Black 2nd Ed Machine Design

ACPL ITEM
DISCARDED
r521.81
B56m
910105
ACPL ITEM
DISCARDED
M
&
ro
3
O
Q
**
&
w
3
M
CD
CJi
CJi
■
n> ~ fc-sf
■OS
MACHINE DESIGN
Machine Design
PAUL H. BLACK
Professor of Mechanical Engineering
Oh io University
Second Edition
McGRAW-HILL BOOK COMPANY, INC.
New York Toronto London
1955
MACHINE DESIGN
Copyright © 1955 by the McGraw-Hill Book Company, Inc.
Copyright, 1948, by the McGraw-Hill Book Company, Inc. Printed in the

United States of America. All rights reserved. This book, or parts thereof,
may not be reproduced in any form without permission of the publishers.
Library of Congress Catalog Card Number 55-7268
THE MAPLE PRESS COMPANY, YORK, PA.

PREFACE
This book is intended to serve as a textbook for courses in general

machine design and as a reference book in mechanical-design offices.
It is assumed that the reader has a knowledge of mechanics, strength of
materials, kinematics, mechanical processes, and engineering materials
including their properties. A statement of the necessary elements of
these background subjects is given for reference.
The analytical method combined with results of experimental investi¬
gation has been used wherever possible, and the empirical method has
been used only when it is impractical to use the analytical approach.
In the latter class of problems, the direction of rationalization is indicated
by references. 910105
Changes in this edition are mainly in fuller explanation of principles,
assumptions, statements, and methods, in the addition of worked-out
examples, and in the inclusion of some applications which were omitted
in the first edition in the interest of a somewhat condensed book. In
addition, the coverage of the field has been broadened by treating more
fully the selection of material, types of power-transmission units, motor
selection, and other phases of comprehensive applications. The stand¬
ards have been brought up to date, including those for screw threads and
the ratings of power-transmission chains and V belts and gears, as well
as new codes and practices.
The problems at the back of the book are representative of the field
covered and each problem is based on an actual installation with empha¬
sis on the design type of problem rather than on the analysis type. In
this revision the number of short problems has been increased; also there
have been added several longer problems which require the related con¬
sideration of items such as application requirements, selection of material,
units of power transmission, motor selection, and other phases of com¬
prehensive problems.
The literature of the field has been drawn upon freely, and a special
effort has been made to acknowledge the sources of the information.
The suggestions of members of the academic and industrial fields is hereby
gratefully acknowledged, in particular to Dr. M. V. Barton of the Uni¬
versity of Texas for his contributions to the organization and early
VI PREFACE
development of the book, to Dr. R. T. Hinkle of Michigan State College

for his suggestions, to Dr. R. E. Peterson of Westinghouse Research
Laboratories for references on stress concentration and for his review of
the section of stress-concentration design factors, and also to C. P.
Kottlowski and W. V. Covert of Diamond Chain Company, and B. L.
Pearce of Link Belt Company for reviewing the chapter on power chains,
to L. T. Bruggeman of Allis-Chalmers Manufacturing Company for
reviewing the section on V belts, and to Dr. R. C. Quisenberry and R. R.
Selleck of Ohio University for suggestions on the chapter on motor
selection.
Paul H. Black
CONTENTS
Preface . v
Chapter 1. Introduction. 1
2. Machine-design Computations. 5
3. Loading, Induced Stresses, and Failure.10
4. Stress Concentration in Machine Members.31
5. Allowable Stresses.52
6. Members That Fail by Buckling.61
7. Engineering Materials.71
8. Mechanical Fabrication and Processes.91
9. Detachable Fastenings.122
10. Springs.146
11. Pressure Cylinders.168
12. Translation Screws.173
13. Shafting.184
14. Belt Drives and Hoists.192
15. Power-transmission Chains.213
16. Shaft Couplings.224
17. Clutches and Brakes.231
18. Spur and Parallel Helical Gears.251
19. Gears for Nonparallel Shafts.278
20. Surface Finish, Friction, and Wear.297
21. Sliding Bearings and Lubrication.313
22. Rolling Contact Bearings.338
23. Metal Fits and Tolerances.353
24. Vibration and Vibration Control.361
25. Motor Selection.376
Appendix I. ASA Abbreviations.383

II. Elements of Sections.384
III. Bending Moment and Deflection Formulas.385
IV. Values of 7rd3/16 and rd3/S2.386
V. Properties of Materials.387
VI. Tensile Strength and Yield Point versus Brinell Hardness . 388
VII. Endurance Limit versus Tensile Strength.388
vii
VIII CONTENTS
VIII. Yield Points for Steel.389

IX. Hardness Numbers Conversion Chart.391
X. Stress-concentration Factors.392
XI. WR2 for Rotating Bodies.407
XII. Method for Determining Lewis Factor.408
XIII. Graphical Integration.409
XIV. Table of Wire Sizes.412
XV. Decimal Equivalents.413
Questions and Problems.415
Author Index.463
Subject Index.465
CHAPTER 1
INTRODUCTION
As one stands before any machine which is operating—such as a plastic¬

molding machine turning out say a hundred switch-box covers each hour,
or a punch press performing a dozen operations simultaneously at each
stroke on a steel part for a computing machine, or the main reduction
gears of the propulsion machinery of a ship which change the speed from
the turbine speed of 3,600 rpm to propeller-shaft speed of 225 rpm while
transmitting 8,500 hp, or a turboalternator running at 1,800 rpm supply¬
ing light and power to a community at a peak load of 50,000 kw at 5 p.m.
on a December day—as one watches such machines in operation, he is
bound to have feelings of awe and reverence for their performance and
their power, both mechanical and economic. He will no doubt notice
the form of the machine and of the visible parts. He may note that
many of the parts appear complicated and that their operation contributes
to the operation of the machine as a whole in a synchronized manner.
Watching such a machine is an inspiration to a mechanically minded per¬
son to understand its construction, operation, and performance.
The operation and performance of the machine as a whole depends on
the individual parts, each of which has individual characteristics just as
the machine as a whole has its characteristics. These individual parts
may be shafts, couplings, bearings, gears, brakes, controls, or many other
units.
The production of such a machine is the result of an enormous amount
of directed work: theoretical, experimental, and manual. The theoretical
work includes the application of the laws of physics, including thermo¬
dynamics, mechanics, and strength of materials, chemistry and, invari¬
ably, of the mathematical methods. The experimental work involves the
verification of the laws and relations that were developed from theo¬
retical considerations, the determination of the properties of materials,
the analyses of stress, strain, and vibrations in complicated structures,
and the performance of complete machines and units of machines, such as
power-transmission units. The manual work involves actual production
in the shop, which makes use of highly developed techniques. Other
factors that must be considered are safety and economics. Each of
1
2 MACHINE DESIGN
these, as well as many others that are not mentioned, is in the prov¬
ince of experts who confine their activities mainly to the fields of their
specialization.
In the production of every machine, someone must have the over-all
responsibility of deciding on the mechanical arrangements to be used, of
making use of the information provided by the specialized fields men¬
tioned above, and of combining them by engineering judgment so that
the machine can be manufactured economically, so that it will perform
its specified function properly, and will have a satisfactory life and main¬
tenance. The man who has this over-all responsibility is the mechanical
designer, or machine designer.
In each design, a number of operations are involved.
1. It may be necessary to decide on the kinematic arrangement of the
parts of the machine. In a new machine, this may call on the creative
ability of the designer, that is, on his ingenuity, inventiveness, resourceful¬
ness, originality, and intuition. In this procedure it is always well to
begin by setting down a clear statement of the problem.1 Creative
ability applies not only to the kinematic arrangement of the parts but also
to other phases of design, such as devising new operating processes, new
materials, and new production methods. It may be an individual or a
group operation. After the kinematic arrangement of the parts has been
established, their motion may require detailed analysis. This analysis
may involve the use of tentative dimensions that may require altera¬
tion as the design progresses and the dynamic characteristics become
established.
2. The maximum forces coming on the parts and the nature of the
forces must be determined. The forces may be due to either static or
impact loading, and they may or may not be applied repeatedly. Inertia
forces may also be induced, and vibration forces which are either locally
induced or transmitted from another member may be present. Fre¬
quently it is not possible to determine the forces exactly and in these cases
the judgment of the designer is required to estimate the maximum forces,
his decision being based on factors which occasionally may be somewhat
indeterminate.
3. The ability of the member to function properly must be considered.
1 For instance, in telephone equipment the soldered connection has proved unsatis¬
factory. The Bell Laboratories set out to devise a method of making connections
which would have the following characteristics: high conductivity, long life, speed in
manufacturing, quality independent of maker, small space, no necessity for clipping
ends of wire, and low cost. That was the statement of the problem. The solution
was the development of the solderless wrapped connection made by an electrically
driven wrapping tool. This connection is replacing soldered connections in Western
Electric communication equipment. See J. J. Kuhn, Sr-, Solderless Wrapped Con¬
nections, Bell Labs. Record, February, 1954.
INTRODUCTION 3
This involves both the material of the member and its size and shape.
The selection of the material requires a consideration of mechanical
properties, of heat-treatment, and of manufacturing processes. The
load-carrying ability, which will be discussed in detail later, may depend
on the yield point of the material, its modulus of rigidity, ductility, heat-
treating characteristics, endurance limit, strength, or creep limit. In
addition, the shape and finish of the member may have a considerable
effect on its strength.
4. The manufacturing process by which the member is to be made must
be selected. This involves consideration of the effects of the process on
the properties of the material and on the cost of the processing. The
latter item is obviously of great importance in any design intended for
production in large numbers.
After the part has been satisfactorily designed to operate properly
(,functional design), it may be necessary to make a production design in
order to reduce the cost to a minimum.1
5. In most designs it is necessary to consider additional items, such as
safety, appearance, cost of operation, standard codes, thermal effects,
contamination, corrosion, and maintenance.
This book attempts to cover the principles of general machine design.
Chapters 3 to 8 cover essentially the design of parts for strength and
rigidity. This section includes the selection of material for making the
parts. Chapters 9 to 12 cover fastenings, springs, cylinders, and transla¬
tion screws. Chapters 13 to 19 may be termed the power-transmission
section. This is a most important section for industrial applications.
Chapters 20, 21, and 22 cover surface finish, friction, wear, lubrication,
and j ournal-bearing design and selection of rolling-contact bearings. The
last chapter, Chap. 25, is on motor selection and is included since many
machines include a motor in the drive and the characteristics of the motor
may have a marked effect on the design of the entire equipment.
The problems in the back of the book are based in most cases on data
from actual installations. They are mainly of the design type rather than
analysis problems. The comprehensive problems at the end of the prob¬
lem section are intended to illustrate the design procedure, the integration
of work covered in the various chapters, and the necessary correlation
between the five steps stated earlier in this chapter. This correlation is
made necessary by the related functioning and performance of each part
of the machine. Experience in the field gives the machine designer an
appreciation of this interesting process.
There are numerous references indicated throughout the text as foot-
1 E. Buckingham, “Production Engineering,” John Wiley & Sons, Inc., New York,
1942; Herbert Chase, “Handbook on Designing for Quantity Production,” McGraw-
Hill Book Company, Inc., New York, 1944.
4 MACHINE DESIGN
notes. While it is realized that some students may not have the oppor¬
tunity to consult these references or may not have them readily available,
they are given to indicate the scope and sources of further developments
and for the direct use of interested designers. Most of these references
carry selected bibliographies for further reference.1
Just as the field of machine design today involves requirements of
manufacture and performance which were unknown or unappreciated a
generation ago, the coming generation must deal with problems which
today are just coming to light. High- and low-temperature and high-
pressure applications of materials are two of the pressing developments.
The advent of the gas turbine and of atomic power production and utiliza¬
tion is requiring performances which were considered impossible a few
years ago. And, finally, the trend today is to require smaller and smaller
things to do bigger and bigger jobs and to do them better. The machine
designer who appreciates each of the five items listed earlier in this chap¬
ter and also a sixth item, to make each of the other five work together effec¬
tively, is in an excellent position to handle the problems which will be
demanded tomorrow by industry.
1 Two excellent general sources of additional information are Marks’ “Mechanical
Engineers’ Handbook,” 5th ed., McGraw-Hill Book Company, Inc., New York, 1951,
and Kent’s “Mechanical Engineers’ Handbook, Design and Production,” 12th ed.,
John Wiley & Sons, Inc., New York, 1950.
CHAPTER 2
MACHINE-DESIGN COMPUTATIONS
2-1 Importance of computations. Design procedure in engineering

offices usually requires the making of computations and notes that become
the property of the firm or company rather than of the individual who
makes them. As a rule the computations are checked by someone other
than the designer; hence they should be made and recorded so that they
can be followed readily by the reviewer; in other words they should, so far
as possible, be self-explanatory.
In order that a designer may acquire the habit of making computations
in good form and of keeping them orderly, he should constantly keep in
mind that they are a record of that phase of the design and that others
may have occasion to refer to them.
It is interesting to look back over a sample of one’s own computations
which were made a year or so previously. If they are readily understood,
it indicates that they have good form. If not, one can easily imagine
that another person would not be impressed with the ability or the cour¬
tesy of their originator.
2-2 Form of computations. To aid the designer in developing good
form, some suggestions are made here. Many of these are undoubtedly
already followed by the reader, some of them may be new, but they are
generally characteristic of a presentable record and it is suggested that all
of them be tried.
1. Each sheet should carry the designer’s name, date, and consecutive
numbers if more than one sheet is used.
2. One side only of the paper should be used and a 1-in. strip on the
left edge of the sheet should be reserved for binding.
3. If possible, a sketch should be drawn for every problem. The sketch
should be neat and correctly drawn, and may be used to show the arrange¬
ment of the parts, and to show such data as dimensions, speeds, materials,
etc. For data that cannot be shown on the sketch, a list should be used.
4. Each step should be set off, labeled or lettered, so that divisions of
the work can be readily identified. This corresponds to using para¬
graphs in ordinary written matter and is helpful in making computations
understandable.
5
6 MACHINE DESIGN
5. All pertinent assumptions should be stated and references given.

6. Equations should be solved for the unknown quantity before numer¬
ical substitutions are made. v
7. The numerical substitutions should follow the literal equation and
be in order, term for term.
8. Units should not be placed in numerical equations. If a check on
units is desirable, a separate dimensional equation should be used.
9. Units should be indicated for every separate value or result that has
units, whether it be data, intermediate calculated value, or final result.
The ASA abbreviations are recommended (see Appendix I).
10. Long equations should be avoided if possible; instead, intermediate
values or particular physical quantities should be calculated separately
before being inserted into the parent equation.
11. An unjustified number of significant figures should not be used.
12. Calculated dimensions of parts should be rounded out to nominal
sizes and indicated, as, for example, d = 2.95; use 3 in.
13. Values in an equation in formal computations should not be
canceled. Since erasures should be avoided, a change is indicated by
drawing a line through the old value and writing the new value over or
under the old. Canceling values would be confusing. Do any canceling
off to one side.
14. All results should be considered from the standpoint of their
reasonability.
2-3 General solution. In solving problems involving several quanti¬
ties, it is usually desirable to make a general solution first, using symbols
for the quantities involved and after the general solution is completed, to
make numerical substitutions for the particular solution that is required.
This method of making a general solution first has several advantages,
some of which are as follows:
1. It provides the possibility of making a dimensional check on the
final expression.
2. It makes it possible to determine the number of significant figures to
which the quantities must be determined in order to obtain a desired
degree of accuracy of the final result.
3. It facilitates making an alternative solution with changed data.
This procedure, which is discussed later, is invariably necessary in
engineering offices.
4. A general solution can be used for making subsequent solutions of
the same problem. If one mixes particular numerical values in a deriva¬
tion, it becomes necessary to repeat the entire derivation if a new set of
data is specified.
2-4 Absolute and relative errors. The difference between a measured
value and its exact value is termed the absolute error. Thus, if the
MACHINE-DESIGN COMPUTATIONS 7
measured value is too large, the error is positive, and if too small, it is a
negative error. If the absolute error is divided by the exact value, the
result is the relative error and is usually expressed as a percentage. In
engineering, the relative error frequently has more significance than the
absolute error. The following comparison will illustrate the difference.
1. An absolute error of 30 mph in determining the speed of an airplane
traveling 300 mph is a relative error of 10 per cent. This illustrates poor
accuracy.
2. A measurement of the distance from New York to San Francisco
with an absolute error of 30 miles is a relative error of 1.2 per cent. This
is good accuracy.
3. Michelson determined that the commonly used velocity of light was
30 mps in absolute error. This is a relative error of 0.016 per cent, which
is highly accurate.
2-5 Errors in computations.1 Let A and B represent the correct
values of two quantities and m and n be, respectively, the relative errors
involved in the measurement of the two quantities.
Addition of two numbers when one contains an error.
A + B — C the correct result

(A + mi) B = D the result containing the error
The absolute error in the result is equal to
D - C = (A + mA) + B - (A + B) = mA
The relative error in the result is equal to

mA _ m
Ta^B = 1 + (.B/A)
The last result is less than m. Therefore, in adding quantities, a rela¬
tive error in one of the quantities does not affect the result by as great a
relative error as in the inexact quantity.
Subtraction of two numbers when one contains an error.
A — B = C the correct result

(A + mA) — B = D the result containing the error
D - C = (A + mA) - B - (A - B) = mA

mA _ m
A - B ~ 1 - (B/A)
1 J. B. Scarborough, “Numerical Mathematical Analysis/’ Johns Hopkins Press.
Baltimore, 1930.
8 MACHINE DESIGN
The last result is greater than m. Therefore, in subtracting two quan¬

tities, an error in one of the quantities affects the result by a larger relative
error than that in the original quantity. In case the two quantities A
and B have nearly the same value, the relative error in the result due to an
error in A or B may be considerable. This magnification of error is
shown by the chart, Fig. 2-1, which illustrates the effect of a 1 per cent
error in the larger of two numbers in subtraction.
Addition and subtraction of two numbers when both contain errors. It
may be shown that in addition and subtraction of numbers containing
errors, the result contains an absolute error equal to the sum of the abso¬
lute errors of the individual quantities.
Difference in numbers,percent
Fig. 2-1. Effect of errors in subtraction of numbers.
Multiplication and division of two numbers when one contains an error.
A X B = C the correct result

(A + mA) X B = D the result containing the error
D — C = (A + mA) X B — AB = mAB
mAB
-AB=m
Therefore, in multiplication as well as in division of quantities, a rela¬

tive error in one of the quantities affects the result by the same relative
error.
MACHINE-DESIGN COMPUTATIONS 9
Multiplication of two numbers when both contain errors.
A X B = C the correct result

(A + mA)(B + nB) = D the result containing the error
The absolute error in the result is

D — C = (A + mA)(B + nB) - AB
= AB(m + n + mn)
The relative error is equal to
AB(m + n + mn)
= m + n + mn
AB
For small values of m and n, the term mn can be dropped to obtain an
approximate result. Therefore, for relative errors of small value in two
(or more)‘numbers, the relative error in their product is approximately
equal to the algebraic sum of the relative errors in the numbers. Thus, if
two numbers each have 3 per cent errors, their product will be in error by
6 per cent, using the approximate method. The actual percentage error
is 6.09.
Raising a number to a power. In this case, the relative error in the
result is approximately equal to the relative error in the number multiplied
by the power when the relative error of the number is not large.
Extracting a root of a number. In this case, the relative error in the
result equals the relative error in the number divided by the root where the
relative error is not large.
2-6 Revising computations. The methods discussed above may be
used to revise computations as required by small changes in specifications
or data. Here, the change in the quantity may be considered as a differ¬
ence between a new and an old value just the same as an error is a differ¬
ence between a measured value and the correct value. In changed data,
the effect of the change on the quantity being solved for depends on the
mathematical operations involved (addition, multiplication, etc.), and
the methods worked out above may be used to alter an old result without
re-solving the equations involved.
As an example, the diameter of a solid circular shaft subjected to a tor¬
sional moment T may be determined by the use of the equation
where s is the allowable shearing stress. Assume that a diameter d has

been calculated for a particular value of T and s. If, for example, the
data are changed so that the torque T is increased 10 per cent, the cor¬
responding diameter of the shaft will be increased approximately 3^ per
cent over that indicated by the original calculation.
CHAPTER 3
LOADING, INDUCED STRESSES, AND FAILURE
3-1 The design problem. In arriving at the most satisfactory design

of a machine, the designer makes use of mechanics, strength of materials,
properties of materials and their heat-treatment, manufacturing proc¬
esses, and other fields of the applied sciences, combining them in a prac¬
tical manner by using ingenuity and design judgment, so that the machine
can be produced in the most economical manner and so that it will per¬
form its required functions satisfactorily.
In the design of machine parts that are severely loaded, the stresses are
kept within limits by using proper size, form, material, and finish for the
parts. Examples of such parts are the head shaft of an elevator which
supports the drum, cables, elevator cage, and passengers or freight; the
shaft of a turboalternator carrying the rotor which may weigh upwards of
a million pounds; and the bolts holding in place the cylinders of an air¬
craft engine.
In the design of another class of parts, strength is likewise important
but also the deflection requirements are equally important and may be
more influential than strength in establishing the size of the parts. A
typical example is a shaft which supports gears. Here the gear teeth
must be well aligned under load, and this usually requires a larger shaft
than if strength alone were the criterion for satisfactory service. In this
case too, vibration characteristics are important and they are evaluated
not in terms of strength of the shaft but of its stiffness.
In another class of parts, production requirements and assembly and
handling are determining factors, such as the housing for a speed reducer.
Here strength calculations are difficult, if not impossible, and in cast
housings they are usually not significant. That is, the thickness of the
housing wall for strength alone may be too thin to cast properly, and it
may become broken in shipment or in handling by the customer if he
drops it, which must be anticipated. Therefore the stresses due to
service loads in a practical casting are usually low because the parts are
made overly heavy. In this case the added rigidity due to the heavy
parts is of benefit in maintaining alignment of the parts.
Finally, in another class of parts, usually small, as in instruments, con-
10
LOADING, INDUCED STRESSES, AND FAILURE 11
trol parts, business machines, and telephone equipment, the parts may
have small service loads so that their size is determined not by strength
requirements but rather by those of kinematics, processing, assembly,
accessibility, adjusting, etc.
The above discussion is not intended to be inclusive of all kinds of parts
which a machine designer may encounter, but points out that strength
and stiffness are not the only criteria in the design of many machine
parts.
In the design of a load-carrying member, three of the quantities that
must be determined by the designer are (1) the applied loads on the
member, (2) the stresses and deflections produced by the applied loads
and their comparison with permissible values, and (3) the material of the
member and its size and form. Some of these quantities are related, and
many of them are of such a nature that they cannot be determined
exactly but must be approximated according to the judgment of the
designer. The exercise of this judgment requires a thorough knowledge
of fundamental principles and an appreciation of the balance and com¬
promises which are invariably necessary.
For example, in the design of a camshaft, the forces on the shaft due to
the action of a cam and its follower depend on the mass of the follower and
frequently also on a spring force, on the speed of operation and the cam
profile, on the clearance between the cam and the follower which may or
may not be properly adjusted, and on the deflection and vibration charac¬
teristics of the shaft itself.
The stresses induced in the shaft may be kept within limits by using a
large shaft, but space and weight limitations may require a small shaft;
also heat-treatment is more effective on small parts and usually small
parts cost less. On the other hand, allowable deflections and vibration
characteristics may require a larger shaft. Then the bearings for the
shaft should be considered, since their selection is influenced by the size of
the shaft which in turn affects the loads on the shaft and bearings. To
save weight, a hollow shaft may be considered but this may be too expen¬
sive for the price class of the machine.1
If the shaft is too large for space requirements, the designer may con¬
sider using alloy steel instead of carbon steel, but the alloy steel and the
required heat-treatment may unduly increase the cost; also alloy steel and
carbon steel have the same deflection properties and if deflection is impor¬
tant, which is frequently the case since it affects vibration characteristics
and bearing performance, alloy steel is no better than carbon steel.
Thus it can be appreciated that the solution of this problem is more
1 In a machine in the appliance field recently placed on the market by one of our
large manufacturing companies, the first item in the specifications was: Retail selling
price shall be $14.95.
12 MACHINE DESIGN
than substituting given values in an equation and using a slide rule for
the result, but instead it becomes an interesting experience of balancing-
several interrelated and generally conflicting characteristics and influences.
The three quantities discussed above, i.e., loads, induced and allowable
stresses, and the corresponding deflections are discussed in this chapter
and in the two following ones in order to provide a background for dealing
with them in the design of machine members.
3-2 Types of loads. A static load is one which does not vary in magni¬
tude, direction, or point of application. It may also be defined as a load
that induces stresses that do not vary. Some ex¬
V///\////t amples of static loads are those on members due to
dead weights, tightening-up loads on bolts, and cen¬
trifugal forces on a disk rotating at constant speed.
Loads that vary somewhat but only infrequently
are often treated as static loads, for example, the
w
II
II
1 1
spring load on a safety valve.
It
A cyclic load is one that induces stresses which
vary in magnitude and/or direction. Cyclic loads
Jz may or may not be accompanied by shock. Ex¬
1 ^ PTiv,
1-L
1
_J, i ♦ —
amples of cyclic loads are those on a gear tooth and
loads on a rotating shaft subjected to a bending
L— moment. Loads caused by vibration are cyclic.
1
This type of load is sometimes referred to as a re¬
Fig. 3-1. Shock load¬
ing. peated load or a fatigue load.
Shock loads1 are due to impact. To investigate
the effect of shock, an elastic system loaded by a falling weight W, as
shown in Fig. 3-1, will be used as an example.
The system shown is one whose elastic displacement 8 of the point of
load application is proportional to the load producing it, that is, P = C8,
where C is a constant for the system.
Let W = falling weight, lb
h = height of free fall, in.
8 = displacement of point of load application, in.
P = impact load, lb
C = P/8 = lb per in. of deflection
By assuming that the support is unyielding, the energy given up by the
falling weight will be absorbed wholly by the system. Assuming also
that the mass of the system is small so that its inertia forces may be
neglected and that the stresses induced are within the elastic limit, the
energy absorbed by the system will be equal to the product of the average
1 S. Timoshenko, “Strength of Materials/’ 2d ed., I, p. 300, D. Van Nostrand
Company, Inc., New York, 1940; also Merhyle F. Spotts, Impact Stress in Elastic
Bodies, Product Eng., vol. 17, no. 3, p. 200, March, 1946.
load and the displacement, or yP8. By equating this energy to that

given up by the falling weight, the following equation is obtained:
y2P8 = W(h + 8) (3-1)
But 8 = P/C where C is a constant, lb per in., for a system. Substi¬

tuting this value of 8 into Eq. (3-1) and solving for P gives
P = W (l + (3-2)
If Eq. (3-2) is written as
W= 1 + yj1 +2W (3'3)
the quantity P/IE may be termed a “shock factor” and represents the
magnification of load caused by impact.
Note that the shock factor can be decreased by lowering C, that is, by
reducing the number of pounds required to deflect the member 1 in., that
is, make the system more “springy” or “soft.”
In the example above, for a bar in tension, 8 = PL/AE, or
r - - - —
By decreasing the area A, or by decreasing the modulus of elasticity E

(using, for instance, rubber instead of steel), or by increasing L, then the
value of C is decreased and the magnitude of the shock load P would be
decreased.
The equation for P/8 depends on the system, whether it be a bar in
tension, a beam in bending, or a helical spring in tension or compression;
in the case of a bar in torsion, the equivalent of C is radians per inch-
pounds of torque. In the latter case, a new equation may be derived for
the shock factor in which the equivalent value for C may be determined
from 6 = TL/JG, that is, C = d/T.
If W in Eq. (3-2) were applied instantaneously and without initial
velocity, h in the equation would be zero and the equation would reduce to
P = 2W
This equation indicates that the effect of applying a load suddenly is

twice as severe as applying it statically. This is termed a suddenly
applied load without velocity of approach. An example approximating
this type of loading is the pressure caused by the explosion of gases in an
internal-combustion engine cylinder which is caused by a rapid building
up of pressure in a time of the order of 0.001 sec.
14 MACHINE DESIGN
In case the time interval of application of the load is appreciable as

compared with the natural frequency of the system, the shock factor
would be less than indicated by the above equation.1
Referring back to the example of Fig. 3-1, the energy of the falling
weight was transferred to the bar as strain energy so that when the weight
reached its lowest position the bar would be in maximum tension. The
tensile force in the bar would then accelerate the weight upward and in
the absence of friction and windage the weight would be returned to its
original position.
If the falling weight on striking the end of the bar became locked to it,
the bar in tension would then accelerate the weight upward until the
stress in the bar became zero. Then the moving weight being locked to
the bar would place the bar in compression until the weight came to rest.
At this point, consider the negative sign for the radical in Eq. (3-2).
In the above-described manner the bar would be subjected to a repeated
load which would cause vibration of the support, granted that actually
the support would not be rigid as assumed. The vibration would con¬
tinue until the energy was dissipated due to damping.
Example 3-1. A rigidly built-in steel cantilever has a cross section in. deep,
% in. wide, and is loaded 12 in. from the support. Determine the maximum bending
stress in the beam, assuming (a) that a load of 8 lb is applied gradually and (6) that an
8-lb weight is dropped on the beam through a distance of z/{ 6 in.
Solution: The section modulus is
I bh2 0.375 X 0.52

= 0.0156 in.2
6 6
(a) Static stress:
M 8 X 12
= 6,150 psi
I/c 0.0156
(6) Impact stress:
The moment of inertia is
bh3 _ 0.375 X 0.53

0.00390 in.3
12 12
From the equation for deflection of a cantilever,
PL3
8 = (Appendix III)
3 El
ZEI 3 X 30,000,000 X 0.00390
= 204 lb per in.
L3 ~ 123
The shock factor, Eq. (3-3), is
P ! i a /i , 2Ch , , . L , 2 X 204 X 0.1875 .

r = l + \(l+lr = l+-\/l + -g- = 4.33
1 See W. M. Murray, Effects of Shock Loading, Product Eng., December, 1954, p.

171,
The impact stress then is
4.33 X 6,150 = 26,600 psi
The actual value of the stress would be less than this value due to the inevitable non¬
rigidity of the support, to the inertia of the beam, and to the local deformations of the
weight and the beam at the region of contact at the impact. These influences could
be taken into account if a more exact analysis were required. Stress concentration at
the built-in end of the beam may be taken into account separately.
3-3 Tension and compression. Tensile stress. The average tensile

stress induced in the body of a simple tension member as shown in Fig. 3-2
is
P
s (3-4)
A
where s = average stress, psi
P = axial load, lb
A = cross-sectional area, sq in.
Compressive stress. By reversing the directions of the loads in Fig. 3-2,
a compressive stress will be induced in the member. Equation (3-4) may
be used to determine this stress, unless the member is so slender that it is
classified as a column, in which case stability considerations will determine
failure. This type of failure is treated in Chap. 6.
Bearing pressure. There is a particular type of local compression that
requires special attention. This is known as “bearing pressure/’ or “con-
P
V
Fig. 3-2. Tension Fig. 3-3. Distribution of bearing

member. pressure for pin connection.
tact pressure,” and occurs between two members held in contact. If the
load is applied to the member in Fig. 3-2 by pins, the distribution of pin
pressure will not be uniform but will be in accordance with the shape of
the surfaces in contact and the deformation characteristics of the two
materials. The distribution of pressure will be similar to that shown in
Fig. 3-3(6).
16 MACHINE DESIGN
Since the actual distribution of pin pressure is difficult to determine

accurately, the calculation of bearing pressure usually consists of dividing
the load by the “projected bearing area” of the pin, i.e.,
(3-5)
where P = load on the pin, lb

t = length of the pin in contact, or thickness of the link, in.
d = diameter of the pin, in.
The pressures as calculated by Eq. (3-5) from results of tests on links
and pins or on riveted connections may be used to determine allowable
pressures in design.
3-4 Bending. Straight beams. The stresses induced by bending are a
particular distribution of tensile and compressive stresses. For a straight
Fig. 3-4. Distribution of bending stress in a straight beam.
beam made of homogeneous material with the same modulus of elasticity

in tension and compression and stressed within the elastic limits, the fol¬
lowing equation applies:
My
s
I
where s = stress at a distance y from the neutral axis, psi

I = moment of inertia of cross section about the neutral axis, in.4
M = applied bending moment, lb-in.
The above equation gives the maximum tensile stress when it is evalu¬
ated for the maximum value of y on the tension side of the specimen.
This value of y may be denoted as ct. The maximum compressive stress
is similarly found using the maximum value of y on the compressive side
as cc. For symmetrical beam sections, ct = cc, which may be denoted as c
(see Fig. 3-4). Hence,
st — sc — i ~ I/c (3~6)
The quantity I/c is known as the “section modulus” and has the units
of in.3
Curved beams. In curved beams, such as crane hooks and punch-press
frames, the neutral axis of the cross section is shifted toward the center
of curvature of the beam, causing a nonlinear distribution of stress, as

shown in Fig. 3-5.
For this case1
— (\ -f ~ —y.—\ (3-7)
AR\^ Z R + y)
where M = bending moment, lb-in., positive when it increases curvature

A = area of cross section, in.2
R = radius of curvature of centroidal axis, in.
y = distance from centroidal axis to fiber with stress s; y is posi¬
tive when measured away from center of curvature, in.
if y
Z = a property of the cross section defined by — -r/ - . — dA
, A J n, ~r y
Fig. 3-5. Distribution of bending stress in a curved beam.
The value of Z may be found analytically or graphically. The deter¬

mination may be complicated, and hence maximum stresses are frequently
determined by using a correction factor with the straight-beam equation.
Hence
(3-8)
where I = moment of inertia of cross section about the centroidal axis,

in.4
c = distance from centroidal axis to inner or outer fiber, in.
K = a factor depending on the shape of the cross section and the
curvature of the beam
Values for K are given in Fig. 3-6. It may be noted that K is nearly
unity for values of R/c greater than 10. For sections that do not approxi-
1 F. B. Seely, ‘ Resistance of Materials,” p. 371, John Wiley & Sons, Inc., New
York, 1937.
18 MACHINE DESIGN
mate a rectangle or a circle, see complete tables,1 or investigate by means

of Eq. (3-7).2
Fig. 3-6. Factor K for curved beam of circular and rectangular cross section.
3-5 Shear. Direct shear. The shear stress is the component of the
stress on a plane section that is parallel to the section. An example of a
Fig. 3-7. Rivet connection and shear deformation.
member that has undergone permanent deformation in direct shear is the

rivet in Fig. 3-7. The average shear stress over the cross-sectional area of
the rivet, or pin, has the value
1 F. B. Seely and J. O. Smith, “Advanced Mechanics of Materials,” p. 148, John
Wiley & Sons, Inc., New York, 1952.
2 C. D. Albert, “Machine Design Drawing Room Problems,” p. 226, John Wiley &
Sons, Inc., New York, 1951.
P
where P = shearing load

A = area in shear
For the rivet shown in Fig. 3-7, where the diameter of the rivet is d,
4P
Ss = (3-9)
7T d2
Torsion. Torsion induces shear stresses on cross sections normal to the

axis of bars and shafts. For circular shafts, the shear stress at any point
a distance r from the center is given by
Tr
Ss =
J
where T = torque on shaft, lb-in.
r = distance from center of the shaft to the point of stress ss, in.
J = polar moment of inertia, in.4
For solid circular shafts, the maximum stress occurs at the outside fiber
and, since the polar moment of inertia for a circular section is 7rd4/32
where d is the diameter, the maximum shearing stress is
16T
s8 = (3-10)
ird*
Where weight must be a minimum, it is common practice to use hollow

shafts. This removes material from the center of the shaft where the
stresses are low, and it also improves the heat-treating characteristics.
Fig. 3-8. Rectangular bar in torsion. Maximum shear stress occurs at a.
Bars in torsion with noncircular cross sections have a complicated stress

distribution for whose determination the theory of elasticity is essential.1
This theory shows that the maximum stresses in a bar of rectangular
cross section occur at the midpoints of the long sides, as indicated at a in
Fig. 3-8.
The maximum shear stress for bars with rectangular cross section may
be determined by means of the formula2
1 S. Timoshenko and J. N. Goodier, “Theory of Elasticity,” 2d ed., McGraw-Hill
Book Company, Inc., New York, 1951.
2 Ibid., p. 273.
20 MACHINE DESIGN
T
ss = - (3-11)
\
abP
where T = torque on bar, lb-in.

b = breadth of rectangular section, in.
t = thickness of rectangular section, in.
a. = coefficient depending on the ratio b/t (see Table 3-1)
Table 3-1. Stress Coefficients for Rectangular Bars
b/t a
1.00 0.208
1.50 0.231
2.00 0.246
3.00 0.267
4.00 0.282
6.00 0.299
8.00 0.307
10.00 0.313
00 0.333
For long and narrow cross sections, the value of a is approximately }/%;
hence for b/t ^ 10
_ 3T
S‘ ~ bt2
3-6 Eccentric loading. If the short prismatic bar in Fig. 3-9(a) is

loaded by a compressive force collinear with its longitudinal axis, the
stress on cross sections of the bar remote from the points of application of
the loads will be uniformly distributed and Eq. (3-4) will apply.
When the point of application of the load is displaced a distance e along
a principal axis of the bar, as shown in Fig. 3-9(6), the stress on cross sec¬
tions of the bar will not be uniformly distributed. As shown at (c), two
forces Pi and P2 equal in magnitude to P may be introduced without
altering the equilibrium of the forces on the bar. It may be noted that
the force P i will induce a uniformly distributed stress whose value is
equal to Pi (or its equal P) divided by the cross-sectional area A. The
couple equal to Pe will bend the bar, and this will induce a compressive
stress at n and a tensile stress at m. By using the principle of superposi¬
tion, the direct compressive and bending stress may be added at n and
subtracted at m. Thus
P pc P
Sc = ~T + 1 (3-12)
Pect P
(3-13)
s‘ - ~1 A
where sc = maximum compressive stress, psi
St = maximum tensile stress, psi
When the loading on the members is tension, the equations above may
be used by interchanging the subscripts c and t.
-4
ler
Pi
" M
rn n
17777ZK777777, T7////////&//. W//*
P p P
(a) (b) (c)
Fig. 3-9. Axial and concentric loading.
Example 3-2. Curved beam design. The sketch, Fig. 3-10, shows the yoke of a
portable hydraulic riveter. The yoke screws on a riveting head (not shown), which
carries the hydraulic cylinder, handle, and manual control valve, and is connected to
the pressure pump by a hydraulic hose. These units have capacities of 10 to 30 tons
and weigh 10 to 15 lb. To keep the size of the unit small for accessibility to the work
and its weight low for ease of handling, it is necessary to make the yoke of alloy steel.
Since each head should accommodate yokes of several sizes, it is customary for con¬
venience in mounting the yokes on the head to make all yokes the same width, dimen¬
sion w.
Data: Capacity P = 10 tons; gap a = 2 in.; reach b = 2 in.; width of yoke w = 1

in. Assume SAE 2340 steel. This steel in 1^-in. thickness and oil-quenched and
tempered at 800 F has an ultimate strength of 180,000 psi and a yield point of 165,000
psi. The endurance limit for repeated stress will be approximately 90,000 psi. For
22 MACHINE DESIGN
design for finite life of the riveter, the endurance strength may be taken between the
infinite-life (say 10,000,000 cycles) value and the yield point (1 cycle). Let us assume
an endurance strength of 150,000 psi and a factor of safety of 2, which will make the
allowable stress equal to 75,000 psi. This value compares favorably wTith calculated
induced stresses in existing riveters. The design problem is to determine the depth of
the yoke 6.
Solution: By introducing at the gravity axis of the section two forces each equal in
magnitude to P and opposite in direction as shown in Fig. 3-10(6), one of these forces
and the original force form a bending moment on the section equal to P X e. This
moment induces tensile stress in bending at A and compression at B. The remaining
force induces a uniform tensile stress over the section and adds to the tensile stress due
to bending at .4 to produce a total tensile stress at A, as given by Eq. (3-12) for tension
loading with allowance for the effect of curvature from Eq. (3-8) as follows:
. P KPe = P_ 6 KPe
St A I/c wh wh2
In this equation, st can be made equal to the allowable stress, P and w are known,
and 6 is the unknown quantity on which K and e depend. The equation is in quad¬
ratic form so that a direct solution for 6 is not convenient.
Let us simplify the above equation as much as possible, as follows:
st _ 1 , 6Ke
P wh wh2
By substituting values for st, P, and h, and simplifying, the following form is
reached:
h2 - 0.1786 = imKe
This equation could be solved by assuming values for h and solving by trial but this
method would be laborious and not enlightening. It is helpful to note from Fig. 3-6
that if we obtain an approximate value for h, K will not change markedly for new
values of h. Let us assume a reasonable value for h and use that value to determine
K and e, then solve the equation for the unknown quantity h. One might expect h to
be, say, 2 in. The following values may then be found: e = 3 in., R = 2 in., R/c = 2
and, from Fig. 3-6, K = 1.6. Therefore
h2 - 0.1786 = 1.07Ke = 1.07 X 1.6 X 3 = 5.14
This equation could be solved directly for 6, but an inspection shows that the term
0.1786 will be small in comparison with the 62 term. Let us neglect for the present
0.1786 and solve for 6 by taking the square root of 5.14, which is approximately 2.25.
This value is larger than the assumed value of 6; therefore let us increase the value to,
say, 2)^2 in. and solve again by trial.
For 6 = 2Y2 in., e = 3.25, R = 2.25, R/c = 2, and K = 1.55.
Then
1.07Ke = 1.07 X 1.55 X 3.25 = 5.4
Now
62 - 0.1786 = 2.52 X 0.178 X 2.5 = 5.8
This value is larger than 5.4; hence, the correct value lies between 2 and 2.5. A
third trial of 6 = 2% yields an almost exact check; hence 6 will be specified as 2% in.
With reference to the initial equation for the maximum stress in the yoke, it is
evident that if the dimensions of the section are known, it is relatively easy to solve
for the stress. This is an analysis problem. But the design problem is one of deter¬
mining the dimensions of the member, using a reasonable value for the allowable stress.
In many cases, a direct solution is not possible; but the problem can be solved readily
by devising a procedure similar to that used in the example above in which relatively
small quantities can be tentatively disregarded in a trial solution, and trial values
may be assumed for other quantities which do not change markedly or for those on
which judgment can be exercised. This general approach is a subtle one, and should
not be confused with the usual conception of trial-and-error, hit-or-miss, or random-
guess methods. The method illustrated here is a powerful tool for the designer and
will be illustrated in other parts of this book.
3-7 Combined stresses. The various cases of loading, i.e., tension,

direct shear, bending, and torsion, as discussed, induce two fundamental
kinds of stress: (1) normal stress, as in tension, compression, or bending,
where the stress is normal to the area on which it acts, and (2) shear
stress, as in the case of rivet shear, or in torsion, where the stress acts
along the area. Whatever the condition of stress at a point may be, it
can always be expressed in terms of two components, the normal compo¬
nent and the shear component.
Most design problems deal with systems in which several types of load¬
ing exist simultaneously, such as bending and torsion in shafts. It is
necessary to determine the combination of stresses induced by the sepa¬
rate loads in order to determine the maximum value of normal stress and
shear stress, so that these may be compared with the allowable values of
the stresses for the material.
The following equations are given here for reference to determine the
maximum stresses in members subjected to loading which induces tensile
stresses in two perpendicular planes x and y and shear stresses in the xy
plane. The equations given here are for the case of combined tension
and shear and are derived in textbooks on mechanics of materials.
Let (st)x, (st)y, and (s8)xy represent calculated values for applied normal
and shear-stress components at any particular point in a member, (s*)i
and (s*)2 represent principal stress components, and (ss)max the maximum
shear stress. Then
2
(s,)i = (S<)l + + yj (&t)x (St) v

+ «?
2
(.St)x ($t)y
M, - (s,)x 2 {s,)“ - yl + S>
;
(St)x ($t)y
max + «.2
The maximum normal stress at the point is given by the first equation
above and the maximum shearing stress by the last equation for the case
of loading to which the equations apply.
24 MACHINE DESIGN
These equations can be represented graphically by a diagram known as

Mohr’s circle, as shown in Fig. 3-11.
In a design problem, the applied stresses (st)x, (st)y, and (ss)xy can be
calculated from the loading on the member; then it is desirable to deter¬
mine the maximum induced tensile and shear stresses to evaluate in
terms of allowable stresses. To draw the diagram for this case, the nor¬
mal stresses are laid off to suitable scale along the horizontal axis, and the
shear stresses along the vertical axis as follows: Lay off (st)x and (st)y as
shown to give points A and B. Lay off (s8)xy equal to AC = BD.
Fig. 3-11. Mohr’s circle for determining maximum normal and shear stresses (st) i and
(s4)max from applied stresses (st)x, (st)y and (sa)xy.
Through points C and D, draw a circle as shown. Then the maximum

normal stress (st) i will be equal to OE, the minimum normal stress (st) 2
equal to OF, and the maximum shearing stress (ss)max will be equal to GC.
The relations shown graphically by Mohr’s circle can be readily verified
by comparison with the three equations which the circle represents.
3-8 Types of failures. Each individual item in a modern junk yard
bears its own evidence of failure. Some of the parts failed early in life,
others outlived their life expectancy, and a few rest in obsolescence. The
cause of failure is not always evident but usually can be traced to a rela¬
tively small part of its structure or environment. Premature failures,
except those due to direct abuse or carelessness by the customer, are
within the responsibility of the designer. A study of the various types of
failure is a useful preliminary step in preventing or delaying failure of
parts or units in new designs.
Excessive deflection. An example of this type of failure was cited in
Art. 3-1 as a shaft which had sufficient strength but lacked necessary
stiffness so that the gears mounted on the shaft were not held in alignment
and failure of the gears was the result. In this example the shaft should
be considered to have failed as completely as though it had broken. Lack
of necessary stiffness may also cause failure of the bearings, or failure of
other parts due to excessive vibration of the shaft. Since steel, which is
used for most severely loaded machine parts, has practically uniform
modulus of elasticity, the control over stiffness lies in sufficient size of
parts and for steel is independent
of composition or heat-treatment.
Other examples of parts which
have rigidity as a critical charac¬
teristic are housings for gear drives,
supports for bearings of rolls where
controlled thickness is necessary as
Fig. 3-12. Failure of a brittle material by a Fig. 3-13. Failure of a ductile material by
static load. a static load.
in making photographic film, printing-press parts, and crankcases which

support crankshaft bearings.
Lack of strength. Failure due to this cause is usually obvious and is
evidenced by permanent set or fracture. Failure by fracture depends on
the type of material and on the loading. The failure of a part made of
brittle material is usually by fracture which occurs when the ultimate
tensile strength of the material is reached. A typical example of this
type of failure is the cast-iron clamp shown in Fig. 3-12. The fracture
occurred at a region of maximum tensile stress due to a static load induc¬
ing direct tension combined with bending due to eccentric loading.
A typical failure of a part made of ductile material and subjected to a
static load is shown in Fig. 3-13 in which fracture occurred on the steel
bolt in tension, and also some bending, as is evident. The fracture
exhibits plastic flow. In a technical sense actual failure of the bolt
occurred before the final stage shown in the figure, as the stress exceeded
26 MACHINE DESIGN
the yield point and the bolt became permanently deformed. This was a
front-bumper bolt on a large, loaded truck in which an attempt was made
to pull the truck, with one rear wheel in a hole, uphill and out of the hole
by a chain pull on the front bumper. When the bolt failed, the rear
wheel dropped back into the hole and the truck overturned, severely
injuring the driver and damaging the truck and cargo. It is an example
of abuse and carelessness by the customer.
Failure of a ductile material under cyclic loading exhibits a different
type of fracture than a static-load fracture. In cyclic loading the frac¬
ture usually begins at a region of concentration of stress and spreads,
usually slowly, as a fatigue crack
from the initial point until the re¬
maining section carrying the load
becomes so small that the part
breaks suddenly. An inspection of
the fracture usually shows the place
where the fatigue crack began, called
the eye. From that point the grad¬
ual spread of the crack is apparent
as a glossy area with no evidence of
plastic flow, and the remaining area
shows sudden failure, usually with
plastic flow as a ductile material
under static load. As the fatigue
crack spread, it may show evidences
of shock load.
The study of a fatigue crack is
very interesting and it presents the
record of events leading to failure.
Fig. 3-14. Failure of a ductile material
The fractured gear teeth of the
by a cyclic load. hypoid pinion in Fig. 3-14 illustrate
a typical failure of this type.
The prevention or satisfactory delay of fracture lies partly in the choice
of proper material and partly in avoiding forms and shapes as well as
finishes of machine members which tempt the formation of a crack. This
phase of failure is discussed extensively in Chap. 4, Stress Concentration
in Machine Members.
Buckling. Members or parts of members subject to compressive
stresses may fail by sideways buckling. Examples are parts of bridges,
ships, and aircraft; and in machinery, columns such as piston rods,
pressure vessels subjected to external pressure, and stiffening webs and
plates. Failure under conditions that lead to buckling is discussed in
Chap. 6.
Corrosion fatigue.1 When stress concentration in cyclic loading occurs

in a corrosive environment, the seriousness of stress concentration is
greatly increased. Even the presence of air reduces the endurance limit
of a member from its value in a vacuum. Ordinary tap water, salt water,
and other corrosive substances markedly reduce the endurance limit. It
is of interest to note that the problem of corrosion fatigue presented itself
Fig. 3-15. Section of boiler strap showing caustic embrittlement crack at rivet hole.
in 1915 in the steel cables of mine sweepers which were acted on by vibra¬
tion forces owing to the flow of sea water. Values are often quoted for
endurance limits in certain substances, such as the brine endurance limit
for steel. It is interesting to note that the corrosion effect on the mem¬
ber may not be great enough to “eat away” the material visibly. In
fact, polished test specimens that have failed in corrosion fatigue in salt¬
water spray appear as polished as the original test pieces, even though the
test may indicate an endurance limit that is 40 per cent of its dry-air
value.
Caustic embrittlement. This type of failure involves stress concentra¬
tion in the presence of a caustic corrosive substance and has appeared in
1 B. B. Weseott, Fatigue and Corrosion Fatigue of Steels, Mech. Eng., 1938, p. 813;
T. J. Dolan, Simultaneous Effects of Corrosion and Abrupt Changes in Section on
Fatigue Strength of Steels, Tran-s. ASME, vol. 60, p. A-141, 1938; Battelle Memorial
Institute, “Prevention of Fatigue of Metals,” p. 68, John Wiley & Sons, Inc., New
York, 1941.
28 MACHINE DESIGN
pressure vessels. Stress concentration has been produced by rivet holes,

unannealed welded joints, or sharp reentrant corners of flanges. The
loading has been of the fatigue type because of the repeated expansion
and contraction of the vessel. The fatigue stress concentration in the
presence of certain chemicals, such as caustic soda in feed water, has led
to embrittlement that has caused many boiler failures.1 Figure 3-15
shows a caustic embrittlement
crack at the rivet hole in a boiler
strap. Some means that are used
to combat caustic embrittlement
are (a) boiler feed-water treat¬
ments to reduce the effect of
caustic soda on the metal and
(b) design of the vessels to reduce
stress concentration.
Mechanical pitting} One type
of stress concentration in fatigue
appears as small pits at regions
of high contact pressures. These
have appeared in the contact
surfaces of heavy rollers, gear
teeth, and balls, rollers, and
races of rolling bearings. Ex¬
perimental evidence3 indicates
that there are three stages in the
formation of the pit: (a) the ap¬
Fig. 3-16. Pit formed by repeated rolling pearance of a small fatigue crack
contact. Enlarged. Motion of roller load
in the surface due to the roller
from top to bottom. (Courtesy of Westing-
house Electric Corporation.)
pressure; (5) in the presence of a
liquid, such as a lubricant which
fills the crack, subsequent applications of the roller pressure causes a
hydrostatic action in the pocket of entrapped liquid which causes the
crack to spread and form a small chip; (c) the chip breaks off and forms
a pit. The theory is based on observations which showed that the pits,
as shown in Fig. 3-16, form only in the presence of a liquid and that they
1 F. G. Straub, Embrittlement in Boilers, Univ. Illinois Bull. 216, 1930; A. E. White,
Changes in a High-pressure Drum to Eliminate Recurrence of Cracks due to Corro¬
sion Fatigue, Trans. ASME, vol. 61, 1929.
2 Called here mechanical pitting to differentiate this type of pitting from electrical
pitting, which is an electrolytic action which affects, for example, bearing surfaces of
electrical machinery. Another type of pitting is found in fluid flow in connection with
cavitation and affects runners, propellers, blades, and valves.
3 S. Way, Pitting Due to Rolling Contact, J. Appl. Mechanics, vol. 2, no. 2, p.
A-49, 1935.
are invariably fan-shaped and of the kind of orientation as predicted on

the hydraulic basis. The pit in Fig. 3-16 formed from a crack at the top,
progressed downward fanwise, and the chip finally broke off at the bottom
curved edge of the pit. The only apparent remedy for mechanical pitting
is to maintain the contact pressures below the surface endurance limit
of the material (see also Fig. 18-8).
Mechanical fretting. This type of damage occurs on metallic surfaces
which are finely finished and closely fitted and which have slight relative
sliding motion. The motion is of very small magnitude and may be due
to elastic deformation, as that between a ball-bearing race and the shaft
on which it is pressed, or it may be a slight oscillatory motion, as that
between the balls or rollers and their races of a variable-pitch-propeller
operating mechanism or between the leaves of a spring during flexing.
The damage manifests itself as a powder, or muddy substance if oil is
present, and a scabby appearing surface. While the damaged area is usu¬
ally in the form of isolated shallow pits, it may be serious enough to
affect the fit between the parts.
This type of damage has been believed1 to be due to corrosion or oxida¬
tion of small particles torn from the contacting surfaces and has been
termed “fretting corrosion.” It is now believed2 that the corrosive
action is secondary and that the primary cause is high pressure at con¬
tacting peaks or areas at the interface of the surfaces which causes plastic
deformation at these areas which in turn forms a mechanical interlocking
of the contacting peaks (see Fig. 20-7). When sliding motion takes place,
the areas of plastic deformation are strengthened (work-hardened). This
makes a bonded interlock stronger than the parent metal, and shearing
failure occurs at a distance beneath the surface so that a particle is torn
off. Loose particles then accumulate in low areas adjacent to the dis¬
rupted peak. These particles may later be drawn into the contacting
regions and may produce abrasive wear. To make matters worse, since
the particles are very small, they may oxidize, thereby becoming harder
and of increased volume so as to accelerate the abrasive action and cause
spreading of the affected area. Neighboring areas similarly affected may
merge and create a sizable affected region. As the regions increase in
size and fill with expanded (oxidized) particles, the pressure at the center
of the region increases owing to crowding and this further increases the
abrasive action so that a pit may form. These pits are of varying area
Tomlinson, Thorpe, and Gough, Fretting Corrosion at Forced Fits, Metal Progr.,
May, 1939, p. 468; Battelle Memorial Institute, “Prevention of Fatigue of Metals,”
p. 160, John Wiley & Sons, Inc., New York, 1941.
2 I-Ming Feng and Rightmire, The Mechanism of Fretting, Lubrication Eng.,
vol. 9, no. 3, p. 134, 1953; Uhlig, Feng, Tierney, and McClellan, A Fundamental
Investigation of Fretting Corrosion, NACA TN 3029, December, 1953.
30 MACHINE DESIGN
and depth and may eventually damage the surface so that the fit is
ruined.
The cure for fretting is not easy. Finer finishing in order to increase
the contacting area and thus decrease the pressure is not a general cure
since the contact is more intimate and fretting may be encouraged.
Decreasing the loads to reduce the pressure and also decreasing the
motion by using larger parts or reducing vibration are effective but not
always possible. Lubricating the surfaces is generally ineffective; in fact
it may aggravate the trouble since slipping may be increased.
The introduction of thin films of rubber and of nylon to cushion the
contact has met with some success in the laboratory. Badly fretted
surfaces have been ground off and rebuilt by plating or metal spraying.
CHAPTER 4
STRESS CONCENTRATION IN MACHINE MEMBERS
4-1 General discussion. In local regions of machine members, the

stresses may not be distributed as indicated by the elementary theory of
strength of materials. Such localized variations in stress distribution are
caused by sharp discontinuities in shape of the member or by surface
roughness or scratches.
These localized stresses are frequently of large magnitude and, under
the action of loads in service, may give rise to a crack at the region of the
localized stresses. Such a crack usually leads to failure of the member by
rupture. It is evident, therefore, that it is necessary for the mechanical
designer to have a knowledge of stress concentrations, especially in
regard to their causes and the means for avoiding them or lessening their
seriousness.
4-2 Causes of stress concentration. Variation in properties of mate¬
rials from point to point in a member. Some examples of this type of
stress concentration are (a) internal cracks and flaws, (b) cavities in
welds, (c) air holes in steel and concrete, and (d) nonmetallic inclusions.1
Pressures at points or areas at which loads on a member are applied.
Some examples of this type of stress concentration are (a) contact between
a wheel and rail, (b) contact between the balls and races of a ball bearing,
(c) contact between a beam and its supports, and (d) contact between
gear teeth (see Fig. 4-14).
Abrupt changes of section. There are two classes of this type of stress
concentration: the first involves surface condition, for example, scratches
due to improper machining or to roughness in handling. In machining,
the dimensions of the scratches or irregularities vary in size from large
in roughing operations to microscopic in fine finishing. Stress concentra¬
tion may also be caused by abrupt changes in form of a member. Stress
concentration of this kind may be extremely serious, and its prevention
lies squarely within the designer’s responsibility. Some common
examples of this kind of stress concentration are considered in Art. 4-4.
1 J. N. Goodier, Concentration of Stress around Spherical and Cylindrical Inclusions

and Flaws, Trans. ASME, vol. 55, p. APM 55-7, 1933.
31
32 MACHINE DESIGN
4-3 Stress concentration defined. The stress at a point in a member

when influenced by one or more of the causes discussed in Art. 4-2 is in
general greater than that determined by elementary strength of materials
and is referred to as stress concentration, or localized stress.
4-4 Stress concentration due to holes and notches. Tension member
with a transverse hole. In a prismatic bar subjected to a tensile load as
shown in Fig. 4-1 (a), the stress is uniformly distributed over a cross sec¬
tion AA of the bar.
If the bar has an elliptical hole, as shown in Fig. 4-1 (b), the stress at a
section CC remote from the hole will be uniformly distributed over the
section, but the stress over the cross section BB through the hole will not
Fig. 4-1. Effect of holes on stress distribution in tension members.
be uniformly distributed over the section carrying the load. The maxi¬
mum stress will be induced at the edge of the hole and it may have a value
several times that of the stress at the section CC.1 The value of the
maximum stress in terms of the stress at the section CC is given by the
expression
Smax — So ( 1 + 2 - (4-1)
in which s0 is the stress at the section CC, and b and c are the semiaxes
of the ellipse perpendicular and parallel, respectively, to the line of the
load as shown in the figure.2
1 In this discussion, the hole is assumed to be so small that it does not appreciably
reduce the cross section of the bar.
2 S. Timoshenko and J. N. Goodier, “Theory of Elasticity,” 2d ed., p. 84,
McGraw-Hill Book Company, Inc., New York, 1951. In the derivation of Eq. (4-1),
it was assumed that the bar is of infinite width. However, for bars of finite width
and with holes not larger than one-sixth the width of the bar, the equation may be
used without introducing large errors. For circular holes larger than one-sixth the
width of the bar, the experimental data in Art. 4-8 should be used to determine the
value of the maximum stress.
STRESS CONCENTRATION IN MACHINE MEMBERS 33
It may be seen from Eq. (4-1) that for large values of b/c, which repre¬
sents an ellipse approaching a transverse slit, smax reaches a very high
value.
For small values of b/c, which represents an ellipse approaching a longi¬
tudinal slit, smax is not markedly increased over s0.
For the case of a circular hole, b = c, and sn = 3s0. The stress dis-
tribution for this case is shown in Fig. 4-1(c).
Tension member with small notches. The stress
1 I 1 I f
concentration in the notched tension member in
Fig. 4-2 is influenced by the depth b of the notch
and the radius r at the bottom of the notch. The
maximum stress may be calculated by the following
equation,1 which applies to members having notches
which are small in comparison with the width of
_L
the bar. ■So
= So 1 + 2 b) (4-2) Fig. 4-2. Notch in

edge of tension mem¬
where So is the uniformly distributed stress at a ber.
section remote from the notch.
4-5 Visualization of stress concentration. The stress distribution in
tension members having transverse holes or notches, as discussed in
Art. 4-4, are representative of the few cases that have been solved by
analytical methods. For determining stress concentration in the maj ority
of members of the forms found in machinery, it is necessary to resort to
experimental methods, some of which are discussed in Art. 4-7. How¬
ever, there are qualitative methods that are useful to the designer in
visualizing probable stress concentrations in members. Two of these
methods of visualizing stress concentration are discussed below.
Shafts subjected to torsion. A qualitative representation of stress con¬
centration in circular shafts of variable cross section subjected to torsion
considers that the shaft is divided into a number of concentric tubes, each
of which carries an equal share of the twisting moment on the shaft.2 On
the section of a typical shaft shown in Fig. 4-3, the left-hand end is shown
divided into five concentric tubes, including the solid central one.3
1 C. E. Inglis, Stresses in a Plate Due to the Presence of Cracks and Sharp Corners,
Trans. Inst. Naval Architects, 1913.
2 L. S. Jacobsen, Torsional Stress Concentration in Shafts of Circular Cross Section
and Variable Diameter, Trans. ASME, vol. 47, p. 619, 1925.
3 The thicknesses of the tubes have been chosen so that the angle of twist per unit
length of each tube has the same value. This division results in a wall thickness h,
which decreases as the tubes become larger in diameter, as shown; and it is necessary in
order that the assembly of concentric tubes will twist without relative displacement
and thus, by fulfilling the condition of continuity, simulate the twisting of the solid
shaft.
34 MACHINE DESIGN
The right-hand end of the shaft is divided into five concentric tubes in
a similar manner. The sets of tubes for the left-hand and the right-hand
ends of the shafts are joined by smooth curves, as indicated in the figure.
Fig. 4-3. Circular stepped shaft divided into equitorque tubes.
It may now be noted that the thickness of the outer tube in the vicinity
of the fillet is very small, which indicates high stresses at the thin section.
It becomes apparent that means for maintaining the uniformity of thick¬
nesses of the equitorque tubes are identical with means for reducing stress
_J^= J^= --
(a)~Or\qlr\a\ ^-Improved ^-Improved

STEPPED CIRCULAR SHAFT
/WAA^VAAAA/^^'"'” ywwvwvwJy^
I
(ct)-Original ('e^-Improved
HOLLOW THREADED SHAFT
^-Original (g)~ Improved

SHAFT WITH TRANSVERSE HOLE
(i)-1m proved d/V-Improved

CRANK AND ARM
Fig. 4-4. Examples of stress reduction in members subjected to torsion.
concentration. Some of these means are illustrated in Fig. 4-4, in which

it may be seen that low stress concentration is indicated by gradual vari¬
ation of the distances between the equitorque tube curves. An interest¬
ing case is Fig. 4-4(j) in which the strength of a crank under repeated load¬
ing was increased 40 per cent by cutting an eccentric circular groove as
indicated in the improved section, although the groove lowered the moment
of inertia of the crankpin by 50 per cent.1 910105
Tension members. A similar method of representing stress concentra¬
tion in tension members makes use of lines indicating the direction of the
principal stresses. Figure 4-5 shows f | f,| f f tttttt tttittt
an axially loaded tension member with
the direction lines of the principal
stresses. In the upper end of the mem¬
ber, the lines are parallel and are drawn
equal distances apart to indicate uni¬
form stress distribution. These lines
are then extended to join with the (a)-0r\q\m\ ^-Improved (c)- mproved
tttittt
(d)- mproved
IMI
te/Mm proved r/7-Improved
Ôriginal ^‘Improved ^-Improved

Fig. 4-5. Stress direc¬ Fig. 4-6. Examples of reduction of
tion lines for a ten¬ stress concentration in members sub¬
sion member. jected to tension.
corresponding lines for the lower end of the member, as indicated.

In the vicinity of the fillet, it may be noted that the changes in slopes
of the stress-direction lines are pronounced and the lines are close together,
both of which may be shown to be an indication of stress concentration.2
As in the case of the equitorque tube visualization of torsional stress
concentration, the stress-direction lines for tension members may be used
to suggest means for reducing stress concentration by adopting shapes of
the member to favor equispaced and low-slope lines of stress direction.
Some of these methods are illustrated in Fig. 4-6.
1 Zeit. Ver. deut. Ing., vol. 76, p. 981, 1934.
2 R. V. Baud, Avoiding Stress Concentration by Using Less Material, Product Eng.,
May, 1934.
36 MACHINE DESIGN
4-6 Seriousness and mitigation of stress concentration. It is evident

from the preceding discussion that high concentration of stress may exist
in certain regions in machine and structural members. Since a crack
which may form at a region* of stress concentration invariably leads to
failure of the member, it is important that the designer reduce the con¬
centration to the practicable minimum. The designer can effectively
reduce stress concentration in most cases by simple expedients, such as
the use of gradual transition curves and generous fillets.
The seriousness of stress concentration depends on the properties of the
material and on the type of loading, namely, whether the load is static or
cyclic.
Static loading. It was noted in Art. 4-4 that the stress at the edge of a
small transverse circular hole in a tension member has a value three times
that in the sections remote from the hole. The distribution of stresses
across the section including the hole is indicated in Fig. 4-1 (c).
If the material of which the member is made has a stress-strain diagram
which is a straight line up to rupture, the proportions of the stress-distribu¬
tion diagram in Fig. 4-1 (c) will not be altered as the load on the member
is increased. This is true because the stresses would increase proportion¬
ately until the maximum stress in the member reached the breaking stress
of the material. At this stage, a crack would form at the edge of the hole.
The crack would have the effect of introducing additional stress concen¬
tration and of decreasing the section carrying the load. Both of these
effects would cause very rapid failure of the member.
Brittle materials, such as cast iron, undergo relatively little yielding;
therefore a member which has stress concentration and which is made of
brittle material will fail substantially as described above, since the con¬
centration of stress, as indicated by Fig. 4-1 (c), remains in the member
until failure occurs.
In considering the failure of a member made of a ductile material, such
as wrought steel, and with stress concentration, it is necessary to note
that the stress-strain diagram for the material indicates a region which
involves plastic flow. The plastic flow occurs beyond the yield point and
allows considerable strain to take place before failure occurs. The stress-
distribution diagram in Fig. 4-1 (c) will be valid if the stresses are below
the proportional limit; however, beyond the yield point, plastic flow at
the region of stress concentration will cause the stresses to be redistributed.
The redistribution of stresses will be one that will approach a uniform
distribution. To illustrate this change in distribution, one-half of the sec¬
tion in Fig. 4-1 (c) is shown in Fig. 4-7. Figure 4-7(a) shows an idealized
stress-strain diagram for a ductile material, and (b) represents a portion
of the bar adjacent to the hole and the stress-distribution diagram for low
stresses in the member. Figure 4-7 (c) represents the diagram for increased
stresses, and (d) shows the stage when the maximum or peak stress is just
equal to the yield point. Figure 4-7(e) shows the diagram for an increased
load on the member, and the corresponding crosshatched area indicates
the region which has undergone plastic flow, leading to the next stage (/).
Finally (g) indicates the diagram at the time the peak stress reaches the
breaking stress of the material, when a crack forms and the member fails.
From the accompanying sequence diagrams, it is apparent that plastic
Plastic state
»
i/i I
IA
<D Elastic Rupture
L 'state
CO
Strain
(a)-Idealized stress-strain diaqram
for ductile material
Fig. 4-7. Redistribution of stress, and plastic flow in tension member made of ductile
material [see Fig. 4-1 (c)].
flow in ductile materials will lessen the seriousness of stress concentration.

