set-points of each controller on the machine and searching
for the positions giving the best overall result. Such a system, the ultimate
in automation, is, of course, very much a pipe-dream at the moment so far
as the paper machine is concerned.
Quality instrumentation at the dry-end almost always requires very
careful standardizing and frequent attention if it is to be kept at an
operational level which is sufficiently useful to be relied on; it is also not
always easy to check how the properties actually measured by the instru-
ment compare with those which it is desired to measure, particularly with
regard to subjective qualities where laboratory instruments can more
readily be compared with subjective assessments by carefully controlled
ranking tests. In the case of devices such as dirt counters and hole detectors
it is arguable whether the machine dry-end is the best place for them any-
way. The main use of this type of equipment is in automatically sorting out
paper which must be rejected because it is below the acceptable quality
level required; on the paper machine itself this function is hardly suitable
unless it is thought practicable to have tabs inserted in the roll according
to the dirt-level, number of holes, etc., registered in a given time. The only
advantage in using such devices actually on the machine would seem to be
the quick availability of information about deteriorating conditions which
would enable rapid steps for correction to be taken. Probably the winder
taking machine rolls immediately after they are thrown out is the best
position to compromise the two functions of rejection indicator and early
warning system.
Normally any quality instrumentation has to be installed between the
calenders and reel-up in order to measure the appropriate paper property
in the condition most closely approximating to the finished, saleable
product. The result is that especially when measuring systems with cross-
web traversing facilities are required the available space between calenders
and reel-up can easily become cluttered with arms and beams, switches and
terminal boxes. This is hardly likely to endear the devices to the dryerman,
particularly when he finds it difficult to feed up the sheet without injuring
his rear or his head on some projecting piece of hardware. But specialized
forms of 'quality' instrumentation are continually growing in scope and
value so some means must be found to accommodate equipment that will
eventually be considered vital to successful operation. As most of it is
electrical or electronic in operation, and is generally sensitive to dust,
ambient air eddies, and sharp knocks (intentional or otherwise), it seems
likely that a large box will eventually be found necessary to enclose all the
various instruments; this box would stretch across the full-width of the
504
machine, with a slit for feeding into from the calenders and appropriately
placed air blowers to lead the tail from an outgoing slit on to the reel-up-
it would, of course, be painted black.
SC . i 4 Control applications
There are a number of control systems often associated with the drying
section, of which the most important involve control of the supply steam
pressure (with developments from this), and various aspects of drying
cylinder and hood (ordinary and H.V.) operation. Control of pressure of
the steam supply to the drying cylinders and other parts of the machine has
been found extremely effective in probably all cases where it has been
installed. The control is arranged quite simply by using a pressure or
temperature signal from the main supply line (see SC .11) to adjust the
position of the main valve; even when up-stream pressure control is very
good the changes in steam demand at a given pressure require that an
integral or reset action be incorporated in the controller to avoid offset, a
condition in which the actual pressure controlled at differs from that
desired.
The main advantage of this simple steam pressure-control system is that it
reduces fluctuations in the drying which manifest themselves at the machine
roll by sudden and erratic periods of dampness, perhaps associated with
blackening or creases. This in turn permits a higher average moisture
content to be attained with all the advantages this carries. With M.G.
cylinders, where the temperature must be raised gradually at start-up to
avoid stressing the metal, a normal steam control system is readily adapted
to ensure a regulated rate of introduction of steam during heating-up. As
with most control systems, the position of the steam control valve should
be shown on the same recorder as the actual steam pressure to permit
operation of the system to be examined. Link-up with moisture meters to
provide a more elaborate control taking account of variations in conditions
affecting the rate of drying, e.g. moisture content of the web entering the
cylinders and ventilation conditions, is discussed in SC. 3.
Other control systems associated directly with the drying cylinders and
with hoods are becoming more common and increasingly elaborate; most
have already been described in the appropriate context and will be men-
tioned here only briefly. Control of drying cylinder operation usually
depends on ensuring that the differential pressure across a bank of dryers
remains steady and this can be achieved either by regulating a supply of
make-up steam into the flash side, or by adjusting a throttling valve in the
condensate or in the flash-steam line; often one or more of these methods
is used simultaneously and separate arrangements allowing for the change
in conditions at a break can be incorporated. The temperature of steam in
any individual section can be controlled exactly as the main supply steam,
and it is possible even to have a control giving a pre-set temperature
gradient along the first few dryers. Similar arrangements are used for
cascade systems, M.G. cylinders, and thermocompressor systems. The level
in all condensate receiver tanks is controlled, also the pressure or vacuum
in the final receiver (usually by varying the quantity of cooling water
505
5C.2 THE DRYING SECTION AND CALENDERS
injected by spray or circulated through coils). Other control facilities, e.g.

automatic venting of flash tanks to keep the air content below a desired
minimum, can also be installed.
With ordinary and H.V. hoods control of supply air conditions with
respect to any desired temperature and humidity is commonplace; exhaust
air conditions can also be controlled, normally to a desired humidity level,
by regulating the quantity of air recirculated. The precise manner in which
this is achieved depends essentially on the system lay-out, whether heat-
exchange equipment is incorporated in the exhaust, and the steam supply
available for heating incoming make-up air. Such controls are often con-
sidered to be an integral part of the design of any type of hood and are
valuable for ensuring successful operation at an economic level.
Two other control applications are occasionally found in the drying
section. One is control of the draw from a measurement of web tension, a
direct development from having some simple tension measuring arrange-
ment. This control can be achieved either mechanically or electrically,
depending very much on the drive system, and due to the sensitivity of
draw settings in the dryers can be a valuable facility for reducing breaks.
The other application is control of a dry-end pulper: this involves prim-
arily level control together with, where possible, control of consistency and
temperature of the stock also; various methods of achieving this have
already been described earlier. The object here is efficient and uniform
disintegration of broke to make its introduction back into the preparation
plant or machine chest as trouble-free as possible. To this end a useful
feature of any dry-end pulper operation is to have a means of reducing the
rate of disintegration when little or no broke is entering the pulper; as it is
normally preferable to maintain continuous operation if only to be ready
for a sudden break at the dry-end, this really precludes the obvious
possibility of using the pulper on a batch system, and some arrangement is
therefore useful to idle the rotor when the flow-rate from the pulper drops
below some pre-set value. Another useful facility is an indication at the
dry-end of the main broke tank level: as the supply of dry broke fed back
into the system needs to be steady to help maintain uniform preparation
conditions, it is evident that the broke tank level must so far as possible be
kept between two points, an upper one above which the broke usage needs
to be increased to prevent the risk of overflowing the chest and a lower
point below which broke usage must be reduced to avoid the possibility of
emptying the tank. By having an indication of broke-tank level at the
dry-end, especially with an appropriate warning light to show when level is
dangerously high, the dryerman is able to regulate addition to the pulper
of broke that has been torn off machine rolls or trim from the winders, etc.;
also he can keep an eye on the level during a prolonged break at the
calenders when the full machine web is falling into the pulper.
SC.2 MAINTENANCE OF THE DRYERS AND CALENDERS
SC. 2 I Felt changing
The changing of dry felts is an essential part of the maintenance of the
drying section, though the problems involved in performing this system-
506
THE DR YING SECTION AND CALENDERS SC.21
atically are by no means as complicated as for wires and press felts. This is
because dry felts invariably last many months on a machine and represent
only a few coppers of the cost of producing a ton of paper. Unfortunately,
it is perhaps for this very reason that on many machines dry felts are
neglected and become an unnecessary source of lost production.
As with other clothing, a record is kept in which the various makes and
types of felts used in each position of the dryers are listed, together with
their cost and general comments on how well they ran on the machine. In
addition there will be some means of comparing the performance of
different types of felt, and commonly this will either simply be the life in
days or weeks, or the tons made on the machine during the felt life. This
sort of information ensures that only suitable felts are used in each position
and in the long run assists considerably in discovering the physical make-
up, material, and other characteristics of the felt best suited to each
particular application.
The weakness of this simple record system, as with press felts, is that it
emphasizes the importance of felt life. Other aspects of performance will
certainly be taken into account; in particular, systematic records will
reduce the possibility of getting felts which are unsuitable for reasons such
as shedding of fibres, tendencies to harden or burn, and undue dimensional
instability which causes the stretch gear to be strained at a break or
necessitates taking a piece out of the felt and re-seaming after a few days on
the machine. These points are certainly of importance, and it is most
desirable to select felts which require as little attention as possible through-
out their life. But especially with dry felts there is absolutely no advantage
gained from prolonging running time on the machine excessively. Many
papermakers mistakenly pride themselves on the life of their dry felts and
will not take one off until it is threadbare and almost in shreds. Certainly
there is little evidence that deterioration even to this extent affects the
overall drying unduly, particularly in the case of felts used towards the
end of the dryers where they play only a relatively small part in the
mechanism of drying; but it is nevertheless evident that a policy based on
squeezing as many weeks as possible from each dry felt will eventually cost
far more in downtime, for repairing unmanageable holes and for un-
scheduled shuts when a felt suddently tears apart, than can possibly be
saved in felt costs.
The author favours a system of scheduled changing for dry felts based so
far as possible on a regular pattern. With such a procedure, each dry felt is
changed after a pre-determined number of weeks and a sequence is devised
so that no two felts (particularly the first two) are changed at any one time.
The main advantages of this are that it allows better planning of the work
involved, enables stores of new felts to be ordered more systematically, and
prevents any sudden change in drying conditions arising from having to put
two or more new felts on at the same time. Whenever possible the same
standard life is chosen for each felt, thus permitting a regular pattern to be
formed: for example, it may be decided to run the eight dry felts on a
machine for a life of 32 weeks each, so one is changed ever 4 weeks.
Accidental damage, trials, or unexpected deterioration will of course throw
507
the sequence temporarily out of gear, but it is always simple to adjust the
changing dates to correct this.
Adoption of a planned changing scheme of this type can mean removing
some felts that look far from worn. But it is likely that an old felt does
begin to have some adverse effect on drying as it hardens and becomes more
impervious with age, even if only in causing some variation in drying
across the machine which affects the moisture level of the machine roll and
thereby necessitates a reduction in the average moisture content; if this
were the case then clearly it may not be uneconomical to remove a felt
before it has reached the end of its practicable running life. But leaving this
point aside, regular removal somewhat earlier than the former average life
invariably pays for itself by reducing to negligible proportions the need for
stopping the machine to change or repair unexpectedly.
This is not to say that once planned changing is introduced no further
effort is made to improve performance; new types of felt are continually
appearing and it is the papermaker's duty to assess their virtues for the
machines in his charge. Provided the new felt runs satisfactorily, it would
be usual practice to relate the average life to the cost of the felt and so
compare the running cost with that for the usual type of felt. In some
positions this can mean that certain hard wearing materials, e.g. synthetic
re-inforced and terylene, are economically preferable because the extremely
long life which they give far outweighs the initial higher cost for the felt.
Even when the longer life of such a felt is more or less balanced by its
greater expense, reduction of labour charges from changing less frequently
and the lower Jikelihood with slower deterioration of a sudden change
being needed make it the preferable choice. But against this, higher storage
costs for replacement felts (especially if a felt is individual to one position
in which case normally at least two must be kept lest one is damaged when
being put on the machine), and the risk of greater loss resulting from
accidental damage, means that little advantage is gained by achieving a felt
life in excess of one to two years.
SC. 2 2 Influence offelts on drying

A feature of dry felt performance which is invariably neglected is the effect
on drying rate and steam utilization efficiency. But much evidence has
already been given to show how different types of felt alter the drying, and
the effect this has on steam costs, or on drying rate when production is
limited by restricted drying pressures, is of obvious importance; as an
example, a change to a thicker and heavier felt of the same material and
weave may well yield a proportionally longer life and be more economical
to run, but both drying rate and efficiency could be reduced and this is
likely to be far more significant. One reason for the lack of attention
devoted to this aspect of dry felts is the complexity of comparing perform-
ance, especially on machines making many grades; but the potential saving
in this respect is far more important than any gain in reducing felt costs,
and no change to a new type of felt should be made without making some
attempt to evaluate changes which occur in the drying. The means of doing
this will now be discussed.
508
THE DR YING SECTION AND CALENDERS 5C.22
Carrying out regular dryer tests, as described in SC. 2 4, is the best way
to have on hand enough data with which to compare the performance in
any particular position of different types of dry felt. Otherwise, for this
specific purpose the main information required for each felt is the drying
rate through the section concerned and some comparison of overall steam
consumption related to water evaporated (see below). The drying rate over
the full drying section may give an indication of some change, but where
more than four felts are in use it is preferable to take samples for moisture
content determination from the middle of the dryers in order to determine
the water removed in the particular section concerned; several determina-
tions through the life of the felt are required and so far as possible, to
avoid complicating the issue in the absence of other data, these need to be
obtained at times when conditions, particularly machine speed, grade of
paper, and steam pressure, are nearly the same (this is of course because of
the dependence of drying rate on steam pressure and moisture content
entering the cylinders).
It must also be remembered that each section usually contains two dry
felts and the effect of changing one may be partly masked by the other.
One way to avoid this and isolate the effect of an individual felt is to
determine the rate of condensate flow from several consecutive top and
bottom cylinders in the section concerned (though see SC. 24 for comments
on this). The relation between the average flow from cylinders covered by
the new felt with those opposite covered by the other felt is then compared
with the same relation for the old felt (it is necessary to take a comparative
figure in both cases because even with identical felts on both top and
bottom cylinders the evaporation rate will not normally be the same due to
the influence of ventilation conditions, felt age, felt drying capacity, and so
forth): for example, if in one position on a machine a new type of top felt
is tried, an improvement in drying rate should be indicated by an increase
in the ratio between the average condensate flow from two or three of the
top cylinders and the average flow from the adjacent bottom cylinders
(still covered it is presumed by the same felt as before). Alternatively,
samples of the web for moisture content determinations can be obtained
from between several consecutive cylinders and used to calculate the water
evaporated by each cylinder, the data being used for comparative purposes
in exactly the same way. The difficulty with this method is devising a
satisfactory sampling technique (see SC .2 4).
It may be argued that a change in drying rate, to be really significant,
should be readily apparent from the change in steam pressure needed to
dry the paper under normal operating conditions. This is perfectly true,
though it is always preferable to have more systematic data on which to
base a decision as to the effect of any particular felt on the overall drying
rate. It is after all a rare occurrence when the substitution of a new felt
makes a difference to the drying rate which is immediately obvious.
The other aspect of felt performance which can be of importance relates
to the efficiency of steam utilization. As steam flow is never measured to
individual felt sections the only means of assessing any change of this
nature is by relying on data for the total steam flow to the whole dryer
509
SC.2 3 THE DR YING SECTION AND CALENDERS
section and relating this to the water evaporated. This can be done by
comparing results obtained in a number of tests on both old and new felts,
in each case using a steam flow meter and sampling the web for moisture
determination entering and leaving the dryers. Alternatively, where a
machine makes virtually the same grade all the time, over a number of
weeks a significant change may be detected in the ratio of the total steam
used in the dryers to the total paper dried (provided downtime does not
vary greatly).
SC. 2 3 Other maintenance

Apart from the dry felts, there are several other parts of the drying section
that require regular inspection and maintenance. Sheehan feed ropes can
cause a surprising amount of downtime if insufficient supervision is given
to their operation. As an item of relatively small cost it is far preferable
that ropes are changed at frequent fixed intervals chosen to reduce to
negligible proportions any attention to them which necessitates stopping
the machine. In between changes the ropes should be regularly inspected
for signs of fraying and loosening splices, and the tension gear should also
be checked for free movement; cross-over points in particular must be set
carefully. A record of the life of ropes in each position is essential to show
up excessive wear due to sticking or misalignment of pulleys, rubbing on
the frame or other ropes, and so forth.
The outer surfaces of drying cylinders and M.G. cylinders, especially the
latter, should be kept as bright and polished as possible and if the polish
cannot be maintained satisfactorily while the machine is running then
periodically the surfaces (particularly of the wet-end cylinders) have to be
scraped or buffed before they become too rough. This is done on drying
cylinders by using a short length of sharp, heavy deckle-blade or some
other appropriate scraping device on the end of a pole which is slowly
pushed across the cylinders as they are crawled round; where there is room
in the cylinder pockets this procedure is speeded up by fixing two or three
blades together in such a way that both an upper and lower dryer can be
scraped simultaneously. On M.G. cylinders special buffing wheels are
used (see 5B. 8). The insides of drying cylinders should also be inspected
more frequently than is generally the case. Not a great deal can be done to
remove uneven deposits of rust and scale, at least during normal scheduled
shuts, and this is one potential advantage of using filming amines; but
siphon tips (including the simple open-ended pipe) are a frequent source of
waterlogging due to rusting, erosion, and damage, and whenever practic-
able one or two cylinders should be opened up on a rota each shut period
for general examination and to check the clearance between siphon and
cylinder.
Regular inspection and a record of the life, pattern of wear, and condi-
tion when removed of all doctors is most important, and this applies
especially to M.G. and calender doctors. These are preferably placed on a
preventive maintenance system to avoid neglect. Alterations to operating
load and position or angle should also be noted; oscillating gear and the
510
loading mechanism need occasional examination. Sprays used for damping
purposes require frequent attention for signs of clogging or wear in the
nozzles and periodically should be tested off the machine for flow rate in
different regions along their length. Rolls in which cold water is circulated,
the drum reel-up, sweat roll, calender rolls, etc., need cleaning at intervals
with an appropriate chemical to remove scale and rust. Hoods, H.V. in
particular, hot-air nozzles, drying and other ventilation equipment all
require thorough cleaning from time to time to remove accumulations of
dirt and dust which block orifices and could become a fire hazard or
contaminate the sheet.
Maintaining the camber on breaker stack rolls, smoothing press, M.G.
pressure roll, etc. can be treated as a similar problem to press rolls.
Calender roll camber maintenance is discussed in 5C. 4 but it is worth
noting here that any rolls in the stack not in use when the machine is
running should always be greased to prevent rusting and hung separately
to avoid the appearance of fiat spots.
SC. 2 4 Long-term records
Checking the drying section from time to time for the purpose of collecting
long-term records of performance is all too rarely done in paper mills.
One reason for this is undoubtedly the difficulty of obtaining sensible data,
especially when different makings are frequent} and the general feeling
that what measurements are made are likely to be neither accurate nor
necessarily representative of the typical state of affairs. Nevertheless this is
the section of the machine which is generally the costliest to run and
potential savings in power and steam consumption are high enough to
warrant some examination of the system at regular intervals. This is
directed broadly towards two distinct ends. One is to keep a check on the
drying rate and steam utilization efficiency to detect any gradual deteriora-
tion and provide data with which to compare performance should any
alteration be made (e.g. type of felt used, operating conditions of felt
dryers or hot-air equipment, etc.); this should embrace the ventilation
system and can be as comprehensive as desired, though strictly speaking
it is necessary to undertake separate checks for each of the major grades
of paper produced and this increases appreciably the work involved. The
other purpose of dryer tests is to ensure that the drying cylinders and
flash system are retaining their general effectiveness and providing
the desired pattern and uniformity of cylinder temperature through the
whole section.
Consider first the question of drying rate. To determine this it is only
necessary to extract samples at several places across the web both entering
and leaving the drying section (for convenience this is done at the same
time as testing the wire and presses, but particularly it is preferable to
avoid times when wet felts are nearing the end of their life otherwise the
drying efficiency can appear higher than it really is); from moisture deter-
minations on these samples the lbs. water evaporated per lb. paper pro-
duced is calculated (using the formula quoted in SA. 2 4). The production
rate multiplied by this figure then gives the appropriate drying rate. As an
511
alternative to breaking the sheet for samples after the last drying cylinder,
it is usually considered adequate to sample off the machine roll and assume
the moisture content is the same; if a sweat roll or spray damper is used,
then it is necessary to revert to sampling the web leaving the dryers and
also allow for the added moisture before using the production rate figure as
this will be based on weight at the reel-up. The drying rate is dependent
closely on steam pressure and also on the moisture content of the web
entering and leaving the dryers. To interpret the data it is therefore neces-
sary to keep note of these figures also.
This simple dryer test provides useful information but in practice the
actual rate of drying on any particular machine is really important only if
operating steam pressure is close to the maximum available making it very
desirable that a close watch is kept on this aspect of dryer performance. A
more interesting figure is the steam utilization efficiency, for which the
total steam usage must also be known. Efficiency is then obtained from
dividing steam used by the water evaporated in the same period; the
theoretical minimum is 1·12 but normal values vary from 1·6 to over 2· O
depending partly on the grade of paper being dried. This figure is relatively
independent of machine conditions, i.e. speed and moisture content of the
web, and provided test data is compared for the same grade of paper it
gives a more useful indication of any trends in efficiency of the drying
section as a whole.
During any test of this nature other available measurements will be
included relating to general running conditions in the dryers: these include
steam pressures in the flash system, felt dryers, and so on; measurements
of air condition in the ventilation system; load on M.G. pressure roll,
breaker stacks, etc.; power consumption in the different sections; draws,
sheet and felt tensions. Ideally a complete heat balance should be derived
showing, possibly in the form of a Sankey diagram though this does not
lend itself to comparative purposes, the heat flow in the machine house as a
whole; this would cover all heat supplied in the machine house (including
general heating of incoming air, of back-water and other ancilliary points),
heat leaving the machine house in the condensate and air exhaust, heat
usefully employed in evaporating water, together with re-circulation of
heat in exchangers. A balance agreement within 10 per cent. can be con-
sidered good (especially when a machine is not hooded) because accurate
sampling of air ducts and other places is difficult, especially to measure
humidity (see ref. 17); it is not practicable anyway to take account of all
heat sources (e.g. from motors and friction) or heat losses (e.g. evaporation
from open surfaces and general building losses). Certain values such as the
temperature of the condensate water vary little and have only a small effect
on the heat calculation; these may safely be taken as standard. Other values
such as air flows are best calculated using fan performance curves. Humid-
ity is most accurately determined with conventional wet and dry bulb
thermometers, though hair hygrometers and dew point measuring devices
are most convenient.
Several key figures can be derived which aid interpretation of such
comprehensive data. In particular it is useful to know the proportion of the
512
heat supplied which is usefully used (another measure of utilization
efficiency), the overall heat losses and their relation to ambient tempera-
tures, the weight of air used in comparison with moisture evaporated (a
measure of ventilation efficiency and dependent on incoming humidity
which governs the amount of moisture it is possible for the air to pick up),
and the relation between power and steam usage with particular emphasis
on hot-air supply units. In addition periodic tests of this thoroughness help
to show up deficiencies in such things as fan operation, heat-exchange
equipment, hot-air nozzles, etc. A good illustration of the determination
and derivation of this sort of information together with a suitable means of
setting out the various figures has been reported by Chalmers (8), to which
reference may be made for further details. Other illustrations and some
case studies appear in an article by Mardon et al. (125).
It is obvious that the large number of different positions at which it is
necessary to check air condition for velocity, temperature and humidity,
determine flows of steam, and so forth means that a substantial amount of
work is entailed. Making provision for such tests by fixing suitable tapping
points in air ducts and installing separate flow meters, pressure gauges,
ammeters, etc. at key points can be quite an expensive business. Possibly
the best solution in the case of a mill venturing into this field for the first
time is to consult a heating and ventilation expert who could advise on
what measurements are needed and where they should be taken. As a
preliminary to this it will be necessary to survey the whole system, including
probably the boiler plant and power generation units, and this may well
produce valuable suggestions leading to economies in running cost. In
connection with this, in this country the National Industrial Fuel Efficiency
Service is an obvious body to tum to (66).
The other aspect of the dryers requiring periodic checks relates primarily
to the effect of drying on the paper. To this end it is most important to
check the rate of heating up of the sheet and the subsequent drying
gradient. The usual method adopted is to determine the surface tempera-
ture of each cylinder using a pyrometer attached to a suitable bar. Although
none of the various makes of pyrometer available can be relied on to give
very accurate measurements of absolute temperature, for comparative
purposes in any one test most types are satisfactory provided cylinders are
not coated with scale (this changes the closeness of contact and emissivity
of the metal surface, both of which affect the temperature sensitive
element); to avoid the influence of frictional heat (and reduce the degree of
discomfort involved) response time of the instrument should be rapid. The
average temperatures of top and bottom cylinders generally differ, but the
difference between adjacent top or adjacent bottom cylinders need not
exceed about 5 deg. F. The gradient of temperature rise at the beginning
of the dryers (and in some cases the fall at the end) can be usefully checked
by this means, but perhaps the greatest benefit derives from pinpointing
waterlogged cylinders.
An alternative measurement is to time the condensate flow from each
cylinder but for this it is necessary to install a T-branch on the condensate
line with suitable cocks to divert the flow either into a steam trap or, when
513
suitable, direct to atmosphere; a bucket and stopwatch suffices to check

the flow rate. The trouble with this method is that under the usual condi-
tions of measurement the pressure on the condensate line from the
cylinder is altered when the flow is diverted: it may be raised if a steam trap
is used or lowered if the flow is allowed to discharge to atmosphere. This
could alter the volume of condensate in the cylinder and affect the condi-
tions inside sufficiently to change the condensation rate. It is however
likely that any change will be relatively small as the main factor governing
condensation rate is heat demand from the paper; provided the tempera-
ture of the inner surface of the cylinder is relatively unaltered, i.e. the
thermal insulation of condensate in the cylinder is unaffected, then the heat
transfer rate and hence the condensation rate should be unaltered too. But
as a precaution when using this method condensate flow from any cylinder
must be allowed ample time to settle down before measurement is made.
In any case the flow rate between cylinders working at substantially differ-
ent pressures (e.g. at a change in flash section) must be compared with
caution. It may be noted that there is a relatively simple means of avoiding
this particular difficulty and this is by making provision for installing a
pressure gauge and a valve in each condensate line; with the flow diverted
the valve is then used to adjust the pressure before measurement to make it
the same as during normal operation.
Other measurements desirable for long-term comparisons of the effect
of drying on the paper are various checks of air conditions within the
pockets and by the side of the cylinders. Temperature and humidity require
measurement with standard instruments at a number of defined positions,
and air velocity can usefully be measured with a velometer type instrument.
The web itself can be sampled at various points in the drying section and
tested for moisture content and temperature; for this purpose it is best to
use two cutters on the wire about 9 to 12 inches apart in such a position as
to allow a piece of the web as far from the edge as possible to be grabbed.
The accuracy of this type of test is not great due to the continually changing
condition of the web in each open draw (in particular temperature falls
rapidly as evaporation occurs in the first few inches after leaving the
cylinder surface), but for rough comparative purposes of the drying rate at
various positions in the section it is quite satisfactory. The width of the
sheet entering and leaving the dryers provides a measure of the cross-
machine shrinkage rate.
To complete information for long-term records, conditions in the
calenders with regard to load, number of nips used, steam or water flow
and pressure, etc., pressure of winding at the reel-up, water flow to
dampers, and of course quality tests on the paper itself are all required.
Cross-web temperature checks for cylinders and calenders may be con-
sidered a worthwhile addition in some circumstances but would not
normally form part of the measurements taken for the purpose under con~
sideration. If differential pressure tappings are available on the steam and
condensate lines to each cylinder these pressures can also be checked, but
this information is more useful for specific trouble-shooting than for long-
term records.
514
THE DRYING SECTION AND CALENDERS 5C.3
SC.3 CONTROL OF MOISTURE CONTENT
It is the dryerman's duty to regulate moisture content of the sheet at the
reel-up, and on most machines this is likely to be the most important of his
tasks and the one which requires most skill. This applies especially to the
need to attain uniform moisture content across the width of the web, a
problem which is one of the most pressing in need of solution. Reasons for
the importance of moisture control and the means available for achieving
uniformity in both machine and cross-direction is the subject of this
section.
SC. 3 1 Moisture content and its relation to substance
If asked which is the most important property of a paper that must be kept
in specification, most production personnel will undoubtedly agree on
substance. And certainly every grade of paper must be matched as closely
as possible in substance if it is to have some semblance of continuity from
one making to the next. But once the machine has been adjusted to give a
substance approximating to that required, it is debatable that moisture
content then becomes the more vital variable to keep under close super-
vision and control.
There are several reasons for this, the most important being that satis-
factory uniformity of substance in any case cannot be achieved without
reasonable moisture control. This is because moisture in the paper is an
integral part of the substance, and if the moisture content can vary
appreciably there is no means of telling whether a change in substance
should be corrected by an alteration to the fresh stuff valve, or the steam
flow valve, or both. To make this clear, suppose that moisture content does
in fact vary on a machine and no means is available for its measurement;
then an observed substance increase (above that required) of, say, 3 per
cent. at the end of a particular machine roll could be caused equally well by
a 3 per cent. increase in fibre entering the machine system as fresh stuff, or
by a 3 per cent. increase in the moisture content, and is more likely to be
due to a combination of both. There is no means of distinguishing the
effect of dry fibre or moisture solely by weighing the paper.
Even when the moisture content is known precisely (a difficult enough
matter, as will be discussed shortly), the remedy for the 3 per cent. increase
in substance will not be immediately obvious. Suppose for the sake of
argument that it is found by end-of-roll testing that moisture has increased
by 1 per cent. at the same time that the substance has shown the increase of
3 per cent. The normal reaction to correct the situation would be for the
machineman to alter the stuff valve to reduce substance by 3 per cent., while
the dryerman alters the steam valve to reduce moisture by 1 per cent., each
thereby correcting the observed errors in the property under his super-
vision. But the result of this combined effort may well be that at the next
roll testing reveals substance to have decreased not by 3 per cent. but by
as much as St per cent., thereby making it 2-! per cent. light instead of
correct weight, while moisture content has decreased not 1 per cent. back
to its former level but by no less than 2t per cent.! Both machineman and
dryerman would then take drastic action, this time to increase substance
515
5C.3 I THE DRYING SECTION AND CALENDERS
and moisture, and the way is open to a session of wild hunting about the
required values that is a common feature of process control charts.
At first sight this example of the changes that might occur in such a
situation seems incredible, but a little careful consideration of what
happens when substance is deliberately altered makes the reason clearer.
Results on the first roll (substance up 3 per cent., moisture content up 1 per
cent.) could well be due entirely to an increase of the order of 2 per cent. in
the flow of fresh fibre to the machine, i.e. in the dry weight of the sheet;
this is because a 2 per cent. increase in weight of fibre to be dried reduces
the overall rate of drying per unit area of web, and this could easily result
in the measured increase of 1 per cent. in moisture content. A 1 per cent.
moisture increase added to the 2 per cent. increase in dry weight produces
the observed 3 per cent. increase in substance. The remedy would therefore
be simply to alter the fresh stuff valve appropriately, leaving the steam
valve untouched. If instead the fresh stuff valve is altered to produce a 3
per cent. drop in substance, this could, if the change were in fact equivalent
to decreasing the dry weight by 3 per cent., lead to a decrease of about
4! per cent. in substance due to the additional reduction of 1t per cent. in
moisture content. Alteration of the steam valve by the dryerman to give a
further 1 per cent. drop in moisture content in an attempt to correct the
original condition leads to the overall decrease of 5-! per cent. in substance
and 2-! per cent. in moisture content.
This example must be subject to some qualification as it depends on the
assumption that a 1 per cent. change in dry weight of fibre will cause an
additional change oft per cent. in the moisture content-a purely hypo-
thetical figure. But undoubtedly some drop in moisture content must occur
when a lower quantity of fibre requires drying, just as it would were the
speed to be decreased slightly; the precise degree of change will depend on
the drying characteristics of the machine, behaviour of steam flow in
response to alteration in demand, and so forth, and also on the level of the
moisture content (higher values being more sensitive to changing conditions
than lower ones). But the example does serve to illustrate the sort of
dependence of substance on moisture which can be expected in practice.
Two points of importance emerge. Firstly, alterations of substance (in
the sense of changing fresh stuff valve position) and of moisture need to be
carefully correlated. The machineman and dryerman cannot go their own
ways independently, the one checking substance, the other moisture, as
their actions mutually affect each other. If the machineman finds cause to
alter the fresh stuff valve, the dryerman should be warned of an impending
change in moisture content. If the dryerman alters the steam valve position,
this may affect the substance to a degree which causes the machineman
unnecessarily to consider changing the fresh stuff valve position. It is not
possible to lay down any rigid procedure as this depends on the methods
available for measurement of both substance and moisture. But probably
the simplest aim is to keep moisture content as steady as possible by one of
the methods discussed below, then the machineman can more easily take
account of probable variations in moisture content before moving the fresh
stuff valve, making an alteration which allows for the fact that moisture
516
will also change and have its own effect on the apparent substance (this he
in fact learns to do in practice by observing how a given movement of the
valve affects substance). For efficient regulation of both properties it is
really necessary to control not substance but dry fibre weight, a subject
that has already been dealt with in lC.3.
SC . 3 2 Other reasons for controlling moisture content
There are several other reasons why it is important that moisture content at
the reel-up is controlled (using the general sense of the word). Firstly, any
sample extracted from the machine roll begins to pick up moisture at a rate
dependent on the initial moisture content and on the ambient air conditions
(the latter affect the equilibrium moisture content the paper would event-
ually attain). Hence when substance is measured on samples torn off at the
reel-up in the usual way, the result is affected by the gain in moisture
content up to the time the weighing is completed. The complication this
adds to the machineman's job of keeping the making substance consistent
and close to the specified value has been discussed more fully in lC. 3. The
less variable the moisture content at the reel-up, the easier it is to overcome
this particular difficulty and make an allowance for the difference between
substance as measured and as required.
Another reason for the importance of moisture control is the dependence
of calendering on this property. Any change in moisture in the paper at the
calenders will create variation in the surface and bulking characteristics of
the sheet. The work of Brecht and Heyn (33) discussed in SA. 4 3 indicates
the existence of an optimum moisture content for maximum smoothness,
but this is at about 15 per cent. moisture and so in practice increasing
moisture content at the reel-up can be expected to yield an overall increase
in smoothness and a decrease in bulk. This in turn affects other properties
of the sheet after calendering. With an M.G. machine, the moisture content
at the M.G. cylinder has a critical effect on the glazed finish of the sheet.
A final reason, probably the most valuable of all, is that generally
speaking it is almost always advantageous to obtain as high a moisture
content as possible in the sheet at the reel-up. To achieve this some form of
control becomes essential because the higher the moisture the more
variable it is liable to be. Attempting to obtain high moisture in the
machine roll without adequate regulation only leads to substantial
wastage either from blackening in the calenders or from creasing at the
winder due to uncontrollable differences in reeling hardness across the web.
The advantages of a high moisture content are worth examining for a
moment because they present in many ways the crux of the need for control
of this property and the potential advantages to be gained.
A high natural moisture content at the reel-up saves running costs
because less steam is used for drying. It also diminishes the likelihood of
static in the sheet and improves fold, stretch, tear and flexibility properties,
which are actually also those most affected by fluctuations in moisture
content resulting from changing ambient air conditions. For paper sold in
reel form high moisture content represents a considerable saving because
water is cheaper than the furnish. When sold in sheet form the same applies
517
but to a lesser extent (because it is more likely that moisture is picked up