Frequently this plastic flow takes place in machine and structural mem¬
bers when the working loads are applied. Structural members with rivet
holes, key ways in shafting, and riveting or welding in ships that are
“broken in” in calm waters before being subjected to service on the high
seas are examples of this kind of reduction of stress concentration.
At times an intentional overload is applied to a member to cause plastic
flow and corresponding stress redistribution. Examples are found in
overstressing steel link chains and stressing generator rotors by overspeed
test operation.
38 MACHINE DESIGN
It should be noted that when a crack develops, the real breaking stress
of the material has been reached. In a ductile material, the real breaking
stress is appreciably higher than that indicated by the conventional stress-
strain diagram, which is expressed in terms of the original area of the test
specimen.
In concluding the discussion of the effects of the properties of material
on static stress concentration, it may be stated that stress concentration
in static loading is very serious in brittle materials, and it is less serious in
ductile materials owing to the relief of stress concentration by plastic flow.
However, the designer should reduce stress concentration wherever possi¬
ble without regard to the class of material of which the member is made.
Cyclic loading. In the preceding section it was stated that in static
loading stress concentration is very serious in members made of brittle
materials and is somewhat less serious in members made of ductile mate¬
rials. In the latter case it was noted that the seriousness of stress con¬
centration was lessened by the local plastic flow, which resulted in a more
favorable distribution of stresses.
In cyclic loading, however, stress concentration is always serious since
the ductility of the material is not effective in relieving the concentration
of stress caused by cracks, flaws, surface roughness, or any sharp discon¬
tinuity in the geometrical form of the member. If the stress at any point
in a member is above the endurance limit of the material, a crack may
develop under the action of the repeated load, and the crack will, in all
probability, lead to failure of the member. It is important to realize that,
even though the region subjected to the peak stress is extremely small,
the crack is liable to form, and once formed will lead quickly to failure of
the member.
The last statement is emphasized by the report1 that approximately 90
per cent or more of the number of cycles of application of load to produce
failure of a member are undergone before the crack may be detected by
laboratory means. It is evident then that the opportunity for an inspec¬
tor to detect a crack under service conditions in time to prevent failure is
relatively slight.
Figure 4-8 illustrates the effect of surface condition on the endurance
limit of test specimens, i.e., the marked reduction in endurance limit due
to surface condition. The figure shows also that high-strength alloy steels
are more sensitive to stress concentration. The lower ductility of the
high-strength alloy steels is partly responsible for the increased sensi¬
tivity, but it is not generally true that ductility is insurance against stress
concentration in fatigue.
The gain in endurance limit is the reason for polishing aircraft-engine
1 Battelle Memorial Institute, “Prevention of Fatigue of Metals,” Appendix 27,
connecting rods and link rods. The over-all polish is expensive, but it is
justified by the increase in safety. Figure 4-9 shows a polished link rod
for a radial aircraft engine.1
Fig. 4-8. Effect of surface condition on endurance strength. (Taken from A. V.

Karpov, Fatigue Problems in Structural Designs, Metals & Alloys, December, 1939.)
Fig. 4-9. Link rod for aircraft engine.
The endurance limit is defined as the maximum stress which a material

can withstand for an indefinite number of repeated applications of load¬
ing without failure of the specimen. This is shown by the endurance-
limit curve in Fig. 5-2.
1 Ibid. This source gives many illustrations of fatigue failure due to stress concen¬
tration and states that the record is replete with failures due to “notches, nicks, key-
ways, oil holes, screw threads, scratches, rough surfaces, quenching cracks, grinding
cracks, sharp changes in section, thin outstanding fins, poor fillets, tool marks, inclu¬
sions in the metal, corrosion pits, and the like, i.e., some localized nucleus from which
failure started.”
40 MACHINE DESIGN
The endurance limit is not strictly a property of a material although

it is frequently referred to as such, because it is affected by size, form, and
finish of the part as well as by the stress range. In the laboratory the
specimens are usually tested on a rotating-beam testing machine which
gives results for repeated, completely reversed bending. The specimens
are machined, according to specifications for one of the standard sizes,1
and polished, to avoid stress concentration and effect of surface roughness.
While it is necessary in machinery to use threads, oil holes, keyways, and
other forms that introduce concentration of stress, it is generally possible
for the designer to devise means for markedly reducing stress concen¬
tration. As has been discussed, this is especially important in mem¬
bers subjected to cyclic loading. When the designer realizes the serious¬
ness of stress concentration in cyclic loading, coupled with the frequent
difficulty of estimating service loads, the chance of occasional or acci¬
dental overloads, and the possibility of large forces due to resonant vibra¬
tion, he will take steps to reduce the loads to a minimum and to avoid the
use of geometrical forms that invite stress concentration. He will also
endeavor to select proper material, to specify proper heat-treatments and
surface finish, and to guard against careless machining, handling, and
inspection.2
Some means for reducing stress concentration that apply especially to
members subjected to repeated loading are illustrated in Fig. 4-10.
Figure 4-10(6) shows a circular fillet used to reduce stress concentra¬
tion. The circular fillet is the usual form, although it has been shown
that a spiral or parabolic fillet, as shown at (e), results in slightly less con¬
centration.3 The gain in the use of the noncircular-arc fillets is generally
not justified in view of the increased manufacturing cost. Figure
4-10(d) illustrates a separate collar used to provide a flat thrust cheek.
The design in (c) is preferred, however, to that in (d) where the flat cheek
is desired.
Figure 4-10(/) shows a cylinder flange bolted to the crankcase of an
internal-combustion engine. It is evident that stress concentration will
exist at the fillet joining the cylinder wall to the flange, at the corner of
the counterbore, and at the roots of the threads of the screw. Figure
4-10(<7) illustrates an improved design using large radii fillets and the
Aero thread employing an insert that reduces the stress concentration at
1 H. F. Moore and G. N. Krouse, Repeated Stress (Fatigue) Testing Machines Used

in the Materials Testing Laboratory of the University of Illinois, Univ. Illinois Eng.
Expt. Sta. Circ. 23, 1934.
2 Inspection marks and numeral stampings have been known to be the cause of
failure. See Product Eng., January, 1941, p. 25.
3 E. E. Weibel, Studies in Photoelastic Stress Determinations, Trans. ASME, vol.
56, p. APM-56-13, 1934.
the roots of the threads on the screw (see Art. 9-3 for discussion of the
Aero thread).
Figure 4-10(h) shows a hub of a gear, wheel, or sheave pressed on a
shaft. The press fit and loading will produce contact stresses in the
(aLOriginal f6l-lmproved (c) -Improved
Parabolic
Cylinder
sleeve
Spherical
washer
Aero-thread
cap screw
Crankcase
Aero-thread insert
Section of
insert
f/V-Originofl ^-Improved
fA)-Originoil ^-Improved Detail of groove

Fig. 4-10. Examples of reducing stress concentration.
shaft.1 The improvement at (i) employs a “relieving groove” cut in the

face of the hub. The groove increases the flexibility of the hub at its end,
thereby cushioning the contact pressures. A desirable form of the groove
is indicated in the sketch at (j).2
1 M. V. Barton, The Circular Cylinder with a Band of Uniform Pressure on a
Finite Length of the Surface, J. Appl. Mechanics, vol. 8, no. 3, p. A-97, 1941.
2 Peterson and Wahl, Fatigue of Shafts at Fitted Members with a Related Photo¬
elastic Analysis, J. Appl. Mechanics, vol. 2, no. 1, p. A-l, 1935; Buckwalter and
42 MACHINE DESIGN
The groups of examples shown in Figs. 4-4, 4-6, and 4-10 should be
regarded as commonly used methods for relieving stress concentration for
the particular cases shown. It is necessary for the designer to use his
ingenuity to devise appropriate methods for each particular case. An
acquaintance with the literature on the general subject of stress concen¬
tration is valuable for this purpose.* 1
In addition to avoiding forms that cause severe concentration of stress,
and selecting suitable material, there are means at the disposal of the
designer for increasing the fatigue strength of members in special applica¬
tions. Some of these are discussed below.
Cold working,2 3 4 especially rolling at fillets on shafts, may materially
increase the fatigue strength. In this process the shape of roller, the pres¬
sure, and the axial feed are important for best results. Good results have
been secured in railway axles by surface rolling.
Shot peening3 with steel shot has been used to increase the life of steel
springs, connecting rods, and gear-teeth fillets. Size of shot and time of
treatment are important for good results. The stress range of a coil
spring has been increased over 40 per cent by shot peening.
Under stressing or over stressing^ may increase the fatigue strength of
members. Overstressing must be induced very cautiously, preferably
by successively increasing loadings. This procedure has been called
“ coaxing.”
The problem of surface treatment5 6 is quite extensive. Beneficial
results are generally greatest for members in corrosive atmospheres.
Improper heat-treatment6 should be guarded against. A very common
result is surface decarburization, which results in a lower endurance limit
for the member. Quenching may cause surface cracks. Grinding off a
surface layer of such members may be beneficial.
Peterson, Locomotive Axle Failures and Wheel Press Fits, Railway Mech. Eng., April,
1935, p. 127.
1 “Three Keys to Satisfaction/’ Climax Molybdenum Company, New York.
2 O. J. Horger, Effect of Surface Rolling on the Fatigue Strength of Steel, Trans.
ASME, vol. 2, no. 4, p. A-128, 1935.
3 F. P. Zimmerli, Flow Shot Blasting Increases Fatigue Life, Machine Design,
November, 1940, p. 62.
4 See Battelle Memorial Institute, “Prevention of Fatigue of Metals,” p. 87, John
Wiley & Sons, Inc., New York, 1941; J. B. Kommers, Effect of Understressing and
Overstressing in Fatigue, Proc. Soc. Exp. Stress Anal., vol. 3, no. 2, 1946.
5 Frye and Kehl, The Fatigue Resistance of Steel as Affected by Some Cleaning
Methods, Proc. ASTM, vol. 26, p. 192, 1938; Dolan and Benninger, The Effect of
Protective Coatings on the Corrosion-fatigue Strength of Steel, Proc. ASTM, vol. 40,
p. 658, 1940; Speller, McCorkle, and Mumma, The Influence of Corrosion Accelera¬
tors and Inhibitors on Fatigue of Ferrous Metals, Proc. ASTM, vol. 28, II, p. 159,
1928.
6 Battelle Memorial Institute, op. cit.
4-7 Experimental methods for investigating stress concentration.

The stress-concentration factor has been defined as the ratio of the
maximum stress in a member to the stress at the minimum section as
calculated by the use of an equation from the elementary strength of
materials.
In determining the maximum stress, it is generally necessary to resort
to one of the experimental methods. A knowledge of these methods is of
great value to the designer since such knowledge enables him to appreciate
stress concentration and the applications and limitations of the factors,
and also since a firsthand acquaintance with the experimental methods
aids the designer in visualizing the effects of various design expedients in
Fig. 4-11. Stress-strain diagram for cured plaster in compression. (From Roark and
Hartenberg, Predicting the Strength of Structures from Tests of Plaster Models, Univ.
Wisconsin Bull. 81, 1935.)
reducing stress concentration. The photoelastic method is especially

useful in this respect and laboratory experience with that method is
invaluable to the mechanical designer.1
Brittle-material method. Some brittle materials have an almost con¬
stant stress-strain characteristic to rupture. Stress concentration in a
member made of such a material will exist until rupture occurs, as dis¬
cussed in Art. 4-6. Since the load at rupture of the member can be
determined by test, this method enables one to compare the peak stress at
the point of stress concentration with the calculated stress and thus to
determine values for stress-concentration factors.
Pottery plaster is one of the brittle materials that is convenient to use
for models of machine members. Figure 4-11 is a stress-strain diagram
for a high-grade pottery plaster. In determining the stress-concentration
factor for a shaft with a shoulder, for instance, a plaster model of the
1 G. H. Lee, “An Introduction to Experimental Stress Analysis,” John Wiley &

Sons, Inc., New York, 1950; M. Hetenyi, “Handbook of Experimental Stress Analy¬
sis,” John Wiley & Sons, Inc., New York, 1950.
44 MACHINE DESIGN
shaft is cast and machined. Figure 4-12 shows such models. The model
is then gradually loaded in a manner that simulates the actual loading of
the member, such as torsion or bending,
and it is closely observed as the load is
applied in order to detect the formation
of a crack. The crack will form at the
most highly stressed region, and when
it forms it is apparent that the stress at
the region is the breaking stress of the
plaster, and is the maximum stress in
the model. By the use of the loading
at the instant the crack appeared, the
nominal stress at the region of the
crack is calculated by the use of the
appropriate equation. The ratio of
the actual maximum stress to the cal¬
culated stress is the static-stress con¬
centration factor.
Fig. 4-12. Cast plaster models of
The brittle-material method may
shafts with fillets. (Courtesy of
be used for one-, two-, or three-di¬
University of Illinois Engineering
Experiment Station.) mensional problems.1 The results are
quantitative.
The rubber-model method. This method may be used for one-, two-, or
three-dimensional problems; it employs rubber models that are ruled
with cross-sectional lines. When the model is loaded, the distortion of
the squares is a qualitative indication of strains that may be interpreted
in terms of stresses. The regions of stress concentration are generally
plainly apparent. Figure 4-13 shows a rubber model in the unloaded and
loaded conditions.2
Short gauge-length method. This method employs a delicate extensome-
ter that has a short gauge length (0.1 in. by R. E. Peterson). The strain
is measured at critical points on the surface of the machine member. The
method requires a highly developed technique but gives very accurate
and informative results.3
1 Seely and James, The Plaster-model Method of Determining Stresses as Applied

to Curved Beams, Univ. Illinois Bull. 195, 1929; Seely and Dolan, Stress Concentration
at Fillets, Holes and Keyways as Found by the Plaster-model Method, Univ. Illinois
Bull. 276, 1935; Roark and Hartenberg, Predicting the Strength of Structures from
Tests of Plaster Models, Univ. Wisconsin Bull. 81, 1935.
2 F. B. Seely, “Advanced Mechanics of Materials,” p. 199, John Wiley & Sons,
Inc., New York, 1932; Brewer and Glassco, Determination of Strain Distribution by
the Photogrid Process, J. Aeronaut. Sciences, vol. 9, no. 1, November, 1941.
3 Peterson and Wahl, Two- and Three-dimensional Cases of Stress Concentration,
and Comparison with Fatigue Tests, J. Appl. Mechanics, vol. 3, no. 1, p. A-l, 1936.
The membrane analogy, or soap-film method. This method is used to

investigate stresses over cross sections of bars subjected to torsion. A flat
plate is used which has a hole which is cut the same shape as that of the
cross section of the bar. A thin membrane, such as a soap film or rubber
sheet, is stretched over the hole and uniformly loaded from one side by
>/////.«/y//,, sfi
//, m-- m
'stfs/stf/v;'//,-.
, , ,
mm ■'
■■ .
Fig. 4-13. Rubber model. Unloaded model above; axially loaded model below.
air pressure. The analogy establishes certain relations between the

deflected surface of the membrane and the stresses in the twisted bar.1
One of these relations is that the slope of the membrane at any point is
proportional to the shearing stress at the corresponding point in the
twisted bar. It is possible to visualize the location of the maximum stress
by picturing the analogous soap film. The sand-heap analogy has been
1 S. Timoshenko, “Strength of Materials,” II, p. 623, D. Van Nostrand Company,
Inc., New York, 1930.
46 MACHINE DESIGN
developed to be used in connection with the membrane analogy to investi¬

gate the formation and development of plastic flow in torsion members.1
The photoelastic method. This method uses polarized light and trans¬
parent models, which are generally cut from sheets of homalite. By
means of images produced by interference of light rays, the stresses in the
loaded model may be investigated. Figure 4-14 is a photograph of the
photoelastic image of a loaded gear tooth. The region of stress concen¬
tration is indicated by the closely spaced fringes at the fillets of the gear
Fig. 4-14. Photoelastic photograph of loaded gear tooth. (Courtesy of University of

Illinois Engineering Experiment Station.)
tooth. The photoelastic method is generally applied to two-dimensional

problems, although it is being developed to solve three-dimensional
problems.2
The electric strain gauge has found extensive use in industry on account
of its accuracy, its adaptability to all kinds of structures and parts to
indicate local surface conditions of strain, and its ability to indicate either
1 A. Nadai, “Plasticity/’ p. 132, McGraw-Hill Book Company, Inc., New York,
1931; M. A. Sadowsky, An Extension of the Sand-heap Analogy in Plastic Torsion
Applicable to Cross Sections Having One or More Holes, J. Appl. Mechanics, June,
1941.
2 M. M. Frocht, “ Photoelasticity/’ John Wiley & Sons, Inc., New York, 1941;
R. D. Mindlin, Review of Photoelastic Method of Stress Analysis, J. Appl. Phys.,
April and May, 1939. The latter gives an extensive bibliography.
static or dynamic strains. In the latter use an oscilloscope or an oscillo¬

graph are necessary. The most widely used type of electric strain gauge
is the resistance-wire type (the SR-4 gauge made by Baldwin-Lima-
Hamilton Corporation) which is bonded to the part under investigation.
Its operation is based on the change in electric resistance of a fine wire as
it changes in length when the part is strained. The gauges are small and
light (so-called postage-stamp gauges) and have been used to measure
strains in many applications, for instance, in fillets of engine crankshafts
of aircraft in flight, in drive rods and crank pins of locomotives, and in
concrete structures such as dams. These gauges are used in many fields,
for example, for measuring torques in rotating shafts and also as internal-
combustion engine pressure indicators.
The use of brittle coatings on machine parts has extensive use for a
qualitative measure of strains and in indicating the directions of principal
stresses. In this method the part is sprayed with a lacquer which becomes
brittle on drying. When the part is loaded, cracks form at the region of
high stress. The pattern of the cracks indicates the distribution of princi¬
pal tensile stresses and the closeness of the cracks may be used to estimate
the magnitude of the stress. This method is frequently used to indicate
the direction of the principal stresses prior to installing electric strain
gauges so that they may be orientated most advantageously. Brittle
coating and equipment is made by the Magniflux Corporation under the
trade name Stresscoat.
4-8 Stress-concentration design factors. Three common types of
loading in machine members are tension, bending, and torsion. Tension
and bending induce normal stresses s, and torsion (as well as direct shear)
induces shear stress ss.
There are two kinds of design factors which we must use to take into
account concentration of stress, one for flat bars in which cases the maxi¬
mum stress at the notch or fillet is a uniaxial system. The appropriate
design factor is termed theoretical stress-concentration factor. These are
theoretical factors because they are determined mathematically or by the
photoelastic method.
The other kind of design factor, as discussed later, applies to grooved
and stepped shafts in which cases the stress system is biaxial and is taken
into account by a notch factor.
The types of simple loading on members without and with stress con¬
centration are shown in Fig. 4-15 and are intended to illustrate what the
theoretical stress-concentration factor is and how it can be used in design
equations. There are of course other kinds of loading and many other
forms of members; those shown are samples. In design, each case is a
special one from the standpoint of loading (tension, bending, and torsion,
or a combination of these); shape of the member (bar with transverse
48 MACHINE DESIGN
hole, shaft with circumferential groove, stepped shaft, etc.); type of

material (brittle or ductile); and kind of loading (static or cyclic). This
article is intended to acquaint the designer with the appropriate methods
for selecting the factor and properly using it and also to help him gain an
appreciation of its importance.
The theoretical stress-concentration factor may be defined as the
maximum stress in a member (at, for instance, a notch or a fillet) divided
by the nominal stress at the same section as calculated by a simple equa¬
tion from strength of materials. The maximum stress is determined
mathematically or photoelastically. The nominal stress is usually calcu¬
lated across the net or minimum section.
The notch factor is defined as the strength of an unnotched member
divided by the strength of a notched member of the same size and
material.
Stress-concentration design-factor charts for several common shapes of
members are given in Appendix X.1
In Fig. 4-15(a) is shown an example of tension on a uniform bar in
which the tensile stress is uniformly distributed over cross sections and is
equal to P/A, where A is the cross-sectional area. At (6) is shown a simi¬
lar bar with notches. As discussed in Art. 4-4, the notches cause a con¬
centration of stress so that the maximum stress smax is greater than P/A
and is equal to Kt X P/A. In this case A is the net area across the
notches. By definition, then, Kt is the ratio of the maximum stress to
the nominal stress P/A. Its value is always greater than unity.
For design purposes, we may solve the equation for the area
P
A = Kt X ——
Smax
In the design of a machine member, the applied load can be determined

from the loading data. The value of smax can be limited to the allowable
stress as discussed in the following chapter. The value for Kt may be
found or estimated from available data such as the charts in Appendix X.
In Fig. 4-15(c) is shown a uniform beam in bending for which the equa¬
tion s = Me/1 applies, and at (d) is a beam with notches for which
Smax = Kt X Mc/I. At (e) and (/) are shown the case of torsion. Here
the stress-concentration factor is Kts, the s in the subscript referring to
shear stress, so that ss(max) = Kts X Tr/J.
1 These charts as well as symbols, nomenclature, and methods are taken by per¬
mission of the author and publisher of “Stress Concentration Design Factors,” by
R. E. Peterson, John Wiley & Sons, Inc., New York, 1953. This source contains dis¬
cussions of the determination of the factors, an extensive bibliography, and some 94
charts for factors for a wide variety of types of loading and shapes of members. Six¬
teen of these charts have been selected for inclusion in Appendix X.
In all cases in computing the area or section modulus, the net section at
the notch, fillet, hole, etc., should be used (along with the stress-concen¬
tration factor) to determine the quantity desired.
In notched flat bars and stepped bars, the maximum stress at the notch
or fillet is a uniaxial system. In a grooved or stepped shaft in tension or
bending, however, in addition to the maximum stress as given by the
equation containing the stress-concentration factor as discussed above,
there is also a stress at right angles to the maximum stress, i.e., the stress
S S max
£= P/A sma,’KtP/A
(a) Tension of uniform bar (b) Tension of notched bar
a -7
(•M A >
m)
>i
s max
M
I/C £ max Kf.
I/C
(o) Bending of uniform beam (d) Bending of notched beam
-T-e
c
k*
s.
J/r
(e) Torsion of uniform shaft (f) Torsion of grooved shaft

Fig. 4-15. Simple cases of loading without and with stress concentration.
system in biaxial. When we have two stresses at right angles, we need a

theory of strength in order to predict failure. The Mises1 criterion gives
values which are in good agreement with results of failure tests of mem¬
bers made of ductile material. By taking into account both stress con¬
centration and the Mises criterion, a new factor has been determined and
labeled Kt>, the prime indicating the combined factor. For members in
which the highest stressed element is subjected to biaxial tension, the
value of Kt> is always less than Kt, so that if the former is not available the
latter may be used. The greatest difference is not over 15 per cent. In
the charts, values for the case of shear (Kt's) are not given for torsion of
1 R. von Mises, Mechanik der festen Korper im plastisch deformablen Zustand,

Nach. Ges. ITTss. Gottingen Jahresher, Geschdftsjahr, Math-physik. Kl., 1913, p. 582;
also A. Nadai, Theories of Strength, Trans. ASME, vol. 56, p. APM-55-15, 1933.
50 MACHINE DESIGN
round shafts because with notches or fillets they are the same as Kts
values.
In cyclic loading, the effect of the notch or fillet is usually less than
predicted by the use of the theoretical factors as discussed above, although
for high-strength steels the full theoretical factor is generally obtained.
The difference depends on the stress gradient in the region of stress con¬
centration and on the hardness of the material. The term notch sensi¬
tivity has been applied to this behavior. Notch sensitivity may be
defined as the degree to which the theoretical effect of stress concentration
is actually reached. The stress gradient depends mainly on the radius of
the notch, hole, or fillet and on the grain size of the material. While
extensive data are not available at present, the curves in Fig. 4-16 may be
Fig. 4-16. Average notch sensitivity. (“Metals Handbook1954 Supplement, Metal

Prog., July 15, 1954.)
used to estimate the value of the notch-sensitivity factor q for two steels.
When data are not available for determining values for q, it can be
neglected; i.e., use q equal to unity, and the design will be on the side of
safety.
In cyclic loading, therefore, the notch-sensitivity factor is used to
modify the stress-concentration factor and a new factor is obtained which
is termed estimated fatigue notch factor. This factor may be obtained as
follows:
Ktf = 1 + q(Kt - 1) (4-3)
Kt'f = 1 + q{Kt> - 1) (4-4)
Ktsf = 1 + q(Kts — 1) (4-5)
where Kt = theoretical stress-concentration factor for normal stress (ten¬

sion or bending)
Kt8 = theoretical stress-concentration factor for shear stress
(torsion)
Kt' = theoretical combined factor which takes account of both

stress concentration and strength theory (Mises criterion)
Ktf = estimated fatigue notch factor for normal stress
Kt'f = estimated combined fatigue factor for normal stress
KUf = estimated fatigue notch factor for shear stress
q = notch-sensitivity factor for cyclic loading
Table 4-1 may be helpful in collecting the various procedures sug¬
gested in this article. For ductile materials under static loading, stress
concentration usually may be neglected; i.e., use a value for the factor
equal to unity, for reasons discussed in Art. 4-6.
Table 4-1. Chart for Guidance in Selecting Stress-concentration
Factors for Various Types of Loading and Materials
Brittle material Ductile material

Loading
Normal Shear Normal Shear
Static Kt Kt. N eglect Neglect

Cyclic Kt Kt. Kt'f Kuf
For round bars with circumferential groove or shaft with shoulder

fillet, if q values are not available, use Ktf values.
Example 4-1. Determine the size of a shaft to be subjected essentially to a cyclic

bending moment of 4,000 in.-lb. The shaft is to be ground and with a semicircular
groove in. in radius. The groove is part of a sealing arrangement. Assume that
the shaft is annealed 1030 steel and that the factor of safety is 2. (Using data from
Chap. 5, the allowable stress may be found to be approximately 16,000 psi.)
Solution: Let us assume that the ratio of r to d (Fig. X-9) is approximately
so that if the shaft is found to be 1 in. in net diameter at the groove, the ratio r/d will
be correct; otherwise the assumption may be modified for a second trial.
If r/d is then D/d will be 1.25 and K\ = 1.62 from Fig. X-9.
From Fig. 4-16,
q = 0.9
From Eq. (4-4),
Ktff = 1 + q(Kt• — 1)
= 1 + 0.9(1.62 - 1) = 1.56
Then
7rdz KffM 156 X 4,000
= 0.393 in.3
16 S(all) 16,000
From Appendix IV,
d = 1^6 in. and D = 1%6 + H = We in.
The actual ratio r/d is 0.125/1.3125 = 0.99 instead of the assumed value 0.125, and
the actual value of q is 0.85 instead of 0.9. Also the actual value of D/d is 1.19. A
check will show that it is not worthwhile to make a second solution.
CHAPTER 5
ALLOWABLE STRESSES
5-1 General. In Chaps. 3 and 4, relations were discussed for deter¬

mining the values of stresses which are induced by applied loads on
machine members. The effects of impact loading, and of sharp changes
in section and other “stress raisers’’ were considered so that the designer
can calculate or estimate the probable value of the maximum stress1 in
the part being designed. This stress is referred to as the maximum
induced stress or simply the induced stress.
In many cases the loads are not known closely or the factors for stress
concentration are not known for the particular shape of the part, so that
in predicting maximum stresses the designer may resort to experimental
stress analysis on models or he may obtain data from similar parts in
service. This is common practice in the design of members of complex
shape, such as automobile connecting rods, crankshafts, parts of firearms,
complicated castings, etc. In any event the designer determines as
closely as possible the expected maximum induced stress in the member
being designed. This is usually in the form of an equation with the
cross-sectional dimensions of the member as the unknown quantities.
The next step is to choose the material for the member, basing the
choice partly on its mechanical properties so that its allowable stress will
be at least equal to and frequently greater than the maximum induced
stress. The allowable stress is thus the upper limit to the induced stress.
In order to provide a margin against failure, it is common practice in
machine design to determine the allowable stress by dividing the failure
stress for the member by a factor of safety. The failure stress depends
squarely on the method of failure and is discussed in the following articles.
5-2 Allowable stresses for static loading. For members made of
ductile material and subjected to static loading, failure occurs either if the
part fractures or if it becomes permanently deformed. Both of these
1 In case there are initial stresses expected in the part due to casting, fabrication, or
heat-treatment, these may also be included in the maximum stress; however, initial
stresses are difficult to estimate, so that if it is not feasible to relieve them by annealing
or other means, their influence may more properly be dealt with by using an increased
value of the factor of safety.
52
ALLOWABLE STRESSES 53
modes of failure can be prevented if the maximum stress in the part is

kept below the elastic limit (or the yield point which for steel has prac¬
tically the same value) as follows, where f.s. is the factor of safety:
«■ - ft (5-1}
where sa = allowable stress, psi

sy = elastic limit, or yield point, psi
The failure load for a statically loaded member made of ductile mate¬
rial is not markedly affected by stress concentration because of the redis¬
tribution of stress that takes place beyond the yield point. Therefore in
the design of a member made of a ductile material to support a static load,
it is permissible, generally, to modify the value of the stress-concentration
factor or to neglect it entirely. The following general rules may be
applied: If it is desired to prevent plastic flow entirely, use the full value
of the stress-concentration factor. If plastic flow at the regions of stress
concentration is not objectionable, as is usually the case, neglect the
effect of stress concentration, that is, use Kt = 1. In some cases it may
be desirable to use values of Kt between its full value and unity.
In Fig. 3-13 is shown a steel bolt that failed in static loading caused by
simultaneous tension and bending. This is failure as a ductile material
and necking down is apparent. Stress concentration caused by the
threads had little or no effect on failure.
For materials which do not have a well-defined yield point, as cast iron,
the allowable stress for static loading may be based on the ultimate
strength with a factor of safety larger than for a ductile material. In
members made of brittle material there is practically no plastic flow of the
material to relieve stress concentration; hence the full value of the factor
should be used.
In Fig. 3-12, the cast-iron C clamp failed in static loading at the point
where the curved-beam effect would indicate maximum tensile stress.
The fracture is typical of brittle material and shows no plastic flow.
5-3 Allowable stresses for cyclic loading. In this type of loading,
failure generally occurs at a stress lower than for static loading. There
are two reasons for this: first, the stress at failure is the endurance limit
of the material which is lower than its yield point. The endurance limit
is usually determined from tests on polished specimens with no abrupt
changes in section. Second, in machine members, stress concentration
is invariably present in some degree owing to necessary changes in section
or to other causes such as rough finish, and, in cyclic loading, the ductility
of the material does not reduce stress concentration; instead a fatigue
crack may develop, which usually leads to rapid failure of the part. It is
therefore necessary in cyclic loading to base the allowable stress on the
54 MACHINE DESIGN
endurance limit of the member and always to use the appropriate stress-
concentration design factor. The endurance limit for a material depends
on the finish of the part (see Appendix VII) and also on the range of stress
for the particular loading, as discussed in Art. 5-4.
The allowable stress for cyclic loading may be determined as follows:
where sa = allowable stress, psi

se = endurance limit, psi
f.s. = factor of safety
In cyclic loading, stress concentration is serious whether the material
is ductile or brittle; hence the full value of the appropriate stress-concen¬
tration design factor should be used.
The hypoid pinion teeth shown in Fig. 3-14 failed in cyclic loading by a
fatigue crack1 which started at the tension fillet of the tooth and spread
until the tooth broke off. Two teeth were thus affected. There was no
plastic flow exhibited on the surface of the fracture.
5-4 Endurance limit and range of stress. The rotating-beam testing
machine2 is used in the majority of fatigue tests; hence the endurance
limits reported as results of these tests are for completely reversed stresses.
Since many machine members undergo different ranges of stress than the
completely reversed, it is convenient for design purposes to know the
effect of the range of stress on the endurance limit.
The type of stress as shown in Fig. 5-1(5) is a completely reversed stress
in which the maximum stress in tension equals the maximum stress in
compression, as in the rotating-beam testing machine.
The type of stress shown at (d) has a range of stress that may be speci¬
fied by stating smin and smax. This stress may be considered as a steady
(mean) stress sm and a completely reversed stress sr superimposed on the
steady component. A reversed stress that is superimposed on a com¬
pressive stress is shown in Fig. 5-1 (a).
The endurance limit of a material for a particular range of stress is the
maximum induced stress that the material can withstand without failure
for an indefinite number of repetitions of the stress. In Fig. 5-2 is shown
a diagram for determining the endurance limits for specimens made of
three steels. These results were determined from rotating-beam tests.
1 The monograph “Fatigue and Fracture of Metals/’ William M. Murray (editor),
published jointly by the Technology Press of Massachusetts Institute of Technology
and John Wiley & Sons, Inc., New York, 1950, contains an excellent review of this
subject.
2 Moore and Krouse, Repeated-stress (Fatigue) Testing Machines Used in the
Materials Testing Laboratory of the University of Illinois, Univ. Illinois Eng. Expt.
Sta. Circ. 23, 1934.
ALLOV/ABLE STRESSES 55
In the design of members subjected to other ranges of stress than com¬

pletely reversed stress, it is necessary to know the effect of the range of
stress on the endurance limit. A formula that fits experimental data
Fig. 5-2. Fatigue test results plotted for endurance-limit determination.
reasonably well is given by Moore and Kommers.1 This formula is given

in terms of the stress ratio r instead of the range of stress. The stress
ratio is used, since in the design of a member the values of the minimum
and maximum stresses are unknown, but their ratio, within the elastic
1 H. F. Moore and J. B. Kommers, “The Fatigue of Metals,” p. 185, McGraw-Hill

Book Company, Inc., New York, 1927; see also J. 0. Smith, The Effect of Range of
Stress on the Fatigue Strength of Metals, Univ. Illinois Eng. Expt. Sta. Bull. 334,
1942. Note that in Art. 5-7 a more refined method is discussed for dealing with range
of stress.
56 MACHINE DESIGN
limit, is equal to the ratio of the minimum to maximum applied loads,

which is known generally.
(5-3)
where s'e = endurance limit for any stress range represented by r

se = endurance limit for completely reversed stresses as found by
the rotating-beam testing machine
minimum stress
r = stress ratio = for completely reversed
maximum stress’
stresses, r = —1; for repeated stress, r = 0; r cannot be
greater than unity
Experimental evidence indicates that the value of se for polished speci¬
mens of wrought ferrous metals is approximately one-half the ultimate
strength in tension and also that the endurance limit of wrought ferrous
metals in repeated completely reversed torsion, or shear, is approximately
0.55se.
5-5 Design for requirements other than strength. Stiffness. While
all machine members must have sufficient strength to prevent either
fracture or permanent deformation, many machine parts have, in addi¬
tion, limits to their deflection. Limits to deflection are particularly
necessary in shafts which carry gears so that excessive wear of the teeth
due to misalignment will be avoided. Also, the shaft must not bend so
much as to bind it in its bearings. Shafts of motors and generators must
be stiff enough to maintain the correct air gap between the rotor and
stator. Machine tool shafts must be stiff enough to give accurate cuts
without chatter marks. In many moving parts, vibration may be objec¬
tionable, and vibration is a function not of strength of a member or a
system but of its deflections. Actually there are more industrial machine
parts that have their dimensions determined on the basis of stiffness than
on strength requirements.
In considering stiffness in the design of a part, two difficulties arise: (1),
the limits to deflection are not as closely known as are limits to stress; (2)
it is generally impossible to establish an expression for the deflection in
terms of the loading in an actual machine part being designed. After the
dimensions of the member have been determined, it is usually compara¬
tively simple to calculate the probable deflections at critical locations.
The use of graphics (see Appendix XIII) is helpful in many cases.
It is a common procedure in designing for stiffness first to determine
the dimensions of the part on the basis of strength, using a value for the
allowable stress determined from analyses of similar existing parts in
similar service which have performed satisfactorily. Many design codes
are written on this basis and frequently in checking back on the factor of
safety it is found to be higher than might be expected; values of 10 or

higher are not uncommon. These values should be thought of, not as
giving a member which is, say, ten times as strong as it need be, but rather
that the corresponding allowable stress gives a member which will prob¬
ably have satisfactory deflection characteristics. After the dimensions
of the member have been thus determined for strength, the deflections at
the necessary critical points can then be determined and compared with
values regarded to be in the range of allowable values. In checking the
vibration characteristics, the natural frequency of the moving part can
be calculated after the deflections have been determined. This procedure
is discussed in detail in Chap. 24.
Pressures. At points of contact between parts, the pressure between
them may be the basis for design in order to prevent failure. In some
cases there is virtually no movement between the surfaces, as in rivets,
rolling bearing races where they contact the housing or the shaft, and in
serrations in fittings. These parts may fail by crushing of the surfaces or
by mechanical fretting.
At points of contact between surfaces that have relative movement, the
surfaces may fail by pitting, by galling, or by other types of wear. Exam¬
ples are gear teeth, rolling elements in rolling bearings, pins and bushings
of power chains, cams and the followers, and wheels or rollers on rails.
These types of failure depend, among other factors as discussed in this
book, on pressure between the parts, which, if kept within proper limits,
may prevent failure or delay it for a reasonable period. Allowable pres¬
sures may be determined on this basis from parts which have behaved
satisfactorily in service and may be used for design purposes.
5-6 Factor of safety. As discussed earlier in this chapter, the factor of
safety is a number used to divide into the appropriate property of the
material to obtain the allowable stress. In design for strength to avoid
fracture and permanent deformation, the appropriate property for static
loading of brittle material is the ultimate strength, and for ductile mate¬
rial is the yield point, while for cyclic loading the appropriate property is
the endurance limit. The allowable stress thus determined is then
equated to the induced stress for the part, and calculations are then made
for the dimensions. The induced stress should include any effects of
stress concentration and shock loading.
In some design offices, the stress-concentration factor and the shock
factor are included in the factor of safety, but that method is not used
here since those two factors can be estimated fairly closely. In addition,
they usually vary from section to section in the member so that if we
included them in the factor of safety, we would have the anomaly of a
different allowable stress for each section of the member.
The above discussions apply to members which are designed for strength.
58 MACHINE DESIGN
For stiffness, the allowable stress is usually determined from design codes
or from analyses of similar members which have performed satisfactorily
under similar service conditions. In this case the designer may well avoid
thinking of the allowable stress in terms of a factor of safety.
The choice of the value for the factor of safety depends largely on the
judgment of the designer and should cover such considerations as the
following:
1. Degree of certainty of the loading
2. Reliability of material
3. Initial stresses
4. Whether assumptions made in the analysis are on the side of safety
5. Whether failure would endanger life, damage expensive machinery
or equipment, or cause expensive shutdowns
When effects of stress concentration and shock loading, if any, are
included in the calculations for the induced stress, or if the induced stress
is estimated from methods of experimental stress analysis, the factor of
safety for ductile material in static loading based on the yield point is
usually in the range of 1.5 to 2.5 for industrial equipment. The lower
values are used when low weight and cost have precedence over high
dependability and insurance against any failure. In many parts made in
production quantities, and as dictated by quality-control requirements,
a certain percentage of failures can be tolerated if such failures of the part
would not endanger life or cause excessive inconvenience or expense.
Usually failures in power-transmission machinery are expensive so that
efforts to prevent any failure are well justified.
5-7 Allowable stresses for combined variable loading. As has been
discussed early in this chapter, we have assumed that failure in a ductile
material subjected to static loading occurs when the induced stress reaches
the yield point, and in cyclic loading when the stress reaches the endur¬
ance limit. If a member is subjected to combined loading, part of which
may be static and part cyclic, the question arises as to how to determine
the allowable stress.
Let us assume the case of a straight bar subjected to a tensile load which
varies from Pmax to Pmin. This loading can be expressed as a steady com¬
ponent Pm (the subscript m standing for mean) on which is superimposed
a cyclic component Pr, as represented in Fig. 5-1 (d). If A is the cross-
sectional area of the bar, the stresses corresponding to the above loads
will be a static stress Pm/A and a cyclic stress Pr/A.
Experimental data on bars with this type of loading are fairly extensive,
and when the failure points are plotted as in Fig. 5-3, they follow generally
the curve indicated by the dotted line. The curve shows that if the static
component is zero, failure occurs at point A, which is the endurance limit
of the material, while if the cyclic component is zero, the bar fractures at
B, the ultimate strength. In the design of machine parts, we wish to

regard failure in static loading as occurring at the yield point in order to
avoid permanent deformation; hence if we regard point C, the yield point,
as indicating failure under static conditions, it will be consistent with
usual design procedure.
Fig. 5-3. Failure under combined static and cyclic loads.
It has been suggested1 that, for design purposes, a straight line be

drawn from A to C so that no failure points will lie below this line and the
end conditions A and C are consistent with usual design practice. Any
point on this line, such as F, will represent the limiting conditions for
design purposes.
From similar triangles2
Sy — Pm/A Pr/A
(5-4)
This equation can be written in the form

Pm -f~ PrSy/s
Sy (5-5)
A
If the factor of safety is introduced and the strength-reduction factor

K't is applied to the variable load component, Eq. (5-5) becomes
Pm ~b K t'Pr$y/$e
(5-6)
f.S. ~AT
1 C. R. Soderberg, Factor of Safety and Working Stress, J. Appl. Mechanics, vol.
52, no. 11, p. APM-52-53, 1930.
2 This approach in writing the design equations follows closely, by permission of the
author and publisher, R. T. Hinkle, A Simple Method of Presenting the Combined
Variable-load Equations, J. Eng. Educ., vol. 41, no. 7, March, 1951.
60 MACHINE DESIGN
An inspection of this equation shows that, if the variable component

Pr equals zero, the equation reduces to A = Pm -s- s„/f.s., and if the static
component equals zero, it reduces to A = (Kt'Pr) -s- s«/f.s. The allow¬
able stresses in these equations then are consistent with Eqs. (5-1) and
(5-2).
If the appropriate symbols are used in Fig. 5-3, the following equations
for bending and torsion are obtained:
I = Mm + KfMrSy/Se
(5-7)
C Sy/f.S.
J _ Tm + K/TrSv/Se
(5-8)
r Sys/f.s.
In Eq. (5-8), sys is the yield point in shear, and su/se for tension can be
used instead of the corresponding values for shear since the ratios are
nearly the same value. This equation applies of course only to shafts of
circular cross section.
The equation for combined bending and torsion both of which may be
variable may be similarly obtained, using Eq. (13-5) to give the following
equation, which has become known as the “ Westinghouse equation” for
solid shafts of circular cross section:
7Td3 _ y/(Mm + Kt'MrSy/SeY + (Tm + Ks’Trsy/se)2

(5-9)
16 Sys/f.S.
where d = diameter of shaft

Mm = mean component of bending moment
Mr = cyclic component of bending moment
Tm = mean component of torsional moment
Tr = cyclic component of torsional moment
Ke = combined factor for bending
Ks> — combined factor for torsion
sy = yield point in tension
sys = yield point in shear
se — endurance limit in reversed bending
Equation (5-9) may be compared with the equation for the ASME
Code for the Design of Transmission Shafting given in Chap. 13. The
latter equation may be used for the design of general purpose transmission
shafting; but if cost, service, and safety requirements are critical, as in
turboalternator shafts which are parts of very expensive equipment and
where outages are serious, then the use of Eq. (5-9) is fully justified.
Deflection and vibration characteristics of such shafts should be investi¬
gated separately.
CHAPTER 6
MEMBERS THAT FAIL BY BUCKLING
6-1 Columns and struts. Short compression members subjected to

centrally applied loads may be designed on the basis of direct compression.
If, however, the member has a length greater than four to six times the
least dimension perpendicular to its axis, it is classed as a column or strut
and failure may be caused by buckling.
Theoretical equations for the design of columns were first developed by
Euler. The Euler formula for the buckling load, or critical load, of a long
column was derived on the assumption that the column bows sideways
while the stresses are within the elastic limit. This type of failure is the
result of elastic instability.
If the column is of less slender proportions, the maximum stress may
reach the yield point before sideways bowing occurs; hence, the Euler
formula does not predict the critical load. This type of failure is the
result of plastic instability. A rational equation for the critical load has
not been developed for this type of failure; hence, formulas based on
experimental results must be used for columns that fail by this method.
A formula that gives a reasonable check on test results for columns of this
type is the J. B. Johnson formula.
A third type of failure for columns made of thin sections, such as tubes,
or for built-up columns is local buckling.
6-2 Design of steel columns. The Euler and the J. B. Johnson formu¬
las are as follows:
The Euler formula:
mr2EI
F = -
cr L
The J. B. Johnson formula:
Fer = Asy( 1 - —T")

V 4n7T2Ep2/
where Fcr = critical load causing failure, lb

A = cross-sectional area, in.2
/ = moment of inertia of area, in.4
61
62 MACHINE DESIGN
L = length of column, in.

p «= least radius of gyration of cross section, in.
n = end-fixity coefficient (see Fig. 6-1)
E = modulus of elasticity of material, psi
sy = yield point of material, psi
mv2E
In the design of a column or strut, the length is usually known but,
since the dimensions of the cross section are not known initially, one can¬
not determine which equation, Euler’s or J. B. Johnson’s should be used.
CASE:
1 Round ends or pin ends, free but guided 1
H Fixed ends 4
ILI Flat ends / to 4
JY One end fixed, other free but guided 2
V One end fixed, other free 0.25
Fig. 6-1. End-fixity coefficients for columns.
As an aid in expressing the range in which the equations apply, and in

checking the results, the dimensionless quantity B/p2 is useful in which
SyL2
B =
mr2E
The Euler and J. B. Johnson equations may be rewritten by introducing

B and p as
The Euler formula (for B/p2 > 2):
(6-1)
The J. B. Johnson formula (for B/p2 < 2):

MEMBERS THAT FAIL BY BUCKLING 63
If these two equations forFcr are equated, it will be found that B/p2 = 2.
For that condition, the column would be at the point of simultaneous fail¬
ure by elastic and plastic instability, that is, for values of B/p2 less than
2, failure would be predicted by the J. B. Johnson equation, while for
B/p2 greater than 2, the Euler formula applies. These ranges are shown
in Fig. 6-2.
Most struts used in machinery are of proportions in the J. B. Johnson
range rather than the Euler range, so that, unless it is evident that the
Fig. 6-2. Ranges for Euler’s and J. B. Johnson’s equations.
column will be long and of small cross section, the J. B. Johnson equation
should be tried as a first trial and the assumption checked by the criterion
of B/p2 after the cross-sectional dimensions have been determined.
In arriving at the design value for the end-fixity coefficient n, it is
necessary to consider the degree to which the actual support approaches
one of the ideal cases in Fig. 6-1. For instance, the fixed-end case is
seldom realized in an actual installation because of the deflection of the
support so that the design value of n may be well limited to 3 or possibly
3.5 instead of the ideal value of 4.
In determining the radius of gyration, it is necessary to consider the
plane in which failure occurs. The plane of failure will be that for which
the combination of bending resistance and end fixity is the least, i.e., the
plane for which the product nl is the least.
64 MACHINE DESIGN
Example. A pin-ended strut 6 in. long is to be made of SAE 1030 steel, of circular
cross section, and is to support a compressive load of 2,000 lb. The diameter of the
loading pins at the ends of the strut is H in. See Fig. 3-2 for the appearance of the
strut. Assuming a factor of safety for the strut of 1.5 and an allowable bearing pres¬
sure at the pins of 10,000 psi, determine the dimensions for the strut.
Solution: For the strut,
SyL2 42,000 X 62
B = = 0.0051 in.
mv2E IXtr2 X 30,000,000
Fcr = f.s. X Fall = 1.5 X 2,000 = 3,000 lb
From the J. B. Johnson equation,

B
Fcr = Asy l 1 —
4p2
If substitutions are made for A = 7rd2/4 and p = d/4 and the equation is solved for
d2, the result is
4FCr n 4 X 3,000
d2 =-b 4B =--- + 4 X 0.0051
trsy tv X 42,000
= 0.091 + 0.020 = 0.111 in.2
d = 0.333 Use % in. in diameter for the strut
As a check on the choice of the J. B. Johnson equation,
B 4B 4 X 0.0051
= 0.145
d2 (0.375)2
Since the value of B/p2 is less than 2, the assumption of the J. B. Johnson equation
was correct.
For the dimensions of the eye, the projected bearing area of the pins may be calcu¬
lated as
2,000
0.2 in.2
10,000
The minimum width of the eye therefore would be
0.2
0.4 in.
0.5
By using in. for the width of the eye, clearance would be provided for machining the
faces of the eye.
The dimensions of the strut would be as follows:
Diameter of body of strut.^ in.

For the eye:
Inside diameter.in.
Outside diameter. 1 in.
Width.A in.
6-3 Design of cast-iron columns. Experimental results indicate that

for brittle material, a straight-line equation is valid for predicting the
failure load. One of these equations giving good results for the safe or
allowable load for a cast-iron column is
Fan = A (9,000 - 40 (6-3)
Refer to Art. 6-2 for notation.

6-4 Factor of safety for column design. The maximum stress that
exists in a column at failure is indeterminate. For this reason it is
customary to specify the factor of safety in column design in terms of
loads rather than stresses:
Fcr = f.s. X Fau
where Fcr = critical or failure load, lb

Fau — allowable or safe load, lb
The value for the factor of safety to be used in design depends on the
installation and the service. Where weight must be kept to an absolute
Fig. 6-3. Eccentrically loaded strut.
minimum and the conditions of loading and quality of material are accu¬
rately known, a value as low as 1.25 may be used. In the usual applica¬
tion, a value of from 2 to 3 or 4 may be suitable.
6-5 Beam columns. Members required to withstand axial loads and
transverse loads simultaneously are called beam columns. Examples of
beam columns are an eccentrically loaded strut, as shown in Fig. 6-3, and
an end-supported beam carrying a weight, as shown in Fig. 6-4.
For the eccentrically loaded strut shown in Fig. 6-3, it is apparent that
the moment at the ends, Pe, will cause transverse deflection of the strut.
The axial compressive load P will then have a line of action eccentric to
the strut by an amount 5 at the center. This force and distance in turn
induce an additional bending moment on the beam. The magnitude of
the moment therefore depends on the deflection of the structure, which
in turn is related to the loading. This interrelation makes the determina¬
tion of the moment difficult; however, if the moment is known, then the
maximum stress may be determined from
P , Me
(6-4)
S = A + —
66 MACHINE DESIGN
where P = axial load, lb

A = area of cross section, in.2
- = section modulus, in.3
c k
M = maximum bending moment due to combined axial and trans¬
verse loads acting simultaneously, in.-lb
It is usually difficult to determine the bending moment M analytically.
Tabulations of some of these bending moments are available for various
combinations of bending.1
Fig. 6-4. Beam with transverse and axial loading.
One case of particular interest is the eccentrically loaded strut in Fig.

6-3. From the consideration of the equilibrium of a strut with axial loads
applied eccentrically, the deflection at the center can be determined, and
the moment will be2
L
M = Pe sec — (6-5)
2P
where P axial compressive load, lb

e eccentricity of axial load from centroid of section, in.
L length of strut, in.
P radius of gyration of cross section, in.
A area of cross section, in.2
E modulus of elasticity of material, psi
1 R. J. Roark, “Formulas for Stress and Strain,” 3d ed., McGraw-Hill Book Com¬
pany, Inc., New York, 1954.
2 Timoshenko and MacCullough, “Elements of Strength of Materials,” 3d ed.,
p. 309, D. Van Nostrand Company, Inc., New York, 1949.
The stress at the center of the beam from Eqs. (6-4) and (6-5) is
P Pec L
s = ~ + — sec —
A I 2p EA
ec L
— sec (6-6)
P* 2P
For the beam with a central transverse load and an axial compressive
load as shown in Fig. 6-4, the bending moment at the center is1
WL tan U
M =
U
TV
where TJ — — \ —
F = critical load from Euler’s equation, lb

L = length of beam, in.
W = transverse load, lb
The stress in this case becomes from Eq. (6-4)
S = P + WcL tan U
(6-7)
S A + 4/ U
6-6 Design of beam columns—approximate method. With the excep¬

tion of long, slender beam columns in which the deflections are relatively
large because of the combined axial and transverse loads, it is sometimes
feasible to approximate the bending moment M due to these loads for
substitution in Eq. (6-4) in order to determine the maximum stress.
A formula2 giving sufficient accuracy is where Mi equals the maximum
bending moment due to the transverse loads alone, and (3 is a coefficient
depending on the method of loading and the kind of supports:
Mi
M = (6-8)
1 - (/3PL2/EI)
Table 6-1
Loading P
Cantilever, end load. H
Cantilever, uniform load. K
Simply supported, center load. H2
Simply supported, uniform load. 5As
Fixed ends, center load. K4
Fixed ends, uniform load. 2
1 S. Timoshenko, “Theory of Elastic Stability,” p. 5, McGraw-Hill Book Company,

2 Maurer and Withey, “Strength of Materials,” 2d ed.; John Wiley & Sons, Inc.,
New York, 1940,
68 MACHINE DESIGN
For the case shown in Fig. 6-4, the equation for the maximum stress
becomes
P c WL (6-9)
Z + / L4 - (PL2/SEI)
When the axial and bending loads are known for a beam column, the
maximum stress can be determined. It is apparent, however, that the
induced stress is not directly proportional to the applied loads. Hence,
in using the factor of safety, it is necessary to determine the critical values
of the axial load and bending load and to apply these critical loads to
Eq. (6-9) as follows:
Pcr = f.s. X P
Wcr = f.s. X W
and
(f.s.)P c (f.s.)IFL
>cr = sy = (6-10)
~A~ + 7 [4 - [(f.s.)PL2]/3P/
The determination of the dimensions of the required section is a matter
of trial and error.
mil i
i
rmr T
(a)- Lateral (b)- Plate, (cPPIate
buckling in compression in shear
(cL)- Buckling of compression flange of I-beam

Fig. 6-5. Failures by elastic stability.
6-7 Flat plates. Flat plates subjected to bending, compression load¬

ing, or shear loading in the plane of the plate may fail by elastic stability.
Examples of this type of failure are shown schematically in Fig. 6-5. The
same general principles that apply to the buckling of struts or columns
also apply to thin plates. The loading on the member, the type of sup¬
port, and the class of material (ductile or brittle) of which the member is
made determine the mode of failure and the value of the critical loading.1
1 S. Timoshenko, “Theory of Elastic Stability,” McGraw-Hill Book Company, Inc.,
New York, 1936.
In many cases such plates will continue to support an increasing load

after buckling has occurred; however, for most designs failure should be
based on the load to produce buckling.
The failure of flat plates subjected to loading normal to the surface of
the plate is due not to elastic stability but to tension failure.1
6-8 Tubes with external pressure. Tubes subjected to external pres¬
sure may fail by localized buckling. As the external pressure approaches
a definite critical value, it is found that the deflections of the walls of the
tube increase indefinitely. At the critical value of the pressure, the deflec¬
tions become very large and the tube tends to collapse.
It was shown in the case of a column subjected to a central compressive
load that the critical load depends on several factors, i.e., modulus of
elasticity, stress at failure (above or below the proportional limit), and
end restraint. These same factors must also be considered in the case of
tubes subjected to external pressure. If the walls of the tube are very
thin compared with the diameter, then failure by buckling can occur
before the proportional limit has been reached. Bryan2 shows that the
pressure to produce buckling is given by
2 Etz
(6-11)
(1 - v2)d*
where pc = critical pressure, psi

E = modulus of elasticity, psi
t = thickness of tube, in.
d = mean diameter of tube, in.
v — Poisson’s ratio = 0.29 for steel = 0.34 for brass
This equation for tubes is analogous to the Euler equation for slender
struts.
In thick tubes, buckling occurs when the proportional limit is reached.
As in the case of columns, it is feasible to determine empirically a formula
that represents failure in this range. An equation proposed by Southwell3
is
Vc (6-12)
Sy( 1 — v2)d2
1 +
Et2
1 See R. J. Roark, “Formulas for Stress and Strain,” 3d ed., McGraw-Hill Book
Company, Inc., New York, 1954, for formulas applying to plates subjected to normal
loading.
2 G. H. Bryan, Application of Energy Test to the Collapse of Long Thin Pipe Under
External Pressure, Proc. Cambridge Phil. Soc., vol. 6, p. 287, 1888.
3 R. V. Southwell, On the Collapse of Tubes by External Pressure, Phil. Mag.,
vol. 29, p. 67, 1915.
70 MACHINE DESIGN
Because of lack of uniformity in thickness and perfection of roundness

in commercial tubing, it is found that the values in Eq. (6-12) are higher
than test values. If Eq. (6-12) is used for the entire range of failure
(elastic and plastic), conservative values will result.
Shortening the tube will provide restraint to collapse, and hence
increases the critical pressure for the tube. For details of these cases and
the code pertaining to their design, the reader is referred to the literature
of the subject.1
In the design of tubing subjected to external pressure, the factor of
safety is used as the ratio of the critical pressure to the operating pressure.
•
1 For instance, H. F. Saunders, and D. F. Windenburg, Strength of Thin Cylindrical

Shells under External Pressure, Trans. ASME, 1931; Proposed Rules for the Construc¬
tion of Unfired Pressure Vessels Subjected to External Pressure, Mech. Eng., April,
1934, p. 245.
CHAPTER 7
ENGINEERING MATERIALS
7-1 Introduction. In designing a machine member, the selection of

material and the manufacturing process by which the part is to be made
should be considered together. For this reason, the mechanical designer
should be familiar with such interdependent factors as the adaptability
of materials to the various processes, the effects of the processes on the
properties of the materials, and the design details involved in the processes.
Since the selection of material and the process for each machine member
represents a special case involving a large number of variables, compre¬
hensive rules cannot be stated. However, some important general con¬
siderations are discussed in this chapter.
The most economical design is arrived at by considering the total cost
of labor, materials, and overhead for each of a number of proposed designs
that are satisfactory from a technical standpoint.
7-2 Factors affecting the selection of materials. Some of the more
important economic factors and physical and mechanical properties that
are involved in material and process selection are discussed briefly in the
following paragraphs.
Availability and cost of materials vary continually, and, as the change
is toward favorable or unfavorable conditions, designs will necessarily
undergo corresponding alterations for economic reasons. At times certain
materials may become unavailable for general industrial use, and the
necessity may arise for substitute designs based on procurable materials.
A forceful example of this expediency occurs during national emergencies
when alternate designs are required to avoid the use of strategic materials,
as well as processes which are affected by the unavailability of equipment.
Strength and rigidity are important properties of materials that are used
for machine members. Strength, as measured by the ultimate strength,
is necessary to prevent failure of the member by rupture. However, some
steels have the desirable property of high ultimate strength coupled with
low ductility, which may be undesirable in members subject to stress con¬
centration. The use of high ultimate strength steels should imply care¬
fully controlled heat-treatment to avoid harmful effects, such as those
caused by concentration of mass, surface decarburization, and quenching
71
72 MACHINE DESIGN
cracks. To guard against permanent deformation of the member, the

elastic limit should be considered in design. For ductile materials, the
yield point may be used ordinarily instead of the elastic limit.
Rigidity is of importance in members whose deflections are limited.
Rigidity depends upon the modulus of elasticity. It should be noted that
all steels have practically the same value for the modulus of elasticity. It
follows that a change from a soft low-strength steel to a hard high-strength
steel will not materially alter the rigidity of the part. It may also be
noted that a member made of cast iron will generally be more rigid than
a member of equal load-carrying ability made of steel, since the larger size
required for the cast-iron member will more than compensate for its lower
modulus of elasticity.
Resistance to fatigue should be the basis for the design of members that
are subjected to cyclic loading. This property is measured by the endur¬
ance limit. If concentration of stress is present in the member, notch
sensitivity and damping capacity of the material should also be considered.
Carefully controlled heat-treatment should be applied to members that
are subjected to fatigue in order to avoid harmful surface effects. In
some cases, the strength of a member may be increased by grinding off a
surface layer after heat-treating.
Damping capacity is defined as the energy dissipated as heat by a unit
volume of the material during a completely reversed cycle of stress. The
specific damping capacity is the percentage of the total energy absorbed
during a cycle. The damping capacity is related to internal friction in
the material and is represented by the mechanical hysteresis diagram.
The diagram in Fig. 7-11 shows the specific damping capacities of cast
iron, of a carbon steel, and of an alloy steel. It may be noted that the
damping capacity depends on the maximum stress. It has been sug¬
gested2 that the significant critical damping capacity for use in design is
the value at the endurance limit. The damping capacity is markedly
affected by magnetic fields.3 Low damping capacity is of value when
resonance is desired, as in the case of tuning forks, bells, or gongs, but high
damping capacity is desirable in most machine members to prevent an
accumulation of serious resonant stresses, to decrease vibration, to
decrease chatter as in machine tools, or to decrease noise.
Resilience should be considered when the material is subjected to shock
loading. A material with a suitable yield point should be selected, and
1 G. S. von Heydenkampf, Damping Capacity of Materials, Proc. ASTM, vol. 31,

II, p. 157, 1931.
2 Battelle Memorial Institute, "Prevention of Fatigue of Metals,” Appendix 23,
3 E. R. Parker, The Influence of Magnetic Fields on Damping Capacity, Trans.
Am. Soc. Metals, vol. 28, p. 661, 1940.
ENGINEERING MATERIALS 73
the member should be designed so as to secure a desirable resilience of the

part.
Hardness and ductility are important in many members. In bearing
surfaces which have relative motion and in which fluid lubrication does
not exist, hardness is of importance to limit wear. Hardness and duc¬
tility generally have a reciprocal relation in ferrous materials. The latter
property is frequently desirable in order to relieve concentration of stress
and it is effective in this respect in static loading but not in cyclic loading.
Weight may be important; often it is desirable, as in the case of founda¬
tions and flywheels; but it is undesirable in other cases, such as aircraft
Fig. 7-1. Comparison of damping characteristics for cast iron and two steels.
parts, where light metals, such as aluminum or magnesium alloys, may be

used.
Electrical properties and thermal properties may be important. The
latter property may be involved in expansion or contraction conditions
that call for a consideration of the coefficients of expansion.
Resistance to wear or corrosion may be determining properties, the former
in unlubricated or poorly lubricated friction surfaces. Generally, like
materials in contact are not satisfactory for friction surfaces, although
there are exceptions. Corrosion resistance is important in members sub¬
jected to a corrosive environment. The effect of corrosion is especially
serious in the presence of stress concentration in cyclic loading.
Casting and forging characteristics are important for materials that are
used for castings or forgings. Some materials, especially alloy steels, are
difficult to cast because of effects of high temperature. Fluidity is of
importance for good-quality thin sections in castings.
74 MACHINE DESIGN
Machinability is frequently a critical factor, for instance, in stock for

parts made by automatic machine tools. Often an expensive material
which is readily machinable is a more economical selection than a lower
priced material which may *be difficult to machine.
Low friction is of importance in bearing materials. The friction condi¬
tions are affected by the materials in contact and the surface finish. Cer¬
tain combinations of materials that are in contact in bearing surfaces pro¬
duce satisfactory results. This subject is considered in the discussion on
bearings.
Friction is desirable in many instances: in friction surfaces of brakes,
clutches, and power-transmission belts.
7-3 Materials. The more common materials of which machine parts
are made are discussed briefly in the following paragraphs. Detailed
properties of these materials may be found in publications pertaining to
engineering materials and to mechanical design.1 Appendix V contains
data for some of these materials.
Cast iron is the cheapest of the cast metals, and in the plain or alloyed
state it can be cast into almost any size and form that is desired. The
three classes of cast iron are as follows: (1) Gray iron, so called from the
gray fracture as cast. Gray iron is easily machinable, but is not as abra¬
sive-wear resistant as chilled iron; it will resist galling wear as in the case
of brake shoes, cylinders, and pistons. (2) White iron exhibits a white
fracture. It is hard and difficult to machine, and it is usually produced
to be made into malleable cast iron. (3) Chilled cast iron is white iron
composing parts of castings that are cooled rapidly to provide hard wear-
resistant surfaces. The depth of chill can be controlled by composition
and heat-treatment. Chilled cast iron is used in such members as dies,
car wheels, plow points, and cam followers.
The damping capacity of cast iron is high, as shown by Fig. 7-1. This
property makes it suitable for beds, bases, and frames of machinery that
are subject to vibration and for crankshafts and camshafts. The effects
of mass and cooling rates on the structure and properties of the casting
should be considered. In high-temperature applications, change in the
physical dimensions known as “creep” and “growth” may be important
to consider.
Cast iron is used for cylinder blocks, heads, crank cases, flywheels,
crankshafts and camshafts, beds, bases, frames and housings, car wheels,
brake drums and shoes, dies, wheels, gears, sprockets, pulleys, rolls, and
many other machine parts of irregular form.2
1 “ASME Handbook, Metals Engineering—Design,” Oscar J. Horger (editor),

McGraw-Hill Book Company, Inc., New York, 1953.
2 “Cast Metals Handbook,” and “Alloy Cast Iron,” American Foundrymen’s
Society, Chicago.
Malleable iron is made by transforming white-iron castings, which are

originally hard and brittle, by a heat-treatment which converts the mate¬
rial into a matrix of ferrite containing nodules of temper carbon, so that
it becomes tough and ductile, resistant to impact, and easily machinable.1
Wrought iron is a mechanical mixture of highly refined metallic iron and
a small quantity of slag. It is resistant to corrosion in ordinary atmos¬
pheres, and welds readily. The endurance limit is low as compared with
other ferrous materials.
Cast steel, either plain or alloyed, is used where castings of improved
properties over iron or malleable castings are desired. Cast steel is
stronger and tougher, weighs less for the same strength or stiffness, and
has a higher endurance limit than cast iron. Corrosion resistance can be
improved by the use of alloys or protective coatings. Some of the uses of
steel castings are as bases for heavy-duty machinery, highly stressed cast¬
ings, frames for transportation and construction equipment, gears, wheels,
and many small and intricate highly stressed machine parts.2
Wrought steel in its various forms is commonly used for machine parts.
Its applications range from members subjected to mild service, where low-
carbon steel having relatively low strength is satisfactory, to members
subjected to severe conditions of loading and environment where high-
carbon steel or alloy steel is used. It is important to consider that the
necessity for proper heat-treatment increases with the carbon and alloy
content of steel. The mechanical properties may be markedly controlled
by the composition and heat-treatment, and the selection problem is
usually one of choosing from the large number of available steels the one
that is most satisfactory for the member being designed.
The common alloying elements used in steel are manganese, nickel,
chromium, molybdenum, vanadium, and wolfram (tungsten). Nickel
increases the strength, yield point, endurance limit, shock resistance, and
hardness without a corresponding loss in ductility.
Wrought steel may be forged, swaged, rolled, drawn, and welded, and
it is especially susceptible to machining and finishing. It finds innumer¬
able applications in members subjected to high stresses due to external
loads either static or cyclic, to shock loading, or to stress concentration,
and in parts subjected to wear, high or low temperature, or corrosion;
it is also used in applications where special properties, such as electrical
or thermal requirements, are involved and where space and weight must
be conserved.
1 See “American Malleable Iron Handbook,” Malleable Founders’ Society, Cleve¬

land.
2 See “Steel Castings Handbook,” Steel Founders’ Society of America, and “Cast
Metals Handbook,” American Foundrymen’s Society, for information regarding
cast steel.
76 MACHINE DESIGN
Some applications of steel in machinery parts are in crankshafts, con¬

necting rods, piston rods, fastenings, keys, pins, rivets and bolts, ball and
roller bearings, springs, shafting, gears, valves, frames of heavy stationary
and transportation equipment, tubes, levers, arms, dies, and rolls.1
Brass and bronze are used in machinery parts in both cast and wrought
form. Brass is used where moderate strength and ductility, resistance to
corrosion, or good wearing qualities are desired. Bronze is superior to
brass in these properties but is more expensive. Phosphor bronze has
relatively high tensile strength, yield point, endurance limit, and shock-
resistant properties.
Aluminum and aluminum alloys. Pure aluminum resists oxidation, is
highly ductile, and has good forming properties, but it has poor casting
characteristics and machining properties. The effects of copper as an
alloying element are to raise the ultimate strength and endurance limit
and to improve the casting characteristics and machinability, but to lower
its resistance to corrosion. Aluminum-copper alloys are used in such
applications as crankcases, transmission housings, and fittings. The alu¬
minum-silicon alloys have better corrosion resistance and mechanical
properties but poorer machinability than aluminum-copper alloys. They
are suitable for marine castings, water jackets, housings, and castings
where a minimum of machining is required.
The aluminum-copper-magnesium-manganese alloy known as “duralu¬
min” is suitable for parts that require severe working, and it has good
corrosion resistance and strength. It is obtainable in sheets, plates, tubes,
rods, wire, extruded sections, bolts, and rivets. The plates are well
adapted to spot welding and are widety used in aircraft construction.
Nonmetallic materials. Wood has some uses in machine members, it is
used where lightweight parts subjected to moderate shock loading are
required, as in circuit-breaker operating rods, or where a nonmetallic bear¬
ing material is desirable, such as lignum vitae in water-lubricated bearings,
and it is used in impregnated friction surfaces, as in wood-block brakes.
Rubber and similar synthetic materials such as Neoprene have a variety
of applications in machinery. Rubber should be protected from high
temperature, oil, and sunlight. It is an excellent material for seals and
diaphragms, for water-lubricated bearings, for parts subjected to vibra¬
tory forces (such as vibration mountings, flexible couplings, and flexible
bearings), and for tubes and hose. Rubber may be bonded to steel by
the use of an intermediate nickel plating.
Plastics have become available for many machine parts. In general,
the plastics are light in weight, hard, and insoluble, are nonconductors of
1 Two sources of considerable information on steel useful to designers are the “SAE
Handbook,” Society of Automotive Engineers, and “Nickel Alloy Steels,” The Inter¬
national Nickel Co., Inc.
electricity and have a high damping capacity. They may be molded

into shapes for small bases, housings, knobs and handles, or bonded with
a filler, such as paper or fabric, to be made into sheets, bars, and tubes,
most of which are readily machinable.
Laminated bakelite, Micarta, and celoron are used extensively for
lightly loaded gears where quietness is desired. The plastic gears are
usually mated with cast-iron gears. This type of plastic is also used as
bushings for bearings. Molded nylon is finding increased use for small
gears, bearings, friction inserts, and many applications where low wear
and friction are desirable.
7-4 Mechanical properties of materials. SAE numbering system.
The numbers for materials are given in Table 7-1.
Table 7-1
Material Number
Carbon steels. 1XXX
Plain carbon. 10XX
Free machining. 11XX
Manganese (intermediate). 13XX
Nickel.;.... 2XXX
3.5% Ni. 23XX
5.0% Ni. 25XX
Nickel-chromium. 3XXX
1.25% Ni, 0.75% Cr. 31XX
3.5% Ni, 1.50% Cr. 33XX
Corrosion and scale-resistant. 30XXX
Molybdenum. 4XXX
C-Mo. 40XX
Cr-Mo. 41XX
Ni-Cr-Mo. 43XX
Ni-Mo (1.75% Ni). 46XX
Ni-Mo (3.50% Ni). 48XX
Chromium. 5XXX
Low Cr (0.5% Cr). 50XX
Med Cr (1.0% Cr). 51XX
Chromium-vanadium. 6XXX
Cr 1%. 61XX
Triple-alloy steels
0.55% Ni, 0.50% Cr, 0.20% Mo. 86XX
0.55% Ni, 0.50% Cr, 0.25% Mo. 87XX
3.25% Ni, 1.20% Cr, 0.12% Mo. 93XX
1.00% Ni, 0.80% Cr, 0.25% Mo. 98XX
Silicon manganese
2% Si. 92XX
Boron
0.0005% B (min). XXBXX
Properties of steel. The selection of a definite steel for a particular

application involves the consideration of the service requirements for the
78 MACHINE DESIGN
part, availability of the material, and the combined cost of the material
and processing of the part. Normally the designer may select for many
parts plain carbon steel on account of its low cost, ease of fabrication, and
any necessary heat-treatment. If the service requirements are severe,
it may be necessary to select alloy steel to secure higher strength, greater
hardenability, greater ductility, resistance to shock loading, and/or better
Fig. 7-2. Mechanical properties of SAE 1045 steel, water-quenched. {Courtesy of

International Nickel Co., Inc.)
machinability. The choice of alloy steel, however, will have the dis¬
advantages of higher cost of material, more exacting heat-treatment, and
the possibility of temper brittleness in some grades.
The relative costs of steel are roughly as follows, taking hot-rolled
plain-carbon steel bars as a basis: hot-rolled alloy steel, 60 per cent
greater; cold-finished plain-carbon steel, 30 per cent greater; and cold-
finished alloy steel, 85 per cent greater.
Steel with less than 0.30 to 0.35 per cent carbon will not have its proper¬
ties altered materially by heat treating. With higher carbon content, the
effects of heat treating are illustrated by the sample chart (Fig. 7-2)
which shows that with low tempering temperature, the tensile strength,
yield point, and hardness are high but ductility and resistance to impact
are low.1 On the basis of service requirements the designer can select the
optimum tempering temperature.
Oil quenching is not as drastic as water quenching and gives lower
strength and hardness but with less distortion of the part (Fig. 7-3).
The size of the part affects markedly its tensile properties, as shown by
the sample chart Fig. 7-4. For the yield point of several carbon and
alloy steels, see Appendix VIII.
Tempering temperature, deg F

Fig. 7-3. Mechanical properties of SAE 1045 steel, oil-quenched. (Courtesy of Inter¬
national Nickel Co., Inc.)
If hardenability is a factor for wear resistance, plain carbon steel with

casehardening may be suitable, or if surface requirements are severe for
resistance to shock and high pressure, a steel of the hardenable class called
H steels may be necessary. Hardenability curves for selecting these steels
are available which give results of standard end-quench tests and are
expressed as shown in the sample graph (Fig. 7-5). These curves are
presently available for about a hundred steels from 1320H to 9445H.2
1 In comparing impact values between the Charpy keyhole notch and the Izod V
notch, it is well to remember that the former usually lies between 15 and 45, while for
the Izod test values are frequently over 100.
2 For a set of compositions and hardenability bands for the H steels, refer to the
“SAE Handbook,” “AISI Steel Products Manual,” or “Physical Metallurgy” by
Clark and Varney.
80 MACHINE DESIGN
350
300 8
<u
250
200 2
150 £
too 2
50
0
0 12 3 4 5 6
Diameter of section; in.
Fig. 7-4. Effect of size on properties of SAE 1045 steel. (Courtesy of International
Nickel Co., Inc.)
Distance from quenched end of

specimen in sixteenth inches
Fig. 7-5. Hardenability band for SAE 2330H steel. (Courtesy of Society of Auto¬
motive Engineers.)
In specifying hardenability, the maximum and minimum hardness

values at a definite distance from the end of the test bar may be stated.
The distance chosen would be that corresponding to the section used in
making the part. For example, in Fig. 7-5, the specifications J4y53 = Y6 in-
means that at Yq in. from the end of the control specimen the hardness
by a Jominy test for SAE 2330H steel would lie between the limits of 42
and 53 Rockwell C.
In determining strength properties for steel to use in design calculations,

the yield point can be determined from Appendix VIII. The tensile
strength can be estimated from the diagram in Appendix VI, and the
endurance limit from the curves in Appendix VII for the required surface
finish. See Example 7-1 for the procedure.
For steel, the ultimate strength in shear is approximately two-thirds
the ultimate strength in tension. There is no general relation to deter¬
mine the yield point in shear but it may be approximated from the values
given in Table 7-2. The shear values in this table were determined from
torsion tests.
Table 7-2. Yield Ratios of Some Steels*
SAE _ . yield point in torsion

Condition Hat io *11 • . • «
No. yield point in tension
1010 Normalized 0.70

2350 Annealed 0.78
3140 Quenched and tempered 0.68
* From “The Torsional and Shear Properties of Quenched and Tempered Nickel Alloy Steels in
Different Sizes,” The International Nickel Company, Inc., New York, 1949.
In the absence of more specific data, the endurance limit in shear may
be estimated for design purposes by multiplying the endurance limit in
flexure by 0.55 for carbon steels and by 0.58 for alloy steels. Appendix V
gives some properties of steel and nonferrous materials.
Example 7-1. Determine the yield points and endurance limits for SAE 1045
steel, oil-quenched and tempered at 900 F, assuming that the parts are in the size
range to 1^ in. with machined surfaces.
From Appendix VIII,
Sy = 86,000 psi
From Appendix VI,
st = 115,000 psi
From Table 7-2,
sy in shear = 0.62 X 86,000 = 53,000 psi
From Appendix VII,

se in tension = 41,000 psi
se in shear = 0.55 X 41,000 = 22,500 psi
Machinabiiity. From the standpoint of production cost, a paramount

property of the material is its machinability. This property at present
82 MACHINE DESIGN
cannot be directly measured in units ;x but it can be satisfactorily expressed

as a combined result of three factors: (a) tool life, (b) quality of the
machined surface, and (c) required cutting power. It turns out that the
most significant of these factors is tool life, since long tool life usually goes
hand in hand with good surface quality and low cutting power. A useful
measure of tool life is the volume of material which is removed from the
parts being machined before the tool becomes dull and must be reground.
It becomes necessary then to establish a criterion for a dull tool.
In a cutting tool, wear occurs chiefly along the tool flank at the side
cutting edge. This wears away the side relief adjacent to the cutting
edge and produces what is called a wear land. When the wear land
becomes large, the power required increases, the surface finish becomes
poor, and tool wear is accelerated. Thus, an arbitrary value can be
assigned to the width of the wear land (perpendicular to the side cutting
edge) and the volume of metal removed from the parts being machined
can be used as an expression of “machinability.”
Cutting tools made of high-speed steels have been run satisfactorily
until the wear land was as much as 0.060 in. and carbide tools to 0.030 in.
To obtain comparative values of machinability for different metals, a
standard wear land may be set for a series of tests and the volume of
metal removed may be used as a basis for comparing the machinability of
various metals. In the “U.S. Air Force Machinability Report,”2 a wear
land of 0.015 in. was selected for most of the tests. In production,
carbide tools can be run with over twice this amount of wear land.
To pass on to the part being machined, it has been demonstrated that
machinability of steel is governed more by its microstructure than by any
other characteristic. The chart in Fig. 7-6 was prepared from U.S. Air
Force test data using 70B carbide cutting tools, a widely used type, with
a limiting wear land of 0.0150 in. These data had little scatter, but simi¬
lar results secured by using tools made of high-speed steel show more
spread than with carbide tools.
In design of heavily loaded parts for production in large quantities, the
problem is to select a material and heat-treatment to satisfy the loading
requirements and then to determine the probable microstructure of the
part to use in determining the machining requirements. Often the
1 The state of our present knowledge of machinability recalls the statement: “I
often say that when you can measure what you are speaking about, and can express
it in numbers, you know something about it; but when you cannot measure it, when
you cannot express it in numbers, your knowledge is of a meagre and unsatisfactory
kind; it may be the beginning of knowledge, but you have scarcely, in your thoughts,
advanced to the stage of Science, whatever the matter may be.” Lord Kelvin (1824-
1907).
2 “United States Air Force Machinability Report,” Curtiss-Wright Corp., Wood
Ridge, N.J., vol. 1, 1950; vol. 2, 1951.
machining requirements and strength requirements call for a compromise

in selection of material and its heat-treatment.
There are also available in tables relative machinability values for
various materials in terms of their chemical composition and Brinell hard-
P is Pear/He F is Ferrj/e
Fig. 7-6. Tool life versus cutting speed. Tool 78B carbide. Wear land 0.015 in.
Depth of cut 0.100 in. (From 11 United States Air Force Machinability Report ,” Curtiss-
Wright Corp., vol. 2, 1951.)
ness numbers.1 These values are usually stated in terms of 100 for SAE
1112 (free machining), with lower values for poorer machining materials.
However, machinability cannot be predicted accurately by chemical
composition and Brinell hardness numbers alone. When production cost
is critical, efforts should be made to select cutting tools and speeds on the
1 Kent’s “Mechanical Engineers’ Handbook, Design and Production,” 12th ed.,

pp. 19-30, John Wiley & Sons., Inc., New York, 1950.
84 MACHINE DESIGN
basis of microstructure. The Air Force report referred to includes a

large number of charts of tool life and cutting time versus cutting speed
for a variety of carbon and alloy steels represented by their composition,
heat-treatment, and microstructures.
High-temjperature applications. Because of the increase in efficiency at
high operating temperatures of heat engines as steam and gas turbines
and turbosuperchargers and of processing equipment as cracking stills,
it has become necessary to develop materials which will have the required
properties at these temperatures. The loss of strength in a carbon steel
Temperature, deg F
Fig. 7-7. Short-time tensile properties of forging steel. Effect of temperature on the
properties of metals. (ASME-ASTM Symposium, 1931.)
at high temperatures is shown in Fig. 7-7, which indicates the inadequacy

of this steel for high-temperature service. While considerable progress
has been made in recent years in developing heat-resistant alloys and
refractory metals, the frontiers of applications in the high-temperature,
high-efficiency field is limited by available materials to retain strength
and dimensions at the high temperatures. At present the maximum
temperature in steam- and gas-turbine casings is limited to about 1000 F,
although experimental work is in the 1400 to 1600 F and higher areas.
It is significant that at the operating temperatures of high-temperature
steam turbines and of gas turbines, the blades are of red-white heat and
their strength, if made of ordinary steel, would be less than of cast iron at
room temperature.
In addition to the loss of strength at high temperatures, steel and other

metals exhibit the phenomenon of creep, which is the gradual elongation
of the entire member at high temperatures over a long period of time.
High temperatures and high stresses increase the creep rate so that at
high temperatures we must limit the stress so that the part will not
elongate beyond a certain amount over a desired life of the machine or
equipment. For instance, a turbine blade may be designed with an
Hours
Fig. 7-8. Tensile creep curves for high and low stress.
allowable stress which will limit the total creep of the blade to not over
0.01 in. in 25 years. In Fig. 7-8 is shown the form of creep curves. In
the use of creep data, it should be remembered that extrapolation of the
curves may be unreliable. This poses a problem since if we are designing
a unit for a life of 10 years, or 87,600 hr, we cannot wait that long to con¬
duct long-time creep tests; instead we may be forced to make judicious
use of short-time tests.
In Fig. 7-9 is shown a series of creep-design curves for a carbon steel.
For extensive data on the properties of metal at elevated temperatures,
see the references below.1
Donald S. Clark and Wilbur R. Varney, “Physical Metallurgy for Engineers,”
D. Van Nostrand Company, Inc., New York, 1953; Carl H. Samans, “Engineering
Metals and Their Alloys,” The Macmillan Company, New York, 1949; G. V. Smith,
“Properties of Metals at Elevated Temperatures,” McGraw-Hill Book Company,
Inc., New York, 1950; “The Nickel Alloy Steels Handbook,” International Nickel
Co., Inc., New York; Samuel L. Hoyt, “Metal Data,” Reinhold Publishing Corpora¬
tion, New York, 1952.
86 MACHINE DESIGN
A classic example of creep is that of a sheet-lead roof originally in.

thick and weighing 16 lb per sq ft which was placed as a cover on the roof
of the Cathedral of St. Maurice at Angers, France, completed in 1247.
The thickness of the lead at1 the ridge is now 3^32 in-> while at the eaves it
has gathered into folds 10 in. high.1
At low temperatures, down to at least —125 F the strength, elastic prop¬
erties, and endurance limit of steel are increased and the ductility changes
little, but the notched-bar impact values usually decrease markedly,
which indicates brittleness. This change in property is important in the
design of low-temperature processing and refrigerating equipment and
Temperature, deg F
Fig. 7-9. Creep data for annealed cast carbon steel. (Courtesy of The Crane Company.)
also in low ambient temperature applications as railroad and ordnance

equipment and in airplanes. Maximum resistance to this embrittlement
is secured with low carbon content but this of course reduces the strength.
The addition of nickel as an alloying element toughens steel markedly at
low temperatures in both cast and wrought steels (see Fig. 7-10).
7-5 Selection of steel. Low-carbon steel. Steel with less than 0.20
per cent carbon has high ductility, is rolled or stamped, and is suitable for
sheet or strip products. Structural steel and boiler plate is usually of
0.15 to 0.25 per cent carbon. SAE 1015 to 1020 steel is known as car¬
burizing grade; it may be formed cold and when carburized has good wear
characteristics. SAE 1020 steel finds wide use in cold-worked parts as
bolts and rivets and cold-drawn parts as bars and shafting. It provides
a good finish for this class of parts. SAE 1022 to 1024 steel is used for
heavier sections such as forged parts and may be readily welded and
brazed.
1 From Product Eng., June, 1941, p. 293.

Medium-carbon steel is of 0.30 to 0.83 per cent carbon. SAE 1030 to

1035 steel has lower ductility than the low-carbon grade but it responds
to heat-treatment. It may be forged cold for small forgings for connect¬
ing rods, shafts, and bar stock. SAE 1040 steel finds extensive use for
tubing, automotive bolts, connecting rods, and crankshafts. SAE 1050
steel is harder and is usually oil-quenched. It is used for gears and heavy
steel forgings.
Temperature;deg F
Fig. 7-10. Effect of temperature on impact resistance of carbon and nickel steel.
Steels with 0.55 to 0.83 per cent carbon have high strength and good
wear properties. They are generally forged hot for spring wire and for
parts having cutting edges. Steel with 0.65 to 0.75 per cent carbon is
used for railroad rails. SAE 1065 to 1085 steel is used for coil springs and
flat springs and SAE 1070 to 1083 is used for plow shares, toothed diggers
and scrapers, and other parts subjected to shock and wear.
In the high-carbon steel group, SAE 1090 to 1095 is widely used for
springs, punches and dies, chisels, saws, cutting tools, steel wire, and
88 MACHINE DESIGN
cables. Steel with 1.0 to 1.2 per cent carbon is used for saws, files, and
razor blades, while steel with 1.2 to 1.5 per cent carbon is used for balls,
rollers, and races for rolling bearings.
Alloy steels. Nickel steel is the most widely used of the alloy group
on account of its high strength, high ductility, and good impact prop¬
erties. SAE 2335 steel is used for highly stressed gears, axles, castings,
nuts and bolts, while SAE 2335 to 2350 is used for many other highly
stressed machine parts. SAE 2515 steel is used for more severe service
where high fatigue strength or good carburizing qualities are necessary;
this steel, however, is costly. It finds use in wrist pins, kingpins, trans¬
mission gears, ring gears, pinions, and cams.
The chromium steels, SAE 50XX and 51XX, have improved hardening
properties and higher strength and wear resistance, but may be temper
brittle. These steels are used in applications as ball bearings.
Nickel-chromium steel has further increased strength and harden-
ability but may be temper brittle. SAE 31XX is used for drive shafts
and axles, transmission gears, and connecting rods.
Carbon-molybdenum steel, SAE 40XX, is a carburizing grade used for
spline shafts, transmission gears, and pinions where the service is not
too severe, while SAE 41XX, chrome-molybdenum, has good deep¬
hardening characteristics, high ductility, and weldability and is rela¬
tively inexpensive.
7-6. Relation of cooling rates to design.1 The serviceability of a
machine part may be greatly affected by the interrelation between heat-
treatment and several factors that are involved in the design of the part,
such as its form and mass.
The nonuniformity of physical properties of steel throughout a member
may be due to unequal cooling rates in various sections and to the result¬
ing effect on the decomposition of austenite. For example, in quench¬
ing a solid sphere heated to a uniform temperature throughout, the
surface will be cooled more rapidly than the interior, and the difference
in cooling rates will vary with the type of coolant. Also, a small sphere
will have a greater difference in cooling rates than a larger sphere, since
the ratio of surface to volume is inversely proportional to the diameter.
It is evident that points in an irregularly shaped member may have
widely different cooling rates, which may result in a corresponding differ¬
ence in the effect of heat-treatment. For instance, the point of a tapered
pin will be cooled more rapidly during quenching than the larger end.
Likewise, a corner of a rectangular bar will be cooled faster than an edge,
and the edge faster than a flat surface, and the surface faster than the
interior. Thus, if uniform properties are required, the designer should
1 See discussion on design and heat-treatment in “ Metals Handbook,” American
Society for Metals.
attempt to shape the part so that the cooling rates in the various sections
will be as nearly uniform as possible. An approach to this condition is
found in the distribution of the material as evenly as possible, i.e., in the
avoidance of large masses of metal. The procedure is illustrated in
Fig. 7-11 where material has been removed from heavy sections.
(a) (b)
Fig. 7-11. (a) Original design of block to be heat-treated. (6) Improved design,
showing reduction of concentration of mass, and substitution of fillets for sharp
reentrant corners, showing methods 1 and 2.
The Effect of Mass charts1 given in handbooks indicate the effect of

size or mass and the quenching operations on the physical properties of
steels. For instance, one of these charts indicates that the tensile
strength of SAE 1045 steel quenched in water at 1500 F and drawn at
1000 F has a value for a 4-in.-diameter specimen which is approximately
15 per cent less than that for a 1-in.-diameter specimen. Other prop¬
erties show corresponding differences.
An example of the mass effect is shown by the comparison between the
strengths of a solid shaft and a hollow shaft. An 8-in. shaft with a 4-in.
axial hole will have its torsional section modulus reduced by 6)4 Per cent
and its weight reduced by 25 per cent, as compared with an 8-in. solid
shaft. Tests on shafts show that the
bored shaft is more than 40 per cent
stronger than the solid shaft with the J
(a) (b)
same heat-treatment.
Internal stresses may be serious in
O
that they may lead to cracks forming
during heat-treatment or under service (c) (d)
Fig. 7-12. Progressive improve¬
loads. A bar or plate of section as
ment in section of bar to be heat-
shown in Fig. 7-12(a) may have cracks treated.
develop at the reentrant corners because
of unequal cooling of the wide and narrow portions. An improvement
results by introducing fillets, as shown in (5), or gradual transition
slopes, as shown in (c). Further improvement may result from the
removal of some material, as shown in (d), for reasons discussed in
the preceding paragraphs. In some cases, however, the use of a large
fillet, as required by considerations of stress concentration, may have its
1 See "SAE Handbook,” Society of Automotive Engineers.

90 MACHINE DESIGN
advantages reduced by reason of the mass effect, so that a medium-sized

fillet may be a compromise.
Distortion and warping may be caused by the unsymmetrical shape
of the part or by nonuniform application of the coolant during quenching.
Thus the slotted ring in Fig. 7-13 may be distorted during cooling, as
shown by the dotted lines in (a). Such distortion may be decreased by
the use of two or more symmetrical slots. Distortion of gear teeth in
heat-treatment may lead to noise and excessive wear. Special heat-
treatment, such as induction hardening or flame hardening, may be used
to give a minimum of distortion in this case.
The dotted lines in Fig. 7-13(6) indicate warping that may be caused
by the sudden application of the coolant from one side of the ring.
CHAPTER 8
MECHANICAL FABRICATION AND PROCESSES
8-1 Introduction. In this treatment mechanical fabrication is in¬

tended to mean the permanent joining of structural or machine parts
that are made as separate units for manufacturing, assembling, or trans¬
portation reasons. Thus large plates are fabricated by riveting or weld¬
ing together small plates of sizes that are easily manufactured. Sim¬
ilarly, a drum may be fabricated from a cylinder and two circular plates
by welding or riveting. Fabrication is intended to represent a process
different from the assembly of units that are essentially different in func¬
tions, such as a gear and a shaft assembled by a force fit or a valve seat
in a cylinder block assembled by a shrink fit.
8-2 Riveting—general remarks. The function of rivets in a joint is
to make a connection that has strength and tightness. Strength is neces¬
sary to prevent failure of the joint. Tightness is necessary in order to
contribute to strength and to prevent leakage; for example, it is required
in a boiler or in a ship hull.
Rivets may be driven by hand or by riveting machines. In hand rivet¬
ing, the original rivethead is backed up by a hammer or heavy bar, and
then the die or set, as shown in Fig. 8-1, is placed against the end to be
headed and the blows are applied by a hammer. This forms the second
head called the point. In machine riveting, the die is part of the hammer,
which is operated by air, hydraulic, or steam pressure.
In structural and pressure-vessel riveting, the diameter of the rivet hole
is usually in. larger than the nominal diameter of the rivet. After
the rivet is driven, it will contract on cooling. The lateral contraction
will be slight, but there will be a longitudinal tension introduced in the
rivet that will draw the plates tightly together in well-driven rivets. This
tension will cause frictional resistance between the plates that will resist
sliding of the plates under load. The resistance to sliding due to inter¬
plate friction is seldom considered in calculations of the strength of
riveted joints. The procedure of neglecting interplate friction is fol¬
lowed because the pressures caused by the driving and cooling of the
rivets and the coefficient of friction between the plates are difficult to
determine, and neglecting those factors is on the side of safety. Also
91
92 MACHINE DESIGN
in high-temperature installations, relaxation may cause the tension in

the rivets to be reduced.
While rivets in some applications carry an external tensile load, their
general use should be limited to shearing loads. If it is necessary for a
tensile load to be carried by a connection, a bolt should be used.
Stress concentration exists in the parts riveted and in the rivets them¬
selves ; however, since the materials used in most riveted parts are ductile,
stress concentration may not be serious. In ductile material plastic flow
around the rivet hole may be relied upon to relieve stress
Point concentration if the loads on the joint are static. Stress
concentration exists also in the rivet, and it may be
serious at the junctures of the shank of the rivet and the
heads. This concentration of stress may be reduced by
countersinking the rivet holes q in. so that a fillet is
formed as shown in Fig. 8-2. Since loads on the rivets
may be variable because of the character of the external
Fig. 8-2. Fil load, the vibration, or the temperature change, the rivet
lets in rivets. may fail owing to stress concentration in fatigue, even
though the rivets are made of ductile material.
The ways in which a single riveted lap joint may fail are illustrated in
Fig. 8-3.
The maximum stresses in a riveted joint at failure are difficult to cal¬
culate. Instead of attempting to use exact stresses, a safe joint may be
designed on the basis of simple tension of the plates, shear of the rivets,
and bearing at the rivets and plates, using allowable stresses determined
MECHANICAL FABRICATION AND PROCESSES 93
from similar existing and satisfactory installations. The allowable

stresses quoted in the following article have been determined on this
basis and are recommended by the structural standards.
8-3 Structural and machine-member riveting. In the riveting of
structural and machine members, the rivet holes may be punched,
drilled, and reamed, or some com¬
binations of these operations may
be used. If punched holes are used
without being reamed, as in struc¬
tural work, it is assumed in calcu¬
lating the strength of the plate that
the punching operation may dam¬
age the plate to the extent of a
J/3 2-m- thick ring around the hole.
Thus, in determining the net area
of the plate between rivet holes in
a tension plate, the rivet hole
should be considered as having a
diameter that is in- larger than
the nominal diameter of the rivet.
Furthermore, in calculating the
strength of a structural riveted
joint, the nominal diameter of the
rivet should be used in determining
the strength of the rivet, since it is
assumed here that the rivet may
not completely fill the hole through¬ (d) (e) (f)
out the length of the rivet. These Fig. 8-3. Modes of failure of a riveted
two assumptions are illustrated in joint: (a) bending of the plate; (6)
rupturing of the plate by tension; (c)
Fig. 8-4.
shearing of the rivets; (d) crushing of the
In structural riveting, spacing of
rivets or the plates; (e) shearing of the
the rivets should be governed by margin; (/) tearing of the margin.
the following considerations: (a)
stability of the joint, (b) the existence of a margin between the rivet
hole and the edge of the plate that is not less than the rivet diameter,
and (c) clearance for the die during riveting.
For structural steel, the following allowable stresses may be used:
For plates in tension. 20,000 psi

For rivets in shear. 15,000 psi
For bearing of rivets and plates
Single shear. 32,000 psi
Double shear. 40,000 psi
94 MACHINE DESIGN
The determination of the strength of a simple structural joint is illus¬

trated by the analysis of the connection shown in Fig. 8-5.
The problem is to determine the safe load that may be carried by the
joint illustrated in the sketch. The angles and the plate are structural
steel, and the rivets are % in. in nominal diameter.
A=Nomina! diameter =d
11 TvXX/ B= Diameter of punched
*
hole -d-t-'/e in.
il
=Diameter to be used in
JnnnnxX calculating the strength
of plate =d+'/8 in.
\*-y1
-+
■*—19—^
— C7 — ►
Fig. 8-4. Structural riveting.
1. The strength of the plate along the section A A is equal to
[3y2- (H + }{)] xy2x 20,000 = 26,250 ib
©
2. The strength of all the rivets in shear is equal to
2
^ X 15,000 = 39,700 lb
3. The strength of all the rivets in bearing is equal to
3(M X Vi) X 40,000 = 45,000 lb
4. The strength in failure of the plate in tension along section BB

combined with failure of the lower rivet (a) in shear and (b) in bearing
is equal to
(а) {[3M - 2(% + H)] Xy2 X 20,000} + 2 X tt/4 (%)2 X 15,000

= 30,750 lb
(б) {[sy2 - 2(M + H)] xy2x 20,000} +%xy2x 40,000
= 32,500 lb
The above calculations indicate that the joint would fail by rupture
of the plate along section A A and that the safe load would be 26,250 lb.
Note that stress concentration has been neglected; a procedure that is
justified by the lack of seriousness of stress concentration because of the
ductility of the structural steel and the relatively high factor of safety
used in determining the allowable stresses.
P' ^2
A
C.6.
T ■J'. V
l 3 (6)
(cl)
Fig. 8-6. Eccentrically loaded connection and rivet loading.
Eccentrically loaded connection. A connection of this type is illustrated

in Fig. 8-6(a). The group of three rivets supports the loads. The first
step in the solution of the problem is to locate the center of gravity of
the rivet group. For convenience, the rivet group is redrawn in Fig.
8-6(5). If the rivets are of the same size and the upper two rivets are
equally spaced with regard to the vertical center line, the center of grav¬
ity of the rivet group will be at the point indicated by CG.1
The next step is to introduce at the center of gravity of the rivet group
two forces which are equal and opposite to P, as shown by Pr and P"
in Fig. 8-6(5), and then to regard P' as inducing a direct shearing load
known as “ primary shear/’ on the rivets, and to consider P and P" with
their moment arm e as producing a moment equal to Pe, which tends to
1 If the rivets are of different diameters or are unsymmetrically spaced, the center of
gravity of the rivet group may be located by the use of a well-known principle of
mechanics, X = 2 AX/S A, as illustrated in Fig. 8-7.
96 MACHINE DESIGN
rotate the rivets about the center of gravity of the group. The loads
induced by the turning moment are known as “secondary shear.”
Now the primary shear on each rivet is equal to P/3 and these loads
are shown in Fig. 8-6(c), where the rivet group has been redrawn.
The turning moment will induce loads on rivets 1, 2, and 3, which are
of different magnitudes and directions. The direction of the secondary
shear on each rivet is at right angles to the direction of its moment arm r.
Thus s[', or the secondary shear on rivet 1,
will act at right angles to ri, and s" and s"
will act at right angles to r2 and r3, respec¬
tively. To determine the magnitudes of
s[r, s", and s3, it is necessary to make an
assumption.
A reasonable assumption is that the two
members riveted together, i.e., the plate
and the channel in Fig. 8-6(a), are rigid
and that the deformation takes place by
the rivets. This assumption is valid for
Fig. 8-7. Center of gravity of
the proportions of plates and the number
rivet group.
and size of rivets that are met with in the
usual structural or machine members. However, it may not be valid for
very small, thin plates and for relatively large rivets.
If the assumption stated above is used, it follows that as a result of
the turning of the plate the deformation of each rivet will be proportional
to its distance r from the center of gravity of the group. Therefore
ri r2 r3
From the condition that the sum of the external turning moment and
of the internal resisting moment equals zero,
Pe = Sjri + s2 r2 + s3 r3
The equations given above are sufficient to determine the secondary

shear on each rivet and, in general, it may be found that
" = PeVn
Sn r\ + r\ + r|
where the subscript n refers to any rivet.

The secondary shear loads are drawn in Fig. 8-6 (c), which shows for
both primary and secondary shear the loads produced by the plate on the
rivets.
The primary and secondary shear loads may be added vectorially to
determine the resultant load R on each rivet.
In the solution of a problem, the primary and secondary shear loads

may be laid off approximately to scale and generally the rivet having the
maximum total shear load will be apparent by inspection. The values of
the resultant load for that rivet may then be calculated. The maximum
loaded rivet becomes the critical one for determining the strength of an
existing connection if all the rivets are the same size, or for selecting the
rivet size or sizes in a design.
Sleeve
A ^zsszzzzS^ SOIL
1 Spreader ring
&
fa^-Assembled
Fig. 8-8. Draw rivets for riveting from one side of plate.
Draw rivets have been used widely for riveting in locations that are
accessible from one side only. Figure 8-8 (a) shows a rivet of this type
known as the Huck rivet, which is composed of three essential parts:
(1) the hardened steel pin, (2) the soft steel sleeve, and (3) the spreader
ring.
The complete rivet is supplied to the user assembled as shown in
Fig. 8-8(a). At (6) is shown the rivet inserted from the right into the
holes of the plates. The sleeve is held against the plates while the pin
is pulled toward the right. This operation spreads the sleeve as shown
in the diagrams at (c) and (d) and finally the pin breaks off, leaving the
completed rivet as shown at (e). Another and simpler draw rivet is the
Cherry rivet.
98 MACHINE DESIGN
8-4 Riveting thin plates. The same general requirements of strength

and tightness that apply to general riveting apply also to thin plates;
however, because of the possibility of failure of thin plates carrying com-
\
!§! ^77777//////////////////////^777//
Fig. 8-9. Buckling of riveted thin plates.
pression or shear loads by buckling of the plates, special requirements are

necessary in thin-plate riveting.
A frequent cause of weakness of thin-plate construction is buckling of
the plate because of the inaccurate spacing of
rivet holes and/or improper riveting. This
wrinkling of the plate is illustrated in Fig. 8-9;
it may be serious in built-up structural mem¬
bers and in columns or beams or in plates
carrying shear loads, such as aircraft structures.
Flush riveting may be used in the wetted sur¬
faces of ship hulls or aircraft structures in
order to reduce skin friction. Special pro¬
cedures are necessary, especially in aircraft
construction, to prevent wrinkling of the plate
Fig. 8-10. Flush rivet de¬
between rivets and “dimpling” of the plate
veloped for aircraft con¬ around rivets. Figure 8-10 illustrates a flush
struction. rivet that has been developed for aircraft
construction.1
Explosive rivets are used in aircraft structures in quantity production
and are used also for locations accessible from one side only of the plate.
Figure 8-11(a) shows a section of an unexploded aluminum rivet with the
Fig. 8-11. Explosive rivets.
powder, or chemical, charge c in the shank. The rivet that has been
exploded by a heated-tip riveting gun placed in contact with the rivet-
head is shown at (b).
1 Berlin and Rossman, Flush Riveting Considerations for Quantity Production,
SAE Journal, August, 1939, p. 328.
8-5 Pressure-vessel riveting. The riveting of drums and tanks

intended to withstand high pressure requires special consideration in
design, materials, and construction. This is especially true for fired
pressure vessels, such as a steam generator drum. The design of pressure-
vessel joints should be based on
strength and tightness.
The strength of the joint should
be as near as possible to the strength
of the unpunched plate. It is possi¬
ble to produce joints of high effici¬
ency, i.e.j more than 90 per cent.
Tightness of the joint is necessary
to prevent leakage, and to ensure
Fig. 8-12. Calking of riveted joint.
tightness the spacing of the rivets
must be limited. Calking of the joint may be necessary to increase the
resistance to leakage. Figure 8-12 illustrates the calking operation.
Joints for vessels subjected to high pressure are usually made of the
double-strap construction shown in Fig. 8-13 (a) in order to minimize
bending of the plate, as illustrated in Fig. 8-3(a). The plates are gen¬
erally of unequal width in order to increase the joint efficiency and to
Fig. 8-13. Double-strap butt joints.
decrease stress concentration in the plate. This stress concentration is

caused by flexure of the plate due to the expansion and contraction of the
shell. A consideration of the stress-flow lines, as discussed in Chap. 4,
indicates that a joint with unequal-width cover plates, as in Fig. 8-13(a),
would have less stress concentration than one with equal-width cover
plates, such as the one in (b). Placing the narrow strip on the outside
of the drum facilitates calking of the joint adjacent to closely spaced
rivets. This latter consideration will be evident from Fig. 8-14.
100 MACHINE DESIGN
In order to design a joint for a pressure vessel, it is necessary first to

determine the type of joint to be used. The choice involves the cost of
construction and its relation to the efficiency of the joint, since the higher
efficiency joints have a larger number of rows of rivets. On this basis,
a type of joint may be selected that will be satisfactory from an economic
standpoint. After the type of joint has been decided upon, the design
of the joint involves the determining of the size of the rivets and the
pitch in order to secure approximately equal strengths of the joint for all
methods of failure. It is evident that with large rivets spaced close
Fig. 8-14. Double-riveted double-strap butt joint.
together the rivets will have a high strength and will favor a tight joint,
but the plate will be unduly weakened by the large, closely spaced holes.
In the treatment of pressure-vessel joints given here a typical standard
joint as specified by the ASME Code1 will be investigated for strengths
under various probable methods of failure.
The following assumptions are made:
1. That each rivet takes an equal share of the load. This assumption
implies that the shell and plate are rigid and that all the deformation of
the joint takes place in the rivets themselves. Because of the actual
deformation of the shell and plates, however, the rivets in the outer row
in a multirow joint are subjected to more than an equal share of the load.
2. The plates are undamaged by the hole-forming operation, and the
rivet fills the hole after it is driven. These assumptions are valid in
1 ASME Boiler Construction Code, Sec. I, Power Boilers.

high-grade construction using drilled and reamed holes and well-driven

rivets. Since the diameter of the hole is 3^l6 in- larger than the nominal
diameter of the rivet, the nominal diameter plus J^6 in- should be used in
all calculations.
3. Failure of the margin by shear or tearing will not take place. This
is true for ASME Boiler Code joints.
4. Finally, stress concentration is neglected. The stress concentration
in the shell, plates, and rivets may not be serious because of the ductility
of the material; however, in the presence of corrosive substances, such as
those found in certain kinds of boiler-feed water, corrosion fatigue may
be serious (see Art. 3-8).
The strengths of a double-riveted double-strap butt joint, shown in
Fig. 8-14, used as a longitudinal joint of a drum vTill be determined for
various methods of failure.
In the figure, p indicates the pitch of rivets in the row having the great¬
est pitch, and indicates the length of a repeating group whose efficiency is
typical of the efficiency of the entire length of joint.
Let p = pitch of rivets in the outer row, in.
d = diameter of rivet after driving, or diameter of rivet hole, in.
t = thickness of plate, in.
t' = thickness of cover plate, in.
st = tensile strength of plate = 55,000 psi
ss = shearing strength of rivet in single shear = 44,000 psi
sc = crushing strength of rivet or plate = 95,000 psi
Assume that
p = 4% in. = 4.875 in.
d = % in. = 0.875 in.
t = % in. = 0.375 in.
t' = î6 = 0.3125 in.
1. The strength of the solid plate is equal to
ptst = 4.875 X 0.375 X 55,000 = 100,550 lb
2. The strength of the plate between rivet holes in the outer row is
equal to
(p — d)tst = (4.875 — 0.875) X 0.375 X 55,000

= 82,500 lb
3. The strength of all rivets in shear is equal to
7rd2 ( 0.8752'
5— S. = 5 7T X 44,000
4
= 132,290 lb
102 MACHINE DESIGN
4. The strength of the plate between rivet holes in the second row plus
(a) shearing strength of rivets in the outer row or (b) crushing strength
of rivets in the outer row is equal to
7T d2
(p — 2 d)tst + = (4.875 - 2 X 0.875) X 0.375 X 55,000
T
0.8752
+ 7T X 44,000
= 90,910 lb
(6) (p - 2d)tst + dt'sc = (4.875 - 2 X 0.875) X 0.375 X 55,000
+ 0.875 X 0.3125 X 95,000
= 90,429 lb
5. Crushing strength of all rivets is equal to
2dtsc + dt'sc = d(2t + t')sc = 0.875(2 X 0.375 + 0.3125) X 95,000

= 88,320 lb
6. Crushing strength of rivets in second row plus shearing strength of

rivets in outer row is equal to
hrd2\ n ( 0.8752\
2dtsc + ( — ) ss = 2 X 0.875 X 0.375 X 95,000 + hr —) X 44,000
= 88,800 lb
The above analysis indicates that the joint will fail by method 2. The
efficiency of the joint is
82.500
X 100 = 82.0 per cent
100.500
The efficiencies of some commercial boiler joints are given in Table 8-1.
Table 8-1. Efficiencies of Commercial Boiler Joints

Efficiency,
Type of joint per cent
/
Lap joints:
Single riveted. 45-60
Double riveted. 63-70
Triple riveted. 72-80
Butt joints:
Single riveted. 55-60
Double riveted. 70-83
Triple riveted. 80-90
Quadruple riveted. 85-94
8-6 Welding—general remarks. There are three general fields of

application of welding:
1. Fabrication in which welding is used as an alternate method for cast¬

ing or forging. In this application, cost of construction and serviceability
are important in making the choice between welding and the alternate
methods. Cost of construction involves a consideration of the following
items:
1. Welding:
Cost of materials, as structural steel
Cost of cutting and forming parts
Cost of positioning parts, welding and machining
2. Casting:
Cost of materials, as iron and steel
Cost of patterns and molds
Cost of casting and machining
3. Forging:
Cost of materials
Cost of dies
Cost of forging and machining
Serviceability involves a consideration of strength and stiffness of the

parts, and, in parts subjected to variable loads and/or vibration, of the
damping capacity of the structure. In general, strength requirements are
favored by the welding of structural steel parts, deflection requirements
are favored by the larger sections of castings, and damping capacity is
favored by larger volumes of cast and forged parts, especially in the use
of cast iron.
In many instances where relatively large deflections are not objection¬
able and high damping capacity is not essential, the use of welded struc¬
tural steel results in a substantial saving in production cost; however,
where deflections are limited, or where impact loads and vibration require
high damping capacity, forged or cast construction is more satisfactory.
Approximate comparisons for a bearing support made in small numbers
by different methods are given in Table 8-2.
2. Fabrication in which welding is substituted for riveting. In this
application, welding is generally favored from the standpoints of strength
and saving in weight. In welding, however, the cost of holding the parts
during welding and the cost of welding itself, including materials, equip¬
ment, and operators, may favor riveting in some applications.
3. Welding used as a repair medium. Welding can be employed to
reunite metal at a crack, to build up a small part that has broken off,
such as a gear tooth, or to repair a worn surface, such as a bearing sur¬
face. Metal spraying also can be used in the latter example.
8-7 Methods of welding. The common methods of welding are forge
welding, electric-resistance welding, and fusion welding.
104 MACHINE DESIGN
Table 8-2. Comparison of Fabrication Methods for Bearing Support
kp-jzJ A i) Li-i_l u lJ
Welded, Forged steel, Cast steel, Cast iron,

per cent per cent per cent per cent
Static strength. 100 100 100 100

Deflection. 259 259 191 100
Damping capacity. 26.1 26.1 27.8 100
Weight. 71 71 76 100
Cost. 45 300 110 100
Thickness t. 54 54 60 100
In forge welding the parts must be heated to the plastic state at the
regions where they are to be joined, and then impact, which is produced
by a hand hammer or a press, causes the parts to unite. Wrought iron
and low-carbon steel may be forge-welded. The process has limited use;
it is employed in the manufacture of wrought-iron pipe and also as a
repair medium.
In electric-resistance welding, the parts to be joined are pressed together
and an electric current is passed from one part to the other until the
metal is heated to the fusion temperature at the joint. A butt joint may
be formed in this manner from relatively thick plates, or if the plates are
thin a lap joint may be formed. In the latter case, if the pressure is
applied by the two electrodes, one on each side of the overlapped plates, a
spot weld is produced, or if two rollers are substituted for the point
electrodes and the plates are pulled between the rollers a seam weld is
produced.
In fusion welding, the parts to be welded must be held in position while
molten metal is placed at the joints, thus fusing the metal of the parts,
called the parent metal. The molten metal then solidifies to form the
welded joint. The weld formed by the welding metal when it cools is
essentially cast metal.
The heat to melt the weld metal and to make the parent metal plastic
is generated in a number of ways, each of which defines the method of
fusion welding as (a) Thermit welding, (b) gas welding, and (c) electric-
arc welding.
In Thermit welding, a mold is built around the joints at the region where
it is desired to confine the molten metal. Then the Thermit, which is
composed of a mixture of finely divided iron oxide and aluminum, is
placed in a reservoir in the mold. The Thermit is ignited and the alu¬
minum reduces the iron oxide to molten steel, at a high temperature,
which flows in the mold, melts the parts, and forms the joint on solidify¬
ing. Thermit welding is used principally to repair heavy cast-iron and
steel parts and to weld heavy sections, such as rails, in the field where
other welding equipment is not available.
Gas welding, often called autogenous welding, uses an oxygen-hydrogen
or oxygen-acetelene gas, which is burned in a torch providing a pointed
flame. The flame heats the parts to be welded and melts the welding
metal, or filler rod, which on cooling forms the joint.
The oxygen-hydrogen process is generally used for welding nonferrous
metals of low melting points, while the oxygen-acetylene process is used
for welding ferrous and some nonferrous metals and alloys in thin sheets,
and as a repair medium. Gas-welding equipment may be modified for
use in flame cutting of plates.
Arc welding employs either a carbon-rod electrode with a separate rod
used for the source of weld metal or a metallic-rod electrode in which the
welding rod itself serves the dual purpose of electrode and source of weld¬
ing metal. The metallic-rod arc-welding process is adapted to both
structural and machine welding because of the consistently high quality
of the welding that it is possible to obtain by the use of the proper weld¬
ing procedure by trained operators of the welding equipment.
The making of high-quality welds requires specially designed equip¬
ment in order to produce a welding arc which has the necessary qualities
for the proper control of the welding in producing joints of desired
properties.
If a bare electrode or filler rod is used, the deposited weld metal while
it is hot will absorb oxygen and nitrogen from the atmosphere, which will
decrease the strength of the weld
metal and lower its ductility
and its resistance to corrosion.
Coating
Shielded-arc welding is employed Gaseous shield.
E/ectrode
to prevent oxygen and nitrogen Deposited metal
Arc
from coming in contact with the
r // i/iillHrd^
weld metal while it is molten. \ //1i/ii/ Base metat
NT'S',/'/"" ' > V
The shielded arc is produced by
using welding rods coated with
solid material which melts and Fig. 8-15. Shielded arc.
vaporizes when heated by the arc
and which forms a shield of inert gas around the molten weld metal; it
also furnishes a flux which floats out impurities to form a slag on the top
of the weld which further prevents harmful oxidizing effect on the weld
while it is cooling. This slag is brushed off after the joint has cooled,
leaving a good-appearing bead.
Shielded-arc welds are used in high-grade construction where strength
is paramount. Figure 8-15 shows details of the shielded arc.
106 MACHINE DESIGN
8-8 Forms of welded joints. Some common forms of welded joints are
shown in Fig. 8-16. When butt welds are used, it is not necessary to
bevel the edges of plates that are ^ in. or less in thickness. For heavier
plates, the edges should be beveled. For plates % in. or more in thick¬
ness, it is desirable to weld from both sides of the plate in order to min¬
imize distortion of the welded part and to reduce the amount of deposited
metal required.
(aj-Single-V Butt f6)-Double-V butt
/c)-Lap (d)- Offset
7^7/
w///m
(e)-Co rner
I
(fh Tee (g)-Edqe
Fig. 8-16. Common forms of welded joints.
8-9 Structural and machine welding. In welding of structural mem¬

bers or machine members, the primary purpose is to reduce the cost over
other fabrication methods and to equal or improve the serviceability of
the structure. Structural-member welding is employed generally as an
alternate to riveting. Machine-member welding may be used as a sub¬
stitute for rivets or bolts, for casting or forging, or for the purpose of
permitting the joining of small cast units to form a large part that would
be inconvenient to cast as a single unit.
In both structural and machine welding, the welds should be sym¬
metrical with respect to the axis of the welded member unless the loading
is unsymmetrical. The latter case is represented in Fig. 8-17, in which,
assuming that the stresses at the welds are uniformly distributed,
La X a = Lh X b
which represents equilibrium of moments of resistance offered by the

fillet welds.
The assumption that the stresses in the welds are uniformly distributed
is not true because of the elongation of the welded member under load.
The deviation from uniform distribution of stress is small for short lengths
of fillet welds, but becomes considerable for long weld lengths; hence the
design of long welds should take into account the concentration of stress
at the load-input end of the bead.
Fig. 8-17. Unsymmetrical welds.
In the design of a welded part, an attempt should be made to arrive

at the most economical combination of structural material, of cutting and
bending, and of deposited welded metal. It should be realized that
deposited weld metal costs per pound from 25 to 50 times as much as
structural steel. From this standpoint welded joints should be reduced
(a.)-Cast iron
Cost 100%
Weld 90 in. Cost 65% Weld 60 in. Cost 42%

Fig. 8-18. Designs for a support.
to a minimum in favor of the use of formed parts. Figure 8-18 illustrates

alternate designs for a support.
Overwelding also should be avoided. In Fig. 8-18(6) and (c), the top
of the support need not be held by a continuous weld. Likewise, in
welding three strips to form a beam similar in section to an I beam,
intermittent welding of the flanges and web is entirely satisfactory and
results in considerable reduction in cost.
108 MACHINE DESIGN
8-10 Welding design.1 There are three types of welding design, as fol¬
lows: (a) substitution method, (6) approximate calculation, and (c) precise
calculation.
Substitution method. Ih the change-over from a cast, forged, or riveted
machine member to a welded design by the use of the substitution
(a)~0riqina\ cast
link
(dJ-Par\s for welded
link
Fig. 8-19. Substitution method in welding design.
method, the welded part is fashioned from plates and structural and
formed parts to simulate the shape of the original member. This method
is illustrated in Fig. 8-19 in the change-over from a cast-iron to a welded
link. If a program calls for changing to welding of a complicated
machine, the change-over is generally made on one part at a time in order
that any trouble that develops because of the substitution of welding
may be localized and dealt with independently.
Approximate calculation. This method is generally applied to a new
design or to a moderately stressed existing machine member. The sizes
of parts and the strengths of the welded joints are determined in accord¬
ance with allowable values that have been found satisfactory in practice.
Allowable loads for normal and parallel fillet welds are given in Table 8-3.
Table 8-3. Allowable Static Loads on Mild-steel Fillet Welds
Allowable static load per linear inch of weld, lb
Size of weld, in. Bare welding rod Shielded arc
Normal weld Parallel weld Normal weld Parallel weld
X 1,000 800 1,250 1,000

Ke 1,500 1,200 1,875 1,500
X 2,000 1,600 2,500 2,000
2,500 2,000 3,125 2,500
% 3,000 2,400 3,750 3,000
X 4,000 3,200 5,000 4,000
5A 5,000 4,000 6,250 5,000
H 6,000 4,800 7,500 6,000
1 See “Procedure Handbook of Arc Welding Design and Practice,” 9th ed., Lincoln
Electric Co., 1950.
It may be noted in the table that the values for normal welds are higher
than those for parallel welds. This is due to the fact that there is a more
uniform distribution of stress for the normal weld than there is along the
length of the parallel weld. The values of shielded-arc welds are higher
than those for bare welding-rod welds for reasons discussed in Art. 8-7.
CccJ-Normal (b)- Para 11e I

Fig. 8-20. Fillet welds.
In determining the length of weld required, x/i in. should be added to

the length of each weld to allow for starting and stopping of the bead.
For welded joints that are subjected to cyclic loading, special precau¬
tions are necessary to avoid harmful effects of stress concentration. The
Fig. 8-21. Stress flow lines in welded joints.
precautions should be in accordance with the principles discussed in Chap.

4. The use of stress-flow lines in the design of a welded connection indi¬
cates that the joint in Fig. 8-21(6) is stronger in fatigue loading than that
shown at (a). Further increases1 in endurance limit may be realized by
the use of the joint similar to that in Fig. 8-21(c). The butt weld shown
at (d) and (e) illustrates improvement in endurance limit secured by
grinding to form a flush joint.
1 V. L. Maleev and J. B. Hartman, “Machine Design,” 3d ed., p. 200, Inter¬

national Textbook Company, Scranton, Pa., 1954.
110 MACHINE DESIGN
Example 8-1. A structural steel plate 3 in. wide and % in. thick loaded statically
in tension by a load of 10,000 lb is welded to a plate with normal welds, as in Fig.
8-20(a). Determine the size of weld required.
Solution: The net length of each weld is 3 — )4 = 2)4 in. The load on the two
welds would then be
10,000
2,000 lb per in.
2 X 2.5
From Table 8-3, 34-in. shielded-arc welds would be satisfactory.
Example 8-2. A 34-in. plate 2 in. wide carries a static tensile load of 8,000 lb and is
to be welded to a plate by parallel fillet welds, as in Fig. 8-20(6). Determine the
length of each weld, using )4-in. shielded-arc welds.
Solution: The net length of welds (Table 8-3) is
8,000
2 in. Use welds 2)4 in. long.
2 X 2,000
Table 8-4. Minimum Size of Fillet Welds for Different Thicknesses

of Plate*
Plate thick- Minimum weld
ness, m. size, m.
34 to 34
34 to
% to % 34
% to i
1)4 to l%
Over 1)4 %
* From C. H. Jennings, Welding Design, Trans. ASME, vol. 58, p. 497, 1936.
Precise calculations. This method is used in exceptional cases where

the geometry of the joint is unusual or where safety is of prime impor¬
tance, and where the savings in cost justifies the additional design
expenses. In this method an attempt is made to approximate the value
of the stresses in the weld and in the adjacent parts of the joint as well
as in the structure as a whole. See Table 8-5 for allowable stresses for
coated-electrode welds and Table 8-6 for stress-concentration factors for
Table 8-5. Allowable Stresses for Coated-electrode Welds on

Low-carbon Steels*
Static loads, Cyclic loads,

Type of Weld
psi psi
Butt welds:
Tension. 16,000 8,000
Compression. 18,000 8,000
Shear. 10,000 5,000
Fillet welds:
Transverse and parallel welds. 14,000 5,000
maximum stresses in these welds. Stress concentration should be con¬
sidered for cyclic loading but may be neglected for static loading for
reasons given in Art. 4-6.
Table 8-6. Stress-concentration Factors for Welds*

Stress-concentration
Type of weld factor
Reinforced butt weld. 1.2
Toe of transverse fillet weld. 1.5
End of parallel fillet weld. 2.7
T butt joint with sharp corners. 2.0
In a normal fillet-welded lap joint (Fig. 8-22), the size of the weld is
specified by dimension ec or cd. These dimensions are equal. The size
of the weld is not greater than the thickness of the plate and it may be
P P
-*e- ->■
P
■<—
P
less. Minimum sizes of welds are recommended in Table 8-4. In

arriving at the size of welds to be specified, it is well to realize that dou¬
bling the size of a weld doubles the strength, but requires four times the
metal to be deposited.
In the weld, maximum shearing stresses occur along the face cd but are
concentrated at c (the heel of the weld) on account of the plate being more
rigid than the weld. In normal fillet welds, the stresses may be assumed
as uniformly distributed along the length of the weld if the load is cen¬
trally applied. In Fig. 8-22 it may be assumed that each of the two welds
carries one half of the load P, so that the normal stress on the section ec
will be equal to P/2Lt. The component of that load which is normal to
the throat section cf is 0.707 times its value across ec. Since the area
of the throat section cf is 0.707 ec, it follows that the normal stress
across the throat is equal to the tensile stress at ec. The 0.707 value
above is the function of the 45-deg angle of the throat section.
112 MACHINE DESIGN
Single-weld lap joint. This type of joint should be avoided whenever

possible because of the opening of the joint on the unwelded side and
because the stresses under load are high and difficult to estimate. The
stresses are high because of the sharp bending of the plate. Figure 8-23
shows a rubber model of such a joint in the unloaded and the loaded
states and also, for comparison, a model of a double-weld lap joint.
(a) Single-weld lap joint
(b) Double-weld lap joint

Fig. 8-23. Rubber models of lap joints. (Courtesy of Lincoln Electric Company.)
8-11 Eccentrically loaded joints. A joint of this type is illustrated

in Fig. 8-24. In the design of such a connection, the center of gravity
of the weld pattern is located, and at that point, two forces equal and
opposite to the external load P are applied. One of these forces produces
shearing stresses on the welds which have a direction parallel to P and
which may be assumed uniformly distributed over the weld pattern.
These are called primary shear stresses.
The other load P and the externally applied load acting at the eccen¬
tricity e produce a twisting moment on the welds which induces additional
stresses on the welds called secondary shear stresses. The analysis of
loads on this joint is similar to that on eccentrically loaded rivet groups
discussed in Art. 8-3.
The polar moment of inertia of the weld pattern may be taken from
Table 8-7. Assuming that the plates are relatively stiff and that the
deformation under load is concentrated at the welds, the secondary shear¬
ing stress at any part of the welds will be proportional to its distance from
the center of twist. It follows that the maximum secondary shearing
(a) Welded bracket

Fig. 8-24. Bracket attached to support by fillet welds.
stress will occur at the corners of the weld pattern. The secondary shear¬
ing stress may be added vectorially to the primary stress to determine the
maximum shearing stress. It can usually be determined by inspection
where the resultant stress will be a maximum at any part of the joint.
The location is k in Fig. 8-24.
Example. Assume in Fig. 8-24 that the load P = 1,500 lb, e = 5 in., a = 3 in.,
and b — 2 in. The latter dimension is in. shorter than the weld to allow for start¬
ing and stopping of the bead. Determine the size of weld required for static loading.
Solution: The polar moment of inertia is
j = bt(3a2 + b2) = 2£(3 X 32 + 22) = 1Q ^

6 6
The radius to the most remote area is
r = y2 Va2 + 62 = % a/32 + 22 = 1.8 in.

The primary shear stress is
P 1,500 375
Sl 2bt 2X21 t
The secondary shear stress is
Tr _ (1,500 X 5) X 1.8 _ 1,310
s2
J 10.3< t
114 MACHINE DESIGN
Table 8-7. Polar Moment of Inertia for Weld Patterns
Weld Polar moment of

pattern inertia J*
t tb3
12
ta3
12
bt(3a2 + b2)
6
bt(a2 + 3 b2)
6
t(a + b)3
6
* Approximate.
The primary shear stress is always directed parallel to P. The secondary shear
stress is maximum at the four corners of the weld and is in a direction perpendicular to
its radius r drawn from the center of twist to the corner. The resultant stress is the
vector sum of the primary and secondary stresses and it is apparent by inspection that
the maximum resultant stress occurs at location k, as indicated in the stress pattern.
The resultant stress R may be determined analytically or graphically and is found to
be equal to 1,550/2.
This shearing stress would act along the leg of the weld (cd in Fig. 18-22). The
throat of the weld, however, is smaller than the leg and to be on the side of safety, it
is customary to use the throat area, even though the shear stresses as calculated do
not lie in that plane. This consideration would require using 0.7072 instead of 2 for the
section.
The maximum stress then would be equal to
1,550 2,190
0.7072 °r 2
Using an allowable stress from Table 8-5 equal to 14,000 psi, the size of the weld is
. _ 2,190 _ n 1Kft .
1 14,000 °'156 m’
A weld would be specified.
8-12 Stress relieving. As the molten metal solidifies and cools in a

welded joint, and as the parts adjacent to the joint cool, stresses are set
up in the parts owing to volumetric contraction. These local stresses
may in some cases become very high, and improperly welded joints may
crack on cooling owing to these residual stresses. If the welded parts
are rigidly attached to restraining members, additional stresses may be set
up because of general cooling of the structure after welding is completed.
If the material of the structure is ductile, the seriousness of the residual
stresses may be alleviated owing to plastic flow. Since the weld metal
itself is essentially cast, its ductility will
generally be lower than that of the parent
metal, and it cannot be depended on to
relieve residual stresses materially, unless YIG; 8"25' Lap weld Wlth Provlslon
... . . „ for lateral contraction,
special electrodes are used, therefore,
in order to prevent residual stresses from becoming serious, the parts
welded should be made of ductile material and provision should be made,
if possible, for flexibility of the parts during cooling of the welds. The
principle involved is shown in Fig. 8-25 in which two members joined by
fillet welds are not placed in contact, so that contraction of the parts will
not be restrained in the vertical direction in the sketch.1 Intermittent
welding may be desirable in long welds in order to prevent concentration
of the effects of heating and cooling of the parts welded and the weld
itself. Preheating of the parts may also be desirable in some instances.
1S. C. Hollister, Stress Relieving, “Welding Handbook/’ American Welding
Society, 1938.
116 MACHINE DESIGN
In a structure made of both welded and riveted parts, the welding should
precede the riveting so far as possible.
Residual stresses may be relieved after welding by suitable heat-treat¬
ment. The temperature/ rate, and time of heat-treatment depend on
the size and thickness of the parts and on the plasticizing rate of the mate¬
rial. Highly stressed parts whose failure would involve life hazard or
expensive structure damage, such as welded pressure vessels and welded
penstocks, should be stress relieved; this applies also to parts which
require close tolerances or those subjected to impact and fatigue
loading.
In welded steam-generator drums, stress relieving is effective in reduc¬
ing caustic embrittlement.
Hand peening of the joint while it is hot is effective in increasing the
strength of the joint, particularly in austenitic steel, but overpeening
should be avoided. Indications are that peening reduces residual stresses
in parts of the weld that may be in the plastic state and that it also intro¬
duces some cold working.
It is not generally necessary to stress-relieve welded machine members
if they have been properly designed with special precautions to avoid
stress concentrations, and if they have been welded properly. However,
in highly stressed parts and in parts subjected to repeated loading, stress
relief may be necessary.
8-13 Testing and inspection of welds. The mechanical testing of
welds in either static or fatigue loading has in general three purposes, as
follows:
1. Comparison of strengths of joints. This comparison may be for (a)
welds of different geometry, for example, in comparing the strengths of
a single-V butt weld with a double-V butt weld, or of a full bead with a
flush bead weld, or of a rectangular splice plate in a beam connection with
a diamond-shaped plate, as shown in Fig.
8-26. Or (b) the comparison may be for
the purpose of determining the effect on
strength of using, for example, welding rods
of different composition, rods of different
sizes in electric-arc welding, or bare versus
coated electrode welds.
2. Testing of filler metal. In this type of
testing of welds, specimens are prepared by
Fig. 8-26. Splice plates. machining from a welded joint a part of the
deposited metal and the adjacent parent
metal. The specimens are used to determine yield point, ductility,
nature of bond between deposited metal and parent metal, and chemical
composition.
3. Qualification tests of welders. Standard joints are made and they are
tested to determine the quality of the welding. These procedure tests are
made periodically and aid in maintaining consistent high-quality welding.
Inspection of welds, sometimes called nondestructive testing, is specified
in some instances. The most commonly used of these methods, each
having its own fields of application, are as follows:
Visual inspection. A method of inspection available to skilled opera¬
tors is known as “visual inspection,” which means that the operator
observes as the bead is “laid down” characteristics of its formation that
are a good indication of the quality of the weld. These characteristics
are rate of burning off of the electrode, fusion and penetration, forming
of the bead in the joint, and the sound of the arc. The evaluation of these
dynamic characteristics involves a certain amount of the “personal ele¬
ment” in the operator, but, taken together, they are one of the best
indications of the quality of a weld as they permit the operator, as it
might be said, to see inside the joint.
X-ray inspection. The x-ray equipment is expensive and is not port¬
able, and hence such inspection is available only where large numbers of
parts must be inspected and where the parts are not too large to be con¬
veniently transported. The limitation of the method is the thickness of
material suitable for penetration in a reasonable time. Two inches is
generally considered to be the upper limit of metal thickness for ordinary
work although up to 5-in. sections have been satisfactorily defined in
photographs.
Gamma-ray inspection. The gamma-ray equipment may be portable,
and hence is adaptable to field work. The cost of the equipment is lower
than the x-ray equipment; however, the radium required makes its use
expensive. The gamma rays penetrate the work more rapidly than
x-rays, hence this method is suitable for inspecting thick plates.
Magnifiux method. This method uses magnetic dust, applied either
dry or in an oil bath. The welded part, which of course must be mag¬
netic, is magnetized and the magnetic dust assumes a pattern in the vicin¬
ity of the joint which gives valuable indications of surface defects and
some subsurface defects, such as cracks and other discontinuities and
inclusions. If residual magnetism is objectionable, as in aircraft parts,
it should be removed by demagnetization.
Stethoscope method. In this method, the soundness of the weld may be
determined by means of the characteristics of the sound or ring produced
by tapping the weld with a light hammer. The sound is analyzed by the
operator using a stethoscope as an aid. The condition of the weld can be
quite accurately determined by skilled operators.
8-14 Rolling, spinning, and seaming. These are analogous operations
in which (usually) circular or cylindrical thin metal parts are rotated on a
118 MACHINE DESIGN
spindle while forming rolls shape the part. Figure 8-27 shows a welded
Y-belt sheave for which the edges of the flange were curled by the form¬
ing rolls.
8-15 Stitching. Thin metal sheets may be joined by stitching, in
which mild steel clips have their ends punched through the overlapped
sheets and then bent over to form the stitch.
8-16 Manufacturing processes.1 The manufacturing processes may
be considered in three groups: (a) forming processes, such as casting, forg¬
ing, rolling, and machining to dimension; (6) fabri¬
cating processes, such as soldering, brazing, and
welding; (c) finishing processes, such as finish ma¬
chining or grinding, honing, lapping, and superfinish¬
ing. A brief discussion of these processes follows.
Casting. Sand casting requires a separate mold for
each piece. Metal-mold casting may be used for
relatively low melting temperature metals; it em¬
ploys a metal mold that can be used for producing
many castings. Die casting is used in quantity
production of small parts made of low-melting-
temperature allo3Ts.
Hand forging, or machine forging using power
hammers, is used for parts whose shape requires
working of the materials in the semiplastic state, or
where it is desired to control the internal structure
as regards grain size and grain direction. Examples
Fig. 8-27. Fabri¬
of forged parts are upset and swaged pieces, crank¬
cated V-belt sheave.
shafts, connecting rods, valves, and gears. Forging
design should strive for as much material as possible in one plane and for
the avoidance of undercut surfaces and deep recesses.
Rolling in a rolling mill is used to form long members of uniform sec¬
tion, such as bars, rails, structural sections, and plates. The material
may be hot-rolled for large reduction in section or cold-rolled for accurate
sizing or grain control. Rolling may also be used to form gear teeth and
screw threads and to form sheet metal to desired shapes.
The cold-metal processes include shearing, blanking, bending, drawing,
which is used primarily for forming rods and shafting and for producing
cup-shaped parts from flat blanks, and squeezing, which includes coining
and extrusion. Metal spinning is a type of drawing used for forming sur¬
faces of revolution including beads and seams. The extrusion process is
1 H. C. Hesse, “Engineering Tools and Processes,” D. Van Xostrand Company,
Inc., New York, 1941; M. L. Begeman, “Manufacturing Processes,” John Wiley &
used to form the softer metals, such as aluminum and magnesium alloys,
into a variety of shapes.
Sintering is a process that is used to combine compressed powdered
metals or a mixture of powdered metals by heat-treatment so that a rigid
piece having desirable properties is formed. Nonmetallic substances may
or may not be included. The process may be used to produce accurately
sized parts which are composed of materials which will not mix or which
cannot be solidified from the molten state, and it may also be used for
parts which can be produced more economically by sintering than by
other processes. The strength and porosity of the finished parts can be
accurately controlled. Some examples of sintered products are bearing
sleeves which are used for bearings inaccessible for lubrication and which
are composed of a mixture of bronze and graphite, porous bronze bear¬
ing sleeves which are impregnated with oil, and bronze and iron parts
which are used as gears for oil pumps and pistons for hydraulic controls;
other examples are refractory products and cutting tips for machine tools.
Soldering, brazing, and welding. Soldering and brazing are used to join
metal parts by the use of a filler metal whose melting temperature is
lower than that of the metals to be joined. Brazing uses a harder filler
metal than is used in soldering and forms a stronger joint. Brazing in a
furnace is very effective and forms strong and neat joints. Welding is
used as a fabrication medium to join parts permanently and to form
built-up members as substitutes for castings and forgings, as a joining
medium to replace fastenings as rivets and bolts, and as a repair medium
to replace broken or worn sections of members. Welding is discussed
in Art. 8-7.
8-17 Design of castings. An important consideration in the design of
castings is volumetric contraction or shrinkage, which takes place as the
metal cools from the molten state to the solid state. In many cases the
contraction is restrained by the mold or cores or by adjacent parts of the
casting that have solidified earlier in the cooling process. The restrained
contraction gives rise to internal stresses that may lead to failure in
service.
In general, the necessity for considering the effects of contraction and
for designing the casting to minimize harmful effects increases with the
melting temperature and the volumetric contraction of the metal. Thus
castings of low-melting-point metals and alloys generally may be designed
without special regard for effects of contraction; iron castings require
medium precautions; and steel castings require great care and special
considerations.
Some general features that should be considered in the design of cast¬
ings are discussed in the following paragraphs. They apply specifically
120 MACHINE DESIGN
to steel castings. However, their incorporation should result in the

improvement of any casting.1
Large risers are desirable to supply liquid metal to compensate for loss
of volume during solidification. To ensure proper flow of molten metal,
the position of the mold during pouring should be considered.
Provision should be made for solidification in a progressive manner
from the lowest points of the casting to the risers. Progressive solidifica¬
tion requires the avoidance of relatively large isolated masses of metal.
Gradual changes in section should be used where possible.
Avoid sharp corners. External corners should be rounded off. Gen¬
erous fillets should be used to join sections, although it is not advisable
to use long fillets with radii greater than the thickness of the sections they
adjoin. An oversize fillet may intro¬
duce the effect of concentration of
mass.
Use properly spaced ribs or brackets.
A rib or bracket has great strengthen¬
ing and stiffening effect in service and
in addition, since it is thin, it cools
quickly and provides support for the
sections it adjoins during cooling,
thus lessening distortion of the casting.
Webs perform the same functions of
lessening distortion of the casting and
(c) (d) provide strength and stiffness in
Fig. 8-28. Reduction of concentration
service. At the junction of sections
of mass.
and a rib, concentration of metal may
be avoided by the use of a cored rib as shown in Fig. 8-28. A T junction
may be cored as shown at (6) to reduce concentration of masses of metal
and possibility of forming shrinkage cavities. An X junction may be
similarly cored as shown at (c) or offset as shown at (d).
When the design of a steel casting becomes complicated, consideration
should be given to the use of smaller cast parts assembled by welding or
bolting.
The precautions and suggestions discussed above are some of the more
important special features that should be incorporated in the design of a
casting, and are intended to be considered in addition to the usual casting
design details, such as providing draft, shrinkage and finish allowances,
pads for locating points, avoidance of sections too thin to cast or control,
caution in the use of inserts and chaplets, and avoidance of deep recesses
and undercuts. It is desirable that the designer have an insight into the
1 See Briggs, Gezelius, and Donaldson, Steel Casting Design for the Engineer and
Foundryman, Trans. Am. Foundrymen's Soc., vol. 46, 1938.
effects of size and shape of the casting on the properties of the metal and
that he consult with the foundryman and the metallurgist during the
design of castings in order that quality and economy be ensured.1
8-18 Surface finish. The surface finish is very important from the
standpoint of serviceability, appearance, and production cost of a machine
member. The type and quality of surface finish has great influence on
friction characteristics, wear, endurance limit, and effects of corrosion of
machine members.
Since the fine finishing operations remove surface cracks, a polished
member will have a higher endurance limit than one with a rough surface.
Highly stressed members subjected to cyclic loading, such as the air¬
craft connecting rod in Fig. 4-9, may therefore be polished.
Members subjected to a corrosive environment may have a coating or
plating to resist the corrosive action.
The appearance of the part, or of the complete machine, may be greatly
improved by the use of proper finishes and the judicious use of decorative
trims.2
8-19 Sampling and inspection. In securing samples for the purpose of
determining chemical composition or physical properties of the material
of which a machine member is made, it is important to have the sample
representative of the metal in the part. For check analyses, samples
should be secured in accordance with established standards. For cast¬
ings, test coupons either attached or separately cast should be provided.
The coupons are subjected to the same heat-treatment as the castings
and used for metallurgical and mechanical testing.
The finished product or samples may be subjected to the usual mechan¬
ical tests to determine hardness, tensile strength, yield point, ductility,
notched-bar or impact-test values, or for ratings, such as Magnaflux rat¬
ings, which indicate seams, cracks, inclusions, and cavities. X-ray or
gamma-ray inspection may also be necessary.
1 The “Cast Metals Handbook,” American Foundrymen’s Society, and the “Steel
Castings Handbook,” Steel Founders’ Society, are sources of considerable information
on the subject of casting design.
2 Harold Van Doren, “Industrial Design,” 2d ed., McGraw-Hill Book Company,
CHAPTER 9
DETACHABLE FASTENINGS
9-1 General remarks. Some machine parts must be so constructed

that they may be readily connected or disconnected without damage to
the machine or the fastening. This requirement may be for the purpose
of holding or adjustment in assembly or service, inspection, repair, or
replacement, or it may be for manufacturing or assembly reasons. The
parts may be rigidly connected, or provision may be made for predeter¬
mined relative motion.
Detachable fastenings perform the functions mentioned above. They
are of many forms and degrees of ease of assembly or disassembly. Some
types require special tools for this purpose, and some have special locking
devices, while others depend solely on friction to hold them in place.
They have in common, however, positive connection or in some cases set
limits of motion, and relative ease of connection and disconnection with
no damage to parts.
Some of the more common types of detachable fastenings are described
in the following paragraphs.
9-2 Screw fastenings. A screw fastening is composed of two elements,
as illustrated by a bolt and nut. Frequently one of the parts to be con¬
nected is so constructed that it becomes one of the elements of the fasten¬
ing. In fact, both of the parts to be connected may serve as the elements
of the fastening as in the case of a handle with a thread turned on one
end to engage in a tapped hole in a lever.
The choice of type of fastening is very important; the location or plac¬
ing also is equally important. Fastenings should be located and placed
so that they will be subjected to tensile and/or shear loads, and bending
of the fastening should be reduced to a minimum. Bending of the fasten¬
ing due to misalignment, tightening-up loads, or external loads has been
responsible for many failures. To relieve fastenings of bending stresses,
the use of clearance spaces, spherical-seat washers, or other devices may
be necessary.
Some common types of screw fastenings are described as follows:
Bolts. A bolt has a head at one end and a nut fitted to the other, as
shown in Fig. 9-1 (a) where it is used to attach a cylinder head to a
cylinder.
122
DETACHABLE FASTENINGS 123
Tap bolt. In Fig. 9-1(6) is shown a tap bolt engaging in a tapped hole
in the cylinder wall. If the cylinder head is removed frequently, the
thread in the hole may become worn, necessitating costly repairs.
Studs as shown in Fig. 9-1 (c) are threaded at both ends. One end is
screwed into the tapped hole in the cylinder wall and the other end
receives the nut.
(ct)-Bolf (b)-Tap bolt (c)-Siud