during cutting, sorting and counting). Usually, however, reels are con-
ditioned to bring up the moisture content closer to its equilibrium value in
air. This is an expensive separate process, though with fine papers it often
becomes essential because it reduces the cockling and curl that otherwise
occurs if the sheet is allowed to pick up moisture naturally. Although in
theory conditioning should ensure that the paper is sold with adequate
water content, in fact it is more than likely that making substance on the
machine will be higher than it need be because the weight increase as
moisture is picked up will be underestimated-hence the weight of a ream
(and the furnish gone into it) will also be greater than necessary. Finally,
the higher the moisture content initially, the less is picked up before
equilibrium is reached; as pick-up of moisture produces deterioration in
smoothness and other qualities this is usually a further advantage to
producing a high natural moisture content at the reel-up.
SC. 3 3 Measurement of moisture; general comments
To be able to control moisture content adequately implies that it must first
be measured with reasonable accuracy. It is precisely here that the biggest
difficulty arises. None of the laboratory methods for determining moisture
content by oven-drying, distillation, vacuum desiccation, hot-plate and
infra-red drying, titration, and so forth are suitable for process control
purposes, mainly due to the time involved in obtaining individual readings.
For this reason a variety of portable instruments for measuring moisture
content have been marketed and these operate in several ways depending
principally on detecting the variation in electrical resistance or dielectric
constant (capacitance) of paper which takes place as the moisture content
alters (see next section).
There are two disadvantages to the use of such instruments for moisture
measurement. In the first place, even with the instrument by the side of the
machine it is extremely difficult to sample and test sufficiently quickly to
prevent a significant pick-up of moisture. Immediately exposure of the test
piece takes place, and this applies particularly in the damp atmosphere of
the machine house, the moisture content increases at a rapid rate: when
paper is made very dry on the machine, say in the 2 to 3 per cent. region, it
can in fact prove almost impossible without elaborate precautions to
obtain a reading below 5 per cent. If the sample is drawn from well below
the surface of the machine roll, then plunged into a plastic bag, weighed,
and moisture content determined by oven-drying, a very accurate result can
be expected. But if the sample is drawn from the same position and one of
the usual testing instruments is used, then the reading is invariably higher
due to the impossibility of avoiding some moisture pick-up before the test
result is obtained. The only way this can be minimised is by cutting a layer
of sheets off the machine roll and keeping the instrument well inside the
layer while measuring selected pieces.
The other disadvantages inherent in the use of an instrument is that it
has to be calibrated, and the calibration is rarely the same for different
types of paper. On machines making few grades this can present little
518
difficulty, but where a variety of loadings, furnishes and substances are
used it becomes an irksome task to check up on each grade. Frequently also
the relationship between instrument reading and moisture content is not
linear, so the task of calibration is further complicated. Under these con-
ditions it becomes inevitable that less reliance is placed on the actual
reading obtained on any sample, provided it is known to bear some
relationship to the absolute value, and more emphasis is given to the
variation from one sample to the next; in other words the instrument
becomes used primarily as a deviation meter.
These drawbacks to the use of off-machine moisture-measuring instru-
ments have led to the demand for instruments for measuring directly on
the paper machine web; these carry the advantage of making available a
continuous reading of moisture content, a facility having obvious value.
For this reason much effort has been made by instrument manufacturers to
develop suitable devices for on-machine measurement.
SC. 3 4 Continuous measurement of moisture content
Few devices for the paper machine have stimulated such a wide variety of
ideas as those directed to improving moisture control. It is not proposed
here to enter into a detailed survey of the different methods designed or
proposed (for many the information available is little better than sales
literature); this has been done already by many authors, notably Hurm
(62). What follows reJ?resents just a very brief summary of the salient
features of each method.
Continuous moisture-measuring devices can be divided broadly into
two groups: those attempting to measure some aspect of the machine or
the paper web which is known to be dependent on moisture content, and
those measuring moisture content of the web directly at a point close to the
reel-up. Into the first group fall the following systems:
(a) The use of a tension roll or bar in an open draw between two
cylinders near the end of the drying section; movement of the roll or
bar depends on varying shrinkage in the web and this is taken as an
indication of moisture.
(b) Measurement of sheet temperature with thermocouples placed at
fixed positions a short distance after the web leaves a pressure-
controlled dryer, the surface temperature of which is also measured
for reference (Hurletron system); this method relies on the tempera-
ture drop of the web leaving a drying cylinder being related to the
moisture content-the higher the moisture content, the greater the
temperature drop resulting from evaporation.
(c) Measurement of the steam (or condensate) flow in a pressure-
controlled pilot drying cylinder (Stamm system); this depends on
the fact that the more moisture there is in the web as it comes on to
the dryer, the slower the web will be to heat up and, as a conse-
quence, the greater will be the heat demand and the steam flow into
the cylinder.
(d) As an alternative to (c) the flow of steam can be controlled and
variation of the inside temperature of the cylinder measured.
519
(e) A further alternative to (c) is to measure the surface temperature of

the pressure-controlled cylinder, this being lower when there is more
moisture in the web; the difficulty with this method is that con-
tinuous measurement of surface temperature by one or other of the
contacting instruments commonly used for spot-checking has not to
date proved satisfactory due to the effects of accumulations of fibre
debris and scale on the cylinder surface and on the instrument
response surface itself (though see references 6 and 34 for designs
which are claimed to overcome this problem).
Each of these systems has serious inadequacies. In particular, each is
dependent on substance and sheet structure (as these affect the drying
characteristics) and on the machine speed and general drying conditions
(ventilation and felt conditions, condensing variables, etc.); also calibra-
tion is difficult and in most cases reliance has to be placed on an empirical
approach in which the measuring device is coupled to an appropriate
control (normally the main steam supply valve) and a working position
found which appears best to stabilize the moisture content at the reel-up.
Nevertheless, any of these systems may be expected to improve the situa-
tion on a machine with hitherto uncontrolled steam admission and so can
be justified as better than nothing, though whether the improvement is
better than can be achieved simply by controlling the steam supply
temperature as discussed in SC. 14 is open to doubt. As control systems they
generally have an advantage over reel-up moisture devices in enabling
earlier detection of changes, thus reducing the time-lag for correction. But
they cannot produce a record which can be taken to represent moisture
content at the place where it matters, the reel-up, nor (except possibly for
the second system) can they be adapted for measurement of moisture across
the web.
The second group, moisture-measuring instruments which are used on
the web close to the reel-up, includes the following systems:
(f) Measurement of the electrical resistance or conductivity across the
paper using d.c. or a.c. current and metal rollers or finger electrodes
contacting the web on both sides (Brown Moist-o-graph, Aquatel,
etc.); this is a well-tested system though the reading is affected to a
greater or lesser degree by the thickness of the sheet, the com-
pressibility (or, alternatively, the pressure of the electrodes on the
web), temperature, and the presence of static and electrolytes in the
web.
(g) Measurement of electrical resistance between electrodes on the same
side of the sheet (Shirley system); this has the advantage that it is
simpler to mount on the paper machine but still suffers from the
same disadvantages as (f).
(h) Measurement of electrical capacity or dielectric constant usually
between concentric or parallel-finger electrodes contacting one side
of the paper web (Lippke, Baldwin, Foxboro, etc.); this too is a
well-tried system which remains sensitive down to moisture values
of the order of 2 per cent., but it suffers from non-linearity of
520
THE DR YING SECTION AND CALENDERS SC.3 4
response to varying moisture content and is affected by substance,
to a lesser extent by the furnish and certain additives, and also in
some cases by the temperature of the head (but hardly at all by
other factors).
(i) Similar to (h) but with two non-contacting electrodes on either side
of the web (EKCO); this technique permits high moisture readings
to be measured and has potential applications at the press as well as
the reel-up.
(j) Measurement of air humidity close to the sheet, a property which
has been found to correspond closely to the moisture content
(Atrometer system); this has the advantage that direct contact with
the web is not needed, but careful shielding is required as the
reading is affected by web (and air) temperature and by draughts,
also the hygroscopi<;: elements are prone to dust and dirt, and
require frequent cleaning.
(k) Measurement of the absorption of microwave energy at specially-
selected wave-lengths (Beloit); this, too, has the advantage that the
instrument need not touch the sheet, but to date relatively little is
known of the performance and accuracy of this method.
(I) Measurement by absorption of ultrasonic sound waves, or by
absorption or reflection of infra-red and ultra-violet radiation (G. E.
Infrarnike and Pier); little is known of the potentials of these
techniques either.
Choice between these various systems of reel-up moisture measurement
is difficult. Many, especially (h), have been gradually improved over the
years in respect of their mechanical construction (particularly in regard to
automatic or rapid retraction out of the way at a break) and different
correction factors have been built-in to remove non-linearity of response,
the effect of substance, and so forth. It remains the case, unfortunately,
that whatever type of instrument is used a great deal of calibration work is
necessary on machines that run a variety of different grades of paper; a
routine standardization procedure is also essential, though this is not
usually difficult or time-consuming. In addition the detecting heads are
frequently subject to damage when they are in exposed positions between
the dryers and calenders or between the calenders and reel-up, and often
need regular cleaning to prevent dust accumulating and affecting the
reading. But with systematic attention from the instrument personnel and
training of the dryermen to use and appreciate the value of the moisture
reading, such instruments have an important role on the machine.
With a continuous measurement of moisture content available the
dryerman very soon becomes aware of two important phenomena. In the
first place he discovers that the variability of moisture in the machine
direction is much greater when the web is reeled damper: with the same
instrument response time the range of fluctuation can be as little as O· I per
cent. or O· 2 per cent. if moisture content is in the 5 per cent. region, rising
to O· 5 per cent. in the 7 per cent. moisture region, and with fluctuations of
up to 2 per cent. or more in the 9 per cent. region. Stability is greater when
521
steam pressure is controlled, and is closely dependent on the amplitude of

short-term substance fluctuations, but on all machines it is found that
moisture content at the reel-up is relatively much more unstable if the web
is run damper. This is one of the main difficulties involved in achieving as
high a moisture content as possible.
The other point which becomes immediately apparent with a continu-
ously-measuring moisture meter is that a difference in response is likely
according to whether the main steam pressure in the cylinders is increased
or decreased. If steam pressure is increased, under typical conditions, the
moisture content can take as long as 10 minutes to stabilize at the lower
level corresponding to the new steam valve setting; but if the steam pressure
is then reduced back, the response is generally much more rapid and the
moisture content stabilizes at its former value in a matter of 2 to 3 minutes.
One of the more important advantages of having a continuous moisture
measurement is that it becomes much simpler to achieve the desired
moisture level quickly and accurately after a wet-end break. When
cylinders are run for even a relatively short time without the web passing
over them, they rapidly heat up and the felts dry out; it is therefore normal
practice with a wet-end break of any duration to prevent excessive heating
up of both cylinders and felts either by reducing the steam pressure to the
main section and the felt dryers (while maintaining adequate pressure
differential to avoid waterlogging) or by switching over to separate steam
valves. When the web is fed through the drying section again the original
steam valve positions must be re-instated at an appropriate time. Ideally,
this should be chosen at the point when the time-lag in response to the
controller action is best balanced against depletion of the reserve of heat
in the cylinders and the re-gaining of moisture by the felts; this will ensure
that the web at the reel-up becomes neither too dry (when it tends to break)
nor too damp (when it sticks to the calenders). Determination of the best
means of dealing with a wet-end break and then feeding-up smoothly, i.e.
finding the steam pressure to which a decrease is desirable when the break
has occurred and the most suitable moment during feeding up when re-
version to the former pressure is arranged, is much simpler when a con-
tinuous moisture meter is available. In some cases it has been found
beneficial to increase the steam pressure temporarily to a value greater
than the normal running condition, thereby giving an initial 'kick' to
attain that condition quickly. Systematic application of a routine procedure
for altering steam conditions at the beginning of a break and when feeding-
up again eliminates a large source of downtime. As an aid to determining
these conditions knowledge of the behaviour of the surface temperature of
selected cylinders as the web is fed through would be of assistance;
unfortunately though this is not the best of times to be wielding pyro-
meters in the dryers. A simpler indication which helps to determine the
feed-up procedure that achieves drying equilibrium quickly is the behaviour
of the felt stretch gear; Roger and Webster (20) showed that moisture
content of dry felts could take 15 minutes or more to approach equilibrium
after a break, and as the moisture content is closely related to the running
length of the felt, observation of the stretch gear assists in setting the steam
522
pressure at a break and gives an indication if much change is still taking
place in drying conditions after the web is fed through again.
SC. 3 5 Automatic control of moisture content
From measurement of moisture content to automatic control is a logical
step, and it would be quite possible to link up any of the measuring devices
listed above with the main steam supply valve. In this way the valve
opening could be regulated to keep the moisture content (or, to be more
precise, the instrument reading) steady at some desired value. In practice it
is rare for such a direct form of control to be installed and it is generally
agreed that a better and more flexible arrangement results from using the
moisture meter reading to regulate the set-point of a separate steam
temperature or pressure controller in what is termed a cascade control
system.
The main advantage of having such an arrangement is that the response
characteristics of the steam controller can be chosen to deal best with
overall fluctuations in supply and demand of the dryers, while the moisture
meter response characteristics are chosen to keep movements of the steam
controller set-point damped sufficiently to take account of the long time-
lag and slow response of this part of the system. Such a combination gives
an overall better performance than a direct control set-up. A further
advantage is that failure of the moisture meter, or the need to shut off the
meter signal temporarily to allow a cross-web traverse or clean the
detector head, does not mean that control of the main steam valve must
become manual; instead, the moisture meter output can simply be locked
and the steam controller continues to function and maintain drying
pressure according to the original set-point. It is also easier to prevent a
sharp change to the steam control valve ('bumpless transfer') either when
the moisture meter is switched back into use or in the event of a change
being necessary to the moisture set-point.
With a cascade system of this type, in addition to recording the moisture
reading it is absolutely essential that a record is kept of the main valve
opening (usually on the same chart as the steam temperature or pressure).
It is relatively simple to arrange for a break-detector system to move the
moisture meter out of the way and lock the signal when the sheet breaks,
while at the same time (in the case of a wet-end break) the steam controller
set-point can be moved to a lower level to suit conditions with no paper in
the dryers, as discussed in the previous section. There would then be
manual or automatic reset to the former working position of the steam
controller as the sheet is fed through, with the moisture meter signal
reintroduced when the web has again reached the reel-up.
Those moisture-measuring systems which effectively average the whole
web appear at first sight to be most suited to this form of moisture control.
But in fact, it is preferable to have a detector head which measures only a
relatively narrow portion of the web in order that the position of the head
can be chosen to effect control over the dampest place where the sheet is
liable to cause the most trouble if moisture content becomes excessively
high. Also, most types of detector head suit a traversing arrangement and
523
thus allow some form of measurement of the cross-web variation (see

following sections).
Although ultimate control is normally thought of as applying to the
main steam supply valve, this regulates the whole drying section and the
large thermal capacity then means that response can at best only be slow.
It is preferable instead, whenever possible, to control the steam supply to a
small bank of four or six cylinders at the end of the drying section, thereby
giving a much quicker and more sensitive response. The danger from this
practice is that the maximum steam pressure available for the control
section may be frequently reached, or alternatively the steam supply may
be shut off to such a degree that the cylinders begin to sweat; for this
reason, some precaution is advisable with this form of control to ensure
that the main steam valve or the main pressure controller is altered should
the valve opening on the separate bank exceed pre-set upper and lower
limits. An alternative system which is claimed to improve response is to
alter the temperature or velocity of air in a high-velocity-air hood; there is
some difference of opinion about which hood operating characteristic is
the more suitable to alter, and the choice must depend on the relative ease
with which the control can be arranged and its effect on the overall
efficiency of the hood. Successful moisture control by changing the air flow
to hot-air blowing felt rolls has also been reported (124).
SC. 3 6 Unevenness of moisture content across the web
Of all the properties of paper, the one which is invariably the least uniform
across the width of the machine is moisture. This has many important
consequences. In the first place the more uneven the moisture, the lower
will be the average moisture for the whole web because the dampest
position must always be kept below the moisture content at which blacken-
ing occurs; further, greater variation from one moment to the next is found
at higher moisture values so the dampest position has in practice to be run
at a moisture which is well below the blackening point.
It becomes more difficult to build a uniform machine roll when moisture
is uneven because damp places tend to be thinner and slackly wound (a
direct consequence of the greater calendering effect on moister paper);
differences in reeling-hardness across the roll become accentuated on the
supercalender or winder, where the regions of uneven moisture are likely to
run into creases. To offset this effect the dryerman has recourse to the
calender air blowers, but frequently substance in the offending damp
positions has also to be reduced to dry the region down sufficiently; this is
all right provided high substance in the relevant positions across the web
were the original cause of the moisture being high, but if the substance
were correct in the first place then the slack area is removed only at the
expense of introducing some other lack of uniformity. The problems
arising from these difficulties are dealt with in more detail in SC. 4.
Most properties, in particular fold, stretch and tear, are sensitive to
moisture content, so that variation of moisture across the web will in turn
introduce variations in these other properties. Also, as regions with
initially higher moisture content pick up less moisture before reaching
524
equilibrium in ambient air conditions, this will introduce differences in the
behaviour of different parts of the sheet and be a source of, amongst other
things, curl and cockling. Curl can in addition be caused directly as a
result of the differential drying rate associated with damp regions in the
web. Finally, the effect of calenders is very sensitive to moisture content
so that areas passing damper through the calenders will, on average, be
smoother.
Cross-web moisture differences thus require very close attention. They
generally arise from two distinct sources. Firstly, there are differences
which must be regarded as more or less permanent and a direct character-
istic of the machine operation. These include general ventilation deficien-
cies, poor cambers on the press rolls, inadequate hot-air supply to cylinder
pockets, and so on; their commonest manifestation is in the form of dry
edges, and numerous ways have been thought of and tried to overcome this
defect. It is not proposed to discuss these in detail as mention was made of
the main approaches that have been attempted when discussing require-
ments of the condensate removal and ventilation systems. The greatest
difficulty in adapting the overall moisture profile is the labour involved in
achieving the right degree of compensation which is at the same time
suitable for all machine operating conditions; also a ready means of
determining moisture across the web is essential to assess the effect of the
changes made.
Even then it is by no means clear what alterations are likely to have the
effect desired. As an example of this, Harrison (60) mentions that in order
to try and improve the moisture profile on one machine amongst other
things the felt air-supply system was altered and enlarged, the Grewin
system was altered, and the front side of the dryer framing was closed in to
correspond with the enclosed gears at the back of the machine-each of
these changes brought little, if any, improvement in the general moisture
profile and it is disheartening to obs_erve how erratic the reel-up moisture
remained even when both the substance and moisture entering the cylinders
were relatively uniform.
The other sources of cross-web moisture unevenness are those which
gradually change in complexion over a matter of hours or days, and
originate from such things as wet or dry felt deficiencies, temporary
substance unevenness, scaling up of the inner or outer surfaces of drying
cylinders, hot-air ventilation nozzles which have become blocked or
knocked out of position, and so forth. These cannot be allowed for by any
permanent alterations to the machine and are often very difficult to trace
and remedy. As there is no means available to correct the moisture content
in the (often narrow) regions of the web affected, all that can be done is
to play with the substance and calender air blowers. It is in this sphere
particularly that moisture measurement tied with some means of correction
at different positions across the web has much to contribute.
SC. 3 7 Measuring moisture content across the web
Moisture content across the web at the reel~up can be measured by careful
sampling and oven-drying, but for regular use it becomes essential to use
525
5C.3 7 THE DR YING SECTION AND CALENDERS
one of the instruments described above. The main purpose is specifically

to compare moisture content at different positions, so absolute accuracy
of reading is not so important provided consistent results can be ob-
tained.
Any of the portable instruments can be used to obtain cross-web
measurements on a strip off the machine roll provided ample precautions
are taken to avoid air getting to the sheet before the reading is obtained.
This makes it necessary to tear off a minimum of a dozen or so thicknesses
of paper, select specimens for testing from the middle of this pile, then so
far as possible thrust the whole instrument within the pile of paper to
obtain the reading. To avoid systematic error, such as may occur if the
series of readings are always obtained in the same sequence from one
side to the other, it is preferable to randomize the order of testing
positions. The spacing at which tests are performed is generally about
6 in., though this may be widened if the procedure becomes too time-
consuming.
A different approach, which is not limited to end-of-roll testing, is to
use an instrument adapted so that it can be held against the machine roll by
hand. Provided reasonable consistency of reading is achieved and the
hand-held instrument is kept relatively free from operator error, such a
device could prove more flexible and much quicker to use than the normal
form of portable instrument. The main disadvantage is that it does not lend
itself to providing a permanent record of results obtained, and there would
be less likelihood of the dryerman being able to apply systematic correction
to moisture differences.
Adaptation of one or other of the various types of on-machine instru-
ments to traverse the web and give a continuous record of moisture at all
positions has obvious advantages, though unfortunately this is apt to
become rather expensive and involve cumbersome equipment stretching
across the machine. The general principles of using different arrangements
for traversing, in particular choice between different methods of presenting
the readings and co-ordination with the measurement of machine-direction
variation, are very similar to those described for traversing beta-ray gauges
(see lC. 34); in fact to a large degree the full value of a traversing moisture
meter is only realized if linked to a traversing beta-ray gauge (see for
example reference (56)). The system used must be chosen to suit the
behaviour of the paper machine in question, the relative variability of
moisture in machine and cross directions deciding, for example, whether
several measuring-heads are placed in fixed positions across the machine or
a single head is used, this being kept for most of the time in one position to
give machine-direction measurement but also traversing the web at fixed
intervals. When using a traversing head, the most usual and flexible
method, the instrument response time requires careful setting in relation to
the traversing speed (which for practical purposes is usually from 1 to 2
minutes) in order to produce a profile which shows up variations that are
possible to identify and correct; too damped a response does not show up
moisture streaks that are present, while too rapid a response produces such
a wildly fluctuating trace that interpretation becomes difficult.
526
SC . 3 8 Control of moisture content across the web
Installing any cross-web moisture-measuring system, however comprehen-
sive and accurate, is unfortunately only half the battle, for the dryerman
has few facilities to make any form of correction other than those afforded
by altering press loading or the substance profile. Indications of damp
streaks will be useful for pinpointing the position of trouble, and encourage
earlier investigation than is likely if the dryerman has to rely on feeling the
roll. But to obtain full benefit from an on-machine cross-web moisture-
measuring system it is necessary to provide some means of correcting the
moisture profile; in some cases this could even be made automatic, though
whether this is worthwhile is open to argument, and must depend to some
extent on how rapidly the cross-web moisture varies.
Many methods of moisture correction have been described, but as yet
none are really outstanding. Sprays, fixed in the early part of the dryers at
close (6 in.) intervals across the machine and giving an adjustable flow of
hot water on to the web, form the basis of one method for which some
success has been claimed (59, 112). It was found advantageous in this case
to make the controlling system automatic and a special unit was devised to
arrange successive alteration of a series of needle valves each of which
regulated flow to a single spray; to keep the total quantity of water added
at the sprays to a minimum it was necessary to ensure that at least at one
position the spray valve was left fully closed. An obvious criticism of this
system is that it is a retrograde step to add water to the web at any stage of
the dryers, but extra running costs involved in this must be compared to
the advantages gained by the control (it is quite possible anyway that
despite the water addition, less steam is used in the dryers because of the
higher average moisture which becomes possible). Also the capital cost of
the equipment is relatively low. Adjustable sprays on a dry felt rather than
the web provide an alternative method that has been suggested, but this
does not appear in practice to prove very effective. Steam jets are a further
possibility.
Another method of controlling cross-web moisture, which could also be
made automatic though this is not usual, is to use a sectionalized H.V.
hood; this hood is divided into 12 in. to 15 in. compartments across the
machine and the air flow to each compartment can be individually
regulated, normally by means of simple dampers. An installation using
this system has been described by Harrison (60) (this involved separate
measuring-heads opposite each compartment-a rather expensive business,
one would think) and also by Attwood and Hitchen (110). Use is also
made of high-velocity air at high humidity in an on-machine conditioning
device (121) which, it is claimed, not only adds moisture to the web but also
makes the profile more even. This, however, must be regarded not so much
as a control device as a means of improving moisture profile. Finally hot-
air blowing felt rolls in association with plastic fabric dry felts have been
used successfully for improving moisture profile (124, 130) though it is
doubtful if true control would be possible.
Other methods of moisture correction have been suggested, including
the use of banks of infra-red heaters or micro-wave dryers the output of
527
SC.4 THE DRYING SECTION AND CALENDERS
which can be varied across the machine. Another idea has been to apply
differential vacuum across the wire, but any effect this had would almost
certainly be all but eliminated in the presses. As yet, however, nothing has
appeared which offers a relatively cheap and simple means of altering the
moisture profile without incurring a very heavy capital cost. One basic
difficulty is to obtain a technique of correction which is sufficiently
adaptable to be able to overcome the quite narrow streaks of dampness
that are the main source of trouble on most machines; sectionalized
hoods, for example, cannot satisfactorily cope with really narrow streaks,
especially when these are severe, and any large alteration of heating
capacity necessary in one compartment has the effect of changing the
capacity of adjacent compartments. The correction should also have a
relatively quick response, i.e. thermal capacity must be reasonably small,
though it is not essential for correction to be made at the end rather than
the beginning of the dryers simply in order to reduce time-lag; probably
the best place is before serious differential contraction effects are intro-
duced, i.e. not later than two-thirds of the way down the drying section.
With any method permitting adjustment of moisture content across the
machine, it is imperative that a clear limit is imposed on the degree of
correction possible in any one position. The control allowed must, in other
words, be regarded as a final trimming device which corrects faults and
unevennesses that it is impossible to remove in any other way. The sub-
stance profile must be set first, then any general lack of uniformity in
moisture profile due to unsuitable press cambers, dry edges, and so forth
must be eliminated so far as possible by other means, and only then should
the moisture correction be used. If this is not done there is a serious danger
that the dryerman will come to depend on any moisture correction system
to remove all the faults that appear, instead of seeking first to eliminate
these at source, and this could well result in the paper lacking uniformity
across the machine even with a good moisture at the reel-up due to the
differential pressing and drying conditions experienced by different portions
of the web. For the same reason, it is essential not only to provide a record
of the moisture profile, but also one which indicates the degree of correc-
tion applied at each position with clearly marked limits showing the
maximum desirable-only with this facility will there be a deterrent to
placing too much reliance on any available control.
SC .4 CONTROL OF THICKNESS
The thickness or caliper of paper at the reel-up is controlled by operation
of the calenders, with, in some cases, assistance from a breaker stack. On
thicker grades of paper and paperboard the thickness is important in itself,
but on all grades, especially those required for printing, the close relation-
ship between surface smo.othness and thickness means that control of the
latter is no less essential. It is the dryerman's duty to supervise this control.
5C . 4 1 General comments on thickness control
Measurement and control of thickness in the machine direction presents
relatively few problems. The standard test method allows the average value
528
of a sample drawn from the top of a machine roll to be rapidly obtained.
Even though some increase in thickness will occur after extraction from the
reel, due partly to moisture pick-up and partly to relaxation of compression
forces (see Fig. 5. 13), this generally seems to be relatively slight and
within the accuracies of working can effectively be ignored. Changes in
thickness on the machine are of a long-term nature so that adequate control
should normally be achieved simply by using an end-of-reel process control
system and altering calender loading as necessary.
Control of thickness in the machine direction thus presents relatively
little difficulty. Unfortunately this cannot be said for control across the
width of the web. This is particularly important at the reel-up because
uniformity of thickness closely affects the evenness of reeling-hardness or
tightness of the roll; although the machine roll may appear to be winding
uniformly, if there are any soft or unduly hard places due to the web being
thin or thick in some positions (or, to be more precise, due to bulk being
lower or higher) these will almost certainly be accentuated in the winder and
will then cause creasing. In this respect the drum-wound reel-up is better
than older types, because the roll can be reeled at tensions closer to those
sustained on the winder and defects in uniformity which may become
troublesome are more readily apparent. The time-honoured method of
detecting hard and soft places is by hitting the roll with the hand or some
special implement and sensing how much it yields. Hard places occur
where thickness is higher, soft places where it is lower than adjoining
regions of the roll; and to correct these variations, the calender air blowers
and other methods of altering calender loading are brought into play.
Generally speaking, this traditional method of evening up reeling-hard-
ness across the reel works satisfactorily so long as the degree of correction
required to different parts of the roll is not great. In practice most machines
go through periods when this is not the case, and due to some fault either in
the calenders or further up the machine it becomes a struggle to reel at
uniform hardness despite clusters of air blowers full on and concentrated
on to one or two particularly bad soft places. This is frequently due in the
first place to unevenness of substance and moisture, bad heavy or damp
streaks making it impossible to build a uniform roll, but unfortunately
even when unevenness of thickness is obviously due to substance and
moisture inequalities, it is by no means simple to decide on an appropriate
corrective action. Added to these sources of variations in thickness are
those originating in the calenders, so it is not surprising that occasionally
the dryerman can lose control of the situation. This topic will be enlarged
on when dealing with the practical aspects of building a uniform reel in
SC.54.
The most systematic approach to this whole problem is to endeavour
first to get the substance profile as even as possible, then the moisture
profile, and finally to make corrections for uneven thickness across the
machine roll by adjusting the calender air blowers; with these three
properties uniform it is reasonable to expect that the roll will also be even
across the width in respect of finish, strength and other properties, at least
so far as it is practicable to make them even, and at the same time the roll
529
SC.41 THE DRYING SECTION AND CALENDERS
should build up without hard or soft spots so that this problem would also
be resolved. In practice there are limits to how far the substance and,
particularly, the moisture profile can be evened up so some discrepancies
in thickness profile originating from unevenness of these qualities will
require correction.
To assist the dryerman to make correction to the thickness profile
systematically, it is valuable to display some indication of the quantity of
air, and hence the degree of cooling, that is being used in each position
across the calender rolls. This also gives early warning if one particular
position begins to require too much correction.
Testing the reeling-hardness by hand has a rough and ready appearance
about it and an attempt has been made to devise a means of measuring this
property once the reel has been thrown out the machine (18). More recently
a concrete hardness tester has been adapted for use on machine rolls but as
yet there are no operational reports on its use. Though a measure of
reeling-hardness on finished reels could be useful as a tool for investiga-
tional work and possibly to improve the profile on a long-term basis, only
some means of indicating hardness variations while the roll is being built
up would be of genuine value to the dryerman. At least one device for this
purpose has been marketed but no reports on its accuracy or practical
value have yet appeared.
As an alternative to testing reeling-hardness it should be adequate in
most cases to consider paper thickness, since evening this up should
automatically counteract unevenness of reeling-hardness provided sub-
stance is also reasonably uniform. Thickness across the web is readily
measured by micrometer or on a sample strip taken off the machine web
with one of a number of thickness-profiling devices, but except on paper-
board this approach appears in practice to be of limited value for the
purposes of regulating evenness on the machine. One reason for this may
be that differences in thickness expansion (due to moisture pick-up) from
time of extraction from the reel to measurement complicate interpretation
of the results especially as the original thickness variations are closely
related to the moisture anyway.
A more satisfactory approach to this problem, and one which permits
control directly on the machine where it is of most value, would be to use a
device which gives a direct measure of the thickness of the web before
reeling at a large number of positions across the machine. Several methods
of doing this have been suggested and marketed, the most popular
depending on amplification of the gap between rotating discs or rolls
contacting both sides of the web; more recently the pressure from a series
of air jets positioned close to the sheet has been used, the jets either
opposing one another on opposite sides of the sheet or suspended across
the machine on a beam rigidly connected to a roll over which the sheet
passes. The latter system uses small manometers to measure back-pressure
of the air and thus lends itself to a form of presentation similar to one that
can be used for calender air-blower pressure, thereby permitting a rapid
means of assessing the effect that a measured change in calender air
pressure has on the paper thickness. Another device uses an electrical
530
measurement of magnetic permeability in the gap between two sections of
a magnetic reactor contacting opposite sides of the paper web; when the
cross-web changes in this measurement are used to vary air nozzles
directed on the paper (these giving more sensitive response than nozzles on
the calender rolls) it has been claimed that considerable improvement in
thickness profile is possible (114, 115). As yet, insufficient information is
available to indicate whether these devices are capable of withstanding
machine conditions. But when this is the case, it may be expected that a
valuable tool will be available which, by permitting thickness differences to
be readily compared with calender cooling across the machine, should
enable the whole business of building an even reel to be placed on a more
systematic and reliable basis.
SC . 4 2 Regulating the calender cambers
Substance and moisture unevenness across the web are apt to cause thick-
ness unevenness (with the related difficulties due to hard or soft spots)
largely in areas of relatively narrow width. It is also quite possible for the
calenders to be a source of relatively narrow unevenness, especially if
localized wear of calender rolls has occurred over a period due to some
relatively permanent difference in the paper web itself, but by and large,
lack of thickness uniformity directly attributable to the calenders exhibits
itself over larger regions. The most common unevenness from calenders is
roughly symmetrical about the centre, i.e. the result of incorrect camber
magnitude or shape, bearings heating up, incorrect loading, etc.; but also
it may happen that loading or roll wear is greater one side compared to the
other. Such deficiencies show up as a tendency, for example, for the edges
of machine rolls to reel hard compared to the middle (producing wrinkling
from the edges towards the middle), or for one side to reel harder than the
other, producing spasmodic creases running right across the sheet.
Attaining an even pressure in each nip of a calender stack is a many-sided
problem. The shape of the camber must make allowance for the journal
overhang, shear deflection of the roll, and the end-effects of temperature
variation; (using a straightforward circular cam to impart the shape on a
grinder is apt to exaggerate the camber between 10 per cent. and 20 per
cent. of the roll length in from both ends, causing soft spots in these
positions). The magnitude of the camber must be adequate for the weight
of the rolls and the load, if any, applied to the top of the stack. The nips
must be horizontal otherwise the paper will pull unevenly from one to the
next. These are the broad requirements, but to solve them creates many
headaches in the mill, especially when furnishes contain abrasive particles
(groundwood and some loadings) which cause calender rolls to wear
rapidly.
A general discussion of the questions of camber shape and magnitude
has already been given in 4C. 3 1. The same principles apply to calenders
and it is not proposed to develop these any further; for an excellent
detailed survey of the subject, reference should be made to an article by
Stone and Liebert (65). Regarding the necessity to achieve horizontal nips,
theoretically the position is quite clear: all necessary camber should go on
531
the bottom or king roll, and if the stack is loaded an appropriate camber is
also needed on the top roll. In practice it is often thought advisable,
especially for heavy and wide stacks, to distribute a small amount of the
camber (I or 2 thou.) on to the queen roll (second from bottom) where its
effect is of course double that of the same camber on the bottom roll.
Other intervening rolls should not generally be cambered.
Even following these precepts, after grinding a new stack of rolls to the
cambers normally found suitable it can happen that the fit is poor. Some
mills try to prevent this occurring by changing only one or two rolls at a
time, spreading the process for the whole stack over a number of weeks or
months; this has much to commend it except that barring patterns and
other deficiencies in the old rolls are liable to transfer to the new. For this
reason, when roll-changing is staggered it is important for the period
involved to be quite short (say no more than four weeks); it is also appro-
priate that the changing procedure is repeated at frequent intervals on a
planned-maintenance basis. Checking the caliper of the larger rolls of the
stack whenever possible, and the compilation of wear curves, as discussed
in regard to press rolls, can assist in determining the correct camber and
the desirable frequency of changing.
When changing the whole stack at once, or when one of the cambered
rolls is changed, either a light test or one of the other means of checking fit
beforehand, such as the B.P.B.I.R.A. technique of passing through a full-
width strip of N.C.R. paper, is essential to show up any major discrepancies
before the stack is put into use. To be of much use, this check must be
carried out with the stack close to normal running temperature, and under
these conditions it is difficult to determine quantitatively the correction
required to eliminate those discrepancies that are observed. It may be
necessary in bad cases to remove one or more rolls for re-grinding off the
machine, or it is possible to grind the rolls into a better fit on the machine
by dripping down the stack an appropriate abrasive grinding compound.
Grinding-in on the machine does not, however, find general favour because
every roll then sustains relatively high wear; it is only suitable as a
relatively quick means of making minor corrections.
SC . 4 3 Temperature variation across calender rolls
The running temperature of calender rolls can vary considerably over quite
a small distance, and this of course means that local nip pressure can also
vary by a large order. Using a surface pyrometer, Howe and Lambert (61)
found differences of up to 50 deg. F. at positions one foot apart, and
succeeded in demonstrating (by assuming a simple relationship between
temperature and roll expansion) that the total effect down a stack of such
temperature differences is closely related to the decrease in thickness and
increase in smoothness occurring in the paper at different positions across
the machine. In fact, on the machine investigated it appeared that thickness
variation across the web at the machine roll largely originated in calender
temperature differences, only a relatively small proportion of the variation
being attributable to local differences in the paper before entering the stack.
Large temperature differences in calender rolls are probably created
532
mainly by the self-exaggerating effect of what may initially be only
relatively small differences in pressure. This is because a slightly greater nip
pressure at some point will cause more heat to be generated, and this in
turn will expand the rolls and increase the pressure there still further;
equilibrium is only achieved at an even greater nip pressure which causes
more thickness reduction in the paper.
It should be the main purpose of the air blowers to correct these tempera-
ture differences. Normally, as discussed above, this correction is attempted
using subjectively-assessed reeling-hardness variations at the machine roll
as the guide, this being the most direct approach to the problem. Though a
more systematic method would be to use on-machine thickness profiles to
indicate where air is needed, the difficulties of obtaining satisfactory thick-
ness measurements of this nature have so far limited developments in this
direction.
An alternative and possibly more practicable approach has been
described by Sawyer (108) who measured calender roll temperatures
directly, using a non-contacting radiant-heat device known as a radio-
meter; this is an instrument which has an obvious value for trouble-
shooting when detection of comparative temperatures is required, though
it is probably of doubtful suitability for obtaining accurate absolute
readings. Regular traverses across the rolls with this instrument showed
that patterns change appreciably, even from day to day, and temperature
differences could be great, especially on rolls high in the stack. Tempera-
ture of the web was also measured by this technique and variations wider
even than the calender rolls were found; this may be due to differences in
moisture which will affect the relative transfer of heat. Developing this
work further, Sawyer arranged an automatic correction system whereby
the temperature profile of a roll as measured by the radiometer was used to
alter the volume of cooling air applied by the blowers. This helped to
reduce reeling faults in the paper and was considered to be a very useful
approach. The main limitation lies in the fact that the temperature of only
one or possibly two rolls can be used as the basis for correction; also some
thickness variation in the finished paper is definitely due to causes other
than calender roll temperature differences, even if the work of Howe and
Lambert indicates that the contribution is relatively small, and this system
will not correct this. One likely approach is to use the method devised by
Sawyer to apply correction for calender roll temperature differences with
one set of air blowers, leaving alleviation of defects in the paper from
uneven moisture and substance to a final correction of the thickness or
reeling-hardness with a second set of air blowers.
SC. 4 4 Altering calender stack load
Even with the best set of calenders, the need to change load or the number
of nips in order to achieve different degrees of smoothing makes it im-
possible to avoid camber difficulties at one time or another. Dryermen have
devised ingenious ways to correct camber deficiencies, from sticking pieces
of emery paper or even machine wire under calender doctors to the use of
wedges and jacks to force apart adjacent roll journals. Most of these
18 533
5C.44 THE DR YING SECTION AND CALENDERS
methods are self-defeating in that they create extra wear on the rolls in
precisely the regions where it is not wanted (for example, heating up an
area of the roll by inserting an abrasive material under the doctor will
temporarily narrow the gap in the nip and hence alleviate a hard spot, but
it will also cause the roll to wear faster at that point, thus tending to
increase the gap, and eventually even greater local heating will be required
to avoid the hard place reappearing). For this reason such expedients
should be avoided completely or used only on a strictly temporary basis.
It has been realized for some time that what is needed is a means either of
altering effective camber during running or avoiding altogether the
necessity for any camber at all. The various methods devised to achieve
this end will now be briefly discussed.
If a roll having a particular camber is rotated very slightly horizontally
relative to the roll above, this has the result of decreasing the effectiveness
of the camber. This can readily be visualized by thinking of two parallel
rolls in line-contact along the nip; if one of the rolls is crossed horizontally
about the centre point of the nip then contact remains only at that point
and elsewhere a gap appears between the two rolls which is progressively
greater from the centre out to the edges. To counteract this effect and
regain line-contact, a degree of negative camber on one or other roll would
be needed. This principle has been used for varying the effective camber on
an operating stack (93). The method involves offsetting each queen roll
journal in relation to the vertical line formed by the other roll journals,
positioning being made precise through a specially-designed electro-
magnetic unit. On a relatively large stack, a movement of one inch can give
an effective change of 25 thou. in the camber on the bottom roll. When the
stack is loaded, compensation also becomes necessary for the top roll and a
similar means of movement of the roll journals is needed there.
It is claimed that this technique of crossing rolls is simple and effective,
requiring only ordinary standard rolls and anti-friction bearings. More
camber than is likely to be necessary must be placed on the bottom and top
rolls initially because it is only possible to relieve the effect of camber, not
to add to it. It is also possible by moving one journal more than the other
to counteract to some extent any tendency for one side of the sheet to be
calendered more than the other. One deficiency of the technique lies in the
fact that the shape of the camber is effectively changed when rolls are
crossed and a profile which initially gives a perfect fit would alter to an
elliptical curve that tends to put too much camber nearer the ends com-
pared to the middle of the roll. However, this effect may well not be notice-
able in the normal course of operation provided alteration to the amount
of crossing of the rolls is kept reasonably small.
A second method of changing effective camber has been reported by
Kettering (105) and involves the use of air diaphragms to apply pressure
between the journals of adjacent rolls. This represents in effect a systematic
application of the favourite practice of wedging jacks between roll journals
to relieve the ends and increase the effective camber. A certain proportion
of any force applied between journals will offset the weight of journals and
bearings and improve line-contact when the normal shape of camber
534
THE DRYING SECTION AND CALENDERS 5C.5
imposed by a grinder is used, but the main danger in this technique (as
with placing jacks between journals) is that it would certainly be used to
correct other deficiencies; over a period this could well create heavier wear
in the middle, making it necessary to increase steadily the amount of
pressure or compensation applied.
Two other methods for camber compensation have the advantage of
eliminating the need for putting any camber on the rolls; they can
probably accommodate nip-pressure operating changes more easily and
automatically, and the absence of camber reduces the necessity for frequent
re-grinding and eliminates any effect resulting from micro-slip at the centre
relative to the edges (the velocity of the outside of a cambered roll being of
necessity greater than at the edges). The first of these is the Kusters
'swimming' roll which employs a stationary centre shaft inside a rotating
shell; oil is used in between these two surfaces under sufficient pressure to
keep the rotating shell floating about the centre shaft, and whatever the
load from above there is automatic compensation at each point of the
shell. The 'swimming' roll is of a relatively complicated and costly con-
struction and when used in the normal position at the bottom of a stack it
is necessary to arrange the drive on to the queen roll. However, several of
these rolls are now operational and a fairly detailed report on their use has
been given by Hillman (102).
The other method of avoiding camber ('Accra-nip') uses the principle of
roll bending. In this the bottom journals are built out at both ends and, in
effect, cantilevered by the application of downward forces outside the
bearing supports. This has the effect of bending the roll upward in the
middle, rather as a ruler bends upward if it is held at each edge with the
thumb on the inside and then pulled down with the fingers; by this means
the natural deflection that the roll would take up under its own weight and
the load of the rolls above is compensated. With a loaded stack the same
principle is applied to the top roll, the force imposed at the journal ends
being upward in this case. Stone and Liebert (65) have described how this
system functions and have illustrated the precision that is possible by using
hydraulic pressure in cylinders to apply the forces on the journal ends.
Furthermore, by suitably sizing and positioning the cylinders, it is possible
to link the top and bottom rolls in such a way that when a change in
running nip-pressure is required this is readily achieved simply by altering
the common hydraulic pressure.
SC.5 PRACTICAL POINTS