Fig. 9-1. Types of screw fastenings.
Of the three types of fastenings described above, the bolt makes the
best connection from the standpoint of construction cost, the tap bolt
and stud result in a saving in space and weight since a smaller flange is
required, and the stud has a replacement advantage over the tap bolt in
the event of wear of the thread.
(a) (b) (c) (d) (e)

Fig. 9-2. Cap screws with various types of heads; (a) hexagonal; (6) filister; (c) button;
(d) flat; (d) hexagonal socket; (/) fluted socket.
Cap screws are similar to small-size tap bolts except that a greater
variety of shapes of heads are available, as shown in Fig. 9-2. Shouldered
cap screws also are in common use.
Machine screws are cap screws slotted for a screw driver and are gen¬
erally used with a nut.
Setscrews are used to prevent relative motion between parts. They
may be used instead of a key to prevent relative motion between a hub
124 MACHINE DESIGN
and a shaft in light power-transmission members, or they may be used

in connection with a key, where they prevent relative axial motion of the
shaft, key, and hub assembly. Some types of setscrews having different
types of heads and points axe shown in Fig. 9-3. Various combinations
of heads and points are available.
Locking devices.1 In a tightened screw fastening there are two oppos¬
ing external forces present: (a) a loosening force due to the action of the
axial load which tends to unscrew the fastening, and (6) the force of
friction which tends to resist the unscrewing motion. Screw fastenings
are designed so that the friction force is expected to prevent unscrewing
or loosening of the fastening.
Ordinary threaded fastenings generally remain tight under the action
of static loads; however, many of these fastenings become loose under
Fig. 9-3. Setscrews.
the action of cyclic loads. Thus fastenings subjected to pulsating service

loads or to vibration require restraint from loosening in addition to that
offered by thread and collar friction.2 A large number of locking devices
are available. Some of these are described as follows: The jam nut, or
lock nut, is shown in Fig. 9-4(a). The castle nut to be used with a cotter
pin is shown at (6). Two types of split nuts are shown at (c); the first has
the split closed after threading as shown, so that screwing the nut on the
bolt will open the split and introduce additional thread friction; the sec¬
ond has the split opened by the small screw after the bolt is tightened.
The Everlock Locknut is shown at (d). The elastic stop nut shown at (e)
employs a hard fiber or nylon collar recessed in the nut that becomes
threaded as the nut is screwed on the bolt causing a tight grip. This
type of nut is especially effective in parts subjected to vibration. Some
types of lock washer are shown at (/).
1 Whittemore, Nusbaum, and Seaquist, The Relation of Torque to Tension for

Threadlocking Devices, Natl. Bur. Standards Research Paper 386, 1931.
2 J. N. Goodier and R. J. Sweeney, Loosening by Vibration of Threaded Fasten¬
ings, Mech. Eng., vol. 67, no. 12, p. 798, December, 1945.
In Fig. 9-4(a), the standard thickness nut is placed next to the work
with the thin nut (jam nut) on top. The reverse arrangement, i.e.,
placing the thicker nut on top is preferred by some designers since, as
can be shown,1 the upper nut carries a greater load than the bottom one
and it is concluded that the top nut should be the thick one with the thin
one next to the work. However, it has been shown by Goodier and
Sweeney that the first few threads next to the loaded face of a nut carry
most of the load; hence, it appears that the positioning of the nuts is not
(a)~Jam nut
J— 3§N
''>1
feLEI o she
stop nut
(TV-Lock washers
Fig. 9-4. Locking devices.
of primary importance from the standpoint of loading. Ease of assembly

is favored by placing the thick nut next to the work where it can be held
by a standard wrench during tightening, whereas placing the thin nut
on first requires a thin wrench to hold it.
Multiple threads may be used in cases in which a combination of fine
threads is required because of space limitations and rapid axial motion,
as in a fountain-pen cap or a quick-locking screw. In all standard fast¬
enings, however, a single thread of small pitch is used in order to secure
a large holding force and to lessen the tendency to shake loose in service.
1 W. E. Horenburger, Evaluating Self-locking Nuts, “ Machine Design,” p. 135,

Penton Publishing Company, Cleveland, 1946.
126 MACHINE DESIGN
Commonly used screw-thread terminology is as follows:

Pitch is the axial distance between corresponding points on two con¬
secutive turns
Lead is the distance the nut advances during one revolution
Diameters:
Major, or outside, diameter
Minor, or root, diameter
Unless otherwise stated, specifications of screw fastenings refer to the
major diameter of the fastening. Thus, a in. bolt refers to a bolt
Yi in. in outside diameter.
9-3 Forms of screw threads for fastenings. The right-hand thread
is always used unless there is a special reason for requiring a left-hand
thread and, unless otherwise stated, specifications for threads imply
right-hand threads.
The American National thread was the standard thread form for fast¬
enings in the United States until November, 1948, when the Unified
Screw Thread Standard was adopted. The Unified Standard now applies
to government and industry in Great Britain, Canada, and the United
States. The threads now standard in the United States adhere to the
Unified and American Screw Thread Standard ASA B 1.1-1949 and are
referred to as the American Unified thread form designated by UN fol¬
lowing the thread size.
All American standard threads whether unified or not are interchange¬
able among themselves and may be used with the British form, although
there may be in some pieces slight interference of metal in the threads
which is not serious. Only Unified threads, however, are interchangeable
in the sense that tolerances are completely maintained.
The differences between the new American Standard and the old
National Standard forms are that the old standard forms had flats at
the crests and roots of the threads, whereas the new standard has rounded
roots and rounded crests. During a change-over period it is optional
whether the external thread crests are rounded or flat. The thread
angle in both the old and new standards is 60 deg.
Another difference is in the tolerance and gauging system which has
been reworked to permit interchangeability, and also there are some
changes in thread series and in the classes; the old class 1 and class 4 have
been dropped and Classes 2A and 2B and 3A and 3B have been intro¬
duced. In the new classes, A refers to the external thread tolerances and
B refers to the internal thread tolerances. Classes 2A and 2B will be
used generally for fastenings (the latter for closer fits), and classes 3A
and 3B for refined fits. Actually the old classes 2 and 3 have been car¬
ried over in the new standards but their use is expected to disappear
gradually in favor of the 2A and 2B and 3A and 3B classes.
The thread form for the Unified and American Standard is shown in
Fig. 9-5.
For the three series, coarse threads are intended for general industrial
purposes; fine threads are for applications which require fine adjustment
Fig. 9-5. Unified and American screw-thread form. Threads shown separated for
clarity. Crests on external threads may be rounded or flat.
or additional tightness and resistance to loosening due to vibration. The

extra-fine series is used for special applications. Table 9-1 shows thread
sizes for the coarse and fine series threads.
In designating screw threads, the nominal size is given first (in inches
or numbers), hyphen, number of threads per inch, letters of the thread
form and series, hyphen, and class of toler¬
ance. For example: 1/2-20 UNC-2A indicates
a Unified bolt having a nominal diameter of
in., with 20 threads per inch, National coarse
series with allowances and tolerances for class
Thread-
2A fit for the bolt. The mating nut would form
carry the same designation except that it would circle
be 2B instead of 2A.
Aero threads. The Aero thread employs an
insert made of wire of section as shown in Fig.
9-6. The insert is shaped like a coil spring and
is screwed into the threaded recess in the nut or __'CD
threaded member, thus forming a thread for the Fig. 9-6. Aero thread.
engagement of the bolt or screw. The thread
spaces of the bolt or screw are approximately semicircular in section, and
as a result, little stress concentration is introduced. The bottom of the
thread in the nut or threaded member is sharp, but stress concentration
at this location is generally not serious, since the nominal stress is low.
128 MACHINE DESIGN
Table 9-1. Unified and American Screw-thread Sizes
Coarse threads Fine threads
\
Size
Threads Root area, Threads Root area,
per in. sq in. per in. sq in.
0 80 0.0015
1 (0.073) 64 0.0022 72 0.0024
2 (0.086) 56 0.0031 64 0.0034
3 (0.099) 48 0.0041 56 0.0045
4 (0.112) 40 0.0050 48 0.0057
5 (0.125) 40 0.0067 44 0.0072

6 (0.138) 32 0.0075 40 0.0087
8 (0.164) 32 0.0120 36 0.0128
10 (0.190) 24 0.0145 32 0.0175
12 (0.216) 24 0.0206 28 0.0226
H 20 0.0269 28 0.0326
He 18 0.0454 24 0.0524
% 16 0.0678 24 0.0809
He 14 0.0933 20 0.1090
y2 13 0.1257 20 0.1486
He 12 0.1620 18 0.1888
X 11 0.2018 18 0.2400
H 10 0.3020 16 0.3513
Vs 9 0.4193 14 0.4805
l 8 0.5510 12 0.6245
iX 7 0.6931 12 0.8118
m 7 0.8898 12 1.0237
m 6 1.0541 12 1.2602
iH 6 1.2938 12 1.5212
m 5 1.7441
2 4H 2.3001
2K 4H 3.0212
2H 4 3.7161
m 4 4.6194
3 4 5.6209
3H 4 6.7205
3H 4 7.9183
3H 4 9.2143
4 4 10.6084
In addition to lowered stress concentration in the screw, the Aero thread

has the advantage that it can be used in materials that are difficult to
thread, such as cast iron and aluminum. The insert can be screwed into
the hole as tapped, thus leaving the insert itself to form the thread for
receiving the screw. The insert, usually bronze, will resist wear and may
be replaced if necessary. A comparison of endurance limits of standard
threads and Aero threads is shown by the diagram in Fig. 9-7. In
Fig. 4-10(gO is shown an Aero-thread assembly for an aircraft-engine-
cylinder flange fastened to an aluminum crankcase.
Fig. 9-7. Curves showing results of endurance tests on steel specimens with Aero
thread and National Coarse thread. (From H. F. Moore, Strength of Screw Threads,
Product Eng., November, 1939.)
Another thread insert known as the “Heli-coil” is similar to the Aero-

thread insert except that the cross section of the wire is diamond shaped
so that it will form a thread that will accommodate an American Standard
threaded part. While the Heli-coil thread does not have any stress-
concentration advantage over the American Standard thread, it does
have advantages for use in materials that are difficult to thread and at the
same time maintains interchangeability with American Standard threads.
An application of this thread is in the spark plugs in cast-iron or aluminum
cylinder heads.
Pipe threads. The American Standard taper pipe thread is formed with
a taper of % in. per ft. The form of the thread is shown in Fig. 9-8. In
addition to their use in pipe fittings, pipe threads are used in attaching
oil-feeding devices, gauges, etc.
130 MACHINE DESIGN
9-4 Rolled threads. The dies for forming rolled threads are of two
types: flat dies, used in pairs which have relative motion of translation,
and cylindrical dies, which rotate and are used in sets of three. The dies
have grooves of the desired thread form with a slope on the die corre¬
sponding to the angle of the helix of the thread. The cylindrical blank
is rolled between the dies, thereby forming the threads. The thread is
formed by depressing the surface of the blank to form the root of the
thread while the displaced material flows outward to form the crest.
Thus the original diameter of the blank, or shank of the bolt, lies between
the outer diameter and the root diameter of the thread. The process is
limited to external threads.
Because of the cold forming of the material, the mechanical properties
are increased, which, coupled with favorable flow lines and uniform and
smooth surfaces, materially increases the strength of the part, especially
in cyclic loading. Close control may be had over dimensions; in fact,
excellent micrometer screws are made by the process. Threads may be
rolled on all sizes of blanks up to 5 in. in diam¬
eter. While thread rolling has been practiced
for over a hundred years, the modern precision-
rolled thread is far superior to the rolled car¬
riage bolt of early days, and in production,
marked savings may be had in labor and ma¬
terial cost as well as improved performance.1
9-5 Height of nut. The approximate thick¬
ness of a nut required for equality of strengths
of a bolt in tension and of the thread in shear
may be determined as follows:
Fig. 9-9. Height of nut. Assumptions: (a) that each turn of the
thread of the nut supports an equal share of
the load; (6) that stress concentration is neglected; (c) that for standard
coarse threads, dr = 0.8d0 (see Fig. 9-9); (d) that for steel, the yield
stress in shear $s is equal to one half the yield stress in tension st.
^‘Engineering Data on Thread Form Rolling,” Reed Rolled Thread Die Co.,
Worcester, Mass.
The strength of the bolt in tension is equal to
The strength of the threads in shear is equal to
Fs = wdrhss
By equating Ft and Fs,
h = 0.4do
Therefore, for standard coarse threads, the threads will be as strong

in failure by shear as the bolt in tension if the height of the nut is 0.4
times the nominal diameter of the bolt. American Standard nuts are
approximately %d0 in height; hence, standard threads will not fail by
shear. If such threads do fail, the failure may have been preceded by
“thread crossing” or wear. The thickness of a standard nut is made
greater than that required for simple strength to provide for sufficient
bearing area between the wrench and the nut.
The assumption that each thread supports an equal share of the load
is incorrect because of the elongation of the bolt and
the compression of the nut under load. It has been
shown that the load is concentrated on the part of
the thread near the base of the nut.1 The form of
nut shown in Fig. 9-10 allows a more equal distri¬
bution of the load and has been used in heavily
loaded bolts. In some instances, during assembly
the bolts are stretched by a bolt puller and the nuts
are screwed on by hand and then loaded as the bolt
Fig. 9-10. Shape of
puller is released. This procedure is followed in nut to reduce locali¬
order to load the bolts more accurately than is zation of thread
possible by tightening up the nuts and in order to loading.
prevent galling of the threads.
9-6 Stresses in screw fastenings. Tightening-up stresses. Tighten¬
ing-up stresses in bolts are highly indeterminate.2 The judgment of the
mechanic in selecting a wrench and in applying the force cannot be pre¬
dicted with accuracy.
An equation3 for estimating the tightening-up load is
F i = kd
1 J. N. Goodier, The Distribution of Loads on the Threads of Screws, J. Appl.
Mechanics, vol. 7, no. 1, p. A-10, 1940.
2 “Torquing of Nuts in Aircraft Engines,” SAE War Engineering Board, Society
of Automotive Engineers, New York, 1943.
3 J. H. Barr, The Stress on Bolts in Service, Sibley Journal of Mechanical Engineer¬
ing, October, 1902.
132 MACHINE DESIGN
where F1 = initial axial load due to tightening up, lb

d = nominal diameter of bolt, in.
k = 16,000 lb per in., as determined for making steamtight joints
(this value may be reduced for fastenings not set up as tight
as for a steamtight joint)
Shearing stresses. When bolts are subjected to direct shearing loads,
they should be located and threaded so that the shearing load comes upon
the body of the bolt and not upon the threaded portion (see Fig. 16-2).
If several bolts are intended to share a shearing load, high-grade construc¬
tion requires finished bolts fitted to reamed holes. In some instances, the
bolts may be relieved of shear loads by shear pins used for this purpose.
Stress concentration. Photoelastic tests indicate that the concentration
of stress at the root of a standard coarse thread is large, as shown by a
static stress-concentration factor that is equal to 5.62.1
This concentration of stress is generally not serious in bolts made of
ductile material and subject to static loads. However, in dynamic load¬
ing, the concentration of stress has
been found to reduce the endur¬
ance limit of standard coarse
threaded bolts by fatigue stress-
concentration factors equal to 2.84
for medium-carbon steel and 3.85
for SAE 2320 nickel steel heat-
treated.2 These factors were deter¬
mined for repeated tension loading
(a) (b)
of bolts threaded with no special
Fig. 9-11. Relative rigidity of bolts and
connected parts.
care toward relieving stress concen¬
tration at the juncture of the thread
and the shank of the bolt. Marked improvement in fatigue strength
may be secured by using undercut relieving grooves, as discussed in
Chap. 4.
Effect of rigidity of parts. The resultant axial load on a bolt depends
on the initial tension and on the external load. It depends also on the
relative elastic yielding (springiness) of the bolt and the connected mem¬
bers. If the connected members are very yielding as compared with the
bolt, as in Fig. 9-11(a) which shows a soft gasket, the resultant load on
the bolt will closely approximate the sum of the initial tension and the
external load. If the bolt, however, is very yielding as compared with
the connected members, as in Fig. 9-11(6), the resultant load will be either
1 S. G. Hall, Determination of Stress Concentration in Screw Threads by the

Photoelastic Method, Univ. Illinois Bull. 245, 1932.
2 Moore and Henwood, The Strength of Screw Threads under Repeated Tension,
Univ. Illinois Bull. 264, 1934.
the initial tension or the external load, whichever is greater. Actual con¬
ditions usually lie between the two extremes.
To estimate the resultant load, the following equation may be used:
F = F1 + —F2
1 + a
where F = resultant axial load, lb

Fi = initial tension due to tightening up, lb
F2 = external load, lb
a = ratio of elasticity of connected parts to elasticity of bolt
For soft gaskets and large bolts, the value of a is high and the value of
a/(1 + a) approximates unity so that the resultant load is equal to the
sum of the initial tension and the external load.
For hard gaskets or metal-to-metal contact surfaces and with small
bolts, the value of a is small and the resultant load is due mainly to the
initial tension (or to F2 in the rare case it is greater than Fi).
The value of a may be estimated by the designer to obtain an approxi¬
mate value for the resultant load.
The designer thus has control over the influence on the resultant load
on a bolt by proportioning the sizes of connected parts and bolts and by
specifying initial tension in the bolt.
For instance, a connecting-rod bolt may be tightened up with initial
tension greater than the external load. The load on the bolt then will be
static and the bolt may be designed on the basis of static failure. If the
initial tension was not high enough, the resultant load on the bolt would
be affected by the external load which is cyclic, and failure of the bolt
would be on the basis of cyclic failure and governed by the endurance
limit of the material and stress concentration, both of which would be
unfavorable.
That is the reason for the practice in the assembly of automotive
engines to tighten up connecting-rod bolts so that the stress approaches
the yield point.
It is difficult, however, to determine accurately initial stresses even
though a torque wrench is used for tightening. This is on account of the
variation in surface finish of the threads, fits, and the coefficient of fric¬
tion at the threads and faces of the boltheads and nuts. For accurate
determination of initial tension, it is necessary to measure the elongation
of the bolt.1
9-7 Bolts subjected to shock loading. When a bolt is subjected to
shock loading, as in the case of a cylinder-head bolt of an internal-combus-
1 J. O. Almen, Bolt Failure as Affected by Tightening, “Machine Design,” p. 133,

134 MACHINE DESIGN
tion engine, the resilience of the bolt should be considered in order to

prevent breakage at the thread.
Resilience is defined as the energy that is returned by a member upon
release of the load. If a member is not stressed beyond the elastic limit,
it is capable of returning all the energy of deformation; if stressed beyond
the elastic limit, it will return only a portion of the energy of deformation,
the energy not returned being consumed in permanently deforming the
member. The term “resilience” applies to a member or specimen; the
units of resilience are inch-pounds.
The resilience of a cubic inch of a material when it is stressed to the
elastic limit is called the modulus of resilience. The modulus of resilience
is equal to the area under the stress-strain curve up to the elastic limit.
It is a property of a material and is a measure of its shock-resisting ability.
Its units are inch-pounds per cubic inch and it is equal to s]/2E, in which
se is the stress at the elastic limit and E is the modulus of elasticity.
If a bolt of the usual form having a full-sized shank and threaded end,
as shown in Fig. 9-12(a), is used to support a tensile load, the stress in
the threaded part will be higher than
that in the shank. If a tensile load
is suddenly applied, the energy ab¬
sorbed by each unit volume of the
bolt will be proportional to the
square of its stress at the same loca¬
tion; hence, a large part of the
energy will be absorbed at the region
Fig. 9-12. Bolts of same strength but
of the threaded part. If the shank
one (6) having greater resilience than
the other (a).
of the bolt is turned down as shown
at (6), the shank of the bolt will un¬
dergo a higher stress, and hence will absorb a greater proportion of the
energy, thus relieving the material at the sections near the thread. If
the shank of the bolt is turned down to a diameter corresponding to that
at the root of the threads, or even slightly less, the static strength of the
bolt will not be decreased.1
The resilience of a bolt also may be increased by increasing its length.
9-8 Keys. The most common function of a key is to prevent relative
rotation of a shaft and the member to which it is connected, such as the
hub of a gear, pulley, or crank. A large number of types of keys are
available and the choice in any installation depends on several factors,
such as power requirements, tightness of fit, stability of connection, and
cost. For very light power requirements, a setscrew may be tightened
against the round shaft or against a flat spot on the shaft. For most
1 See “Torquing of Nuts in Aircraft Engines,” SAE War Engineering Board,
Society of Automotive Engineers, New York, 1943.
requirements a positive connection, such as that offered by a key, is

necessary. A setscrew is frequently used to seat the key firmly in the
keyway and to prevent axial motion of the parts.
Square keys as shown in Fig. 9-13 (a) are common in general industrial
machinery.
Flat keys, Fig. 9-13(6), are used where added stability of the connection
is desired, as in machine tools. Square or flat keys may be of uniform
cross section or they may be tapered. In tapered keys the width is uni¬
form and the height of the key is tapered by x/% in. per ft. Tapered keys
may have gib heads, as shown in Fig. 9-13 (c), to facilitate removal.
□ i-n
fa)-Square key
□(b) -F\aF
r :_!□key
B tz: (d)-Kennedy key

(c) -Taper key with
gib head
ffj-Round ^-Spline
key fitting
Fig. 9-13. Types of keys.
The Kennedy key is used for heavy duty and consists of two keys driven
as shown in Fig. 9-13 (d). The hub is bored to fit the shaft and is then
rebored off center as shown. The keys force the shaft and hub in con¬
centric position.
The Woodruff key, shown in Fig. 9-13 (e), requires a key seat formed by
a special side-milling cutter. This key will align itself in the key seat.
It has the disadvantage of weakening the shaft more than by the use of a
straight key.
The round key, or pin, Fig. 9-13(/), introduces less stress concentration
at the key seat in the shaft and is satisfactory except for the necessity
of drilling the hole to accommodate the pin after assembly of the hub
and shaft. This may be a disadvantage in production and prevents
interchangeability.
A spline fitting, as shown in Fig. 9-13 (g), is composed of a splined shaft
formed by milling and a mating hub with internal splines formed by
136 MACHINE DESIGN
broaching. The splines in reality are a number of keys integral with the
shaft. Splined fittings are adaptable to mass production and are used
where radial space must be conserved.
Straight-side splines are being replaced at an increasing rate by stub
involute splines. These splines have the advantages of greater strength,
a self-centering feature, and production economy. The standards specify
fittings with ^ to 48/96 diametral pitches, from 6 to 50 teeth, all with a
30-deg pressure angle.1
Keyways for straight keys are formed either by a side-milling cutter,
which forms a sled-runner key way [Fig. 9-14 (a)] or by an end-milling
cutter forming profiled keyways, illustrated at (6). The sled-runner key¬
way requires a longer space between the end of the key and the end of
the keyway than does an end-milled keyway. This favors the end-milled
keyway in locations near a shoulder. However, the end-milled keyway
(a)-Side-milled or End-milled
sled-runner or profiled
Fig. 9-14. Types of keyways.
reduces the endurance limit of a shaft more than does the sled-runner
type.
Feather keys are used where it is necessary to slide a keyed gear or pulley
along the shaft. The key is generally tight in the shaft and clearance is
provided between the key and the hub keyway.
When a gear or pulley must be moved axially along the shaft while
power is being transmitted, it is desirable to provide for a minimum of
axial force. The use of two feather keys equally spaced requires a smaller
axial force than the use of one key. To show this, consider the axial force
required to slide the gear hub along the shaft of diameter d when a torque
T is being transmitted for the following cases: (a) when one feather key
is used and (b) when two feather keys are used.
Because of clearance between the bore and the shaft and between the
key and the keyway, the shaft will assume a position approximating that
shown in Fig. 9-15(a). For this case,
T = Fi ^ approximately
1 See ASA B5.15, 1946, and also “SAE Handbook.”

or
2T
Fl =
d
The axial force required to move the hub axially is

4/T
2fF, =
d
where / is the coefficient of sliding friction.
(a)'One feather key 76V-Two feather keys

Fig. 9-15. Forces for one and two feather keys. Clearance is exaggerated.
Where two feather keys are used, the shaft will assume a position as
shown in Fig. 9-15(6). For this case
T = F 2d approximately
or
F^-d
The axial force necessary to move the hub axially is
2fT
2fF2 =
d
By comparing the two equations for the axial forces, it is evident that
the axial force required for two feather keys is one-half that for one
feather key.
Stresses in keys. When a key is used in transmitting torque from a
shaft to a rotor or hub, the distribution of stresses in the key is three-
dimensional and is complicated in nature. These stresses are caused by
forces that are of two classes: (a) forces due to the fit of the key in its
keyway, as in a tight-fitting straight key or in a tapered key driven in
place. These forces produce compressive stresses in the key and it is
impossible generally to predict their magnitude. (6) Forces that are
caused by the torque transmitted. These forces produce compressive
and shearing stresses in the key. The distribution of the forces along
138 MACHINE DESIGN
the length of the key is not uniform, since the forces are concentrated near
the torque-input end. The nonuniformity of distribution is caused by
the twisting of the shaft within the hub. The exponential curve in Fig.
9-16 shows the approximate distribution1 of the
stress along the length of a key.
In order to obtain a relation between stresses and
loads that may be used for design purposes, it is
customary to neglect the forces due to the fit of the
key and to assume that the distribution of forces
along the length of the key is uniform.
Proportions of a key. A proportion that has given
good results in practice makes the width of a key
Fig. 9-16. Approxi¬ one-quarter the diameter of the shaft. The thick¬
mate distribution of
ness of a key for equal strengths of the key in
tangential load along
length of a key. failure by shearing of the key and compression on
the key may be determined for steel by considering
the corresponding allowable stresses in shear and compression.
The forces acting on a key for a clockwise torque being transmitted
from a shaft to a hub are shown in Fig. 9-17. The forces F' on the top
and bottom of the key resist tipping of the key. The force F between the
Fig. 9-17. Forces on a key due to transmitted torque.
side of the key and the key way in the hub is due to the resisting torque T
and may be determined from the relation
approximately
On the basis of the assumptions stated above, the strength of the key
for failure by shearing of the key is
1 J. P. Den Hartog, Trans. ASME, vol. 54, p. APM-54-10, 1932.

and its strength for failure by compression of the key is
rp _ Fd _ tLscd
Jc ~ T 4
By equating the two torques and using the relation that sc = 2ss, which
is approximately true for steel, it is found that w = t; hence a square steel
key is approximately as strong in failure by shear and crushing.
A square key whose sides are one-quarter the diameter of the shaft
may be stronger or weaker than the shaft, depending on the length of
the key. The length of the key required for equality of strengths of the
key and the shaft may be determined by equating the strength of the
key in shear to the strength of the shaft in torsion. Therefore, neglecting
all stress concentrations,
wLssd ird3
- = — So
2 16
Substituting d/4 for w, it is found that
approximately
Thus for equal strengths of the key and shaft, the length of the key
should be 1.5 times the shaft diameter.
The above discussion is the basis for the proportions of a key that has
wide industrial use, i.e., a square key whose sides are one-quarter of the
shaft diameter and whose length is at least 1^2 times the shaft diameter.
For gears, and for rotors of large diameter, stability may require a longer
hub, and a key of corresponding length may be used. Also, shafts
designed on the basis of stiffness are generally overly strong, and hence
short keys may be used.
In order to limit the number of stock sizes of keys, standard key sizes,
which may vary somewhat from one-quarter shaft diameter, are approved
by the ASA and are given in Table 9-2. Flat-stock key dimensions are
also given in the table.
It may be noted that a flat key is weaker in compression than in shear;
however, the use of flat keys is generally confined to equipment with the
shaft sizes that are required for stiffness, which usually have an excess of
strength, so that the standard size flat key has ample strength.
The assumption that stress concentration may be neglected is no doubt
justified since, as shown in Fig. 9-16, the key is subjected to a concentra¬
tion of load and stress; also, the key way introduces stress concentration
in the shaft, so that the stress concentrations will affect both elements
and can be neglected to obtain approximate relations of dimensions. In
140 MACHINE DESIGN
Table 9-2. Standard Dimensions of Plain Parallel Keys
Key dimensions, in.
Shaft diameter, ins

Thickness
(inclusive)
Width
Square key Flat key
X~Xe X X X2
%-Vs He He X
15Ae-ni X X He
He He X
V/ie-m Hs X X
11SH 6-2H X X X
2He~2H X X He
21He-SH H H X
3H6-3H Vs Vs X
3%-4^ l l H
4% e-5X IX IX Vs
5%e-6 IX ix l
addition, the effect on seriousness of stress concentration of the ductility

of the material should be considered. In a steel shaft and key made of
ductile material and subjected to a static load, stress concentration may
not be serious. However, with variable loading, as in a shaft with tor¬

sional vibration, the stress concentration may be serious.
9-9 Knuckle joints and cotter joints. Knuckle joints. These joints
as illustrated in Fig. 9-18 are used to connect two rods that are under the
action of tensile loads, although if the joint is guided the rods may support
a compressive load. They are used where it is desired to have a joint
that can be readily disconnected for adjustments or repairs. The ends
of each of the rods are formed by upsetting the rod and machining to
form the eye, as shown at the right-hand part of the joint in Fig. 9-18,
and the yoke or fork, shown at the left in the figure. The knuckle pin
forms the connecting element and is held in place by a small cotter pin or
other fastening.
The methods of failure of the joint are indicated below, and the strength
of the joint for each method of failure is indicated. Stress concentra¬
tion and nonuniform distribution of loads are neglected; therefore the
strengths given below will be approximate; however, they serve to indi¬
cate a well-proportioned joint. The safe loads for the rod would require
an appropriate factor of safety.
1. Failure of the solid rod
2. Failure of the knuckle pin in shear
3. Failure of the rod end by shearing
F = (d2 — di)ass approximately
4. Failure by shearing of the fork ends
F = (d2 — di)2bss
5. Failure of the eye in tension
F = (d2 — d\)ast
6. Failure of the fork ends in tension
F = (d2 — di)2bst
7. Failure of the rod end in compression by bearing against the pin
F = adisb
8. Failure of the fork ends in compression by bearing against the pin
F = 2bd\Sb
9. Failure of the pin in bending. Assuming that the pin is loaded as

shown by Fig. 9-19, the maximum bending moment on the pin occurs at
the midsection of the pin and is equal to
F
approximately
2
142 MACHINE DESIGN
Therefore
37rd\st
F =
4(3a + 46)
A well-designed joint is one that has equal strengths in failure for all
the above methods of failure. For a particular material, such as steel,
the relation between the stresses would be fixed,
which would be used to determine the dimensions
of the joint. From the above equations, it is
Ta. obvious that dimension 6 should be equal to one-
half a. However, 6 is generally made somewhat
1 greater than one-half of a in order to prevent
b deflection or spreading of the forks, which would
±.
introduce excessive bending of the pin.
Cotter joints. A cotter joint is used to connect
Fig. 9-19. Loading of
two rods rigidly. The joined rod may carry ten¬
pin in knuckle joint.
sile or compressive loads. As shown in Fig. 9-20,
there are three parts to the joint, the rod end (left-hand end), the socket end
(right-hand end), and the cotter that fits into the tapered slot. The
taper of the slot as well as the cotter is usually on one side, the right as
shown in the figure. Clearance between the cotter and the slots in the
rod end and socket allow the driven cotter to draw together the two parts
of the joint until the socket end comes in contact with the shoulder on
the rod end at a. Further driving of the cotter will bend it in the plane
of the side view, and this bending will aid in keeping the joint tight under
the action of a variable load. The dimensions that may be used in inves¬
tigating the relative strengths of the rod in various methods of failure are
indicated by symbols on the sketch.
Tightening of the cotter introduces initial stresses that are difficult to
estimate.
Failure of the joint may occur in several ways, the most probable of
which are as follows:
1. Tension failure of the rods at the diameter d
2, Tension failure of the rod or socket across the slot
3. Shear failure of the rod or socket between the slot and the end of the
rod or socket
4. Shear of the cotter
5. Bearing between the cotter and the slot in the rod end or in the
socket
6. Failure of the cotter in bending
7. Failure of the collar of the rod end in shear due to tightening of the
cotter or to any compressive load to which the joint may be subjected
9-10 Pins and snap rings. Pins and snap rings are two examples of a
large number of fastenings of a general type that is finding increasingly
popular application in machinery, especially in light machinery parts.
Pins may be used in straight, tapered, or grooved forms. In function,
they may be classified as locating pins, called dowel pins, which fix the
Enlarged
Section
Fig. 9-21. Grooved pin.
relative position of two parts; shear pins which transmit service loads;
safety pins which have the function of a shear pin but which are designed
to fail at a predetermined load and thus protect expensive parts from
damage. Safety links, designed to fail in tension under the action of an
overload, perform the same function as a safety pin.
Pins should be accurately fitted into reamed holes and located so that
bending of the pin will be eliminated or reduced to a minimum. Provi¬
sion should be made for pins to be removed. This may require a through
hole so that the pin may be punched through, or it may require a plain
or a threaded head for drawing.
The grooved pin has a rolled groove, generally tapered, as shown in
Fig. 9-21. The rolling of the groove deforms the pin, as shown by the
enlarged sectional view, so that when the pin is driven or pressed into a
drilled hole, a tight fit is produced between the pin and the hole. The pin
deforms elastically, and it is effective under the action of variable loading
without loosening of the pin. Some applications of pins are shown in
Fig. 9-22.
Snap rings are of two general types, external and internal, as shown in
Fig. 9-23 (a) and (6). The relations between the dimensions of the ring
and of the recesses that accommodate them should be considered so as
144 MACHINE DESIGN
used as a dowel
f^J-Shear pin for bearing)
(d)-Grooved pin
used as a key
Fig. 9-22. Application of pins.
(b^-Internal snap rings

Fig. 9-23. Snap rings.
to provide for sufficient strength of the ring against permanent set during
insertion or under load and to provide for ease of insertion or removal.1
Snap rings, and also wire clips that perform the same function, are avail¬
able in various forms and sizes. Figure 9-24 shows snap rings used to
retain a ball bearing in place. The larger ring supports an axial load on
the shaft to the right in the figure. The smaller ring retains the inner race
of the bearing against the shaft shoulder. Since ease of insertion is nec¬
essary in production, clearance must be provided as the rings cannot be
seated under axial load. If end play is not permitted, screwed locking
fastenings should be used instead of snap rings.
1 Peter F. Rossman, Designing Snap Ring Fastenings, Machine Design, May, 1941.
CHAPTER 10
SPRINGS
10-1 General discussion.1 Springs have a variety of applications in

machinery, as follows:
1. Springs are used to control forces due to impact or shock loading
and to control vibration. The energy of impact loading involves the
product of a force and a distance that corresponds to the deflection of the
resisting member. Thus, to absorb a given amount of energy, increasing
the deflection decreases the transmitted force. A spring allows consider¬
able deflection for a moderate load, and therefore may be used to reduce
Fig. 10-1. Helical compression spring.
the magnitude of the transmitted force in shock loading. Examples of

springs for this use are buffer springs for elevators and for car-track ter¬
minals, springs for railway cars and automobiles, and springs in cushioned
gears and sprockets.
2. Springs are employed to control motions and to apply forces to
members. Examples of such uses are valve springs in internal-combus¬
tion engines, springs to produce pressure in brakes and clutches, and
flexible supports for machinery or equipment.
3. Springs are used for storing energy, for example, in clocks, switch¬
throwing devices, and starters.
4. Springs are employed to measure forces, as in scales.
1 For a complete discussion of springs, see A. M. Wahl, “Mechanical Springs,”
146
SPRINGS 147
Springs may be classified in accordance with their shape, such as helical,
spiral, leaf, etc. The characteristics of some of the more common types of
springs are discussed in the following articles.
10-2 Helical springs. In Fig. 10-1 (a) is shown a helical coil spring
with ends adapted to support a compressive load. At (6) the spring has
been deflected by the axial load P, which may be assumed to compress
the spring as between two parallel plates. The following notation
applies:
Let P = axial load, lb
D = mean diameter of the coils, in.
d = diameter of the wire, in.
p = pitch of the coils, in.
8 = deflection of the spring, in.
n = number of active coils
C = spring index = D/d
G = torsional modulus of elasticity, psi
ss = shearing stress, psi
The number of active coils, n, represents the coils in the spring with
the exception of those which lie flat against the
compression plates. These inactive coils do
not contribute to the deflection of the spring.
Stresses in a helical spring made of round
wire. In Fig. 10-2 is shown a part of a com¬
pression spring that supports a compressive
load P and a section of the wire cut by axial
plane. The part of the spring shown in the
figure is in equilibrium under the action of the
two forces P and the resisting torsional mo¬
ment T, as shown in the figure. The latter Fig. 10-2. Loading on helical
spring.
equals
T =
The shearing stress due to the torque T is
8 PD
ss = - (10-1)
7rd3
In addition to the stress given by Eq. (10-1), the direct shearing stress
due to P should be included. By adding the direct shearing stress, which
equals ^P/ird2, to the torsional stress SPD/irdd, the maximum shearing
stress, which is located at the inner side of the curved wire, may be deter¬
mined by
8PD 4P_ _ 8PD / J_\
(10-2)
7rdz 7rd2 7rd3 \ 2C/
148 MACHINE DESIGN
It is evident from Eq. (10-2) that for springs of small index C the effect
of direct shear, (1/2C), is appreciable.
The effect of curvature of the wire as it forms the coil should be con¬
sidered also.
In order to include the effects of both direct shear and wire curvature,
a stress factor has been determined by the use of approximate analytical
methods by A. M. Wahl1 which may be used with Eq. (10-1) to deter¬
mine the maximum shearing stress in the wire as follows:
, = ^8PR = (10-3)
* ird3 ird2
where
4C - 1 0.615
~ 4C - 4 + C
Values for K may be read from Fig. 10-3. The values of K as indicated
by the Wahl equation have been verified by experimental means using
Fig. 10-3. Wahl stress factor for helical springs.
strain measurements secured by sensitive extensometers, and have been

verified also by exact calculations based on the theory of elasticity, which
indicates that results are correct within 2 per cent for spring indexes
greater than 3. This includes most springs used in machinery. The
Wahl stress factor has been adopted generally for the design of helical
springs.
Deflection of helical spring. An equation for the axial deflection of a
helical spring in terms of the axial load, spring dimensions, and a materials
constant may be conveniently determined by equating the work required
to deflect the spring to the strain energy in the twisted wire. For close-
1 A. M. Wrahl, Stresses in Heavy Closely Coiled Helical Springs, Trans. ASME,

vol. 51, p. APM-51-17, 1929.
SPRINGS 149
coiled springs the bending of the wire is small and the strain energy of
bending may be neglected.
The axial load as shown in Fig. 10-1 increases linearly from zero to P,
and hence the work required to compress the spring is the average force
P/2 times the deflection, or P8/2.
The strain energy in a bar twisted by a torsional moment T through
a total angle 6 is TO/2. The total angle of twist 0 = TL/GJ, where L
equals the length of the twisted wire and J equals the polar moment of
inertia of the wire section. The active length of wire in the helical spring
equals irDn/cos a, where a is the lead angle of the helix, which for close-
coiled springs is of the order of 5 deg; hence cos a is approximately unity.
For springs of indexes as used in machinery, namely, 2^ or 3 and
above, the strain energy in the curved wire forming the coils will be close
to the strain energy in a straight wire; hence from the above relations the
following equation for 8 may be obtained:
8PDzn _ SPC3n
(10-4)
~GdT ~ GdT
Equation (10-4) gives values that have been accurately verified by

test.1 The value of G for spring steel is approximately 11,500,000 psi.
By solving Eq. (10-4) for P/8, a useful quantity is obtained, i.e.,
P Gd
(10-5)
8 ~ 8CH
It may be noted that the right-hand side of this equation is constant

for a particular spring; hence P/8 is constant. Its units are pounds per
inch, its physical interpretation is the number of pounds required to
deflect the spring 1 in., and it is known as the spring rate. The spring
rate represents the slope of the load-deflection line as shown in Fig. 10-4.
Any difference between the ideal line and the actual curve is due prin¬
cipally to the change in number of active coils as the spring is deflected,
to elastic hysteresis of the material, and to friction between the end coils
in contact with each other or with the loading plate.
Compression springs in which the free length is more than four times
the mean diameter of the coils may fail by sidewise buckling. If space
limitations make the use of such slender springs necessary, they may be
guided on a central rod or mounted in a hole or tube, but the coils are
apt to bind and cause wear. Instead, the spring may be resolved into a
series of short springs separated by guided bushings. The designer
1 A. M. Wahl, Further Research on Helical Springs of Round and Square Wire,

Trans. ASME, 1930.
150 MACHINE DESIGN
Fig. 10-4. Load-deflection characteristics for an indicator spring in compression.
Mean diameter
Fig. 10-5. Buckling will occur at points above curve.
should be aware, however, of the increase in diameter of a spring when it

is compressed. This is owing to the changing of slope of the coils and to
the unwinding of the coils.1
1 See SAE War Engineering Board Manual, “Design and Application of Helical and
Spiral Springs for Ordnance,” Society of Automotive Engineers, New York, February,
1943.
SPRINGS 151
The curve in Fig. 10-5 may be used to determine the possibility of

buckling of a helical compression spring with squared and ground ends,
mounted so that the ends move only in the axial
direction.
Springs of noncircular wire. Helical springs may
be made of noncircular wire, such as square or
rectangular, wound flatwise or edgewise, for two
purposes: (a) to provide greater resilience in a
given space, or (5) to provide for predetermined
altering of the spring rate by grinding off the out¬
side of the coils. With regard to the latter use of
rectangular-wire springs, the required calculations Fig- 10-6. Helical
become very complicated for round-wire springs spnng made of rec-
. i .• i . i /» , , . tangular wire,
but are relatively simple for rectangular wire,
especially for edgewise winding as shown in Fig. 10-6.
The maximum shearing stress in a helical compression or tension spring
made of square wire is equal to
2A0PD
SS = K 63 .
The axial deflection is equal to

* 5.58 PCzn
5 = m
where P = axial load on the spring, lb
D = mean diameter of the coils, in.
b = width or thickness of square wire, in.
C = spring index = D/b
K = stress factor (see Fig. 10-3)
10-3 Spring material.1 To aid the designer in the selection of spring
material, the service for which helical springs are used may be broadly
classified as severe, average, and light. While the classification in a par¬
ticular application depends largely on the designer’s judgment, in general,
severe service includes rapid continuous loading where the ratio of min¬
imum to maximum stress is one-half or less, as in valve springs. Average
service includes the same stress range as in severe service but with only
intermittent operation, as in engine-governor springs. Light service
includes springs subjected to loads that are static or very infrequently
varied, as in safety-valve springs.
1 A. M. Wahl, General Considerations in Designing Mechanical Springs, Machine
Design, April, 1938.
152 MACHINE DESIGN
Music wire (ASTM A228-41) is suitable for small springs made of wire
up to 0.028 in. in diameter, or if necessary for strength and finish up to
0.125 in. Oil-tempered wire (ASTM A229-41, about SAE 1065) is suit¬
able for low stresses and noncritical uses. Valve-spring wire (ASTM
A230-41) is used when long life is important, or alloy steel wire (SAE
6150) may be used. For high-temperature applications, tungsten high¬
speed steel (18-4-1) or molybdenum type (6-6-2) is suitable. Nonferrous
materials include phosphor bronze, beryllium copper, silicon bronze,
nickel-silver alloys (German silver), or nickel alloys, such as Monel,
K Monel, Iconel, Iconel X, and Z nickel may be used. As to forming,
springs made of wire less than Y in. in diameter are formed cold; wire
over in. is hot formed.1
10-4 Allowable stresses. Static loading. For helical coil springs sub¬
jected to static loading, the elastic limit in torsion may be used as a basis
for determining the working stress. For springs subjected to light serv¬
ice, a factor of safety of 1.5 is suggested. The factor may be used to
divide into the yield-point stress to obtain the working stress, or to mul¬
tiply the actual spring load to obtain a limit load to use for design pur¬
poses. In case failure of the spring would have serious consequences,
the factor of 1.5 should be raised, whereas if breaking of the spring would
cause only minor inconvenience, the factor may be lowered somewhat.
In determining the maximum stress induced in the wire, the effect of
ductility of the material should be considered. The value of the Wahl
stress factor K includes both the effect of curvature of the wire and of
direct shearing stress in the wire. The direct shearing stress is uniformly
distributed over the wire section. However, the stress augment due to
wire curvature is concentrated at the inside of the coil. This latter stress
has the nature of a localized stress or a stress concentration and, as dis¬
cussed in Chap. 4, this type of stress concentration is not serious in ductile
materials subjected to static loads. Therefore the use of the full value
of K in Eq. (10-3) will result in conservative design. Equation (10-2)
will give satisfactory results for springs subjected to essentially static
loads. Figure 10-7 indicates approximate values for elastic limits in tor¬
sion for chrome-vanadium spring steel wire, SAE 6150.
Fatigue loading. The same general principles that govern working
stresses for members subjected to fatigue loading apply to springs; i. e.,
the working stresses should be based on the endurance limit of the mate¬
rial, and full consideration should be given to stress concentration, sur¬
face finish, range of stress, notch sensitivity of the material, and the pos¬
sibility of corrosion fatigue. The maximum stress induced in the wire
can be satisfactorily determined by the use of Eq. (10-3), although for
1 See F. P. Zimmerli, Proper Use of Spring Materials, SAE Meeting, French Lick,
Ind., June, 1947.
SPRINGS 153
some ranges of stress and for some types of materials this equation gives
results that are somewhat high.1
In order to determine allowable shearing stresses, several methods have
been proposed to allow for the effect of range of stress, i.e.t the ratio of
minimum stress to maximum stress in the wire. The most promising of
these methods1 is based on the consideration that the actual stress is com¬
posed of a variable stress superimposed on a steady stress (see Art. 5-4).
For example, as shown in Fig. 5-1, an assumed stress range in a spring of
sraax = 30,000 psi to smin = 20,000 psi would be considered as a variable
stress sv = ± 5,000 psi superimposed upon a static stress sm = 25,000 psi.
Diameter of wire, in.

Fig. 10-7. Torsional elastic limit for SAE 6150 spring wire.
As a further step, the effects of stress concentration and of the “sensitivity

index ” are considered separately in their influence on the two components
of stress, since one component is static and the other variable.
Since endurance data for spring wire under all representative conditions
of service are not available to date, it appears satisfactory at the present
time to use Eq. (10-3) with the full value of the Wahl stress factor for
determining the induced stress and to use available endurance data, as
in Fig. 10-8, to determine allowable stresses. This procedure will result
in conservative design.
In Fig. 10-8 are shown allowable torsional stress ranges for SAE 6150
steel. The ordinate between the 45-deg line and the line for a particular
size wire indicates the allowable stress range for any value of the minimum
1 See A. M. Wahl, Analysis of Effect of Wire Curvature of Allowable Stresses in

Helical Springs, Trans. ASME, vol. 61, p. A-25, 1939.
154 MACHINE DESIGN
stress. For example, if smin = 35,000 psi, and the wire size is 0.344 in.,
then smax = 67,000 psi, or the stress range is 35,000 to 67,000 psi as
indicated in the figure.
Fig. 10-8. Allowable torsional-stress range for SAE 6150 steel.
For general-purpose helical springs, design stresses used by Westing-

house Electric Corporation are given in Table 10-1. These stresses
should be used with Eq. (10-3) and apply to SAE 6150 oil-tempered hot-
wound springs heat-treated after forming. The stresses are conservative
and may be increased if conditions warrant.1
Table 10-1. Spring Design Stresses, psi
Wire diameter, in. Severe service Average service Light service
Up to 0.085 60,000 75,000 93,000

0.085-0.185 55,000 69,000 85,000
0.185-0.320 48,000 60,000 74,000
0.320-0.530 42,000 52,000 65,000
0.530-0.970 36,000 45,000 56,000
0.970-1.5 32,000 40,000 50,000
1 A. M. Wahl, “Mechanical Springs,” p. 135, Penton Publishing Company, Cleve¬

land, 1944; “Handbook of Mechanical Spring Design,” Division of Associated Spring
Corp. Bristol, Conn., 1948; Kent’s “Mechanical Engineers Handbook, Design and Pro¬
duction,” 12th ed., sec. 11, John Wiley & Sons, Inc., 1952.
SPRINGS 155
Example 10-1. A helical valve spring is
to be designed for an operating load range
of approximately 20 to 30 lb, that is, 20 lb
when the valve is closed and 30 lb when
open. The deflection of the spring (valve
lift) in the above load range is to be about
0.3 in. Assuming severe service and a
spring index of 10, determine the size of
wire, size and number of coils, and pitch you
would recommend.
Note: Since the engine for which this
valve is to be used will operate intermittently Deflection, in.
and since failure of the spring would not
Load-deflection chart.
have serious consequences, the design stress
may be assumed as about 25 per cent in excess of the recommendations in Table 10-1.
Solution: From Eq. (10-3),
8KPC 8 X 1.15 X 30 X 10 880

d2 =
TTSg TTSs
Size of wire: For a trial, assume the size of the wire will be between 0.085 and
0.185 in. in Table 10-1. Then the allowable stress will be 1.25 X 55,000 = 68,700 psi.
880
d2 = 0.128 in.2
68,700
or
d = 0.113 in.
Use No. 11 wire, diameter = 0.120 in., from Appendix XIV. If the calculated size of
the wire had not been in the range assumed, further trials would have been necessary.
Number of coils: The figure shows that the spring rate is
P 10
— = — = 33.3 lb per in.
5 0.3
The spring rate represents the slope of the load-deflection line and is constant over the
entire range.
Equation (10-5) may be written as
Gd 11,500,000 X 0.120
n = -— = —-1- = 5.18
8C3(P/<5) 8 X 103 X 33.3
Use 5^ active coils or 7J>£ total number of coils to allow for one coil at each end of
the spring for squaring and grinding as in Fig. 10-10(d).
The actual spring rate is
P _ Gd 11,500,000 X 0.120
31.4 lb per in.
!) ~ SChi 8 X 103 X 5.5
This differs from the previous value because of rounding out the size of wire and
number of coils.
Pitch of coils: A valuable feature of a compression spring is that there is an auto¬
matic stop to the deflection when the coils come into contact and the spring becomes
156 MACHINE DESIGN
solid. If the shear stress in the wire at this condition is below the yield point, the
spring will not be damaged due to overstressing the wire in shear. It is therefore
desirable to specify the pitch of the coils so that if the spring is accidentally or care¬
lessly compressed solid the stress will not exceed the yield point in torsion. From
Fig. 10-7, the yield point in torsion in 135,000 psi. The corresponding load on the
spring is
7rd2ss _ 7T X (0.120)2 X 135,000
66.4 lb
8KC ~ 8 X 1.15 X 10
The corresponding deflection of the spring is
66.4
= 2.11 in.
31.4
From the equations in Art. 10-5 for spring lengths, the compressed-solid length of
the spring is
{n + 1 )d = (5.5 + 1) X 0.120 = 0.78 in.
The free length of the spring is the compressed-solid length plus the deflection
required to compress it solid; thus the free length of the spring is
0.78 + 2.11 = 2.89 in.
This is equal to (np + d), from which the pitch may be found equal to 0.50 in. This
is the pitch which would allow a stress at coil closure equal to the yield point. The
actual pitch, however, should be less than 0.50 in., to provide a margin of safety as
well as to conserve space. By assuming a pitch equal to %% in., the free length of
the spring becomes
np + d = 5.5 X 0.3125 + 0.120
= 1.84 in. Use V/& in.
The outside diameter of the spring is
(C + 1 )d = (10 + 1) X 0.120 = 1.44 in. Use in.
Care must be used in reducing the pitch of the coils as was done above so that the
spring does not close up before the maximum service load is reached.
The specifications for the spring are: outside diameter, 1%6 in.; size of wire,
No. 11 W and M gauge; pitch of coils, %6 in.; 7Y2 coils with squared and ground ends.
10-5 End connections for helical springs. A compression helical

spring has a number of advantages over a tension spring as follows:
1. There is less stress concentration at the end coils of a compression
spring than at the loop or other attaching device of tension spring.
2. Breaking of the wire of a compression spring will not necessarily
result in failure of the spring, since the coils adjacent to the break will
contact and permit the spring, if it is held in line, to function until repairs
can be made.
3. Deflection of a compression spring to the solid condition acts as an
automatic limit to the stress in the spring.
The free length of a coil spring is approximately np + d and the com¬
pressed solid length is approximately (n + l)d (see Fig. 10-1).
SPRINGS 157
Washburn and Moen wire gauge is usually standard for steel spring
wire (see Appendix XIV); Brown and Sharpe is standard for nonferrous
materials. The wire size should be specified in
decimals.
A spring used in compression for a tension load is
shown in Fig. 10-9. End constructions for helical
compression springs are shown in Fig. 10-10,1 and
tension-spring connections are shown in Fig. 10-11.
There may be a large stress concentration introduced
at the loop of a tension spring, as shown in Fig.
10-11(a). The stress concentration depends on the
ratio of ro/ri. One spring manufacturer2 states
that a stress factor equal to r0/ri used with Eq.
(10-1) gives satisfactory results.
Tension springs3 are generally wound with initial
tension to facilitate handling, so that the spring does
not begin to extend until the external load over¬
comes the initial tension.
In all springs there is eccentricity of loading
introduced by the end connections. This eccen¬
Fig. 10-9. Compres¬
tricity is generally not appreciable in compression sion spring arranged
springs having squared and ground ends, large spring for tension rod.
index, nd six or more active coils.4
10-6 Surge in springs. When a helical spring rests on a rigid support
at one end and the other end is deflected by a suddenly applied load, the
coils of the spring will not momentarily share the deflection equally, since
time is required for the propagation of stress along the spring wire. An
analysis of the progression of the deflection is that the end coil in contact
with the applied load deflects, and then this coil transmits a large part
of its deflection to the adjacent coil. Thus, in progression, a wave of com¬
pressed coils travels along the spring to the supported end, where it is
“reflected” and travels back to the deflected end. In the absence of
damping, this wave travels along the spring indefinitely. The presence
of damping, however, causes the wave to die out. If the time required
for the wave to travel from one end of the spring to the other and return
coincides with the time interval between load applications, a condition of
1 See Design Work Sheets, Product Eng.

2 Barnes, Gibson, Raymond, “The Art and Science of Spring Making/’ 4th ed.,
p. 47, Division of Associated Spring Corp., Detroit and Ann Arbor, Mich, 1941.
3 “Handbook of Mechanical Spring Design,” Division of Associated Spring Corp.,
Bristol, Conn., 1948.
4 C. T. Edgerton, Stresses in Helical Compression Springs—Present Status of
Problem, Trans. ASME, vol. 61, 1939.
158 MACHINE DESIGN
resonance is reached and very large deflections of the coils will be pro¬
duced with correspondingly high stresses. Under this condition failure
of the spring may result. To avoid the possibility of resonance or surge
in springs, it is advisable to ensure that the natural frequency of the spring
be considerably removed from the frequency of application of the load.1
(c)-Squared (ct)-Squared
and ground
10-10. Compression spring ends.
(b)
Fig. 10-11. Tension springs.
The condition of surge has been a problem in valve springs in internal-

combustion engines in connection with spring breakage and valve
fluttering.
The following equation may be used to determine the natural frequency
of a steel coil spring, as described above.
_ 761,500d
Jn nD2
1 W. M. Griffith in Engineering Standards for the Design of Springs, Product Eng.,
July, 1939, states that the natural frequency of the spring should be at least 20 times
that of the frequency of the applied load.
SPRINGS 159
where /„ = natural frequency, cpm

d = diameter of spring wire, in.
D = mean diameter of coils, in.
Stranded-wire springs. In some applications of springs, such as in
machine-gun mechanisms, the velocity of application of the load is high
and the reflection of stress waves in the spring may give rise to resonant
stresses. The low damping capacity of steel may permit these stresses to
reach large magnitudes which may result in short life of a conventional
spring.
Springs made of stranded wire have shown improved performance in
these applications. The springs are usually wound with wire made of
three strands of preformed wire in which the direction of winding the
(a)-Flat spring
Fig. 10-12. Forms of leaf springs.
strands is opposite to the direction of winding the coils to form the spring,
so that the deflection of the spring tends to wind the individual wires
closer together, which introduces frictional damping and thus decreases
the magnitude of the transient stresses. The fatigue life of stranded-
wire springs has been increased as much as four times over the life of
conventional springs.1
10-7 Leaf springs. Leaf springs may have the form of a single leaf,
as shown in Fig. 10-12(o), or they may be laminated, as shown at (b).
The laminated form is used to secure large resilience within a small space.
A common trouble with laminated springs is fatigue failure of the leaves.
Some factors contributing to this type of failure are the weakening effect
of the hole if the center bolt is used, contact pressures produced by U bolts
and rebound clips, stress concentration caused by improperly shaped
leaf ends, initial curvature and relative change of curvature of the leaves
during loading, vibration of the spring ends during rebound, which may
1 H. H. Clark, Stranded Wire Helical Springs for Machine Guns, Product Eng.,
July, 1946, p. 154.
160 MACHINE DESIGN
cause breakage of a leaf at an unloaded location, and improper heat-

treatment, especially surface decarburization. Each of these factors has
at least a partial cure, which
: -r11 h should be incorporated in leaf¬
- h—r n=i-1—1 spring design.1
= -M-
WJJtfcdt—1 Stresses and deflections in leaf
springs. The equations for maxi¬
(o) (01 !
mum stress and for deflection in
o
)
J J J 1 leaf springs may be closely deter¬
(a) Leaf spring mined by the method shown in the
following example of a cantilever
leaf spring. In Fig. 10-13(6) is
shown a flat cantilever made up
by splitting the leaves of the ac¬
tual spring shown at (a) and laying
them side by side to form a beam
of uniform thickness t, as shown
by the dotted line in (6) which is
a beam of uniform strength. The
breadth at the fixed end is 6, which
is the sum of the widths of the
(b) Equivalent flat cantilever composite leaves.
Fig. 10-13. Leaf spring and equivalent The equations are for the maxi¬
cantilever. mum stress
6 PL
$
bt2
and for the deflection at the free (loaded) end
= 6PL3
6 ” Ebt3 %
where E is the modulus of elasticity in pounds per square inch and the
other notation is as shown in Fig. 10-13.
Laminated springs permit a saving in material and a greater deflection
than a spring of constant depth, that is, they have greater resilience and
shock-absorbing ability.
10-8 Other types of metallic springs. Concentric helical springs.
These springs are usually used in compression and provide a large amount
of resilience in a small space, and are also used for safety in the event of
1V. L. Maleev and J. B. Hartman, “ Machine Design,” p. 305, International
Textbook Company, Scranton, Pa., 1954; “ Manual on Design and Application of Leaf
Springs,” SAE War Engineering Board, Society of Automotive Engineers, New York,
1944.
SPRINGS 161
failure of one of the component springs. The direction of winding of the
two helices is opposite to avoid locking of the coils in the event of sideways
misalignment or buckling. The design of concentric helical springs
should be based on equal maximum stresses in the wires of the two
springs.
Conical and volute springs. Springs of this type are made of round or
rectangular wire with coils of decreasing size to provide for telescoping of
the coils. The decrements of coil sizes in relation to the wire size may
provide for partial or complete telescoping in springs made of round wire,
or for complete telescoping of springs made of rectangular wire wound
with the long side of the wire parallel to the axis of the spring.
A comparison of the construction of a conical and a volute spring shows
that the conical spring is wound with a constant axial pitch, resulting in
Fig. 10-14. Conical and volute springs.
a spring with a variable pitch angle and an outline of a truncated cone,

as shown in Fig. 10-14(a), tvhereas a volute spring is wound with a con¬
stant pitch angle and a variable axial pitch, resulting in a spring with the
outline of a truncated paraboloid, as shown in Fig. 10-14(6) and (c).
Both conical and volute springs have a constant spring rate until the
largest active coil “bottoms” or becomes inactive. Thereafter the num¬
ber of active coils decreases as the load is applied and, as a result, the
spring rate increases. Thus such springs become “ stiffer” as they are
loaded.1 This is a very desirable characteristic for some applications.
The compactness of conical and volute springs and their variable spring
rate are desirable in some applications; however, an accompanying dis¬
advantage is an inefficient distribution of stress in the spring. The dis¬
tribution of stress is complicated in that it varies both along the wire
and over the cross section of the wire. The stress along the wire is a maxi¬
mum at the largest active coil and it decreases toward the smallest coil.
However, as the load on the spring increases, the larger coils “bottom”
and become inactive, and hence the remaining active coils support the
1 Bernhard Sterne, Characteristics of the Volute Spring, SAE Journal, vol. 50,
p. 221, 1942.
162 MACHINE DESIGN
increasing loads that induce increasing stresses. The maximum stress

that is possible occurs when the smallest coil is the only active one, and
this condition should be the basis for design if the spring is likely to be
compressed solid. The Wahl stress factor for the coil of smallest index,
i.e., the smallest coil, should be used.
The influence of the stress factor becomes considerable in springs wound
of rectangular wire “on edge,” as in Fig. 10-14(c). In some volute
springs, the wire is tapered toward the small-coil end, giving a somewhat
more favorable stress distribution.
Fig. 10-15. Torsion springs.
- H
, .J1
—- k
-r1
(b)-Plate spring
Fig. 10-16. Springs having high rate.
To determine the deflection of a conical or volute spring, the equations

for the deflection of a helical spring may be used, in which the mean coil
diameter is used as the average of the mean diameters of the active coils.
Torsion springs are of either helical or spiral form, as shown in Fig.
10-15. These springs may be used in flexible drives, for cushions and
return springs in starters, and in various electrical devices and mech¬
anisms, and for spring-closed covers.
Ring springs and plate springs as shown in Fig. 10-16 are used where
high spring rates are desirable, as in flexible support for heavy equip¬
ment, in large flexible couplings, and in large valve springs. In Fig.
10-16(a), a compressive load causes the external rings to stretch and the
SPRINGS 163
internal ones to be compressed. The slope of the contact areas is critical
in order to provide for proper ring deformation and allowable compressive
forces between the rings. The compressive forces
are limited by compressive stresses, and frictional
forces to permit self-release of the rings.
Belleville springs. A type of plate spring
known as the “disk,” or Belleville, spring has
found many uses in applications requiring high
spring rates and compact spring units.1 A nest
of these springs is shown in Fig. 10-17. The unit
may be held in alignment by a central bolt or
tube.
Fig. 10-17. Nest of disk
The Neg’ator spring. This spring is called the
springs.
neg’ator spring because it can be designed to
produce a force-deflection characteristic with a negative slope, as
shown in Fig. 10-18(a); also it can
be designed for a zero slope (con¬
stant force) or for a positive slope.
The spring is composed of a flat
steel band which is prestressed so
that it has a natural curvature and
(ct) Force-deflection characteristics of which is wound on a flat-faced
Negator springs
reel. A force must be exerted on
the end of the spring to unwind it
from the reel and in turn the spring
tends to wind itself back on the
reel. Depending on the degree of
prestressing, the force on the spring
Storage bushing. may be constant or it may increase
or decrease as the band is unwound
from the reel. Figure 10-18(6)
shows such a spring.
At (c) in the figure is an applica¬
tion of a constant-force spring used
Output bushing / in an aircraft as a counterbalance
Cable drum
to provide for ease in lowering and
(c) Application of Negator spring raising an instrument from the
Courtesy of Hunter Spring Co. and Republic Aviation Corp.
fuselage for inspection so that it
Fig. 10-18. Neg’ator springs.
will remain in place wherever it is
stopped. The neg’ator spring wound on the storage bushing (or reel)
provides a constant tangential force on the output bushing. The output
1 See Almen and Laszlo, Disk Springs Facilitate Compactness, Machine Design,
June, 1936.
164 MACHINE DESIGN
bushing is geared to the cable drum around which are wound cables A,
B, and C which run over pulleys and provide three-point suspension
for the instrument. In this application, the instrument weighs 36 lb and
may be lowered a distance of 18 in. from the latched position. In this
case due to space and weight limitations dead-weight counterbalances
are not possible.
Neg’ator springs have a large variety of applications in the industrial
field as well as in mechanisms of all types, such as in business machines,
office, household, and building equipment and in servomechanisms. The
spring units are light in weight, occupy small space, and have versatile
characteristics.1
Torsion-bar springs. The torsion-bar spring has a range of applications
from precision instruments to vehicle suspensions. The bars are usually
solid of circular cross section, although hollow tubes and rectangular bars
are used. The torsion bar is usually fixed at one end, with a radial
bearing and crank arm at the other end. The suspended load at the end
of the crank twists the bar. The “spring rate” is usually expressed as
torsional moment in inch-pounds per degree of twist (T/d). The usual
equations for torsional stress and angular deflection may be used in the
design of torsion-bar springs.
The stresses are limited by “settling” in service or by fatigue failure.
Proper selection of material and heat-treatment2 is effective in securing
suitable endurance properties and for limiting settling. Shot peening is
effective in raising the endurance limit but will not reduce settling.
For fastening the ends of the bar, splines are suitable for heavy service
but the rod should be upset at the ends so that the outside diameter of
the spline may be 20 to 30 per cent greater than the diameter of the bar.
The length of the spline is usually about 0.4 times the diameter of the
spline. To avoid stress concentration, the transition from the body of
the bar to the spline should be gradual.
In many applications the use of torsion bars makes possible a saving in
weight and space as well as desirable load-deflection characteristics.
10-9 Nonmetallic springs. Air springs secure resilience by means of
air confined in a cylinder with the load supported by a plunger. The
position of the plunger and the spring rate can be altered by controlling
the air pressure. One application is in passenger busses so arranged that
as the bus becomes loaded with passengers, the air pressure is externally
1 See W. J. Cook and P. C. Clarke, The Neg’ator Spring—A Basic New Elastic
Member, Product Eng., July, 1944; F. A. Votta, The Theory and Design of Long-
deflection, Constant-force Spring Elements, Trans. ASME, vol. 74, p. 439, 1952.
Also see literature from Hunter Spring Co., Lansdale, Pa.
2 See “SAE Manual of Design and Manufacture of Torsion Bar Springs,” SP-26,
1947.
SPRINGS 165
and automatically increased in order to maintain the normal level of the

body.
Rubber springs. A type of flexible support that has advantages in
some applications employs rubber or a material with similar properties
Deflection,in.
Fig. 10-19. Load-deflection characteristics of a rubber mounting.
to provide suitable load-deflection characteristics. The usual support of

this type employs rubber in compression and/or shear.
Rubber in shear instead of compression is particularly adaptable to
large deformations per unit load. The comparison is shown by Fig. 10-19.
One type of a rubber mounting uses a
series of shear units which provide a spring
rate comparable with that of a steel coil
spring. One of these mountings designed for
supporting an axial compressive load is shown
in Fig. 10-20 in which the rubber, bonded to
the sleeves, is in shear and compression.1
As contrasted with metallic springs which
have practically no damping, the nonmetallic
springs have advantages in their inherent
damping properties which are of considerable Fig 10_2o Rubber spring
value in vibration loading. The rubber
should be protected from oil and high temperature. If it is not prac¬
ticable to keep oil from the rubber, a material such as neoprene should be
used.
1 F. L. Haushalter, Rubber as a Load-carrying Material, SAE Journal, January,
1939; J. F. Downie Smith, Rubber Springs—Shear Loading, J. Appl. Mechanics,
June, 1939.
166 MACHINE DESIGN
10-10 Additional considerations.

Protective coatings. For springs
subjected to repeated loading in corrosive environment, including inter¬
mittent applications of water, the problem of corrosion fatigue appears.
The use of nonmetallic protective coatings is in general of doubtful
advantage, and even metallic coatings may yield disappointing results.
Cadmium plating gives comparatively good results, as shown by the fol¬
lowing fatigue limits of a specimen of spring steel:1
Uncoated, in air. 70,000 psi

0.0004 in. cadmium plate, in water. 42,000 psi
0.0002 in. cadmium plate, in water. 35,000 psi
Uncoated, in water. 18,000 psi
Interleaves. To prevent wear in leaf springs due to interleaf friction, oil

or grease may be used as a lubricant. The lubricant may be squeezed out
from between the leaves, leaving the surfaces of the plates dry except for
an adsorbed film. This film may be broken through at regions of exces¬
sive pressure, and abrasive or galling wear may result. One attempted
cure for this wearing is the introduction of thin wood interleaves between
the spring leaves to absorb lubricant, to prevent leaf contact, and to
reduce wear.
Shot peening. An interesting development in recent years has been the
process of shot peening. In this process, coil or leaf springs are subjected
to small steel shot blasted from an air nozzle or impelled by mechanical
means. The shot impinges on the surface of the springs, which thereby
undergoes a surface cold working.2 The cold working raises the endur¬
ance limit of the spring material at the surface; also the peening
introduces surface compression, which has the effect of preventing the
spread of a fatigue crack and this contributes to raising the endurance
limit.
It has been found that the endurance range of small helical springs
made of carbon and chrome-vanadium steels may be increased from
20,000 to 70,000 psi up to 20,000 to 115,000 psi. Comparative increases
are realized in other steels used for either coil or leaf springs. It is of
interest to note that the size of shot and the time of blasting have a bear¬
ing on the results. Also, it has been shown that shot peening loses its
effectiveness at high temperatures and that the increase in fatigue
strength is completely lost at a temperature of 825 F. Too, it has been
shown that a shot-peened spring is more susceptible to effects of corrosion,
1 D. J. Me Adam, Jr., Fatigue and Corrosion Fatigue of Spring Material, Trans.

ASME, vol. 51, p. APM-51-5, 1929.
2 F. P. Zimmerli, How Shot Blasting Increases Fatigue Life, Machine Design,
November, 1940, p. 62.
SPRINGS 167
and thus if shot-peened springs are to retain their high endurance ranges
they should be protected from corrosion.
Shot peening has also been employed on connecting rods of automotive
engines to raise their endurance limits and on link rods of aircraft engines
as a substitute for the expensive and time-consuming polishing process.
Indications are that shot peening may have beneficial results in many
other applications.
CHAPTER 11
PRESSURE CYLINDERS
11-1 Introduction. Cylinders have a variety of uses in machinery.

They may be classified variously, each group requiring particular equa¬
tions for their design and the class of material for their construction, as
follows:
Dimensions (thin- and thick-walled cylinders). The ratio t/d, which is
the ratio of wall thickness t to cylinder diameter d, is the criterion. Thin-
walled cylinders are used for pressure cylinders in small presses, tanks,
and low-pressure processing equipment. They are generally made of
welded or riveted steel or of seamless steel tubing. Thick-walled cyl¬
inders are used in high-pressure cylinders, tanks, gun barrels, etc.
End construction (open end or closed). A simple cylinder with a pis¬
ton, such as a cylinder of a press, is open ended and has circumferential, or
hoop, stresses induced by the fluid pressure. A tank has closed ends,
and has longitudinal stresses in addition to circumferential stresses. In
either open or closed cylinders, the restraint offered by the ends and the
relative rigidity of the parts of the cylinder often induce indeterminate
stresses at the juncture of the cylinder and the head which are higher
than those indicated by elementary equations. These stresses may be of
such magnitude in high-pressure installations that accurate calculations
based on the theory of elasticity are demanded. Also, stresses due to
welding or riveting may be introduced. Such fabrication stresses should
be considered when the factor of safety is selected, and reduced to a
minimum by appropriate design features, including annealing.
Material. The material may be brittle, such as cast iron, or ductile,
such as steel. The type of material may dictate the method of failure
and thereby the appropriate equation for its strength.
Service (conditions of temperature, pressure, and environment). Low-
temperature or high-temperature operation may influence the choice of
material and the value of the allowable stress, and it may also influence
the basis for design, i.e., on elastic state for normal temperatures, or on
creep if the temperature for steel cylinders is above 650 F. In addition,
thermal stresses may be induced owing to heating or cooling of the cyl-
168
PRESSURE CYLINDERS 169
inder,1 and in some cases caustic embrittlement is present (see Art. 3-8).
High-pressure cylinders may require special considerations because of
weight limitations or safety requirements.2
Whether the pressure is applied internally or externally will also affect
the design. In some pressure vessels which normally have internal pres¬
sure, failure because of external pressure should be guarded against in the
event of a partial vacuum being formed inside the cylinder, for example,
because of condensation in a steam drum or evaporator.
The case of external pressure being applied on the side or ends of cyl¬
inders requires design on the basis of buckling.3
Fig. 11-1. Cylinder with internal pressure.
In chemical apparatus or refining equipment, stainless steel or a cor¬

rosion-resistant lining may be required.
The designs of some of the more common types of cylinders subjected
to internal pressure are discussed in the following articles.
11-2 Thin-walled cylinders. The analysis of stresses induced in a
thin-walled cylinder as shown in Fig. 11-1 will be made by neglecting the
effects of curvature of the cylinder wall and by assuming that the tensile
stresses are uniformly distributed over the section of the walls. This
assumption is valid for thin cylinders whose ratio t/d is equal to or less
than 0.1. For values of t/d greater than this value, the variation from
uniform stress distribution becomes appreciable and thick-cylinder design
must be used.
Circumferential tensile stress. The fluid force acting on a longitudinal
section of unit length is equal to pd, and for equilibrium of forces may be
1 S. Timoshenko and J. N. Goodier, “Theory of Elasticity,” 3d ed., McGraw-Hill

Book Company, Inc., New York, 1951; C. H. Kent, Thermal Stresses in Thin-walled
Cylinders, Trans. ASME, vol. 53, p. APM-53-13, 1931.
2 ASME Boiler Construction Code, sec. VIII, Unfired Pressure Vessels.
3 S. Timoshenko, “Theory of Elastic Stability,” art. 83, McGraw-Hill Book Com¬
pany, Inc., New York, 1936; Saunders and Trilling, Collapse by Instability of Thin
Cylindrical Shells under External Pressure, Trans. ASME, vol. 56, p. 819, 1934; R. G.
Sturm, A Study of the Collapsing Pressure of Thin-walled Cylinders, Univ. Illinois
Bull. 329, 1941; ASME Boiler Construction Code, loc. cit.
170 MACHINE DESIGN
equated to the resisting force equal to 2tei, where S] represents the cir¬
cumferential, or hoop, stress. Thus
(11-D
Longitudinal tensile stress. The fluid force acting on a ring section is

equal to \^jrd'2p and for equilibrium of forces may be equated to the
resisting force ttdts2, where s2 represents the longitudinal stress. Thus
(11-2)
In Eqs. (11-1) and (11-2),

Si = circumferential stress, psi
s2 = longitudinal stress, psi
p = internal pressure, psi
d = internal diameter, in.
t = wall thickness, in.
11-3 Thick-walled cylinders. The distribution of stress in thick-
walled cylinders cannot be as¬
sumed to be uniformly distributed
as in the case of thin-walled cylin¬
ders. Analysis indicates that the
'4b tangential stress in a thick-walled
(a) (b) cylinder is considerably greater at
the inside surface than at the out¬
(a) - Tangential
stress side and that the radial stress is of
(b) -Radial sufficient magnitude to be signifi¬
stress
cant. The distribution of stress is
Fig. 11-2. Stress distribution in wall of
indicated in Fig. 11-2. Equations
thick cylinder.
have been developed for cylinder
material and conditions as given below. The notation for these equa¬
tions is as follows:
Let d0 = outer diameter of cylinder, in.
di = inner diameter of cylinder, in.
p = internal pressure, psi
t = wall thickness = (dQ — di)/2
v = Poisson’s ratio
= tangential stress, psi
sr = radial stress, psi
s'8 = maximum shear stress, psi
s't = allowable tensile stress (limiting value of s't), psi
ss = allowable shear stress (limiting value of s'), psi
Lame’s equations for cylinders with internal pressure. Using the assump¬
tion that plane transverse sections of a cylinder remain plane during load-
PRESSURE CYLINDERS 171
ing, i.e., that all longitudinal fibers are equally strained, Lame has shown
that the maximum tangential stress at the inside fiber of the cylinder is
y{df + df)
sj(max) =
dl - df
and the maximum radial stress at the same point is
Brittle material. Since the tangential stress is a principal stress and is

a maximum for the biaxial system, in designing a cylinder of brittle mate¬
rial and in accordance with the maximum-stress theory of failure, this
stress should not be exceeded, or
p{df + df)
st = sj(max)
dl ~ df
If t = {d0 — df) /2, the thickness of a cylinder made of brittle material
di ( lst + V
t = (11-3)
2 W st - p
Equations for ductile materials based on maximum-strain theory. The
stresses in a thick cylinder may be determined by the use of Lame’s
equations; however, in some cases, for example, in open-end cylinders,
such as gun barrels, or in shrink fits, the allowable stresses cannot be
determined by means of the maximum-stress theory of failure. For this
case, the maximum-strain theory shows closer coincidence with experi¬
mental results. According to this theory which states that failure occurs
when the strain reaches a limiting value, the strain is expressed in terms
of a limiting stress st, and Birnie’s equation for this case, when solved for
the thickness, becomes
di f lst + (1 - v)p _ \
(11-4)
2 W* - (1 + v)p /
Clavarino’s equation based on the same theory of failure applies to

closed-end cylinders as follows:
( Ist + (1 - 2v)p
(11-5)
^ St — (1 + v)p
The following outline is given for reference to select the equation to be

used in the design of thick cylinders:
Brittle material, open ends—use Eq. (11-3)
Brittle material, closed ends—use Eq. (11-3)
Ductile material, open ends—use Eq. (11-4)
Ductile material, closed ends—use Eq. (11-5)
172 MACHINE DESIGN
11-4 Cylinder heads. Bending of the shell due to the restraint of

the heads is shown exaggerated in Fig. 11-3. It is evident that the
stresses in the plate induced by fluid pressure, as determined by the cyl¬
inder equations in Arts. 11-2 and
11-3, are increased because of the zzzzzzzz:
•a-
m
bending of the plate, and this bend-
ing becomes greater in proportion ^SoKdTcast (b)- Flanged (c)-Flanged
through bolt stud bolt
to the relative stiffness of the head
and shell. The bending stresses
may be allowed for in low-pressure
cylinder designs by the use of an
faO-Riveted- (e)- Riveted
appropriate allowable stress; how¬ convex concave
ever, in high-pressure cylinder de¬
signs the bending stresses should
be reduced to a minimum by the
use of special head construction, as (fj-Seamless tube- fgd-Seamless tube-
shown in Fig. 11-4(A). through bolts screwed head
Openings for manholes, hand¬

holes, nozzle attachments, and
other connections markedly affect
(JzJ-Heoid joint machined (£)- Flat head-

for welding welded
(j)-Welded
Fig. 11-3. Buckling of shell due Fig. 11-4. Types of cylinder heads.
to end restraint.
the stresses in pressure vessels and should be carefully considered, espe¬

cially in high-pressure and high-temperature installations.1 Examples of
head constructions and means of attachment are shown in Fig. 11-4.
1 Maulbetsch and Hetenyi, Stresses in Pressure Vessels, Design Data, Trans. ASME,
vol. 58, p. A-107, 1936; Taylor and Waters, The Effect of Openings in Pressure Vessels,
Trans. ASME, vol. 56, p. APM-56-3, 1934; “Welding Handbook,” American Welding
Society, pp. 1306 ff., 1942.
CHAPTER 12
TRANSLATION SCREWS
12-1 Introduction. Translation screws are used to move machine

parts against resisting forces, for instance, in a screw-operated tensile-
testing machine, jack, press, or lead screw of a lathe. The usual source
of power is a rotating shaft or crank. The design of a translation screw
requires considerations of strength of the screw in tension or compression,
shearing of the screw itself or of the threads, wearing of the threads, and
p |
r
h
t
(tz)-Sc{uare thread (6)-Acme thread
Fig. 12-1. Forms of translation screw threads.
power requirements. The latter consideration involves frictional resist¬

ance, which may account for a large proportion of the power supplied to
the screw.
In most translation screws, the nut has axial motion against the resist¬
ing axial force while the screw rotates in its bearings. In some screws,
however, the screw rotates and moves axially against the resisting force
while the nut is stationary, and in others the nut rotates while the screw
moves axially with no rotation.
12-2 Forms of threads. Square threads. This thread has the propor¬
tions shown in Fig. 12-1(a). The square thread is more efficient than
angular types, such as V or acme threads; but it is more difficult to cut,
173
174 MACHINE DESIGN
and it is not so adaptable to a split nut, and also it has no means of adjust¬
ment for wear. A common proportion of square-threaded screws is the
Sellers square thread, as given in Table 12-1.
Table 12-1. Sellers’ Square Threads
Diameter, in. Threads per in. Root area, sq in.
X 10 0.0207
Ke 9 0.0375
% 8 0.0555
Ke 7 0.0767
hi 6K 0.1049
%6 6 0.1364
% 5K 0.1709
% 5 0.2063
K 5 0.2597
% 4K 0.3000
X 4X 0.3638
% 4 0.4058
1 4 0.4804
ix 3X 0.6013
IK 3X
3
0.7854
1% 0.9201
IK 3 1.1462
IK m 1.3414
ik 2X 1.5394
IK 2X 1.8265
2 2K 2.0422
2K 2K 2.7245
2K 2 3.3410
2K 2 4.2000
3 IK 4.9087
3M IK 5.9396
3K IK 6.8930
3X iX 7.8853
4 IK 8.8434
Acme thread. This thread, which is a form of the V or angular thread,

has the proportions shown in Fig. 12-1(6) and is used where a split nut
is required and where provision must be made to take up wear as in the
lead screw of a lathe. Wear may be taken up by means of an adjustable
split nut. Proportions of the acme thread are given in Table 12-2. The
efficiency of the acme thread is slightly less than that of a square thread.
Buttress thread. This thread has the form shown in Fig. 12-1 (c) and
is used for translation under load in one direction only. This thread com¬
bines the higher efficiency of the square thread and the ease of cutting
and the adaptability to a split nut of the acme thread; furthermore it is
stronger than other forms because of the greater thickness at the base of
TRANSLATION SCREWS 175
the thread. While this thread has limited use for power transmission,1
it is used in some cases as a screw fastening.
Table 12-2. Acme Screw Threads
Threads Depth of Thickness

per in. thread, in. at root, in.
1 0.5100 0.6345
1M 0.3850 0.4772
2 0.2600 0.3199
3 0.1767 0.2150
4 0.1350 0.1625
5 0.1100 0.1311
6 0.0933 0.1101
7 0.0814 0.0951
8 0.0725 0.0839
9 0.0655 0.0751
10 0.0600 0.0681
Multiple threads. Translation screw with multiple threads, such as

double, triple, etc., are employed when it is desired to secure a large lead
with fine threads or high efficiency. An applica¬
tion is found in high-speed actuators.
12-3 Efficiency of screws. Square threads.
Consider that the nut of a screw is moved against
an axial load by the rotation of the screw. The
load on the nut will be transferred to the screw as a
distributed load on the surface of the threads in
contact. For the purpose of analysis, this distrib¬
uted load may be assumed to be concentrated at
a point on the mean circumference of the thread,
as at o in Fig. 12-2. The following notation will
be used:
Let Q = axial load, lb
d = diameter of mean helix, in. Fig. 12-2. Forces on
a = lead angle square-threaded
</> = friction angle screws.
2(3 = included thread angle [see Fig. 12-1(6)]
/ = coefficient of thread friction = tan 0
L = lead of threads, in.
T = torque required to overcome thread friction and to move the
load, lb-in.
T0 = torque required to move the load neglecting friction, lb-in.
e = efficiency of screw
1 G. H. Atwood, A New Type Screw-luffing Crane for Shipbuilding, Mech. Eng.,
September, 1944, p. 569.
176 MACHINE DESIGN
Under static conditions, the direction of the load on the thread will be
normal to the thread surface, as along ao in Fig. 12-2. When the screw
rotates so that the nut is moved against its external load Q, the line of
action of ao will be rotated through the angle of friction <j> to bo, as shown
in the figure.
For equilibrium of forces, the component of bo parallel to the axis of
the screw is
Q = bo cos (a + </>)
The component of bo at right angles to the axis of the screw is
F = bo sin (a + </>)
Hence
F = Q tan (a + </>)
and
T = F~ ^ tan (a + 0) (12-1)
_ Qd hrfd + L\
(12-la)
2 \Td - fLJ
To obtain an expression for the efficiency of the screw, let T0 represent

the torque required to raise th3 load, assuming that friction is not present.
Thus from Eq. (12-1)
Qd
1o = tan a
and the efficiency

T0 tan a QL
(12-2)
T tan (a + </>) 2tT
Angular, or V, thread. Since the normal loads on the thread of an

angular, or V, type is oblique to the axial load, the normal load will be
greater than that for a square thread. Hence, the frictional force will be
correspondingly greater, and the friction terms in Eqs. (12-1) and (12-2)
should be divided by cos /3, or
Qd /irfd sec j8 + L\
(12-3)
2 \7rd — JL sec )
and
tan qj[1 — (/ sec jS tan a)]
(12-4)
tan a + (/ sec (3)
Reverse motion. If the rotation of the screw moves the nut in the same
direction as the load, as in the case of a screw jack in lowering the load,
the vector bo in Fig. 12-2 will make an angle with the axis of the screw
that is equal to a — 0. The equations for the torque and efficiency as
derived for direct motion will be altered. The torque required to lower a
load for a square-threaded screw becomes
Qd Qd Md - L\
T tan (a — 0)
2 2 \ird + fL/
It may be noted from the above equation that if 0 < a, the torque T
required to lower the load is negative, i.e., an effort must be applied to
the screw to resist the tendency of the load to descend.
If, however, 0 > a, the torque T will be positive, indicating that an

effort must be applied to lower the load. Such a screw is known as a
“self-locking” screw.
Efficiency versus lead angle. It is evident that the efficiency of a square-
threaded screw depends on the lead angle a and the friction angle 0.
Figure 12-3 shows the variation of the efficiency of a square-threaded
screw for direct motion (raising the load) with the mean lead angle a.
It may be noted that the efficiency reaches a maximum value for a:
that is equal to 45 deg approximately.
A qualitative explanation of the increase in efficiency to a maximum
value followed by a decrease as a increases may be made by considering
that the work of thread friction is equal to the frictional force along the
thread multiplied by the distance through which this force acts. For a
low angle, the slope of the thread is small, and the distance through which
the friction force acts is great compared with the lead of the thread.
178 MACHINE DESIGN
Hence, the work of thread friction will be great as compared with the
useful work, and a low efficiency will result.
For a large lead angle, say 70 deg, the normal thread
force becomes large and likewise the force of friction
and the work of friction become large as compared with
the useful work. Between the two extremes that have
been discussed there is a value of lead angle that yields a
minimum value of work of friction as compared with the
useful work, and hence a maximum efficiency also.
It should be remembered, however, that if the screw
is required to be self-locking, a low lead angle is neces¬
sary in order to introduce intentionally a frictional force
that is sufficient to prevent reverse motion.
12-4 Collar friction. The axial force produced on a
screw by the nut must be resisted by a collar or its
equivalent. This action and reaction is illustrated in
Fig. 12-4.
To determine the total torque required to turn the
screw, the frictional torque of the collar must be added
Fig. 12-4. Axial
to the screw-thread torque, as given by Eq. (12-1) or
forces on a
screw. Eq. (12-3).
The frictional torque of the collar may be determined
by assuming that the frictional force on the collar acts at its mean
diameter. Therefore
fQdc
Tc = (12-5)
where Tc = torque of collar friction, in.-lb

Q = axial load on the screw, lb
/ = coefficient of collar friction
dc = mean diameter of the collar, in.
The efficiency of a screw including both thread friction and collar fric¬
tion may be obtained by dividing the torque required to move the load,
neglecting friction, by the torque required to move the load including
thread and collar friction.
12-5 Coefficient of friction. Values for the coefficient of friction for
the threads of translation screws have been determined by investigations1
and have been found to depend on the quality of materials, workmanship
in cutting the threads, degree of “running-in” of the threads, and lubrica¬
tion. The value of the coefficient of friction does not vary markedly with
different combinations of material, load, or rubbing speed, except under
starting conditions.
1 Ham and Ryan, An Experimental Investigation of Thread Friction of Screw
Threads, Univ. Illinois Bull. 247, 1932.
1. For high-grade materials and workmanship and for well run-in and
lubricated threads, the coefficient of friction may be taken as 0.10.
2. For average quality material, workmanship, and conditions of opera¬
tion, the coefficient of friction may be taken as 0.125.
3. For poor-quality material and workmanship and for newly machined
surfaces which are indifferently lubricated and which have slow motion,
the coefficient of friction may be taken as 0.15.
4. The coefficient of friction for starting conditions may be taken as
ljî times the value for running conditions.
5. The coefficient of collar friction may be taken as the same as for
thread friction.
Example 12-1. A 10-ton screw jack with a maximum extension of 4 in., is to have
double square threads, (a) Using an allowable compressive stress of 5,000 psi,
determine the size of screw, size of collar, and
length of nut required. (The allowable compres¬
sive stress in the screw is a low value in order to
give a screw which will accommodate a well-pro¬
portioned thread.) (6) Determine the torque
required to raise the load and also the efficiency of
the jack.
Solution:
(a) Dimensions of screw, collar, and nut: The

root area of the screw equals
Q = 20,000
= 4 in.2
sc 5,000
From Table 12-1, a 2%-in. screw with 2 threads

per inch is satisfactory. The pitch of the threads
is % in. and the lead (double thread) is 1 in.
The outside diameter of the collar is usually
larger than the outside diameter of the screw to
give a reasonable bearing pressure on the collar.
Assuming an outer diameter for the collar equal to
3^ in. with a pilot pin 1 in. in diameter, the collar
bearing pressure is Screw and collar for jack.
_
4Q 4 X 20,000
Vc = = 2,270 psi
7r (dl — d\) 7t(3.52 — l2)
This value is in the safe range as compared with values in Table 12-3 for screw
threads. If trouble due to galling is expected, a thin bronze washer may be interposed
between the rubbing parts of the collar.
The length of the nut may be determined from the number of threads required to
limit the bearing pressure on the threads to a permissible value.
The area of one thread is equal to
7rdh =7r X 2.5 X 0.25 = 1.97 in.2
where h is the depth of the thread.

180 MACHINE DESIGN
Using an allowable thread bearing pressure p = 2,500 psi, the number of threads
required is
20,000 _ . AA ,
2,500 <X 1.97 4-06threads
Since there are 2 threads per inch, the length of the nut in inches is one-half of 4.06.
On the basis of bearing pressure, a 2-in. nut would be satisfactory, but for stability of
the screw it is practice to use its length at least equal to the diameter of the thread.
Three inches would be a reasonable length for the nut.
(6) Torque to raise the load and the efficiency of the jack: The mean diameter of the
thread may be found from the figure for the outside diameter of 2% in. and pitch of
y2 in. Considering the thread as square in section, d — 2% — 34 = 2^j in.
The torque to raise the load and to overcome thread friction, using a coefficient of
friction / = 0.125 from Art. 12-5, is
Qd /irfd + L\ _ 20,000 X 2.5 At X 0.125 X 2.5 +

400 in.-lb
T 2 \ird - fL) 2 \tt X 2.5 - 0.125 X l) 6’
The torque for the collar is
m fQdc 0.125 X 20,000 X 2.25 0 . 1U

Tc = - = -0- = 2,810 in.-lb
The total torque is

6,400 + 2,810 = 9,210 in.-lb
The efficiency of the screw and collar is
QL 20,000 X 1
- = 0.35, or 35 per cent
2ttT 2x X 9,210 P
12-6 Stresses in screws. Tensile or compressive stresses in the body of

a screw due to the axial load may be determined by dividing the axial
load by the minimum cross-sectional area of the screw, unless the screw
is long and slender and subjected to a compressive load in which case
column formulas must be used.
The torsional stress in the screw owing to the torsional moment required
to rotate the screw may be determined by considering the minimum cross
section of the screw.
Shearing stress and bearing pressure are generally determined by assum¬
ing that the loads are uniformly distributed on the threads in contact.
Let Q = axial load on screw lb (see Fig. 12-5)
d0 = major diameter of threads, in.
dr = minor diameter of threads, in.
t = width of thread, in.
n = number of threads in engagement
ss = shearing stress, psi
Sb = bearing pressure, psi
The shearing stresses for the threads of the screw and of the nut are,
respectively, equal to
Q
ss (screw) =
mrdrt
Q
ss (nut) =
mrdot
The bearing pressure on the threads is equal to
4Q
Sb
mr(dl — d2r)
The bearing pressure is limited by lubrication conditions. In Table

12-3 are given some limiting values of bearing pressures.
Fig. 12-5. Screw-thread dimensions.
12-7 Other types of screws. Differential screws. When slow advance

or fine adjustment is necessary and it is not advisable to use fine threads,
a differential screw may be used. As shown in Fig. 12-6, this screw is
composed of two component screws, one screw a is keyed to the hand-
wheel and has threads of pitch px. The other screw b is threaded into the
first one but is prevented from rotating by the feather key c. The
threads on this screw have a pitch p2, which is smaller than px.
If the handwheel is turned right-handed one turn, the screw a will
move downward in the frame a distance pi. As a rotates, the screw b
will be screwed upward into a and will move relative to a a distance p2.
182 MACHINE DESIGN
Thus the total movement downward of b relative to the frame will be

Pi — P2. It is apparent that if the two pitches p2 and pi are equal, the
screw b will not move; and if p2 is made less than px, as discussed above,
Table 12-3. Safe Bearing Pressures for Screws
Material
Safe bearing
Type Rubbing speed
pressure, psi
Screw Nut
Hand press. Steel Bronze 2,500-3,500 Low speed, well lubricated

Jackscrew. Steel Cast iron 1,800-2,500 Low speed < 8 fpm
Jackscrew. Steel Bronze 1,600-2,500 Low speed <10 fpm
Hoisting screw. Steel Cast iron 600-1,000 Medium speed 20-40 fpm
Hoisting screw. Steel Bronze 800-1,400 Medium speed 20-40 fpm
Lead screw. Steel Bronze 150-240 High speed > 50 fpm
the motion of b will equal their difference, which can be made as small
as required.
If the hand of one of the threads is reversed, the motion of b will be
equal to p2 + p2. This type of screw is known as a ‘‘compound screw.”
Fig. 12-7. Ball-bearing screw and nut. (Courtesy of Saginaw Steering Gear Division,
General Motors Corporation.)
Ball-bearing screw and nut. A type of screw has been developed

recently which interposes steel balls between properly formed threads on
the screw and nut, and thus rolling friction is substituted for sliding
friction.1 In Fig. 12-7 is shown a ball-bearing screw and nut.
1 R. K. Allan, “Rolling Bearings,” p. 20, Sir Isaac Pitman & Sons, Ltd., London,
1945; Ball Bearing Screw and Nut, Saginaw Steering Gear Division, General Motors
Corp., Saginaw, Mich.
The threads in both screw and nut are approximately semicircular in

section. When the nut is translated as the screw rotates, the balls run
out of the threads on the nut and are returned by a recirculating groove
to the other end of the nut.
Efficiencies as high as 90 per cent are claimed for this screw with cor¬
responding increase in life over conventional screws. Applications to
date include steering gear for automotive vehicles1 and power actuators.
Francis W. Davis, Power Steering for Automotive Vehicles, SAE Journal, vol. 53,
no. 4, p. 239, April, 1945.
CHAPTER 13
SHAFTING
13-1 General considerations. The design of a shaft may require the

interrelated considerations of a number of factors, such as the following:
material and heat-treatment; strength for power and loading require¬
ments; stiffness as affecting, for instance, bearing performance, gear
operation, timing and critical speeds; weight and space limitations; and
stress concentration. Some general design methods that involve these
considerations are discussed in the following articles.
13-2 Terminology. In machinery, the general term “ shaft ” refers to a
member, usually of circular cross section, which supports gears, sprockets,
wheels, rotors, etc., and which is subjected to torsion and to transverse or
axial loads, acting singly or in combination.
Fig. 13-1. Stepped shaft for induction motor.
Specifically a shaft is a member which rotates and which is subjected to

torsion accompanied by transverse or axial loads; an axle, with few
exceptions, is a member which is subjected to transverse loads but not
torsion; a spindle is a short shaft.
Most shafts are not of uniform diameter, but are stepped to provide
shoulders for locating gears, pulleys, bearings, or other attached or con¬
tacting parts. In Fig. 13-1 is shown a stepped shaft for a motor with
shoulders for locating the rotor, ball bearings, gear, etc.
13-3 Materials. The most common material for shafting is mild steel.
For high-strength requirements, an alloy steel, such as nickel, nickel-
chromium, or chrome-vanadium steel, is used.
Commercial shafting is made of low-carbon steel similar to SAE 1015
formed by hot rolling and finished to size by cold drawing or by turning
and grinding. Cold drawing produces a stronger shaft than hot rolling,
but with higher residual stresses. The residual stresses may cause distor¬
tion of the shaft when it is machined, especially when slots or keyways
are cut. Shafts for special purposes may be forged.
184
SHAFTING 185
13-4 Commercial sizes of shafting. Formerly shafting was made of

bars rolled to nominal dimensions of even quarter-inch intervals and then
finished 34 6 in. under the nominal size. Sleeve bearings are available to
accommodate shafts }{q in. under the nominal size. Shafting may be
obtained in the following sizes. Dimensions are in inches.
Table 13-1. Transmission Shafting
Diameter Increment
^6-2^6 H
2^6-5% y2
Table 13-2. Machinery Shafting
Diameter Increment
H-2H He
2^-4 H
4-6 H
13-5 Design considerations. The size of a shaft for a particular appli¬
cation may be determined on the basis of strength, or both strength and
rigidity.
In designing a shaft on the basis of strength, it is necessary to consider
the following: type of loading, i.e., static, shock, or cyclic; the weakening
effects at points of stress concentration due to key ways and shoulders;
the combination of loading, such as bending and torsion. The size of the
shaft must be sufficient to prevent the induced stresses from exceeding
the allowable stress for the material.
In many cases the rigidity of the shaft is an important design feature.
The twisting of the shaft may be limited in order to provide accurate pre¬
scribed timing or motions, as in the camshaft of an internal-combustion
engine. Transverse deflections may be limited, for instance, to maintain
proper bearing clearances or gear-tooth alignment. Both torsional and
transverse rigidity are important in vibration.
13-6 Determination of shaft sizes on the basis of strength. The action
of the loads on a shaft is generally one of the following: (a) torsion, (6)
bending, (c) torsion combined with bending, and (d) torsion combined
with axial tension or compression.
Torsion. For a simple twisting moment Ton a solid circular shaft of
diameter d, the maximum shear stress is
16 T
(13-1)
ird3
By assuming that failure is based on the maximum-shear theory, the
maximum shear stress must not exceed the allowable shear stress, ssa.
Then
rrd3_T
(13-2)
16 Ssa
186 MACHINE DESIGN
Bending. For a simple bending moment M on a solid circular shaft of

diameter d, the maximum bending stress is
32M
st = (13-3)
7T d3
By substituting the allowable tensile stress sa for st

7T d3 M
(13-4)
32 sa
(See Appendix IV for values of section moduli for standard-size shafts.)

Torsion combined with bending. Most rotating shafts carry gears, pul¬
leys, sprockets, or sheaves that cause bending of the shaft in addition to
torsion. An effort should be made to mount the gears, pulleys, etc., as
near to the bearings as possible in order to reduce the bending moment.
The design of shafting made of ductile material is based on the maxi¬
mum shear-stress theory; hence it is necessary to determine the maximum
combined shear stress in the shaft due to the applied twisting and bending
moments. By using the combined stress equation,
ss (max) =
and substituting the values of ss and st from Eqs. (13-1) and (13-3) the
maximum shearing stress becomes
ss (max) = —— vV + M2
7T a3
By limiting the maximum shear stress to the allowable shear stress for
the material and solving for the section modulus,
ttcT _ Vt2 + M2
(13-5)
16 SSa
The term \/T2 + M2 is called the equivalent twisting moment and is

defined as the fictitious twisting moment that will induce the same maxi¬
mum shearing stress in the shaft as the combined action of the actual
twisting moment and the actual bending moment.
Torsion combined with axial tension or compression. Some shafts, such
as propeller shafts, are subjected to torsion combined with direct axial
loads. The shear stress due to torsion for a solid circular shaft is given
by Eq. (13-1). The direct stress due to an axial load P is
SHAFTING 187
These stresses may be combined by the use of the combined stress

equation as
ss (max) (13-6)
7T d3
By substituting the allowable shear stress for ss (max), the above equa¬
tion may be solved by trial. If the axial load is a compressive load, it
should not exceed the critical buckling load.
For shafts on which the torque and bending moment are cyclic, stress
concentration should be considered. If the fatigue-notch factor for
shear stress (KtSf) is applied to Eq. (13-1), and the combined factor for
normal stress (Kt'f) is applied to Eq. (13-3) and both substituted in
Eq. (13-5), it becomes
7T dZ V (KufTY + (K,fMy
(13-5a)
T6~ Ssa
In the use of Eq. (13-5a), it should be realized that it is a design equa¬

tion for strength. In the design of most shafts in machinery, however,
stiffness of the shaft usually requires a larger shaft than strength.
Since most shafts are not uniform in diameter, it is usually impossible to
design them directly on the basis of stiffness. Hence, it is common prac¬
tice to design such shafts on the basis of strength, using allowable stresses
as determined from analyses of similar existing shafts which have per¬
formed satisfactorily in service. The allowable stresses so determined
usually are low in value and the shafts are “overdesigned” for strength
but have the required stiffness characteristics. Allowable stresses in
shear equal to 5,000 to 8,000 psi are common in the design of such shafts.
An uncertainty in using Eq. (13-5a) is that available values for stress-
concentration factors are determined separately for torsion and for bend¬
ing and it is not usually known whether the factors so determined apply
to combined stresses due to the simultaneous action of torsion and bend¬
ing. However, if stiffness is not an important requirement, Eq. (13-5a)
may be used with an allowable stress determined by dividing the endur¬
ance limit in shear by an appropriate value for the factor of safety.
In applications where stiffness is critical, it is usually satisfactory to use
Eq. (13-5) with low allowable stresses to determine the size of shaft.
After its dimensions are fixed, the deflection characteristics may be deter¬
mined by the use of the graphical method in Appendix XIII.
The maximum stresses in the shaft at critical sections may be inves¬
tigated by the use of the combined stress equation, Eq. (5-9), and if the
stresses are too high, the material or its heat-treatment may be modified
accordingly. A change in class of steel or its heat-treatment will not
188 MACHINE DESIGN
affect, however, the deflection characteristics of the shaft. This proce¬
dure is illustrated in Example 13-1.
13-7 Allowable stresses. The allowable stresses for the design of
shafting should be chosen with consideration for the factors discussed in
Chap. 5. Specifically, this bases the allowable stress on the yield point
if the loading is static and on the endurance limit if the loading is cyclic.
In many cases of combined loading, one of the loadings may be static
and the other repeated. An example of this is a rotating shaft on which
are mounted gears transmitting uniform torque in which the bending
stresses are completely reversed while the shearing stresses due to the
torque may be essentially static. The true allowable stress in this
instance should be based on an endurance limit determined from tests
simulating the exact combination of stresses that exist in the shaft.
Only rarely are such values available. To consider both of the loadings
as completely reversed would be on the side of safety. Particular effort
should be made to reduce stress concentrations to a minimum in shafting.
This is especially necessary in shafts subjected to vibration, either self-
induced or parasitic. Such vibration causes fatigue stresses which are
superimposed upon those due to power loads.
Rough surfaces should be eliminated. Press and shrink-fit collars and
hubs should be carefully considered and their effects minimized, as sug¬
gested in Chap. 4.
Example 13-1. The shaft of a speed reducer is mounted on ball bearings A and B,
as shown in the figure. The stub end of the shaft is fitted with a flexible coupling.
The gear is 5 in. in diameter and 5 hp are transmitted at 120 rpm. Assuming an
allowable stress ssa — 6,000 psi, determine the size of shaft required.
Solution: An inspection of the sketch shows that the bending moment on the shaft
is cyclic and has a maximum value at the mid-point of the gear. The torque on the
shaft between that section and the stub end is equal to
63,030 hp 63,030 X 5
T = = 2,620 lb-in.
rpm 120
SHAFTING 189
The tangential force on the gear is
2T _ 2 X 2,620
1,050 lb
D 5
The bending moment at the center of the gear may be found equal to 2,100 lb-in.
From Eq. (13-5),
ird3 VV + M2 yj2,6202 + 2,1002

0.559 in.3
I6~ Sga 6,000
From Appendix IV, the required shaft diameter is 134 in. As an extension to the
problem, assume that the diameter of the shaft at the right of the gear is 134 in-, and
determine the maximum stress at the fillet (r = 34 in.).
Referring to Eq. (5-9), the bending moment is completely reversed so that Mm = 0,
and the torsional moment is steady so that Tr = 0. Equation (5-9) then reduces to,
solved for the allowable stress,
f.s. 7rd3/16
Assuming that the shaft is made of SAE 1020 steel,
Suit = 62,000 psi (from Appendix V)

sv = 35,000 psi (from Appendix V)
Se = 25,000 psi (from Appendix VII)
Sys = 0.71 X 35,000 (from Table 7-2)
= 24,800 psi
Therefore
sy _ 35,000
se ~~ 25,000
From Fig. X-14, using D/d = 1.2, r/d = 0.10, Kt’ is found equal to 1.5.
By substituting the above values in the right-hand term of the equation above, the
maximum induced stress is found to be equal to 13,350 psi. The left-hand term of the
equation is the allowable stress and for an assumed value of the factor of safety (f.s.)
equal to 1.5, is
Sys = 24,800
16,600 psi
f.s. ~ 1.5
Hence, the maximum stress at the fillet is satisfactory. If it were larger than the
allowable stress for the assumed material, the latter could be raised by assuming a steel
of higher carbon content or an alloy steel with the appropriate heat-treatment. A
change in material would not, however, affect the deflection characteristics of the
shaft; hence the nominal diameter of the shaft, 134 in., would remain unaltered.
13-8 Determination of shaft sizes on the basis of rigidity. In many

machines the sizes of shafts are determined by the limits that are placed
on their deflections. Two kinds of rigidity must be considered, torsional
and lateral.
190 MACHINE DESIGN
Torsional rigidity. The total angle of twist in radians for a circular

shaft of uniform cross section is given by the equation
4 TL
r ~ JG
For a solid circular shaft of diameter d, this equation becomes
584L77
(13-7)
“ Gd4
where 6 = angle of twist, deg
L — length of shaft, in.
T = torque on shaft, lb-in.
d = shaft diameter, in.
The permissible amount of twist depends on the particular application.
In drive shafts of machine tools, the twist should not exceed 0.08 deg
per ft. In line shafts, 0.75 to 1.0 deg per ft may be used as limiting
values.
Lateral rigidity. The diameter of shafts may be determined by per¬
missible lateral deflections as required, for instance, to maintain proper
bearing clearances or gear-teeth alignment. The deflections of a shaft of
uniform section may be readily found by the use of the appropriate equa¬
tion from mechanics of materials. If the shaft is of variable cross section,
the deflections may be determined from the fundamental equation for the
elastic curve of a beam, namely,
d2y _ M
dx2 ~ an
Two integrations of this equation will yield the deflection y. This inte¬
gration process is not difficult if the moment of inertia of the shaft is
uniform and if the distribution of the bending moment along the shaft
can be expressed analytically. In most cases, however, neither of these
conditions exist. However, a solution may be conveniently reached by
use of graphical methods (see Appendix XIII).
13-9 Code for design of transmission shafting.1 The code is based on
the failure of ductile ferrous shafting according to the maximum-shear
theory. The equation for the outside diameter d is
d = Jl6 V[KmM + }4aFd(l + K2)]2 + K,T)2 (13-8)

' 5TS,o(l - X4)
1 Code for the Design of Transmission Shafting, Engineering and Industrial Stand¬
ards, ASME, 1927.
SHAFTING 191
where ssa = allowable shear stress, psi

Km = combined shock and endurance factor for bending
Kt = combined shock and endurance factor for torsion
K = di/d = ratio of internal to external diameters
a = column-action factor
F = axial tension or compression, lb
Table 13-3
Km Kt
Stationary shafts:
Load gradually applied. 1.0 1.0
Load suddenly applied. 1.5-2.0 1.5-2.0
Rotating shafts:
Load gradually applied. 1.5 1.0
Load suddenly applied, minor shock. . 1.5-2.0 1.0-1.5
Load suddenly applied, heavy shock. . 2.0-3.0 1.5-3.0
This equation considers all the combined actions of loading and is

simplified if some of the loads are absent. Hence, if F — 0 and M = 0,
the equation for a shaft in torsion is obtained.
The column-action factor a is given as unity for a tensile load. For
a compressive load the following applies:
_1_ L
for — < 115
a “ 1 - [0.0044(L/p)] P
«-As(-y
7v2nE \p)
, L
for - > 115
P
where n = 1 for hinged ends and n = 2.25 for fixed ends. For ends that
are partly restrained, as in bearings, n = 1.6. p is the radius of gyration.
The allowable shear stresses to be used with code for commercial-steel
shafting are
ssa = 8,000 psi for shafts without key ways
Ssa = 6,000 psi for shafts with keyways
13-10 Vibration of shafts. In many rotating shafts vibrations are set

up that become troublesome or destructive at certain speeds. Frequently
violent vibrations are caused by the centrifugal forces on the shaft because
of unbalanced gears, pulleys, disks, or other rotating parts (see Chap. 24).
CHAPTER 14
BELT DRIVES AND HOISTS
14-1 General considerations. In the transmission of power from one

shaft to a parallel one, the types of mechanical drives that may be used
are flat-belt, V-belt, chain, or spur-gear drives. Some comparative fea¬
tures that should be considered in selecting the type of drive for a par¬
ticular application are the following:
Center distance. Belts and chains are suitable for long center distances
as compared with gears, which may be used for comparatively short center
distances. V belts are generally short-center drives.
Velocity ratio. Because of the slip and creep of belt drives, the driven
pulley will rotate at a speed less than that determined by the use of ratios
of pulley diameters. This loss in speed may be compensated for by a
small change in diameter of one of the pulleys. In chain drives the veloc¬
ity ratio is not constant during one revolution of the smaller sprocket.
In circular gear drives, the velocity ratio is always constant.
Shifting. Flat belts with relatively long center distances may be
shifted by the use of tight and loose pulleys. Gears may be shifted when
provided with feather keys, splines, or clutches.
Maintenance. The maintenance of belt drives usually amounts to peri¬
odic center-distance adjustment to compensate for stretch of the belt, and
replacement of the belt when it is no longer serviceable. In chain drives,
center-distance adjustment is necessary to compensate for wear of the
chain but is generally required at infrequent intervals only. In chain
and gear drives, lubrication is an important item in connection with
maintenance.
Cost. In general, the comparative first cost of the drives under discus¬
sion is in the order of flat belts, V belts, roller chains, silent chains, and
gears. In the positive drives, i.e., chains and gears, the necessity for
greater accuracy in shaft alignment and for provision for lubrication
results in further increase in first cost.
14-2 Materials. In belt and rope drives, the materials used must be
strong, flexible, and durable and must have a high coefficient of friction.
The following materials are most commonly used for these drives:
192
BELT DRIVES AND HOISTS 193
Leather. Oak-tanned leather is a standard material for flat belts.

Chrome leather may be used where a very pliable material is desired.
Fabric and rubber belts. Fabric, such as canvas or cotton duck usually
impregnated with a filler, is a relatively cheap material and may be used
for light power transmission. Such fabric may be vulcanized to form
rubber belts for use where exposed to oil or sunlight.
V belts. V belts are made of fabric and cords moulded in rubber and
are generally covered with fabric. As shown in Fig. 14-1, the section that
carries the power load is located near
^ \ Tension
the neutral axis of the section where the \\\— — --------z) // I in bend in a
y --r/// ' *
stresses owing to the bending of the
belt around the pulley will not be large.
\\\---r.-r.-z--ir-_r_r-—z//
j\ Loota tension
// \ Compression
The fabric near the upper surface ) in bending
carries the tension due to bending and Fig. 14-1. Section of V belt.
the rubber portion at the bottom is
compressed during bending. The belt is vulcanized to form a loop of
the exact length required for the particular installation.
In a recent development, the load-carrying element is composed of
small endless steel cables, which give high strength and low stretch.
Rope. Manila, hemp, and cotton ropes are used occasionally to trans¬
mit power. The continuous-rope, or the multiple-rope system may be
used. The rope sizes vary from % to 2 in. in diameter.
Material for pulleys. Pulleys are generally made of cast iron, cast steel,
or pressed steel. The cast materials have good friction and wear charac¬
teristics. Pulleys made of pressed steel are lighter than cast pulleys,
but in many instances have lower
friction and may produce exces¬
sive belt wear.
14-3 Angle of contact and
length of belts. The maximum
power transmitted by a belt drive
is limited either by stretching of
the belt or by excessive belt
Fig. 14-2. Geometry of open-belt drive.
slippage at the pulleys. Stretch¬
ing may be controlled throughout
a reasonable belt life by the use of low allowable stresses in the belt.
Since slippage of the belt would occur at the pulley having the smaller
angle of contact, the smaller angle of contact should be used in the
design of a belt drive.
In the open-belt drive shown in Fig. 14-2, the smaller angle of contact is
at the smaller pulley. From the geometry of the figure it may be seen
that
D-d
6 = 7T — 2a = 7T approximately (14-1)
C
194 MACHINE DESIGN
In a horizontal flat-belt drive, it is customary to use the top side of the

belt as the slack side during the transmission of power.
The length of an open belt may be shown to be approximately equal to
\ (D + d) + 2C + (Z)~,(14-2)
For a crossed belt, the angle of contact is the same on both pulleys and,
approximately, equal to
e = 7T + (14-3)
and the length of the crossed belt is, approximately,
+ a) (D + d) + ViC2 - (D + d)2 (14-4)
14-4 Flat-belt drives. If the drive shown in Fig. 14-3 is stationary

and the belt is tightened by adjusting the center distance, the tensions
around the belt will be equal, as shown by the initial tension vector Fq.
Now if the smaller pulley is assumed to rotate as a driving pulley in the
direction shown to transmit power, the tension in the upper side will be
increased and that in the lower side decreased, as shown by F1 and F 2.
In a belt transmitting power, these tensions are known as the “tight
tension” Fi and the “loose tension” F2. Their difference (Fi — F 2) is
the net tension, which may be determined from the horsepower being
transmitted and the velocity of the belt.
The ratio of the belt tensions may be determined from the following
analysis.
Ratio of belt tensions. The forces acting on a belt are shown in
Fig. 14-4.
Let Fi = tight tension, lb
F2 = loose tension, lb
6 = angle of contact, rad
r = radius of pulley, in.
/ = coefficient of friction
b = width of belt, in.
t thickness of belt, in.
w weight of belt material, lb per cu in.
V belt speed, fps
Fig. 14-4. Forces on element of belt.
The forces acting on an element of the belt shown crosslined in the

sketch are as follows:
Let F = force on left-hand end
F + dF = force on right-hand end
P = pulley force
fP = frictional force
C = centrifugal force
From the conditions of equilibrium,
2F„ = P + C - F sin — - (F + dF) sin — = 0

2 2
or
P + C - Fd6 = 0
also
„ dd dd
XFh = F cos — - {F + dF) cos — + fP = 0
or
-dF +fP = 0
Now
C = mass X acceleration
r dd btw I2v2
=-= Fc dd
a r
where
12 wbtv2
9
196 MACHINE DESIGN
By substituting C = FcdO into Eq. (a), eliminating P between Eqs. (a)

and (b), and assuming/ as constant,
dF
¥ - F, = f
Therefore
or
(14-5)
Power transmitted. For convenience in solving belt problems, the belt

tensions may be expressed in terms of pounds per square inch of belt
cross section. Equation (14-5) then becomes
fl-fe efe (14-6)

h~fc
where fi = maximum stress in the belt, psi
/2 = minimum stress in the belt, psi
„ 12 wv2
Solving Eq. (14-6) for (/] — /2), it becomes
(/l - /.) = (/l - fc) "-jiP t/
The horsepower that may be transmitted per square inch of belt cross
section is
hp per sq in.
(/1 />2
550
v(fi ~ fc) efe - 1
(14-7)
550 efd
where fj = allowable stress in the belt, psi

12 wv2
fc =
g
W = weight of belt material, lb per cu in.
0.035 lb per cu in. for leather

V = belt speed, fps
g = acceleration of gravity = 32.2 fps2

f= coefficient of friction
6 = angle of contact of belt with pulley, rad
The cross-sectional area of the belt required may be found by dividing

the horsepower to be transmitted by the horsepower per square inch
[Eq. (14-7)]. The width and thickness of belt may then be selected
from Table 14-1.
Table 14-1. Leather-belt Data
Thickness, in.
Grade of Increments of
belting Single Double Triple Quad¬ width, in.
to 8 in. to 12 in. to 24 in. ruple
A to A
1 by
Light. X X 1 to Fi
4 by
4 to 7 by y2
Medium. .. 5A2 V2 8 to 30 by 1
32 to 56 by 2
Heavy. He % %6 % 60 to 84 by 4
Initial tension. The relation between the initial tension F0 and the ten¬
sions Fi and F2 depends on the stress-strain characteristics of the belt
material. For leather, the stress is not a linear function of strain, and
hence the relation between Fh F2, and F0 is not a simple one. The fol¬
lowing equation may be used to estimate the effect of initial tension on
the power tensions:1
y/F~i + \/F~2 = 2 \/Fo
A suitable initial tension for leather belts is 200 to 225 psi.
14-5 Leather-belt data. Data applying to leather belts are given in
Table 14-1. The ultimate strength of oak-tanned leather varies from
3,000 to 4,500 psi. In order to secure reasonable life of a belt, it is neces¬
sary to limit the maximum working stress to a value considerably lower
than its ultimate strength. On this basis, the value of /3 in Eq. (14-7)
should be limited to from one-tenth to one-eighth of the ultimate strength,
say 400 psi for average conditions.
Coefficient of friction. The coefficient of friction / depends on the mate¬
rials of the pulley and the belt, on the slip, and on the belt speed. For
leather belts on dry cast-iron or wood pulleys a recommended value of
the coefficient of friction is 0.40 to 0.50.
Belt velocities used in practice for maximum power and economy range
from 5,000 to 6,000 fpm, but for longer belt life, 3,000 to 4,000 may be
used. A considerably lower value may be required, however, on account
of low shaft speeds and limited pulley sizes. At high speeds, the pulleys
should be dynamically balanced.
Unless the belts are endless, some type of belt joint or fastener must be
used. A cemented joint is practically as strong as the material joined.
1 C. Barth, Trans. ASME, vol. 31, p. 29, 1909.
198 MACHINE DESIGN
Wire or rawhide lacing decreases the strength of the belt, but if the means
of joining the ends does not affect the stretching of the belt or considerably
decrease its strength, the belt should operate satisfactorily if low allowable
stresses are used in the design' of the drive.
14-6 V-belt drives. Grooves in V-belt sheaves. On account of the
wedging action of a V belt or rope in the groove of the sheave or pulley,
the traction force (force transmitting power) is greater than in a flat belt
running on a flat-face pulley. To ensure wedging action in the groove,
the belt should make contact with the sides of the groove but not the
bottom. It is evident that the
wedging action and the traction
force is large for small groove
angles (see Fig. 14-5); however,
the force required to pull the belt
out of the groove as it leaves the
sheave is large for small groove
angles, resulting in loss of power
and excessive wear of the belt.
Fig. 14-5. Forces on flat and V belts.
The selected groove angle is there¬
fore a compromise to secure large traction force without unduly large
force to pull the belt from the groove. The standard angle between
the sides of V belts is 40 deg. The groove angle in the sheave is less than
the belt angle to allow for change in shape of the belt cross section on
wrapping around the sheave. Groove angles of 32 to 38 deg are used.
Standard sizes of belts for power transmission have been adopted and
designated by sizes A, B, C, D, and E as indicated in Table 14-2. For
low bending stresses as the belt wraps around the sheave, small belts are
better for long belt life; however, a large number of small belts requires a
wider sheave which increases the load overhang (distance from the belt
force to the bearing) which in turn increases the shaft stresses and bear¬
ing loads, all of which raises the cost of the drive. Thus optimum design
requires compromises in belt selection.
V-belt sheaves. The use of large-diameter sheaves and the correspond¬
ing high belt speeds decreases the load on the belts required to transmit
the power, but the gain with high belt speeds is partly nullified by the
increase in centrifugal forces due to the high speeds. Again a compromise
must be reached. The accompanying tables will be helpful in selecting
the most economical drive.
The sheaves mav be made of cast iron, which has excellent friction
characteristics on V belts, or of pressed steel, which is lighter and cheaper
but may give rise to excessive belt slip, wear, and noise. Sheaves may be
purchased with single or multiple grooves and are available also with
variable pitch for speed control.
Table 14-2. Selection and Dimensions of V Belts
Motor speed, rpm

Normal
horsepower 720
1,800 1,200 900
and below
K t° i K A A A A
2 to 5 A or B A or B A or B A or B
10 A or B B B B or C
15 B B or C B or C C
20 B or C C C C
25 to 30 C C C c
40 C C or D C or D D
50 to 60 C or D C or D C or D D
75 C or D C or D C or D D or E
100 C D D D or E
125 D D D or E
150 to 200 D D E
250 E E E
300 and over E E
Section A B c D E
Width, in. H % % IK 1M
Thickness, in. Me 1/'3 2 1/^2 K 2%2
Types of drives. In Fig. 14-6 are shown common arrangements of

V-belt drives. A disadvantage
of the vertical drive at (b) is vibra¬
tion of the slack side of the belt
under load, which may cause
vibration of the equipment and
decrease belt life. In the V-flat
drive at (c), the small sheave is
grooved but the V-belt runs on
the flat ungrooved rim of the large
sheave or flywheel. This drive
avoids grooving of the flywheel
and is used where the speed ratio
is large and the center distance
relatively short. Slipping of the
belt on the larger sheave is usu¬
ally not troublesome since the
angle of contact is large.
In Fig. 14-6(d) is shown a
“double-V” drive used to permit
200 MACHINE DESIGN
rotation of sheaves in opposite directions. The belt has short life due to
high bending stresses, since the belt is double thickness, and to reversed
flexure.
Selection of V belts. The b^lt section may be selected from Table 14-2,
using the rated horsepower of the drive.
The design horsepower used in determining the number of belts is the
rated horsepower multiplied by the service factor from Table 14-3. The
service factor takes into account the severity of load transmitted which
depends on the characteristics of the driving and driven units. When
shock and peak loads are light, use the lower values from the table; for
severe loads, use the higher values.
Where space permits, it is good practice to use a size of driving sheave
to give belt speeds in the range of 3,500 to 4,500 fpm. Table 14-4 may
be used bo select a recommended size of driving sheave, and Table 14-5
for available stock sheaves. In these tables, the diameter given is the
Table 14-3. V-belt Service Factors

Agitators, paddle propellers. 1.0-1.2
Bakery machinery. 1.0-1.2
Flour, feed, and cereal mills. 1.0-1.4
Machine tools. 1.0-1.4
Conveyors. 1.0-1.5
Generators and exciters. 1.2
Laundry machines. 1.2
Paper machinery. 1.2
Printing machinery. 1.2
Compressors. 1.2-1.4
Screens. 1.2-1.4
Brick and clay machinery. 1.2-1.6
Oil-field machinery.. 1.2-1.6
Textile machinery. 1.2-1.8
Fans and blowers. 1.2-2.0
Pumps. 1.2-2.0
Crushing machinery. 1.4-1.6
Mills.. 1.4-1.6
Line shafts. 1.4-2.0
Rubber plant machinery. 1.4-2.0
Table 14-4. Recommended Size Driving Sheave
Minimum pitch diameter, Recommended range of small

Section
in. pitch diameters, in.
A 3.0 3.0- 5.0

B 5.4 5.4- 7.4
C 9.0 9.0-13.0
D 13.0 13.0-17.0
E 21.0 21.0-28.0
pitch diameter of the belt in the groove. The AB section is a composite

sheave for either A or B section belts. The diameter given is for the A
section; for B section, the pitch diameter should be increased by 0.4 in.
The horsepower ratings of single strand of V belts may be determined
from Eqs. (14-8).
Table 14-5. Stock Sheave Sizes
AB section* B section C section D section
1 to 6 grooves 2 to 10 grooves 3 to 10 grooves 5 to 12 grooves

pitch diameter pitch diameter pitch diameter pitch diameter
3.0 5.0 9.0 5.4 7.0 16.0 8.5 10.0 20.0 13.0 18.0
3.2 5.2 10.6 5.6 7.4 18.4 9.0 10.2 24.0 13.4 22.0
3.4 5.4 12.0 5.8 8.6 20.0 9.2 10.6 30.0 13.8 27.0
3.6 5.6 15.0 6.0 9.4 25.0 9.4 11.0 36.0 14.2 33.0
3.8 5.8 18.0 6.2 11.0 30.0 9.6 13.0 44.0 14.6 40.0
4.0 6.0 19.6 6.4 12.4 38.0 9.8 16.0 50.0 15.0 48.0
4.2 6.2 24.6 6.6 13.6 15.4

4.4 6.4 29.6 6.8 15.4
4.6 7.0 37.6

4.8 8.2
* Composite sheave for A or B section. Diameters shown are for A section. For B section, add 0.4
in.
Belt
Section Horsepower Rating per Strand
1.95 3.80
A hp = V ■J70.09
- 0.0136F2 (14-8o)
kd
3.43 9.83
B hp = V ■j/0.09
- 0.0234F2 (14-86)
kd
6.37 27.0
C hp = V - 0.0416F2 (14-8c)
J/0.09
kd
13.6 93.9
D hp = V yo.09 - 0.0848 F2 (14-8d)
kd
19.9 178
E hp = V yo.09 - 0.122F2 (14-8e)
kd
where hp = maximum horsepower recommended for one strand of belt

of standard quality (multiply by arc-of-contact factor and
length-of-belt factor from Fig. 14-7 to obtain corrected
horsepower)
202 MACHINE DESIGN
V = .belt speed, thousands of feet per minute

k = small-diameter factor for speed ratio of drive (Fig. 14-7)
d = pitch diameter of small sheave, in. (the maximum values
of kd for belt sections are as follows: A, 5 in.; B, 7 in.;
C, 12 in.; D, 17 in.; E, 28 in.)
In the above equations the first term in the bracket depends on the
strength of the belt, the second term is an allowance for bending stress
as the belt passes over the sheave, and the third term is an allowance for
centrifugal force. In the bending-stress term, the diameter factor k
(see Fig. 14-7) allows for the smaller bending stress as the belt passes
over the driven sheave if it is larger than the driving sheave. If the
Speed rafio
Arc of con- Belt length, in.

tacf, deg
Fig. 14-7. Factors for horsepower ratings of V belts.
speed ratio is 1:1, k is unity, but as the speed ratio increases, the driven
sheave becomes larger and the diameter factor increases.
The power transmitted may be limited by slipping of the belt on the
small sheave. This is taken into account by the arc-of-contact factor
given in Fig. 14-7 in terms of the arc of contact of the belt on the smaller
sheave. The factor is unity for 180-deg contact and decreases with the
arc of contact.
A final factor is necessary to allow for the effect of length of belt on the
frequency of bending. Length-of-belt correction factors are given in
Fig. 14-7.
The horsepower rating from Eq. (14-8) is corrected by multiplying by
the small-diameter factor, the arc-of-contact factor, and the belt-length
factor. The corrected rating (horsepower per belt) is then divided into
the design horsepower to obtain the number of belts required.
If the number of belts is excessive, a larger belt section and/or increased
sheave diameters may be considered. A moderate number of belts is
desirable so that if one belt breaks or stretches excessively, the remaining

belts may carry the load until replacement of the belts can be made.
In replacing the belts, a complete set of new belts should be used rather
than replacing a single damaged belt, since a new (unstretched) belt
would carry more than its share of the load and would have a short life.
In V-belt drives, if assembly or replacement requires means for dis¬
connecting the belt, a metal coupler may be used, or belts made up of
removable, laminated leather V-belt sections may be used. In the latter
case, tension may be adjusted by adding or removing sections. An appli¬
cation requiring such a belt is a belt-driven generator or compressor
mounted under the bed of a passenger railway car and driven by a sheave
on the car axle.
Example 14-1. Select a V-belt drive for a 10-hp 1,160-rpm induction motor to
drive a ventilating fan at approximately 400 rpm. The minimum center distance
between the sheaves is 40 in.
Solution: From Table 14-2, a B-section belt is recommended.
From Table 14-3, assume the service factor is 1.4.
The design horsepower is 1.4 X 10 = 14.
Sheave selection: From Table 14-4, the recommended range of size for the small
sheave is 5.2 to 7.4 with 5.4 as a minimum. From Table 14-5, assume 7.0 in. as a trial
pitch diameter.
The belt speed is
7rd rpm 7r X 7.0 X 1,160

2,120 fpm
~V2 = 12 *
The diameter of the large sheave is
1,160
X 7.0 = 20.3 in. Use 20 in.
400
The speed of the fan will then be 406 rpm.

The speed ratio is approximately 3:1, which gives the small-diameter factor k, from
Fig. 14-7, equal to 1.13. Thus kd = 1.13 X 7.0 = 7.91 in. Since the maximum
value of kd is 7 in., this value will be used in Eq. (14-86).
3.43 9.83
hp per belt = V - 0.0234F2 J
po .09
kd
3.43
- ^ - 0.0234 X
“ 2'120 [2. 1200 09
= 2.120(3.2 - 1.4 - 0.10) = 3.61
The arc of contact, from Eq. (14-1), is
7r = 7T 2.81 radians, or 161 deg
From Fig. 14-7, the arc-of-contact factor is 0.95.

204 MACHINE DESIGN
The belt length is, from Eq. (14-2),
dV
L — “ (D + d) + 2C + —
2 4C
tv (20 - 7)2
= _(20 +7)+ 2 X 40 + T3^-
= 42.4 + 80 + 0.1 = 122.5 in.
(A standard belt length should be selected from manufacturer’s data.)