SC. 5 1 Start-up
Drying and calendering conditions at the time the sheet is first fed through
must be as close as possible to normal running, otherwise there will be
endless delays and frustration as breaks recur and broke piles up and
impedes access to the machine. The large thermal capacity of the drying
cylinders makes it necessary to heat up gradually from ambient tempera-
ture over a period of hours, so steam is let into the dryers (including those
for felts) as soon as it becomes available; especially if the system is not
535
automatic the temperature of the dryers and the flash system should
obviously be inspected at intervals to ensure that it is rising steadily. Air
vents on the condensate lines or flash tanks are kept open for as long as
seems necessary, but to be assured of complete removal of non-condensible
gases from the cylinders it is desirable to set each section crawling round as
soon as the necessary power can be used; this also prevents the felts
getting hot unevenly. When the facility is available, steam is let into the
calenders from quite early on, but in any case half-an-hour or so before the
machine is ready (earlier if a new bottom or queen roll has been installed)
the calenders should be set in motion so that temperature of the rolls
reaches near normal. The calenders are usually started at full-speed, and
this is best done whenever possible with the rolls lowered rather than by
skidding upper rolls on to a stack already running which is liable to cause
damage; doctors are cleaned and brought into contact straightaway when
the stack is up to speed. The ventilation system and H.V. hoods are also
started soon enough to ensure air is available at the normal temperature.
Once the cylinders have been set crawling round prior to feeding
through, it is essential for the dryerman to devote some attention to several
other parts of the dryers. All doctors must be placed in contact with the
dryers and M.G. cylinder, and set appropriately for load, oscillation, etc.
Cylinder surfaces, especially M.G., require examination. Grewin nozzles
and other air blowers should be felt to ensure there is no blockage, and the
monitored temperature and pressure conditions checked to be normal.
Felts and Sheehan ropes must be inspected to see that running is steady,
the guides working, seams and splices not pulling apart, and the tension
gear functioning and set correctly.
When the cylinders are eventually speeded up in preparation to receive
the sheet, felt guides and the stretch gear on ropes and felts may need
readjusting. The power taken by each section should be checked as an
indication of poor bearing lubrication, excessive doctor pressure, or
waterlogging; the draws are pre-set to conditions anticipated. Other
equipment demanding supervision, such as a dry-end pulper, water
damper, sweat roll, M.G. pressure roll, reel-up drum and so forth should
form part of a check list to make sure the dryerman gives them the neces-
sary attention. Immediately prior to feeding through the sheet a last-
minute inspection of the steam and ventilation system is worthwhile and
each cylinder should be felt for signs of waterlogging. The compressor
providing air for blowing the tail across the draws should be up to pressure
and each blower quickly checked over, and the calender air-blowing fan
started. The lubrication system and cooling of bearings, cylinders and
calenders requires inspection, though this is usually the job of an engineer.
When the sheet is ready to be fed through the dryers, at an appropriate
moment steam pressure in the cylinders is increased to normal operating
level (or slightly above, as drainage rate and pressing efficiency is likely to
be lower than normal at a start-up, giving a higher moisture content
leaving the presses). The importance of getting suitable conditions in the
dryers to ensure the web arrives at the end of the section with as near as
possible correct moisture content has already been stressed; if sufficient
536
THE DRYING SECTION AND CALENDERS 5C.5 1
attention is not paid· to this, feeding up can become a nightmare with the
dryerman altering the steam pressure violently to counteract the web
becoming too damp or dry, then losing control of conditions altogether
when the sheet breaks and the process of feeding up must be repeated. A
4 to 6 inch tail may be passed by hand from the last press on to the first
cylinder, where it can temporarily pile up on the doctor, and then on
through the cylinders; on machines running above 500 or 600 feet per
minute a Sheehan rope system is in general use and the tail is blown off the
press straight into the ropes, which then carry it through to the end of the
dryers. The tail is always followed down the machine by the dryerman and
when he sees it is passing smoothly along, the sign is given for the cutter to
be brought across to widen the sheet to full-width; should the sheet break,
the dryerman is on the spot and immediately signals for the web to be
broken off at the press or wire before he begins clearing away the broke.
With an M.G. cylinder it is usual to feed over the whole sheet at once. The
ventilation system may have to be shut off or air velocities reduced
temporarily while the tail is fed through, otherwise it tends to flap un-
controllably and break.
The sheet is doctored off the last cylinder, which is invariably a top one,
and to prevent the possibility of the web adhering to the last dry felt and
being carried over the top, a series of air nozzles may be installed close to
where the felt and dryer separate to blow the tail section down on to the
cylinder doctor. After allowing a short period for moisture content of the
web to approach its normal value (and in this respect a dry-end pulper
under the calenders is particularly beneficial) a new tail is cut, by means of
a knife, on the top cylinder and passed or blown under the dancing bar or
roll and between guide plates into the top nip of the calenders; on slower,
narrow-deckle machines it is possible simply to grab the whole web off the
last dryer and either cram it straight into the calender nip or, as is done in
some cases, lead it over the top roll and into the nip from the opposite side.
The sheet is fed down the calenders by hand, being snapped double into
each nip by an acquired flick or simply with the fist held against the roll as
far as the nip guard allows, or with a tail the process may be made auto-
matic by means of steel fingers or air blowers on each roll; frequently,
immediately prior to threading through the calenders, the rolls are sprayed
with paraffin oil or some other lubricant which cleans them and softens any
paper that tends to stick or be stamped out. Feeding down the calenders
can be a critical operation and on no account must a wad of paper be
allowed to jam in the stack causing the rolls to jump, or worse to stop
altogether in which case scorching and the appearance of fiats are
likely.
The tail will normally be carried or blown straight across from the
calenders to the reel-up although on fast machines ropes can be used; with
a drum reel-up the core is brought into contact with the drum as little
before this as possible to prevent roughening up the reel-up drum surface,
especially when the roll is damped by sprays on the drums. Alternatively
the tail can be widened (by hand or automatically) by taking across the
knife on the last dryer so that the full sheet is allowed temporarily to come
537
off the bottom cylinder roll; then a fresh tail, cut by hand on the bottom
calendar roll with a sharp-pointed spike, can be led or blown across.
Once the sheet is reeling satisfactorily the first thing the dryerman must
check is the moisture content; unless this is near the correct value, breaks
are likely. He must assess how evenly the roll is building up, load the
calenders by the usual amount, and make preliminary adjustments to the
calender air blowers which should have initially all been off or only part
open. As soon as possible the first roll is thrown out to allow cross-width
checks on the quality, particularly the substance, to be made. When the
machineman has got the substance uniform, and drying and ventilation
conditions have settled down, further adjustments to the air blowers will,
of course, be necessary. This topic is enlarged upon in SC. 5 4.
The dryerman must also at the first opportunity examine the progress of
the web down the dryers. If the web is tending to stick or pick on the first
few cylinders, the temperature must be reduced; build-up of dust and fibre
on cylinder doctors indicates if picking is occurring and also shows how
evenly doctors are in contact with the cylinders. Any tendency to flap in the
open draws, especially at the edges, should be remedied before creasing or
turned-over edges occurs; correct alignment of cylinders and felt rolls is
important in this respect, and hot-air nozzles require careful adjustment.
Draws are usually altered visually to give correct tension: if slack the sheet
tends to wrinkle and fold over, if tight to crease lengthways. Spreader bars
or expander rolls should spread out the sheet uniformly and smooth out
any unevenness of pull. The steam and flash system should be inspected as
soon as practicable, attention being paid particularly to the ·pressure
differential across each bank of cylinders. Felts also require inspection
and movement of the guide roll and stretch gear should be positive but
slight.
SC. 5 2 Shut-down
Procedure for shutting down the drying section and calenders for a
scheduled shut is relatively straightforward. The sections are brought to a
stop as soon as the sheet has run out and the ventilation system, H.V. hood,
and steam supply to dryers, felt dryers, and calenders turned off. Loading
on M.G. pressure roll, breaker stack, and calenders is relieved and for a
lengthy shut each calender roll, particularly the bottom and queen rolls,
should be supported separately or at most in pairs on the journals to avoid
the possibility of fiats appearing; the rolls may also be lubricated with
paraffin. Felt tension is slackened, doctors removed from contact (especial-
ly on calender rolls), hoods raised, and other parts of the system, con-
densate pump, water in spray damper or sweat cylinder, the calender air
cooling system, dry-end pulper, and so on are attended to as necessary.
An M.G. cylinder is usually kept rotating slowly during a long shut to
avoid stressing the dryer.
Felts and ropes are now inspected for flaws, for which it is necessary to
crawl the cylinders round. At the same time the state of the drying cylinders
is assessed and it may be necessary to remove scale and fuzz from several
538
cylinders, particularly at the wet-end, by using suitable scraper-knives.
One or more cylinders may be opened up for examination of siphons and
the general condition inside. Calender rolls are examined carefully for marks
and roughness, or signs of barring on the surface; calenders usually generate
a considerable amount of dust so cleaning, preferably by vacuum cleaner,
requires careful attention. All pieces of broke must be cleared away from
the dryers either by hand, or by hooking or (when gears are completely
enclosed) by blowing out with compressed air; doctors and associated
trays on top cylinders are cleaned out with a long-handled brush or
vacuum cleaner.
During a wet-end break of any length, a similar procedure is necessary
to when the machine is being shut. The drying cylinders may be kept
rotating, but calenders should definitely not be allowed to run without
paper for too long because of the heat generated by the contact that takes
place between rolls. Reduction of the steam supply for a break has already
been discussed. An H. V. hood is not usually turned completely off and
allowed to cool, but normally the heat supply is greatly reduced and the
fans allowed to idle; in a similar way it may be arranged that the ventilation
system is kept warm, though a Grewin nozzle system is often turned off to
ease feeding up.
SC . 5 3 Changing and running dry felts

There are still several machines using endless dry felts and for these the
method of changing is basically similar to wet felts, but can be very
cumbersome and time-consuming. Nowadays dry felts are seamed and the
procedure for changing them involves slackening back the tension roll,
cutting the old felt across at an appropriate position, and roughly stitching
an end of the new felt on to the old. The cylinders are then rotated slowly
round, the new felt being drawn out of its folds and the old one collected
up, and finally the old felt is detached and the ends of the new one pulled
square and seamed together. In the event of the old dry felt falling off the
machine it is necessary to thread ropes along the felt run and use them to
pull the new felt on. Boards and other equipment needed for dry felt
changing should, of course, be kept clean and always stored in the same
place.
Methods of seaming vary according to the felt material and running
conditions. On old, slow machines a linen thread stitch can suffice, while on
faster machines split copper rivets may be used in several staggered rows
(this is apparently particularly suitable with asbestos felts where seams pull
out easier than with cotton felts). Other seaming methods involve glueing
or cementing the two ends (popular with synthetic felts) and the specially-
designed clipper seam which joins together alternate metal loops placed on
each end. With all except the last method which leaves a small gap between
the two ends of the felt, a lap is formed of one end over the other in such a
way that the trailing end is on the outside. The choice of seam and the skill
with which it is put together are very important, as a large proportion of
downtime originating from dry felts is caused by seams pulling apart to
539
make a large hole, or completely breaking in two. It is for this reason that
careful and frequent inspection of the seam is essential.
A new felt must be started with care, and is first slowly crawled round to
check that it is pulling evenly and the seam is holding. The tension is then
steadily increased, but before the section is brought up to speed the guide
and tension gear are inspected, the felt edges watched for signs of un-
ravelling or soiling, and a check is made that all felt rolls are turning al-
right. When full speed is reached the guide is again checked, and when of
the automatic variety it can be brought to a central working position by
means of a hand guide roll, following the well-known principle that the felt
moves towards the side of the roll it touches first. Alternatively the stretch
roll is used to adjust the position of the felt, in which case the felt tends to
move to the side that is slackened off (this being the side the felt first
contacts); however, this will also cause the seam to pull ahead on the slack
side due to shortening of the run. Ideally, the seam should keep perfectly
straight across the machine due to the rigidity of the felt, but if it shows
signs of leading at one edge this is an indication that a roll is out of align-
ment. For a more detailed discussion of this topic reference should be made
to an article by Woodside and Macmillan (119).
Once paper is drying, the stretch roll and tension gear must be carefully
watched. There should always be a steady, positive motion occurring if the
automatic tension is working correctly; otherwise when the gear is sluggish,
excessive tension can easily build up, threatening the seam and causing the
felt to become baggy in places. Shrinkage of a dry felt as it becomes damp
is common and under these conditions, especially with a new felt, the
stretch roll position is inspected regularly for some hours; occasionally a
felt shrinks excessively causing the edges to overlap the roll ends and get
oily, or it may have to be removed as the stretch roll reaches the end of its
travel and tension becomes so excessive that roll bearings get overheated.
Other dry felts stretch when new and if this becomes excessive it may be
necessary to stop the machine, cut across the seam, and re-seam with a
piece taken out. In very bad cases the felt may have to be removed al-
together because its deckle becomes too narrow to cover the paper. When a
new dry felt is joined up, experience of the behaviour of similar types
allows the position of the stretch roll to be appropriately selected when the
ends are seamed.
Running of dry felts usually presents little difficulty unless the felt is
accidentally damaged, or if the web passes under a cylinder doctor during a
break and finishes up in a thick wad which permanently bulges the felt. At
best this is liable to cause paper coming under the bulge to be held off the
cylinder by a steam pocket, creating a damp, stretched patch; but if the
bulge is really bad, as it passes over the stretch roll, the felt will be liable to
run into two short diagonal creases. Wrinkles in dry felts, especially newer
ones of the plastic fabric types, are also caused by felt rolls being out of
alignment (to check this involves taping round cylinders and rolls at both
ends and also measuring diagonally) or by the stretch roll being tighter in
movement or binding at one or both sides; a temporary remedy is to
slacken back the side to which a diagonal wrinkle runs. Persistent moisture
540
THE DR YING SECTION AND CALENDERS 5C.5 4
differences across the sheet can affect a dry felt by causing it to be damper,
and of different length, in one position across the machine compared to
another.
SC. 5 4 Building a good roll
A good measure of the skill of a dryerman in his job is the quality of the
rolls he produces from the machine. The evenness of reeling-hardness
across the web, absence of obvious shade differences on the roll which
betoken lack of uniformity in moisture and calendering, absence of hair
creasing or cuts, and evenness of the edge of the roll, all these are to the
dryerman what even substance, good formation, and (in general) lack of
two-sidedness are to the machineman: a general comparative test of the
control and understanding he has of the machine.
The business of building a good reel has been discussed in several places
already. It depends basically on maintaining a uniformity across the web in
respect of substance, moisture, and thickness. If, in addition, an even
tension is held in the draws, alightment of rolls is correct, and the reeling
pressure is kept steady, then a really satisfactory roll is certain to be
produced.
The job of obtaining an even substance across the web is the machine-
man's, but all too often he relies on the dryerman to indicate to him that
the weight at some place across the reel appears wrong. This comes about,
for example, when as much calender air as possible has been directed on to
some soft spot and still fails to remove it; the dryerman may then ask the
machineman to tickle up the substance in that region to try to correct this.
Unfortunately this does not always prove to be the remedy and the soft
spot can be worsened. It all depends on the origination of the low thickness
causing the soft spot, and this can be the substance, moisture, calenders,
or any combination of the three.
To illustrate the complexity of the situation, consider the possible sources
of a soft spot which appears at some place across the machine roll. This
may, in the first place, be due simply to the calenders being too hot in that
region (camber too great, overheated bearings at the edges, poor doctor
application, etc.). The correct remedy in this case is to use air from the
calender blowers to cool the appropriate position down. On the other hand,
the soft spot may originate from the paper being too damp at that particu-
lar position across the web when entering the calenders (uneven press
action, worn dry felt, cylinders scaled up, etc.); the calenders then have
more compressional effect on the damper region, thus producing the soft
spot. To remedy this correctly would necessitate a means of correcting
cross-web moisture in the drying section. Thirdly, the substance may be
light due to a ridge in the wire or bad adjustment of the slice; this in itself
could produce a tendency towards low thickness and a soft spot, though
low substance would be associated with comparative dryness at the reel-up
and this should counteract the effect to a greater or lesser extent. The
remedy for uneven substance obviously lies at the slice. Any combination
of these three basic sources could produce the soft spot, so more than one
correction may well be needed.
541
It should be observed in this context that an associated complication is

created by uneven shrinkage of the paper at different points across the web.
If one position should dry quicker than the rest of the web, this causes it to
stretch due to the greater restraining force it experiences and has the effect
of producing softness in the machine roll; this is irrespective of the fact
that the region may well retain lower moisture content through to the
calenders which should partly offset the softness. This situation is often
found at the edges (particularly the front edge), where on machines running
well-beaten grades, slackness can be very troublesome and in extreme cases
causes a sort of wavy fluting on the roll edge; it originates from excessive
stretching due to quicker drying at the edges for which, as discussed
earlier, no really satisfactory remedy yet exists. Slack edges as such can,
however, be overcome using spreader bars or expander rolls to help
stretch the middle up to that sustained at the edges, but often the substance
is run permanently heavy at the edges to offset the more rapid drying.
While the various possible causes of a soft spot are thus well-known,
identifying the source in any particular case is practically impossible
without instrumental aids. Generally speaking the dryerman relies simply
on the adjustment at the calenders to even up the reeling-hardness, and
when this fails he turns for assistance to the machineman. As already
stated, the most systematic procedure is first to adjust the substance as
evenly as possible across the web, then the moisture, and finally use the
calender air blowers to even up reeling-hardness. Only if the air blowers
fail to correct sufficiently under these conditions, can temporary adjust-
ment of the substance (necessarily reducing uniformity of this property) be
justified to keep the machine running; if this procedure is followed, it is
evident that the calenders are more readily identified as the source of the
trouble and the fault can be remedied at the first opportunity. To achieve a
method of working along these lines requires accurate information about
the cross-machine substance and moisture profiles and, above all, a means
of correcting moisture differences across the web at least as adequate as is
available on most machines for correcting substance. The state of develop-
ment of each of these desirable features has already been discussed in detail
and it is evident that there is a long way to go. In the meantime the dryer-
man must work as best he can with the adjustments available to him, plus
his knowledge of the idiosyncrasies of the machine.
Differences in reeling-hardness due to uneven nip pressure at the calen-
ders are usually not so local and random in nature as those originating
from lack of uniformity of substance and moisture. Because of this an
isolated soft or hard spot is most likely due to lack of uniformity in moisture
and substance, and comparison with obvious differences in dampness of
the roll can provide some clue as to the origin of the unevenness (it is, of
course, not difficult to detect fairly damp regions across the roll by feeling
the web with the palm, or by viewing the roll at an angle when damp streaks
appear darker). If a soft spot coincides with a damp region, there is a
strong likelihood that the cause of softness is in the drying or pressing. On
the other hand, if the soft spot does not appear any different in moisture
content from the rest of the web, it is more likely that low substance in that
542
THE DRYING SECTION AND CALENDERS 5C.5 5
region is the cause. A similar tendency shows up with hard spots: if
moisture appears normal, the substance in that region is probably heavy.
However, there is a complication in that, as the dry-weight of a straight-
forward heavy streak grows, a point seems to be reached where the web
becomes so much damper that reduction in thickness at the calenders, due
to the added moisture, more than offsets any increase in thickness resulting
from the additional fibre; when this happens the substance unevenness is
shown up by a soft spot, not a hard one, and the interpretation likely to be
placed on the coincidence of softness and dampness (that drying is at fault)
becomes incorrect.
The close association that substance and reel-up dampness have on
reeling-hardness makes it imperative that the machineman and dryerman
co-operate closely in their efforts to produce good rolls and correct cross-
web faults at the reel-up. The machineman always indicates when any
changes of speed, substance, furnish and so on affecting the overall machine
conditions are due, but keeping the dryerman in touch at times when it is
necessary to make any important alterations which may affect the cross-
web uniformity, is equally important. If, for instance, the machineman
alters one or more of his jackscrews to change the substance at some place
across the machine according perhaps to the appearance of the dry-line or
couch take-off line, when the alteration has been completed he should warn
the dryerman to keep an eye on the appropriate position. The same applies
to other running adjustments likely to alter the cross-web evenness:
depending on how good the machine system is, this could cover anything
from changing press load to altering head at the slice. Experience is the
only guide in these matters. Neither should information be one-way, for
the machineman generally has to rely, in his efforts to achieve uniform
substance across the roll, on test results which do not take account of
moisture difference; hence the dryerman should always indicate the exist-
ence of damp regions, and inform the machineman if he has altered drying
conditions to an extent which could have an effect on the moisture content
at different places across the web.
SC. 5 5 Checking the dryers and calenders during running
Apart from inspecting the dry felts and ropes, and endeavouring to build a
satisfactory machine roll, the dryerman must constantly be checking over
many other parts of the drying section and calenders. An engineer is
normally assigned specifically to inspect the steam supply, flash system and
condensate tanks, the individual drives, the ventilation and hood system,
and the bearing lubrication and cooling, but it is nonetheless the dryer-
man's ultimate responsibility to ensure these are behaving correctly and
catering adequately for the needs of the sections.
On a tour of inspection the dryerman will first examine round the
calenders, feeling the journals and air nozzles, and looking at the doctors
to see that the fit is even, then wiping them across with a rag. As he walks
alongside the dryers, each cylinder should be touched as a rough check
against waterlogging; the surfaces are inspected and if rings are forming,
the doctors are moved laterally or weighted. Grewin nozzles are tested in
543
5C.5 5 THE DRYING SECTION AND CALENDERS
case of blockage. The surface of an M.G. cylinder and the various doctors
are given particular attention, and evenness of the take-off line from the
cylinder surface often shows up irregularities. Particular note is taken of
the line of the sheet in the open draws between cylinders: drive irregulari-
ties, dryer eccentricities (which should be observable by eye or by a regular
change in the tone as the dryer rotates), and misalignment of rolls due to
gradual wear of the bearings, are some causes of an uneven pull which
shows up a regular flapping of the whole sheet; but more commonly,
flapping occurs only at one or other edge and is a product of uneven
drying at the edge or poorly directed air blowing from nozzles into the
pockets. Dispersion of steam in the cylinder pockets should be steady and
reasonably even across the dryers. Towards the wet-end, the cylinders are
looked over for any signs of being too cool and sweating, or too hot, in
which case a rapid build-up of fibre dust is observed in the top doctor trays
as the surface of the sheet is picked.
Returning down the back of the machine, the dryerman inspects the
load on each section and the draw positions, and notes any fluctuation
indicating that running is not so smooth as it should be. The flash and
condensate system should be functioning satisfactorily, with liquid-level
gauges between marked limits and registered pressures within the usual
range; fluctuations in the flash system resulting from surging of pumps or
build-up of condensate in the cylinders should be noticeable from the
various indicators, especially when the drive is being affected. Ventilation
instruments are checked, and any steam leaks from dryer journals noted.
Breaks in the drying section are caused mainly by the action of excessive
tension on weak places in the web. Excessive tension occurs when the draw
is uneven or if shrinkage forces within an individually-driven section
become too great; machine-direction strain lines can generally be seen
when tension is too high in any particular draw. Weak places in the sheet
originate from wet-end defects (pitch or slime spots, rough edges, light
substance streaks, etc.) which have survived and possibly been weakened by
the couch and press draws. Other weaknesses may be caused in the dryers
themselves if some regions remain very damp as the remainder of the web
dries. Such damp regions can occur if a piece of hard dirt is embedded in
the sheet (this appears to prevent the felt pressing evenly on the paper
surface giving an area round the particle which dries slower) or from
condensation dropping, cylinders sweating, or if the felt does not give an
even pressure on the web (bulges caused by damage or poor running in,
poor seams, holes, etc.).
Certain parts of the drying section are more prone to breaks than others :
the first few cylinders where the web, especially when too damp, may tend
to stick to the dryer surface and blister if the temperature is too hot or the
cylinder dirty, is a frequent location of breaks; also susceptible are the
cylinders towards the end of the section where the shrinkage rate is
relatively great and the web tension is more sensitive to changes in beating
and making conditions. It is useful for the dryerman to keep a record of the
number of the dryer which immediately precedes the point where each
break occurs; over a period this record shows up the positions which give
544
most trouble, and serves to indicate that attention is needed to such things
as altering the flash system to change the temperature of the first few
cylinders, checking the alignment of one or more cylinders and felt rolls,
making provision for more draws, and so forth.
Breaks in the calenders are almost entirely due to defects in the web
which can have originated anywhere down the machine. A very damp place
of relatively small area may well survive the dryers, but when the calenders
are reached this region of the web becomes distorted into a light patch
trailed by corrugations giving a miniature creping effect, or it may be
stamped out completely to produce a hole. Creases or strained regions of
the web readily fold over in the calenders to produce a permanent cut in
the paper which may tear apart where the web leaves the bottom nip. A
regular transparent spot at one point of the web is usually caused by a piece
of fibre or pitch adhering to a calender roll and passing under the doctor.
The draws to and from the calenders are particularly liable to fluctuation
due to changes in load caused by cylinders waterlogging, belts slipping, and
so forth, and also due to changes in tension as the machine roll builds up;
this makes these draws very sensitive to any defects in the web. Dust can be
a problem at the calenders but can be reduced by running at a higher reel-
up moisture content, avoiding spreader bars in favour of expander rolls,
and employing vacuum doctors to suck the dust away.
Fortunately, on most machines, more defects in the paper web actually
survive intact through to the reel-up than cause a break; if this were not the
case few paper machines would make any profit. Many of these defects are
admissible in the saleable paper, provided they are relatively small, but
many must be removed before the customer takes delivery. When paper is
destined to be cut up in sheets and sorted either by hand or machine, this
presents little difficulty; but when the paper is for sale in roll form it is
important that as many defects as possible are removed at the winder. As
this runs at greater speed than the paper machine, it is obviously not so
easy for the winder crew to spot any holes, calender cuts, damp blotches,
and so forth as it is for the dryerman who can observe the sheet travelling
down the calenders and across to the reel-up. All good dryermen, even when
relaxing and talking, do so facing the reel-up so that their eye catches any
defects and they can then mark the side of the roll or insert a tab to ensure
closer inspection and possibly removal at the winder. A roll with a great
many tabs sticking out of it does not by any means condemn a dryerman
for not remedying some recurring defect, which may not anyway be within
his power; rather it shows up consideration for his colleague on the winder,
even though this man may not regard it that way when he sees the roll.
545
REFERENCES
(1) Attwood, B. W. and Smith, S. F.: 'An Experimental Machine for Investigating
the Drying Process.' Proc. Tech. Sec. 31, 3, 577. November, 1950.
(2) Smith, S. F.: 'Dried-in Strains in Paper Sheets.' Proc. Tech. Sec. 31, 1, 107.
February, 1950.
(3) Arlov, A. P. and Ivarsson, B.: 'Effects of Draws and Felts on Mechanical
Properties.' Svensk. Pap. 54, 21, 729. November 15, 1951.
(4) Rance, H. F.: 'The Influence of Papermaking Variables upon Sheet Character-
istics.' Proc. Tech. Sec. 33, 1, 173. February, 1952.
(5) Jordan, J. C.: 'Paper Mill Ventilation with Closed Hoods at Iroquois Falls.'
P. and P. Mag. Can. 54, 3, 253. Convention Issue, 1953.
(6) Lukianovitch, S.: 'Moisture Control with Closed Hoods.' P. and P. Mag. Can.
54, 10, 126. September, 1953.
(7) Smith, S. F. and Attwood, B. W.: 'Paperboard Drying Investigations by Means
of an Experimental Dryer.' TAPPI 36, 11, 481. November, 1953.
(8) Chalmers, G. J.: 'Performance Testing the Dryer Section of a Paper Machine.'
P. and P. Mag. Can. 55, 3, 326. Convention Issue, 1954.
(9) Cooke, B. D.: 'The Behaviour of Condensate and the Operation of Dryer
Siphons.' P. and P. Mag. Can. 55, 7, 119. July, 1954.
(10) Jordan, J.C.: 'Report on Closed Hood Installations at Iroquois Falls.' P. and P.
Mag. Can. 55, 3, 229. Convention Issue, 1954.
(11) Montgomery, A. E.; 'Variation of Drying Rate of Individual Driers Through
the Drier Section.' TAPPI 37, 1, 1. January, 1954.
(12) Nissan, A. H.: 'Functions of the Felt in Water Removed on the Papermaking
Machine.' TAPPI 37, 12, 597. December, 1954.
(13) Rance, H. F.: 'Effect of Water-Removal on Sheet Properties.' TAPPI 37, 12,
640. December, 1954.
(14) Brecht, W. and Pothmann, D.: 'The Effect of Tension During Drying on the
Behaviour of Paper.' Das Papier 9, 13/14, 304; 17/18, 429. 1955. (In German.)
(15) Carlander, A.: 'Development of Continuous Thickness Measurements on Paper
Machines.' Svensk Pap. 58, 16, 578. 1955. (In Swedish.)
(16) Carlander, A. T. and Jordansson, L.: 'Examination of a Surface Pyrometer for
Measuring Drying Cylinder Temperature.' Svensk Pap. 58, 17, 616. 1955. (In
Swedish.)
(17) Knowles, M., Mardon, J., and White, R.: 'Some Observations on Drying on
Yankee-type Cylinders. Part III. Air and Heat Balances in the M.G. Hood.'
Proc. Tech. Sec. 36, 1, 129. February, 1955.
(18) Mardon, J., Gavelin, G., and Logan, K. C.: 'The Effect of Some Machine
Characteristics on Paper Uniformity.' P. and P. Mag. Can. 56, 3, 275. Convention
Issue, 1955.
(19) Nissan, A. H. and Kaye, W. G.: 'An Analytical Approach to the Problem of
Drying of Thin Fibre Sheets on Multicylinder Machines.' T APPi 38, 7, 385.
July, 1955.
(20) Rogers, W. A. W. and Webster, G.: 'Measurement of Moisture in a Running
Dry Felt,' P. and P. Mag. Can. 56, 3, 241. Convention Issue, 1955.
(21) Brecht, N., Gerspach, A., and Hildenbrand, W.: 'The Influence of Drying
Tension on Some Paper Properties.' Das Papier 10, 19/20, 454. 1956. (In German.)
(22) Dreshfield, A. C. Jr. and Han, S. T.: 'The Drying of Paper.' TAPPI 39, 7, 449.
July, 1956.
(23) Lippke, P. and Rau, W.: 'The Measurement and Control of Moisture with
Hygrotester Instruments.' P. and P. Mag. Can. 57, 3, 252. Convention Issue,
1956.
(24) Nissan, A. H.: 'A Theory of Drying of Sheet Materials using Heated Cylinders.'
Chem. and Ind. 13, 198; and J. Textile Institute 47, T48. 1956.
546
THE DRYING SECTION AND CALENDERS
(25) Soinenen, M.: 'On Paper Drying Theory.' Paperi ja Puu 38, 10, 491. October.
(In Finnish); also Paper Trade Journal 142, 19, 36. May 12, 1956. (In English.)
(26) White, R. E.: 'Residual Condensate, Condensate Behaviour, and Siphoning in
Paper Driers.' TAPPI 39, 4, 228. April, 1956.
(27) Bell, J. R., Robinson, W. F. E., and Nissan, A. H.: 'The Role of Woollen Felts
in the Removal of Water in the Drying Section of a Paper Machine.' T APPi 40,
7, 558. July, 1957.
(28) Maynard, C. E. G. and Newman, J. A. S.: 'The Prevention of Machine Shrinkage
and its Effects on Moisture Expansion.' TAPPI 40, 7, 177A. July, 1957.
(29) Preston-Thomas, H. and Dauphinee, T. M.: 'Some Factors Affecting Drying on
Paper Machines.' P. and P. Mag. Can. 58, 5, 159. April, 1957.
(30) Snider, E. H.: 'Newsprint Drying Questionnaire Summary.' Canadian Tech.
Sec. Proc. p. 166. 1957.
(31) Wultsch, F.: 'Factors Affecting the Smoothness of Paper.' Alig. Pap. Rund. 12,
607. 1957. (In German.)
(32) Bowden, J.M.: 'Dry Felts. Review of Dry Felts at Present Available and Results
found from their Use in the Manufacture of Fine Paper.' Proc. Tech. Sec. 39, 2,
235. June, 1958.
(33) Brecht, W. and Heyn, D.: 'The Smoothness of Calendered Paper and its
Dependence on Atmospheric Humidity.' Woch fur Pap. 86, 23/24, 1037. 1958.
(In German.)
(34) Hopkins, L. F.: 'A New Drying Cylinder Surface Temperature Thermometer.'
Proc. Tech. Sec. 39, 1, 11. February, 1958.
(35) Janson, L. and Nordgren, B.: 'Drying of Paper on Felt Covered Cylinders.'
Svensk Pap. 61, 19, 834. October 15, 1958. (In Swedish.)
(36) Jepson, M. D.: 'Some Applications of Forced Convection Drying in the Paper
Machine Industry. World's Paper Trade Review 150, 25, 2179. December 18,
1958.
(37) Preston-Thomas, H.: 'Uniformity of Drying Across a Sheet.' P. and P. Mag.
Can. 59, 2, 73. February, 1958.
(38) Race, E.: 'A Machine for Determining the Influence of the Properties of Felts on
the Drying of Paper.' Proc. Tech. Sec. 39, 2, 215. June, 1958.
(39) Bhargava, R. L. and Robertson, J. H.: 'Infra-red Drying of Paper.' TAPPI 42, 5,
385. May, 1959.
(40) Brauns, 0. and Ponton, S.: 'Drying of Paper on Felt-covered Cylinders. The
Importance of the Felt for Transport of Heat and Water.' Svensk Pap. 62, 7,
245. April 15, 1959. (In Swedish.)
(41) Brecht, W.: 'The Behaviour of Papers when their Moisture Content Changes.'
Parts I and II. Das Papier 13, 7/8, 130; 13, 9/10, 201. April, May, 1959. (In
German.)
(42) Brecht, W., Raderer, P., and Weitzel, W.: 'Tendency of Paper to Curling.' Das
Papier 13, 11/12, 237. June, 1959. (In German.)
(43) Campbell, J. E. and Hughson, G. D.: 'Improvement of Moisture Profile by
Dryer Insulation.' P. and P. Mag. Can. 60, 3, T-77. March, 1959.
(44) Daane, R. A.: 'An Experimental Study of Some Drier Drainage.' TAPPI 42, 3,
208. March, 1959.
(45) Eisele, K. M.: 'Infra-red Radiation Applied to Paper Drying.' P. and P. Mag.
Can. 60, C, 151. Convention Issue, 1959.
(46) Hill, L. E. Jr.: 'Minton Vacuum Driers.' TAPPI 42, 5, 375. May, 1959.
(47) Ponton, S.: 'Drying of Paper on Felt Covered Cylinders.' I. 'A Comparison of
Different Types of Felt.' Svensk Pap. 62, 19, 692. October 15, 1959. II. 'A
Comparison Between Two Wool Dry Felts of Different Weight.' 62, 20, 749.
October 30, 1959. (In Swedish.)
(48) Preston-Thomas, H.: 'Recent Studies in Paper Drying.' P. and P. Mag. Can. 60,
l, T-17. January, 1959.
(49) Race, E.: 'Experience in the Manufacture and Use of All-Synthetic Drier Felts.'
TAPPI 42, 10, 827. October, 1959.
(50) Barnscheidt, N. and Schadler, M.: 'Removal of Condensate from Drying
Cylinders.' Das Papier 14, lOa, 600. 1960. (In German.) Summary in Paper
547
THE DR YING SECTION AND CALENDERS
(51) Chapman, J.: 'General Factors Concerning M.G. Operation.' Paper Technology
l, l, 33. February, 1960.
(52) Glynn, P., Jones, H. W. H., and Gallay, W.: 'Drying Stresses and Curl in Paper.'
Canadian Tech. Sec. Proc. p. 378. 1960.
(53) MacLaurin, M. I.: 'New Instruments for Control of the Paper Machine.' Paper
Technology 1, 4, 381. August, 1960.
(54) Nissan, A. H. and Hansen, D.: 'Heat Transfer and Water Removal in Cylinder
Drying.' I. 'Unfelted Cylinders.' TAPP! 43, 9, 753. September, 1960.
(55) Schoonen, J. P.: 'The Effect of Filming-Amines on Heat Transfer in Paper
Drying Cylinders.' Paper Technology 1, 4, 373. August, 1960.
(56) Smith, B. W.: 'Scanning Basis Weight and Moisture Gauge Systems on Paper
Machines.' TAPP! 43, 3, 226. March, 1960.
(57) Burgess, B. W.: 'The Drying of Web Materials.' P. and P. Mag. Can. 62, 6,
T-303. June, 1961.
(58) Gardner, T. A.: 'Performance of Slotted Orifice Air Dryers.' P. and P. Mag.
Can. 62, 6, T-327. June, 1961.
(59) Gardner, R. C. and Church, F.: 'The Automatic Control of the Transverse
Moisture Profile of a Paper Web.' World's Paper Trade Review 156, 9, 748.
August 31, 1961.
(60) Harrison, G. A.: 'An Interim Report on a Sectional Performance High Velocity
Hood Installation.' P. and P. Mag. Can. 62, 3, T-205. March, 1961.
(61) Howe, B. I. and Lambert, J.E.: 'An Analysis of the Theory and Operation of
High Speed Steel Roll Calender Stacks.' P. and P. Mag. Can. 62, C, T-139.
Convention Issue, 1961.
(62) Hurm, R. B.: 'Principles of "On the Machine" Moisture Measuring Systems.
TAPPI Project 775.' TAPPI 44, 6, 125A. June, 1961.
(63) MacLaurin, M. I.: 'Instrumentation-its Success and Delusions.' Svensk Pap.
64, 10, 394. May 31, 1961.
(64) Prince, F. J., Wagner, 0. R., and Mack, E. S.: 'Radiant-Heating-Its Character-
istics, Generation and Application.' TAPP! 44, 5, 208A. May, 1961.
(65) Stone, M. D. and Liebert, A. T.: 'Crown Control of Paper Calenders.' TAPP!
44, 5, 308. May, 1961.
(66) Tyler, J. H.: 'Factors Affecting the Thermal Efficiency of Paper Mills and their
Influence on Production.' P. and P. Mag. Can. 62, 3, T-123. March, 1961.
(67) Whittaker, J. D.: 'A New Forced Convection Drying Hood.' P. and P. Mag.
Can. 62, 1, T-11. January, 1961.
(68) Wultsch, F.: 'The Effect of Calender Rolls on the Quality and Printability of
Paper.' Woch. fur Pap. 89, 4, 143. 1961. (In German.) .·
(69) Brauns, 0. and Larsson, T.: 'High-Efficiency Hoods for Multi-Cylinder Drying.
Part 2, Machine Experiments.' Svensk Pap. 65, 12, 488. June 30, 1962.
(70) Brecht, W. and Raderer, P.: 'The Causes of Curling in Papers.' Woch. fur Pap.
90, 373. April, 1962. (In German.)
(71) Crook, 0. M. and Bennett, W. E.: 'The Effect of Humidity and Temperature
on the Physical Properties of Paper.' British Paper and Board Industry Research
Association. Report No. RA/T/90. February, 1962.
(72) Emerton, H. W., Page, D. H., and Hale, W. H.: 'Structure of Papers as seen in
their Surfaces.' Formation and Structure of Paper. Transactions of the September
1961 Oxford Symposium, p. 53. B.P. and B.M.A. Tech. Sec. publication, 1962.
(73) Gallahue, W. M.: 'New Developments in Stiffening and Stabilisation of Paper.'
(74) Glynn, P. and Gallay, W.: 'Further Studies of the Mechanism of Curl in Paper.'
(75) Hultgreen, 0.: 'Raising Quality and Production through High-Speed Drying.'
T APPI 45, 6, 178A. June, 1962.
(76) Larsson, T.: 'High-Efficiency Hoods for Multi-Cylinder Drying. Part I, Litera-
ture Survey.' Svensk Pap. 65, 5, 1964. March 15, 1962.
(77) McCaffrey, K. A.: 'Paper Machine Dryer Control with Jet Compression.' P. and
P. Mag. Can. 63, 4, T-237. April, 1962.
(78) Means, J. A.: 'Water Removal by High Velocity Air Hoods.' TAPP! 45, 6, 527.
June, 1962.
548
THE DR YING SECTION AND CALENDERS
(79) Nissan, A. H., George, H. H. Jr. and Hansen, D.: 'Heat Transfer and Water
Removal in Cylinder Drying. II Felted Cylinders.' TAPP! 44, 3, 213. March,
1962.
(80) Page, D. H. and Tydeman, P.A.: 'A New Theory of Shrinkage Structure and
Properties of Paper.' Formation and Structure of Paper. Transactions of the
September 1961 Oxford Symposium, p. 397. B.P. and B.M.A. Tech. Sec. publica·
tion, 1962.
(81) Allander, C. and Eneroth, J.M.: 'Heat Transfer when Blowing Through Hole~
and Slits onto a Flat Surface.' Svensk Pap. 66, 9, 349. May 15, 1963. (In Swedish.)
(82) Blanchard, R. L. et al.: 'The Change of Paper Properties through Machine
Calender Stacks.' Part I. 64, C, T-119. Convention Issue, 1963.
(83) Cuffey, W. H.: 'Newsprint Calendar Vibration as it affects Machine-Direction
Caliper Uniformity.' P. and P. Mag. Can. 64, 9, T-379. September, 1963.
(84) Fahey, D. J. and Chilson, W. A.: 'Mechanical Treatments for Improving
Dimensional Stability of Paper.' TAPP! 46, 7, 393. July, 1963.
(85) Gates, E. R. and Kenworthy, I. C.: 'Effects of Drying Shrinkage and Fibre
Orientation on Some Physical Properties of Paper.' Paper Technology 4, 5,
485. October, 1963.
(86) Hendry, I. F. and Newman, J. A. S.: 'Effect of Machine Variables on the Curl
of Paper.' Paper Technology 4, 4, 381. August, 1963.
(87) Howe, B. and Cosgrove, J. C.: 'Calender Stack Barring on Newsprint Machines.'
P. and P. Mag. Can. 64, 6, T-259. June, 1963.
(88) Hoyle, R.: 'Thermal Conditions in a Steam Drying Cylinder.' Paper Technology
4, 3, 258. June, 1963.
(89) Hurm, R. B.: 'Factors Affecting Dryer Cylinder Uniformity.' TAPP! 46, 9, 531.
September, 1963.
(90) Jepson, M. D.: 'Drying Paper on Cylinders with Accelerator Hoods-Principles
and Practice.' Paper Technology 4, 3, 268. June, 1963.
(91) Kerler, J. D.: 'Operation of the Ross Sectional-Performance, High-Velocity
Hood.' P. and P. Mag. Can. 64, 2, T-53. February, 1963.
(92) Kirk, L. A., Hudson, F. L., and Hearle, J. W. S.: 'Mechanism of Moisture
Transfer from Paper to Felt During Hot Surface Drying.' Paper Technology 4,
3, 251. June, 1963.
(93) Kuehn, H. E.: 'Calender Crown Compensation by Precision Control.' TAPP!
46, 6, 193A. June, 1963.
(94) Luckins, J.: 'Theoretical Benefits from New Methods of Drying.' Paper Tech-
nology 4, 3, 245. June, 1963.
(95) Mercer, J. D.: 'Operation of the Ross Sectional-Performance, High-Velocity
Hood.' P. and P. Mag. Can. 64, 2, T-53. February, 1963.
(96) Pye, I. T.: 'Calender Barring in Newsprint.' P. and P. Mag. Can. 64, 4, T-194.
April, 1963.
(97) Robey, L. A. and Webzell, A. B.: 'Steam Supply and Condensate Removal.'
Paper Technology 4, 3, 235. June, 1963.
(98) Wahlstrom, P. B., Larsson, K. 0. and Asklof, C. A.; 'Calender Barring, its
Mechanism and possible Elimination.' P. and P. Mag. Can. 64, 4, T-265. April,
1963.
(99) Blatchley, C. G. and Stratton, H.J.: 'The Performance ofThermocompressors as
related to Paper Machine Dryer Drainage Systems.' P. and P. Mag. Can. 65, 7,
T-301. July, 1964.
(l 00) Gardner, T. A.: 'Minimum Pressure Differentials required for Dryer Drainage.'
P. and P. Mag. Can. 65, 4, T-188. April, 1964.
(101) Gardner, T. A.: 'Engineering Aspects of High-Velocity Drying on Paper
Machines.' P. and P. Mag. Can. 65, 4, T-191. April, 1964.
(102) Hillman, C. H.: 'Calender Roll Crown Compensation by the Swimming Roll
Method.' TAPP! 47, 8, 141A. August, 1964.
(103) Kenworthy, I. C., 'Drying Strains and Shrinkage in Paper.' Paper Technology 5,
4, 259. August, 1964.
(104) Jackson, M. and Ekstrom, L.: 'Studies concerning the Compressibility of
Paper.' Svensk Pap. 67, 20, 807. October 31, 1964.
(105) Kettering, G. H.: 'Calender Nip Relieving.' TAPP! 47, 8, 137A. August, 1964.
549
THE DRYING SECTION AND CALENDERS
(106) Mardon, J., et al.: 'The change of Paper Properties through Machine Calender
Stacks. Part II.' P. and P. Mag. Can. 65, 11, T-481. November, 1964.
(107) Page, D. H. and Sargent, J. W.: 'The Structure and Physical Properties of Paper.
Part II The Structure of Paper in Cross Sections.' British Paper and Board
Research Association. Report No. RA/T/112. September, 1964.
(108) Sawyer, W. C.: 'The Dynamics of Heat Flow in Paper Machine Calender Rolls.'
TAPP! 47, 3, 145A. March, 1964.
(109) Wahlstrom, P. B. and Larsson, K. 0.: 'Factors determining Condensate
Removal.' P. and P. Mag. Can. 65, 5, T-203. May, 1964.
(110) Attwood, D. and Hitchin, D. F.: 'Control of Moisture Profile with Cross-
Control Drying Hoods.' P. and P. Mag. Can. 66, 1, T-27. January, 1965.
(111) Burnham, L. A.: 'Pocket Ventilating Rolls to optimize Drying.' TAPP! 48, 7,
82A. July, 1965.
(112) Church, F.: 'Moisture and Substance Control on the Paper Machine.' P. and P.
Mag. Can. 66, 1, T-3. January, 1965.
(113) Cole, E. J., Drapeau, C. I., and Whitney, C. G.: 'New Aspect in the Theory and
Use of Calendars on the Machine.' La Papeterie 87, l, 30. January, 1965. (In
French.)
(114) Kahoun, J. B.: 'The Development of an On-Machine Caliper Profiler.' P. and P.
Mag. Can. 66, 3, T-195. March, 1965.
(115) Kahoun, J. B.: 'System for On-Machine Measurement and Control of Paper
Caliper.' TAPP! 48, 7, 60A. July, 1965.
(116) MacLaurin, M. I.: 'Control of the Final Moisture Content of the Paper Web.'
P. and P. Mag. Can. 66, 1, T-11. January, 1965.
(117) Parker, J. R.: 'Corrugation of Calender Rolls and the Barring of Newsprint.'
Paper Technology 6, 1, 33. January, 1965.
(118) Schaffrath, P. M.: 'New Attachment improves Calender Stack Operation.'
Paper Trade Journal, 149, 47, 57. November 22, 1965.
(119) Woodside, L. M. and MacMillan, H.J.: 'Guiding Felts-Which Way the Right
Way?' Pulp and Paper Mag. America 39, 50, 35. December 13, 1965.
(120) Anon: 'Strengthening of the Paper Web and Dimensional Stability.' Allg. Pap.
Rund. 23, 820. June 21, 1966. (In German.)
(121) Barbour, D. R. et al.: 'Stress Relieving of Paper on the Paper Machine.' P. and
P. Mag. Can. 67, 6, T-292. June, 1966.
(122) Calkins, D. L.: 'The Effects of Siphon Location on Paper Drying.' P. and P.
Mag. Can. 67, 4, T-225. April, 1966.
(123) Jackson, M. and Gavelin, G.: 'The Role of Nip Temperature and Surface
Moisture in the Calendering and Supercalendering Processes.' Svensk Pap. 69,
5, 131. March 15, 1966.
(124) Jender, B. and Gavelin, G.: 'Sheet Moisture Control on Paper Machines with
Air Blowing Rolls.' P. and P. Mag. Can. 67, 12, T-528. December, 1966.
(125) Mardon, J. et al.: 'The Organisation, Carrying Out and End Results of Heat
and Material Balances on Papermaking and Coating Machinery.' P. and P. Mag.
Can. 67, 1, T-37. January, 1966.
(126) Mardon, J. et al.: 'Dynamic Consolidation of Paper during Calendering.'
Consolidation of the Paper Web. Transactions of the September 1965 Cambridge
Symposium, p. 576. B.P. & P.M.A. Tech. Sec. publication, 1966.
(127) Page, D. H., Sargent, J. W., and Nelson, R.: 'Structure of Paper in Cross-
Section.' Consolidation of the Paper Web. Transactions of the September 1965
Cambridge Symposium, p. 313. B.P. & P.M.A. Tech. Sec. publication, 1966.
(128) Page, D. H. and Tydeman, P.A.: 'Physical Processes occurring during the Drying
Phase.' Consolidation of the Paper Web. Transactions of the September 1965
Cambridge Symposium. p. 371. B.P. & P.M.A. Tech. Sec. publication, 1966.
(129) Parker, J. R.: 'Effects of Dry Pressing on Printing Properties of Uncoated Paper
Webs.' Consolidation of the Paper Web. Transactions of the September 1965
Cambridge Symposium, p. 959. B.P. & P.M.A. Tech. Sec. publication, 1966.
(130) Race, T. et al.: 'Hot-Air Blowing Rolls with Conventional Dryer Felts and
Open-Mesh Dryer Fabrics.' TAPPI 49, 9, 59A. September, 1966.
(131) Simmons, T.: 'Condensate Removal on High-Speed Papermaking Machines.'
Svensk Pap. 69, 23, 813. December 15, 1966. (In Swedish.)
550
PART 6
PRODUCTION CONTROL
INTRODUCTION
61 In the foregoing Parts the Fourdrinier paper machine has been sub-
jected to detailed examination and an attempt has been made to present
information which will be useful to production personnel in their efforts to
uvercome operational problems and improve output and quality. In this
final section, the paper machine will be looked at from a wider viewpoint
and treated as the main part of a manufacturing concern, the paper mill,
which is required to operate as profitably as possible. Emphasis will, in
other words, be on the paper machine as a whole, with particular attention
given to methods of monitoring production and quality and to the detec-
tion of inefficiencies in the system and assessment of their importance.
These aspects of operating a paper machine come under the general
heading of production control and involve examining the process on a
longer time-scale than has been considered hitherto. Concern here is with
the gathering and analysis of statistics and the presentation of data in a
form which can readily be interpreted by those concerned with managing
the plant. In the broadest sense, the objective of such data is to point up
weaknesses in the production system which, with the aid of various depart-
ments in the mill, can then be tackled in a programme of long-term
improvement.
This aspect of running the paper machine has of course been implicit
throughout the book and has already been discussed in various guises.
Better understanding of the process from a proper appreciation of available
theoretical knowledge, improved awareness of sensitive spots on the
machine, consistent operation and a planned sequence for starting and
shutting a machine, proper planning of maintenance schedules, sensible
organization of grade changes, the use of instrumentation to guide
operatives in maintaining steady conditions, all these in one form or
another relate to improving the system. Even more directly concerned are
the keeping of careful records of clothing changes, suction box dressing,
calender and press roll grinding, and so on.
Of particular importance in any programme of steady improvement is
the gathering of data for the purpose of establishing over a long period
standards of operation which act as a touchstone to check conditions on
some subsequent occasion. The data it is desirable to collect in this way has
been detailed under the heading of 'Long-term records' in each section
dealing with maintenance. It is important that this information is obtained
at times when the machine system is running well and making paper of
satisfactory quality (unfortunately this is the precise occasion when the
papermaker is reluctant to cause a break); ideally it should be performed
on each of the major grades of paper produced. In times of trouble on the
machine or serious quality deterioration, the existence of standards which
can form a basis of comparison can be invaluable and the expense of
taking the necessary tests at intervals and making short interruptions to
553
61. l PRODUCTION CONTROL
production are rapidly repaid. Results obtained from these tests should of
course be presented to the production personnel in a manner which allows
ready interpretation of changes and trends; in many cases the most
appropriate form is a straightforward graph on which, over a period, it may
be possible to draw action limits to assist taking a decision whether or not
to enquire closer into any sudden change in level.
61. 1 Monitoring machine performance
It is proposed in what follows to treat the subject of production control in
two parts. The first part will be concerned with methods of monitoring
production, the second with ways of improving it and the means of
assessing their significance.
With a piece of machinery as large and complex as a paper machine it is
only to be expected that improvements to the process are continually being
sought. Many such improvements are the result of overcoming obvious
snags and replacing worn out pieces of equipment. In these cases it is
usually plain for all to see that a particular course of action is desirable and
even, in some cases, imperative. But sometimes there may be other time-
wasting equipment and practices that are not immediately evident. How
can these be detected ?
The answer is by comprehensive and accurate monitoring of time lost in
operating the machine. Properly designed records of downtime can indicate
the greatest sources of wasted time, and so point the need to concentrate
effort in certain spots. For this reason, an essential feature of any paper
machine should be the keeping of a systematic analysis covering down-
time, speed, production, and efficiency (the latter under various headings).
With such records it will be possible to tell just how well progress is being
made in improving operation. This alone is a valuable facility, if only to
disillusion one when anticipated increases in speed or reduction in broke
do not materialize.
Monitoring machine performance means not only keeping a check on
the actual quantity of paper produced but also on the quality. This
impinges on quality control, a subject that has already been sufficiently
written about. In the present context attention will be given more to some
broader issues relating to quality control and rejection schemes, and also
to the long-term monitoring of quality.
61. 2 Improvement of machine performance
At the root of all improvement is, or should be, the potential financial
advantage. Too often this all-important point is ignored, particularly it
may be said by technical personnel. Often it is simple to detail the technical
advantages of some innovation. But to be valid these must be brought
down to hard cash before they can really be justified. Unfortunately this is
very often difficult and sometimes impossible to fulfil.
The ability to provide relevant financial data for such assessments is one
of the prime advantages of a detailed costing system. The most convenient
form this can take seems to be that of standard costing as this lends itself to
providing answers to the many questions that must be asked before it is
554
PRODUCTION CONTROL 61.2
possible to present an economic justification for a particular course of
action. To be able to assess whether any improvement is economically
valid it must be possible to relate increases in machine speed, reductions in
broke and downtime, raising moisture levels in finished paper, and so on to
pounds, shillings and pence. Associated with this is the need to be able to
determine the actual cost of producing different grades of paper and
different substances of the same grade. Too often it is the case that the
profitability of different grades, especially when specific requirements
create difficult machine conditions or necessitate a prior shut, are known
only hazily.
These are the subjects which will be discussed under the general heading
of production control. It is surprising how little has been written about this
topic and how few references there are in the technical literature. Batch
production processes seem to have been given far more attention than
continuous processes like papermaking, and unfortunately the methods
used in control of batch processes are rarely relevant. It is hoped that the
discussion in the following pages will help in some small way to remedy this
deficiency.
555
CHAPTER 6A
MONITORING MACHINE PERFORMANCE
6A .1 ANALYSIS OF DOWNTIME
Checking the downtime occurring on a paper machine is probably the most
simple yet useful method readily available to monitor performance. It has
the merit of being largely independent of the type of paper made on the
machine, and so overcomes the problem of analysing overall machine
performance when a variety of grades are made in rapid succession. Also
the cost of an hour's lost time can usually be estimated with some accuracy
even in mills with only a rudimentary accounting procedure, so it is a
relatively simple matter to determine the financial loss resulting from
downtime due to different causes.
A really comprehensive downtime analysis has many advantages. It
shows up the main sources of interruption to steady production and allows
the benefits of any improvement from an alteration made to the machine or
to operating procedure to be readily determined. It also, when properly
presented, draws the attention of management to deficiencies in the produc-
tion system and assists in defining areas to which technical resources should
be directed. If certain relevant parts of the data are m1.de available in an
appropriate manner to machine crews, it is possible that operating discip-
line will improve. Though in this respect careful consideration of mill
labour conditions is needed before publicizing shift-wise comparisons as it
is not always either desirable or wise to attempt to inculcate an open
competitive spirit. Finally, where a variety of grades is run on a machine an
accurate breakdown of time lost in preparation and in running each grade
permits more detailed costing figures to be accumulated.
6A. 1 1 Collecting information on downtime
The traditional form of report compiled by the machineman or foreman at
the end of each shift is completely inadequate for any reliable analysis of
downtime. Estimates of lost time from different causes and on specific
occasions will almost certainly be inaccurate, and a strong tendency to
report fewer breaks and lower times than actually occurred would be an
understandable human failing. In cases involving clothing changes or a
routine wash-up it is quite likely that a standard time is reported regardless
of what actually happened. Nevertheless some sort of report is obviously
essential and the form this takes depends on how other data on downtime
is obtained.
The simplest method of improving the accuracy of available data is to
link to the machine a recorder which clearly indicates the periods when no
paper has been passing to the reel-up. This recorder can serve solely to
provide downtime data, or may be used in addition for some other purpose,
for example to show the machine speed. It is important that all breaks
556
PRODUCTION CONTROL 6A.l l
occurring at the reel-up are indicated and this can be arranged in a straight-
forward manner by using a photoelectric or ultrasonic web detector
immediately prior to the reel-up. In some cases changes in power input to
the calender or reel-up can· be used for the same purpose, or a micro-
switch on a small feeler resting on the web can act as a sensing device.
When either of these systems is employed, however, some difficulty can
occur at reel changes when flapping of the web inadvertently actuates the
break-detector mechanism. Continuous substance or moisture meters are
an alternative method of obtaining a record of breaks though, if the
instrument heads are removed manually from the web at a break and then
replaced when the sheet is up again, there is a likelihood of consistent error
in the length of time recorded. In addition to providing a simple record of
breaks, a useful facility is to arrange for these to be added together on a total-
izing counter. For convenience in checking downtime on different grades,
this totalizer can be in two parts, one of which is re-set manually to zero
at the start of each grade and logged at the finish.
The breaks shown on such a recorder will vary in length from a blip
where the web has broken at the calenders and been immediately fed back,
through to the really long breaks where the machine has been shut. A
variety of data can be extracted from the record of these different breaks
(this is dealt with below) but the most valuable information obviously
depends on the ability to relate each break to some specific cause. It is here
that the machine report becomes important and must be compiled using
the recorder chart as a basis. The recorder should be considered by the
machine crews and supervisory staff as an aid to more accurate reporting
of downtime, not in any way as a means of assessing their individual
performance (hence the reservations with regard to publishing shift-wise
comparisons) otherwise co-operation could become difficult to secure. In
any event, it is usually only practicable for reasons to be given for breaks
which last for some time, say over ten or fifteen minutes, otherwise
compiling the machine report becomes a lengthy business.
It is also valuable to know whereabouts on a machine breaks occur most
frequently and to compile this information necessitates careful reporting.
Elaborate equipment has been devised to do this automatically by using a
succession of vacuum switches and photoelectric or ultrasonic web
detectors placed at strategic points down the machine to sense when it is
operational (14, 15, 17). The signals from these detectors are then fed to a
specially-designed data-logging device which produces a simple read-out
listing time lost due to breaks in various positions, with cumulative totals
over appropriate periods and calculations of operational efficiency for each
section. The same signals can of course be used for actuating other opera-
tions integral with the machine (couch pit pump started, cutter brought
over, etc.) and for giving audible and visual indications of a break.
Whether the expense involved in installing and maintaining such equip-
ment is justified depends on the value any individual mill places on having
data of such precision available. It is still necessary to report the reasons
for breaks occurring, so close co-operation with operating staff is vital.
There may also be a danger of accumulating too much information so that
557
6A.l 2 PRODUCTION CONTROL
the main points of interest become lost in a sea of figures. On the other
hand it is very useful to know the relative proportion of breaks occurring
at different positions on a machine, and careful discipline is required if this
is to be noted at all accurately by one of the operators. Possibly it is useful
to have one or two break detectors at, say, the couch and presses and at a
size press or on-machine coater to provide some independent picture of
what is happening, and these can be linked to quite simple counting
devices. But apart from this an attempt should be made by the operators
to list breaks occurring at different positions; in particular, for breaks in
the drying section it is useful to note the cylinder immediately preceding
each break because this data built up over a period can assist in pinpointing
fluctuating and uneven draws, and water-logging cylinders.
An essential feature in encouraging accurate reporting is to design a
suitable form for the shift machineman or supervisor to fill in, rather than
just expect him to provide the information required in a blank book.
Examples of such forms have appeared in the literature (see 9, 12, 14, 15)
and must be tailor-made to suit each machine. Broadly-speaking, it is
important to keep the demand for details down to a minimum and design
the form so that a simple sequence of facts has to be filled in for each break.
This would include: time break started and finished (preferably taken from
recorder chart); position break occurred (tick in appropriate column);
reason for break (again a tick in an appropriate column with a space for
miscellaneous reasons); and finally space for notes on grade changes,
machine conditions, equipment requiring attention, loss of quality, and so
on. In certain circumstances when there is doubt about the cause, it is
desirable that whenever possible the ends of breaks are kept for inspection
by production supervisors. A sequence of short breaks at the calenders
would not be individually listed. An alternative to using a machine report
for detailed listing of breaks is to employ a device whereby one of a
number of selected causes can be chosen on a telephone type dial which
marks the appropriate number on to the recorder chart. This has the merit
of making subsequent analysis much easier but cannot completely replace
a machine report.
A well laid-out machine report also permits simple clerical analysis, a
further asset. It is important, however, that this analysis should always be
performed in close association with someone in the production department
who understands what is happening and can interpret the reasons given for
breaks in a sensible manner. The cause of any doubtful breaks must be
discussed with the personnel concerned otherwise friction can develop
between the production and other departments in the mill over the basic
reason for a break occurring. Often too a particular break can be attributed
to a number of causes and some sort of apportionment has to be agreed.
On such occasions it is wise to avoid intense post-mortems so long as a
reasonable splitting does not strain amicability.
6A. 1 2 Analysing downtime data
With a record of breaks occurring on the machine, and a reasonably
comprehensive report from the supervisory personnel, it is possible to
558
PRODUCTION CONTROL 6A.l 2
accumulate really invaluable information relating to all aspects of per-
formance. The data collected in this way is all essentially long-term and is
not usually suitable for checking on day-to-day production difficulties. It is
proposed now to discuss the sort of procedure that can be adopted when
analysing breaks records, and the type of information obtained. In this
respect, much depends on machine conditions and on the length of runs
between makings, and what follows is based on the author's experience
and must be considered only as typical.
The most practicable arrangement for a breaks recorder is to use a strip
<;hart with a speed of one inch per hour. This is conveniently analysed at
weekly intervals and a special rule marked in sixtieths of an inch can be
used to determine the length of individual breaks. The chart would first be
marked off clearly into shifts, and each shift identified with the appropriate
crew (this is particularly important when a four-shift system is used and
crews are continually changing their working hours). The number of
breaks occurring in each shift is counted, and using the rule the length of
time occupied by each break is noted on the chart. Blips on the chart can
simply be regarded as a break of one-minute duration. The next step is to
compare the breaks record with the machine report and account for all
breaks which lasted for longer than some generally agreed period, probably
ten or fifteen minutes as mentioned above. With properly prepared working
sheets this whole procedure can be made into a perfectly straightforward
clerical operation.
From this basic analysis of the chart, a variety of statistics can be
produced. Firstly depending on whether the mill shuts at week-ends or is in
continuous production, the time taken for starting-up and shutting down
or for a scheduled maintenance shut is noted. Likewise the time for grade
changes necessitating a shut for washing-up are listed, and also any shuts
for routine machine testing and for experimental reasons. A graph showing
the week-by-week time for starting-up or for scheduled maintenance shut
periods is useful for indicating trends, and also, in conjunction with an
arbitrary action limit, can be used by management as the basis for initiating
an enquiry into why a particular shut took longer than usual. Averaged
over suitable periods, this data is also particularly useful for accounting
purposes, especially where a machine runs a variety of grades some of
which require more elaborate preparation. For this purpose, to permit
accurate costing shuts must be carefully related to the appropriate grade,
and if an individual period is excessively long due to some peculiar con-
dition, to avoid unfair bias to the average it is preferable to discount it.
The breaks themselves can be analysed in a variety of ways. One useful
and simple indicator of performance is the number of breaks occurring per
shift. This is easy to determine, and provided planned breaks for grade
changes, maintenance shuts, and machine testing are discounted, and only
the figures for full shifts during which the machine has been running are
used, then some useful statistics can be obtained for long-term comparison.
For instance, the number of breaks occurring during shifts operated by
different crews provides a simple basis for comparing performance. In
contrast to comparison on a time basis, this has the merit of avoiding bias
559
6A.12 PRODUCTION CONTROL
caused by the chance occurrence of long shuts perhaps for reasons outside
the control of the machine crew. Provided a sufficiently long period is
chosen (three or six months) then an inattentive and slack crew will
gradually accumulate a slightly higher average number of breaks per shift.
It is also interesting to observe that on the night shift, irrespective of which
crew is working, there are almost always on average fewer breaks. Theories
to account for this are left to the reader.
A graph showing the grand average number of breaks per shift from
week to week can give a useful indication of general trends. This graph
should either be on semi-logarithmic paper, or the logarithm of the number
of breaks should be plotted. The purpose of this is to give equal percentage
changes the same value (a reduction from 6 to 3 breaks per shift is equally
as good as a reduction from 12 to 6).
Breaks lasting for different durations can be grouped into several
categories, for example those under five minutes, between 5 and 15 minutes,
between 15 and 60 minutes, and over 60 minutes. Summed over a long
period, this shows how much time is lost due to breaks in each group and
thereby indicates whether, for example, a large number of breaks of short
duration occurring in the dryers and calenders is making a greater total
contribution to lost time than the occasional long shut for clothing changes
or other problems. This sort of analysis sometimes provides unexpected
information because it is only too easy to ignore the waste in production
caused by a large number of short breaks. The only difficulty in arriving at
a realistic assessment of the time lost from short breaks is the limitation in
determining from the record chart the length of each break. In practice one
minute is the shortest realistic time that can be given for a simple blip,
particularly in view of the quantity of paper actually reeled up that is likely
to be torn up afterwards when there is a succession of short breaks. An
alternative and more accurate approach, suitable when a totalizer is
associated with the breaks recorder, is simply to work out the total time
lost less the sum of time lost from all itemized longer breaks.
Comparison of the time lost due to breaks resulting from different causes
usually yields the most valuable information. With a well-designed machine
report in which the reasons for all longer breaks are clearly given, the task
of analysing and allocating breaks in this way can become a routine
procedure requiring only the occasional query with the production depart-
ment. The categories into which breaks are divided must be carefully
considered, understandable, and tailored to conditions in each mill and
even to individual machines. Obvious examples of the sort of categories
normally chosen are:
mechanical (all faults caused by circumstances under the jurisdiction of
the engineers-belts, pumps, bearings, guides, etc.)
electrical (starters, motors, supply, etc.)
instruments (controllers, photo-cells, relay systems, etc.)
wire (all faults directly due to deficiencies in the wire-patching holes,
repairing cracks, changing, poor drainage, etc.)
wet felts (similarly for wet felts-made-up, edges, patching, changes,
etc.)
560
PRODUCTION CONTROL 6A.2
dry felts (similarly for dry felts-seams, worn, edges unravelling, etc.)
washing up (time lost cleaning up which is attributable to wet-end
conditions-dirt, pitch, slime, cleaners or screens clogged, etc.)
wet breaks (all otherwise unspecified breaks occurring at couch and
presses-dandy picking, sticking to wire or press rolls, tears, etc.)
dry breaks (similarly for dryers and calenders-wads in cylinders,
creasing, calender stamps, etc.)
In addition breaks occurring at specific points of the machine which are
directly attributable to operation of the equipment concerned can be listed
separately, for example size press, breaker stack, coater, H.V. hood, and
so on. Finally, and inevitably, a category for miscellaneous and accidental
breaks is needed, if only on occasion to avoid lengthy arguments.
The data yielded by this sort of analysis requires periodic examination
and it is useful to prepare reports over periods of three or six months
summarizing such information as the average number of breaks per shift
and the average time lost per week (or per hour if the working week
changes periodically) due to breaks in each of the main categories. This
gives management a clear indication of any changing trends on a machine
and highlights the main reasons for lost production in so far as it has been
caused by downtime on the machine.
6A.2 PERFORMANCE DATA
The analysis of downtime data provides a useful indication of the running
efficiency of the paper machine. But this of course tells only part of the
story because it would be possible for a machine to run with few breaks
yet produce little saleable paper. Other attributes of operation, overall good
production, speed, efficiency, and so on, have to be considered. These more
general indices of performance will now be discussed in detail. Each is
important not only for examining week-to-week variations but for
investigating long-term trends and comparisons with other machines and
also for deriving other data which can be used to examine the profitability
of running in different conditions.
6A . 2 1 Production analysis
Once attention is turned to the performance of a paper machine in terms of
production it becomes impossible to ignore the different grades of paper
produced. For a machine turning out nothing but a standard newsprint or
M.G. wrapping, analysis of production can follow the same sort of weekly
procedure suggested for analysis of downtime. But most machines make a
variety of grades which, due to the different demands of processing, have to
be run at a wide range of machine speeds. The reason for this and the
differences in production created by various machine limitations will be
examined in more detail later. For the present it is sufficient to note the
advisability of carrying out a production analysis separately for each
major grade of paper produced. Normally grades where machine conditions
are essentially similar can be lumped together, for example where the
differences involve such minor attributes as colour, calendering degree,
and, within reasonably wide limits, loading content. Different substances of
561
the same grade can be kept together when they cover a fairly narrow range,
say over 10 per cent. of the average, but otherwise it is convenient to form
separate groups about the most popular substances.
For each individual grade (using this term for convenience to refer to
each grouping of grades and substances) the first piece of information
required for each making or over a week, whichever is the shorter, is the
average running speed of the machine. There is always a speed indicator on
a machine and frequently it is adequate for the machineman's booking of
this speed to be used. However, because of its importance it is desirable
that the correctness of this speed indicator is given a periodic checking,
perhaps weekly depending on its reliability, and this is most easily carried
out with accuracy by timing with a stop watch a fair number of rotations of
an M.G. or drying cylinder the circumference of which has been very
accurately taped. Any persistent bias in the indicator reading can then be
allowed for. Where speed changes are liable to occur quite frequently
during a making it is useful to have a speed recorder. The chart from this
can then be used to give more accurately the times of changes and the
different speeds run. From this data an average speed is calculated from the
individual speeds weighted appropriately according to the time paper was
made in each run.
It is quite straightforward, knowing the average speed and the deckle and
average substance at the reel-up, to calculate the weight of paper produced.
To do this the downtime analysis is used to determine the duration of time
paper was actually reeling up. However this procedure is subject to several
inaccuracies, in particular due to the fact that the average substance as
weighed is only representative of a number of samples. Also if, for example,
two widely different speeds had been run for a similar length of time, and
practically all the lost time had occurred during the higher, calculation of
production in the way suggested would produce a figure greater than the
true value because implicit in this calculation is the assumption that time
has been lost in proportion to the length of run. This error can only be
overcome by analysing lost time separately at each speed, which com-
plicates the calculation somewhat.
For these reasons it is important actually to weigh the paper produced on
each reel. This is sometimes carried out on the lifting crane, which is not
very accurate. A preferable though not so convenient procedure is to use a
standard scales set in the floor. It is important to note in this context that
when a downtime analysis is used to characterize machine performance,
then this accounts only for paper that did not appear at the reel-up. Thus
for complete accountability, all paper that has been reeled-up must be
weighed, i.e. weighing must be done before any is stripped off.
The gross production figure for each making, determined either by
calculation or by weighing, represents all paper actually reeled up, be it of
sound quality or otherwise. Accounting for the portion that proves unsale-
able or is spoiled in further processing is important in itself (see below).
But for the purpose of assessing performance of the system as a whole and
for such essential tasks as estimating future sales potential, the figure of
greatest importance is the actual saleable production. This, divided by the
562
appropriate number of hours actually occupied by the making (normally
including preparation time), produces a figure of so many lb. or tons
saleable production per hour which is the most direct measure of how well
the whole process has operated.
Where all the paper produced is despatched in reel form, large differences
in trim waste at the winder (a feature out of the hands of machine crews)
can have a substantial effect on this saleable production figure. Accordingly
for the purpose of long-term and inter-mill comparison of machine as
opposed to process performance, allowances for this are sometimes made
by determining the lb. production per hour per inch width of saleable
deckle, a measure of production advocated in the U.S. Newsprint Service
Bureau's Manual of Instructions.
6A. 2 2 Accounting for broke losses
The term 'broke' embraces waste paper in a wide variety of situations.
'Wet broke' is used to refer to the web doctored off presses, 'dry broke' to
paper either removed at a break in the cylinders and calenders or stripped
off the machine roll. In individual mills reference will be made to 'coated
broke', 'sized broke', 'salle broke' and so on according to the origin. In the
present context the term 'broke' is used to cover only paper removed or
wasted between the paper machine reel-up and despatch of the saleable
product, i.e. after the machine roll has been weighed for subsequent
determination of the gross production. Thus, broke pulled off at the
calenders is not included, but broke stripped off the machine roll is. The
reason for making this distinction is quite simply that broke at the calenders
is made during downtime, and so would be covered in the system of
accounting advocated here. Paper stripped off the machine roll, on the
other hand, is included in the gross production figure. If this distinction is
not carefully followed then there is a strong likelihood that some broke
originating at the dry-end of the machine would in effect be accounted for
twice.
This definition of 'broke' is the most useful when comparing the overall
performance of different machines. The survey conducted by the British
Paper and Board Industry Research Association has shown how the
percentage of broke to gross production varies enormously from one mill
to another (13). Out of some 57 mills, 10 reported a loss of under 5 per cent.,
but 12 had losses greater than 20 per cent. This variation is of course a
reflection of the degree of converting required by different grades of paper,
but even so comparisons of similar grades showed considerable variation
between mills: coarser grades from 2 to 8 per cent.; fine papers from 5 to
15 per cent. for a low finish, 15 to 25 per cent. for a high finish; coated
papers from 15 to 35 per cent. To encourage reduction of these figures the
Association has set up a Broke League in which monthly comparisons
between mills (suitably coded) are published.
Before broke can be reduced in a systematic manner it must be possible
to identify where it is coming from and the reasons it is caused. It is
surprising in how many mills it is customary to collect together broke
from all sources, taking care only to separate it into grades and colours, and
563
dump this straight back into the preparation plant or into storage with
little or no attention to the quantity involved. This is particularly the case
where broke can be added back in large quantities without adversely
affecting drainage characteristics on the machine wire or strength of the
finished sheet. Until all broke is carefully separated and individually
weighed or the quantity calculated it is unlikely that any effort to reduce it
will yield anything but a temporary improvement.
The first necessity is to determine gross production of the machine.
When a machine is equipped with a pulper under the calenders there is no
problem in keeping separate broke which is made before the reel-up.
Otherwise some care has to be exercised to prevent it being included with
broke stripped off the machine roll. This is, however, not likely to be too
troublesome when it is accepted practice for the full machine roll to be
removed and weighed before any is stripped off. Some reel changes on
slow machines necessitate pulling out the web beyond the reel-up spindle
before breaking and wrapping it round to start a new roll; the paper pulled
out in this way will not be shown as a break on the usual type of recorder
and so, strictly speaking, should be included in weighing of the machine
roll.
An alternative system that has been proposed to eliminate the need for
weighing of rolls relies for its basis on yardage measurements. Thus, the
length of paper passing on to the machine roll is compared with the length
coming off the winder or at any separate process such as the coater or
supercalender. This could be advantageous where there are difficulties in
separate weighing, and it also overcomes the problem of accounting for
increased substance in coating, sizing, or impregnating. An additional
feature is to measure the equivalent yardage produced at the couch or
presses (using vacuum switches to detect presence of the web); this can
provide an alternative indication of production losses in the dryers to that
which can be derived from working on a time basis, as discussed in the
previous section.
Whichever method is used, care in determining gross production off the
machine is important because subsequent broke losses have to be compared
to it. Defining the centres at which each broke loss is to be accounted by
separate weighing and booking is the next task. This must depend on
practicability in relation to mill conditions and on the degree of finishing
and converting which is carried out. To begin with, there is no point in
accumulating a detailed balance sheet accounting for every piece of broke
unless this is presented in a form which ensures that improvement is
continually sought and persistent offenders are identified. Nonetheless it
would seem essential for efficient working that each main finishing opera-
tion is treated separately and in each the main causes of broke are identified
by the operators concerned. Weekly examination of the data derived from
this should be made by both production and finishing house managers. For
long-term records, graphs for each main grade expressing the losses at
different centres as a percentage of gross production are very useful.
The first point in most systems at which broke must be carefully checked
is a winder when this is used. All paper stripped off the top of machine rolls
564
and removed from within the roll for various reasons (holes, creases, damp
patches, dirt spots, substance changing from one order to the next, shade,
etc.) should be separately weighed. Most faults at this point are attributable
to the paper machine and so careful accounting is important because the
losses can be compared to gross production to ensure that increases in the
latter are not brought about uneconomically at the expense of lower
quality. On fast machines making paper mainly for despatch in reel form,
it is customary for whole reels cut on the winder to be sent for re-winding
if their quality is suspect or if too many sections have to be cut out of the
reel in the time available. A proportion of these reels will be recovered for
eventual sale and this must be subsequently credited in the winder broke
return. The same applies to reels with quality deficiencies which are held
over for closer examination pending a management decision. Reels re-
classified to a lower grade due to poor quality, incorrect substance, and so
on are likewise credited, though if this is a frequent occurrence it is wise
to keep a separate record of the tonnage involved so that a proper assess-
ment of the monetary loss incurred can be made.
A separate aspect of winder operation which is particularly important
with fixed-deckle operation is trim removal. This depends entirely on
balancing order requirements and the efficiency of the order department in
reducing loss to a minimum. As such it requires separate checking and for
convenience it is usual simply to compare the deckle used off the winder
with the deckle available on the machine roll (which even on fixed-deckle
machines must be checked from time to time due to changes in cross-
machine shrinkage). A straight calculation based on gross production then
determines the maximum tonnage going forward from the winder. With
this procedure trim must be kept separate from paper removed for other
reasons which has to be weighed; normally this is no problem for trim is
usually removed manually, or blown by fan, to separate barrows or direct
to a pulper. With variable-deckle machines it is customary to keep dry
trim to a minimum by adjusting the width of trim cut off at the wire to suit
the deckle needed on the machine roll. This obviously affects the gross
tonnage the machine is capable of producing and should be taken into
account when considering machine utilization (see next section) but
otherwise is a convenient facility for reducing the handling and processing
costs incurred with dry broke. Often with variable-deckle operation a fixed
allowance is made for dry trim though it is wise to check this occasionally
or over a period it will come to be exceeded.
Other obvious points requiring separate broke accounting are reel
finishing operations such as supercalendering, coating, sizing or laminat-
ing, and particularly salle operations such as cutting, sorting and counting,
guillotining, embossing, and so on which can account for quite the
majority of the total broke. In each case the major faults causing broke
should be individually listed to provide maximum information. Where
substance changes, as in coating and sizing, account must be taken of this
in the broke analysis. Paper sold as seconds or retree should be separately
accounted for as over a long period the percentage in relation to paper
passing through the salle gives an interesting basis for checking on changing
19 565
standards. Conditioning operations produce an apparent substance