From Fig. 14-7, the belt-length factor is 1.07.
The corrected rating of one belt is
3.61 X 0.95 X 1.07 = 3.67 hp
The number of belts required is
14
3.82 Use 4 belts
3.67
Thus, four B-section V belts with sheaves 7.0 and 20 in. in diameter would be
specified.
14-7 Tooth-belt drive. The sheaves have axial grooves which engage
teeth on the belt, as shown in Fig. 14-8. The belt employs a number of
small steel cables which carry tension
under load, thus permitting a light
drive and high speeds. The drive is
positive, and the tension to transmit
the power does not depend on initial
tension. Sheave diameters may be
as small as in* to as large as re¬
quired, and belt speeds of the order
of 15,000 fpm are practical. Appli¬
cations include business machines,
sewing machines, timing drives, port¬
able woodworking equipment, and
power-transmission units. Besides
being positive, the drive is compact,
light, quiet, versatile, and low in
maintenance. However, it is more
sensitive to misalignment than flat
belts or V-belt drives. The first cost
Fig. 14-8. Timing belt drive. (Cour¬
is higher on light-duty belt drives but
tesy of United States Rubber Company.) may be more economical on large
drives.
14-8 Pivoted motor mountings. When a belt transmitting power
passes over its sheaves, the tension in the belt varies from the maximum
value Fi to the minimum value F2, and it is the difference in tensions
Fi — F2 which accounts for the power transmitted. At full power Fi is

large and F2 small as compared with the values at partial power. It is
evident that the initial tension in the belt must exceed the minimum value
of F2 in order to hold the belt against the sheaves. When the drive is at
rest, the initial tension must remain in the belt in order to be available
for full power requirements; thus, in time, stretching of the belt will
shorten its life.
Pivoted motor drives were developed to reduce the initial tension to a
minimum and to provide automatically for increasing tension as required
by the power being transmitted. Thus when the drive is idle or working
at partial load, the belt tensions are comparatively lower than in the
fixed-center type of drive and belt life is correspondingly increased.
In Fig. 14-9, the motor pulley is rotating clockwise. The belt tensions
F1 and F2 tend to overturn the motor in a clockwise direction about the
base pivot point. Neglecting centrifugal tension, the overturning
moment is F\a + Fb2. The moment which resists the overturning
moment is produced in part by the weight of the motor acting at the
lever arm c, that is, Wc, and in part by the reaction torque T on the frame
of the motor which is counterclockwise and equal to the torque on the
rotor. Thus
Fia + F2b = Wc + T (14-9)
The effect of the reaction torque T is to rotate the motor counterclock¬

wise about the pivot point and thus to increase the belt tensions. If the
value for the torque T is divided by the distance from the pivot point to
the center of the motor, that is, the lever arm g in the figure, the result
206 MACHINE DESIGN
will be a force acting at the center of the motor which in turn is carried
by the belt. This action performs the same function as initial belt ten¬
sion, but is induced only as needed by the torque on the motor in trans¬
mitting power. The force acting at the center of the motor will depend
on the lever arm g; the smaller the lever arm, the greater will be the force,
and vice versa.
Two types of pivoted motor mountings are in use. In one, the Rock-
wood-type drive shown in Fig. 14-9, the distance c and the distance g are
both large, so that most of the belt tension is due to the product of W and
c and that due to the torque reaction is small. The lever arm c may
be altered by shifting the motor along the pivoted support to provide
belt tensions as required for the maximum power to be transmitted.
When this power is exceeded, the
belt will slip on the sheave, which
will prevent overloading the motor
but may damage the belt.
In the second type of pivoted
motor mounting, the American
drive shown in Fig. 14-10, the dis¬
tances c and g are small so that
the belt tensions Fi and F2 are pro¬
duced mainly by the reaction
torque T. This torque varies with
the power transmitted so that
there is never more tension in the
belt than the load requires. In
order to locate the pivot point near
the center of the motor, the motor
is mounted in a cradle so that the
motor will swing about the pivot point shown in the figure. In this
drive, belt tensions are built up as required by the load. When the
drive is idle or is operating at partial load, the tensions will be low;
hence, belt life will be increased. However, if the drive should be over¬
loaded, the belt tensions will be increased accordingly and the belt may
be broken or the bearings may be damaged.
14-9 Arrangements of hoisting tackle. A simple hoisting tackle is
shown in Fig. 14-11(a), in which a rope passes over a sheave so that a
load Q may be raised by a downward effort P. To reduce the effort, an
arrangement of pulley blocks may be used as showm at (b), in which the
lower, or hook, block and the upper block each has two sheaves. To
avoid confusion in representing such a hoist, a developed diagram (c) may
be used. The hoist at (c) is known as a “four-part” line hoist, since four
lines of rope lead to the hook block.
It is evident that the more lines leading to the hook block, the less
will be the effort P required to raise a load Q, but that more rope must be
reeled off at P to raise the load a given distance.
Fig. 14-11. Hoisting tackle.
14-10 Ratio of rope tensions. When a rope passes over a stationary

sheave, as in Fig. 14-12(a), the loads P and Q will be equal. Note that
the moment of the effort about the sheave center is PD/2 and is equal to
QD/2.
Fig. 14-12. Rope passing over sheave.
If it is desired to raise Q, it is common experience that it is necessary

to apply an effort P which is greater than Q. The increase in P is due
to several causes, as indicated at (b) in the figure. When the rope runs
onto the sheave at /, the resistance of the rope to bending will shift its
208 MACHINE DESIGN
center line a distance e to the right of the vertical tangent, as shown in

the figure. The moment Qe is that required to bend the rope. As the
rope assumes the curve of the sheave in passing from / to g, the wires
or fibers of the rope shift on4 one another and adjust themselves, so that
when the rope arrives at g it will be naturally curved and must be
straightened as it leaves the sheave. This shifts the rope to the right
a distance e', as shown in the figure. Thus the lever arm of the resistance
Q has been increased and that of P has been decreased.
In Fig. 14-12(a), the support (P + Q) provided by the sheave pin was
drawn through the center of the pin; but when the sheave rotates, as in
(6), the load (P + Q) will be displaced a distance/d/2, as shown where
/ is the coefficient of bearing friction and d is the diameter of the pin,
i.e., /d/2 is the radius of the “friction circle.”1 Moments may now be
taken about the line of action of (P + Q), for instance, and the following
equation obtained for the relation between the
effort P and the resistance Q:
where C has a value greater than unity. Values

of C depend on the size of rope, on the relative
size of sheave and pin, and on the coefficient of
friction, and have been determined experi¬
mentally. For manila rope, an average value
Fig. 14-13. Loads on a
hoist.
for C of 1.14 may be used; for wire rope, see
Table 14-9.
If the angle of wrap is different from 180 deg, the values of C as given
may be used without introducing appreciable error.
14-11 Loads and efficiency. As an example of determining the loads
on the various ropes of a hoist and its efficiency, the arrangement shown
in Fig. 14-13 will be used.
Raising the load. Label the ropes Ti, T2, etc., to represent their ten¬
sions. From the analysis in Art. 14-10
P = CT, = CbT,
Tb = CT4 = CAT,
Ta = CT, = C3Ti
T3 = CT2 = C2T1
t2 = ctl
1 C. D. Albert, “Machine Design Drawing Room Problems,” p. 411, John Wiley &
From the equilibrium of forces on the lower block
Q = T1 + T2 + T, + T4 + T6
C6 - 1 Cb - 1
= Ti( 1 + C + C2 + C3 + C4) = Tx = P
C - 1 C\C - 1)
or
P = Q (14-11)
C( 1
Without friction, P0 = Q/5. Hence, the efficiency of the hoist is
Po X unit distance _ Cl 1
(14-12)
P X unit distance 5C5(C — 1)
Lowering the load. For this case
T, = CT2
Tz = CTs
T3 = CT4
T4 = CTb
T5 = CP
C - 1
P = (14-13)
C(Cb - 1) Q
In the above example, note that in raising the load, the rope P is the
maximum loaded one, while in lowering the load, the rope is the maxi¬
mum loaded one. The capacity of a hoist depends on the most heavily
loaded rope, as will be discussed in Art. 14-14.
Table 14-6. Ultimate Strength of Manila Rope

/d = rope diameter, in. \
\FU = ultimate strength, lb /
d Fu d Fu d Fu d Fu
% 1,270 5A 4,000 1 Vs 10,500 20,000
H6 1,870 H 4,700 1H 12,500 m 25,000
V2 2,400 Vs 6,500 15,400 2 30,000
He 3,300 l 7,500 iK 17,000 2K 37,000
14-12 Manila rope. Manila rope is used for hoisting only in very small
capacity hoists and where safety is not a prime consideration. Allowable
loads for a reasonable life of rope may be found by dividing the ultimate
strength of the rope Fu in pounds by a factor of safety equal to 7 for rope
speeds up to 100 fpm, increasing to 18 at 300 fpm and to 36 up to 800 fpm
(see Table 14-6 for ultimate strengths of manila rope).
210 MACHINE DESIGN
14-13 Wire rope. Construction. Wire rope is constructed of strands

each of which is made of small wires twisted together. This construction
permits the rope to be wrapped
around a sheave without undue
bending stresses in the wire.
Two types of winding are in use:
Ca)-Requ'\ar lay
one t3^pe (a) (see Fig. 14-14), in
which the direction of twist of
the wires in the strands is opposite
to the direction of twist of the
strands, is known as “ regular
(b)-Lang lay
lay.” When the directions of
twist of the wires and strands are
fcj-Section of 6 x7
coarse laid
the same, as in (b) of the figure,
it is known as “lang lay.” The
regular lay will not untwist or
Fig. 14-14. Winding of wire rope.
spin under load as much as lang
lay, but tests have shown that it has a shorter life.
Wire rope is made in several standard constructions of which the fol¬
lowing cover most services:
6 X 191—standard hoisting rope

6 X 37—extra-flexible hoisting rope
8 X 19—extra-flexible hoisting rope
6 X 7—standard coarse laid rope
The hemp core is saturated with lubricant to retard rusting of the wires
and to reduce friction between the wires. The rope should be periodically
cleaned and re lubricated.2
Table 14-7. Increments of Standard Sizes of Wire Ropes
Rope Diameter, in. Increment, in.
6 X 19 34 6
and ^-234 34
6 X 37 234-224 X
8 X 19
and
H-Hg 346
6X7
x-ix 34
1 The first figure represents the number of strands and the second the number of
wires in each strand. For example, a %-in. 6X7 wire rope is 24 in. in diameter and
has 6 strands each with 7 wires [see Fig. 14-14(c)]. The core shown in black is hemp.
2 ASA Safety Code, Elevator Wire Rope Maintenance, Mech. Eng., February, 1943.
Standard sizes of wire ropes may be found from Table 14-7. Ultimate
tensile strengths in pounds for plow-steel wire rope and weights in pounds
per foot may be approximated from the equations in Table 14-8.
Table 14-8. Approximate Wire Rope and Sheave Data
(d = rope diameter)
Ultimate Recommended
Weight diameter
Rope strength, dw A
lb per ft
Fu, lb Minimum
Average
6 X 19 72,000d2 1.60d2 0.063d 0.38 d2 45d 30 d

6 X 37 68,000d2 1.55 d2 0.045d 0.38 d2 27 d 18d
8 X 19 62,000d2 1.50d2 0.050d 0.35 d2 31d 21d
6X7 68,000d2 1.45d2 0.106d 0.38d2 72 d 42d
14-14 Bending stresses in wire ropes. As a wire rope is curved to fit

its sheave, the outer wires are subjected to tension in bending. This is
in addition to the tension due to the service load. As the rope passes
over the sheave, the bending stress is relieved by the readjustment of the
wires. When the rope is straightened as it leaves the sheave, the inner
wires are subjected to tension in bending.
Table 14-9. Values of C for Wire Rope
Rope diameter C Rope diameter c

% 1.090 % 1.064
Kt 1.083 H 1.054
A 1.076 Vs 1.046
He 1.070 l 1.040
A consideration of the mechanics involved leads to the conclusion that

the bending stress is directly proportional to the diameter of the wire and
inversely proportional to the diameter of the sheave, or
s = k^ (14-14)
where k is the constant of proportionality.

If the bending stress is multiplied by the net cross-sectional area A
of the wire in the rope, the product sA represents a load that may be
referred to as a bending load Fb. Its value from Eq. (14-14) becomes
F„ = kA ^ (14-15)
The sum of this bending load and the external, or service, load on the
rope should not exceed the allowable load on the rope, or, expressed as
212 MACHINE DESIGN
an equation,
~ ^F„ + F, (14-16)
l.S. v
Fb = kA^
where k = 12,000,000 psi

dw = diameter of wire in rope, in.
D = pitch diameter of sheave, in.
A = net area of wire in rope, in.2
Fu = ultimate strength of rope, lb
Fs = service load on rope, lb
Table 14-10. Recommended Minimum Safety Factors for Wire Rope

(American Steel and Wire Co.)
Track cables. 3.2
Guys. 3.5
Mine shafts:
Depths
To 500 ft. 8
1,000-2,000 ft.. ... 7
2,000-3,000 ft. 6
Over 3,000 ft. 5
Miscellaneous hoisting. 5
Haulage ropes. 6
Overhead and gantry cranes. 6
Jib and pillar cranes. 6
Derricks. 6
Small electric and air hoists. 7
Hot ladle cranes.. 8
Slings. 8
Equation (14-16) is based on the assumption that the significant prop¬

erty of a wire rope is its strength. The factor of safety is chosen so that
the rope selected on the basis of strength will have a reasonable life. This
means that the rope should not fail as a result of fatigue or wear.
An important matter contributing to both fatigue and wear is the com¬
pressive stress1 on the wires caused by bearing against the sheave. This
condition, among others, indicates that the analysis given here is approx¬
imate and must require the liberal factors of safety suggested in Table
14-10. The values in this table include an allowance for bending stress,
hence are higher than the factors applying to Eq. (14-16).
1 Drucker and Tachau, A New Design Criterion for Wire Rope, Trans. ASME, vol.
67, p. A-33, 1945.
CHAPTER 15
POWER-TRANSMISSION CHAINS
15-1 Features of chain drives? Important advantages of the chain

drive are (a) that it is adapted to long or short center distances and (6)
that one chain may be arranged to drive more than one unit. Some
arrangements are shown in Fig. 15-1. It is important that the shafts be
in good alignment, especially for high-speed chains and for wide chains.
Chain drives that are properly selected and installed operate at high effi¬
ciency and with low maintenance cost.
B B
(a,)- Shorf center

distance
Sprockets:
A-Driver
B~ Driven
C~ Idler
Fig. 15-1. Arrangements of chain drives.
15-2 Chain speed. The action of a chain as it runs with a sprocket

may be likened to that of a nonslipping belt running with a prism.
Assuming that the driving prism, or sprocket, rotates uniformly, it is
evident that the speed of the chain will vary from a minimum value, as
shown by the solid lines in Fig. 15-2(a), to a maximum value, as shown by
the dotted lines. Because of this variation in chain speed, the driven
sprocket will not rotate uniformly unless it has the same number of teeth
as the driving sprocket. By increasing the number of teeth on the
sprocket, the variation in chain speed will be reduced.
21 3
214 MACHINE DESIGN
For a sprocket having 11 teeth, the variation in chain speed amounts

to about 4 per cent; for 17 teeth, the variation is 1.6 per cent; and for
24 teeth, the variation is less than 1 per cent. Thus if the minimum
number of teeth in the small sprocket is 17 or preferably 24 teeth, the
drive should operate smoothly.
Vmax
The average chain speed is equal to the length of chain reeled off the
sprocket in unit time, or
pTN
V = (15-1)
12
where V = chain speed, fpm
p = pitch of chain, in.
T = number of sprocket teeth
N = sprocket speed, rpm
Although there is a trend toward higher chain speeds for ordinary
applications, economical speeds are of the order of 2,500 fpm for roller
chains and 4,000 fpm for silent chains. Both types of chains have been
operated satisfactorily up to 6,500 fpm.
It is usually desirable to use small pitches in a chain drive to prevent
surging of the chain; however, excessively small pitches may increase the
first cost unnecessarily. As an aid in selecting a pitch, the following
empirical formula which has given good results in practice may be used.
(15-2)
where p is the pitch in inches and N is the speed of the small sprocket
in rpm.
15-3 Number of teeth on sprocket. The most desirable number of
teeth on a sprocket depends on several considerations. As shown in
Fig. 15-3, the angle through which a chain link turns on its pin as it
POWER-TRANSMISSION CHAINS 215
engages with the sprocket is equal to 180 deg divided by the number of
teeth on the sprocket. Thus with a small number of teeth, the angle of
rotation will be large and wear of the pin and bushing will be rapid. This
consideration requires a minimum of 17 or preferably 24 teeth for steady
chain loading. Finally, for a particular pitch of chain, the size of the
sprocket—and hence the chain speed—increases with the number of teeth.
This means that the load on the chain required to transmit a given horse¬
power will be less with a large sprocket than with a smaller one, and hence
a smaller chain may be used. There is a limit to the chain speed, how¬
ever, because of dvnamic effects and lubrication difficulties.
The above discussion indicates that a large number of teeth on a
sprocket is desirable from an operating standpoint. However, the cost
Fig. 15-3. Action of chain engaging sprocket teeth.
of larger sprockets and cases, space limitations, and lubrication considera¬

tions generally favor a small number. Hence, the number of sprocket
teeth specified must represent a compromise. Recommended minimum
numbers of sprocket teeth are given in the tables for various types of
chains.
It is usually preferable to use an odd number of teeth on a sprocket
so that wear will be uniformly distributed over the teeth. This is desir¬
able particularly for exposed or partially lubricated drives. An odd
number of teeth favors the condition where an old chain is replaced by a
new one in that a sprocket with an odd number of teeth would show uni¬
form wear of the teeth, whereas a sprocket having an even number of
teeth would show alternate teeth having greater wear than the inter¬
mediate ones.
To determine the pitch diameter D of a sprocket, the geometry of
Fig. 15-3 may be used, i.e.,
.
oin —
d V
— — -i—
D
-
216 MACHINE DESIGN
or
V V
D = (15-3)
sin (6/2) sin (180/T)
For sprocket cutting, the pitch diameter is calculated to thousandths

or tenths of thousandths of an inch, modifying if necessary the calculated
value to provide clearance.
15-4 Design horsepower. The useful life of power-transmission chains
is usually terminated by troubles arising from wear at the joints. The
wear is caused by the rotation of the links as they are seated on the
sprocket, i.e., through the angle 6/2 in Fig. 15-3. The wear may be
aggravated by shock loads due to speed variations caused by the driving
unit or the driven equipment or both. Wear at the joints increases the
pitch of the chain so that the links will not seat properly on the sprocket
teeth which in turn causes increased dynamic loads on the joints. When
this occurs, the rate of wear is increased and the useful life of the chain
will be near its end.
To allow for shock loads and hours per day of use, the transmitted
horsepower should be multiplied by a service factor to obtain the design
horsepower (see Table 15-1).
Table-15-1. Service Factors for Power-transmission Chains
Roller chain, Inverted-tooth chain,

hr per day hr per day
Type of load
10 24 10 24
Uniform load, average conditions. 1.0 1.2 1.0-1.2 1.3-1.5

Moderate shock. 1.2 1.4 1.3-1.5 1.6-1.8
Heavy shock. 1.4 1.7 1.6-2.0 2.0-2.5
15-5 Block chains. These chains as shown in Fig. 15-4 are used
mainly for conveyor applications. The chain is relatively noisy and
wear is rapid because of the impact between the blocks and the sprocket.
Fig. 15-4. Block chain.
15-6 Roller chains. The side plates for the roller chain shown in
Fig. 15-5 are blanked from cold-rolled steel. The pins, bushings, and
rollers are generally made of alloy steel, hardened and ground. The
chains are rugged and durable and if properly selected, installed, and
lubricated will give excellent service.
Roller chains are manufactured in standard pitches, as shown in
Table 15-2. In the chain number, the right-hand digit 5 indicates a
Table 15-2. Recommended Maximum Rpm of Sprockets for Roller Chains
Chain No. 25 35 41 40 50 60 80 100 120 140 160 180 200 240
Pitch 34 % 34 X % % 1 1M ix m 2 234 234 3
Teeth Rpm
11 4,310 2,260 1,020 1,690 1,220 920 580 415 325 235 200 165 145 110
12 4,960 2,590 1,170 1,940 1,400 1,050 670 475 375 270 230 190 165 125
13 5,540 2,900 1,310 2,180 1,570 1,180 750 535 415 305 260 215 185 140
14 6,070 3,170 1,430 2,380 1,720 1,290 820 585 455 335 280 235 205 155
15 6,530 3,420 1,540 2,560 1,850 1,390 880 630 490 360 305 255 220 165
16 6,940 3,630 1,630 2,720 1,960 1,480 935 670 520 380 325 270 235 175
17 7,290 3,810 1,720 2,860 2,060 1,550 985 700 550 400 340 285 245 185
18 7,590 3,970 1,790 2,980 2,150 1,610 1,020 730 570 415 355 295 255 195
19 7,840 4,100 1,850 3,080 2,220 1,670 1,060 755 590 430 365 305 265 200
20 8,050 4,210 1,890 3,160 2,280 1,720 1,090 775 605 440 375 315 270 205
21 8,230 4,300 1,940 3,230 2,330 1,750 1,110 790 620 450 385 320 280 210
22 8,370 4,380 1,970 3,290 2,370 1,780 1,130 805 630 460 390 325 280 215
23 8,480 4,430 2,000 3,330 2,400 1,800 1,150 815 640 465 395 330 285 215
24 8,560 4,480 2,020 3,360 2,420 1,820 1,160 825 645 470 400 330 290 220
25 8,610 4,510 2,030 3,380 2,440 1,830 1,160 830 650 475 400 335 290 220
30 8,580 4,490 2,020 3,370 2,430 1,830 1,160 825 645 470 400 335 290 220
35 8,200 4,290 1,930 3,220 2,320 1,740 1,110 790 615 450 380 320 275 210
40 7,580 3,970 1,780 2,970 2,140 1,610 1,020 730 570 415 355 295 255 195
45 6,820 3,570 1,600 2,670 1,930 1,450 920 655 515 375 320 265 230 175
50. 5,950 3,110 1,400 2,330 1,680 1,270 805 575 450 325 275 230 200 150
55 5,010 2,620 1,180 1,970 1,420 1,070 675 480 375 275 235 195 170 125
60 4,020 2,100 950 1,580 1,140 860 545 390 305 220 185 155 135 100
218 MACHINE DESIGN
rollerless bushing chain, 1 indicates a lightweight chain, and 0 indicates

a chain of usual proportions with roller. The number to the left of the
right-hand digit is the number of 34 in. in the pitch. The recommended
maximum speed of the small sprocket is shown for various numbers of
teeth in the small sprocket. Sprockets are available from 5 to 159 teeth.
Fewer than 11 but preferably 24 are not recommended.
Equation (15-4) gives the recommended horsepower rating for a single
strand of roller chain. The ratings are based on a service factor of unity.
By dividing the horsepower per strand from the equation into the design
horsepower (which includes the service factor) the number of strands for
the chain is arrived at. Roller chains may consist of a single strand or
of two or more strands. The usual maximum width is four strands, but
wider chains with special design features for pin support are available.
In Eq. (15-4), the first term in the bracket depends on the allowable
bearing pressure on the pins and bushings to limit the rate of wear at these
points. The angle of rotation of the links as they become seated on the
sprocket affects the wear and is allowed for by the function of 6, and the
centrifugal tension is allowed for by the F1-41 term:
V \y i.4i
hp per strand = p2 (1+25 vers 9) (15-4)
23.7 1,057
where p = pitch of chain, in.
V = chain speed = pTN/12, fpm
T — number of teeth in small sprocket
N = rpm of small sprocket
180
Q
T
vers 6 1 — cos 6
Tables of horsepower ratings for various speeds of small sprocket,
chain pitch, and number of teeth on the small sprocket are given in the
American Standards Association
publication ASA B29.1-1950.
Considerable design information
is also given in that source, includ¬
ing recommended types of lubri¬
cation for various chain speeds.
Before the final selection of chain
is made, the manufacturers’ data
Fig. 15-6. Morse silent chain.
should be consulted.
A,
15-7 Inverted tooth chains. The inverted tooth, or silent, chain is

made of overlapping links connected by pins in the Reynolds type of chain
or by rocker joints in the Morse chain (see Fig. 15-6). Since the loading
on the pin or rocker is nearly uniformly distributed over its length, bend¬
ing of the pins is practically eliminated, and thus a small connection may
he used. The connection will therefore be so flexible that the chain
links will accommodate themselves to small inaccuracies in the sprocket
teeth, and pin breakage is uncommon. Thus the width of a silent chain
is not so closely limited as in a roller chain. Silent chains over 30 in.
wide have been used.
Inverted tooth chains are manufactured in standard pitches, as shown
in Table 15-3. In the chain number, SC refers to silent chain and the
numeral following is the number of in. in the pitch. The table gives
the recommended maximum speed of the small sprocket for various
pitches and number of teeth on the small sprocket. For best results, the
small sprocket should have at least 21 teeth. Sprockets are available
with 17 to 150 teeth.
Equation (15-5) gives the recommended ratings for inverted-tooth
chains in horsepower per inch of width of the chain. The computed value
may be divided into the design horsepower to give the width of chain
required. The range of chain widths should be between p and lOp but
preferably between 2p and 6p. The widths vary by 3^-in. increments
and maximum standard widths are given in Table 15-3.
Table 15-3. Maximum Rpm of Small Sprocket for Inverted Tooth Chains
Chain No. SC3 SC4 SC5 SC6 SC8 SC10 SC12 SC16
Pitch, in. X X % X 1 m IX 2
Max. width, in. 4 7 8 10 14 20 24 30
Number of
Rpm
teeth
17 4,000 3,500 2,500 2,000 1,200

19 5,000 3,500 2,500 2,000 1,500 1,200 1,000 700
21 6,000 3,500 3,000 2,500 1,800 1,200 1,000 700
23 6,000 4,000 3,000 2,500 1,800 1,800 1,200 800
25 6,000 4,000 3,500 2,500 1,800 1,800 1,200 900
27 6,000 4,000 3,500 2,500 2,000 1,800 1,200 900
29 6,000 4,000 3,500 2,500 2,000 1,800 1,200 900
31 6,000 4,000 3,500 2,500 2,000 1,800 1,200 900
33 6,000 4,000 3,500 2,500 2,000 1,800 1,200 900
35 6,000 4,000 3,500 2,500 2,000 1,800 1,200 900
37 5,000 3,500 3,000 2,500 1,800 1,200 1,000 800
40 5,000 3,500 2,500 2,500 1,500 1,200 900 800
45 4,000 3,000 2,000 2,000 1,500 1,000 900 700
50 3,500 2,500
hp per inch of width

2,000
The notation is the same as for Eq. (15-4).

1,800
pv_
53
,_
1,200 1,000
425(7’ - 8)
800 600
(15-5)
220 MACHINE DESIGN
In Eq. (15-5), the first term depends on the allowable bearing pressure
on the pins and the last term allows for the centrifugal tension and also
for the number of teeth in the small sprocket, which is related to the
angle of rotation of the linksqn seating on the sprocket.
The American Standards Association publication, ASA B29.2-1950,
and manufacturers’ data give tables of horsepower ratings of inverted-
tooth chains for various pitches and speeds of sprockets with various
numbers of teeth.
15-8 Maximum sprocket bore. In small sprockets, it is necessary to
determine whether the number of teeth and the pitch selected will result
in a sprocket large enough to be mounted on the shaft with due allowance
for a keyway. On this basis, the following relations may be used to
determine the minimum number of teeth that may be used with a given
pitch for a sprocket bore d.
Roller-chain sprocket:
4d
Tmin =-b 5 for pitches to 1 in. (15-6a)
4d
= — + 4 for pitches 134 to 2^ in. (15-66)
Silent-chain sprocket:
rp 4d | n
1 min =-b 6 for pitches to 2 in. (15-6c)
V
15-9 Length of chain. The following equation may be used to deter¬
mine the approximate length of a chain:
Tl + T2 2C p(T1 - T,y
(15-7)
2 “h p ^ 39.5(7
where L — length of chain in links

Ti, T2 = numbers of sprocket teeth
C = center distance, in.
p = pitch, in.
The length in links should be used as an even number unless it is desired
to use an offset, or hunting, link. The advantage of these links is that
adjustment of the center distance may be in units of one link instead of
a pair of links. A roller-chain offset link is shown in Fig. 15-5. The
center distance for sprockets should not be less than the sum of the diam¬
eters of the sprockets.
If the exact chain length is required for a specified center distance or
sprocket combination, the following equation may be used:
T Ti + T2 , 2C cos a t d(Ti + T?)

(15-8)
L ~ - + d + 180
The notation is the same as for Eq. (15-7); a is the angle between the
center line and the tangent to the sprocket pitch circles (see Fig. 14-2).
D\ — D %
sin a = —2Q—
15-10 Adjustment for chain tension. Means for adjusting the center
distance should be provided to secure proper chain tension when the
drive is installed and to make adjustment for wear of the chain in service.
There should be a small amount of slack in the chain, but excessive slack
may permit surging of the chain to take place. In some cases it may be
necessary to use idler sprockets or shoes installed on the slack side of the
chain to take up slack.
15-11 Chain cases and lubrication. Chain drives should be enclosed
to provide for proper chain lubrication, to keep dirt and grit from the
chain and sprockets, and to ensure the safety of the operator. The case
should be easy to install and to open for inspection and lubrication.
Except at very low speeds the chain should not run through the oil. A
high-speed chain running through oil causes an excessive amount of
churning which raises the temperature of the oil and which may create
high pressure in the case so that leakage may occur. Oil disks or rings
may be used to distribute the lubricant.
Example 15-1. Roller-and silent-chain selection. Select a chain drive to connect

an 1,150-rpm induction motor to a centrifugal pump rated at 1,000 gpm against
a head of 30 ft at 575 rpm of the pump. The efficiency of the pump is 63 per cent
and the efficiency of the chain drive may be taken as 98 per cent. The fluid pumped
is water and the duty is 24 hr per day.
Roller chain:
1,000 X 8.33 X 30
Transmitted hp = = 12.2
33,000 X 0.63 X 0.98
Select a 15-hp motor.
Design hp = 1.2 X 12.2 = 14.6.
Maximum pitch = 0^ = (n^j)§ = °-85 in-

From Table 15-2, assume %-in. pitch.
Assume 21 teeth on small sprocket.
From Table 25-2, the motor-shaft diameter is 1% in. From Eq. (15-6a),
T . _ 4d ,_ 4 X 1.875 . . _
J- min — | n rrr I O — Id
p 0.75
The assumed 21 teeth is therefore satisfactory for bore of small sprocket.

From Eq. (15-1), the chain speed is
_ pTN 0.75 X 21 X 1,150

1,510 fpm
V 12 12
222 MACHINE DESIGN
The angle of link rotation is
a 180 180
0 = ~y = ~2i = 8.5 deg
vers 6 = 1'— cos 6 = 1 — 0.989 = 0.011
V r V1A1~\
hp per strand = p2 ^7 — (1 + 25 vers e> 1^57 J
r 1,510 _ n + 25 X 0.011) l,510x 41
= (0.75)2
L 23.7 k 1,057 ]
30,360\
= 0.563 ( 63.7 - 1.275
( 1,057 ) ~ 15,t
No. of strands = t+t; = 0.955 Use 1 strand of No. 60 chain
15.3
Sprocket diameters:
Small sprocket:
V 0.75
£>1 = = 5.07 in.
sin (I8O/T1) sin (180/21)
Large sprocket:
T2 = X 21 = 42
575
V 0.75
D* = = 10.05 in.
sin (I8O/T2) sin (180/42)
Minimum center distance:
C = Di + D2 = 10.05 + 5.07 = 15.12 Use 16 in.
Length of chain:
T _Tr + T2 2C p(T2 - Tx)2

2 ^ V ^ 39.5C
21 + 42 2 X 16 0.75(42 - 21)2
2 + 0.75 + 1.39.5 X 16
= 31.5 + 42.7 + 0.525 = 74.7 links Use 76 links
Inverted-tooth chain: The design horsepower, assumed number of teeth, pitch, and
chain speed are the same as for the roller chain.
For the inverted-tooth chain,
vV T VI
hp per inch = 53 [ 1 ~ 425(T _ 8) J
0.75 X 1,510 r 1,510 1
53 L 425(21 - 8) J
14.6
Chain width = -+77 = 0.935 in. Use 1-in. wide chain
15.6
The sprocket diameters and chain length will be the same as for the roller chain.
15-12 Tension chain linkages. This application is a chain which has

lineal motion but not continuous in direction. The chain need not be
formed as an endless belt. In Fig. 15-7(a) is shown an hydraulic lift
with a chain passing over a floating sprocket. In this arrangement the
length of the cylinder is one-half the vertical lift of the platform. At
(6) in the figure is a spring-controlled device to reposition a shaft that

has been rotated through a limited angle. Other applications of chains
in tension linkages are chain hoists, chain earth-drilling rigs, elevators
for materials handling, lift trucks, and servomechanisms. For the latter
class of applications, a miniature mechanical chain which has a pitch of
0.147 in. is available in roller and inverted-tooth construction, made of
stainless steel or of beryllium copper (nonmagnetic) with sprockets as
small as 0.3401 in. in pitch diameter.
The advantages of chains versus wire cables in tension linkages are
positive connection between chain and sprocket, ease of flexure of chain,
(a) Hydraulic
lift
Fig. 15-7. Tension linkages.
simplicity of end connection of chain to attached member, and ease of

lubrication.
The allowable load in pounds on a single strand of a standard roller
chain used in a tension linkage under steady conditions (no shock) may
be determined from the equation
Allowable load = l,900p2 (15-9)
where p is the pitch of the chain in inches. For the lightweight chain,
No. 41, use 1,000 instead of 1,900 in the above equation.
The above equation is based on the ultimate strength of the chain and
an average factor of safety of 8. The real factor of safety is lower,
however, because the ultimate strength of the chain is not alone the
criterion of failure, since bearing pressure, cyclic loading, stress concen¬
tration, and endurance limit of the links and pins are also involved.
Any effects of shock loading should be allowed for by introducing a shock
factor in accordance with the judgment of the designer.
CHAPTER 16
SHAFT COUPLINGS
16-1 Introduction. Shaft couplings are used in machinery for several

purposes, the most common of which are the following: (a) to provide
for the connection of shafts of units that are manufactured separately,
such as a motor and a generator, and to provide for disconnection for
repairs or alterations; (6) to provide for misalignment of the shafts or
to introduce mechanical flexibility; (c) to reduce the transmission of
shock loads from one shaft to another; (d) to introduce protection against
overloads; (e) to alter the vibration characteristics of rotating units.
Some commonly used types of shaft couplings are described briefly in
the following articles.
Circumferential
Ca)-A\‘ujned f/ange '' -Radial
— flange
-r
i
(b)-Lotte roil misalignment
(c)-Anqu\ar misalignment
Fig. 16-1. Alignment of shaft center lines. Fig. 16-2. Flanged shaft coupling.
16-2 Rigid couplings. This type of coupling has no flexibility or resil¬

ience; hence it is necessary for the shafts that are to be connected to be
in good alignment, both laterally and angularly, in order to avoid exces¬
sive loads on the coupling, on the shafts, or on the shaft bearings. The
two types of shaft misalignment are shown in Fig. 16-1.
Flange coupling. This is a common form of rigid coupling; it is com¬
posed of two mating halves, which are shown assembled in Fig. 16-2.
The coupling bolts should be ground and fitted to holes that are drilled
and reamed in the radial flanges after the coupling halves are assembled.
This procedure is necessary in order that the bolts should share equally
the loads imposed on them when the coupling transmits torque. The
purpose of the circumferential flange is to avoid danger of exposed rotating
224
SHAFT COUPLINGS 225
nuts and boltheads. The coupling should be designed to avoid failure

of the key, failure by shearing off of the radial flange at its juncture with
the hub, and failure of the bolts.
Compression coupling. This cou¬
pling utilizes two split cones, as shown
at c in Fig. 16-3, which are drawn
together by the bolts b in order to
produce a wedging action which
tightens the parts of the coupling and
the shafts. The coupling in Fig. 16-3 Fig. 16-3. Compression coupling.
is the Sellers coupling. Some types
of compression couplings do not have keys, but depend entirely on friction
produced by the compression pressure to prevent slipping of the parts in
transmitting torque.
16-3 Couplings with kinematic flexibility. Oldham coupling. This
coupling may be used to connect shafts that have lateral misalignment.
The three parts of the coupling shown
in Fig. 16-4 are the two slotted hubs and
the central floating part. The floating
part has two tongues, one on each face,
at right angles to each other. The
-
tongue a fits into the slot of the right-

hand member and allows for to-and-fro
relative motion of the shafts, while the
tongue b fits into the slot of the left-hand
member and allows for vertical relative
motion of the parts. The resultant of
these two components of motion will
accommodate lateral misalignment of
Fig. 16-4. Oldham coupling. Up¬
the shafts as they rotate.
per view shows parts; lower view
shows assembly. The American flexible coupling shown
disassembled in Fig. 16-5 utilizes the
Oldham principle. By widening the slots and tongues in the Oldham
coupling, the central floating member may be formed into a block as
shown in the figure so that it may ride between the faces of the jaw flanges
of the driving and driven members. The bearing surfaces of the floating
block are provided with replaceable nonmetallic strips which are lubri¬
cated from the grease reservoir in the hollow block. This coupling may
be used for shafts that have lateral misalignment, and may be used also
with some angular misalignment if the coupling halves are properly
mounted with axial clearance.
The Amerigear flexible coupling transmits torque from one shaft to the
other by means of two hubs and a floating sleeve. The hubs have exter-
226 MACHINE DESIGN
nal gear teeth cut on their peripheries and the sleeve incorporates match¬
ing internal teeth. The two hubs are separated axially, and the gear
teeth are of a special design with involute flanks and faces curved in
both directions so that they {nay accommodate angular misalignment by
a ball-and-socket action. This coupling has wide application in railway,
automotive, air-transportation, and marine equipment as well as
extensive industrial use.
Chain coupling is composed of two sprockets mounted face to face on
the ends of the shafts and coupled and connected by a length of chain
wrapped around the sprockets. This type of coupling is made by both
silent-chain and roller-chain manufacturing companies. Small lateral
Fig.16-5. American flexible coupling. (Courtesy of American Flexible Coupling

Company.)
and angular misalignments may be accommodated. The couplings are

enclosed in a case filled with grease.
The universal joint, or Hooke’s coupling, is used to join two intersecting
shafts. Angular misalignments up to 15 or even 30 deg may be accom¬
modated. The angular velocity ratio for two shafts connected by one
universal joint is not constant, but may be made constant by the use of
an intermediate shaft and two universal joints properly mounted.1
A constant-velocity universal joint that utilizes a group of rolling steel
balls as the intermediate member has been developed. Two types are in
use, namely, the Rzeppa2 and the Bendix-Weiss3 universal joints. Shafts
with angles of from 30 to 40 deg may be coupled by a single joint of this
type. Applications include automotive drives and machine tools.
16-4 Flexible couplings with resilient parts. This type of coupling
achieves flexibility by means of resilient parts or members interposed
1 Albert and Rogers, “Kinematics of Machinery,” John Wiley & Sons, Inc., New
York, 1938.
2 The Gear Grinding Machine Co., Detroit, Mich.
3 Bendix Products Corp., South Bend, Ind. See H. Id. Mabie, Constant Velocity
Joints, Machine Design, May, 1948.
SHAFT COUPLINGS 227
between the coupled parts. The resiliency has several functions, such as
allowing for small lateral or angular misalignment, reducing the trans-
missibility of shock loads from one unit to another, and altering the vibra¬
tion characteristics of the connected parts.
Ajax coupling. This coupling shown in Fig. 16-6 is composed of two
halves with one set of pins rigidly attached to one half of the coupling
Fig. 16-6. Ajax flexible coupling.
and the alternate pins attached to the other half. Each steel pin that is
attached to one-half of the coupling projects into a bronze bushing that is
mounted in a rubber sleeve in the other half. The coupling halves are
mounted with axial clearance. A small amount of angular and lateral
misalignment will be taken up by the rubber sleeves.
Falk coupling. This coupling is composed of two slotted members
shown in Fig. 16-7(a) which are connected by a continuous steel spring
9~
Cc r—“htc— -1 - cr
1 L
(C
1
-J L
l» c
AD cm UL Position oit no load
/
(C 1 CT
__;;
« —i i— —r
±_ .n1 i— ■ -- :.l»
—ram
Position at full load
(ct) (b)
Fig. 16-7. Falk coupling.
which lies in the slots. The elements of the spring provide the flexibility
of the coupling.
The sides of the slots in which the spring lies are formed as shown at (6)
in the figure. It may be noted that the effective length of the elements
of the spring changes from a maximum g at no load to a minimum g'
at full load, so that as the torque on the coupling increases the coupling
becomes “ stiff er.” This is a very desirable characteristic for installations
228 MACHINE DESIGN
in which torsional vibration of the shaft is an important consideration.

The coupling parts may be readily disconnected by removing the steel
spring.
The Westinghouse-Nuttall coupling uses helical springs as a resilient con¬
nection between the two halves of the coupling shown in Fig. 16-8. This
Fig. 16-8. Westinghouse-Nuttall coupling.
coupling is used to reduce the transmission of shock loads and to alter

the torsional vibration characteristics of rotating members. This type of
resilient connection is also frequently interposed between the hub and
rim of a gear sprocket or clutch to form a self-contained shock-absorbing
member.
Leather-link coupling. This coupling employs leather links to connect

the driving and driven members, as shown in Fig. 16-9. It should be
noted that one set of links are pulling links for one direction of rotation
while the alternate set are trailing links. In Fig. 16-9 the links a are
pulling links for clockwise rotation. The trailing links are necessary to
prevent backlash in the event of any tendency to reversal of motion, such
as shock loading.
SHAFT COUPLINGS 229
Flexible disk coupling. This coupling is similar in construction to the

leather-link coupling, except that a disk made of leather or rubberized
canvas is used instead of links. This type of coupling has extensive
application in drives transmitting low power.
Lord coupling. The two halves of this coupling, Fig. 16-10, are con¬
nected by a section made of rubber or Neoprene that is bonded to the
halves of the coupling.
16-5 Slip couplings. This coupling is de¬
signed to permit relative rotation, or slip,
between the driving and driven parts at a
predetermined torque for the purpose of pre¬
venting damage to parts because of over¬
loading. There are many variations in form
of this type of coupling. A common form for
medium and large capacities is illustrated in
Fig. 16-11. The flanged hub a is keyed to
Fig. 16-10. Lord coupling. Fig. 16-11. Slip coupling.
the driving member. The driven member b is composed of a flanged hub

with friction disks attached to both sides of the flange. Several springs,
as shown at c, are loaded by tightening the bolts and thus create pres¬
sure on the friction surfaces.
16-6 Fluid couplings. The fluid coupling is composed of a primary,
or input, element called the impeller and a secondary, or output, element
called the runner. The enclosed fluid provides the connection between
the impeller and runner. The impeller and runner are essentially iden¬
tical in construction. The principle of operation is as follows: The fluid,
usually oil, begins its cycle at the inner radius of a cell that it occupies,
as at a in Fig. 16-12(a). When the impeller rotates, the fluid flows out¬
ward toward b because of centrifugal force. Since the runner would be
initially at rest or rotating at a speed lower than that of the impeller, the
centrifugal force of the fluid in the impeller cells which causes fluid flow
in a clockwise direction, as shown, will be greater than the centrifugal
force of the fluid in the runner cells which tends to cause counterflow.
Hence, when the impeller is rotating, there will be fluid flow in a clock¬
wise direction, as shown at (a) in the figure.
230 MACHINE DESIGN
The above discussion accounts for flow of the fluid from the impeller
to the runner and back to the impeller. Now as shown at (6) in the figure,
the fluid at position a will have a tangential velocity Va. When the fluid
arrives at position b, it will have a tangential velocity Vb that is greater
than Va by the ratio of the corresponding radii. A unit mass of fluid
will therefore have gained kinetic energy in flowing from a to b, and this
gain represents the input to the coupling.
The next phase in the operation is the transfer of the kinetic energy,
which the mass of fluid has gained, to the runner as the fluid passes
radially inward through the cells of the runner while its tangential veloc¬
ity is reduced from Vb to Va. The energy thus transferred to the runner
imparts rotation to the driven member.
The difference between the impeller and runner speeds is known as slip
and is usually expressed as a percentage of the impeller speed. For well-
designed fluid couplings, the slip may be as low as 1 per cent at rated
capacity.
In a fluid coupling, the fluid connection between the impeller and run¬
ner results in a marked decrease in transmission of torsional vibrations.
By controlling the quantity of fluid in the coupling, the speed of the
runner may be varied. This feature makes the fluid coupling useful in
speed control. Other applications of this coupling are due to its torque-
speed characteristics.
In order to avoid self-induced vibrations set up by the coupling on
account of the “ register,” or matching, of the vanes in the impeller and
runner, the number of vanes in one half of the coupling is one more or
less than the number of vanes in the other half of the coupling.
CHAPTER 17
CLUTCHES AND BRAKES
17-1 Introduction. In the operation of & friction clutch during engage¬

ment, the initial condition is that the driving member is moving while
the driven member is at rest; and in the final condition both members are
moving at the same speed, i.e., with no relative motion. In the operation
of a brake during braking, the initial condition is that one member, such
as the brake wheel or drum, is moving while the braking member is sta¬
tionary; and in the final condition both members are at rest and have no
relative motion.
231
232 MACHINE DESIGN
It is thus apparent that the principle of operation of both a friction

clutch during engagement and of a friction brake during braking is to
bring two members having relative motion to the state of no relative
motion. The operation of a * clutch is therefore essentially the same as
that of a brake; however, there are structural differences in the two units
because of control requirements and the necessity for providing for heat
absorption or dissipation in brakes.
17-2 Plate clutches and brakes. Plate dutch. In an automotive
type of plate clutch shown in Fig. 17-1, the left hand part, which is the
flywheel, is connected to the engine.
The driven plate is attached to the
hub, which in turn is keyed to the
driven shaft. Pressure on both
sides of the driven plate is pro¬
duced by the coil spring, which acts
through the levers, as shown. To
release the clutch, the sleeve a is
moved to the right, which com¬
presses the spring, removes the load
from the levers, and allows the
driving plates to rotate free of the
driven plate. Note that both sides
of the driven plate are friction surfaces, so that there are two pairs of
friction surfaces for this clutch.
Force analysis. In Fig. 17-2 there is shown a friction disk with outer
and inner diameters, D and d, respectively. The mating disk, not shown,
produces a pressure on the surface because of the axial load P.
Elementary surface area dA = 2tvr dr

Normal force on dA = p2irr dr
Frictional force on dA = fp2irr dr
where p = the surface pressure

/ = coefficient of friction, assumed as constant
Therefore
'D/2
P = 2t f pr dr (a)
d/2
and
'D/2
T = 27r/ / pr2 dr (b)
d/2
Before Eqs. (a) and (6) may be integrated, it is necessary to make an

assumption regarding the distribution of pressure. Two cases will be
considered.
CLUTCHES AND BRAKES 233
1. Uniformly distributed pressure. For new clutches and rigid mount¬

ings, the pressure may be assumed to be uniformly distributed over the
contact area; hence, p may be regarded as constant.
Therefore, from Eq. (a),
p = ^ (£>2 - d2)
and, from Eq. (b),
T = — (£>3 - d3)
24 v '
By eliminating p from Eqs. (c) and (d) and solving for Ty
T =fP(Dl ZiA (17-1)

3 \D2 — dy
where T is the torque for one pair of friction surfaces in contact.

2. Uniform axial wear. An inspection of worn clutch plates reveals
that the plates are not of uniform thickness. To account for the non¬
uniformity of wear, it is necessary to make an assumption regarding wear.
The mechanism of wear is the gradual tearing away by interlocking
(friction) of particles of the rubbing surfaces. Thus friction and wear
are copartners. The rate of wear depends on the friction force at the
surfaces in contact which depends on the pressure, and it depends on the
rubbing velocity and on the hardness of the surfaces. Thus the wear
on an element of area in one revolution depends on the product of fp
and the distance 27rr. If the coefficient of friction is assumed as constant,
the wear is proportional to the product pr. This quantity is proportional
to the work of friction on the unit area; hence, the assumption may be
stated as the normal wear of the surface is proportional to the work of friction.
If n is the normal wear (perpendicular to the surface) and k is the constant
of proportionality, then n = kpr.
In friction clutches and brakes, it is customary to employ a hard
metallic surface such as cast iron for one surface with a replaceable lining
or facing made of softer material with a high coefficient of friction for
the contacting surface. The latter may be composition fiber, compressed
woven asbestos, wood, cork, sintered metal, etc. In the following dis¬
cussion it is assumed that the wear will be concentrated on the softer
material.
In Fig. 17-3(a), the pressure is assumed to be uniformly distributed as
for new plates; hence the normal wear at the outer radius is equal to kpr2
and at the inner radius is equal to kpri. It follows that wear will be
greater at the outer radius; hence the plate will wear until it has assumed
a shape such as the one shown at (b) in the figure. This changed shape
234 MACHINE DESIGN
will result in a redistribution of the pressure, so that pi at the inner radius

will be greater than p2 at the outer radius. Wear will take place in this
manner until the pressure distribution is such that the product p2r2
equals p\r\. Then the wear will progress uniformly along parallel
surfaces.
It is apparent that wear of a friction plate takes place in two stages:
(a) the initial wear as the stable pressure distribution becomes established,
Fig. 17-3. Disk pressure and wear.
which is shown in Fig. 17-3 (c); (6) as the plates wear during the life of
the plates, the surfaces follow the dotted lines as shown in the figure.
When the equilibrium condition is reached,
n = kpr = a constant
and
By substituting the value of p from Eq. (e) into Eqs. (a) and (6), integrat¬
ing, and eliminating the constant C, the equation for the torque becomes
T=f^(D + d) (17-2)
where T is the torque for one pair of friction surfaces in contact.

In Eq. (17-1) which is for new clutches and brakes, the term
1 /D3 - d*\
3 \D2 - d2)
may be called the friction radius. In Eq. (17-2) the friction radius is
+ d). A comparison of these equations shows that the friction
radius for new clutches is slightly larger than for worn-in clutches and the
percentage difference may be expressed in terms of the ratio D/d, as
shown in Fig. 17-4. The ratio D/d for industrial clutches and brakes
and for automotive clutches is of the order of 1.5. For this proportion
the difference in the two equations is very low and is much smaller than
the variation in the value for the coefficient of friction; hence on the basis
of accuracy the choice between the two equations is unimportant. How¬
ever, Eq. (17-2) gives values for the axial force which are on the side of
D/d
Fig. 17-4. Comparison of Eqs. (17-1) and (17-2).
safety, it applies to most of the life of the plates, and it can be more
directly related to space limitations during design, and is therefore recom¬
mended for use rather than the uniform-pressure equation.
Multidisk clutch. Where large torques must be transmitted, a multi¬
disk clutch, as shown in Fig. 17-5, may be used to limit the operating
force. The torque transmitted by this clutch may be determined by

multiplying the right-hand member of Eq. (17-2) by the number of pairs
of friction surfaces in contact.
Disk brakes. Disk brakes are used to a limited extent in some installa¬
tions where speed control is necessary. In Fig. 17-6, there is shown a
236 MACHINE DESIGN
multidisk brake used with a hoisting drum. The member a is keyed to

the shaft b by means of the feather key c, thus causing a to rotate with
the shaft but permitting it to slide axially during engagement and disen¬
gagement. One set of friction disks is splined to the member a, while the
Fig. 17-6. Multidisk brake.
other set slides axially on the bolts d but rotates with the drum. When it
is desired to raise the load, a force on the operating lever slides the member
a to the left and connects it with the drum, and the unit acts as a clutch.
The shaft is prevented from rotating backward by a ratchet. In order to
Fig. 17-7. Cone clutch.
lower the load, the operating force is decreased, which allows the load
to be lowered. During this phase the unit functions as a brake.
17-3 Cone clutches. Cone clutch. In Fig. 17-7 is shown a cone clutch
in which the outer cone a is the driving member and the inner cone b is
the driven member. The compression spring provides the force to engage
the clutch. In order to disengage the clutch, a shifting fork on a collar
that runs in the groove c may be used to slide the inner cone to the right,
thus separating the cones.
Force analysis. The relation between the torque transmitted and the
axial force may be obtained by following a similar procedure to that for
the plate clutch in Art. 17-2. The friction surface of the inner cone of
the clutch is shown in Fig. 17-7(6). The cone-face angle is a, as shown,
and the axial force is P.
2ttr dr
Elementary surface area dA
sin a
p2ier dr
Normal force on dA
sin a
fp2irr dr
Frictional force on dA
sin a
’D/2
P = 2tt I pr dr (a)
d/2
and
[D/2
/ pr2 dr (b)
sin a J /2
From the assumption that “the normal wear is proportional to the work
of friction,” as discussed in Art. 17-1, it may be deduced that
By substituting the value of p from Eq. (c) into Eqs. (a) and (6),
integrating, and eliminating the constant C, the following equation is
determined for the torque transmitted by a cone clutch
fP(D + d)
T = (17-3)
4 sin a
By making the cone angle a equal to 90 deg, the cone clutch becomes a
plate clutch and Eq. (17-3) reduces to Eq. (17-2). It may be noted that
a given torque may be transmitted by a relatively small axial force if the
cone-face angle is decreased. There is a lower limit to the angle a, how¬
ever, since the frictional force that must be overcome in releasing the
clutch increases as a decreases. Thus a clutch with a small cone angle
requires a relatively small force to engage the clutch but a large force for
disengagement. The SAE recommends an angle a = 12.5 deg for cone
clutches faced with leather or asbestos or having cork inserts.
Cone brakes are similar to cone clutches in construction and operation.
238 MACHINE DESIGN
17-4 Block brakes and clutches. In this type of brake or clutch one
or more blocks, or shoes, are forced against a wheel that usually has a
cylindrical surface.
Single block brake. In Fig. 17-8 is shown a single block brake in which
the block attached to the operating lever is forced against the rotating
wheel. The frictional force pro¬
duced by the block on the wheel
will retard the rotation of the
wheel.
Force analysis for friction block
brake. A cylindrical wheel, or
drum, that is assumed to rotate as
indicated is shown in Fig. 17-9.
The block is forced against the wheel by a radial force P. The relation
between the force P and the torque T may be determined as follows:
Let P — operating force on the block in a radial direction
D = diameter of wheel
T = torque on wheel
6 = one-half the angle of contact surface of the block
b = width of wheel
/ = coefficient of friction for materials of block and wheel
p = pressure between block and wheel
D
Elementary area of contact dA = b df)
D
Normal force on dA = p b d<j>
The component of this normal force parallel to P is equal to
dP = P~^b cos </> d<t>
or
+e D Db f+0
P = I p -^ b cos <t> d<j> = ~2~ I p cos </> d<j> (a)
Force of friction on elementary area = fp~^b dcj>

r +e D2
T = /_. fPT bd4> (b)
By assuming that the coefficient of friction is constant, the equation

for T becomes
D2b f + e
T =/— / V (17-4)
4 e —
In order to obtain an expression for the pressure in terms of 4>, it is
necessary to make an assumption, a rational one being that the “normal
wear is proportional to the work of friction.”
In the usual construction of a block brake or clutch, the wear takes
place mainly on the block or lining attached to the block. As wear
occurs, the block or lining will retain the cylindrical shape of the wheel,
as shown in Fig. 17-9(6). The component of wear in the direction of P,
i.e., ab will be constant. Therefore the normal wear ac = ab cos (f>.
(AJ-Wear of block
Fig. 17-9. Single block clutch.
Now the work of friction on an elementary area is proportional to the

pressure p, and hence from the assumption that the normal wear ac is
proportional to the work of friction
p is proportional to ab cos </> or p = C cos </>
where C is a constant of proportionality. From Eq. (a)
CDb f + e
P = ~Y~ / 0 cos2 </> d<f)
= (20 + sin 20)
and from Eq. (6)
T = CjpPc„4,d4,-2Cfl ,(!)’*,«
By eliminating C from the two preceding equations,
240 MACHINE DESIGN
The tangential frictional force on the block may be found by dividing

the torque by the radius, or F = 2T/D; hence,
4 sin 6
F =fP (17-6)
k26 + sin 2d,
6, degrees
Fig. 17-10. 4 sin 0/(20 -f- sin 20) plotted against the semiblock angle 0.
In the above equation, the quantity/[(4 sin 6)/(26 + sin 26)] may be
termed the “equivalent coefficient of friction” and denoted by/'; hence
F = f'P (17-7)
In Fig. 17-10 is shown the variation of the function (4 sin 6)/(26 + sin 26)
with the angle 6.
The forces acting on the operating lever of the single block brake in
Fig. 17-8 are shown in Fig. 17-11
for counterclockwise rotation of
the wheel.
For the equilibrium of moments
about 0 of the forces acting on
the lever, assuming that the re¬
sultant of the frictional forces
may be assumed to be a force F
acting as shown,
Fig. 17-11. Forces on brake lever.
A(a + b) - Pb -Fc = 0 (17-8)
From Eqs. (17-7) and (17-8) the equation for the torque in Fig. 17-11
becomes
fAD(a + b)
T = (17-9)
2(6 + fc)
In Eq. (17-8), the moment Fc is in the opposite direction to the moment

due to the operating force A, and hence the self-induced moment Fc
retards application of the brake.
For clockwise rotation of the brake wheel in Fig. 17-11, the force F
will be reversed and the moment Fc will aid the application of the brake.
Hence, such a brake is known as “self-energizing.” If the distance c is
Fig. 17-12. Double-block spring-set brake. (Courtesy of General Electric Company.)
great enough, the moment Fc will be sufficient to apply the brake without
an appreciable operating force A. This may be considered an advantage,
but such a brake may “grab” and result in uncontrolled braking.
By locating the pivot point above the line of action of F (at O'), the
“self-energizing” feature will be present for counterclockwise rotation.
In Fig. 17-12 is shown one form of a double-block spring-set brake.
Note that the linkage at the top of the brake between the yokes is
arranged so that the compressive force of the spring pulls the yokes
together and applies or “sets” the brake. The floating bell crank at
242 MACHINE DESIGN
the left in rotating clockwise spreads the yokes against the spring force
and releases the wheel. The brake lining is compressed woven asbestos
and is fastened to the cast brake shoes by brass (nonscoring) rivets.
The brake releasing force ôn the bell crank is provided by a General
Electric “thrustor” which is a hydraulic unit operated by the electric
motor shown. This control is especially suitable for large brakes on
account of its smooth operating characteristics. In smaller brakes a
solenoid is suitable. Other releasing devices are magnets, torque-motor-
driven screws, and hydraulic cylinders.
The brake shoes in Fig. 17-12 are pivoted on the yokes to provide
uniform contact between the lining and the wheel. The tangential fric¬
tion force f'P on the brake shoe introduces a moment on the shoe which
makes the distribution of pressure somewhat different from the case in
Fig. 17-9; however, if the moment arm is kept small by locating the
pivot as near as possible to the contacting sur¬
face of the brake lining, i.e., as near the line of
action of the friction force, the turning moment
on the shoe will be small and Eq. (17-7) may be
used. Thus, the force diagram in Fig. 17-13 may
be used without serious error.
An electric device connected in series or shunt
with the motor is an ideal arrangement so that
the brake is released when the motor is started,
Fig. 17-13. Double-block and the brake is applied by the spring when the
brake.
motor switch is opened. In some applications a
time-delay relay is used to allow the moving parts to slow down before
the brake is applied.
In Fig. 17-13, the relation between the spring force S and the torque on
the brake wheel may be determined as follows:
SM0I = S(a + b) + Fic - Pib = 0

2M02 = S(a + b) - F2c - P2b = 0
The relation between F and P may be determined by using Eq. (17-7),

or
El = El = p
Pi P2 J
From the above equations, Fi and F2 may be determined. Then
T = (F1 + F,) - (17-10)
Because of the difference in values of F\ and F2 the wear of the two

brake linings will be unequal if the drum always rotates in one direction.
The unequal wear is usually not serious, since the wear is very slight in
high-quality linings, and also many brakes operate in both directions of

rotation in service, which tends to equalize the wear of the linings of the
two shoes. If the pivot points 01 and 02 in Fig. 17-13 are located on the
lines of action of F1 and F2, the shoes will have equal wear. The effect
of placing the pivot points as shown in Fig. 17-13 is that a smaller spring
force is required to resist a given torque than if the pivot points were
located on the lines of actions of the friction forces F1 and F2.
Fig. 17-14. Double-block clutch.
Double block clutch. The construction of a block clutch is similar to

that of a block brake except that provision must be made for rotating
the blocks and for controlling the load on the blocks. A double block
clutch is shown schematically in Fig. 17-14. The driving plate a carries
the two pivoted levers, to each of which is attached a block. The operat¬
ing forces A are applied by a suitable linkage.
17-5 Band brakes and clutches. Band brakes. A band brake is
shown in Fig. 17-15 in which the band is tightened by means of the force
244 MACHINE DESIGN
A on the operating lever. This band may be made of leather, of canvas

impregnated with rubber, or of a steel band faced with wood blocks.
For clockwise rotation of the wheel as shown, the tension Fi in the
band attached to 0 will be greater than the tension F2. The sum of the
moments of the forces acting on the lever is
Sd/o = F10 T F 2a — A (a T b) =0
The relation between Fi and F2 may be determined from Eq. (14-5) by

deleting the centrifugal force term Fc,
— = efe
F, e
The torque on the brake wheel is given by the following equation
T = (Fi — Ft) |
From the three equations above, the relation between the force on the
operating lever and the torque on the brake
wheel becomes
2 Ta
A = (17-11)
D(a + b)(e'6 - 1)
Band brakes have very extensive use in

machinery. They are simple and depend¬
able and are especially suitable for large
and rugged installations.
Band clutches. A band clutch in which
a conical sleeve a in Fig. 17-16 slides on
the shaft and rotates the lever b which in
turn tightens the band is a type used fre¬
quently in large, rugged equipment, such as that used for road building
and field construction.
Differential band brake. In this type of band brake, the tension in the
band assists in applying the brake, as shown in Fig. 17-17. For this
brake,
A (a + b) — aF2 + cF i = 0
The equation may be reduced to
F1 a
A =
a b ~ Ffe ~ c
In this equation, if the quantity a/efe is less than c, the brake will be
self-locking for clockwise rotation of the wheel. This feature is undesir-
able in a speed-control brake because of “ grabbing,” but it may be used

to advantage in a “ backstop/’ which is a differential band brake used to
permit rotation of a shaft in one direction but to prevent backward
rotation. An application is on a drive for an inclined conveyor to prevent
backward motion in case the power fails while the conveyor is loaded.
The brake is placed on the motor shaft where the torque is the least and
therefore requires a smaller brake than if it were placed on a lower speed
shaft of the drive.
17-6 Energy considerations in brakes. The function of a mechan¬
ical brake is to control the speed of a machine or a moving body by trans¬
forming the mechanical energy of the moving parts into heat energy and
then to dissipate this energy. In the design of a brake it is necessary to

determine the amount of energy to be thus transformed and also the time
allowed for a braking cycle. The latter is necessary so that the tempera¬
ture rise of the brake will not be excessive.
Energy equations. The energy that a brake must absorb is equal to
the difference between the energy given up by the moving parts and the
energy losses in bearings, gears, etc. The latter energy is usually rela¬
tively small.
The equations given below apply specifically to a motor-driven hoist¬
ing drum. A cycle of hoisting and lowering a load, such as that for a
traveling crane, is shown in Fig. 17-18. The exact shape of the accelera¬
tion curve Oa depends on the torque-speed characteristics of the hoisting
motor. The most severe loading on the brake occurs at be as the weight
is brought to rest. The following equations may be used to determine
the required torque capacity for the brake.
246 MACHINE DESIGN
Let Q = weight being lowered, lb

v = initial velocity of load, fps
Et = total energy to be absorbed by the brake, lb-ft
T = torque on brake wheel, lb-ft
t = time of application of brake, sec
N = initial speed of brake wheel, rpm
6 = angle through which brake wheel turns during the time t, rad
a Acceleration
The initial kinetic energy of the load is

Qv2
(a)
2g
The change in potential energy of the load is
Ep = y%Qvt (&)
The initial kinetic energy of rotation of all the rotating parts, such as
the drum, gears, motor rotor, etc., is
WR*ca*
Er =
29
where W = weight of each rotating part, lb

R = radius of gyration of each rotating part, ft
co = angular velocity of each rotating part, rad per sec
The sum of Eqs. (a), (6), and (c) equals the total energy to be absorbed
by the brake:
Et = Ek -f- Ep T" Er (d)
This energy may be equated to the work done by the brake during
the time of its application, or
Et = TO (e)
From Eq. (e),

£ e = 2 X 60Et
6 27rNt
_ 60Et
(17-12)
7rNt
Heat dissipation. The mechanical energy that is transformed into heat

energy by a brake will raise the temperature of the brake. The tempera¬
ture rise depends on the mass of the parts, on the ratio of braking time
to the rest time, and on the heat-dissipation capacity of the brake. The
maximum temperature should be limited to prevent rapid deterioration
of the materials forming the friction surfaces. The maximum tempera¬
ture should not exceed the following values:
Leather, fiber, and wood facing. 150-160 F
Asbestos. 200-220 F
Automotive asbestos block lining. 400-500 F
Since the temperature rise of a brake is difficult to predict, it is usually

satisfactory in the design of brakes to make use of such design coefficients
as foot-pounds of energy absorbed per square inch of surface per minute.
For brakes of similar construction with similar materials for the friction
surfaces, the pressure on the friction surfaces may be used as a guide for
satisfactory temperature conditions. Some values for coefficients of fric¬
tion and allowable pressures are given in Table 17-1.
Table 17-1. Coefficients of Friction and Allowable Pressures*
Coefficient of friction Allowable

Materials in contact pressure,
Dry Greasy Lubricated psi
Cast iron on cast iron. 0.2-0.15 0.10-0.06 0.10-0.05 150-250

Bronze on cast iron. 0.10-0.05 0.10-0.05 80-120
Steel on cast iron. 0.35-0.25 0.12-0.07 0.10-0.06 120-200
Wood on cast iron. 0.35-0.20 0.12-0.08 60-90
Fiber on metal. 0.20-0.10 10^0
Cork on metal. 0.35 0.30-0.25 0.25-0.22 8-15
Leather on metal. 0.5-0.3 0.20-0.15 0.15-0.12 10-40
Wired asbestos on metal. 0.5-0.35 0.30-0.25 0.25-0.20 30-80
Wired asbestos on metal short action 0.25-0.20 200-300
Steel on cast iron, or cast iron on cast
iron, short action. 0.10-0.05 200-300
♦From V. L. Maleev and J. B. Hartman, “Machine Design,” International Textbook Company,