increase from moisture absorption and this can be shown as broke
recovery. Even without conditioning the process of cutting and sorting will
allow some moisture pick-up but normally this would be small enough not
to affect the broke balance to any noticeable extent.
It does not require a great deal of organization to ensure that there is
adequate weighing and booking of broke at each important point of the
system. Each week a balance sheet can be drawn up showing gross produc-
tion of each grade off the machine, and subsequent losses at each of the
broke centres. It is unlikely that broke will be weighed at all centres so
some will be determined by subtraction from the others, but when broke is
individually weighed at all points the discrepancy between the total and the
difference between gross and saleable production provides a useful guide
to accuracy. Detailed information of the individual causes can also be
included. When a winder is used with a fixed-deckle machine the loss at
each centre subsequent to this can be expressed as a straight percentage of
gross production less trim loss since this represents the paper going forward
for finishing. But generally it is simplest particularly with variable-deckle
operation to express broke at each centre as a percentage of gross produc-
tion off the machine. Over a period, an upper limit for broke at each centre
can be set, and whenever the quantity exceeds this limit this can be
regarded as showing the need for closer enquiry.
6A. 2 3 Efficiency measurements
The information made available by the various indices of production and
analysis of broke losses provides all the basic statistics needed to ensure
that operation of both the machine and finishing processes are adequately
monitored. But it is often useful to have a concise figure which gives a
quick and simple indication of overall performance. For this purpose
various 'efficiency' figures can be determined (1, 6, 9, 14) and those which
in the opinion of the author are most useful will now be discussed.
Unfortunately there is no standardized procedure for calculating
efficiency. Methods vary from one mill to another, and reference to
'machine efficiencies' is generally unhelpful when commenting on the effect
of some operational change. A sub-committee of the TAPPI Fourdrinier
Committee has examined the problem of deriving an efficiency figure that
will allow inter-mill comparison. They decided that the simplest definition
for universal use is
Paper machine production time ratio =
Actual running time x 100/Scheduled running time
This definition is appropriate for 7-day operation and for a working week
of fewer hours conversion factors are suggested which allow for the differ-
ent conditions. 'Actual running time' covers the time when paper was
being made on the reel reduced pro rata according to the percentage of
subsequent broke. 'Scheduled running time' is all the time for which orders
are available to run the machine and with delays for any reason including
order changing but making a reasonable start-up allowance. This defini-
tion of efficiency is as good as any for the purpose of inter-mill comparison
566
PRODUCTION CONTROL 6A.2 3
and takes account of losses from all sources. On this basis efficiencies for
newsprint machines vary from 85 per cent. to as high as 97 per cent. The
Committee considered 80 per cent. the minimum tolerable but it is likely
many mills run at 60 per cent. or lower efficiency.
The above definition is equivalent to
time paper at reel-up saleable production x
100
total time available x gross production
where the terms 'saleable' and 'gross' production are respectively the
weight of paper actually sold and the weight of paper produced at the
reel-up. This efficiency figure is therefore readily calculated from the
analyses of downtime and production already described.
Using the time paper is at the reel-up compared to the scheduled time
available is a direct and simple method of measuring efficiency of the
machine itself. A similar figure can be derived to give a measure of
efficiency of just the wet-end of the machine by taking the time the web
passes over the couch. But efficiencies calculated this way represent only
one aspect of operation. They take no account of speed and it is well
known that running a machine faster causes on average a greater number
of breaks. If no account is taken of speed it will be found that a machine
appears to have a higher efficiency the slower it runs, and in any absolute
sense this is patently ridiculous, even though in some circumstances a
greater production may be turned out at a lower speed due to a higher
proportional reduction in lost time. Hence use on its own of this definition
of 'efficiency' will be misleading, and a mill attempting to tie bonus agree-
ments to such a definition will be jeopardizing its future.
It is preferable to use the term 'machine time efficiency' to cover the
percentage of time paper is at the reel-up in relation to scheduled time
available. For internal use, however, a definition is needed which applies
only to the paper machine and yet takes account of speed and substance
variations, one in fact which assesses as truly as possible the actual weight
of paper produced at the reel-up compared to what could have been
produced in ideal circumstances. The difficulty lies in deciding what 'ideal'
circumstances are since this implies deciding a maximum runnable speed
for each substance. In practice it does not really matter at all what value
is taken as an upper speed limit, provided once fixed it is kept at the same
figure. Calculated efficiencies will be higher or lower depending on how
low or high is the chosen maximum speed, but this is immaterial so long
as they are directly comparable. Many mills in fact prefer to use a reason-
able average speed actually run in the past as a basis for calculating
maximum production; this is prefectly all right so long as it is rarely
altered and, of course, so long as efficiencies are always under 100 per cent.
For machines running a single grade with a low substance range, a single
production figure based on some maximum speed, substance and deckle
can be used for direct comparison with gross production off the machine.
But on most machines if the efficiency figures are to be at all comparable
some realistic maximum speed must be decided for each grade and sub-
stance range. Methods of doing this are discussed in detail later.
567
6A.2 3 PRODUCTION CONTROL
The efficiency figure produced in this way can be termed the 'machine
production efficiency', and this will therefore represent a percentage
comparison of actual gross production at the reel-up with what could have
been produced in the scheduled time available at some agreed maximum
speed for each substance. Taking this a stage further, 'overall production
efficiency' becomes the net saleable production related to the same denom-
inator. As a final refinement the net saleable production can be expressed
as a percentage of the paper that could have been produced at maximum
speed for each substance in the complete working week, including shuts
for maintenance and normal starting-up allowance. This may be termed
'plant efficiency'.
These various definitions are summarized in descending order of
magnitude as follows:
Wet-end time efficiency =
time web at couch x 100/scheduled time available
Machine time efficiency = .
time paper at reel-up x 100/scheduled time available
Machine production efficiency
gross paper at reel-up x 100/maximum possible paper
Overall production efficiency =
saleable paper produced x 100/maximum possible paper
Plant production efficiency = saleable paper produced x
100/maximum possible in complete working week
where 'scheduled time available' is the operating time including grade
changes, a minimum start-up allowance, and all sources of downtime;
'maximum possible paper' is the weight of paper that could be produced
were each grade and substance run at maximum speed and deckle for the
time actually occupied by the making; and 'full working week' includes
time allowed for start-up and maintenance.
These definitions of efficiency cover the machine system as a whole, but
the efficiencies of individual parts of the system are probably more con-
veniently expressed in terms of their own maxima. Thus the significance of
'trim efficiency' is best appreciated if it is defined as:
trim efficiency actual deckle average x
100/maximum deckle available
where 'actual deckle average' represents the average of actual deckles
produced at the machine roll or winder weighted according to the tonnage
produced at each deckle. Similarly, for a coater, efficiencies might be
defined as:
coater time efficiency = time paper at coater reel-up x
100/scheduled time available
coater production efficiency = gross coated paper at reel-up x
100/maximum possible coated paper
where 'maximum possible coated paper' is the weight of base paper used
increased pro rata according to the coating weight added.
568
The main point is that each definition should be readily understood,
considered fair by the operatives, and available to assist supervisors and
management to make an early detection of deteriorating conditions or tell
when there has been an improvement. As with the other indices of per-
formance, a weekly graph with arbitrary upper limits to guide when
remedial action is desirable should be plotted for each efficiency figure.
For long-term appraisal an average covering three to six monthly periods
should be circulated to all relevant departments and comment made on any
trends.
There remains one final aspect of efficiency which is of a rather different
nature, that referring to the usage of raw materials in relation to the output
of paper. When for one reason or another broke cannot be re-used
(fluorescent dyes, unsuitable colour, parchment, etc.) it may have to be
sold at a scrap price or even dumped or burnt. Obviously it is then desirable
to keep a separate account of this as it represents a direct loss to the mill
and reduces the efficiency of raw material usage. Fortunately as it is
common for all broke to be used in one way or another, so all fibre and
loading fed in to the system in the preparation plant leaves as saleable
paper apart from the proportion that is drained to waste in the backwater
system and elsewhere. The more closed the backwater system and
fewer the grades necessitating prior wash-up involving draining the wet-
end and the remains of chests, the less will be the loss of fibre and
loading.
Derivation of this 'furnish usage efficiency' can be performed in two
ways, neither very satisfactory. The first involves a straightforward com-
parison of the weight of pulp and loading fed in with the weight of paper
sold. Although this obviously fluctuates from week to week as more or
less broke remains in storage or in the system for processing, over a fair
period the average should give a reasonable measure of furnish usage
efficiency. Note, incidentally, that weight of broke used in any particular
furnish is irrelevant to this calculation. The main source of inaccuracy is
that normally the weight of bales and their moisture content is known only
from a sample average, and similar problems apply to loading batches.
Also some allowance must be made for moisture present in the finished
paper, yet this fluctuates especially when sold unconditioned.
Such a calculation is not feasible anyway when a mill produces its own
pulp, unless this is dried and stored before use in which case it is possible
to add up the various beater charges and allow for moisture content from
sample determinations. An alternative method is to make a direct deter-
mination of the fibre and loading passing to drain. This has been discussed
in TAPPI Information Sheet No. 598.01. Provided all waste is channelled
to a single drain it is quite possible to sample from this manually at
periodic intervals or to use an automatic sampling device which every few
minutes draws a small volume into a storage jar. Subsequent analysis of the
mixture of samples yields an average value for the consistency and loading
content of the drain water, and this together with a flow average gives a
tolerable estimate of losses (3). This can then be compared to the saleable
paper produced and expressed as a furnish usage efficiency.
569
6A. 3 PRODUCTION CONTROL
6A. 3 MONITORING QUALITY

The first concern of the papermaker is to secure maximum output from his
machines. Following a close second is the need to maintain quality of the
paper produced. In many cases production and quality are very closely
linked, for improvement of the one is difficult to secure without deteriora-
tion of the other. The whole subject of quality has in fact become far more
important than ever before, and now demands a considerable proportion
of the papermaker's time. Various general aspects of this will now be
discussed.
6A . 3 1 The scope of quality control
The need to apply some sort of standard to the grades of paper produced
on a machine has been evident for a long time, and from early days it has
been customary to extract samples from the end and occasionally from the
middle of machine rolls in order to check that substance is close to what
is required. It has also been traditional to issue to the machine crew a
sample either of the previous making or of sheets laid down as standard for
the grade concerned, and expect this to be matched as closely as possible.
The rising quality demanded by customers and increasingly competitive
nature of the paper trade has exposed the inadequacies of this simple
system so that nowadays it is general practice to apply a more detailed
examination of paper at the reel-up.
The main object of this examination is to indicate when quality is not
up to the standard required so that action can be taken by the machine
crews to correct this. The term 'quality control' has come into general use
in the industry to describe this operation, though a more accurate term
would be 'process control'. The distinction is a fine one, because 'process
control' implies that the purpose is solely to indicate the need for action in
order to reduce variation to a minimum and keep as closely as possible to
the standard required. 'Quality control' on the other hand implies that
some sort of inspection and rejection of the product takes place before it is
passed to the customer. This does not happen when straightforward
'process control' is applied, though naturally it contributes to that end.
To distinguish more carefully requires using the term 'process control' to
cover end-of-reel sampling and testing, while the term 'product control' is
often used to denote an inspection and rejection procedure. In deference to
general usage, however, it is proposed to label these functions respectively
'quality control on the machine' and 'rejection scheme'.
Although initial emphasis has been on the need to apply quality control
on the machine, and this indeed is the main topic of the present section,
there is increasing realization that this alone is insufficient to ensure an
adequate product. There must also be a general application of quality
control throughout the mill, particularly to incoming raw materials. Pulp,
loadings and chemicals should all be subjected to routine testing against a
specification. Also quality control needs to be applied in all areas of the
mill where variation can affect the final product, though generally the
application of instrumentation and conventional process controllers are
taking over this role. Most important of all, the status and organization of
570
the department responsible for quality control must be such that it is
regarded as making an essential contribution; not an unproductive liability
but an integral part of the papermaking process.
It is not proposed to discuss these points further, but to turn now to
what must be considered the key to successful monitoring of quality, the
scheme for checking at the end of the machine. The mechanics of this will
be discussed later, for the moment attention is turned to the fundamental
question: what aspects of quality should be brought into the scheme?
In the first flush of enthusiasm for quality control it was not uncommon
to find literally dozens of tests being applied to samples drawn from the
end of machine rolls. Each grade would have standards applied for as many
tests as it was physically possible to accomplish in a reasonable time, and
wherever possible limits would be applied within which the machine crew
was expected to keep. Occasionally a customer might make a complaint
about some specific property, so the matter would be dealt with by intro-
ducing a further test. After a few makings had established a typical level
for results of this test, it would join the others with limits set, for good
measure, well towards the high side.
This sort of procedure commended itself to management because it
seemed that here was an excellent means of keeping up quality levels.
Unfortunately it soon became apparent that nothing of the sort happened.
The papermaker would find not only that he was supposed to watch too
many qualities but that frequently altering one to bring it within limits
would cause another to go out of limits. Some properties would appear to
set themselves at a particular level at the start of a making and no con-
ceivable alteration to operating conditions would yield an improvement.
In time nobody would take any notice of half the tests performed, and the
inevitable acceptance of this by management would affect the attitude of
the machine operatives and jeopardize the value of the remaining tests.
The basic error in all this was lack of realization on the part of the
enthusiasts introducing quality control of just how empirical the process of
papermaking still is. There was, whether consciously or not, the expecta-
tion that the average operative possesses the wisdom of knowing precisely
which action influences what, combined with the speed of a computer to
determine how much alteration to each variable is needed to achieve the
best compromise. In practice it is useless to apply quality control to any
property unless a definite corrective action within the power of the machine
crews is available. When control of some property is vital and au obvious
corrective action is not known, then research and experimentation on the
machine is needed to fill the gap in knowledge. Likewise, when a particular
corrective action influences more than one property, the extent of the
interaction should be tabulated for the operative's guidance. When two or
more properties are obviously closely linked, only one of them should in
any case be subjected to quality control.
Unless these simple points are followed the machine crews will inevitably
find they have an impossible task on their hands. One or more specific
properties will be impossible to control or raise to the required standard
and after repeated attempts this becomes accepted, with the danger that in
571
time really vital properties come to be regarded in the same way. Quality
control is likely to be successful only when it is possible to demand and
receive a valid reason for a property being persistently outside the levels
laid down, and then to devise a means of preventing a recurrence.
Properties that fall into this category are mainly the basic ones control of
which is fairly straightforward. These include: substance, moisture con-
tent, thickness, smoothness, loading content and the simpler strength
properties such as burst and standard average tensile. Certain other
properties are also readily controlled depending on the type of paper being
made and include: porosity, colour and brightness, gloss (of coated papers),
wet strength, crush resistance, and water penetration for sizing. In each of
these cases there are more or less standard testing instruments available
which have to conform to a detailed specification. Equally important, a
controlled corrective action to the paper machine or the preparation
system can readily be made to alter the test value of the paper produced to
a new level.
Properties that should be avoided for quality control purposes include
those for which test methods are vague or their significance doubtful, and
those which are basically traditional and assess attributes of paper that are
difficult or impossible to measure. Above all, properties that relate in some
dubious way to the end-use of the paper and for which ingenious tests have
been devised should never be applied to quality control; if in some mis-
guided moment the apparatus concerned has been purchased as a gesture
of co-operation with the customer, then it should not leave the hands of the
mill laboratory. Examples of the properties which fall into these categories
include: folding endurance, softness, handle, twist-resistance, water-
vapour permeability, water absorption, abrasion resistance, and hygro-
expansivity. Each has its place in an investigation, but if the test concerned
is all-important to the end-use of the paper then either the relation between
it and some of the earlier tests mentioned should be determined, or some
ready means of correcting the level of the property by altering operating
conditions on the paper machine must be devised. There remains other
tests, tear, printability, curl, opacity, formation, and so on that fall some-
where in between the two extremes, and the position with regard to these
must be considered somewhat open at the present time.
6A. 3 2 Applying quality control
Once the properties to come under quality control on paper machine rolls
have been decided, there are a number of further important aspects to
consider. Broadly speaking these cover the mechanics of the quality con-
trol scheme and involve such questions as how and in what manner samples
of paper are to be obtained, how these are to be tested, what limits are to be
placed on the results, and what action is the machine crew then expected to
take? These are subjects that have received a great deal of attention in
numerous articles and one or two books and will be discussed here only
briefly.
The manner in which a sample of paper is obtained from the top of a
machine roll must be related to the type of properties to be tested in the
572
quality control laboratory and the number of individual test operations
required for each property. For each property a compromise must be
reached in the number of individual test results to be performed which best
satisfies the need for accuracy of the average with the time available for
carrying out the operations. The value of quality test results diminishes
with the length of time they need to execute and normally this issue settles
how many individual tests are made. It is, for example, a simple matter to
make several thickness and substance tests but each separate measurement
of tensile strength, ash content, and folding endurance is time consuming.
A systematic approach is possible and has been outlined in the British
Paper and Board Makers' Association Technical Section Proposed
Procedure No. 47: Guide to Sampling from Rolls for Quality Control.
This discusses the requirements of sampling and the practical problems
involved, and also illustrates with examples how the number of individual
test readings for properties with different variability can be decided. The
influence of cross-machine variability on the sampling procedure is also
discussed.
Once a testing procedure has been in operation for some time, average
results for a succession of machine rolls become available and the question
of setting limits on either side of the desired mean value arises. The
purpose of these limits is to give a guide to the machine crew when action
is needed to prevent the property in question straying too far from the
required mean, and this then has the effect of keeping variation in that
property to a minimum. For this task the actual value of the limits is not of
crucial importance so long as they assist in defining a set course of action
for the operatives to follow. Numerous ways of determining suitable limits
have been devised (see, for example, reference 2), and many sophisticated
systems are available. But for papermaking it appears unnecessary and
undesirable to attempt anything too elaborate.
Probably the most straightforward method is in the first instance to set
limits for each property on the basis of the average variation that has
occurred in the past during a number of separate makings. This will
include all the variations inherent in the system (which by definition it is
not within the power of the machine crew to reduce), plus variations
deliberately introduced in the course of making alterations to the machine
for one reason or another. Ideally the latter should be discounted, but
except where a gross change of level in the figures is obvious this is not
possible. Statistical analysis of the results of past makings will permit
calculation of the average variation in the form of the standard deviation,
a. Limits imposed are usually plus and minus 2 a and it is then intended
that these serve as a warning (when the property value falls outside the
limits) that something may have upset the level of the property and a
correction is needed. Further limits of plus and minus 3 a are also some-
times used to indicate when corrective action is absolutely essential.
Machine crews are usually advised not to touch the process when test
readings fall within the 2 a limits, though in practice trends from previous
results are sometimes discernible and an alteration may be made earlier.
In the case of some properties the only requirement will be for either an
573
upper or a lower limit, not for both. It is then necessary to decide whether
to use the appropriate single limit or nonetheless impose upper and lower
limits. This decision depends mainly on the importance of the property. If
it is largely irrelevant except to meet a specification, then a single limit can
be used. But if there is the likelihood of incurring higher making costs by
permitting the property concerned to move too far in value away from the
mean, then obviously double limits are needed. The best example of this is
in the matter of strength, where the end use will often require only a
minimum value, but a maximum limit is also needed so that furnish or
power costs are kept to the minimum necessary.
Once corrective action for some property is clearly required, the opera-
tive has to decide how much alteration to make to the appropriate machine
variable. This procedure is commonly a haphazard one in which an
operative adjusts a valve or alters the position of an indicator by an amount
that he adjudges necessary. He will invariably aim at re-setting the property
value exactly to the required standard, i.e. in the middle of the limits, but if
he is at all over-enthusiastic the degree of correction applied will be too
great and the test result off the following roll will show that a counter-
correction is needed. Avoidance of hunting in this way is very important,
and one method that has been advocated on theoretical grounds is to
provide two target values to aim at when making a correction. One of
these values would lie above the required mean and is intended for when a
result has appeared above the upper limit, and the other below the mean for
when a result is under the lower limit (11). Whenever possible, however,
there should be some guidance provided on valves and indicators to show
how much movement produces a given change in property value. This,
unfortunately, is rarely possible unless a detailed log is kept of all altera-
tions made on a machine, though with the advent of closed-loop on-line
computer control this type of operation will become at one jump complete-
ly automatic.
6A . 3 3 Setting quality standards
The care needed in selection of which tests are used for quality control has
already been mentioned. Equally important is the value of each property
chosen to be the standard for a particular grade. In certain cases this is
clearly settled by trade agreement or British Standard requirements, for
example the substance of many grades, the loading content of printing
papers, electrical insulation of condenser paper, and so on. Several other
properties will, from custom, require to be kept at certain values, but with
many there will be a tendency when first deciding on a stand<:ird to
examine the scatter of averages for past makings and select a value on the
high side. This is done in the belief that a higher burst, finish, thickness, or
what have you is synonymous with higher quality. In some cases this may
be so, but not in all, and before a high level is agreed on other aspects must
be considered. For example, if a higher strength than the former average is
chosen, is this going to lead to higher making costs? Also, would a lower
strength not be sufficient for the customer's needs anyway, or have there
been complaints in the past of low strength? A further point regarding
574
standards is that there is never any virtue in requiring, for the sake of
argument, a particular porosity level for a fine writing when this is both
irrelevant to the end use and of no particular significance in determining
the basic characteristics of the paper.
When a customer suddenly starts requiring a definite standard for
certain properties the matter can become rather complex. In the first place
he is almost certainly going to ask for a tighter specification than he needs.
And what he needs will invariably be based on what he has received
previously. So the manufacturer is faced with having to make to a standard
which equals or exceeds the best he has achieved in the past. In some cases
this could well necessitate greater making costs to reach consistently the
levels required. When this is the case, obviously careful negotiation is
necessary and a polite but firm enquiry into the real needs of the customer
has to be made.
Particularly invidious is the situation where the customer demands that
the degree of variation in the test results of a particular property is reduced
to a lower level than hitherto. It is no use the manufacturer simply reducing
the width of quality control limits applicable to the property concerned,
because it is not within the operative's power to reduce this variation at
will. Faced with the imposition of narrower limits, there may well be an
initial improvement until the operatives come to feel they face an im-
possible task as readings out of limits appear with monotonous regularity.
In time the application of this particular control will fall into abuse and the
limits become meaningless. Whenever making conditions are known to
have a natural variation which is too wide to meet a customer's specifica-
tion, then either the papermaking process has to be modified in some way
to permit the specification to be achieved, or the specification limits must
be widened to a more realistic value. It is often the case, however, that
tight specifications can be met provided a lower machine speed is run. As
this almost certainly implies lower production, it is important that
management is made aware of the necessity and agrees to the situation.
Brauns (16) has presented some interesting reflections on the whole
question of quality levels for competitive grades of paper. He draws
attention to the arbitrariness with which it is first decided to test certain
properties, and how a level is agreed for these properties which depends
mainly on what can be achieved in relation to competitors and which can
often be largely irrelevant to the end-use requirements. As a result of the
establishment of particular levels for various properties, the quality from
different manufacturers will gradually become similar, but whether this is
in fact the best compromise between the manufacturing cost and the
genuine needs a customer has it is impossible to tell. Eventually one
manufacturer will cut the level of a property and reduce his price. The new
level will be found by the customer to make no obvious difference, so he
will demand a similar cut in price from competitors who may take some
time to discover that the necessary economy can be achieved by lowering
quality standards.
Inevitably the needs of the customer must influence quality standards,
and quite rightly so. For this reason a careful record of complaints by
575
customers should form an integral part of a quality control department.

To aid tabulation, the use of a standard form for representatives to fill in is
useful, though a tendency to draw attention to complaints simply because
they are listed as possibilities must be avoided. It is preferable that a
complaint originates specifically from a customer than that he is questioned
as to whether he has experienced difficulty from this or that. Some com-
plaints will always be made, and the manufacturer who has conscientiously
made paper to an agreed specification will nevertheless find the customer
producing trivial complaints of damage or of some difficulty in use which
may often be due to the customer's own negligence anyway. Such circum-
stances then become more a question of business relations than a serious
technical matter.
6A . 3 4 Organization of quality control
Successful operation of quality control depends much more on efficient
organization and documentation than on the precise calculation and
frequent revision of limits. The whole procedure of sampling, testing, and
recording must be rigidly defined, and the action required in the event of
different circumstances occurring has to be clearly laid down. It is not
intended to deal with this aspect of quality control in any detail as there
are many excellent articles (see, for example, reference 18), which discuss
the merits of different organizational structures for quality control and
give examples of the sort of forms that can be used for standard cards,
recording test results, summary sheets, control charts, etc. Discussion will
be confined to a few of the more important points that in the opinion of
the author need careful attention.
Perhaps the most important single assurance that a quality control
scheme works well is that a sound attitude to it is adopted by management
and production supervisors. Continual vigilance is needed and whenever
quality falls an immediate enquiry should automatically follow. Many
mills have daily meetings during which quality deficiencies occurring over
the previous twenty-four hours are discussed and decisions taken regarding
rejected batches. Above all it is essential that the machine crews whose job
it is to operate the quality control scheme are made aware of the serious-
ness with which bad and persistent infringements of the limits are regarded.
This is not to say that in this event ~n operative must be blamed for slack-
ness, but that a genuine enquiry is made into the reasons for low quality
and when this is no fault of the operator (as is frequently the case) a
willingness is shown to get to the bottom of the matter and prevent a
recurrence. On such occasions careful questioning of machine crews often
brings to light surprising explanations for poor quality and closer in-
vestigation shows up the need for some alteration to equipment or
operating procedure.
The manner of presentation of quality control results to machine crews
is important and it is generally agreed that some form of simple graphical
method is preferable. Graphs on paper can be kept for record purposes in
the quality control laboratory, and one or other of the patented systems of
presentation on magnetic or pegboards used in the mill. For each property
576
the limits must be clearly shown, together with the last few results. Usually
some means of drawing attention to the results for properties which look to
be going out of control is also incorporated. Speed of operation is essential
and it is useful to method study the whole procedure of testing to devise the
best routine and lay out for the laboratory; whenever practicable rapid
methods of testing should be used, leaving more rigorous techniques to be
carried out on humidified samples for standardizing purposes. Transmis-
sion of results from the laboratory to the machine must be by some quick
and accurate method, and in some mills it has been found useful to use
close-circuit television for this purpose.
It is especially important that the quality control laboratory is, and is
seen to be, a model of efficiency. The machine crews must be able to rely
implicitly on the accuracy of the results presented to them and there should
never be suspicion that they may be in error. Once crews begin to question
the validity of any results, the whole climate of working is liable to change
for the worse. it is advantageous to rotate the shifts of the laboratory
testers in the opposite direction to those of the machine crews. To ensure
accuracy of test results it is essential that all instruments used in the
laboratory are subjected to frequent comprehensive checks following the
standard methods laid down by national organizations and in some cases
by the International Standards Organization. Any instrument suspected of
bias must be immediately replaced or taken temporarily out of use. A
record of repairs and alterations to each instrument is worthwhile keeping.
To reduce operator error a good standard of training is required and
detailed operating procedures for each test should be drawn up. The
importance must be stressed of recording only results that are actually
determined, and never results that at the time may seem more realistic or
likely; the value of honesty in this matter cannot be too lightly dismissed
for it is inevitable that pressures are brought to bear on testing operatives
to produce figures that do not show too wide a deviation from those
required. At the same time the possibility of error in the laboratory can
never be discounted and it is useful to have a recognized procedure for
re-testing and, if necessary, re-sampling whenever a single test result or an
average is very different from previous values. For individual test results a
number of statistically valid techniques giving an objective criterion for
rejection of the result are available, and one or other of these should be
adopted in preference to leaving such a decision to the discretion of the
tester.
The results accumulated by the quality control laboratory are useful not
only for their immediate purpose on the paper machine but for providing
longer-term information on quality. For each making, or over a period of a
week, it is a simple matter to determine the average value attained by each
of the properties tested. Some laboratories also determine the standard
deviation but this involves considerable calculation which, for the addition-
al information provided, is of doubtful value unless the process is perform-
ed by a computer. A simpler method giving a ready indication of the
degree of variation within a long making is to determine the percentage of
results which appeared outside 2 a upper and lower limits. Graphs of the
577
average weekly quality levels are useful for indicating trends, and when a
number of machines making similar grades from the same furnish are
involved such graphs can on occasion serve to eliminate possible reasons
for a drop in some property because the same trends are discernible on
different machines. Errors due to sudden bias in an instrument can also be
occasionally detected. At three or six monthly intervals overall property
averages for certain common qualities provide useful information to
present alongside downtime and production data. Occasionally all
standards should be reviewed with regard to the average level and the
limits imposed and alterations made to bring them into line with recent
performance.
These are a few of the more important considerations that make for good
quality control practice. They should help to ensure that the cost of run-
ning a laboratory is amply paid for and that the paper leaving a mill has as
good and uniform a quality as can be economically provided. In time it is
to be expected that most aspects of quality control will gradually be
replaced by the use of continuous testing instruments at the end of the
machine roll, but in many instances it may well be found that the cost of
investing in sophisticated equipment of this type together with the qualified
supervision needed for it would be higher than the savings in manual work
involved in the usual form of testing.
6A. 3 5 Rejection procedure
It is valuable for the machine crews, particularly the dryermen, to feel that
the work of the quality control laboratory is genuinely helping them in
their own task. A simple means of encouraging the right attitude is to
involve the dryerman in the quality control operation. This is necessary in
only a small way, as for example if the dryerman is given the job of
signalling to the laboratory a short time before a roll is ready or is made
responsible for marking the results up on a chart or board. In some mills
the dryerman is also made responsible for drawing the sample for testing
and provided this is performed in a carefully defined manner from the top
of the roll it is unlikely to be abused. Many mills, indeed, like to encourage
dryermen to do as much as possible of the testing themselves, arguing that
it should be within their sphere of responsibility for producing a satis-
factory product. This procedure does not, however, find very general
acceptance partly because it tends to load more work on to the dryerman
than seems justifiable, but mainly because of a suspicion that the job
would not be fairly performed. Nevertheless in some circumstances, and
especially when there is independent and subsequent inspection testing, it
can be an acceptable procedure.
The most useful incentive to comply closely with quality control results
occurs when they are associated with some sort of rejection scheme. In this
case the test results are used not just in an endeavour to minimize variation
and keep quality closer to the required standard, but also as part of an
inspection procedure. In other words product control is incorporated and
the whole system becomes a quality control in fact as well as name.
Frequently, as in the case where sales are in reel form, there will be no
578
further testing after the end of the machine to check for matching to a
specification, and so a rejection procedure associated with the conven-
tional form of quality control becomes essential.
Rejection limits used on tests taken at the end of the machine are
frequently fixed at plus and minus three standard deviations from the
mean. When a tight specification has to be met, narrower limits can be
necessary but this is an undesirable state of affairs which could involve
excessive waste. It is probably of some benefit to incorporate rejection
limits on all properties, whether or not the paper concerned is liable in each
case for rejection from being out of specification. The mill is then applying
its own inspection standards and not simply having them imposed. With
properly determined statistical limits application of rejection limits at this
level should not lead to a much greater loss of saleable paper and can be
advantageous in securing more reasonable specifications with new cus-
tomers.
When paper is destined to be inunediately slit and cut in the salle, it
would be usual to have some sort of inspection procedure operating on all
individual sheets or more conunonly on a random sample of the sheets.
Reels with some property outside a rejection limit are then often simply put
to one side pending a decision from management on their future. Further
testing may be required to check the original value, and if the property is
extremely important there may be no alternative but to strip the reel, but
often it will be possible to sell it as retree or seconds, or to reduce it to a
lower grade. The appropriate action depends on how comprehensive the
sampling testing procedure is in the salle. If there is a well designed scheme
incorporating statistically designed acceptance/rejection limits it could
well be cheaper to let a low-quality reel pass through the system and hope
to recover as much good paper as possible from it.
When rolls off the paper machine pass directly to a finishing operation or
to a winder for slitting to size and packing, then a rejection procedure must
operate immediately whenever tests from a sample at the top of the roll
indicate that a property lies outside the rejection limits. Different methods
are in use, but a straightforward approach would be to hold the roll
temporarily and sample again for a check. Ideally this should cover all the
usual properties tested because when one property is suddenly well out of
specification a clue as to the reason for this can sometimes be derived from
a comparison of other property values, but usually shortage of time will
preclude the checking of any properties other than the one most concerned.
If the average of the original test and the check are inside the rejection
limit, the roll can then be passed. But if rejection is confirmed, there can
be standing instructions requiring immediate stripping of a layer 2 to 3
inches deep followed by further sampling and checking, and so on until a
value within the limit is reached. The main trouble with this procedure is
the delay it creates in the steady flow of rolls to the next process, but there
should normally be sufficient slack in the system to permit this. If a roll has
already gone forward there is little that can be done \.Vithout interrupting
the process except eventually to draw a sample from wherever the roll has
unwound to, and then await results from there.
579
When a property is out of rejection limits at the top of a machine roll,

it must also be out of limits at the start of the new roll. Hence whenever a
property is reported out of rejection limits immediate action is required to
make a correction. So long as there is some confirmatory evidence on the
machine, correction should be made before a re-check of the property has
been obtained. It is now necessary to tear a slab from the machine roll,
even if machine speed makes a break inevitable as a result, as soon as it is
likely that the property has been brought back within limits. Several
samples from within the roll may be necessary before this is finally achiev-
ed, and when the roll is later unwound all the paper below where the final
sample was drawn usually has to be torn up. A report on the total quantity
of paper rejected and the reason for this must, of course, be submitted by
the supervisor or machine operative and it is useful to compile a weekly
summary of these losses.
Such a rejection procedure may well seem a bit drastic, but in a well-
ordered mill with properly determined limits it is not likely to be used very
often. If, for instance, plus or minus 3 a limits are used for rejection, then
an average of one roll in a little under 500 would appear to require rejection
when in fact this was not the case (checking would then in practically all
cases confirm that the roll was alright). This is a small price in wasted
effort to pay for guarding against the odd occasion when a genuine and
undetected disturbance to the machine does cause a property to stray in
value well away from what is required. In any event, not every rejection
limit need operate as a true criterion for rejection when it is not part of the
specification required by the customer, and if some specific property is
persistently beyond a rejection limit the time will have come for a close
investigation into the reasons for this and into the value of the limit itself.
580
CHAPTER 6B
IMPROVING MACHINE PERFORMANCE
6B .1 GENERAL SURVEY
In a production system like papermaking there is inevitably over the years
a steady decline in the performance of equipment due to gradual wear and
tear. At the same time bad operating habits can gradually creep into the
work of machine crews and poor quality, once accepted, can slowly
become the norm. Supervisors, with the help of engineering and technical
personnel, attempt to arrest this insidious process of deterioriation by
training and encouraging operatives to improve their standards and by
maintaining equipment well and periodically renewing it. Throughout the
mill, supervisors and operatives alike should be on the look out for ways of
making improvements and any suggestions should be taken up as a matter
of course once they can be shown to have economic merit.
Certain specific aspects of this continual quest for improved performance
will be dealt with later. For the moment it is intended in the present section
to examine, in the broadest possible sense, the various ways in which losses
in production can occur and how these may be overcome. Not every mill
will suffer from the same sources of inefficiency, but the basic principles
should still apply. Attention is naturally concentrated on the Fourdrinier
machine itself, though some consideration is given also to operation in
other parts of the paper mill, especially when these impinge on production.
6B . 1 1 Improving working procedures
There are a number of situations on a paper machine where an improve-
ment in working procedure can serve usefully to reduce downtime and
broke. This applies particularly to start-up, shut down, grade changes,
clothing changes, and shuts for routine maintenance.
Starting up a machine is a frequent enough occurrence yet the sequence
of operations carried out is likely to vary appreciably from one machine
crew to the next. Once it is accepted that a machine requires so many hours
to get under way, and that for quite some time afterwards the paper is
anyway fit only to be stripped for broke, then inevitably this becomes the
normal standard. The very familiarity of the operation makes it difficult
for operatives and supervisors to see any fault in their customary proce-
dure, and when something unusual happens (more often than not to the
quality of the paper) it often seems on the surface that nothing has been
done which is different to the usual method employed, and so there is no
obvious explanation.
An efficient start-up depends primarily on two features: proper design of
the system and adequate training of the crews. Of these, the first is more
important because ideally a system should be so arranged that it is
impossible to go seriously wrong when starting it up. For new paper
581
6B.l l PRODUCTION CONTROL
machines the techniques of systems engineering can be brought to bear in

such matters as deciding the size of chests and pits, the critical points to
apply control, and the precise sequence of operations to minimize dis-
turbances. On existing machines a careful analysis of the customary start-
up procedure will often indicate obvious sources of delay. For instance, it
might be that the fresh stuff is ready well in advance of when the machine-
man can use it due to the length of time needed to heat up drying cylinders
or fresh water in the backwater silo. Provision of a separate steam supply
line for start-up would then be an obvious remedy.
Closely connected with this aspect of an efficient start-up is the time
required to reach equilibrium conditions once stock is on the wire. It is
evident that when fresh water is used for the first dilution of fresh stuff,
some time will be needed before re-circulation builds up the fines content to
a level near that eventually reached. But usually this in itself does not
cause any delay because the machineman has many points to inspect
before feeding across from the couch and by then equilibrium conditions
will have been all but established. Subsequent delays in feeding through
presses and dryers can often be due to inadequate preparation (wetting out
of press felts, decreasing dryer temperature at an appropriate time to
prevent overdrying the web). It is, however, in less obvious attributes of
consistency, colour, pH, and so on that troublesome disturbances can arise
before equilibrium conditions are achieved; for instance, whitewater
circulated through a save-all and returned to the stock preparation system
can introduce a variation in these characteristics that needs many hours to
reach equilibrium. Often, too, inadequate care is taken to try to ensure that
the first batch of fresh stuff is supplied at the correct pH and temperature,
or that the water used to dilute it approximates in these respects to what
will be used once recirculation is established. All these considerations
contribute to the usual situation where it is accepted that feeding through
on the first occasion always produces numerous breaks, and when the
sheet is finally up the first paper appearing at the reel-up is in any case not
up to standard for one reason or another.
The need for adequate training of crews in starting-up procedure is self
evident, but this is assisted by providing each operative with a check-list of
points to go through at each start-up. This should cover not only routine
inspection and cleaning (as outlined earlier in the relevant section on each
part of the machine) but also such matters as pre-setting of all machine
variables (valve positions, draws, loads, etc.) to suit their normal levels
for the particular grade to be run. The duties and area of responsibility of
each operative should also be carefully defined whenever ambiguity is
possible. Where it is particularly important that a definite sequence of
events is followed (to prevent overflow, too high stock consistency, excessive
power demand, etc.) it is preferable that mechanical or electrical interlock-
ing devices are incorporated in the appropriate actuating mechanisms.
Much the same points apply also to shut-down and grade changing
procedures. In both these cases accurate estimation of the time of com-
pletion of an order or a machine roll contributes to reducing wastage at
the end. The sequence of operations when changing grades by running
582
PRODUCTION CONTROL 6B.l 2
through from one order to the next should be carefully planned to reduce
the quantity of paper made in the transition period. This is not difficult to
arrange when only a substance change is required (see the discussion on
this point in lC. 44), but for a colour change much greater care is needed to
prevent a lengthy period of gradual change on the roll before the final
match is achieved.
In some cases of colour change it may be economically preferable to shut
the machine, drain the system and start up again, even though it would be
possible to run through. This is necessary anyway for some colour changes
where a thorough wash-up is needed. The procedure for handling this
situation should be defined to keep the time involved to a minimum, and
particularly when it is a frequent occurrence study of the customary routine
could well show up defects in the machine design which causes delays (pits
too slow to drain or fill, inadequate hose pressure, etc.).
Scheduled shuts for maintenance come more into the province of the
engineering department. It is evident that careful planning of a system of
preventive maintenance with staggered examination of different items over
a period of weeks will contribute to keeping the time required to a mini-
mum. The production personnel should be adequately informed of work to
be carried out on the machine so that they can plan their own jobs to avoid
too many men working in one area or requiring at the same time such
equipment as the overhead crane.
Clothing changes can be a particularly fruitful source of lost time and in
this regard emphasis has already been placed on the necessity for accurate
and comprehensive records and the desirability in many cases of adopting
a scheduled changing of wires and felts. The actual procedure for changing
clothing is frequently made unnecessarily lengthy by inadequate prepara-
tion and the tendency to gather too many men round in the belief that this
hastens the process. The routine for changing each piece of clothing should
be carefully thought out, and is worth close examination using method
study techniques. In arranging a wire or press part considerations of
clothing change form a basic requirement affecting the engineering design,
and close thought is always necessary to make certain that no inherent
difficulties are likely. But even so it is inevitable that there are some snags
on any machine and it is in overcoming these by provision of special
changing equipment and brackets for fixing to the machine that some
ingenuity is required. When the mill organization permits it, it is preferable
for the clothing changes to be carried out by the same gang of men who
specialize in this. When clothing changes are, as is usually the case, the
responsibility of the machine crew who happen to be on shift, then the
quality of supervisory personnel, careful designation of individual duties,
and the training given to crews are most important for ensuring efficiency.
6B . 1 2 Improving operational efficiency
However well-trained the crews and however carefully defined the working
procedures, loss of production and waste still occur due to inadequacies in
the equipment and in the general machine conditions. With regard to
equipment, the isolation and recognition of the causes of downtime on the
583
6B.l 2 PRODUCTION CONTROL
machine, as discussed in 6A. 1 2, assist in pinpointing the trouble spots, but

the aim should be to prevent these occurring in the first place. Poor making
conditions are manifested in quality deficiencies leading to out-of-specifica-
tion paper and broke and even contributing directly to downtime.
Breakdown of equipment can be completely unpredictable, but on the
other hand there is no doubt that a sound preventive maintenance and
inspection programme dramatically reduces the likelihood of it occurring.
Gradual deterioration of performance causes trouble that is more difficult
to avoid (especially when it appears as heavier vibration and increased
variation of drive output) and can be very hard to detect when it becomes
responsible for introducing weaknesses in the paper web that create breaks
in another part of the machine. A long-term aid in reducing the severity of
this general problem lies in the keeping of careful and comprehensive
records of each piece of equipment. These should include not only obvious
details of date of purchase, spares kept, lubrication and servicing, location
of drawings, and so on, but also a careful tabulation of repairs made and
any modifications introduced. This sort of information becomes particular-
ly invaluable whenever a quick repair is needed.
Paper machines, in comparison with other pieces of industrial equip-
ment, last a very long time and during their lifetime are likely to undergo
numerous alterations and increases in running speed. This is inevitably a
serious source of trouble because at any one time there are likely to be
several parts of a machine seriously overstrained and approaching the time
when their renewal is imperative. When this eventually takes place, and
whenever a major modification is made to introduce a new lease of life to
an old piece of equipment, a record of this change should be kept in a
special book devoted to the machine as a whole. This general record of
major alterations can be very useful in identifying changes in operating
conditions which subsequently become noticeable in the long-term data
provided from machine tests, breaks and production analysis, and quality
records. In this connection data going back many years can occasionally be
found to give an important clue, and investigations are assisted by having
all these various records available in a similar form and averaged when
appropriate over the same three or six monthly periods. During enquiries
it is frequently found that instrument and other charts relating to machine
conditions would be helpful but these have long since been jettisoned
because of the space they take up. A useful compromise is to arrange to
retain charts and detailed records covering, say, a couple of weeks during
two characteristic periods of each year. It is then possible to get a general
picture of the operation at different times in the past and extract data that
might not hitherto have been thought worth keeping.
Occasionally it is a useful exercise to make a general analysis of a
particular paper machine system for comparison with other machines.
This is a task which is necessary when any general reconstruction is under
consideration and can be most valuable if carried out in conjunction with
one of the major machine suppliers who have a wider range of experience
in the use of equipment for making different .classes of paper at different
speeds. Such an analysis would involve a detailed examination of the
584
PRODUCTION CONTROL 6B.l 2
function and condition of each piece of equipment, and the present
performance in such matters as power and speed in relation to the original
specification. At such times it would also be appropriate to draw on down-
time records for information and to determine the economic possibility of
renewal (see 6B. 2 5). Especially with the stock preparation system, a
detailed flow diagram and an analysis including power and manpower
requirements and the flow of work through the building is also useful for an
overall examination.
Quality deficiencies are generally caused by excessive process variation
as a result of poor control of the system and the lack of defined operational
procedures. Several specific aspects of this have already been discussed, for
example in relation to starting-up, shutting and changing grades. Other
relevant matters such as the control of cross-web variation and obtaining
as high a moisture content at the reel-up as possible have also been dealt
with. It is obvious too that much of the effort directed at reducing down-
time from faulty equipment will have a beneficial side-effect in reducing
quality variation. Reduction of machine direction substance variation is an
important item in this context. This is partly because lower substance
fluctuations tend to reduce variation in practically every other property,
but also it becomes simpler to attain a mean substance closer to what is
required. This can present a substantial economy over a long period since
(in the face of high substance variation) there will inevitably be a tendency
to produce a sheet slightly on the heavy side both to make running easier
and other properties less likely to fall below standard.
There are several other general points relating to improvement of
operational efficiency that are worth mentioning. Excessive trim loss can
become too easily tolerated and this represents an appreciable wastage.
Modern approaches to trim reduction by applying linear programming
techniques are well worth investigation though there are still a great many
problems in their general application. Sometimes the problem of trim
waste is aggravated by too great a variation in the deckle of rolls coming
off the machine, a result primarily of changes in shrinkage rate in the
dryers. When this cannot be reduced sufficiently, consideration should be
given to closer control and the use of expander rolls.
It is customary to strip a certain amount of paper from the top of each
roll because the outer few layers can: become saturated with water or
unevenly wound and creased during the change. Unfortunately this routine
readily seems to arouse some primitive stimulation from using a knife and
many more layers are stripped off than is necessary. This common habit
can come to represent a substantial wastage of good paper and it is worth-
while giving consideration to providing a special knife with a form of
runner or some device which prevents cutting below a relatively small
depth. Whenever a single shallow cut is insufficient to remove damaged
surface paper, it is not troublesome to strip off a further layer.
The need to keep a check on the overall fibre wastage from a machine
system has been touched upon, and this is one source of loss that can often
be reduced quite simply with a little thought. The reduction of fresh water
entering the system in pump glands, sprays, seals and so on will bring about
585
6B.2 PRODUCTION CONTROL
a proportionate reduction in the loss of fibre in the overflow. To be

permanent it is desirable that water rotameters are provided in key posi-
tions and instructions given as to setting the flow.
Dirt in its various forms is usually responsible for some loss of produc-
tion, either because it creates a break or spoils the appearance of the paper,
and much of the trouble caused by this comes from poor housekeeping and
inadequate cleaning. Insistence on a reasonable standard of cleanliness and
tidiness in the mill not only contributes to reducing this loss but should
also be psychologically beneficial, a matter too often underestimated in
importance. In this respect the mill management can also assist by such
small but valuable items as arranging frequent painting of walls and
machines, keeping working conditions tolerably comfortable, tidying up
areas of the mill which have become disused, and other obvious ways
relating to the welfare and comfort of machine crews.
6B.2 FINANCIAL CONSIDERATIONS

Decisions to replace equipment or to run a new grade have aspects which
are obviously financial as well as technical. To a lesser but nonetheless
important extent, the running of a machine faster or with a different
furnish or with less broke is not just a matter of technical feasibility but has
an important bearing on the economics of the process. This is the subject
of the present section: how, to put it simply, is it possible to put a value on
making a change in the system?
Proper cost data for a paper mill, and particularly for the machine itself,
is essential for giving a realistic estimate of the economic potential of
making changes to operating methods and equipment, and of running new
grades. The same data assists immeasurably in assessing the comparative
worth of different proposals and it can also help to indicate areas and
operations which are ripe for improvement in efficiency. The most common
form that such cost data takes is known as standard costing and this is
discussed first, followed by examples of how the standard cost figures
derived can then be used to put a value on different operational changes.
Of these, the installation of a new piece of equipment requires particularly
careful costing and this is given special attention.
6B . 2 1 Standard costing
The basic principle of the standard costing system is a simple one. It
involves first the splitting of all costs to various centres and the relating of
these to the total cost of producing a ton of each grade of paper. These
costs are collected over, say, a six monthly period to give a reasonable
average for each item and thereby take account of fluctuations from week
to week in machine efficiency, maintenance requirements, accidents to
clothing, and so on. The figures derived over this period are then taken as
the basis for predicting costs for the coming period, with due allowance
made for known variations which are expected to occur (for example, an
increase in pulp prices or running at a higher machine speed as a result of
some alteration just taking place). These figures then represent the standard
586
PRODUCTION CONTROL 68.21
costs for that period and their main use is to act as a standard against
which to compare actual costs incurred, the difference (or variance) figures
being used to indicate the need for investigation in the event of actual costs
being appreciably different from those anticipated. Standard cost figures
are also used for a variety of other purposes such as budgeting, planning
capital expenditure, and (of specific interest in the present context) for
permitting more accurate estimates to be made of the value of operational
changes in running the machine.
The most difficult step in working this system of costing is to decide on
the different cost recovery centres (i.e. functions costed separately) and the
extent to which relatively small variations in costs between different grades
or finishes are taken into account. In what follows it is proposed to give an
illustration of the sort of breakdown that can be adopted for the most
general case: a multi-machine mill with its own pulp preparation and
finishing plant. In deriving this, information and suggestions from several
sources have been used, notably references (4), (5) and (7).
The first point to clarify is the distinction between fixed and variable
costs and this is done most simply by giving examples. Fixed costs are, in
effect, overheads that are unchanged whatever grade of paper is made, and
indeed whether paper is made at all (at least over a reasonably short
period). They include the cost of such items as depreciation on buildings
and equipment; rent, insurance and rates; general office and sales expenses
and salaries; laboratory and other technical services; ancillary services
such as security, first aid, welfare, sports and canteen; direct labour for
machine crews and associated operatives in the preparation and finishing
plants; packing, storage, handling and transport facilities; and material and
labour involved in machine maintenance and repairs including depreciation
of engineering equipment and stores charges. Other fixed costs will include
similar charges for units such as bleaching plant or coaters used by a
proportion of the grades of paper made in the mill and chargeable only to
these grades.
Variable costs are all those charges which have a direct relationship to
the quantity and grade of paper actually produced. They comprise costs
for raw material, including chemicals and loadings; steam, power and
watrr supplies to pulp and paper preparation plant and the machine
(though see page 591); coating, size and other materials used in after~
processes; non-returnable centres, wrapping and packing materials; and
despatch and carriage charges.
The relation between the fixed and variable cost contribution for a
particular grade affects the value of making different improvements to the
system. For instance, if variable costs account for the larger proportion of
the total costs, then reduction in charges for such things as furnish and
finishing can have a very important significance. By contrast, when fixed
costs are relatively high an increase in production can be more valuable
because this spreads the overheads over a greater tonnage. These points
will be considered in greater detail below.
Costs derived from individual centres can be used to build up a picture
of the change in cost of a product as it passes through the mill. The
587
original raw material cost is increased to a slightly higher cost by the time
it enters the pulp preparation plant due to the storage and handling it has
undergone. When pulping is complete, further costs have been involved in
labour, supplies, services, maintenance, depreciation, etc., and this con-
tinues through the stock preparation plant, the paper machine itself, and
finishing. Each part of the process involves further cost and so, in effect,
adds to the value of the product at that point. This is the 'added value'
concept. It might better be termed 'added cost' because the value that can
be attached to something depends on market conditions.
The 'added value' can even be found for such processes as formation on
the wire, pressing, drying and on-machine sizing, as well as for distinct
operations, like supercalendering or off-machine coating. This involves
taking certain arbitrary decisions regarding such matters as the splitting of
maintenance and supply costs, but is nonetheless a straightforward exercise
if the necessary data is available. For instance, Holt (5) performed this
calculation 'at considerable labour' for several machines in a pulp and
paper mill and found that the relative costs were greatest, surprisingly, for
pressing ($14,58 per ton), then formation ($12,36 per ton) and then drying
($11,94 per ton). The main reason for the relatively high pressing cost was
due to high supplies (mainly felts) and services, though the fixed charges
based on original capital cost were also high.
A different type of cost analysis of the Fourdrinier has been made by
Barbour and Tweedie (10) who presented figures showing that typical
operating costs per hour (steam and power only) to run the dryers of a
machine are far higher than similar costs for the wire, which in turn are
higher than the presses. Related to the quantity of water removed the
order of costs involved changed to dryers highest, then presses, then the
wire part.
Other figures presented by these authors show that the initial capital cost
of a machine per foot of length varies little from one section to the next, so
as dryers are the longest once again costs are highest for this part of the
machine. Holt also looked at the capital cost of a machine from the point
of view of deckle, and reached the conclusion that there can be an optimum
width above which it is uneconomical to go.
The value of data such as this is more useful for comparison with other
machines than in the costing of an individual machine where it is not
particularly relevant to know the individual costs of separate parts of the
process when these are completely interdependent. Compared with other
machines, however, it may be possible to gain an indication of where
technical improvement could most usefully be sought, though ih this
respect comparison of downtime and performance figures would usually be
more relevant and simpler to compile.
6B . 2 2 Cost centres
The division of costs into fixed and variable is not always so easy as it
might at first appear. When there are several machines in the same mill
sharing the same services and perhaps the same pulping and finishing
plant, costing machines separately presents some problems. It is appro-
588
PRODUCTION CONTROL 6B.22
priate at this point to consider one or two aspects of the derivation of cost
figures in more detail.
In the first place, choice of cost centres has to be made very carefully.
The precise selection depends entirely on the system, but as a general
principle it is important to be able to cost separately such distinct opera-
tions as supercalendering, coating, dyeing, laboratory services, instrument
department, in fact any part of the mill which makes a clearly-defined
contribution of one sort or another. Excessive analysis has, of course, to be
avoided, and the possible use of each individual cost in controlling future
expenditure, costing new lines, and in assessing the value of modifying the
process must always be borne in mind. In the manufacturing plant itself,
there is little point in costing separately any part of the process that cannot
be by-passed, provided data relevant to operating conditions (power and
steam usage, maintenance time, etc.) can be obtained for analysis if
necessary.
The different costs involved can be considered under several headings,
and it is usual to take the paper machine itself as the main basis. There are,
firstly, the fixed charges on the building and equipment including rent,
insurance and rates. Where two machines share the same building, com-
bined costs would usually be allocated according to area. Where pulp
preparation, paper preparation, and finishing plant are shared, cost-
splitting to different machines and grades (of all fixed charges including
those discussed below) can be on a time utilised or tonnage basis, whichever
is most appropriate.
Depreciation on capital cost of manufacturing equipment presents some
accounting problems because if the practice is followed of writing off a
fairly high perc~ntage in the first year, then on paper a new machine
will often make a heavy loss. One method suggested (7) is to have nil
depreciation in the first year, 21- per cent. for the next two years, and only
then rise to the normal tax-allowance levels of 7t per cent. and i 0 per
cent. of value. However, the precise manner and rate at which capital is
discounted is a problem for the accountant, and as current government
legislation allows an immediate refund in tax relief of between 20 and 45
per cent. of the capital expense it is now possible to allow greater deprecia-
tion in the first year.
To these basic fixed charges are added all the operating overheads from
each of the service departments: engineering, electrical, instrument,
general office, sales, personnel, transport, laboratory, work study, and so
on, together with the various ancillary services. These overheads comprise
depreciation of buildings, labour costs, stores, and general running costs for
each unit. The total of these can be regarded as common to the mill, and
must then be apportioned to the various units of the manufacturing plant
as an addition to the basic fixed charges. The way this is done depends on
many factors including tonnage and value handled, number of operatives
involved, capital cost and age of equipment and so on; a certain degree of
arbitrariness is bound to creep in and decisions of this type must depend
largely on the detail available. For instance, in addition to individual
overheads and direct labour costs, it should be possible to charge main-
589
tenance costs involving labour, time and materials to each manufacturing