Scranton, Pa., 1954.
The design coefficient mentioned above is usually expressed as the prod¬

uct pV, where p is in pounds per square inch and V equals the rubbing
velocity, fpm.
248 MACHINE DESIGN
In block brakes, the pressure is usually expressed as pounds per square

inch of projected area. Recommended values of pV vary from 30,000
to 80,000, depending on conditions of service and heat dissipation.
The lower value of pV mhy be used for continuous operation in close
surroundings and the higher value for intermittent operation in well-
ventilated locations.
In the design of a block brake it is customary to assume a wheel diam¬
eter which will be in good proportion to the size of the unit to which it is
connected. If the brake is to be attached to a motor shaft, it may be
desirable to have the distance between the center of the brake wheel
above the floor the same as the center of the motor shaft above the floor.
From this dimension a reasonable diameter of brake wheel may be assumed.
After the arrangement of yokes and the location of the spring have
been decided on, the spring force required to set the brake against the
rated torque may be determined as well as the remaining forces on the
shoes and yokes. If PL and PR are then determined, the width of the
shoes may be calculated from the pV value. The ratio of the shoe width
to the wheel diameter b/D may then be determined. This ratio is usually
held between and ^ in well-proportioned brakes.
If the ratio is less than it indicates a shoe so narrow that it may not
be held in good alignment against the wheel and the lining may wear
unequally across the face. If b/D is higher than the lining may also
wear unequally across the face because of the difficulty of maintaining a
uniform pressure across the face because of its width; also the brake may
not be in good proportion to cool properly.
If the tentative assumption of the wheel diameter yields a b/D ratio
outside the limits, the wheel diameter should be changed in the second
trial solution.
In case PL and PR are different, as is usually the case, an average value
may be used in connection with the pV value. The justification of this
is that the pV value is a heat criterion and, owing to the transfer of heat
by the wheel from the shoe that has the larger normal force P to the shoe
with the smaller force P, both shoes will operate at approximately the
same temperature; hence, an average value of the normal forces may be
used.
Example 17-1. The proposed layout of a brake to be rated at 175 lb-ft torque at
600 rpm is shown in Fig. 17-19. The assumed drum diameter D is 8 in.; the angle of
contact for each shoe is 120 deg. The coefficient of friction / may be assumed as 0.3
and for the conditions of service a pF value of 50,000 ft-lb per sq in. of projected area
per minute may be assumed. Determine the spring force S required to set the brake
and the width of shoes.
Solution: Spring force: For the semiangle of contact 6 = 60 deg, the function from
Fig. 17-10 is 1.17; therefore
f = 0.3 X 1.17 = 0.351

or
F
p = 0.351 and P = 2.85 F
The equation for the sum of the moments of the forces on the left-hand yoke about
its pivot point Ol is
2;/ol = 125 -f 2Fl - QPl = 0
= 125 + 2Fl - 6 X 2.85FL = 0
or
Fl = 0.7955
The corresponding equation for the right-hand yoke is
2M0r = 125 - 2Fr - QPr = 0

= 125 - 2Fr - 6 X 2.85FR = 0
or
Fr = 0.6285
For the rated torque,
„ , „ 2T 2 X 12 X 175 _0_ n
Fl + Fr = -jj = -g- = 525 lb
5(0.795 + 0.628) = 525 or 5 = 369 lb
The spring force required to set the brake is 369 lb, but when the spring is com¬
pressed further to release the brake the spring force will be greater than 369 lb.
Shoe width:
Pl + Pr 2.85 (Fl + Fr) 2.85 X 525

= 745 lb
2
250 MACHINE DESIGN
The projected bearing area for one shoe is
6X8 sin 60 = 6.436
where 6 is the width of the shoe. v

The rubbing velocity is
7tD rpm 7r X 8 X 600

V 1,260 fpm
12 12
745 X 1,260
pV 50,000
6.436
or
745 X 1,260
6 2.91 Use 3 in.
50,000 X 6.43
b_
| = 0.375
D O
which is between the limits of and Y; hence the proportions of the shoe width and
wheel are satisfactory.
CHAPTER 18
SPUR AND PARALLEL HELICAL GEARS
18-1 Introduction. As defined by the AGMA, gears are machine ele¬

ments that transmit motion by means of successively engaging teeth.
The gear drive is therefore positive, which gives it an advantage over
friction drives, such as friction wheels and belts.
The first gears had cast teeth and in their day were satisfactory. Mod¬
ern requirements for greater loads and higher speeds have demanded
improvements, however, in tooth forms, cutting methods, materials, and
heat-treatment. Progress by the gearing industry is illustrated by the
performance of railway-motor gears, which operate under severe loading
and adverse lubrication conditions for more than IK million miles with
negligible wear.
In the design of gears for industrial service, some important items that
must be considered are quietness and smoothness of operation, available
gear-cutting equipment, strength of the teeth under static-loading and
under dynamic loading, and satisfactory life. The latter is affected by
wear.
In this treatment, the general relations of the above considerations to
the design of gearing are discussed.
18-2 Spur gearing. General characteristics, (a) The drive is positive
and, with circular gears, the angular velocity ratio is constant. (6) The
center distance may be relatively short, thus making a compact drive,
(c) Provision may be made for shifting gears and in some cases for inter¬
changing them to change the speed of the driven member. (d) The
efficiency is high, since the loss of power may be 1 per cent or less of the
power transmitted. (e) The maintenance of the drive is inexpensive and
the life is long.
Spur-gear terminology. Spur gears have cylindrical pitch surfaces and
operate on parallel axes, and the teeth are straight and parallel to the
axis. The terms most frequently used in spur gearing are given here for
reference (see Fig. 18-1). For a complete discussion, see references on
kinematics of gearing.1
1 C. W. Ham and E. J. Crane, “ Mechanics of Machinery,” 3d ed., McGraw-Hill
Book Company, Inc., New York, 1948; Gear Nomenclature, AGMA Standard 112.02,
1948.
251
252 MACHINE DESIGN
The pitch surface is the surface of the rolling cylinder that the gear may
be considered to replace.
The pitch circle is a right section of the pitch surface.
The addendum circle is the circle bounding the ends of the teeth.
The dedendum circle is the circle bounding the bottom of the spaces
between the teeth.
The addendum is the radial distance between the addendum circle and
the pitch circle.
The dedendum is the radial distance between the pitch circle and the
dedendum circle.
Clearance is the difference between the dedendum of one gear and the
addendum of the mating gear.
Backlash is the difference between the tooth space of one gear and the
tooth thickness of the mating gear measured on the pitch circle.
Circular pitch is the distance from a point on one tooth to the corre¬
sponding point on the adjacent tooth measured on the pitch circle. Its
symbol is p; the units are inches.
Diametral pitch is the number of teeth on a gear per inch of its pitch
diameter.1 Its symbol is P.
Note that the product of the circular pitch and the diametral pitch
equals 7r, i.e., pP = tt.
Tooth forms. It is important that a satisfactory tooth form be chosen
for each application. In general, quietness is favored by the low-
pressure-angle form, but they are weaker than the high-pressure-angle
ones. Common tooth forms are the following:
1 Frequently the term “pitch” is used. This should be understood to mean “dia¬
metral pitch.”
SPUR AND PARALLEL HELICAL GEARS 253
1. AGMA 143/2-deg, full depth; quiet running; less than 30 teeth are
undercut, but from 24 to 30 may be used
2. AGMA standard 20-deg, full depth; stronger than (1) and as low
as 16 teeth without undercutting
3. AGMA standard 20-deg, stub; stronger than (2) and as low as 14
teeth without undercutting
4. Long and short addendum, 14^-deg, full depth; for less than 30
teeth, stronger than (1), quiet running but not interchangeable
5. Long and short addendum, 20-deg, full depth; stronger than (4);
for less than 16 teeth; noninterchangeable
6. AGMA standard composite 143^-deg; may be cut with milling
cutters
7. Fellows standard 20-deg stub
Design considerations. The specifications for a gear drive generally
include (a) the horsepower to be transmitted, (6) the speed of the driving
gear, and (c) the speed of the driven gear or the velocity ratio. Fre¬
quently the center distance is specified. The usual drive requires a
speed reduction, for example 2 to 1, which means that the speed of the
driving gear is twice that of the driven gear. The reason that most drives
are reductions rather than increases is that the usual sources of power,
i.e., motors, turbines, and high-speed engines, operate at speeds higher
than those required by driven units, such as machine tools, pumping or
blowing equipment, transportation and propulsion machinery, or con¬
veying equipment. Occasionally, however, speed increases are required.
In the design of a gear drive there are several requirements that must
be met, as follows: (a) The gear teeth should have sufficient strength so
that they will not fail under static loading, such as that at high starting
torques, or under dynamic loading during normal running conditions.
(6) The teeth should have good wear characteristics so that their life will
be satisfactory, (c) The use of space and material should be economical.
(d) The alignment of the gears and deflections of the shafts must be con¬
sidered because of their effect on the performance of the gears, (c) The
lubrication of the gears must be satisfactory. The above general require¬
ments indicate that the final specifications for gears involve many inter¬
related factors. Compromises are usually necessary in order to secure a
desirable balance between good service and minimum cost.
18-3 Strength of gear teeth—Lewis equation. The determination of
the maximum stresses in a loaded gear tooth is complicated by the varia¬
tion in magnitude and direction of the load on the tooth during contact
and by the shape of the tooth, since it has varying width and is joined to
the body of the gear by a fillet. At the first point of contact, the load
Wn in Fig. 18-2 acts normal to the profile of the tooth in accordance with
the fundamental law of gear-tooth action. This normal load may be
254 MACHINE DESIGN
resolved into tangential and radial components; the tangential component

producing a bending moment on the tooth, and the radial component
inducing compressive stresses on sections across the tooth. Owing to
the radial component being eccentric to the center line of the tooth, the
compressive stress is not uniformly distributed across the tooth.
As the point of contact moves along the profile, the magnitude of Wn
changes as well as the moment arm of the bending force. Furthermore,
along the line of contact, the number of pairs of teeth which share the
transmitted load may vary according to the contact ratio. Finally,
because of inaccuracies of tooth spacing and tooth form and because of
deflection of the tooth under load, dynamic forces may be induced.
These dynamic forces are cyclic and may in some cases be larger than the
steady force which transmits power.
As to the gear tooth itself, it is not a simple cantilever because of its
stubbiness and nonuniformity of width. Also stress concentration at
the fillets enters the scene.
It can readily be appreciated from the above situation that the max¬
imum stresses in an existing gear tooth are difficult to determine. The
designer of gears has also to consider that at the beginning of a design, the
sizes of the gears and the teeth are not usually established so it becomes
necessary to approach the design of gears in a somewhat indirect manner.
It should be appreciated in addition that fracture of gear teeth is not
the only mode of failure to be considered, but that wear of the teeth is
also a criterion; in fact more gears fail by wearing of the teeth than by
fracture.
In 1892, Wilfred Lewis made simplifying assumptions1 regarding the
strength of gear teeth which resulted in an equation which has been used
extensively by industry in determining the size and proportions of gears.
1 Wilfred Lewis, Investigation of the Strength of Gear Teeth, Engineers’ Club of
Philadelphia, 1892.
In this investigation, Lewis assumed that the worst position of loading

of the tooth occurred at the first point of contact and that one pair of
mating teeth carried the entire load. As shown in Fig. 18-3, the normal
load Wn was translated to the center line of the tooth and then resolved
into a radial component Wr and a tangential component IF. The tan¬
gential component IF produces a bending moment on the tooth. In
addition, the radial component Wr was neglected.
Lewis then assumed that the gear tooth could be regarded as a can¬
tilever and made use of the principle that a beam of parabolic outline is
one of uniform strength at all sections. He then inscribed within the
tooth a parabolic outline drawn through o and tangent to the tooth pro¬
file at the fillets. Since the actual tooth is stronger than the parabolic
Fig. 18-3. Forces on gear teeth.
beam at every section except TC, this is the critical section of the tooth
and the flexure formula may be applied to establish a relation between the
load IF and the stress of this section.
The equation for W will be used here as a basis for determining the
proportions of gear teeth. After the proportions are established, the
Buckingham equations will be used for determining the hardness for
fatigue-strength and wear requirements.
Using the face width of the gear as F in Fig. 18-3(6),
sFt2
Wh = (18-1)
IT
from which
sFt2
W =
6h
The circular pitch may be introduced as follows:

256 MACHINE DESIGN
where
= f_
^ 6/ip
The quantity y is known'as the “form factor.” To determine its
value, t, h, and p may be determined analytically or scaled from a drawing
similar to Fig. 18-3(6). Note that if the gear is enlarged, the distances
t, h, and p will each be increased proportionally; hence the value of y will
remain unchanged. The value of y then is independent of the size of the
tooth and depends only on the number of teeth on a gear and the system
Table 18-1. Lewis Form Factors y
External gears Internal gears
Number of 14^-deg 20-deg 20-deg 20-deg full-depth system

teeth composite full-depth stub
system system system Pinion Internal gear
a = 1/P * a = 1/P a = 0.80/P a = 1/P a = 1/P
10 0.055 0.064 0.088 0.103

11 0.062 0.072 0.093 0.104
12 0.067 0.078 0.099 0.104
13 0.071 0.083 0.103 0.104

14 0.075 0.088 0.108 0.105
15 0.078 0.092 0.111 0.105
16 0.081 0.094 0.115 0.106

17 0.084 0.096 0.117 0.109
18 0.086 0.098 0.120 0.111
19 0.088 0.100 0.123 0.114
20 0.090 0.102 0.125 0.116
21 0.092 0.104 0.127 0.118
22 0.093 0.105 0.129 0.119

24 0.095 0.107 0.132 0.122
26 0.098 0.110 0.135 0.125
28 0.100 0.112 0.137 0.127 0.220

30 0.101 0.114 0.139 0.129 0.216
34 0.104 0.118 0.142 0.132 0.210
38 0.106 0.122 0.145 0.135 0.205

43 0.108 0.126 0.147 0.137 0.200
50 0.110 0.130 0.151 0.139 0.195
60 0.113 0.134 0.154 0.142 0.190

75 0.115 0.138 0.158 0.144 0.185
100 0.117 0.142 0.161 0.147 0.180
150 0.119 0.146 0.165 0.149 0.175

300 0.122 0.150 0.170 0.152 0.170
Rack 0.124 0.154 0.175
* a = addendum; P = diametral pitch.

of the teeth. A convenient method for determining the value of y for a

given tooth profile is discussed by M. A. Durland (see Appendix XII).
Values for some standard tooth forms are given in Table 18-1.
The value of W in Eq. (18-2) may be used as the torque on a gear
divided by its pitch radius. This gives a value somewhat greater than
W in Fig. 18-3, but the difference is relatively small for gears of usual
proportions.
In deriving Eq. (18-1), it was assumed that the load was uniformly
distributed across the face of the gear. The actual distribution in service
depends on the accuracy of cut of the gears, on the initial alignment and
on the deflections of the tooth, on the shafting, and on the bearings.1
For ordinary industrial installations, the face width F is usually made
from two to five, or preferably, between three to four times the circular
pitch for spur gears.
An extensive investigation of a wide variety of industrial gears that
have behaved satisfactorily in service was made by the use of the Lewis
equation and values for the allowable stress determined (see Table 18-2).
Table 18-2. Values of Basic Stress for Gears
Material s0) psi
Cast iron, ordinary. 8,000
Cast iron, medium grade. 10,000
Cast iron, highest grade. 15,000
Cast steel, 0.20% C untreated. 20,000
Cast steel, 0.20% C heat-treated. 28,000
Forged carbon steel
SAE 1020 casehardened. 18,000
SAE 1030 untreated. 20,000
SAE 1045 heat-treated. 30,000
Alloy steel
SAE 2320 casehardened. 50,000
Bronze SAE 62. 10,000
Phosphor bronze SAE 65. 12,000
Meehanite metal, grade GA. 12,500
Rawhide, fabroil. 6,000
Bakelite, Micarta, celoron. 8,000
If these allowable stresses are in turn used with the Lewis equation to
design similar industrial gears, the resulting gear should have the size
and proportion to behave satisfactorily in service.
1 See Poritsky, Sutton, and Pernick, Distribution of Load along a Pinion, Trans.
ASME, vol. 67, p. A-78, 1945.
258 MACHINE DESIGN
As to dynamic stresses and wear, these can be allowed for by alterations

of the composition of the material and its heat-treatment.
Thus, the sins committed in the assumptions for the derivation of the
Lewis equation affect the allowable stress, and in using these allowable
stresses in similar designs, the sins are neutralized, and thus the designer
may establish a pair of gears of reasonable size and proportions.
The gears may then be investigated for dynamic loads, wear of the
teeth, AGMA ratings, and other criteria, as discussed later in this chapter.
While most of these equations are relatively easy to apply to an existing
set of gears, they are not of such form that they can readily be used for
direct design. The Lewis equation in its original form, or modified for a
Fig. 18-4. Velocity factor.
special class of gears or service, is a most helpful aid in this phase of gear
design.
The value of the allowable stress in gear design depends of course on
the material for which a basic stress s0 may be determined. In order to
make allowance for the dynamic effects due to the velocity of tooth action,
the Barth equation may be used, i.e.,
Sail = SoCv (18-3)
where saii = allowable stress, psi

s0 = basic stress, psi (see Table 18-2)
Cv = velocity factor = 600/(600 + V) for spur gears
V = pitch-line velocity, fpm
The variation of the velocity factor with the pitch-line velocity is
shown in Fig. 18-4. The solid line shows the usual range for spur gears.
There are other velocity factors which are frequently used and which
like the Barth equation are arbitrary. They are not given here, since the
Lewis equation will be used only to determine the proportions of gears,
i.e., diametral pitch, pitch diameters, and face, for sufficient strength, and
Eq. (18-3) is satisfactory for that purpose.
A shortcoming of the velocity factor as given is that it does not take

into account the fact that the pinion teeth are loaded a greater number of
times than the gear teeth. The number of cycles is important in regard
to the endurance limit. A method for considering the relative number of
cycles has been proposed by Merritt.1 This method includes the influ¬
ence of the number of hours per day of service but does not include the
effect of velocity.
In the design of a gear drive, Eq. (18-2) may be used as such, but a
solution by trial is necessary. The equation may be placed in more con¬
venient forms for design purposes by introducing the diametral pitch P.
Solving Eq. (18-2) for the induced stress gives
W
Sind = ™— (18-4)
Fpy
Let T = torque to be transmitted, lb-in.

D = pitch diameter, in.
n = number of teeth
P = diametral pitch
F
k = — = 3 to 4 for ordinary service
Substituting for F and p in Eq. (18-4) gives
WP2
Sind = — (18-5)
kirzy
Equation (18-5) is a convenient form of the Lewis equation for the

solution of problems in which the center distance and the velocity ratio
are specified.
In many designs, the center distance is not specified, and Eq. (18-5)
may be further transformed by substituting for W, as follows:
2T 2 TP
W =
D n
Hence, Eq. (18-5) becomes
2TP3
Sind (18-6)
kTT2ny
Equation (18-6) is a convenient form of the Lewis equation for the

solution of problems in which the center distance is not specified. Since
there is not a unique solution, the final design will depend on the initial
assumptions. The quantities to assume should be those for which judg¬
ment is most easily exercised. For this reason the number of teeth on
1 H. E. Merritt, “Gears,” p. 254, Sir Isaac Pitman & Sons, Ltd., London, 1943.
260 MACHINE DESIGN
the pinion is a desirable quantity to assume since it affects quietness of

operation markedly. By using an average value of k equal to 3.5, the
induced stress can be expressed in terms of the diametral pitch. Then
by comparison with Eq. (18-3), a probable diametral pitch may be estab¬
lished. This procedure is shown by the example below.
Values for common diametral pitches are given in Table 18-3.
Table 18-3. Common Standard Diametral Pitches
1 2 5 9 16
ix 2^ 6 10 18
IX 3 7 12 20
m 4 8 14 24
If the gear is to be keyed to a shaft, the maximum permissible bore

should be investigated. An equation for this purpose was determined by
photoelastic means and has given satisfactory results:1
d = D(0.50 + 0.0344 y/n - 12) (18-7)
where d = maximum permissible bore

D = pitch diameter
n = number of teeth
The equation allows for a standard square key with the center line of
the key way in line with the center line of a tooth space. This is the best
location for the keyway. In addition it applies to gears having 12 to 24
teeth. For gears having more than 24 teeth, use n equal to 24 in the
equation.
Example 18-1. It is required to determine the proportions of a spur-gear drive to

transmit 10 hp from a shaft rotating at 1,170 rpm to a low-speed shaft with a speed
reduction of 3 to 1. Assume that the teeth are 20-deg stub system with 24 teeth on the
pinion. The pinion is to be SAE 1045 and the gear SAE 1030 steel. Assume that the
torque at starting is 150 per cent of torque at rating.
Solution: To determine whether the pinion teeth or the gear teeth are weaker, a
consideration of Eqs. (18-2) and (18-3) shows that the product sQy for the pinion and
gear determines the relative strength.
So n y s0y
Pinion. . 30,000 24 0.132 3,960
Gear. . . 20,000 72 0.158 3,160
The values of say show that the gear is the weaker, and hence will be the basis for the
design. Using 150 per cent rating,
1 P. H. Black, An Investigation of Relative Stresses in Solid Spur Gears by the

Photoelastic Method, Univ. Illinois Eng. Expt. Sta. Bull. 288, 1936.
63,030 hp 63,030 X 10 X 1.5
To 2.425 lb-in.
rpm 1,170/3
2 TP3 2 X 2,425P3
Sind ~ j ~ 12.35P3
kir2ny 3.5tr2 X 72 X 0.158
Now
600 600
Sail ~ So 60Q + y 20,000
600 + V
In the above two equations there are two unknowns, the diametral pitch P and the
pitch-line velocity V. From the above equation for the induced stress
Sind = 4,240 for P = 7

= 6,330 for P = 8
= 9,000 for P = 9
Note from Fig. 18-4 that the velocity factor varies from 0.5 to 0.2 for the usual range
of spur-gear pitch-line velocities. By using tentatively an average value of, say,
for CVJ then the tentative allowable stress will be approximately X 20,000, or say
7,000 psi. By comparison with the above values for the induced stress, it is seen
that the diametral pitch may be 8 or 9. Using a trial value of P equal to 9, then
D = 2.67 in.
7tD rpm 7r X 2.67 X 1,170
V = = 816 fpm
12 12
600 600
Sail — So = 20,000 8,480 psi
600 + V 600 + 816
Thus P = 9 may be satisfactory, since the induced stress is near the allowable stress.
The computed face width equals
kp = k — = 3.5 — = 1.22 in.

y P 9
This face width may now be altered in order to reduce the induced stress to the allow¬
able stress, as follows:
1.22 — = 1.22 = i.29 in.

Sail 8,480
To use a commercial dimension, F = 1}^ in. would be satisfactory.

Note that the computed face width may be decreased by the ratio Sind/saii in case the
induced stress is less than the allowable stress. In any event, the face width should
not be less than 3p or greater than 4p.
Results:
Dp = 2.67 in.
Dg = 8.00 in.
F = 1.25 in.
Center distance = 5.33 in.
18-4 Strength of gear teeth—Buckingham equation. In the preceding

article the velocity factor was used to make approximate allowance for
the effects of dynamic loading. The dynamic loads are due to (a) inac¬
curacies of tooth spacing, (6) irregularities in tooth profiles, and (c) deflec¬
tions of the teeth under load.
262 MACHINE DESIGN
A closer approximation to actual conditions may be made by the use

of equations based on an extensive series of tests,1 as follows:
Wd = W ± Wi (18-8)
where Wd = the total dynamic load

W = steady load due to transmitted torque
Wi = increment load due to dynamic action
Fig. 18-5. Dynamic loads on gear teeth.
In Fig. 18-5(a) there is shown a representation of Eq. (18-8), where Wi

is shown as a cyclic load. In many cases, the value of Wi is greater than
W as shown at (6) in the figure where e-f indicates that the teeth separate.
The increment load Wi depends on the pitch-line velocity, the face
width, the material of the gears, the accuracy of cut, and the tangential
load IF. For average conditions of mass of the gears and connected parts,
the following equation applies:
0.05V (FC + IF)
Wd = W + (18-9)
0.05F + VFC~VW
where Wd = maximum dynamic load, lb

IF = steady transmitted load, lb
1 See the manual of gear design, “Design Loads on Gear Teeth,” by Earle Bucking¬
ham, The Industrial Press, New York, 1931.
F = width of face of gears, in.

C = deformation factor (Table 18-6)
The value of the deformation factor C depends on the error in action
between teeth, the class of cut of the gears, the tooth form, and the mate¬
rial of the gears.
“Error in action” refers to the maximum deviation of the actual tooth
profile from the ideal profile. The deviation is due to inaccuracies in
cutting the profiles and in spacing of the teeth, and also the deflections of
the teeth under load. Tables 18-4 and 18-5 may be used to estimate the
maximum error in action. The values of the deformation factor C may
be determined from Table 18-6. These tables are taken from the manual
of gear design, as cited, and apply to gears having average conditions of
mass, elasticity, accuracy, and pitch-line velocity.
Table 18-4. Maximum Allowable Error in Action between Gears
V Error V Error V Error
250 0.0037 1,750 0.0017 3,250 0.0008

500 0.0032 2,000 0.0015 3,500 0.0007
750 0.0028 2,250 0.0013 4,000 0.0006
1,000 0.0024 2,500 0.0012 4,500 0.0006
1,250 0.0021 2,750 0.0010 5,000 and over 0.0005
1,500 0.0019 3,000 0.0009
Table 18-5. Maximum Expected Error in Action between Gears
Diametral pitch P Class 1 * Class 2 f Class 3 t
1 0.0048 0.0024 0.0012

2 0.0040 0.0020 0.0010
3 0.0032 0.0016 0.0008
4 0.0026 0.0013 0.0007
5 0.0022 0.0011 0.0006
6 and finer 0.0020 0.0010 0.0005
* Class 1—well-cut commercial gears,

f Class 2—gears cut with great care.
t Class 3—ground and lapped gears.
If the dynamic load as determined above is substituted in Eq. (18-2),
the calculated stress should not exceed the flexural endurance limit se/.
To provide a margin of safety, the factor / should be introduced:
Se/ = (18-10)
The factor/ has the following values:
/ = 1.25 for steady loads

= 1.35 for pulsating loads
= 1.50 for shock loads
264 MACHINE DESIGN
Values of flexural endurance limits are given in Table 18-7.
Table 18-6. Values of Deformation Factor C

\
Materials, Tooth Error in action, in.
pinion and gear form 0.0005 0.001 0.002 0.003 0.004 0.005
Cast iron and cast iron 400 800 1,600 2,400 3,200 4,000
Steel and cast iron • 143^ deg 550 1,100 2,200 3,300 4,400 5,500
Steel and steel / 800 1,600 3,200 4,800 6,400 8,000
Cast iron and cast iron 20 deg 415 830 1,660 2,490 3,320 4,150
Steel and cast iron ► full 570 1,140 2,280 3,420 4,560 5,700
Steel and steel depth 830 1,660 3,320 4,980 6,640 8,300
Cast iron and cast iron 20 deg 430 860 1,720 2,580 3,440 4,300
Steel and cast iron stub 590 1,180 2,360 3,540 4,720 5,900
Steel and steel tooth 860 1,720 3,440 5,160 6,880 8,600
Table 18-7. Fatigue Limits of Gear Materials
Flexural Surface
Material BHN endurance endurance
limit sef limit sea
Gray cast iron. 160 12,000 90,000
Semisteel. 200 18,000
Phosphor bronze. . . . 100 24,000
Steel.. .. 150 36,000 50,000
200 50,000 70,000
240 60,000 88,000
280 70,000 103,000
300 75,000 110,000
320 80,000 118,000
350 85,000 130,000
360 90,000
400 100,000
GA Meehanite. 80,000
18-5 Wear of gear teeth—Buckingham equation. The maximum load

that gear teeth can carry without premature wear depends on the radii
of curvature of the tooth profiles and on the elasticity and the surface-
fatigue limits of the materials. The research work of Buckingham has
led to the following equation for estimating the maximum or limiting load
for satisfactory wear of gear teeth.
Ww = DFKQ (18-11)
where Ww = limiting load for wear, lb

2r
Q = ratio factor = (for external gears)
r + 1
2r
= ratio factor = (for internal gears)
r — 1
r = velocity ratio
K = load-stress factor
D = pitch diameter of pinion, in.
ses = surface endurance limit, psi
4> = pressure angle
F = face width of gears, in.
Ei, E2 = moduli of elasticity of materials, psi
s2es sin </>

K =
1.4
Table 18-8. Values of Load-stress Factors K
Surface
K, K,
Material Material endurance
Bhn Bhn 14^-deg 20-deg
in pinion in gear limit,
systems systems
Ses in psi
Steel. 150 Steel. 150 50,000 30 41

Steel. 200 Steel. 150 60,000 43 58
Steel. 250 Steel. 150 70,000 58 79
Steel. 200 Steel. 200 70,000 58 79
Steel. 250 Steel. 200 80,000 76 103
Steel. 300 Steel. 200 90,000 96 131
Steel. 250 Steel. 250 90,000 86 131
Steel. 300 Steel. 250 100,000 119 162
Steel. 350 Steel. 250 110,000 144 196
Steel. 300 Steel. 300 110,000 144 196
Steel. 350 Steel. 300 120,000 171 233
Steel. 400 Steel. 300 125,000 186 254
Steel. 350 Steel. 350 130,000 201 275
Steel. 400 Steel. 350 140,000 233 318
Steel. 400 Steel. 400 150,000 268 366
Steel. 150 Cast iron. 180 50,000 44 60
Steel. 200 Cast iron. 180 70,000 87 119
Steel. 250 Cast iron. 180 90,000 144 196
Steel. 150 Phosphor bronze. 100 50,000 46 62
Cast iron. 180 Cast iron. 180 90,000 193 264
GA Meehanite. ... GA Meehanite.. . • • • 80,000 105 144
Steel. ... GA Meehanite.. . ... 80,000 90 123
266 MACHINE DESIGN
Values of K for various tooth forms and materials are given in Table
18-8. In the wear equation, Eq. (18-11), the value of Ww should not be
less than the dynamic load W d. In gear design, Wd from Eq. (18-9)
may be taken as the minimum allowable value of Ww and when used as
such in Eq. (18-11) gives the minimum value of load-stress factor K. The
required Brinell hardness numbers may be determined from Table 18-8.
In using the equations for static loads, dynamic loads, and wear loads,
it should be realized that the results are not exact but serve for comparison
between expected performance of gears being designed and that of the
gears of the reported tests.
Example 18-2. Determine the Bhn for the pinion and gear for the data of Example
18-1 required on the bases of (a) dynamic load and (b) wear load. Assume class 2
gears.
(a) Dynamic load:
For 150 per cent rating
0.05F(FC + W) Auxiliary Calculations
Wd = W +
0.05F + VFC + IF _ 33,000 hp
40.8 X 2,756 W ~ V '
= 606 +
40.8 + 52.5 33,000 X 10 X 1.5
= 1,811 lb 816
_ =fWj = 606 lb
ef Fpy 0.05F = 0.05 X 816
1.25 X 1,811 = 40.8 fpm
1.25 X 0.349 X y C = 1,720 for class 2 gears
FC = 1.25 X 1,720 = 2,150 lb
= 39,200 psi for pinion
= 32,800 psi for gear FC + IF = 2,150 + 606
= 2,756 lb
From Table 18-7
VFC + W = 52.5
Bhn for pinion = 175
Bhn for gear = 150 7T
0.349 in.
9
(6) Wear load:
2r 2X3
For 100 per cent rating, the dynamic load Wd may be Q = 1 +r
found equal to 1,544 lb. 1 +3
Ww 1544
K = = 309
DFQ 2.67 X 1.25 X 1.5
From Table 18-8, Bhn for pinion = 400
Bhn for gear = 350
Since the latter two values are higher than are required on the endurance-limit basis,
they should be specified. However, the possibility of heat-treating the pinion and
gear to obtain these hardness numbers should be investigated. With the grade of steel
originally specified, it would be difficult to harden either the pinion or the gear through¬
out, but through hardening may not be necessary. A form of surface hardening would
in general be sufficient to obtain the required wear resistance.
18-6 Parallel helical gears. Because of gradual engagement of the

teeth, helical gears run quieter than spur gears and may be operated at
pitch-line velocities up to 10,000 fpm and over. High pitch-line velocities

mean low tooth loads, which in turn promote quietness and high efficiency.
Because of end thrust, the helix angle xp (see Fig. 18-6) is usually limited
in single-helical gears to 15 deg. For smooth operation, one end of a
tooth should be advanced a distance over the other end by a distance at
ip - He Hx. anq/e
F- Face
p - Circular pitch in plane of rotation
= Normal circular pitch
Fig. 18-6. Parallel helical gears.
least equal to the circular pitch, that is, at least equal to p/tan xp in Fig.
18-6. The AGMA specifies a minimum overlap of 15 per cent by stating
that the face width of a single helical gear shall be at least 1.15p/tan xp.
When helical gears of opposite hand are mounted in pairs on a shaft,
or when herringbone gears are used, the axial loads are balanced, and
tooth
Fig. 18-7. Herringbone gears.
hence larger helix angles may be used to secure greater pitch overlap and
the accompanying increase in quietness and strength over single-helical
gears. For herringbone gears as shown in Fig. 18-7, helix angles are used
from 20 to 30 deg for industrial gears and up to 45 deg for turbine drives.
The minimum face width for herringbone gears as recommended by
AGMA is 2.3p/tan xp. For both single- and double-helical gears, the
maximum width of face should be from 1.5 to 2 times the pinion diameter,
268 MACHINE DESIGN
although 2.5 has been used.1 In the latter case the mountings were very
rigid and accurately aligned.
Strength of parallel helical gears. According to the AGMA, a modifica¬
tion of the Lewis equation, Eq. (18-2), may be used for helical gears:
W = S°FP?C'1 (18-12)
where W = tangential tooth load, lb

s0 = basic stress from Table 18-2
p = circular pitch in the plane of rotation, in.
y = form factor; the pressure angle is in the plane of rotation
F = face width, parallel to axis, in.
C = wear and lubrication factor; for enclosed gears, continuously
lubricated with proper oil, C = 1.15; for scanty lubrication,
regularly inspected, C = 1.25: for indifferent lubrication,
C = 1.35
1,200
Cv for accurately cut and ground gears
1,200 + V
78
cv for hardened steel, ground and lapped gears
78 + y/V
The relation between the pressure angle </>n in the normal plane and <j>
in the plane of rotation is
tan (j)n = tan <f> cos \[/
The surface durability of the teeth may be determined by using a modi¬

fication of the Buckingham equation, Eq. (18-9),2 or the AGMA standard
rating.3
The dynamic load on helical gears may be estimated from an extension
of Eq. (18-9) by substituting F cos2 \p for F and by multiplying the last
term (increment-load term) by cos \p. The wear-load equation, Eq.
(18-11), should be divided by cos2 ^ for helical gears.4
18-7 Additional considerations in rating of gears. Profile contact ratio.
In accurately cut gears of high contact ratio, more than one pair of teeth
1 W. P. Schmitter, Determining the Capacity of Helical and Herringbone Gearing,

Machine Design, June and July, 1943.
2 C. D. Albert, “Machine Design Drawing Room Problems,” 4th ed., p. 382, John
Wiley & Sons, Inc., New York, 1951.
3 AGMA Standard Rating, Surface Durability of Helical and Herringbone Gears,
AGMA Standard 211.01-1944.
4 See also AGMA Standard Ratings, Strength of Helical and Herringbone Gear
Teeth, AGMA Standard 221.01-1948, and Surface Durability of Helical and Herring¬
bone Gears, AGMA Standard 211.01-1944.
may be counted on to share the load so that the worst loading on a tooth
may not occur at the first point of contact, as assumed in the derivation
of the Lewis equation, but instead may take place between that point
and the pitch point. For this condition,1 the inscribed parabola will be
shorter than shown in Fig. 18-3 and the form factor will be greater than
the value in Table 18-1. In critical gear applications, such as aircraft
gears, marine gears, and stationary geared power-plant drives, it may be
desirable to consider this more refined procedure, instead of the simple
Lewis method, and to include also stress concentration and other influ¬
ences, as discussed in the remaining sections of this article.
Stress concentration. The following equations2 for stress-concentration
factors may be used to determine the magnification of maximum stress
at the fillets. In the equations r is the fillet radius and the other notation
is the same as in Fig. 18-3.
For 20-deg involute teeth
and for 14J^-deg involute teeth
K = 0.22 +
Connected loads. In high-speed gears in which the rotating parts have

large moments of inertia, as in a motor or turbine with large and heavy
rotors driving a propeller, flywheel, generator, or any mass with large
WR2, the driving and driven units tend to resist changes in velocity due
to irregularities in the gear teeth and to the transfer of load from one pair
of teeth to the next. Large dynamic loads may be induced in this case.3
Allowable stresses. When the refinements of contact ratio, stress con¬
centration, and dynamic loads as affected by tooth errors in action and
by connected masses are considered, the allowable stresses in the teeth
to prevent failure by fracture may be higher than the stresses used with
the simple Lewis equation. It is not possible to state general rules for
determining these allowable stresses since the requirements for each class
of application of gears are unique, so that the safe procedure in selecting
allowable stresses is by experiment and experience. A correct equation
for allowable stresses would also include a term to express the degree of
necessity of the manufacturer to avoid any failure of the gears.
1 See D. W. Dudley, “Practical Gear Design,” p. 44, McGraw-Hill Book Company,
2 T. J. Dolan and E. I. Broghamer, A Photoelastic Study of Stresses in Gear Tooth
Fillets, Univ. Illinois Eng. Expt. Sta. Bull. 335, 1942.
3 See E. Buckingham, “Analytical Mechanics of Gears,” Chap. 20, McGraw-Hill
270 MACHINE DESIGN
For instance, the failure of a gear in a juke box may not cause more than
inconvenience until it is replaced; in fact, the result of such a failure may
bring pleasure to some listeners.
The failure of a gear in an automobile would probably not endanger
life or cause extensive damage; yet if many gears failed in any one make
of automobile it would jeopardize the reputation of the manufacturer.
The failure of a gear in an industrial speed reducer would generally
not have fatal results and may not be excessively expensive to repair.
It might, however, result in expensive shutdown.
The failure of any one of most gears in an aircraft would have good
chances of causing a serious accident.
A failure in the main propulsion gear of an ocean-going ship may cause
a disaster or at best may necessitate an extended layover in drydock with
the necessity, if it were a bull-gear failure, of cutting through several steel
decks or the hull to remove the ruined gear and replace it with a new one.
To the loss in revenue or service should be added the cost of manufac¬
turing a new gear and drydock charges of the order of $1,000 per hour.
Thus, owing to the widely varying requirements in applications of
gears and to the many modes of failure, the choice of allowable stresses
requires careful consideration.
Scoring. In heavily loaded gears which are operated for a long period
of time, pits may develop due to excessive compressive stresses at the
region of contact between the teeth. The pits may become large so that
a considerable area is affected which may lead to scoring of the teeth.
Scoring is characterized by radial scratches and may be associated with
other types of wear, such as abrasion.
A formula that has been used by automotive-gear designers to evaluate
scoring is known as the PVT formula in which P is the Hertz contact
pressure in pounds per square inch, usually calculated for the first and last
points of contact on the pinion tooth profile and taking into account the
contact ratio. The sliding velocities at the points where P is calculated is
represented by V and expressed in feet per minute. The distance in
inches along the line of action from the point where P is calculated to the
pitch point is represented by T. Values of PVT to prevent scoring are
of the order of 1J4 million.
Another method for determining limits to avoid scoring and which
may be a better criterion than the PVT formula uses the “ flash tempera¬
ture” at the contacting surfaces of the teeth. An equation for the flash-
temperature limit to prevent scoring,1 known as the Kelley formula,
includes terms for the rolling velocities of the tooth surfaces in contact,
the coefficient of friction, and the rms value for surface finish after the
gears have been run in.
1 B. W. Kelley, A New Look at Scoring Phenomena of Gears, AGMA 219.04,
October, 1952.
AGMA ratings. These ratings expressed in horsepower cover strength

of spur, helical, and bevel gears and surface durability of spur, helical,
bevel, and worm gears. There are also AGMA thermal-horsepower
capacity ratings for enclosed helical-gear speed reducers.
18-8 Nonmetallic gears. In installations where it is desirable to
reduce noise and vibration or where the loads are light and long life is a
factor, the pinion of a pair of gears may be made of nonmetallic materials,
for example, rawhide, fabroil, molded nylon, or a laminated phenolic,
such as textolite, Micarta, bakelite, or celoron. Rawhide and fabroil
gears are usually assembled by compression between two steel plates.
Rawhide is especially good for shock loading. To avoid overheating
rawhide gears, the pitch-line velocity should normally be limited to 1,750
or 2,000 fpm with a maximum of 2,500 fpm. Since rawhide is affected
by most oils, a paste made of flake graphite and linseed oil forms a good
lubricant. High temperature should be avoided.
The laminated phenolic materials are not affected by water, oil, or mod¬
erate temperature. Pitch-line velocities up to 3,000 fpm are conservative
for industrial installations with commercial standards used in cutting.
With precision-cut gears well mounted and balanced, pitch-line velocities
up to 5,000 fpm may be used safely.
The nonmetallic pinions are generally run with cast-iron gears and
they may be designed by the use of the Lewis equation, Eq. (18-2), and
Eq. (18-3) where
(18-13)
18-9 Failure of gear teeth. Gear teeth may fail by breaking as a

result of static load, shock load, or fatigue load, or they may fail by
excessive wear. The latter is the usual type of failure, especially in gears
that operate continuously. The general types of wear of machine parts,
as discussed in Art. 20-7, may be expanded to apply to gear teeth as
follows:1
Normal wear Galling

Initial pitting Burning
Progressive pitting Rolling
Abrasion Cracking and checking
Scratching Chipping
Scoring Gouging
Normal wear occurs when the gears are new, but it should cease after
the gears are “run in” under proper conditions.
1 See AGMA Standard Nomenclature, Gear Tooth Wear and Failure, AGMA
110.02, 1951.
272 MACHINE DESIGN
Initial pitting may not be serious but if it leads to progressive pitting

the gear may fail (see Figs. 3-16 and 18-8). This is a common type of
failure of gears. Abrasion is generally due to foreign matter in the
lubricant.
Fig. 18-8. Example of pitted gear teeth. (Courtesy of Westinghouse Electric Cor¬
poration.)
The above list gave 12 types of failure which may occur singly but
which generally appear in combination. They may be caused by or
aggravated by such conditions as the following:
Incorrect center distance Excessive temperatures

Shaft misalignment or deflection Excessive speeds
Improper lubrication Vibration
Excessive loads Poor surface finish
18-10 Bearing loads. The three mutually perpendicular components

of the tooth load on a parallel helical gear are Fx, Fy, and Fz, as shown
in Fig. 18-9. These components may be determined as follows*
p _ 2T 33,000 hp
D V
Fv = Fx tan ^
Fz = Fx tan c/>
where T = torque transmitted, lb-in.

D = pitch diameter of gear, in.
i = helix angle
4> = pressure angle
To determine the loads on the bearings A and B, the type of support

for the shaft must be known. If the shaft is relatively short with long
sleeve bearings, a distributed load must be assumed. If the shaft is rela¬
tively long with short sleeve bearings or ball bearings, the shaft may be
assumed to be simply supported.
The loads on the bearings A and B may be determined by considering
the components Fx, Fy, and Fz separately. Note that the bearing reac¬
tions due to Fy and Fz are in the plane of the left-hand view in the figure,
while the reactions due to Fx are perpendicular to that plane.
If there is an “ overhanging load,” such as Fo, because of a driven gear,
sprocket, or belt, its magnitude may be determined; and, if the direction
is known, it can be combined with the bearing reactions caused by FX)
Fy, and Fz. If the direction of F0 is not known, the most unfavorable
direction should be assumed, i.e., in the same direction as the resultant
of the reactions due to Fx, Fy, and Fz. In general, the maximum loads
that it is possible to impose on bearings A and B should be determined
regardless of the direction of rotation of the gears.
Overhanging load. The maximum overhung load which a customer
may apply to the stub end of the shaft of a speed reducer depends on the
type of drive and on the minimum size of sprocket, gear, or sheave which
is keyed to the shaft. The following equation may be used to estimate
the load:
126,060if X hp
(18-14)
D X rpm
where F0 overhung load, lb

K load factor; for chain drive, K = 1, for gear, K = 1J4, for
V belt, K = 1)^2, and for flat belt, K = 2}/%
hP rated horsepower
rpm shaft speed
D diameter of sprocket, gear, or sheave, in.
274 MACHINE DESIGN
If D is not known, a value equal to twice the shaft diameter may be

used. The location of the line of action of F0 may be determined on the
assumption of its being located one shaft diameter beyond the shaft seal,
or at the mid-point of the shaft extension, whichever is the greater dis¬
tance from the adjacent bearing.
18-11 Lubrication of gears. In the discussion of the lubrication of
journal bearings in Art. 21-2, it is shown that in order to have fluid-film
lubrication exist between two surfaces having relative sliding motion it is
necessary for the film to be converging. A pair of gear teeth coming into
action provides a converging film so that fluid-film lubrication is possible.1
Since the surfaces of mating gear teeth are both convex, the width of the
band of the load-carrying film is small, and the pressures developed will
be very high as compared with pressures in a sleeve bearing. In heavily
loaded gears, the thickness of the film may become so small that the film
is ruptured by the irregularities of the tooth surfaces. In this event,
boundary lubrication will exist and special “extreme-pressure’1 lubricants
should be used. Oils of low viscosity should be used for high speeds
and light loads, and oils of high viscosity should be used for low speeds
and heavy loads.
Grease and oil are the lubricants most frequently used for gears. Oil
is generally considered better but it requires an oiltight housing or case
with shaft seals. The same oil may be used for the lubrication of the
gears and the bearings in many cases. When the oil is circulated by a
splash system, the gear should dip into the oil about 1 in. A greater sub¬
mergence may cause churning of the oil and an excessive temperature rise.
18-12 Parallel-shaft speed reducers. These standard units are manu¬
factured by many companies for general industrial service and whenever
possible they should be used rather than a special unit. Frequently,
however, gear reducers must be “built in” the machinery of which they
are a part.
The standard units may be classified as follows:
1. Number of reductions, such as single, double, triple; single reduc¬
tions may be had generally from 1:1 to 8 or 10:1; double reductions from
8:1 to 30 or 45:1
2. Vertical or horizontal (this refers to the plane in which the shafts
are located)
3. Type of gears
4. Type of bearings, such as sleeve, ball, straight roller, or tapered roller
5. External gears, internal gears, or cyclic gear train
6. Geared motor reducer (these units are part of the driving motor)
1 E. K. Gatcombe, Lubrication Characteristics of Involute Spur Gears, Trans.
ASME, vol. 67, p. 177, 1945.
In addition to the above types there are also worm-gear reducers,

bevel-gear reducers, and variable-speed changers.
A single-reduction, single-helical horizontal speed reducer using roller
bearings is shown in Fig. 18-10 and a herringbone reducer in Fig. 18-11.
Service factors. The AGMA recommends the following service factors,
Table 18-9, for application to enclosed reducers. It is necessary that the
reducer have a capacity at least equal to the actual horsepower multiplied
by the service factor.
Fig. 18-10. Section of helical-gear speed reducer. (Courtesy of Timken Roller Bearing
Company.)
Thermal horsepower rating. When a speed reducer is operated, its tem¬

perature will rise until the housing attains a temperature above ambient
temperature sufficient to dissipate the heat corresponding to the power
loss of the unit. If the temperature of the housing is limited in order to
maintain proper lubrication, then the horsepower transmitted is limited
and is known as the “thermal capacity’’ of the unit. The AGMA
specifies a thermal rating in terms of the center distance and the face
width of the gears. For operation at periods of less than 2 hr with long
rest periods, the thermal rating may be ignored, but for operating periods
of more than 2 hr, the transmitted horsepower should not exceed the
thermal horsepower rating.
276 MACHINE DESIGN
Fig. 18-11. Single-reduction herringbone-gear speed reducer, lower view with top half
of housing removed. (Courtesy of DeLaval Steam Turbine Company.)
Table 18-9. Service Factors
Character of load on
driven machine
Prime mover Duration of service
Moderate Heavy
Uniform
shock shock
Electric motor .... 8 to 10 hr per day 1.00 1.25 1.75

24 hr per day 1.25 1.50 2.00
Intermittent, 3 hr per day 0.80 1.00 1.50
Occasional, 3^ hr per day 0.50 0.80 1.25
Multicylinder in- 8 to 10 hr per day 1.25 1.50 2.00

ternal-combustion 24 hr per day 1.50 1.75 2.25
engine Intermittent, 3 hr per day 1.00 1.25 1.75
Occasional, 3^2 hr per day 0.80 1.00 1.50
Single-cylinder in- 8 to 10 hr per day 1.50 1.75 2.25

ternal-combustion 24 hr per day 1.75 2.00 2.50
engine Intermittent, 3 hr per day 1.25 1.50 2.00
Occasional, 3^ hr per day 1.00 1.25 1.75
For a single-reduction, helical-gear speed reducer, the thermal-horse-

power rating (AGMA Standard 420.02-1951) is as follows:
hp = 3.265C \/F - 0.483C1-6F
where C = center distance, in.

F = face width, in.
CHAPTER 19
GEARS FOR NONPARALLEL SHAFTS
19-1 Types of gears. Refer to Fig. 19-1. A brief description of some

gears is given in the following list:
1. Straight bevel gears. The axes intersect, generally at right angles.
The elements of the teeth are straight lines that converge at the apex of
the pitch cone.
• i i
^ A
2. Zerol bevel gears. The axes intersect, and the teeth are curved and
can be ground.
3. Spiral bevel gears. The axes intersect, and the teeth are cuved and
oblique.
4. Face gears consist of a spur or helical pinion in combination with a
conjugate gear of disk form.
5. Skew bevel gears. The axes are nonparallel and nonintersecting, and
the teeth are straight.
6. Hypoid gears. The axes are nonparallel and nonintersecting, and
the teeth are curved.
7. Crossed helical gears. The axes are nonintersecting and are at any
angle. The teeth have the same or opposite hand (see Fig. 19-8).
278
GEARS FOR NONPARALLEL SHAFTS 279
8. Worm gears include worms and their mating gears. The axes are
usually at right angles, as in Fig. 19-12.
In Fig. 19-2 are shown the pitch cones for various relations of axes in
the same plane from the external gear to the internal gear.
19-2 Bevel-gear terminology. An axial section of a pair of right-angle
bevel gears is shown in Fig. 19-3. The shaft angle 2 equals 90 deg,
T is the pitch angle, ro is the face angle, and TR is the root angle. D is
the pitch diameter, and Do is the outside diameter. The addendum is a,
and the dedendum b is measured at the large ends of the teeth. The pitch
cones are represented by oef and ofg and the back cones by jef and
hfg. In the figure, the subscripts P and G refer to the pinion and gear,
respectively.
fbJ-Acute CcJ-Right-angle
bevel,.£<90° bevel,£=90#
(ct)-Miter gears fe>0biuse (f)-Crown (g)- Interna I

velocity ratio bevel 2 >90° £ = 180°
I:I £ = 90° 2>90°
Fig. 19-2. Relation of axes for gears where 2 is angle between shafts.
19-3 Strength of bevel-gear teeth. Distribution of load along teeth. A

half section of a bevel gear is shown in Fig. 19-4(a), the large end of a
tooth at (b), and a top view of the tooth at (c). Note that the elements
of the tooth converge at the apex of the pitch cone, and hence the dimen¬
sions of the tooth at any section are proportional to the distance x of the
section from the apex of the pitch cone 0.
The dotted lines at (c) show the deflected tooth under the load. If
the tooth does not distort under load, the elements of the deflected tooth
pass through the apex 0. Hence, the deflection of an element dx of the
tooth under the load dW will be proportional to x.
The deflection of the element of the tooth as a cantilever is
dWh3 !2dWh3
° ~ 3El ~ 3Et*dx {}
where h and t are the effective height and thickness, respectively, of the
tooth at the section at x (see Fig. 18-3). Since the ratio of h and t for
MACHINE DESIGN
Mounting distance
Fig. 19-3. Bevel-gear dimensions.
every section of the tooth is constant, and since 5 is proportional to x,

it follows that
dW
Cx (19-2)
dx
where C is a constant. Equation (19-2) states that the variation of the

load W with respect to x is proportional to x; hence the tooth load is
greater at the large end of the tooth, as shown in Fig. 19-4(d).
The maximum bending stress at the base of the elementary section of
the tooth is
6h dW 6h ~
s = luF = F Cx
or since h and t are proportional to x, it follows that the stress s is constant
along the base of a bevel-gear tooth.
Strength of bevel-gear teeth. From the Lewis equation, Eq. (18-2), the
load dW on an elementary section of a bevel-gear tooth may be expressed
as
dW = s dx pxy (19-3)
where px = circular pitch of the element of the tooth

y = Lewis form factor
The moment of dW about the axis of the gear is equal to Rx dW and its
integral equals the torque T transmitted by the gear. Hence
dT = Rx dW = syRxpx dx (19-4)
Since Rx and px may be expressed in terms of the length of the pitch-

cone element L, and the circular pitch p and the radius R, both measured
at the large end of the tooth, Eq. (19-4) becomes
T = sy^R f x2 dx (19-5)
L JL-F
= sF-pyR (l - j + fjf) (19-6)
By dividing the torque T by the radius R, the equivalent tangential

force W acting at the large end of the tooth may be found, thus
T ( F F2\
W = ^ = SFvy[l-l + m) (19-7)
In bevel gears, the ratio of width of face to the pitch-cone element,

F/L, is usually less than hence the last term in the parentheses of the
above equation is small compared with the other terms and may be
282 MACHINE DESIGN
neglected, and hence Eq. (19-7) becomes
W = sFpy 1 - 0 (19-8)
Note that a spur gear may be considered as the limiting case of a bevel
gear for which L is infinitely long. For a spur gear the ratio F/L is zero
and Eq. (19-8) becomes identical with Eq. (18-2).
The tangential force W in Eq. (19-8) may be determined by dividing
the torque on the gear by the radius R. The force W is not an actual
force but an equivalent force acting at the large end of the tooth for use
in designing the tooth and specifying it at its large end. The strength
of the tooth at other sections will be equal to that at the large end, as
discussed earlier in this article.
In the use of Eq. (19-8) for design purposes, the value of the allowable
stress s may be taken from Eq. (18-3) and Table 18-2.
For satisfactory operation of bevel gears, the ratio F/p should be
between 2 and 3, and the ratio F/L should not exceed The form
factor y should be based on the formative number of teeth, i.e., the number
of teeth on a gear whose radius equals the back-cone radius.
19-4 Resultant tooth force. For determining the loads on the shafts
of bevel gearing, it is necessary to use the actual resultant tooth force
that acts at some point along the tooth face rather than the equivalent
force at the large end of the tooth, as discussed in Art. 19-3. Since the
tooth is loaded heavier at the outer end, the resultant force W0 acts
between the mid-point of the tooth and the outer end.
From Eq. (19-2),
dW = Cx dx
or
W„ = C \ xdx = £ [L2 - (L - F)2] (19-9)

Jl-f *
Referring to Fig. 19-4, the moment of dW may be taken about 0 as
dM = xdW
or, using Eq. (19-2) again,
dM = Cx2 dx
or
M = C x2 dx = % [L3 - {L - F)3] (19-10)

J L—F 3
The distance of the point of application of the resultant load W0 from

the apex 0 is found by dividing M by WQ; hence
L3 - (L - Fy
(19-11)
L2 — (L — F)2
The radius R0 corresponding to L0 equals
D _ Lo p
rCo -K (19-12)
By combining Eqs. (19-11) and (19-12),

F F2 x
1 “ L + 3L2
R0 = R = ZR (19-13)
x-L
2L
The resultant tooth load W0 is equal to the torque on the gear divided
by R0, or
W0 = -T (19-14)
tC0
This is the tangential component of the actual tooth load. The sep¬
arating component equals TF0 multiplied by the tangent of the pressure
angle.
The values of Z in Eq. (19-13) for various values of F/L are shown in
Fig. 19-5.
n
<D
O
l_
CD
CL
Fig. 19-5. Comparison of exact and approximate radius factors for bevel gears.
From the geometry of Fig. 19-4, it may be shown that the radius corre¬
sponding to the mid-section of the tooth is
Rm = R (19-15)
The values of 1 — (F/2L) are plotted in Fig. 19-5, and also the per¬
centage of difference between Z and 1 — (F/2L). For bevel gears it is
evident that the error in considering that the resultant tooth load acts at
284 MACHINE DESIGN
the mid-point of the face is small, and therefore Eq. (19-15) may be used,
except for very precise determinations of shaft loads. In these cases the
effect on location of tooth bearing of inaccuracies, deflection of teeth, shaft
and mounting, friction, and "other deviations from the ideal case should
be included.
19-5 Gleason system of bevel gears. As a result of considerable the¬
oretical and experimental investigation as well as field experience, the
Gleason Works has developed a system for generating bevel gears that
has given excellent results. In this system the pressure angle and the
addendum depend on the numbers of teeth in the pinion and gear.
Fig. 19-6. Bevel-gear mountings.
19-6 Mounting of bevel gears. The shafts for bevel gears and the
bearing supports should be sufficiently rigid so that the deflections at the
teeth under load will not be excessive. Also, the gears should be properly
adjusted when they are installed. Only by fulfilling the two requirements
of mounting and adjustment, and by providing proper lubrication, can
bevel gears be made to operate satisfactorily in respect to quietness and
life.
The bearings should be designed with due consideration of the thrusts
along the axes. In spiral bevel gears, the directions of the thrusts depend
on the hand of the spiral, the direction of rotation, and whether the gear
is the driver or driven member.
A pair of bevel gears mounted in plain bearings are shown in Fig.
19-6(a). A ball bearing mounting with the pinion straddle mounted is
shown at (6). Note that the teeth of the gear are cut on a steel ring that
is attached to a separate center.
A typical aircraft bevel-gear mounting is shown in Fig. 19-7. In a

mounting of this type, the distance between the pinion bearings should
be at least 2.5 times the overhang x.
Fig. 19-7. Aircraft bevel-gear mounting. (Courtesy of Gleason Works.)
19-7 Crossed helical gears. The most common arrangement of crossed

helical gears is with the shafts at right angles, as shown in Fig. 19-8.
The contact between teeth is, kinematically, point contact. Because of
the elasticity of the material the contact point becomes a small area.
Because of the high tooth loads and the small contact area the contact
286 MACHINE DESIGN
pressure is very high and wear is comparatively rapid. Hence, these

gears are suitable for transmitting light power at moderate speeds. The
efficiency is low, seldom more than 85 per cent.
The velocity ratio of any type of toothed gearing is equal to the inverse
ratio of the numbers of teeth, hence
co2 fti Di cos i/q

(19-16)
COi 712 D2 cos \J/2
where co = angular velocity

n = number of teeth
D = pitch diameter
\p = helix angle
(for subscripts, refer to Fig. 19-8).
Crossed helical gears are used to drive camshafts and auxiliaries on
small internal-combustion engines, feed mechanisms on machine tools,
and other units that require small power.
19-8 Worm gearing. This type of gearing is extensively used to trans¬
mit power at high velocity ratios between nonintersecting shafts that are
generally but not necessarily at right angles. Velocity ratios as high as
300:1 or over are being used in a single reduction. Worm drives may
be used as speed increasers, but their usual use is as speed reducers, in
which the worm is the driver and the worm wheel the driven member. If
the driven machinery has large inertia and if there is possibility of the
driving power being cut off suddenly, the worm drive must be reversible,
or a slip coupling or other device must be placed between the driven unit
and the worm wheel to prevent damage to the drive. When the lead
angle of the worm is greater than the friction angle of the surfaces in con¬
tact, the worm is called reversible, or overrunning, i.e., the worm can be
freely driven by the worm wheel. However, under & fluctuating load, any
worm drive may be reversible in the sense that it may creep backward,
as when a worm is used as a brake to hold a load. A similar condition
appears in the loosening of bolts under fluctuating axial loads, which was
referred to in Art. 9-2.
The speed ratio of a worm drive is equal to the number of teeth on the
worm wheel divided by the number of threads on the worm. From one
to four threads are commonly used on worms, i.e., single-, double-, triple-,
and quadruple-threaded worm.
19-9 Terminology and construction. Single- and double-threaded
worms are shown in Fig. 19-9. The axial pitch p is equal to the circular
pitch of the mating worm wheel. Note that the lead L of an n-threaded
worm is equal to n times the pitch; hence the lead and the axial pitch
are equal in a single-threaded worm. The pitch lead angle X is the angle
between the tangent to the pitch helix and the plane of rotation. The
pitch helix angle is the angle between the tangent to the pitch helix and
an element of the pitch cylinder and is the complement of the lead angle.
The pressure angle is (3. The AGMA states the following standard axial
fol-Single thread ^-Double thread
Fig. 19-9. Worm dimensions.
pitches: 34, %6, H, 34, %, Y±, 1, 134, 1/4, 1%, an(i 2 in. Pressure angles
are as follows:
head angle A Pressure angle (3, deg
* To 12 deg... 14^
To 20 deg. 20
To 25 deg. 22^
Over 25 deg. 25
Reversible drives. 30
The development of the pitch helix is shown at (c) in Fig. 19-9; this
may be used to determine the relation
between X, L, and the pitch diameter
of the worm d, that is tan X = (L/ivd).
The normal lead L0 is equal to L cos X.
The straight worm shown in Fig.
19-9 is the type in most common use.
The Hindleg worm, Fig. 19-10, is
used to some extent, but it requires
extremely accurate alignment.
Various types of worm-wheel rim
construction are shown in Fig. 19-11.
The form at (a) is cut with a form cutter and is suitable for light service.
288 MACHINE DESIGN
The types at (b), (c), and (d) may be hobbed, the latter being a separate
ring for attachment to a cast-iron or steel spider.
Materials. The worm is generally made of steel while the gear is made
of bronze or of cast iron for light service. Some recommended combina¬
tions of materials are given in Table 19-1.
19-10 Design of worm gearing. In the design of a worm drive, the

quantities usually specified are the horsepower, speed, velocity ratio, and
service. In some cases the exact or the approximate center distance is
specified, the former in a replacement drive, and the latter in a drive
where the center distance may be limited by space conditions. If the
designer is limited in the choice of the center distance, it is usually impos¬
sible to achieve the highest efficiency.
Table 19-1. Worm and Worm-wheel Combinations
Service Worm Worm wheel

Light. SAE 1040 Cast iron or SAE 63
Medium. SAE 2320 or SAE 3120 Phosphor bronze
(casehardened)
Heavy.. . Molybdenum or SAE 65
chrome-vanadium
steel (hardened)
In addition to the center distance, the quantities that must be deter¬

mined are the lead angle, the lead, and the number of threads on the
worm. In addition, consideration must be given to the materials of the
worm and worm wheel as affecting strength and wear of the teeth, sizes
of the shafts as affecting their strength and stiffness, mounting of the
shafts, design of a substantial housing of sufficient heat-dissipating capac¬
ity, and lubrication. Most of the above items are related, and many of
them are not susceptible to exact analysis; hence it is apparent that the
design of a worm drive demands the judicious use of theory and experi¬
mental and field data.
Proportions of worm drive. To determine a satisfactory combination of

lead angle, lead, and center distance, the following method may be used.1
The center distance C may be expressed in
terms of the worm and worm-gear diameters
d and D, and it may be further expressed in
D
terms of the axial lead L, the lead angle X, and
the velocity ratio r as follows (see Fig. 19-12):
'7
c = *±° = — (cot X + r) 1 ■■
c
2 271-
testI±
In terms of L0, the above equation becomes
u
C
L0
1 1
27r \sin X
+ )
cos X/
—
X
(19-17) Fig.
gear.
19-12. Worm and worm
Since the velocity ratio r is specified, Eq. (19-17) contains three vari¬
ables C, L0, and X. The right-hand member may be evaluated for various
velocity ratios and the results plotted as in Fig. 19-13.
In Fig. 19-13 the low point on each of the curves represents the lead
angle which corresponds to the minimum value of C/L0. This minimum
1 Courtesy of Thomas P. Colbert, University of Wisconsin, and O. A. Leutwiler,
University of Illinois.
290 MACHINE DESIGN
value represents the minimum center distance that can be used with a
given lead, or likewise the maximum lead that can be used with a given
center distance.
Since the numbers of threads that may be used with various velocity
ratios have been fairly well established by common use (Table 19-2), and
since axial pitches have been standardized (Art. 19-9), it is possible to
determine the combination of lead angle, lead, center distance, and diam¬
eters to suit specifications. The following example illustrates the
procedure.
Table 19-2. Worm Data
Velocity ratio Number of threads
20 and over Single
12-36 Double
8-12 Triple
6-12 Quadruple
4^10 Sextuple
Example. Determine the pitch diameters, number of threads, and axial pitch for a
worm drive having a velocity ratio 14:1 and a center distance of approximately 6 in.
Solution: From Fig. 19-13, the lead angle X is 22 deg and the value C/L0 is 2.83.
Therefore
C 6
L0 = 2.12 in.
2.83 2.83
Lo _ 2.12
L = = 2.29 in.
cos X 0.927
From Table 19-2, assume triple threads, that is, n = 3. Then
0.763 in.
The nearest standard axial pitch is 24 in. Then
L = np = 3 X 0.75 = 2.25 in.

Lo = L cos X = 2.25 X 0.927 = 2.09 in.
C = 2.83L0 = 2.83 X 2.09 = 5.91 in.
2.25
d = = 1.774 in.
7r tan X 7r X 0.404
D = 2C-d = 2X 5.91 - 1.77 = 10.05 in.
Strength of teeth. Since the teeth of a worm gear are weaker than those
of the worm threads, the design for strength may be based on the Lewis
equation, Eq. (18-2), applied to the worm-gear teeth. The velocity
factor Cv may be taken as 1,200/(1,200 + V), where V is the pitch-line
velocity of the worm gear in fpm. Form factors given in Table 18-1 for
full-depth teeth may be used.
Since the Lewis equation gives the strength of a single tooth, it may be
desirable to consider the number of teeth in contact. The number of
teeth in contact may be determined by dividing the angle of action of the

teeth by the angle subtended by the circular pitch.
Wear of worm gears. Failure at the surfaces in contact is the most
prevalent cause of failure of worm gears. The failure may be prevented
or satisfactorily delayed by using proper materials and hardness and finish
of the surfaces, maintaining bearing pressure within allowable limits and
providing correct lubrication. Because of the lack of a standardized
method for determining the load-limit for wear of worm gears at the
present writing, further treatment will not be given here.1
Thermal rating of worm gears. The amount of heat that must be dis¬
sipated in a worm drive is equal to the power loss at the gears, i.e.,
Hi = 33,000 X hp X (1 — efficiency) (19-18)
where Hi = power loss, ft-lb per min.