unit separately (coating, alum plant, bleaching, sizing, etc.), so that proper
account can be taken of the true cost of these facilities when they are used
for particular grades. Under this heading would be included clothing costs
for the machine and other items requiring regular renewal such as Sheehan
ropes and suction box covers. Specialized equipment such a roll-grinding
machine should really only be charged to the units making use of the
facility.
So much for the fixed charges. With these an accounting period of one
year a1lows the inevitable weekly cost variations caused by sudden
repairs or replacements of an expensive piece of clothing to be smoothed
out. The fixed charges for each machine can be used to produce an hourly
standard running cost, on the assumption of a known number of working
hours each week.
Variable costs are principally those for raw materials, supplies of various
kinds, and materials used in finishing and despatch. Water may also be
included as a variable cost though variation in this from one grade to
another would not usually be of significance, so it is simpler to treat water
costs as a fixed charge. Raw material covers pulp costs, either as the
price of bales or of raw materials plus the separate cost of pulping,
together with other additives such as alum, resins, dyes, loadings and other
chemicals for bleaching, wet-strengthening, pitch dispersion, and so on, the
cost for each of these taking into account the cost of preparation. Materials
used in finishing and despatch include coatings, sizes, wrappings, etc. It is
evident that reasonable accounting of variable costs is only possible if
accurate figures are known for consumption of all these materials. For
instance, it is important, especially in multi-machine mills, that appropriate
meters are available on such common lines as those supplying steam, liquid
alum, and other continuous additives to the machines. Power used in the
preparation plant, on the machine, and in finishing processes must also be
known.
The main variable cost is, of course, the pulp itself and certain difficulties
can arise in accounting for consumption of this. Generally speaking, when
bales of pulp are used it is necessary to weigh each of these individually,
classified according to type, and correct for a previously determined
average moisture content. When pulp is prepared in the mill measuring the
quantity passed to the paper machine for a particular making is not so easy
unless it is the custom to dry the pulp and store it in steeps. In this case
determination of moisture content of the pulp and checks on the weight of
loads taken to the beaters can be used to give a reasonably accurate
measure. Otherwise resort has to be made to flow measuring devices and
consistency measurements, or to an accurate determination of fibre
drained to waste.
Broke also presents a problem, especially when it is kept for long periods
for use in specific makings. It is possible to take account of broke made
during a making and offset the value of this, but usually the quantity of
broke it is possible to use in the furnish cannot alter much from one making
to the next and it is possible to discount it completely from the costing point
590
of view. Over a number of similar makings, variations in the quantity of
broke made and re-used should average out sufficiently to be absorbed
with the other variations that occur. Only when broke is switched from one
machine to another or re-used for different grades is it essential to keep an
accurate account of the tonnage involved and to attach a definite value to it.
When a mill produces most of its own power and steam requirements
costing of these, even when usage is accurately known, requires careful
thought. The overheads and running costs of the boiler plant can be
apportioned to the machines and other manufacturing units in accordance
with either the actual process steam usage or the amount that has to be
available when required (a sort of standing charge), or a combination of the
two. Similarly the cost of steam engine or turbine plant can be apportioned
according either to actual power usage or to the provision necessary, or
both. These costs can be presented as fixed charges to the manufacturing
unit based on average usage over the accounting period. The coal used
in the boilers is not quite so simple to deal with because the steam it
provides is in many mills used for both power generation and process
heating. A reasonable compromise is then to split the coal cost on an
average heat drop basis, i.e. in accordance with the amount of energy used
up for the two functions, and arrive at separate charges for process steam
and power steam usage. The full cost for out-of-balance steam would be
charged to the function responsible, e.g. to the engines or turbines when
it is necessary to blow off some steam not required for process heating. In a
mill which does not generate its own power, running costs are obtained
directly: the coal consumption for process steam and the local electricity
board charges for power. These costs for process steam and power can be
added into the other fixed charges, but when steam and power consump-
tion vary considerably from one grade to the next, it is obviously preferable
for them to be treated as semi-variable charges, thereby giving more
accurate data for arriving at comparative costs of different grades. Provided
a mill possesses reasonably accurate steam and power distribution records,
there is no difficulty in this and it also accords more with reality for even
when a machine is shut it is then still charged with overheads incurred in
the boiler and turbine plant but not with the steam and power it does not
use.
When different makings follow one another in quick succession on a
machine it is not practicable to keep separate account of supplies used for
each individual making. Normally makings of similar grade and substance
would be grouped together, and variable costs assumed the same for each,
account being taken only over the complete succession of makings in the
same group. Provided it is reasonable to assume there is no significant
change in such features as the refining, pH, retention on the wire, or the
steam consumption per lb. paper, this practice is perfectly satisfactory.
The same principle can apply equally to grades differing only by small
degrees of pulp composition, loading content or dye addition, because the
different costs involved in changes of this type are readily accounted for.
From records of the average tonnage made per hour, the fixed charges
(in which costs are per hour) and the variable charges (in which costs are
591
68.2 3 PRODUCTION CONTROL
normally recorded per ton) can be brought together for each separate
grade. An overall picture thus emerges showing precisely how the cost of
each grade is built up. When these costs are compared one grade with
another discrepancies will inevitably appear. Sometimes these differences
between similar grades are perfectly legitimate, as for instance when a
particular making requires a prior wash-up so downtime for this reduces the
machine time efficiency. In other cases differences may be more difficult
to account for but nonetheless fair, for example when a slight alteration of
furnish leads to an increase in trouble from pitch. But it is quite possible
also for discrepancies to arise by accident, as for example when a bias is
introduced because a particular grade is always run first after starting up
(and is therefore more liable to inefficient operation) or because another
grade is always run towards the end of the week when there are liable over
a period to be more stoppages for felt washing or wire repairing. The
different costs thrown up in such cases unfairly load the grades concerned
and should really be discounted, but often it is very difficult to be certain
when differences are in fact caused by such matters. Other cost differences
arise when certain grades are repeatedly run only for short makings due to
a greater proportional waste of pulp and loading (in cleaning out after the
making) and of time spent at the start of the making in achieving correct
quality. Such differences are, however, a perfectly fair reflection of the true
making costs. These aspects of machine operation will be investigated in
greater detail in 6B. 3 where the various differences that can appear between
the operational efficiency of one grade compared to another are examined.
68 . 2 3 Costing improvements to the paper machine
It is now appropriate to consider some of the ways in which standard cost
figures can be useful for costing improvements and alterations to the paper
machine. With certain alterations, such as change in furnish or coating
mix, or a reduction in labour force, it is comparatively straightforward to
calculate their influence on the overall cost of the end-product. But in other
instances, such as those involving increased machine speed, reduced broke
or downtime,' or increased selling moisture content, a little care is needed
to evaluate the true effect on the costing figures.
To illustrate these points it is proposed first to present some purely
hypothetical cost figures which, for simplicity, are taken to represent the
standard costs in a mill generating its own steam and power and making an
uncoated and machine-finished grade of paper prepared from bales of pulp
and despatched direct in reel form. This gives a rather higher relative
furnish cost than would be usual in many mills but this means that the
economic effect of the improvements examined is not exaggerated and can
in fact be regarded as closer to the minimum relevant for the particular
selling price concerned. (In practice costs would be available for a variety
of grades and a weighted average of these might be taken based on the
anticipated production requirements on the machine for each grade.)
It is assumed that the machine output of saleable paper related to the
total available hours is 2 tons per hour, that broke amounts to 13-33 per
cent. of production off the machine, that the standard week (i.e. the time
592
available for making paper after a reasonable allowance for starting-up and
shutting) is 130 hours, and that average downtime on the machine amounts
to 10 hours per week. This in fact means that the machine produces paper
at the rate of 2· 5 tons per hour for the equivalent of 120 hours of the week
giving a total output of 300 tons of which 13· 33 per cent. or 40 tons is
broke and the remainder, 260 tons, is saleable. Standard costs are:
Standard Costs
£per ton £per hour
Mill overheads (offices, sales, insurance, services, rates, etc.) 6 12
Depreciation of equipment and buildings 5 10
Direct labour 4 8
Maintenance, repairs, clothing 4 8
Boiler and turbine plant 3 6
Total fixed charges £22 £44
Pulp .. 50 100
Loadings and chemicals . . .. 8 16
Finishing, wrapping, storage and distribution, etc. 7 14
Coal for process steam and power 3 6
Total variable charges £68 £136
Manufacturing costs (from above) .. 90 180
Profit 10 20
Selling price .. £100 £200
With these figures it is now possible to investigate the financial effect of
various improvements that can be made to the operating conditions. First,
what is the situation when an increase of 1 per cent. in speed is achieved?
If other conditions are unchanged, i.e. there is no alteration in downtime
nor in broke ratio, and also no additional maintenance and clothing
needed as a result of running faster, this yields an additional 1 per cent. of
260 or 2· 6 tons per week of saleable paper. Now it must be assumed at this
point that this additional output is in fact saleable, in other words that
order books are full and demand exceeds (or with a sales drive can be made
to exceed) current capacity. This is a central point in all the cost considera-
tions that follow, except for that applying to increased moisture in the
paper, because in all other cases an actual improvement in output occurs.
In the circumstances when capacity exceeds demand a higher output may
simply mean shutting a machine earlier in the week with little resultant
savings. If this is the present situation, or is predicted for the near future,
then obviously any plans based on raising production would be very much
less likely to prove fruitful.
A straight production increase of one ton in the week will give an
additional profit equal to the difference between the selling price and the
sum of variable expenses. This is obvious because all the fixed charges have
already been absorbed in the calculation of standard costs based on current
output. The variable costs represent what has been termed the marginal
cost for small production· increases, in other words the additional cost
incurred for each extra ton produced. The smaller the proportion of
variable costs in production to the selling price, the greater the additional
593
profit from increasing speeds. In the example, the additional profit per ton
is £(100 - 68) = £32, and for a 1 per cent. increase in speed or 2· 6 tons
per week, this will amount to 2· 6 x £32 = £83· 2 per week. This represents
a 3· 2 per cent. increase in the total profits of £2,600 per week.
A similar case is presented when downtime is reduced. This also has the
effect of increasing output, but the additional profit from each extra ton is
very slightly higher than when brought about by a speed increase. The
reason for this is that when a machine is down a certain amount of steam
and power is consumed to keep drying cylinders warm and turn different
sections over. This might amount to a coal usage of about one-third the
normal consumption when producing paper. Consequently, the marginal
cost for an additional ton of paper is in this case equal to the variable costs
(£68) less the coal cost during downtime (one-third of £3), giving £67.
Additional profit per ton is thus £(100 - 67) = £33. A 10 per cent. reduc-
tion in downtime from 10 to 9 hours per week will make available 2 tons of
paper and so increase profit by 2 x £33 £66 per week.
Reduction in broke from improved deckle distribution, more consistent
quality, and so on has a different effect on profits. In this case the saving
(other conditions being assumed unchanged) which results from having a
ton of saleable paper instead of a ton of broke is the cost of reprocessing
the broke to make it saleable. This is, in effect, the total cost of producing
the paper less the cost of the furnish which the broke is substituted for, i.e.
£50 for the pulp plus £8 for loading and chemicals. The cost of producing
the paper is the total manufacturing cost (£90) less the finishing costs of £7
which have to be incurred anyway (unless the broke is from the salle in
which case it is actually less than £7). The saving is therefore £(90 7) -
£(50 + 8) = £25 per ton. Additional profit on the ton of saleable paper is
this saving plus the usual profit of £10, i.e. £35. A reduction of 5 per cent.
in the broke level from 40 to 38 tons will make available a further 2 tons of
paper which will increase profit by 2 x £35 = £70 per week.
This calculation assumes that quality of the broke is as good as the
original furnish and that the cost of breaking down the broke and adding
it to the system is not significantly different from the cost of breaking down
and treating bales of raw pulp. The latter supposition is not unreasonable
for the order of accuracy it is possible to achieve in this sort of calculation,
but the equivalence of the broke to the original furnish is unlikely. A
better approximation of the true saving from reducing broke is to assume
that it normally takes the place of the cheapest and weakest constituent of
the furnish, e.g. mechanical pulp in a printing or news grade, and requires
the addition of some loading to make up for retention losses. The broke,
in other words, is less valuable than the calculation above assumes, so the
saving will in fact be greater than £25 per ton.
As a final example, consider the effect of obtaining a O· 5 per cent.
increase in the value of moisture content of the saleable paper. This is an
exercise more relevant to paper sold in reel form, but to a lesser degree it is
also pertinent to paper sold in reams. An increase of O· 5 per cent. of
moisture in effect substitutes water for the same amount of furnish in the
paper. The saving, other conditions being equal, will thus be O· 5 per cent.
594
of £58 or £0· 29 per ton of saleable paper. There will also be a small
additional saving in steam consumption if the moisture increase is achieved
by drying to a higher value at the reel-up, but if it is added in later the
opposite applies. For a weekly output of 260 tons, the additional profit
from an increase of O· 5 per cent. moisture in the paper is therefore 260 x
£0·29 = £75·4 per week.
Summarising the four calculations above:
Improvement in operating conditions Additional profit per week
1 per cent. increase in speed .. £83-2
10 per cent. reduction in downtime £66
5 per cent. reduction in broke .. £70
0·5 per cent. additional moisture content £75'4
In other words an additional profit of about £100 per week will be
derived from each of the following alternatives:
(i) increase in speed of 1·2 per cent.
(ii) reduction in downtime by 15 per cent.
(iii) reduction in broke level by 7 per cent.
(iv) raising moisture content by O· 65 per cent. moisture.
It is instructive to ponder the implication of these figures, which are by
no means untypical. The value of a little extra water in the paper is
particularly revealing, and reflects the relatively high furnish cost which
causes any small reduction in furnish requirements to make a large con-
tribution to profits. A modest increase in speed and quite a sizeable
reduction in downtime, both of which are often not so easy to achieve,
make a similar contribution to profits. Such data as these should be
available for all paper machines and are obviously of crucial importance to
any plans for improving performance.
68 . 2 4 Installing new equipment
New equipment is installed for a variety of motives, sometimes purely for
its psychological impact on machine crews or for the benefit of a particu-
larly awkward customer. If the circumstances demand it, there is nothing
inherently wrong in this. But whatever the situation one thing should be
clear, and that is the cost of carrying out the operation. For any new piece
of equipment a clear picture of the financial implications should be drawn
up before the order is placed, and afterwards the accuracy of this picture
should be checked so that lessons are learnt from the inevitable mistakes
that occur.
Many things contribute to the successful introduction of a new piece of
machinery and can have a big impact on the eventual financial benefit of
the operation, so it is appropriate first to devote a few words of general
comment to this topic. It is of course obvious that full details of any new
equipment must be obtained before a decision can be taken to purchase it.
These details must cover points such as maintenance requirements,
servicing arrangements, commissioning agreements, services needed, supply
of all ancillary equipment, and a full technical specification. Design and
use of the equipment must be considered by the engineering and produc-
tion departments, and at an appropriate time electrical, instrument,
595
maintenance and other departments should be given the opportunity to

comment.
Where several different types and design of equipment can fulfil the
same role, a detailed comparison must be drawn up and many aspects
looked at apart from the simple ones of price and delivery. Inspection of
the equipment working in other mills is frequently of benefit if only
because a much clearer idea is obtained than comes from examination of
drawings and photographs. The paper industry is blessed by a particularly
co-operative attitude in these matters and the mutual benefit that can in the
course of time be reaped from allowing visitors is recognized by all but a
few insular (and almost always old-fashioned) mills. Information from
impartial sources such as the British Paper and Board Makers' Research
Association can also be invaluable at this stage. Occasionally it is possible
to borrow the equipment from the manufacturer and when doubt persists
this can be a useful way of postponing judgement until as much information
as possible is available.
Once a decision to purchase has been made, the preparation work and
other equipment that has to be ordered or fabricated at the mill should be
listed, and a programme devised for carrying out the whole operation. At
this point it might be beneficial if the job is large enough to decide on man-
power requirements and draw up a critical path analysis for carrying out
the installation. Similar considerations can be brought to the actual timing
and placing of orders for different parts of the ancillary equipment in order
to prevent tying down capital unnecessarily early. In this respect, however,
the potential loss from not being able to complete a job for lack of some
small and relatively inexpensive component generally makes it advisable to
leave an ample margin over the suppliers' estimated delivery times. As the
various items arrive at the mill rigorous inspection against drawings and
specifications is advisable. The position regarding spares must be con-
sidered and adequate safeguards taken to carry a suitable stock.
It is also important at this stage to take every opportunity to acquaint
other members of the production staff and machine crews with the im-
pending installation of new equipment. If necessary, time has to be
devoted out of normal shift hours to explaining the purpose and operation
of the installation. This is not only more likely to secure the eventual
interest and co-operation of the operators, but involves them in the project
at the outset and makes it clear they have some responsibility for its
success. When the time comes for perseverance and ingenuity in the face of
those problems which arise with any new equipment, the attitude of the
production personnel makes a considerable difference to the speed and
effectiveness with which a solution is found. Discussion of the new equip-
ment with operatives is also likely to produce points of view which had not
previously been thought of, and refinements may be introduced which
eventually save a great deal of time. It will also be valuable when compiling
for the benefit of supervisors and operatives a general information notice
and operating instructions-an absolutely essential task.
With any new piece of equipment some changes in operation of the
whole paper machine system and of the paper quality are to be expected
596
PRODUCTION CONTROL 68.2 5
and assessment of these is important. It is no use requiring tests to be taken
of appropriate properties when the equipment has gone in; this must
obviously be done beforehand as well, so the implications of the new
equipment must be discussed with the mill laboratory and a suitable
programme of tests devised. Generally speaking, a straightforward before
and after comparison does not present any problems in experimental
interpretation though care is needed when the values of some properties
are not at a static level but are subject to a steady trend resulting perhaps
from a continual programme to improve quality. In the latter case, so long
as information is available for a number of individual occasions, and this
is one advantage of comprehensive long-term records derived from com-
plete system checks, then graphs can be drawn showing property values
over the period spanning the installation and these should make it unlikely
that incorrect conclusions are drawn.
Once the equipment is installed comprehensive tests may be necessary
to ensure it is operating in the most beneficial way, and that performance is
up to specification. This may require a series of checks to be made with
operating variables at different levels in order to determine optimum
conditions. Analysis of variance techniques can then be applied to give a
correct statistical assessment of the results. Occasionally there will be
danger of causing a loss of production by carrying out this more formal
type of experiment and then other techniques of evaluation, such as
Evolutionary Operation and its offspring Simplex Evop, can be employed.
Failing all else, evaluation of the effect of running with different operating
conditions can be made by examining the accumulated data of ordinary
production runs. In this case multi-regressional analysis is a valuable
statistical tool which can pick out the most important variables from a
large number and assess their effect on the property in question. The
tedious calculations formerly needed before this method could be applied
are now obviated by the use of straightforward computer programmes.
Routines for maintenance, lubrication, servicing, operating tests and
general inspection of the new equipment by papermakers, engineers,
electricians and so on should be worked out to reduce to a minimum the
likelihood of lost production from future malfunctioning. Manufacturers
are usually only too willing to help with this task because trouble-free
operation is as much in their interests as the mill's. It is unfortunately too
easy to neglect new equipment once its initial novelty has worn off and it
appears to be giving a satisfactory performance. But a thorough post-
mortem is necessary some time not too long after the installation has been
satisfactorily completed, and this is an appropriate time to carry out
revision of operating maintenance instructions in the light of experience,
and note down modifications made to the original design. At the same time
a full report on the operational benefits and financial gain should be drawn
up to guide future installations.
6B. 2 5 Costing new equipment
Before any new piece of equipment is purchased a thorough assessment of
the financial implications should be carried out. This may be done in the
20 597
very early stages purely as a feasibility exercise to see if the project ought td
be taken further. In this case the estimated cost for the installation might be
based only on budget prices and subject to considerable error, but the
method of approach is still basically similar. If the exercise is performed
after the equipment is working satisfactorily, then a really complete and
accurate picture is obtained, but again the principal steps in evaluation are
the same. These will now be considered.
It is useful to divide equipment costs into three parts: the costs of
acquisition C, of operation 0, and of maintenance M. In each case it must
be understood that costs associated with any existing equipment which is
being replaced are allowed for. Thus, the scrap value of old equipment and
any spares should be subtracted from C, as should be current operating
and maintenance costs from 0 and M respectively.
The cost of acquisition C comprises the purchase price of the equipment,
plus any ancillary items that are required. Where new standby equipment
is purchased or existing equipment for another purpose is utilized, then a
part share of the cost of this determined on the basis of estimated use
should be added. Installation and commissioning costs are also included,
plus mill labour and materials used from the stores during erection.
Operation cost 0 comprises direct and indirect labour charges; power,
steam and other services; insurances and licences; and the cost of space
occupied by the equipment. Maintenance cost M comprises lubrication and
cleaning charges; an estimate of repairs; and the cost of keeping spare
parts in the stores. Costs for both 0 and Mare taken over a year.
A fourth and most important item is the yearly saving S which is found
or anticipated as a result of installing the new equipment. This can come
under many headings and be relatively simple or difficult to determine
depending on the complexity of the changes occurring or likely to occur.
The profits from increases in machine speed or over-all production,
reductions in downtime or broke, and increases in moisture content have
been detailed earlier in this section and are fairly straightforward. Putting a
value on improved cleanliness of the sheet, better strength characteristics,
and other quality benefits is an almost impossible task and depends entirely
on mill conditions and relationships with the customers concerned. There
are indeed circumstances when these particular factors outweigh all others
and make an apparent increase in production costs worthwhile sustaining.
The four values C, 0, M, and Scan now be used to assess profitability of
the new equipment. There are various ways of doing this and the usual one
adopted (at least when trying to assess the potential worth) is the simplest,
if only because the accuracy of some of the figures relating to expected
improvements are inevitably hazy and make more sophisticated analysis
rather pointless. Usually it is only possible to work out the profitability
taking a reasonable view of the likely improvements, and for this experi-
ence and an objective element of judgement are needed to avoid being over-
optimistic.
The easiest method is to determine the 'pay-back' period. This is quite
simply C/(S - 0 M), which represents the number of years that must
elapse before the initial capital outlay C has been repaid by the annual
598
PRODUCTION CONTROL 6B.2 5
profit of saving S less operating cost 0 and maintenance cost M. One year
would certainly be regarded as really excellent, two as very good, three as
reasonable, and four as nearing the limit of use of the formula for assessing
the value of short-term investments. This method has the supreme merit of
simplicity and is readily understood by operating personnel, even though
accountants would advocate that it is treated with great care because it
ignores interest, depreciation and other essential elements. However, as a
means of distinguishing between the merits of different schemes or
deciding the order of priority of different stages in a large project, it is
certainly perfectly adequate.
A more sophisticated estimate of profitability is the MAPI method
which defines the 'rate of return' on investment. In the first year this is
100[(8 - 0 - M) + (C1 - Co) - D]/Co
where C 0, C1 are respectively capital values when new and at the end of the
first year, and Dis the depreciation over the year. The term (C1 - C 0) is
included to take account of a rise in prices, otherwise the calculation
amounts to determining the ratio between profit (allowing for deprecia-
tion) and original cost, and this is clearly a measurement of the rate of
return.
An example will make these methods clear. Suppose we are considering
replacing with more modern equipment some old inefficient screens which
are always breaking down and running to capacity (occasionally limiting
production). The total cost of the new screens is £10,500 and sale value of
the old equipment is £500, so C or C0 = £10,000. Additional operating
cost amounts to the difference in power consumption which we will estimate
is 20 h.p. or about £1,000 a year (assuming 2d. a kilowatt, continual
working and 20 days shut a year). There is no change in direct labour costs
or floor space required. Maintenance cost should be less and this is
estimated as worth £500 per year. After one year it is reasonable to expect
that the new equipment will be about 5 per cent. dearer or cost £11,000 but
the old screens might then be only worth £300. Hence C 1 = £11,000 -
£300 = £10, 700. Depreciation on equipment of this type may be taken as
pretty heavy, say 20 per cent. over five years. Finally (and this will be sheer
guess-work) assume that an anticipated increase in machine speed coupled
with reduced downtime and a little less broke work out at being worth
£3,000 a year. We thus have C or C 0 = £10,000; 0 = £1,000; M =
- £500; C1 = £10, 700; D £2,000; and S = £3,000. Payback period is
therefore £10,000/(£3,000 - £1,000 + £500) or exactly four years. The
other formula gives a return of
100[(£3,000 - £1,000 + £500) + (£10,700 - £10,000) £2,000]/£10,000
= 12 per cent. This would probably represent a marginal operation. The
main merit in the second formula is that it takes account of changing
circumstances and can sometimes illustrate the disadvantages of not taking
any action. This is because allowance is made for the rising cost of new
equipment and the lessening value of what is at present in the mill.
As an alternative to both these methods, a complete profit analysis can
be undertaken assuming straight-line or more sophisticated rates of
599
depreciation over the years and allowing for tax concessions on capital
equipment, interest payments on borrowed money and other financial
implications of the operation. But this enters the accountant's sphere with
a vengeance and he would certainly prefer to be consulted for a profession-
al opinion.
6B.3 PERFORMANCE UNDER DIFFERENT OPERATING
CONDITIONS
Very few machines continually make the same grade and substance of
paper. On the vast majority a variety of grades are made during the course
of a week, indeed the extent of this variety is a good index of the age of the
machine. It is usual that a gradual progression takes place over the life of a
paper machine from a single product made at (for the period) high speeds,
to the time when the machine falters along making a succession of special
grades in ever-decreasing tonnages. However, this does not necessarily
imply growing inefficiency or smaller profit margins. Indeed, old machines
can often make substantial profits if a suitable grade is found; low tonnages
are no barrier to this, particularly as depreciation will be negligible unless
charged on a replacement basis. But perhaps in this connection the accent
should be on the word 'can', because profits founded on a basically
inefficient process are apt to be ephemeral. It only requires another com-
pany with a more modern and faster machine to set out their stall and the
source of profit vanishes.
In the foregoing sections various aspects relating to the improvement and
costing of machine performance have been discussed, particular regard
being taken to the general assessment of profit for any grade of paper run
on a machine. It is proposed now to tum to a more confined but nonethe-
less highly important question: for a single grade of paper how does the
profitability of running a paper machine depend on the substance and the
machine speed? The situation in which a machine makes one or perhaps
two grades of paper though a wide range of substances is very common,
but not all of these substances are likely to be equally profitable. Likewise,
if attention is confined to a single substance, it is not always best to run at
the fastest possible speed. Discussion of the factors which govern changes
in running costs at different substances and speeds, and identification of
the most profitable lines forms the subject of this section.
6B . 3 1 Machine conditions at different substances
As the substance of a particular grade of paper is raised, conditions under
which it is made on the same machine alter. These alterations stem from
two considerations. Firstly, limitations of the equipment force changes in
operating conditions, and secondly at higher substance even of the same
grade of paper certain properties become less critical so the specification
usually in effect becomes less tight.
The best example of these two points lies in the beating operation. As
substance is increased, an early limitation to production appears both in
the capacity of the preparation system in respect of beaters or refiners and
in the drainage capacity on the wire (which of course depends on the
600
PRODUCTION CONTROL 6B.3 l
degree of wetness of the stock). Accordingly, it is usual to find that the
fresh stock is not so heavily beaten which implies that less energy per ton
of paper is used in the beating process. At the same time the slower
drainage on the wire resulting from a combination of wet stock and heavy
substance forces a gradual compromise in the amount of backwater added
for dilution. The graphs presented by Mardon (9) and reproduced in
Figs. 6. 1, 6. 2, and 6. 3 illustrate these changes.
The effect of reduced beating per ton is clearly a lower development of
potential strength in the paper, together usually with a change in the value
of several other properties such as porosity. In other words, if standard
25 /o5 145
!5ue.sTANC£ lbs/loco ~r 1
Fig. 6. 1. Relation between substance and freeness of fresh stock (after Mardon)
hand sheets are made from the fresh stuff, for papers of higher substance
these are the physical changes which would be found. But as a direct result
of the rise in substance, strength, air resistance, and other properties of the
sheet will also have higher values without any change being made in the
degree of beating. So unless the particular grade demands a pro rata
increase with substance in these properties, which would be unusual, it is
quite possible to reduce beating and accept only a relatively small change
in the properties. Greater substances are frequently demanded purely for
handling and prestige purposes, and clearly when this is the case there is no
value in increasing such properties as strength to greater levels than the
basic minimum required.
Changes of this sort in the operating conditions can have an important
effect on the speed at which it is possible to run a machine when making a
particular substance. But apart from this it is obvious that there are several
other specific restrictions which can limit the speed. The limitation in any
particular circumstance may, for instance, be the main drive motor speed,
or it may be drying capacity, preparation plant, open breast box height,
601
6B.3 1 PRODUCTION CONTROL
drainage on the wire, open couch draw tension, or quite a number of

different factors. It is instructive to consider how these limitations for
different substances can affect the maximum possible machine speed.
A good analysis of this has been given by Chedzey (1) who was interested
basically in defining standards for a paper machine in order to apply work
study bonus techniques. Although this motive would not find much
favour these days, the methods of deriving theoretical limits are still highly
relevant both as a means of comparing actual speeds with those which
should be possible for different substances, and also for the present purpose ·
of determining potential profitability. Assessing the maximum potential of
a machine under different limitations can be fairly complex and it is not
5
>-- 4c
\!)
•tt
Ill
~ 5so
~
z
i
ii: 300
tU
Clt
~s1
..__2~s~~~~-4~~~~~~-~~·s--~--~-a~
1 a~
sue,sTANCE. l'cs/3000 .ft-~

Fig. 6. 2. Relation between refining energy consumption and substance (after Mardon)
proposed to go into the details here. As an example, Fig. 6. 4 reproduces
the calculated theoretical maximum speeds on a machine making mechan-
ical printings and wallpapers due to limitations in drying capacity. The
curve is derived from an assumption that drying capacity is constant, so
that substance x maximum speed must also be constant, but also takes
into account changes in moisture content of the web entering the dryers at
different substances. It is also known, however, that in the later dryers
evaporation rate decreases as the thickness of paper increases due to the
difficulty of conducting heat through a greater bulk of dried fibres; hence
it is likely that if extended far enough the curve should in fact show a
downturn at higher substances.
602
Fig. 6. 5 shows a composite graph derived by Chedzey for a wide
substance range in which two other limitations to machine speed, those of
breast box height and preparation plant are also included.
Variations in the maximum running speed as a result of these limitations
obviously mean that the maximum possible production level changes also.
Fig. 6. 6 shows a graph drawn up by the author to indicate how production
might in fact alter with substance over a fairly wide range. Being purely
hypothetical and illustrative no scales are shown, but the curve is derived
from Fig. 6. 5 and is labelled to identify the specific limitation in successive
"
>
\)
."',
zO•li
...tl
~ O·S
\.J
SS 15 .9S 115
Fig. 6. 3. Relation between breast box consistency and substance (after Mardon)
substance ranges. Here speed of the main drive motor is taken as governing
the upper speed limit (this being a more common reason than breast box
height), while allowance is also made for an important limitation pointed
out by Mardon (9), that of web strength at an open couch draw which could
necessitate reduction of speed at very low substances.
The fixed costs per ton will vary through the substance range approxi-
mately in accordance with the inverse of the actual tonnage produced. In
other words the graph in Fig. 6. 6 could, if turned upside down, represent
the relation between cost per ton and substance. But allowance has also to
be made for the fact that relative sorting and wrapping costs are dearer for
lower substances (because of the greater number of individual sheets cut
per ton). Further, at higher substances steam costs per ton are greater due
to the decreased evaporation rate, though possibly this effect would in
603
practice be offset by reductions in maintenance and power costs as a result

of the lower machine speeds associated with higher substances. Thus to
represent cost per ton a modification has to be made to the shape of the
production/substance curve in Fig. 6. 6 in such a way that the slopes in
ABC and to a lesser extent in DE are accentuated. Variable costs per ton
can be expected, to a first approximation, to remain relatively unchanged
unless furnish mix alters. Taking these points into account, the probable
relationship between cost per ton and substance is shown in Fig. 6. 7.
The relative importance of the increased costs associated with substances
at the ends of the making range depends entirely on the relation between
the fixed and variable costs. Nonetheless, it is evident that outside the range
2o 25 3o 4-S
SlJf)STANCE lbs. D.C.-480.
Fig. 6 .4. Theoretical limitation imposed on machine speed by drying capacity (after
Chedzey)
of substances where the curve forms a relative plateau of costs in CD,

operation of the machine is less economical. This is particularly prominent
at the lower end of the substance scale where speed is limited by couch
draw and the drive, but applies also at higher substances once the region is
reached where drying capacity controls the speed. Above all, the profitability
of running low substances requires careful attention. Just how far this
curve can be considered relevant to the majority of paper machines is
difficult to say, but it seems likely that a fair plateau of costs such as that in
CD does exist in some form on most machines, and awareness of the
substance range it covers would indicate where sales should be concen-
trated.
604
6B . 3 2 Machine conditions at different speeds
It. is commonly believed, by papermakers if not by accountams, that the
faster a machine is run and the higher the production, the more profitable
is the whole operation. This is in fact a complete if excusable fallacy, and
this final section will be devoted to explaining just why this is so.
The best way to illustrate this is by giving an example, and for this
purpose it is proposed to use as a basis the machine for which cost figures
were given on page 593. In that example, the basic conditions provided
a a.. THEQt:tETICA.\.. e,RE.A$T e,ol< LIMIT.
b b. ~flOCUlT\CA.1... PREPARATl:ON PI.A.NT \..\M\T.
c.c..Tt:u;oRETIGA.\.. ORYING l..\MIT.
O---~-r-~~.,.-~--...~~....-~--.~~-.-~~...-~......-
0 lo 2o 5o 4o So oo 7o 8o
sue,,sTANCE- lbs. 0.C. --4eo.
Fig. 6. 5. Theoretical limitations imposed on machine speed by different restrictions
(after Chedzey)
for an output of two tons per hour saleable paper with an average of 10
hours a week downtime. The standard costs worked out at £90 per ton for
manufacturing, leaving £10 profit to make up the selling price of £100 per
ton. Of the manufacturing costs, £4 per ton was due to machine mainten-
ance, repairs, and clothing changes, and £3 per ton came from coal costs
for power and steam, making a total of £7 per ton for items which (as
discussed below) can be regarded as dependent on machine speed.
It will be supposed that these conditions refer to a machine speed in the
region of 1,200 feet per minute and for the purpose of the exercise the
relative conditions pertaining to a speed range from 1,000 to 1,250 feet per
minute will be investigated. The graphs that follow are of course purely
hypothetical but are nonetheless based on typical machine conditions with
allowances made for effects that are likely to influence the results.
Firstly, Fig. 6. 8 shows how, other things being equal, downtime increases
with machine speed. It is a familiar phenomenon that at higher speeds
605
breaks occur more frequently and it becomes more difficult to feed straight
through after a break. This is of course a feature of most pieces of
machinery, the harder they are pushed the more often breakdown occurs.
Such seems particularly to be the case when a paper machine has been run
for many years at one speed and then an attempt is made to run faster. In
such circumstances even a small increase beyond the familiar speed can
often be sufficient to cause bearings to overheat, felts to come over the side,
journals to fracture, gear wheels to split, and so on. But even when a range
of speeds is commonly run to cope with a variety of grades, it is usual for
z
0
....
v l..W\\Tl"44 CoNOlTlo".
:::> A. f,: OPEN C:::ouc.te OA4W.
0
0 e. C.: .t\A1M Oawe.
a c o: P1u:.PAflA~101'( P~A...rt
0..
OE: DR'flNG CA.PAC:.n'Y
Su&STANC.E.
Fig. 6. 6. Graph illustrating how machine production could be limited by different

operating conditions
there to be a small but steady rise in downtime as speed is increased,

followed at some point by a rapidly increasing downtime as speeds
approaching the limit of the machine capability are reached.
The effect of increasing downtime at higher machine speeds is to reduce
the rate of increase in production. In addition it is to be expected that
difficulties will appear in maintaining uniform quality, for instance because
at higher speeds there will be wider fluctuations in the drive output.
Variation in quality results in an increasing broke percentage and this is a
further factor reducing the rate of increase in production. The relation
between production and machine speed might well therefore appear as in
Fig. 6. 9, showing a marked drop in the rate of rise towards the upper limit
of the speed.
It can indeed happen that a point is reached where increases in machine
speed are achieved at such cost in additional downtime and broke that total
production actually begins to decline. Should this occur it is obvious that
606
r.
0 c
~
Cl
Ill
0.
I-
V)
0
u
SUf>STANCE
Fig. 6. 7. Graph illustrating changes with substance in cost per ton of paper, based on
production curve in Fig. 6.6
&. --~~~-r-~~~-r---~~-.-~~~--~~~~
/,090
/lfACNINE SPeE:D. h: p,zr 111/11.
Fig. 6. 8. Relation between downtime and machine speed

607
running at speeds beyond the point of maximum production is of no value.

But apart from this, it can in fact be shown that for achieving optimum
profit it is in all likelihood necessary to run at a lower speed than the one
which gives maximum production.
This appears when account is taken of a further effect of running at
higher speeds. It is invariably the case that the cost per ton of paper of such
items as machine maintenance and repairs, clothing, and the cost of steam
and power all rise as speed is increased. There are several reasons for this.
Increased maintenance and repairs is a direct consequence of greater
a~'--~~-.-~~~-r--~~---~~~--~~--...
l,ooo {oSo ~/oo ~/$0 ~2.So.
MACHINE: SPEEO, ft.pv inin.

Fig. 6. 9. Relation between production and machine speed
downtime. Increased clothing costs occur because wear is relatively faster

at higher speeds and stronger more expensive felts may be needed; also
a sudden failure which necessitates an unscheduled shut becomes more
likely. A rise in steam consumption per ton of paper derives from the
greater heat loss associated with the higher steam temperatures necessary
to achieve the required drying rate. At the same time it is more than
likely that dryness of the web entering the cylinders gradually drops, and
it becomes necessary to dry the paper down to a lower moisture at the
reel-up to counteract a less uniform cross-web profile, both of which lead
to further increases in steam usage. Power demand can also rise faster than
machine speed due to such factors as the higher head needed at the mixing
pump and greater heat loss from bearings, which in turn means a rise in the
cost of power per ton of paper produced. The relation between the machine
608
speed and the cumulative cost per ton of these particular items is shown in
Fig. 6. 10. In addition, the cost per ton at higher speeds may well increase
due to the need for greater strength of the wet web, this being provided
only by a more expensive furnish mix and heavier beating.
The net effect on the total profit at each machine speed is shown in Fig.
6. 11. This has been calculated by taking the production in tons per hour
multiplied by the difference between selling price and variable costs, then
subtracting the fixed costs per hour. For example, at 1,100 feet per minute
production is 4,150 lbs. per hour and cost of steam power, etc., is £5· 8 per
ton, making variable costs £66-8 per ton (see cost figures on page 593).
Thus, profit is £(4,150/2,240) x (100 - 66·8) - 44 or £17·5 per hour. It
is clear that an optimum speed exists at which profit is greatest. This
1'
l &
j
a' 1
~
f
~
~
t
~
"
~
a..
~
5
~-
()
\J 4
11000 1,o5o l,loo l,tSo l,200 l,2.So
MACH\ME 6PU:o. fr. pet- Mi,,.
Fig. 6.10. Relation between machine speed and cost per ton for steam, power, clothing
and maintenance
point, in the example at a speed of about 1,200 feet per minute, is clearly
not coincident with the speed at which maximum production is obtained.
This may well seem surprising, but is purely a consequence of the fact that
the profit per ton diminishes at greater speed levels and this, associated
with a growing decline in the rate of rise of production at the higher levels
of machine speed, is sufficient to bring about the downturn in total profit.
The calculation of total profit, as opposed simply to total production or
profit per ton, is essential for producing a correct analysis of the effect of
different machine conditions. This point must always be borne in mind,
particularly when comparing performance at different machine speeds. A
good example of the ease with which it is possible to fall into error in this
matter is due to Holt (5) and is illustrated in Fig. 6.12. The curve ACDB in
this figure represents a hypothetical relation between the operating costs and
609
68.3 2 PRODUCTION CONTROL
production, so that the swing upwards reflects the rising costs associated
with higher production levels. The upper curve shows the cost per ton of
paper derived from curve ACDB and this has a minimum at the point P.
The production level at P is associated with point Con the production/cost
curve in such a way that OC is a tangent to the curve (total cost divided by
production is then a minimum). This point does not, however, indicate
where profit is greatest.
To find the production giving maximum profit it is necessary to draw a
line OAB representing the value of the paper produced. The total value is
based on the selling price and has a straight line relationship with the
~
:s 2
()
.s::
~
(L
"1 le.
t--
u.
()
ex 1'1
a..
J
~
i2 14
MAOHNE SPEE.D. fr. pe..,. Min.

Fig. 6 .11. Relation between machine speed and total profit
tonnage produced. Points A and B on the production curve define the

region in which a profit can be made; production levels below A or above
C result in an overall loss. The point of maximum profit is reached when
the difference between the operating cost and the value is greatest, i.e. at
point D when the tangent to the curve ACDB is parallel to OAB. Point D
is always at a higher production level than C where cost per ton is least, the
difference between the two being greater the higher the selling price and the
more curved the relation between production and operating cost.
The dependence of overall profit on the selling price in this example is
particularly important to note. It is evident from the graph that the lower
the selling price (and hence the less the gradient of OAB) the lower is the
production level leading to maximum profit. Once the production/cost
610
curve has been identified for different grades of paper run on a machine, it
will be clear that those grades carrying a lower profit margin per ton will
yield maximum profit when run at lower production levels. This is contrary
to popular belief that the lower the profit margin the greater the production
needs to be to make operation worthwhile. Such is the case only if produc-
tion levels of all grades are relatively low and fall in the region AC. If a
machine is consistently run close to point D for one particularly common
MACHlNE PQ!OOUCTION I DAY

Fig. 6.12. Relation between operating costs and production (after Holt)
grade, and it is then proposed to produce another similar grade at a lower

selling price, then maximum return is clearly provided at a lower production
level.
There is, fortunately, a final point of consideration for the papermaker
who may now begin to worry about matters that formerly seemed straight-
forward. The examples just presented have been chosen to highlight little-
known effects. In practice it is unlikely that on most machines the point of
greatest profit is very precisely defined. Conditions can change so rapidly
that it often cannot be possible to locate the position with any degree of
accuracy. Equally important is the sensitivity of the machine system to
small changes in speed. In Fig. 6. 11 it is evident that profit does not alter
significantly through a range of almost 60 ft. per minute, i.e. from around
1,170 to 1,230 ft. per minute. If the profit/speed curve were flatter, this
range would be much wider and precise location of the optimum speed
becomes far less critical.
Nonetheless such a curve as that shown in Fig. 6 .11, or its alternative
611
relating production to total operating costs, should ideally be derived for

the major grades run on any machine. For its very shape is useful to know
as this should affect how the machine is run. For instance, if there is no
pronounced peak in the profit/speed curve, obviously there is no particular
merit in running at the higher end of the speed range as opposed to the
lower end, if only because over a long period greater wear and tear on the
machine at the higher speed will begin to tell.
If running conditions indicate that the normal speed falls in the early
part of the curve where profit increases with rising speed and production,
then different considerations have to be taken into account. When a
machine has a limited order book, it could well in such circumstances pay
to run as fast as possible and then shut for a longer period. This state of
affairs is, however, not very desirable because it leads to an unstable
business: when orders are plentiful high profits can be accumulated, but
when they are limited profit is low. It is preferable that operation normally
falls in the region where quite large changes in speed and production do
little to affect the overall profit. When this is the case, it is possible for the
mill to tolerate appreciable swings in demand without feeling the pinch.
And this, after all, is most important to the peace of mind of the paper-
maker.
612
REFERENCES
(1) Chedzey, C. S.: 'Standard Outputs of Fourdrinier Machines.' Proc. Tech. Sec. 37,
3, 439. December, 1956.
(2) Hendry, I. F.: 'The Calculation of Quality Control Limits.' Paper Maker.
International number, p. 46. 1959.
(3) Muhonen, E. J.: 'The Principles of Loss Calculation in the Paper Industry.'
Paperi ja Puu. 41, 3, 97. March, 1959.
(4) Anon: 'Standard Costs in a Paperboard Mill.' Paper Maker and Paper Selling
80, 5-6, 7. 1961.
(5) Holt, S. G.: 'A Method of Economic Analysis of the Papermaking Process.'
TAPP! 44, 5, 325. May, 1961.
(6) Turner, C. 0. and Filleul, H. W.: 'Reporting on Paper Machine Utilisation.'
APPITA Proceedings 14, 4, 15. January, 1961.
(7) Williamson, F.: 'Mill Management and Costing.' World's Paper Trade Review
155, 16, 1423. April 20; p. 1643, May 4; p. 1837, May 18; p. 2045, June 1; p. 2247,
June 15; p. 2419, June 29, 1961.
(8) Gaudet, E. J.: 'Better Control of Broke and Downtime Costs.' Paper Trade
Journal 146, 43, 26. October 22, 1962.
(9) Mardon, J.: 'The Technical Audit Principles: Promulgation and Practice.' Paperi
ja Puu 45, 12, 689. December 31, 1963; 46, 4, 305. April 30, 1964; 46, 6-7, 403.
July 15, 1964.
(10) Barbour, D. R., and Tweedie, J.: 'Economical Design Considerations for Paper-
machines and their Alteration.' Paper Technology 5, 4, 377. August, 1964.
(11) Nuttall, G. H. and Ainge, D.S.: 'Quality Control: correcting a property outside
limits.' Paper Technology 5, 2, 121. April, 1964.
(12) Stevens, J. Jr. and Newell, S. R.: 'Waste Control in a Fine Paper Mill.' TAPP!
47, 7, 157A. July, 1964.
(13) Wells, L. C.: 'Around the Mills. No. 2 Broke.' What We Are Doing: Bulletin of
the B.P. & B.I.R.A., No. 37. December, 1964.
(14) Cummins, 0. J.: 'Instrumental Approach to Evaluating Paper Machine Operation
and Efficiency.' TAPPI 48, 8, 47A. August, 1965.
(15) Mardon, J. and Mardon, D. H.: 'The Techniques and Utility of Lost Time
Analysis.' TAPP! 48, 10, ?IA. October, 1965.
(16) Brauns, 0.: 'Commercial and Economic Viewpoints on Quality Assurance in the
Paper Industries.' Svensk Pap. 69, 14, 455. July 31, 1966.
(17) Cummins, 0. J.: 'Automatic Monitoring can improve Paper Machine Efficiencies.'
(18) Harding, H. A.: 'The Development of Quality Control in Paper Mills.' Paper
Maker. Part III: 151, 3, 73; Part IV: 151, 7, 85. March, July, 1966.
613
INDEX
Abrasive material (in stock) 2B.66; 3B.42

Accelerator hood see Hood, H. V.
Accra-nip roll 4C. 31 ; SC .44
Aeration of stock tA.16; tB.12; tB. 7; 2A.23
- reduction of lB.41; 1B.52;2B.12; 2B. 31; 2B. 63;
3B.10 2
Air-bleed suction press 4B.32; 4B.43
Air bubbles on wire 1B. 74; 3A.2; 38.52; 38.10 1;
3B.10 2
Air cap see Hood, H. V.
Air, moisture in SA.11; SA.3 12
see also Humidity, air
Air permeability see Porosity
Air resistance see Porosity
Air, temperature of SA.3 12; SC.11
Alum lC.12; tC.14; tC.41
Anisotropy 3A.43; 3B.14; 3B.3 et seq.; SA.34
see also Fibre orientation
Apron 3B. 10 1 ; 3B. 10 2; 3C. 23
- position of over breast roll tB.61; tB.64; 3B.ll; 38.14
see also Slice; Slice jet
Backwater system ll.1; lA.21; tA.23

stability of lA.24 et seq.; 2A.17; 28.13
Baffles, table roll 3B.23; 3B. 72
Baggallay box 3A.41; 3B. 63
Barring, in calender stacks SB.93
Barring, wire tB.82; 3B.22
Basis weight see Substance
Basket cleaner 2B.42
Beta ray gauge IC. 32 et seq.
see also Substance, control of
Blowing 4B.31
Breaker stack SB.10 5; SC.11
Break detectors 3C. 31 ; 4B. 92; SB. 10 3 ; 6A . 11
Breaks analysis see Downtime, analysis of
Breaks, couch 3A.43; 3A.44
Breast box 11.1; lB.5 et seq.
- approach system tB.4 et seq.; tB.81; tC.34
see also Pressurized breast box
- stock consistency tC.13
- - effect of foils on 3B.23
- effect on formation of 3A.24; 3B. 31; 3B.8
relation to drainage conditions of tA.23; tB.12; 3A.16; 3B. 21; 3B. 8
- - relation to substance of 6B.31
-stock flow lC.12
- - control of tB.31; tB.54; lC.14
- stock temperature lC.12; tC.21
- - effect on pressing of 4B.8
- - importance at start-up of tC.41; 6B.11
- relation to drainage conditions of lB.12; 3A.15; 3B.21; 3B.8
615
INDEX
Breast roll 3B.10 1; 3C.23

- discharge from lB.53; 3B. ll; 3B.12; 3B.14; 3C.13
- position relative to slice of lB.64; lC.2; 3B. ll
Broke analysis 6A.22; 6B.23
Broke, dry-end lC.14; lC.41; SB.10 3; SC.14;
6A.22
wet-end lC.14; lC.41; lC.44; 4B.92; 6A.22
Bulk of paper see Thickness of paper
Burst of paper see Strength properties of paper
Calcium chloride lB.13

Calenders SB. 11 ; SB. 9 et seq.
Calendering, effect of machine variables on SA.4 et seq.
- effect of web moisture on SA.42 et seq.
Calender rolls SB.91; SC.42
- camber correction of SC.44
-cooling of SB.92, SC.41; SC.43; SC.54
measuring temperature of SC.43
Camber bar 4A.26
Cambering, roll 4C.31;SB.81;SC.42
Cascade drying system SB.12; SC.11; SC.12; SC.14
Cavitation lA.15
Clarifier see Save-all
Cleaners 21.1; 2B. 1 et seq.; 2C. 1 et seq.
- assessing efficiency of 2A.3 et seq.
-cyclone 2A.21 et seq.; 2B.12; 2B.6 et seq.
- cylindrical 2A. 21 et seq. ; 2B. 5 et seq.
- theory of 2A.2 et seq.
Clothing records 3C. 22; 3C. 23 ; 6B. 11
Cockling 3A.21; 4C.33; SA.37; 5B.51
Colour 1C.41 ; 1C.44; 6B. 11
- two-sidedness of 3A.28; SA.12
Compressibility of paper SA.42; SA.43; SA.44
Computer, on-line lA.2; lC.13; lC.33; lC.44; SC.13;
6A.32
Condensate, removal of SB.14; SB. 3 et seq.; SC.24
-system SB.1 et seq.; SC.12
Conditioning SB.10 5; SC.32
- effect on quality of SA. 3 12
relation with paper moisture content SI.l; SA.43; SC.38
Consistency regulation lB.21; lC.14
Costing operational changes 6B.23; 6B.25
Costing, process 3C.21;6B.21;6B.22
Couching 3A.4 et seq.; 3B.6 et seq.
Couch roll, plain 3B. 61 ; 3B. 65
- suction see Suction couch
Creping SB.81; SB.83
Cross-flow distributor lB.43
Curl 3A.211; SA.38 et seq.; SB.51
Cutters, wire jet 3B.10 2; 3C.31
Cyclones see Cleaners, cyclone
Cylinder, drying SB.32; SB.10 1
- condition inside SB.32; SB.33
- operation of SA.12; SA.21
- surface condition of SA. 37; SB . 14; SC. 23
Cylinder, M.G. SB.11; SB.33; SB.51; SB.8 et seq.
effect on gloss of SB.83
- operation of SA.12 SA.15; SB.12
- surface condition of SB.81 SC.23
Cylinder, pony 3B.65 SA.12; SB.10 1
616
INDEX
Dampers, web 5B.10 5; SC.11; SC.24

Dancing roll SB.10 3; SB.10 5
Dandy 3B. 5 et seq.; 3C.13; 3C .23
Dandy-type roll 3B.24
Deckles 3B.10 2; 3C.23
Deculator lB.74; 2B.12; 2B.63
Deflectors see Table roll baffles
Dew point SA.11
Dirt counting 2A.32; 2A.33; SC.13
Distributor roll see Perforated roll
Doctors, calender SB.10 2; SC.23
- dryer SB. 41 ; SB. 10 2; SB. 10 5; SC. 23
-M.G. SB.81; SC.23
-press 3B.21;4B.92;4C.23
-wire part 3B.22; 3B.62; 3B.10 1; 3C.23
Downtime, analysis of 6A.1 et seq.; 6A. 21; 6A. 23
- measurement of SC.11; 6A. ll
- relation to machine speed 6B.32
-use of data 3C.21; 6B.23
Drainage on wire 3A. l et seq.; 3A.2 10; 3B.1 et seq.;
3B. 22 et seq.
- effect of air in stock on lB.71; tB.74
effect of operating variables on 1B. 12; 3B. 8; 3B. 9
effect of shake on 3B.31
- relation to formation on 3A.22
- relation to machine speed of 3B.9; 6B.31
Draw, couch 3A.41; 3A.43 et seq.
- breaks of web at 3A.44; 6B.31
- observation of 3C.ll; 3C.35
- with jacket couch 3B.61
with suction couch 3B.62
Draws, machine SB.23; 5B.10 1
- control of SB.62; SC.14
- effect on curl of SA.39
- effect on watermark of 3B.53
-measuring 3C. 11 ; SC: 12
- relation to web shrinkage SA.34; SA. 35; SA. 36
Draws, press 4B.51; 4C.12
Dried-in strains see Shrinkage, restriction of
Drive, couch 3A.43; 3B.41; 3C. ll
Drive, machine 3C. 11 ; SB. 23 ; SB. 10 1
Drop-leg on suction boxes 3B.44; 3C.12
Dryer see Cylinder, drying
Dry felts SB.2 et seq.; SC.53
- changing of SC.21; SC.53; 6B.11
- effect on drying of SA.14; SB.22; SC.22
Dry felt tension, effect on drying of SB.23
- effect on shrinkage of web of 3B.53; SA.35; SA.36; SA. 37; SA. 39
- setting of SB.10 3; SC.12
Dry-line position lC.14; lC.44; 3B.8; 3C.31
- effect on dandy of 3B.51
- movement of 3C.35
- relation to suction boxes 3A.34; 3C.12
Dryness see Moisture content
Drying, effect on paper of SA.3 et seq.
see also Shrinkage of paper
Drying rate see Evaporation rate
Drying Section, removal of water in SA.1 et seq.; SA.2 et seq.
Dump valve lB.52; lC.14; lC.42; 2B.12
Dyes, addition of lC.12; lC.44
617
INDEX
Dyes and two-sidedness 3A.28; 5A.12
Economiser see Heat exchanger

Eductors, cyclone lB. 74; 2B.63
Efficiency, machine 6A.23
Effluent lC.12
see also Save-all
Efflux ratio (of slice jet) lA.26; lC.11; lC.14
- effect on paper quality of 3B.14; 3B.31
Electro-doctor SB. 81 ; SB. 10 2
Equilibrium moisture content see Moisture content, equilibrium
Equipment records 6B.12; 6B.24; 6B.25
Evaporation in dryers 5A.ll; 5A.12; 5A.14 et seq.;
5A.22 et seq.
Evaporation rate, determination of 5A.25; 5C.24
- dependence on machine variables 5B. ll; SB.14; SB. 4 et seq.; SB.61
- effect of dry felts on SB.22; SB.24; SC.22
- improvement of SB. 51 et seq.
Evener plates lB.53
Evener roll see Perforated roll
Evolutionary operation, use of 3B. l; SC.13
Expander rolls
use for regulating web shrinkage of 5A.36; SB.10 3; SC.54
use on fabric wires of 3B.74
use on presses of 4A.26; 4C.23
Expansivity, moisture SA.32; SA.33
- relation to sheet defects 3A.21; SA.37; 5A.39
Explosion chamber lB.43
Fabric dry felts 5B.21;5B.22;5B.41;5B.51

Fabric wires 3B.74
Fan pump see Mixing pump
Feed pump see Mixing pump
Felt cleaning see Press felts, cleaning of
Felt dryers SA.14; SA.15; SB.11; SB. 24
Felt roll, hot-air drying 5A.14; SB.24
- use for ventilation of SB.41
- use for moisture profile control of 5C.35; SC.38
Felt suction box 4B.24; 4B.31
Felt wringer press 4A.24; 4B.24
Fibre classification lC.23
Fibre orientation 3A.29 et seq.; 3B.14; 3B.3 et seq.;
SA.34
Fibre suspensions, flocculation of lA.13; lB.13; 3A.23; 3A.24
flow of lA.l;lA.12
Fibre yield see Furnish yield
~& ~ Loading
Flash steam see Cascade drying system
Flat boxes see Suction boxes
Flexible couplings lB.81
Flocculation see Fibre suspensions
Flo-vac belt 3B.43
Flow box see Breast box
Flow in pipes lA.14; lA.15; lB.81
- measurement of 1B.22; lC.23
Foam lB. 71; 1B. 73
Foils 3A.23; 3B.23; 3B.24
Forced convection hood see Hood, H. V.
Formation 3A.2 et seq.; 3B.1 et seq.; 3B.23;
3B.3 et seq.; 3B.Sl
618
INDEX
Formation, measurement of 3A.21; 3B.31 et seq.; SC.13

Forming board lB.82; 3B.13; 3B.14; 3B. 72; 3C.23
Forming boxes 3B.24
Forward drive roll 3B.64; 3B. 71; 3B. 73; 3B.10 1
Freeness measurement lB.11
Fresh stuff flow lB.22; lC.13; lC.14; lC.23; lC.44
see also Hydraulic disturbances
control from beta-ray gauge lC. 32; lC. 33
Fresh water addition lB. 32; lC.12; lC.13; lC.41; 2B.13
Furnish yield 6A.23; 6B.12
Gloss, effect of calenders on 5A.41; 5A.43; 5A.44; 5B.91

- effect of M.G. cylinder on 5B.83
- on-machine measurement of SC.13
Grade changes lA.26; lC.44; 6A.12; 6B.ll
Grewin system 5A.11;5B.24;SB.41 ;SB.42;5C.11;
SC.12
Grooved roll 3B.24
Guard board, jacket 3B.61; 3C.23
Guide, felt 4C.43; SB.10 3; SC.53
Guide, wire 3B.42; 3B.72; 3B.10 2; 3C.23
Hardness, machine roll SC.12; SC.32; SC.36
- correcting profile of 5C.41; SC.42; 5C.54
Head box see Stuff box; Breast box
Heat exchanger 5B.42; 5B.43
Hi-I press 4A.26
Hog-pit lC.42; lC.43; 4B.36; 4B.92
Hog roll see Perforated roll
Hole detector SC.13
Holey roll see Perforated roll
Hood, H. V. 5A.11; 5B.51; SB.52; 5B.53
- operation of 5C.11; 5C.14
- use for moisture control of SC.35; SC.38
Hood, machine SB .11; SB .42; SC.11; SC.14
Hornbustel hole lB.54
Humidification see Conditioning
Humidity, air SA.11; SA. 39; SA. 3 12; SB.43;
SC.11
Huyck-Smith porosimeter 4B.23
Hydraulic disturbances lB.8 et seq.; 2B.31
Hygroexpansivity SA.32; SA.33
Infra-red heating 4B.8; SB.54; SC.38
Jacket, changing of 3C.34

Jacket couch 3A.41; 3B.61; 3C.13; 3C.34
Karaya gum lA.13

Kusters roll see Swimming roll
Lead roll, couch 3B.61; 3B.63; 3C.31

-press 4B.31; 4B.34; 4B.35
Lick-up felt 3B.64; 3B.65
Loading, addition of lC.12; lC.14; 2B.66
- distribution through sheet of 3A.25 et seq.
- effect on retention of lB.12; lB.13
Lukenwald dryer SB.10 1
Lump-breaker roll see Presser roll
619
INDEX
Machine speed, effect on drainage of 1B. 12; 3A. l l et seq.

- effect on drying of SB.61
effect on pressing of 48.5
effect on production of 6B. 23 ; 6B. 31 ; 6B. 32
Magnetic flow meter lB.22; lC.12; lC.13
Maintenance 6A. 12; 6B . 11 ; 6B. 25
- of drying section 5C.2 et seq.
- of presses 4C.23
- of screens and cleaners 2C.2 et seq.
- of wet-end system lC.2 et seq.
- of wire part 3C.23
Manifolds see Breast box approach system
Minton vacuum dryer SB.54
Mixing box 11. 1 ; lA. 16; 1B . 22; 2B . 11
Mixing pump 11. 1 ; 1B. 22; 2A. 26
-air in lB.74
- control of lA.16; lC.14
- pulsations from lB.81
M.G. cylinder see Cylinder, M.G.
Moisture content, at calenders SA.42; SA.43
- control of SC. 3 et seq.
- entering dryers 41. 1; 4C. 13; SI. 1 ; SB. 7
- equilibrium Sl.l; SA.32; SA.39; SA.312;
SA.42
- in dryers SA.13; SA.24; SA.35; SB. 7; SC.22;
SC.24
at reel up Sl.1; SA.37; SA.42; SC.51; 6B.23
representation of 41.2; 51.1
Moisture meters 4C.13; SC. 33 et seq.
Moisture profile, correction of SB.24; SB.41; SB. 51; SB. 54;
SB. 10 5; SC. 38
- variation in 5A.37; SB.51; SC.36; SC.54
Monkey roll see Perforated roll
Opacity SC.13
Orifice plates lB.81; lC.12
Perforated roll lI.1; lB.51; lB.53

Picking, dandy 3B.52
Pick-up, suction 3B.64; 3B.66; 3B.9; 3C. ll; 3C.13;
4B.36
Pipes, avoidance of air in lB.41
- friction loss in lA.12
pH 1C. 12; 1C. 14; 1C .41 ; 2B. 13; 68. 11
- effect on pitch of lC.22
effect on presses of 3B.72
effect on retention of lB.13
Pit, backwater lB.31
Pitch lB.71; lC.12; lC.22; 3B.52;
3B. 74; 3C.35
Pony cylinder see Cylinder, pony
Porosimeter, Huyck-Smith 48.23
Porosity of paper 3B.51; 3B.23; 3B.9; SA.312;
SA.41; SC.13
Press, arrangements of 4B. 3 et seq.
- cross-web variation in 4B.23; 4C.23; 4C.3 et seq.
- double-divided 4A.25
- effect of load on 4A.12; 4A.21; 4B.l et seq.; 4C.ll
- effect of temperature on 48.8
fabric 4A.25; 4A.26
620
INDEX
Press felts 4A.2 et seq.; 4B.2 et seq.; 4B.5;

4C.32; 4C.43
- changing of 4C.21; 4C.43; 6B. ll
- cleaning of 4B.23; 4B.24; 4C.14; 4C.23; 4C.33
- guiding see Guide, felt
- tension of 4C. 12; 4C. 43
Press, felt wringer see Felt wringer press
-Hi-I 4A.26
- influence of entering moisture on 4B.6
- influence of paper properties on 4B.7
- measuring efficiency of 4C.13; 4C.14; 4C. 22
plain 4A.22; 4B. 31
rolls 4B.91; 4C.23; 4C.31; 4C.33
sleeve 4A.26
-smoothing SB.10 5; 5C.ll
- suction 4A.22; 4B.32 et seq.; 4B.91; 4C.14
see also Vacuum, suction press
- suction boxes see Suction boxes, press
- theory of 4A.1 et seq.; 4A. 2 et seq.
- transfer 3B.66; 4B.36
- Venta-nip 4A.26
Presser roll 3A.42; 3B.62; 3C.13
Pressing, effect on paper of 3A.26; 41; 4B.11
Pressure roll, M.G. SB.82; SC.11
Pressurized breast box lB.54; lC.14
Printability of paper SA.41; SA.43
Product control see Quality control
Production analysis 6A.21; 6A.23; 6B.23; 6B.32
Profile bar lB.61
Profitability 6B.3 et seq.
Pulper, dry-end broke 5B.103;5C.14
Pumps tA.16
see also Mixing pump
Quality control lC.31; SC.13; 6A.3 et seq.
Radiclone unit 2B.61; 2B.66
Rectifier roll see Perforated roll
Reel, machine SA.37
Reel-up SB.10 4; SC.12; SC. 23
- moisture content see Moisture content at reel-up
Regulating box 11.1
Rejection scheme 6A.35
Report, machine 6A. ll
Retention, effect of additives on IA. 13 ; 1B . 11 et seq.
- effect on substance of IA.22
effect of wire on 3B. 7; 3B. 74
relation to drainage of lB.12; 3B.9
Reynold's number IA. 11; lA.12
Riffier 2B.41
Roll grinding 4C.31
Rolling friction in calendering SA.44
Rolls see under individual headings
Rolvac unit 3B.43; 4B.24
Rosin size lB.13; lC.12
Rotabelt 3B.43
Rotameters, use of lC.12; lC.13
Rotary belt (over suction boxes) 3B.43
Sand table 2B.41
Sankey diagram lC.23; SC.24
621
INDEX
Save-all, whitewater lA.25; lB.32; lC.14; lC.44;

3B.72
Screens 21.1; 2A.l et seq.; 2A.3; 2C.l et
seq.; 3A.22
- relation to approach flow system lB.41; lB. 82; 2B. l et seq.
Screen, enclosed 2A.14, 2B.3 et seq.; 2C.1 et seq.
-flat 2A. 14; 2B. 21 ; 2C . 1 et seq.
- rotary 2A.14; 2B.22; 2C. l et seq.
- secondary 2A.15 et seq.; 2B.21
Separator tank, vacuum in 3B.44; 3C.11
Service box see Stuff box
Shadow-marking 3A.42; 3B.62; 4A.24; 4A.26; SB.82
Shake 3A. 22; 3B. 11 ; 3B. 3 et seq.
Sheehan ropes SB.23; SC.23; SC.51
Shives 21.2; 2A.16; 2A.24; 2A.33
- removal in screens 2B.33; 2B.34
- removal in cyclones 2A.24; 2B.62
Showers see Sprays
Shrinkage, cross-machine SA.33 et seq.; SB.51; SB.62;
SB.10 3; SC.54
- of paper 3A.43; SA.31 et seq.
- - effect of draws on SB.62; SB.10 1
- - effect of dry felts on SB.23
- - effect of H.V. hood on SB.51
- - effect on watermarking of 3B.53
restriction of SA.33 et seq.; SB.10 1
Shut-down 6A.12; 6B. l 1
of drying section SC.52
- of presses 4C.42
- of screens and cleaners 2C.32
- of wet-end system lC.43
Silica sol lA.13; lB.13
Siphons, condensate removal SB.14; SB.32; SB.33; SC.23
see also Condensate, removal of
Sisal hairs 2B.22
Slice lB.6 et seq.; lC.2; lC.34
see also Apron
- head, relation to jet velocity tA.15; lA.26; lC.11
- - control of 1B. 31 ; 1B. 54; 1c. 14
- jet, character of lA.13; tB.53; lB.64; lC.11;
3A.23; 3B.10 2
- - impact on wire of 3A.22; 3B.11; 3B.12
- - velocity of tA.15; lA.26; lB.42; lC.11;
3B.14
see also Slice head
Slime lB.43; lB. 71; lB. 74; lC.21;
3C.35
Smoothness of paper SA.3 12; SA.42; SB.10 5
- and calendering SA.41; SA.43; SA.44; SB. 91
Smoothing press SB.10 5; SC.11
Sodium aluminate lA.13
Solids content see Moisture content
Specifications 6A.33; 6A.35
Spray dampers see Dampers, web
Sprays, and wet-end system 1B. 32; 1B. 73 ; 1c. 13 ; 1c .42
press 4B.92; 4C.14; 4C.23
-screen 2B.13; 2B.21; 2B.22
-wire 3B.52; 3B.62; 3B. 72; 3B. 73
Spreader rolls see Expander rolls
622
INDEX
Spring roll see Dancing roll

Squeeze roll, jacket 3B.61
Start-up 6A.12; 6B. ll
of drying section SB . 13 ; SB. 15 ; SC. 51
of presses 4C.41
of screens and cleaners 2C.31
of wet-end system lC.41; lC.42
Static SA. 3 12; SB.10 5
Stave roll lB.43
Steam, air in SB. 13 ; SB. 14; SB. 15
- blowing-through of SB.12; SB.14; SB.32
- for pit heating lC.12; lC.14; lC.41; 6B.ll
- in drying section SA.2 et seq.; SB. I et seq.; SC .12;
SC.51
- pressure differential for condensate re- SB.13; SB .14; SB. 32; SC.11 ;SC.14
moval
pressure in dryers SC.11; SC.14; SC.35
for M.G. SB.81
- - relation to drying SA.22 et seq.; SB.11 et seq.
- - relation to paper SB. 7; 6B.31
- - relation to machine speed SB.61
- sprays 3B.52; 3B. 74; 3B.10 2
- superheat in SB.15; SB.52
-traps SB.14; SC.24
- use of amines in SB.15
Stock, definition of 11.1
see also Breast box stock
-jump 3B.1 et seq.; 3B.21; 3B.23
see also Slice jet, impact on wire of
Strainers see Screens, rotary
Strength properties of paper SA.3 12
effect of calendering on SA.41; SA.43
- effect of drying restraints on SA.34; SA.35
- effect of efflux ratio on 3B.14
- effect of foils on 3B.23
Stretch roll, dry felt SB.10 3; SC.12; SC.53
- press felt 4C.12; 4C.43
-wire 3C.11
Stuff box 11.1
Stuff, definition of 11. l
-valve see Fresh stuff flow
Substance, control of lC.13; IC. 3 et seq.; lC.41; lC.44;
SC.13
- cross-web lB.4; lB.61; lB.63; lC.34; 3B.11;
SC.54
- effect on manufacturing costs of 6B . 31 et seq.
- effect of wet-end design on lA.2; 2A.17; 2B.13
- machine-direction variation of lB.8; 2B.31; 3A.43; 3C.12; SB.93;
6B.12
Suction boxes 3A.3 et seq.; 3B.4 et seq.; 3C.12
see also Vacuum, suction box
- cover material 3B.41; 3B.42; 3C.23
- dressing of 3C.24
- press 4B.32; 4B.33; 4B.4 et seq.; 4B.92;
4C.23.
-wet 3B.24
Suction couch 3B.62; 3B.63; 3C. ll
- and open draw 3A.41; 3A.42
and wire wear 3B.71
623
INDEX
Suction press see Press, suction

Sulzer centrifugal exhauster 4B.43
Surge tank lB. 81; lB. 82
Sveen glue lA.13; lB.13; lC.14; 3A.25
Sweat roll 5A. 3 11 ; 5B. 10 5; 5C. 23 ; SC. 24
Swimming roll 4C. 31 ; 5C .44
Table rolls lB.82; 3A.23; 3A.26; 3B.2 et seq.

- discharge from 3A. 11 et seq. ; 3B. 11 ; 3C . 13
- baffles for 3B. 23 ; 3B. 72
-dandy type 3B.24
-grooved 3A.23; 3B.24; 3C.23
Taper flow inlet lB.43; lC.12
Tear properties of paper 3B.14; SA.3 12; SA.41
Temperature of web SA.43; SB.11
Thermocompressors SB.12
Thickness of paper, measurement of SC.13; SC.4 et seq.
- effect of machine variables on 3B.14; SA.3 12; SA.41; SA.42;
5A.44; SB.10 5
- profile at reel-up of SB.92; SC.4 et seq.
Trays under wire 3B. 73; 3B.10 1; 3B.10 2
Trim from winder SB.103; 5C.14; 6A.22; 6B.12
Two-sidedness 3A.22; 3A.25 et seq.; 3B.23;
3B.32; 3B.51; SA.12
Urea-formaldehyde resin see Wet-strength resins
Vacuum assisted foils see Foils

Vacuum, suction box 3A.3 et seq.; 3B.44; 3B.9; 3C.ll;
3C.12
- suction press 4B.41; 4B.42; 4B. 91; 4C.11
Valve, breast box flow control lB.31
Vapour pressure SA.11
Venta-nip see Press, Venta-nip
Ventilation, of cylinder pockets SB.24; SB.41; SB.61; SC.11; SC.24
- machine house SB. 43 ; SC. 11 ; SC. 12; SC. 24
Venturi meter tC.12
Vibrations lB.82
Vickery felt conditioner 4B.24
Vorvac system lB.74
Wake effect lB.53

Wash-up of wet-end lC.43
Water/fibre ratio see Moisture content
Waterlogging of dryers SA. 23; SB. 13; SB. 14; SC. 24
Watermark 3B.53; SB.10 5
Wedge effect (in presses) 4B.5
Wedge-flow spreader lB.43
Wet felts see Press felts
Wet-strength resins lB.13; lC.12
Wet suction boxes 3B.24
Whitewater system 11.1; lA.13; lA.24; lC.14; lC.41
closure of lA.25; lB.32; lC.22
see also Effluent
Wire, changing 3C.21; 3C.33; 6B. ll
- characterstics lB.12; 3B. 7; 3B. 74; 3B. 75
- drainage on see Drainage on wire
-guide see Guide, wire
-life 3B. 71; 3B. 72; 3B. 74; 3C.21; 3C.22
- effect of suction box friction 3A.24; 3A.31; 3B.41 et seq.
624
INDEX
Wire, mark 3A.25; 3B. ll; 3B.51; 3B. 74

return rolls 3B. 64; 3B. 72; 3B. 73; 3B. 10 1
slope of 3B.75
- tension 3B.10 2; 3C.11
- - effect on life of 3B. 71; 3B. 72
- - effect on table roll disturbances of 3A.11; 3B.11; 3B.12
Yankee cylinder see Cylinder, M.G.
625
THEORY AND OPERATION OF THE
FOURORINIER~~~~OOMACHINE
by Geoffrey H. Nuttall
This work is intended primarily for persons who are, upset the process. Particularly in this section the
or will be, directly associated with the production of results of experimental work have been freely
paper or with technical aspects of papermaking. drawn on. The intention is to place in perspective
It assumes some basic knowledge and a fair degree the influence of operational variables associated
offamiliarity with paper machines, and the aim is to with running the paper machine in order to indicate
provide a comprehensive background of information how best different quality requirements can be
which can be continually referred to but is at the fulfilled and the whole operation brought under
same time sufficiently readable to be suitable as a more systematic control.
text book for gaining advanced knowledge of the
process. In scope it fills the gap between The third section is concerned essentially with
general books on papermaking which give a broad practical aspects of paper machine operation. The
introduction to the subject and the large reference value and use of process instrumentation is
works which are mainly concerned with giving a discussed, maintenance requirements are detailed,
comprehensive description of all equipment and and the main tasks of machine crews with
technique~ in use. Students following higher particular emphasis on the problems of running the
courses in Paper Technology will find thatthe machine to produce a consistent product are
material is presented at the right level for their needs. described. There is also some discussion of the
Pulp suppliers, paper machine equipment basic practical tasks involved in machine operation.
manufacturers, and paper users will find this book
an invaluable source of reference.
Previously unpublished material on
Production Control Methods
The Divisions and sub-divisions of The final part of the work, concerned with
the Work Production Control Methods, outlines the data
The work is divided into six parts dealing, relating to machine operation that are essential if
respectively, with the Wet-end Flow System, adequate financial control is to be achieved. The
Screens and Cleaners, the Wire part, the Press regulation of speed, downtime, broke and so forth
Section, Dryers and Calenders, and Production to reduce operating costs, and ways of comparing
Control Methods. All except the last part are the economics of producing different grades on a
divided into three main sections. The first section machine are described. This section includes much
describes theoretical aspects of the process at a new material hitherto unpublished.
level which can be understood by anyone
possessing a basic knowledge of physics; it is The work has been serialised in 'The Paper Maker',
essentially descriptive ratherthan analytical and is London, during the years 1962-1967. The articles
intended to give a general background of the have been completely revised and brought up to date
fundamental principles so far as they are known. for publication in book form. The text-book runs to
approximately 250,000 words with over 100 line
The second section deals with operational aspects illustrations. A comprehensive index and reference
and details the various factors which can affect and list is included.

Theory and Operation of The Fourdrinier Paper Machine

Uploaded by

Copyright:

Available Formats

Theory and Operation of The Fourdrinier Paper Machine

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Theory and Operation of The Fourdrinier Paper Machine

Uploaded by

Copyright:

Available Formats

What are the main sections covered in the work?

What are the main sections covered in the work?

What is the intended audience of the book?

What is the intended audience of the book?

THEORY AND OPERATION

G. H. NUTTALL, M.A. (Cantab.)

PRINTED IN GREAT BRITAIN BY

PART 4: THE PRESS SECTION

PART 5: THE DRYING SECTION AND CALENDERS

PART 6: PRODUCTION CONTROL

R.H. Clapperton and W. Henderson. Modern Paper Making. Waverley

H. Hardman and E. J. Cole. Paper Making Practice. Manchester

To the authors of these books, and of the many references quoted in

THE WET-END FLOW SYSTEM

lA. 1 1 Effect of velocity on the pattern of water flow

~IkTIE~{f- V;L.OC\iY P~OFll..lS

IA. 1 2 Modification to flow pattern produced by fibre

lA. 1 3 Flow of individual fibres and flocculation

verified by Steenberg (92). It appears that flocculation is primarily a

lA. 1 4 Desirable flow velocities at the wet-end

_ 7·66 x B.D. production in long tons/24 hrs.

cause the increase in consistency to be closely balanced by the consequent

IA. 1 5 Relation between pressure and flow

system. Pressure losses of water in pipework under different conditions

1A. 1 6 Centrifugal pumps

Fig. 1. 3. Performance of a typical centrifugal pump with varying water flow

alteration in performance with varying water flow of this type of pump.

lA.2 INFLUENCE OF WET-END DESIGN ON OPERATIONAL

1A.2 1 Backwater system balance at the wet-end

1A. 2 2 Alteration in retention conditions

Fig. 1. Sc. Conditions after third pass round backwater circuit

Fig. 1. 5d. Final equilibrium conditions with new retention

e>~Mfi:IT e,o.I( ~~00

Fig. 1. 6c. Conditions after third pass round backwater circuit

Fig. 1. 6d. Final equilibrium conditions with new backwater flow

lA. 2 3 Alteration in backwater flow

lA. 2 4 Stability of the paper machine flow system

needed to establish a new equilibrium. It is therefore preferable to take all

1A. 2 5 Degree of closure of whitewater systems

will generally be used on a machine requiring frequent colour changes

lA. 2 6 Formation and wire conditions

1B .1 RETENTION ON THE WIRE

18. 1 1 Composition of the furnish and the effect of beating

It is important to note particularly the high loading content (china clay)

lB . 1 2 Drainage conditions on the wire

IB. 1 3 Chemical conditions

lB.2 FRESH STUFF FLOW AND MIXING

lB. 2 1 Consistency regulation

lB.3 BACKWATER AND WHITEWATER SYSTEMS

18. 3 2 Fresh water addition and the whitewater system

introduction of fresh water will increase the overfl.ov> to drain. In this

1B. 4 BREAST BOX APPROACH SYSTEM

lB. 4 1 General approach system

lB. 4 2 Comparison of different approach systems

18. 4 3 Summary of approach design characteristics

Fig. 1. 9a. Simple manifold approach

Fig. 1. 9b. Multiple manifold approach

Fig. 1. 9c. Wedge flow spreader with guide vanes

Fig. 1. 9e. Tapered side-inlet approach