This heat must be dissipated by the gearbox or removed by other
means, such as artificial cooling in the form of cooling coils in the housing
or external oil coolers. The dissipation of heat from the housing depends
on the difference in temperature between the surface of the housing and
the surrounding air, i.e., heat-dissipation capacity H2 in foot-pounds pei
minute is
H 2 — KA{ti — t2) (19-19)
where A = the housing area, sq ft

ti = the temperature of the lubricant, deg F
t2 = the ambient temperature, deg F
K = the coefficient of heat dissipation, ft-lb per min per sq ft per
deg F
The maximum temperature of the lubricant should not exceed 180 F.
Values for K times square feet may be estimated from Fig. 19-14.2
19-11 Forces on worm gears. The force components on a worm are
similar to those on a power screw, since the axial force on a screw corre¬
sponds to the axial thrust on the worm, and the tangential force on the
screw and on the worm are similar. In addition, the worm drive is sub¬
jected to separating forces that tend to force the worm and worm wheel
out of mesh. The three force components on the worm are shown in Fig.
19-15 and may be determined as follows:
1 H. Merritt, Worm Gear Performance, Proc. Inst. Mech. Engrs. {London), vol. 129,
p. 127, 1935; see H. Merritt, “Gears,” Sir Isaac Pitman & Sons, Ltd., London, 1943.
2 V. M. Faires, “Design of Machine Elements,” p. 270, The Macmillan Company,
New York, 1941; H. Walker, The Thermal Rating of Worm Gear Boxes, Proc. Inst.
Mech. Engrs. {London), vol. 151, p. 326, 1944.
292 MACHINE DESIGN
„ w n 2 X torque on worm
tangential force t z = --
Fz
Axial force on worm Fv — --7
tan X
Separating force Fx = Fy tan 0
The forces on the worm gear are equal in magnitude but opposite in
direction to those shown in the figure.
Fig. 19-14. Heat dissipation of worm gearboxes.
Fig. 19-15. Forces on worm teeth. The forces on the worm-gear teeth are in the
opposite direction.
The loads on the worm-shaft bearings may be readily determined by

taking moments of the forces Fx, Fy, and Fz and the overhanging load F0
due to a gear, chain, or belt. For determining overhanging shaft loads,
the AGMA recommends the following factors: chain, 1; gear, 1%; V belt,
13^2; flat belt, 2. If the direction of rotation of the worm and the direc¬
tion of the overhanging load are not specified, their directions should be
assumed to give the maximum possible loading on the bearings for use
in selecting or designing the bearings. The loads on the worm-gear shaft
bearings may be similarly determined.
To determine the efficiency of a worm drive, Eq. (12-2) may be used
for approximate results, i.e.,
. tan X
Efficiency = -~~ (19-20)
tan (X + (pi)
where X = the lead angle
<t>i = the friction angle (see Fig. 19-16 and Table 19-3 for values of
tan </>i)
Equation (12-2) was derived for square-threaded screws; hence it does
not strictly apply to worms having pressure angles other than zero
degrees. The error in using the above equation will be slight, however,
for worms of usual pressure angles. The values of the coefficient of fric¬
tion in Fig. 19-16 include allowance for bearing friction and may be used
for design purposes. Figure 12-3 shows the variation of efficiency with
the lead angle. Note that the maximum efficiency is secured with a lead
angle of 45 deg, but the variation in efficiency is not great for lead angles
of 25 to 60 deg. Care must be exercised in design in this instance since
the loss in efficiency, i.e., (100 — efficiency), represents the heat that must
be dissipated by the gearbox. The difference between 98 and 97 per cent
in efficiency is about 1 per cent, but the difference in loss in efficiency is
50 per cent; hence 50 per cent more heat must be dissipated by a gearbox
for a 97 per cent efficient drive than for one of 98 per cent efficiency.
19-12 Mounting of worm gears. The requirements for mounting
worm gears are the same, in general, as for other types of gears except
that they are more exacting for worm gears. Since the worm-gear shaft
is of relatively large size and its bearings are generally close together, its
deflection is usually not excessive. The worm shaft, however, is smaller
in diameter because of the low torque it transmits and because the size of
the worm shaft should be a minimum because of its great influence on
over-all size of the drive. Thus the relatively small worm shaft and the
necessary large distance between its bearings make its deflection relatively
large.
Table 19-3. Correction Factors for Coefficient of Friction in Fig. 19-16
Multiply values
For Working with
in chart by
Casehardened steel. . . Phosphor bronze 1.0

Cast iron. Bronze 1.15
Cast iron. Cast iron 1.33
Hardened steel. Aluminum 1.33
Steel. Steel 2.0
294 MACHINE DESIGN
It is necessary to limit the deflection of the shafts to avoid excessive

heating at the teeth. For this reason, the shafts should be stiff, the bear¬
ings well supported, and the housing rigid. It is important also to mount
the shafts so that a rise in temperature will not cause the shaft to buckle,
which would in turn cause more heating and result in failure of the teeth.
Well-mounted worm and worm gears are illustrated in Figs. 22-12 and
22-20.
19-13 Concluding discussion. From the considerations of gears in
Chaps. 18 and 19, it is striking that there are so many types of gears,
Fig. 19-16. Coefficient of friction for worm gears (see Table 19-3). (From Merritt,
“Gears,” Sir Isaac Pitman & Sons, Ltd., London, 1943.)
classes of applications, and modes of failure. The development of the

many types of gears has been necessitated by the different requirements
in the various fields, as the automotive, aircraft, marine, industrial, and
other fields. As an illustration of a special application of gears, a differ¬
ential gear drive for a Gravure press is shown in Fig. 19-17. The input
shaft is connected to the main way shaft of the press and there may be
as many as six or more of these differential gearboxes each coupled to the
main way shaft. The output shaft extends through the differential
assembly as shown, and the only gear keyed to it is the 35-tooth helical
gear. The 50-tooth hypoid gear and the differential spider carrying the
adjusting worm gear and the 22-tooth and 23-tooth helical gears all float
on the output shaft, which is supported on bearings outside the gears.
The adjusting gear and the worm (not shown) are used to justify the rolls
to ensure aligned copy.
As to modes of failure, there are three common ones, namely, fracture
of the teeth, excessive wear of the teeth, and failure of the lubricant.
Fracture of the teeth may be due to large shock loads and/or to exceed¬
ing the fatigue life of the material (see Fig. 3-14). This type of failure is
not common.
Differential adjusting worm gear
Output shaft
HT. hypoid pinion
50 T. hypoid gear-
input shaft
Fig.19-17. Differential gear drive for Gravure press. (Courtesy of Kidder Press
Company.)
Wear is a cumulative process and may be satisfactorily delayed by

proper design of the teeth, including limiting the contact pressure and
using proper lubrication. A lubricant is coming to be classed as a mate¬
rial of engineering the same as steel, and rightly so. As an example, the
wrong grease in the differential housing of an automobile would probably
lead to failure of the hypoid gears within a few blocks of travel, and the
automobile would be immobilized the same as if the drive shaft broke.
The temperature of the lubricant for gears may be controlled by cooling
by convection of heat from the gear housing, using cooling fins, or by the
use of forced air or fluid cooling if necessary.
296 MACHINE DESIGN
The AGMA standards are available for strength and durability ratings
for spur, helical, and herringbone gears, straight and spiral bevel gears,
for the durability of cylindrical and double-enveloping worm gears, and
for the thermal ratings of enclosed helical and herringbone, bevel, and
worm-gear units. The standards are used for rating of gear drives.
They are precise and should be read carefully including, like legal papers,
the fine print.
CHAPTER 20
SURFACE FINISH, FRICTION, AND WEAR
20-1 Surface irregularities. The deviations of the actual surface of a

body from its nominal surface may be classified as follows:1 (1) Waviness,
as characterized by more or less regular and recurrent deviations of the
nature of waves is shown in Fig. 20-1 (a). (2) Roughness, as shown at (b),
is characterized by small irregular deviations, as may be felt when the
fingernail is drawn over a ground surface. (3) Surface flaws, as shown
at (c), generally occur at random intervals, for instance, scratches, checks,
etc., and are frequently caused by the processing operations. (4) A sur¬
face with all three types of deviations is shown at (d). (5) Lay means
the direction of the predominant surface pattern.
Cc)-Surface flaws id)-Combination

Fig. 20-1. Types of surface irregularities.
20-2 Measurement of surface deviations. There are a number of

methods that have been devised for measuring surface roughness, or the
degree to which a surface deviates from a plane. None of these methods
is entirely satisfactory, but the commonly used ones, as described below,
yield reasonably consistent readings, and on a comparative basis are
valuable for judging the surface-finishing methods and for controlling
production.
The profilograph is a laboratory instrument in which a stylus with a
sharp tracer point with a tip radius of 0.00005 in. is delicately held against
and normal to the surface being measured, and then is drawn over the
surface at a rate of approximately 0.001 in. per sec. The movement of
the stylus is magnified by a reflected light beam and is recorded on a
moving sensitized film. Thus a graph is secured that is an indication of
the approximate profile of the surface. Since the tracer point has a finite
1 See Surface Roughness, Waviness, and Lay, ASA B46.1-1947.
297
298 MACHINE DESIGN
(a)- Ground surface
^
r 0.000065"
'll * ;« r:'*
T fc./-6round surface superfinished 10 seconds
p O.OOOOl!L
W-Ground surface superfinished 50 seconds

Fig. 20-2. Series of profilographs, each representing about }£-in. linear travel of
stylus over the surface. (Courtesy of Chrysler Corporation.)
radius, and since it cannot explore beneath overhanging projections, the

trueness of the graph is thereby affected. A surface which is described
by 5 microinches (abbreviated 5 /tin.) irregularity by a profilograph means
that the graph indicates a distance from the highest peak to the lowest
valley which is equal to 5 millionths of an inch. A series of profilographs
is shown in Fig. 20-2.
SURFACE FINISH, FRICTION, AND WEAR 299
The profilometer is similar to the profilograph in that a tracer point is

used. It differs, however, in that the motion of the stylus normal to the
surface is translated into electric voltage which is amplified by a cathode-
ray oscillograph arranged to record motions on a moving film. The
record on the film is not a graph of the surface but it indicates the “ run¬
ning average of surface deviations, ” i.e., a root-mean-square value (rms).
Thus 5 pin. by a profilometer (recorded as 5 pin., rms) means that the
square root of the average of the squares of a number of deviations is
5 millionths of an inch. It is thus apparent that the number of micro¬
inches as recorded by a profilograph and by a profilometer cannot be
compared directly; however, each is an indication of the irregularities of a
surface. The profilometer is a portable instrument.
A recent development is the Surfagage (General Motors) which meas¬
ures the arithmetic mean value of surface deviations rather than the rms
value. The Surfagage is a portable tracer-type instrument which con¬
verts the motion of the stylus normal to the surface through a thin viscous
fluid film in shear to a transducer tube. The amplified impulses yield
the arithmetic mean value which may be read directly on the gauge.
The arithmetic mean value is preferred by some manufacturers since
it is felt that it is easier for production and inspection personnel to under¬
stand readily its significance; also it is closer to what may become the
international standard. The British standard now uses the arithmetic
mean value. Fortunately, the difference between the two readings is
small; for most machine surfaces the arithmetic mean is about 10 per cent
less than the rms value.
Other methods1 of determining surface irregularities are the Brush
surface analyzer, the electrical profile recorder, piezoelectric pickup,
reflection of light from surface, profile photomicrographs (which requires
plating of the surface), and the surface dynamometer.
20-3 Damage to surfaces due to finishing. It is evident from metallo-
graphic analyses that the structure of a crystalline metal is changed by
the machining process to an appreciable distance below the surface. In
some of these processes, such as turning, planing, shaping, milling, and
rough grinding, a damaged layer of considerable depth remains on the
part as a result of the operations. An ideal finishing process is one that
removes the damaged layer and produces a geometrically true surface
without introducing a new damaged layer. Some of the finishing proc¬
esses, such as fine grinding and honing, produce a surface that has only
small irregularities, i.e., a good finish; but at the same time generally
introduce a new damaged layer from a metallographic standpoint.
The two main types of surface damage will now be discussed, including
1 A. M. Swigert, “Story of Superfinish,” pp. 91 ff., Lynn Publishing Company,
Detroit, Mich., 1940; Surface Finish, ASME, 1942.
300 MACHINE DESIGN
some features of the surface-finishing methods that avoid these damaging

actions.
Metallurgical damage. The high pressure required and the high tem¬
perature produced by grinding cause a change in structure of an annealed
steel specimen, indicated in Fig. 20-3. Fine grinding affects the structure
to a distance of the order of 0.001 cm below the surface.
This depth of damage is not surprising in view of the high temperatures,
well above 2000 F, regardless of whether or not a coolant is used. The
Fig. 20-3. Metallurgical effects of surface-finishing methods. Arrows show depth of

damage. (From J. Wulff, Proceedings of Special Summer Conferences on Friction
and Surface Finish, Massachusetts Institute of Technology, 1940, p. 14.)
wet grinding affected the surface only about 0.0007 cm below the surface.
The last three finishes shown, i.e., lapped, polished, and superfinished,
affected the surface about 0.0001 cm, a very thin layer, since the pressures
and temperatures were low.
Mechanical damage. When a metal-cutting tool, such as a planer tool,
makes a cut, the chip that is formed curls over the leading edge of the
tool. Since the chip is newly cut metal at high temperature, it removes
any film of lubricant or oxide from the face of the tool and produces condi¬
tions favorable for particles of the chip to adhere to the tool. Under some
conditions a sizable accumulation of metal will build up at the cutting

edge of the tool, and will result in the so-called “ built-up edge,” as illus¬
trated in Fig. 20-4. This built-up edge will grow until it becomes so
large that it breaks off owing to the cutting forces. Thus the built-up
edge periodically forms and breaks off in more or less regular cycles. The
built-up edge has the effect of changing the shape of the tool point and
likewise the form of the surface being machined and is one cause of the
Fig. 20-4. Built-up edge on planer tool. (Courtesy of Cincinnati Milling and Grinding
Machine Company.)
familiar cross marks on machined surfaces, as shown by Fig. 20-5. To

reduce such irregularities, finer cuts may be taken, but the built-up edge
will always be present even though reduced in size. The built-up edges
may appear in all metal-cutting tools, such as planers, shapers, milling
cutters, etc.
In the use of a grinding wheel made of a bonded abrasive, cutting is
performed by small projections of the abrasive. Small built-up edges or
points may appear at the cutting edges or points of such cutters. This is
one reason for the lines that appear on ground surfaces in the direction
of the path of the abrasive over the piece.
30 9. MACHINE DESIGN
In lapping, the motion of the lap should not be one direction as in

scrubbing but should periodically change direction, thus removing the
built-up points on the abrasive before they have an opportunity to reach
a damaging size. One reason for the perfection of the superfinishing
process is that the motion of the abrasive stone is random, i.e., continually
changing in direction, thus avoiding built-up points on the abrasive.
20-4 Production of smooth surfaces. The production of a smooth sur¬
face is necessarily a step-by-step process. First, rough machining fol¬
lowed by fine machining is necessary to bring the part to approximate
dimensions. Then the part may be rough-ground, followed by fine grind¬
ing. In the latter operation, the dimensional accuracy of the part is
established. For the finishing operation, the part may then be honed
Fig. 20-5. Planed surface showing effects of built-up edge.
lapped, or superfinished. These operations are similar in that during

the passage of the abrasive element over the surface, the high points of
the surface are removed and the surface gradually approaches one with¬
out irregularities (see Fig. 20-2).
In honing, a relatively high pressure is used and a corresponding high
temperature at the surface is developed. The combination of high tem¬
perature and pressure usually damages the material below the surface
being produced; also, in the absence of a sufficient degree of random motion
of the stone, built-up points may cause mechanical damage to the surface.
A lubricant is necessary in honing to cool the surface.
In lapping, the peaks of the surface irregularities are progressively
removed, as in honing; but since the pressure between the lap and the
work is comparatively low and the motion slow, a new damaged layer
will not be introduced. Also, since the motion is random, scratches due
to built-up points are avoided. In hand lapping a considerable time may
be necessary to finish a surface.
Superfinishing is similar to lapping except that the relative motion of

the abrasive stone and the part is the result of from three to six separate
superimposed motions so that the path of a particular abrasive point over
the work is very complicated—so-called “ multimotion.” The pressure
is relatively low, being of the order of a few ounces per square inch. A
continuous flow of lubricant is used to flush away particles of the abrasive.
When the irregularities are removed, the lubricant forms a continuous
fluid film over the surface that prevents further finishing, and the opera¬
tion will cease. Thus the superfinishing process produces a surface that
has only small irregularities, within a few microinches of a true surface,
and one that is metallurgically and mechanically undamaged base mate¬
rial to within less than a microinch of the surface. In production, less
than a minute is required to superfinish a surface to within 2 /fin. rms.
The surface that is described above is useful in reducing wear. This
is true since there will not be large surface irregularities to promote failure
of the oil film under adverse conditions of load and speed. Also, for finely
finished surfaces, the wear which takes place during the life of the machine
is greatly reduced, so that it is possible to make the parts originally with
the clearances which are required for the best operating conditions instead
of being forced to use too small clearances when the parts are new and
too large clearances after wear has taken place.
In Table 20-1 are shown some surface-finish values which are obtainable
in various machined and natural finishes.
Table 20-1.
Roughness, rms
Finish or surface 1000 500 250 12 5 63 32 16 8 4 2 1
Cutting torch chip and saw

Hand grind
Disk grind;fi'le ^\co
o \
Lathe, shaper,mill
Bore o V^r
Drill \ cT»
Surface grind
Cylindrical grind
Hone or lap
Polish
Super-finish
Sand cast surface

Foraina
Rolled surface
Dip castings
Extrusions
304 MACHINE DESIGN
20-5 Specification of surface finish. Improvement in performance of

many machines requires finer finished surfaces, and therefore it has
become necessary to establish standard methods for designating surface
finishes. ASA Standard B46-1 uses a designation system to grade sur¬
face roughness as rms values from 0.00000025 in. (J4 /dn-) to 0.063 in.
(63,000 ptin.) in 13 steps. A designating symbol is used on the drawing,
for example, f4 indicates that the surface is to be finished so that its
roughness is limited to a maximum of 4 juin., rms.
It has been shown, however, that surfaces of the same material finished
to the same degree of roughness but by different methods sometimes
behave differently with respect to fatigue strength, friction, wear, and
corrosion. This is true because other characteristics of the surface than
its roughness affect the performance. Among these are (a) direction of
irregularities in the surface plane called lay and (6) the pattern of the
irregularities. In some instances, therefore, it is desirable to specify the
finishing method as well as the degree of roughness.
Sets of machine-finished specimens or their replicas are available on
the market. These are used in a comparative manner so that the surface
specified by the designer may be produced by the shop mechanic and
checked by the inspector.1 In this system, 10 steps of roughness are
designated by the usual finish symbol with an attached number and,
when desirable, the finishing method, for example, “f4 buff,” where the
4 represents the degree of smoothness.
Roughness-width cutoff. In some applications it is desirable to dis¬
tinguish between roughness and waviness of the part in specifying the
surface finish. As an example, where the contact area between two
mating surfaces is important, waviness and roughness both are significant,
the former to ensure uniform pressure between the parts, and the latter
to aid in maintaining the film of lubricant and in controlling wear. In
this case a relatively long distance should be used in obtaining the surface-
finish value. As another example, in a part subjected to cyclic loading,
surface scratches and roughness are dangerous as stress raisers, and in
this case waviness is unimportant. Here, a short distance may be used
to inspect the surface irregularities. The distance over which the surface
irregularities are to be averaged to obtain the surface-finish value is
known as roughness-width cutoff. Standard cutoffs are 0.003, 0.010, 0.030,
0.100, and 1.000 in. The value preferred for most applications is 0.030 in.
A complete indication of surface finish is shown by the symbols in
Fig. 20-6. In (a), one number denotes average roughness. Two num¬
bers in (5) show maximum and minimum roughness values. The
waviness-height rating in inches is added at (c), and at (d) the 0.01 value
1 Walter Mikelson, Surface Finishes, Gen. Elec. Rev., vol. 46, no. 3, p. 185, 1943
is the roughness-width cutoff value, and the lay is indicated as perpendic¬

ular to the edge. Other symbols for lay are two parallel lines to indi¬
cate parallel to edge of surface indicated, X for angular in both direc¬
tions, C for approximately circular pattern relative to center of piece,
M for multidirectional and R approximately radial relative to center.
0.0002-0.25 00002-0.25
0.010
50
W////Z/A V///////A
(a) (b) (c) (d)
Fig. 20-6. Surface-finish symbol.
20-6 Effects of surface finish on friction and wear. When two surfaces
are pressed together, a force parallel to the surfaces is necessary to cause
sliding or relative motion of the two surfaces. The sliding force is
opposed by a force acting along the surfaces that is called the force of
friction.
(a.)-Light pressure
Heavy pressure
Fig. 20-7. Contact between two bodies.
To understand what causes metal-to-metal friction and why it is asso¬

ciated with wear, consider two bodies with irregular surfaces held together
with light pressure. The peaks of the irregularities will be in contact,
as shown exaggerated in Fig. 20-7(a). Now when the surfaces are pressed
together, the pressure at the contacting peaks will be high because of the
small contact areas, and the material at these points may undergo flow
306 MACHINE DESIGN
as shown in (6) in the figure. The pressure at which this flow occurs is
known as the “pressure of fluidity.”1
If now one body is slid relative to the other, two things will probably
occur. One is that interlocking projections, as at D in the figure, may
be detached in more or less sizable particles. This constitutes wear.
The other is that at locations as at E the local regions under the action
of the high pressure may be smeared over to fill in adjacent valleys.
The combination of these two actions, after considerable “wearing in”
has taken place, may result in surfaces shown separated in Fig. 20-7(c) in
which the dotted regions represent parts of the surface that have under¬
gone plastic flow known as “smear metal.” This latter may cover most
or all of the contacting surfaces and if the wearing in process has been
gradual to avoid galling, the surface may be relatively smooth and hard
and satisfactory from the standpoint of wear under mild service condi¬
tions. However if the service is severe, the metallurgically damaged
surfaces may not withstand the high pressures, and rapid wear may be
the result.
It is evident from the above discussion that the resistance to motion,
or force of friction, is dependent on the materials of the surfaces and
their finish. Also as the surfaces become “worn in,” frictional resistance
changes.
Another condition that affects the friction and wear properties of the
surfaces is the presence or absence of a lubricant. The above discussion
has implied clean surfaces, i.e., surfaces free from a fluid film of lubricant
or from an adsorbed film. The adsorbed film forms on newly finished
surfaces by a combination with the surface of cutting oils or fluids or
from handling or even from condensed atmospheric vapors. The
adsorbed film is nearly always present on metallic surfaces so that the
friction of two surfaces in contact is rarely due to base metal in contact
with base metal. In experimental work where it is necessary to remove
this film, cleaning with solvents is not sufficient, and it is necessary to
resort to mechanical means, such as the use of rouge compounds.
From the work of Langmuir2 it may be stated that films of extraordi¬
nary stability consisting of a layer of atoms chemically combined with
the underlying atoms of the solid and, according to later work, molecules
under the surface are affected so that an adsorbed film of more than one
molecule thick is produced.
20-7 Types of wear and its measurement. In order to control wear
in a machine, the designer must give consideration to the materials of
1 H. O’Neill, “Hardness of Metals and Its Measurement,” p. 132, Chapman & Hall,
Ltd., London, 1934; F. P. Bowden and D. Tabor, “The Friction and Lubrication of
Solids,” Oxford University Press, New York, 1950.
2 Irving Langmuir, J. Am. Chem. Soc., vol. 49, p. 1852, 1917.
which the parts are made, their initial surfaces, and the operating condi¬
tions of the rubbing parts. In this treatment, the recognized types of
wear will be discussed briefly and methods suggested for their alleviation.
It should be remembered, however, that the various mechanisms of wear,
which are relatively simple in themselves, seldom occur independently,
but are usually the result of combinations which take place either simul¬
taneously or in sequence.
Cutting wear. This type of wear takes place when a hard material with
a rough surface rubs over a softer one. Experience shows that the softer
material is worn away in a manner similar to cutting in a machining opera¬
tion. It is apparent that the rate of cutting wear depends on the relative
hardness of the materials in contact, the smoothness of the surfaces, and
the velocity of rubbing. Because of the use of hard materials and smooth
surfaces, this type of wear in machines is relatively unimportant.
Abrasive wear. When a particle of grit is carried between two surfaces,
abrasive wear may occur. If the material of one of the surfaces is soft,
the particle of grit may be completely embedded in the surface and no
wear will be caused. If the material of the surfaces in contact is some¬
what harder, however, the particle of grit may become partly embedded
therein and act as an abrasive, causing scoring of the mating surfaces.
It is interesting to note that in abrasive wear, the harder of the two mate¬
rials of the rubbing parts is worn. If the materials of the surfaces in
contact are very hard, so that the particle of grit will not be embedded in
either, the particle will roll between the surfaces and they may be scored
but not seriously damaged.
Abrasive wear thus depends on the relative hardness of three materials,
i.e., of the two surfaces and of the grit. Wear of this type is very com¬
plicated and can be controlled only by a judicious choice of the hardness
of the rubbing parts and by excluding grit from the surfaces so far as
possible by using devices such as dust screens and oil filters.
Wear due to galling. When two surfaces that are absolutely clean are
pressed together, mating portions of the surfaces in contact may adhere
so that when the surfaces are separated or one slid on the other sizable
pieces may be torn away, producing what is known as “galling.” If the
materials are hard, the galled spots may be small, but if the materials
are soft the galling may extend over a large area. Since the sticking
together that is referred to is of the nature of local welding, the tendency
for galling is greater at high pressures and high temperatures.
In machine parts, the surfaces are seldom clean because of the presence
of the adsorbed film. While the adsorbed film may be only a few mol¬
ecules in thickness, it is sufficient to prevent clean metal-to-metal contact.
Only when the conditions of pressure and temperature are effective in
breaking through the adsorbed film is galling likely to occur. It has been
308 MACHINE DESIGN
found, however, that grit in the lubricant between very smooth surfaces
is frequently the cause of removing the adsorbed film and thus promoting
galling. For this reason, if finely finished surfaces are used, it is more
important to ensure clean lubricant than with rougher surfaces. The
presence of 0.01 per cent of grit in a lubricant has been found to promote
galling of a lapped steel surface so that it becomes no better than a ground
one.
The antigalling characteristic of cast iron is due principally to the
graphite, which serves to maintain the adsorbed film. An approach to
this condition is attempted in smooth surfaces of hard steel parts by sub¬
jecting the surface to a brief sandblast in order to provide small widely
separated reservoirs for the lubricant.
Galling (known also as “scuffing”) is found frequently in gear teeth,
screw threads, cam followers, piston rings, cylinder bores, and splines.
Mechanical pitting. When a roller or ball rolls over a surface, pits in
the surface may appear if the surface endurance limit of the materials in
contact is exceeded. Pitting is classified as a type of wear and may lead
to failure of gear teeth, balls, rollers, etc. (see Art. 3-8 for a discussion
of mechanical pitting).
Corrosive wear. This type of destruction is due to chemical action and
in general causes decrease in sizes of members when a protecting film is
not formed by the results of corrosion. Corrosion generally accelerates
other types of wear, although galling is frequently retarded by corrosion.
Corrosion may be due to active chemicals or gases, water, brine, etc.
Special attention should be paid to corrosive agents in lubricant oil.
In addition to the type of corrosion mentioned above, other types are
galvanic corrosion caused by the action of a liquid on the adjacent sur¬
faces of two unlike metals, and “concentration-cell” corrosion caused by
the action of a fluid of nonuniform concentration affecting the surface.
Measurement of wear. Since the actual quantity of material removed
by wear is usually very small, measurement of wear may be difficult.
Relatively large wear may be determined by comparison of the dimen¬
sions of the parts as measured by a micrometer before and after wear.
In the laboratory, wear of small parts may be determined by comparing
weights before and after wearing. In service, a wear gauge may be used.
One type (the McKee wear gauge)1 makes use of a small indentation in
the surface. The indentation is a diamond-shaped pyramidal, shown in
Fig. 20-8. The length of the indentation is measured with a microscope.
The difference between the lengths before and after wear has taken place
is a measure of the wear normal to the surface. The method may be used
1 C. S. Bruce and J. T. Duck, Cylinder Wear Measured with a Microscope, SAE

Quart. Trans., vol. 1, no. 3, July, 1947.
for flat or curved surfaces. It is stated that the sensitivity of this gauge
is 0.000015 in.
The radioactivity method for measuring wear has been developed to a
high degree of sensitivity. In this method, the part whose wear is to be
measured is irradiated; then during test the wear particles pass into the
lubricant and the concentration of wear particles in samples of the lubri¬
cant can be very accurately determined as a measure of rate of wear.
The method requires elaborate laboratory preparations for handling the
equipment and for safety of the personnel, but it is precise and does not
require dismantling of the parts being tested for measurements.
Under conditions of boundary lubrication (see Art. 21-6) in which a
thin film of lubricant is present at the surfaces in contact, friction and wear
Fig. 20-8. Shape of indentation of McKee wear gauge. Original length slightly less
than 1 mm.
do not follow either laws for dry or for lubricated (thick film) conditions.
Wear under these circumstances is markedly affected by the composition
of the lubricant as well as the material of the surfaces. This field is
under extensive investigation since wear of machine parts is largely under
these conditions. The references1 below give results of research work to
date on this subject.
20-8 Solid friction. When two solids are pressed together, a force
parallel to their surfaces in contact is necessary to cause sliding or relative
motion. This force is resisted by a force acting along the surfaces that
is known as the “force of friction.”
If the surfaces are clean and dry, the solid material of the two surfaces
will be in contact, and the force of friction will be caused by the over¬
coming of the interlocking effect of the surface irregularities. These
irregularities may be large, as in visibly rough surfaces, or they may
be of molecular size.
1 Andrew Gemant, “Frictional Phenomena,” Chemical Publishing Company, Inc.,

New York, 1950; F. P. Bowden and D. Tabor, “The Friction and Lubrication of
Solids,” Oxford University Press, New York, 1950; John T. Burwell, Jr. (editor),
“Mechanical Wear,” American Society for Metals, 1950.
310 MACHINE DESIGN
In connection with Coulomb’s laws, it is important to remember that

they apply to surfaces that are clean and dry. In determining solid fric¬
tional resistance by experiment, the surfaces should be thoroughly cleaned
to remove all contamination, including the adsorbed film. Since the
adsorbed film is very difficult to remove completely, many experimental
results reported for friction characteristics of clean surfaces were in reality
determined for partly clean surfaces. This fact probably accounts in
part for the variation in values as quoted by different experimenters.
For solid friction of clean surfaces, the laws of Coulomb apply and are
stated as follows:1
1. The frictional resistance is approximately proportional to the load
on the rubbing surfaces.
2. The frictional resistance is slightly greater for large areas and small
pressures than for small areas and large pressures.
3. The frictional resistance, except for low speeds, decreases as the
velocity increases.
These laws of friction may be expressed by the formula
F =fN
where F = the force applied tangent to the surfaces in contact to over¬

come the resistance due to friction
N = the normal force on the surfaces in contact
/ = the coefficient of friction
If F is the force necessary to start sliding of the surfaces, / is the coeffi¬
cient of static friction; and if F is the somewhat smaller force necessary
to maintain sliding, / is the coefficient of kinetic friction. The transition
from static friction to kinetic friction is gradual, as indicated by the curve
in Fig. 20-9.
Some coefficients of static friction for dry surfaces are given below:2
Table 20-2. Coefficients of Static Friction for Dry Surfaces
Wood on wood. 0.3-0.5
Metal on metal. 0.3 average
Metal on wood. 0.2-0.6
Leather on wood. 0.3-0.5
Leather on metal. 0.3-0.6
Brake lining on steel. 0.4-0.5
Clinging friction. In applications where a lubricated pin or bearing

has intermittent operation with periods of rest, it has been found that
the coefficient of friction is increased after the rest period. For instance,
1 Archbutt and Deeley, “Lubrication and Lubricants/’ 5th ed., p. 49, Charles
Griffin & Co., Ltd., London, 1927.
2 Kimball and Barr, “Elements of Machine Design,” p. 99, John Wiley & Sons, Inc.,
New York, 1935.
an electric switchgear may be required to operate quickly and with little

resistance after an interval of rest, possibly several months. The friction
hindering motion may be greater than that determined from short-time
tests. Another example is a plastic molding machine in which the dies
are forced together by a screw and after
being held together under the required
load (and usually at a high temperature)
it may require more torque on the screw
to release the dies than was required to
set them. 01____
0 Speed
The friction in these cases is known as
Fig. 20-9. Variation of friction
clinging friction. Few data are availa¬ with speed of rubbing.
ble for coefficients of friction after rest
periods; however, one paper1 2 states that in the first 24 hr the increase in
coefficient of friction in pivot bearings is of the order of 15 to 40 per cent,
most of the increase occurring in the first 2}z% hr, and a small increase in
the subsequent 4 days. The tests covered a range of 60 sec to 90 days.
20-9 Rolling resistance. When a roller made of an elastic material is
pressed against a plane surface of a similar material, both the roller and
the plane are deformed so that surfaces in contact appear as shown in
Fig. 20-10. Deformation of surfaces.
Fig. 20-10, where the dotted lines indicate the original outlines. If the
roller is rotated, it may be noted that points A and B on the roller move
with different linear speeds with respect to the center of the roller since
their radii are not equal. Since all points of the plane are substantially
at rest, there must be sliding between the roller and the plane along the
surfaces in contact. The corresponding friction of sliding results in the
work of rolling friction A
It is noted that rolling friction is of the same nature as solid sliding
friction, as discussed in the preceding article, but with the difference that
1 M. C. Hunter, Static and Clinging Friction of Pivot Bearings, Proc. Inst. Mech.
Engrs. {London), 1942, p. 274.
2 See M. C. Shaw and E. F. Macks, “Analysis and Lubrication of Bearings,” p. 422,
McGraw-Hill Book Company, Inc., 1949.
312 MACHINE DESIGN
rolling friction is less than friction of sliding an equal weight over the
surface.
For the purpose of analyses involving rolling friction, it is convenient to
consider that the torque due to frictional resistance is represented by a
couple Wa, as shown in Fig. 20-11,
where W = weight on the roller or ball
F = force required to cause rolling
r = radius of roller
a — distance (coefficient of rolling friction)
For equilibrium of forces for the roller or ball,
Fr = Wa
or
Fr
a (20-1)
W
The value of a may be determined experimentally and it has been shown

that it depends on the materials of the roller or ball and the surface over
Fig. 20-11. Forces on roller.
which it rolls, but it is generally considered independent of the size of

the roller and of the speed and is known as the “ coefficient of rolling fric¬
tion. ” It has dimensions of length and is given in handbooks in inches.
Some values are given below:
Table 20-3
Annealed steel rolling on annealed steel. 0.02
Hardened steel rolling on hardened steel. 0.003-0.004
Cast-iron or steel rolling on wood. 0.10
Ball bearings. 0.0008-0.0012
Roller bearings. 0.0020-0.0040
20-10 Fluid friction. When a fluid film separates two surfaces having
relative motion, solid friction will not exist, and the only friction present
will be that within the fluid film. This type of friction is discussed in the
following chapter.
CHAPTER 21
SLIDING BEARINGS AND LUBRICATION
21-1 Introduction. In a general sense, the function of machine bear¬

ings is to permit constrained relative motion of rigid parts. In sliding
bearings a lubricant is generally inserted or supplied between the mating
surfaces to reduce friction and wear and in some cases to carry away the
heat generated. In ball and roller bearings, rolling motion is utilized.
Some common types of sliding bearings are the following: (a) journal
bearing, (b) full journal bearing, (c) partial journal bearing, (d) clearance
bearing, (e) fitted bearing, (/) thrust bearing, and (g) slipper, or slider,
bearing.
Journal bearings are used to furnish lateral support to rotating shafts.
In a journal bearing, the journal is the part of the shaft which runs in
the bushing, which is usually stationary and supports the journal, as

shown in Fig. 21-1. In some applications, the journal is at rest while
the bushing rotates, and in others, such as connecting rod bearings, both
the journal and the bushing have motion.
In a full journal bearing, the angle of contact of the bushing with the
journal is 360 deg. This is the type most commonly used in industrial
machinery to accommodate bearing loads in any radial direction.
In a partial journal bearing, the angle of contact is 180 deg or less,
120 deg being a common value. The partial journal bearing may be used
when the direction of the load does not change materially; it is used for
structural simplicity, for convenience in applying the lubricant, and for
313
314 MACHINE DESIGN
the reduction of frictional loss. Railroad-car bearings represent an exten¬

sive application of the partial journal bearing.
Clearance in a bearing refers to the thickness of the space allowed for
the lubricant that separates the parts having relative motion. A clear¬
ance bearing is one in which the radius of the journal is less than the radius
(a)-Clearance bearing
Fig. 21-2. Shape of oil film.
of the bushing. Most journal bearings are of this type. The shape of
the oil film in a clearance bearing is shown in Fig. 21-2(a).
A fitted bearing is one in which the radius of the journal and bushing
are equal. It is evident that a fitted bearing must be a partial bearing
and the journal must run eccentric with the bushing in order to provide
space for the lubricant, as shown in Fig. 21-2(6). Fitted bearings are
more common in England than in the United States.
(cl-Conical bearing
Fig. 21-3. Thrust bearings.
A bearing designed to support an axial load is called a thrust bearing.

This type of bearing may be broadly divided into step bearings, see
Fig. 21-3(a), in which the end of the shaft is in contact with a bearing
surface; collar bearings, at (6), in which a collar is attached to or formed
integral with the shaft; and pivoted-segment bearings. The latter type is
discussed in Art. 21-3.
SLIDING BEARINGS AND LUBRICATION 315
Conical bearings as shown at (c) will support a transverse load as well

as an axial load.
A slipper, or slider, bearing is one in which the two surfaces are flat and
nearly parallel and the relative motion is translation.
Figure 21-4 is a schematic diagram of a slider bearing that shows a
wedge-shaped film of lubricant. It will be shown in Art. 21-2 that a
wedge-shaped, or converging, film is necessary in order that a load may
be supported by a fluid film.
A lubricant in sliding bearings is fed or inserted between the surfaces
to reduce friction and wear. Oils and greases are the most common
lubricants but other substances may be used, such as water in hydraulic
equipment, air in high-speed spindles, and milk in cream separators.
In addition to reducing friction and wear, the lubricant may carry away
heat generated in the bearing.
Moving surface
Fig. 21-4. Diagram of slider bearing.
21-2 Mechanism of fluid lubrication. By fluid lubrication is meant a

condition of operation in which the film of lubricant is so thick that metal-
to-metal contact between the parts is prevented, and the only friction is
that which occurs within the fluid film. Fluid friction is considerably
less than metal-to-metal friction. Since fluid lubrication is essential if
friction and wear are to be a minimum, the conditions that are necessary
to establish and maintain fluid lubrication are important and will be
developed in this article. The slider bearing will be used for this purpose,
and later it will be shown how the results arrived at for the slider bearing
may be applied to other types of bearings, for example, the journal
bearing.
A slider bearing is shown in Fig. 21-5. The following assumptions are
made:
1. That the fluid completely fills the space between the two surfaces
2. That the fluid “ wets” or adheres to the surfaces so that the velocity
of the fluid at each surface is the same as that of the surface1
3. That flow in the direction perpendicular to the velocity U is neg¬
ligible
1 Archbutt and Deeley, “Lubrication and Lubricants,” p. 81, Charles Griffin & Co.,
Ltd., London, 1927.
316 MACHINE DESIGN
In Fig. 21-5(a), the surface of the shoe is parallel to that of the moving
surface. When the moving surface has a horizontal velocity U, the hori¬
zontal layers, or laminae, of fluid move with velocities that are directly
proportional to their distances from the fixed surface as indicated. It
should be noted that the area of the triangle oab in the figure is propor¬
tional to the volume of fluid passing a unit width of the film in a unit time.
Since the volume passing all cross sections of the film must be equal, it
follows that the areas of the velocity-clearance diagrams must be equal,
as indicated in the figure.
Shoe
faJ-Uniform thickness film
Shoe
(b)~Converging film
Fig. 21-5. Variation of velocity across film.
The slope of the line ob at (a) is known as the “velocity gradient”

and may be expressed as du/dy} where y is the distance measured along
the film thickness oa.
From Newton’s law of viscous flow, which states that the shearing stress
F is proportional to the rate of shear (velocity gradient), we may write
„ du
(21-1)
~ M dy
where /jl is the viscosity of the fluid.

In Fig. 21-5(a), the velocity gradient is constant; hence the shearing
stress between the horizontal laminae is constant at all points of the film.
In Fig. 21-5(5), the surface of the shoe is inclined to the moving surface
and the thickness of the film varies from /i2, where the fluid enters the
film, to hi where it leaves. This is known as a “converging film.” It is
evident from the figure that the shape of the velocity-clearance diagrams
cannot be triangular always, as shown by the dotted lines, since their
areas could not be equal as required. The variation of the velocities as

indicated by the curved lines would result in figures of equal areas, and
it will be shown that this variation is compatible with velocity-pressure
relations. Note that these curved lines represent variable velocity gradi¬
ents as compared with the uniform gradient for the constant thickness
film.
If an elementary volume of unit width, as shown at A in Fig. 21-5(6),
is drawn as in Fig. 21-6(a), it may be referred to axes ox and oy, as shown,
where bx, and by are dimensions of the element parallel to the ox and oy
axes, respectively. As originally assumed, flow in the ox direction only
will be considered. As shown in Fig. 21-6(a), the lower face of the ele¬
ment will be assumed to move with the velocity u; hence the upper face
y , du , ,
u+^dy
*+i
-V
i 1
P ox
6y
!
o -
F
<-dx-
(a} (b)
Fig. 21-6. Element of fluid.
will move with the velocity u + (du/dy)by. The forces acting in the x
direction are shown at (6). The left face is subjected to the fluid pres¬
sure P and the right face to P + (dP/dx)bx. The shear stress on the
lower face is represented by F and on the upper face by F + (dF/dy)by.
For equilibrium of the horizontal forces on the element,
or
dF _ dP
(21-2)
dy dx
This equation states that the change in shear stress in the y direction
is equal to the change of pressure in the x direction.
From Eq. (21-1),
du
318 MACHINE DESIGN
Therefore
dF d2u
i
dy ^ dy2
From the above equation and Eq. (21-2),
dP d2u
(21-3)
dx ^ dy2
This equation states that the change of pressure in the x direction is

equal to the coefficient of viscosity multiplied by the second derivative
of u with respect to y. It is now apparent that when the velocity gradient
Fig. 21-7. Pressure distribution in converging film.
du/dy is constant, as in Fig. 21-5(a), d2u/dy2 will be zero, and therefore

dP/dx will be zero, and there can be no change in pressure along the film
in the x direction; hence no pressure can be built up in a parallel film.
It can be shown by referring to Fig. 21-7 that for the velocity gradient
at the entering end of the film, d2u/dy2 is positive; hence dP/dx is posi¬
tive, and the pressure in the film will increase in the ox direction. The
increase will be positive until the section is reached where the velocity
gradient is constant when dP/dx is zero, indicating a maximum pressure.
Further along ox, dhi/dy2 will be negative; hence dP/dx will be negative,
and the pressure will drop. The variation of pressure is shown by the
curve in Fig. 21-7.
From the foregoing, it follows that a positive pressure can be built up
and a load supported by a fluid only by the use of a converging film. All
bearings supporting a load by means of a fluid film must make use of the
converging film, as for example, the simple slider bearing shown diagram-
matically in Fig. 21-4, the pivoted-segment thrust bearing, and the jour¬
nal bearing. The latter types of bearings are discussed in the following
articles.
21-3 Pivoted-segment bearing. According to the discussion in the
preceding article, a converging film is necessar}^ in a bearing if the advan¬
tages of fluid friction are to be realized. It is apparent that the thrust
bearings shown in Fig. 21-3 cannot provide converging films, and hence
the axial loads will not be supported by a fluid film. Instead there will
exist at best partial metal-to-metal contact and the friction will be com¬
paratively high. If the pressure between the rubbing parts is high, the
lubricant may be forced from the surfaces in contact and rapid wear or
seizure of the parts may result. Thus the load-carrying capacity of this
Fig. 21-8. Pivoted segment thrust bearing, and pressure distribution on a segment.
type of bearing is limited. The use of multiple collars on the collar thrust
bearing is an attempt to increase the load-carrying capacity; however,
such a bearing occupies considerable space, and it is difficult to secure
equal division of the load among the collars.
The pivoted-segment bearing was developed in order to provide a con¬
verging film in a thrust bearing. This bearing was developed independ¬
ently by Kingsbury in the United States and Michell of Australia in about
1905. As shown in Fig. 21-8, the segments are separately pivoted so
that they can tilt to form the converging film. Because of leakage of the
oil from the sides of the segments, the flow of oil will be as represented on
the face view of a segment. The distribution of pressure is also shown.
The segments may be pivoted on a single point pivot by a rocker, or
the entire segment may be supported by a number of springs. Loads in
excess of a million pounds are supported in this manner by a single thrust
bearing.
21-4 Journal bearings. An oil film of a journal bearing with the clear¬
ance exaggerated is shown in Fig. 21-9. At (a) in the figure the journal
is at rest with metal-to-metal contact at a on the line of action of the sup-
320 MACHINE DESIGN
ported load. When the journal rotates slowly in the direction indicated
at (6), the point of contact will move to b, so that the angle aob will be
the angle of sliding friction of the surfaces in contact at b. In the absence
of a lubricant, there will be dry metal-to-metal friction. If a lubricant is
present in the clearance space, a thin adsorbed film of the lubricant may
partly separate the surface, but a continuous fluid film completely sepa¬
rating the surfaces will not exist because of the slow speed.
As shown at (c), the speed of the journal has been increased so that a
continuous fluid film is established, and the center of the journal has
moved so that the minimum film thickness is at c. It may be noted that
from d to c in the direction of motion, the film is continually narrowing,
and hence is a converging film, as was found necessary in the discussion in
Art. 21-2 to support a load. The curved converging film may be con¬
sidered as a wedge-shaped film of a slipper bearing wrapped around the
Fig. 21-9. Formation of a continuous oil fim.
journal. It may be noted also in the figure that from c to d the film is
diverging and cannot give rise to a positive pressure or a supporting
action. In fact, a negative pressure will be developed in this region that
may serve to draw lubricant from the source of supply if conditions are
favorable, or the negative pressure will be prevented from forming if a
lubricant is supplied at the proper point by forced feed.
Two views of the bearing of Fig. 21-9(c) with the ideal variation of
pressure in the converging film are shown in Fig. 21-10(a). Actually,
due to side leakage, the angle of contact on which pressure acts is some¬
what less than 180 deg. The distribution of pressure in the axial direction
is shown at (b) in the figure.
21-5 The hydrodynamic theory. In the discussion of the mechanism
of film lubrication in Art. 21-2, it was shown that a converging film is
necessary in order that a positive pressure can be built up in the film to
enable it to support a load. For a quantitative determination of the pres¬
sures in a bearing corresponding to given operating conditions, it is neces¬
sary to resort to the hydrodynamic theory, or to a simplified expression
of that theory. A brief resume of the development of the hydrodynamic
theory will be given here, followed by a simplification of the theory lead¬
ing to results that may be used in bearing design.
In 1883, Petroff published results of an experimental investigation

showing a connection between viscosity and friction in a journal bearing.
In 1883 and later years, Tower conducted experiments which were sub¬
sequently interpreted by Reynolds in his classical paper1 in which he
developed his general equation of the hydrodynamic theory, which can
be written as
d / h3 dp\ d / h* dp\ U dh
(21-4)
dx \12/x dx) dz \12/x dz) 2 dx
where U is the surface speed, p is the hydrostatic pressure at any point

in the film whose coordinates are x and z, pi is the viscosity of the lubri¬
cant, and h is the film thickness at any section. U is in the x direction,
and h is in the y direction.
Fig. 21-10. Pressure in a journal bearing.
Reynolds integrated Eq. (21-4) for the case of constant viscosity and
negligible side leakage.
In 1904, Sommerfeld applied Reynolds’ equation to the journal bearing
and succeeded in integrating the equation for all values of bearing eccen¬
tricity. In 1914, an important contribution was made by Hersey,2 who
applied dimensional analysis to experimental work on bearings. The
resulting interpretations compare favorably with the hydrodynamic
theory. The dimensional analysis referred to is the basis for the ZN/p
relations, which are discussed in the following article and which simplify
the design of a journal bearing. In this procedure, the effect of side leak¬
age in the bearing is dealt with by means of introducing an experimentally
determined correction. In addition, the viscosity of the lubricant is
assumed to be independent of the pressure in the bearing and to have a
1 O. Reynolds, On the Theory of Lubrication and its Application to Mr. Beauchamp
Tower’s Experiments, Phil. Trans. Roy. Soc. {London), vol. 177, 1886.
2 M. D. Hersey, “Theory of Lubrication,” pp. 70 ff., John Wiley & Sons., Inc., New
York, 1938. This book contains an excellent discussion of the history and theory of
lubrication.
322 MACHINE DESIGN
value corresponding to an average temperature of the lubricant1 in the

film of oil. The errors due to these assumptions have been shown to be
small for ordinary bearings, although they may be large for some types
of bearings, particularly for‘those having very high pressure.2 Albert
Kingsbury and H. A. S. Howarth have made significant contributions to
bearing design as a result of theoretical and experimental investigations.3
21-6 Friction in journal bearings. The coefficient of friction is impor¬
tant in bearing design, since it affords a means for determining the loss
of power due to bearing friction and for estimating conditions of thermal
equilibrium. The coefficient of friction in journal bearings is defined as
T_
Pr
where / = coefficient of journal friction

P = load supported by bearing, lb
F = summation over the bearing area of the shear stress exerted
by the lubricant, lb
T = torque required to overcome journal friction, lb-in.
r = radius of journal, in.
By the use of the dimensional analysis referred to in Art. 21-5, it is
found that the coefficient of friction for a journal bearing can be expressed
as follows:
where / = coefficient of friction

</> = a functional relationship
Z = absolute viscosity of the lubricant, centipoises
N = speed of journal, rpm
p = bearing pressure in psi of projected bearing area
d = diameter of journal, in.
c = difference between diameter of bushing and diameter of jour¬
nal, in.
L = length of bearing, in.
The quantity ZN/p is termed the bearing characteristic number and
is dimensionless if the quantities are expressed in terms of a consistent
system of units. The variation of the coefficient of friction with operat¬
ing values of ZN/p is shown in Fig. 21-11. The part of the curve ab
1 A. E. Norton, “Lubrication,’’ p. 88, McGraw-Hill Book Company, Inc., New
York, 1942.
2 L. J. Bradford, Bearing Design in the Light of Oil Film Pressure Investigations,
Penn. State Coll. Eng. Expt. Sta. Tech. Bull. 14, 1931.
3 See also Hunsaker and Rightmire, “Engineering Applications of Fluid Me¬
chanics,” chaps. XII and XIII, McGraw-Hill Book Company, Inc., New York, 1947.
represents the region of fluid-film lubrication. Between b and c the vis¬

cosity Z or the speed N are so low, or the pressure p is so great, that their
combination ZN/p will reduce the film thickness so that partial metal-to-
metal contact will result. Between c and d on the curve, boundary lubri¬
cation, or imperfect lubrication, exists. This is the region where the vis¬
cosity of the lubricant ceases to be a measure of friction characteristics,
but the so-called “oiliness” of the lubricant is effective in preventing com¬
plete metal-to-metal contact and seizure of the parts.
It may be noted that the part ab of the curve represents stable operating
conditions, since from any point of stability, a decrease in viscosity Z,
for instance, will reduce ZN/p. This will result in a decrease in/, fol¬
lowed by a lowering of bearing temperature that will raise the viscosity Z.
Fig. 21-11. Journal-bearing friction as a function of ZN/p.
The minimum friction for the bearing occurs near point b on the curve,
but it is necessary to operate the bearing in service sufficiently far to the
right of b to avoid any possibility of breaking down the film of lubricant
because of any unpredicted change in operating conditions. It is custom¬
ary to design bearings for a value of ZN/p at least five times that corre¬
sponding to point b on the curve.
The following equation may be used for estimating values of the coeffi¬
cient of friction/for well-lubricated full journal bearings
_ 473 (ZN\
(21-5)
J ~ 1010 \ p )
where Z = viscosity of the lubricant, centipoises

N = speed of journal, rpm
V = bearing pressure in psi of projected bearing area = load on
journal -f- Ld
L — bearing length, in.
324 MACHINE DESIGN
d — journal diameter, in.

c = difference between bushing diameter and journal diameter, in.
K = factor to correct for end leakage = 0.002 for L/d ratios of
0.75 to 2.81
Operating values for ZN/p should be compared with values given in
Table 21-1 to ensure a safe margin between operating conditions and
the point of film breakdown.
21-7 Bearing loads. An important consideration in the design of a
journal bearing to operate under conditions of film lubrication is the con¬
ditions affecting the breaking down of the film, which occurs at operating
conditions between b and c in Fig. 21-11. In addition to the character¬
istic number ZN/p, other factors that affect the film breakdown are as
follows:
Clearance. The clearance in a bearing should be small enough to pro¬
vide the necessary velocity gradient, so that the pressure built up will
T
$
--L -
(a)-L/d -2 (b)-L/d=l
Fig. 21-12. Effect of length to diameter ratio, L/d, on average bearing pressure.
support the load. This consideration requires a small clearance, which

has the additional advantage of decreasing side leakage. However,
allowance must be made for manufacturing tolerances in the journal and
the bushing, for deflection of the shaft, and for space to permit foreign
particles, such as grit and flakes of metal, to pass through the bearing.
These latter considerations require large clearance. Thus both consid¬
erations require a compromise. A commonly used clearance in industrial
machines is 0.001 in. per in. of journal diameter. The minimum thick¬
ness of the oil film may be of the order of 0.0002 in. per in. of diameter.
Length to diameter ratio. Because of side leakage of the lubricant from
the bearing, the pressure in the film is atmospheric at the ends of the
bearing. As shown in Fig. 21-12, the average pressure, pav, will be higher
for a long bearing as shown at (a) than for a short bearing as shown at (6).
1 S. A. McKee and T. R. McKee, Friction of Journal Bearings as Influenced by

Clearance and Length, Trans. ASAIE, vol. 51, p. APM-51-15, 1924; see also M. C.
Shaw and E. F. Macks, “Analysis and Lubrication of Bearings,” McGraw-Hill Book
Company, Inc., New York, 1949.
Table 21-1. Bearing Design Practices
Maximum Operating
No. Machinery Bearing c/d L/d
V, Psi Z ZN/v
1 Automobile and air- Main 800-1,700 7 15 0.8-1.8
2 craft engines Crankpin 1,500-3,500 to 10 0.7-1.4
3 Wristpin 2,300-5,000 8 8 1.5-2.2
4 Gas and oil engines, Main 700-1,200 20 20 0.001 0.6-2.0

5 4-stroke Crankpin 1,400-1,800 to 10 0.001 0.6-1.5
6 Wristpin 1,800-2,200 65 5 0.001 1.5-2.0
7 Gas and oil engines, Main 500- 800 20 25 0.001 0.6-2.0

8 2-stroke Crankpin 1,000-1,500 to 12 0.001 0.6-1.5
9 Wristpin 1,200-1,800 65 10 0.001 1.5-2.0
10 Marine steam Main 500 30 20 0.001 0.7-1.5

11 engines Crankpin 600 40 15 0.001 0.7-1.2
12 Wristpin 1,500 30 10 0.001 1.2-1.7
13 Stationary, slow- Main 400 60 20 0.001 1.0-2.0

14 speed steam en- Crankpin 1,500 80 6 0.001 0.9-1.3
15 gines Wristpin 1,800 60 5 0.001 1.2-1.5
16 Stationary, high- Main 250 15 25 0.001 1.5-3.0

17 speed steam en- Crankpin 600 30 6 0.001 0.9-1.5
18 gines Wristpin 1,800 25 5 0.001 1.3-1.7
19 Reciprocating Main 250 30 30 0.001 1.0-2.2

20 pumps and com- Crankpin 600 to 20 0.001 0.9-1.7
21 pressors Wristpin 1,000 80 10 0.001 1.5-2.0
22 Steam locomotives Driving axle 550 100 30 0.001 1.6-1.8

23 Crankpin 2,000 40 5 0.001 0.7-1.1
24 Wristpin 4,000 30 5 0.001 0.8-1.3
25 Railway cars Axle 500 100 50 0.001 1.8-2.0
26 Steam turbines Main 100-275 2-16 100 0.001 1.0-2.0
27 Generators, motors, Rotor 100-200 25 200 0.0013 1.0-2.0

centrifugal pumps
28 Gyroscope Rotor 850 30 55 0.0013
29 Transmission Light, fixed 25 25 100 0.001 2.0-3.0

30 shafting Self-aligning 150 to 30 0.001 2.5^.0
31 Heavy 150 60 30 0.001 2.0-3.0
32 Cotton mill Spindle 1 2 10,000 0.005
33 Machine tools Main 300 40 1 0.001 1.0-4.0
34 Punching and shear- Main 4,000 100 0.001 1.0-2.0

35 ing machine Crankpin 8,000 100 0.001 1.0-2.0
36 Rolling mills Main 3,000 50 10 0.0015 1.0-1.5

326 MACHINE DESIGN
Thus from the standpoint of side leakage, a bearing with a large L/d is
preferable. However, space requirements, manufacturing tolerances,
and shaft deflections are better met with a bearing of short length. Thus
a compromise in the L/d ratio is necessary, resulting in a value of from 1
to 2 for general industrial machinery. In crankshaft bearings, the L/d
ratio is frequently less than unity, since short bearings are required from
space considerations, and since the shaft diam¬
eter must be large, which is required by the
strength, stiffness, and vibration characteristics.
(See Table 21-1 for values of L/d for various
types of service.)
Oilholes and grooves. Oilholes and grooves in
a bearing limit the pressure at their locations to
that of the supply pressure, which is generally
OH groove considerably less than the hydrodynamic pres¬
Fig. 21-13. Effect of cir¬ sures required to support the bearing load.
cumferential oil groove
Thus a bearing with a central circumferential
on average bearing pres¬
sure. Compare with Fig.
groove will in effect be equivalent to two narrow
21-12(a). bearings, as shown in Fig. 21-13, with consider¬
able reduction in load-carrying capacity. It is
evident that an oilhole in the region of high pressure reduces the load¬
carrying capacity of the bearing. Thus the location of oilholes and
grooves is an important consideration. From the standpoint of load¬
carrying capacity the following items represent good practice:
1. Locate oilholes and necessary axial grooves in the region of low
pressure.
2. Avoid diagonal grooving.
3. If circumferential grooves are necessary, locate them near the ends
of the bearing.
Surface finish. While the degree of smoothness of the surfaces of the
journal and the bushing does not have a marked effect on a bearing
(a)- Rough surfaces. Smooth surfaces

Fig. 21-14. Effect of surface finish on film thickness.
operating with fluid-film lubrication, it does have an effect on the point

of breaking down of the film. This will be evident from a consideration
of the effect of the high points of a rough surface breaking through a
film, as shown in Fig. 21-14, at a lower value of ZN/p than for a bearing
with smooth surfaces.
The diagram in three dimensions in Fig. 21-15 shows that a bearing

operates in the region of fluid friction at a lower value of ZN /p and with
a lower value of minimum friction with a smooth surface than with a
rough one. For this reason, smooth surfaces are desirable to decrease
the wear of a bearing that must operate continually or pass through a low
value of ZN/p frequently.
21-8 Bearing materials. From the preceding discussion of the mech¬
anism of film lubrication, it is apparent that, when the surfaces of a bear¬
ing are separated by a fluid film, the materials of the bearing do not have
any effect on its operation so long as the parts have sufficient strength
to withstand the imposed loads and sufficient rigidity to maintain
alignment.
ZJV/p
Fig. 21-15. Three-dimensional chart showing effect of surface finish on point of fluid-
film rupture.
On account of strength and rigidity requirements, journals are generally

made of steel. For the bushing, however, there are other requirements,
as discussed below, that must be considered in selecting the best material.
Low friction properties. When a journal is starting, stopping, or run¬
ning at low speeds, lubrication will be of the boundary type and metal-to-
metal contact will be prevented only by the adsorbed film. Metals and
alloys that favor the formation of the adsorbed film are therefore desirable
for bearings in order to limit wear. The oiliness of the lubricant should
be considered in this connection. The low friction and wear character¬
istics of babbitt metals are partly responsible for their widespread use in
bearings.
Compressive strength. The bearing material should have high compres¬
sive strength to prevent extrusion of the material from the bearing.
Since the maximum bearing pressure may be considerably greater than
328 MACHINE DESIGN
the pressure expressed in pounds per square inch of projected bearing

area, yielding may occur in localized regions of the bearing to the extent
that the clearance is materially changed.
Fatigue strength. Under seVere conditions of loading and temperature,
surface fatigue cracks may develop. These cracks may cause checking
and pitting of the surface, thus leading to its deterioration to the extent
that the operation of the bearing is affected. Thus a bearing material
having a high fatigue strength is desirable.
For severe operating conditions, a very thin bearing metal bonded to a
steel back has been found to prevent the spreading of the fatigue cracks.
Such a combination also prevents extrusion of the bearing material
because of the restraining effect of the hard backing. In some automo¬
tive engines a tin-base babbitt lining 0.003 in. in thickness is used.
C onfor mobility. Because of small inaccuracies in the form of the jour¬
nal and its deflection under the imposed loads, the material of the bearing
should adapt its shape to that of the journal. This change of shape may
be accomplished by plastic flow, by wearing away, or by local melting.
Plastic flow is most desirable to achieve conformability, since wearing
away and local melting are accompanied by excessive heat, which may
burn out the bearing.
Embeddability. While all bearings should be designed so that foreign
particles will be excluded, this is difficult to accomplish in plain bearings
lubricated with oil, since grit, sand, and metal particles may be introduced
with the lubricant or ventilating air. If the bearing materials are hard,
the particles may score the surfaces and produce undue wear. But if the
bearing lining is soft so that the particles are completely embedded, this
trouble may not be serious.
Bonding. A bearing material that will bond readily and permanently
with the steel or bronze back is necessary for long bearing life. For the
thin linings mentioned above, an intermediate layer of sintered alloy is
used to secure a good bond.
Corrosion resistance. Bearing materials have various degrees of resist¬
ance to corrosion; this should be considered where corrosive oils must be
used.
The preceding discussion indicates the desirable characteristics of bear¬
ing materials. All of these have not been incorporated to a high degree
in any particular bearing material, and it is evident that the choice of a
material for any application must represent a compromise. This is the
reason for the large number of bearing materials which have been devel¬
oped, each of which is most suitable for particular applications.1 The
tin-base and lead-base babbitts are in widespread use since they satisfy
most requirements for general applications. However, where the loads
1 See “SAE Handbook.”
are very high, bronze and brass bushings may be used. Where high com¬
pressive and fatigue strengths are necessary, copper, lead, and tin alloys
may be used. Where the bearing is inaccessible or lubrication is infre¬
quent, “oilless” bearings, composed of porous sintered bronze or iron with
graphite or oil filling the porous structure, may be used.
21-9 Viscosity and oiliness. Newton’s law states that at any point in
a fluid the shearing stress is proportional to the rate of shear, or
F = fiR (21-6)
where F = shearing stress, or tangential force per unit area

R = rate of shear, or velocity gradient
In Eq. (21-6), the proportionality constant p is known as the “vis¬
cosity” of the fluid. In cgs units, it is the absolute viscosity and is equal
to dynes per square centimeter with a velocity of 1 cm per sec at a distance
of 1 cm. The unit of absolute viscosity is the poise, but is usually used as
one-hundredth of a poise termed a “centipoise.”
The standard method of measuring viscosity in this country is by the
use of the Saybolt Universal viscometer by which is determined the time
required for a standard volume of oil at a certain temperature to flow
under a certain head through a tube of standard diameter and length.
The time so determined in seconds is the Saybolt Universal viscosity.
Machine oils at room temperatures have a range of viscosity of about
100 to 300 centipoises.
To convert the Saybolt Universal viscosity to the absolute viscosity
the following formula may be used
Z = Pt (o.22 S (21-7)
where Z = absolute viscosity at temperature t, centipoises

S = Saybolt Universal viscosity, sec
pt = specific gravity at temperature t
p60 — 0.000365(£ — 60)
p6o = specific gravity at 60 F
The viscosities of some lubricating oils are given in Fig. 21-16.
Oiliness is a joint property of a lubricant and the bearing surfaces in
contact, and it is a measure of the lubricating qualities under boundary
conditions where base metal-to-metal contact is prevented only by the
adsorbed film. There is no absolute measure of oiliness. The four-ball
testing apparatus1 gives a comparative measure of this property. Bab-
1 D. Clayton, The Use of the Four-ball Extreme-pressure Lubricant Testing Appa¬

ratus for.Ordinary Lubricants, in “General Discussion on Lubrication,” vol. 2, p. 274,
Institution of Mechanical Engineers, London, 1938.
330 MACHINE DESIGN
No. p60
1 Transmission oil SAE /60 0.9365
2 Gear oil 0.9/53
3 Transmission oil, SA E 110 0.9326
4 Airplane oil 100, SAE 60 0.8927
5 Automobile oil, SAE40 0.9275
6 Automobile oil SA E 20 0.9254
7 Ring-oiledbearing 0.9346
8 All year automobi/e oil SAE 20 0.9036
9 Turbine oil, ring- oi/ed bearing, SAE /O 0.8694
10 Turbine oil, ring-oiled bearing 0.8877
Fig. 21-16. Viscosity-temperature chart and specific gravity of oils. (Courtesy of The
Texas Company.)
bitt metals favor the establishment of the adsorbed film, which prevents
or retards actual base metal-to-metal contact. Lard oil has better oil¬
iness than mineral oils and has the effect on the coefficient of friction as a
function of ZN/p that is shown in Fig. 21-17.
21-10 Imperfect lubrication. In bearings in which the operating value
of ZN/p is too low to ensure fluid-film lubrication or where there is an
insufficient supply of lubricant in the bearing, the friction will be due to

metal surfaces in contact. This condition is represented by cd in Fig.
Fig. 21-17. Effect of oiliness on point of fluid-film rupture.
21-11. The following formula1 may be used to estimate the coefficient

of friction / under this condition:
, _ CiC2 4/p
(21-8)
J 250 VV
where Ci, C2 = values given in Table 21-2

p = bearing pressure on projected bearing area, psi
V = rubbing velocity, fpm
21-11 Strength and stiffness of journal bearings. The diameter of the
journal is usually determined by making it equal to or somewhat less than
(a)-A\iqned (b)~ Shaft

deflected
Fig. 21-18. Effect shaft deflection bearing alignment.
the adjacent diameter of the shaft. In most cases the shaft diameter is
determined from a consideration of the strength and stiffness and possibly
of the vibration requirements. However, the deflection of the journal in
the bushing is an important consideration in order to prevent undue wear
and failure of the oil film. In Fig. 21-18 there is shown diagrammatically
a perfectly aligned bearing at (a), the effect of shaft deflection at (6), and
a self-aligning bearing at (c), which may be necessary in some instances.
1 Louis Illmer, High Pressure Lubrication, Trans. ASME, 1924, p. 833.

332 MACHINE DESIGN
Table 21-2. Factors for Equation (21-8)
Values of Factor Ci
Method of lubrication Workipanship Attendance Location Ci
Bath, flooded or oil ring High grade First class Clean and protected 1
Oil, free drop, constant Good Fairly good Favorable (ordinary 2
feed conditions)
Oil cup or grease, inter¬ Fair Poor Exposed to dirt or other 4
mittent feed unfavorable conditions
Values of Factor C2
Type and example of bearing C2

Rotating journals, such as rigid bearings and crankpins. 1
Oscillating journals, such as rigid wrist pins and pintle blocks. 1
Rotating journals lacking ample rigidity, such as eccentrics. 2
Rotating flat surfaces lubricated from the center, such as annular step or pivot
bearings. 2
Sliding flat surfaces wiping over ends of guides, such as reciprocating crosshead
shoes. (Use two for long and three for short guides.). 2- 3
Sliding or wiping surfaces such as marine thrust bearings and worm gears. 3- 4
Long nuts for power screws and similar wiping parts. 4- 6
The strength and stiffness of the bearing, bearing cap, and housing
should also be considered so that they will support the imposed loads
and will not deflect or distort beyond permissible limits. It is especially
important to give careful consideration to the deflection of bearings and
shafts that are required to support gears whose teeth must be kept accu¬
rately in mesh. This is especially true in bevel-gear mountings.
21-12 Thermal equilibrium. The heat generated in a bearing is due
to fluid friction or to friction of the parts having relative motion. This
heat can be readily determined when the coefficient of friction is known
Ih = fPV (21-9)
where Hi = heat generated, ft-lb per min

P = radial load on bearing, lb
V = rubbing velocity, fpm
After thermal equilibrium has been reached, heat will be dissipated at
the outer surface of the bearing at the same rate that it is generated in
the oil film. The heat dissipated can be expressed in terms of the product
of the dissipating area and the temperature drop between the surface and
the surrounding air.
For convenience in bearing design, the actual heat-dissipating area may
be expressed in terms of the projected area of the journal. Thus, the
equation becomes
H2 = CA(th - ta) (21-10)
where H2 = heat dissipated by the bearing, ft-lb per min

A = projected bearing area, Ld, in.2
tb = temperature of bearing surface, deg F
ta = temperature of surrounding air, deg F
to = temperature of oil film, deg F
c = heat-dissipation coefficient, ft-lb per min per sq in. of pro¬
jected bearing area per deg F
Values of C have been determined by O. Lasche. The values depend
on the type of bearing, its ventilation, and the temperature difference,
Figure 21-19 gives values of C(tb — ta).
Fig. 21-19. Heat dissipation for plain bearings.
Experiments1 show that the temperature at the outer surface of indus¬

trial bearings is approximately midway between the temperature of the
oil film and the ambient temperature, i.e.,
(tb — ta) = }i(t0 - to)

In some bearings, as in cold-water pumps, there may be conduction of
heat from the bearing, while in steam turbines there may be conduction
to the bearing. Because of the low heat conductivity of steel, however,
conduction by the shaft is generally relatively small.
1 M. D. Hersey, “Theory of Lubrication,” p. 102, John Wiley & Sons, Inc., New
York, 1938.
334 MACHINE DESIGN
21-13 Design of journal bearing. Because of the large number of vari¬

ables involved, the design of a journal bearing usually requires making
reasonable assumptions and then applying available equations to establish
their validity. The following procedure is suggested when the bearing
load and the diameter and speed of the shaft are known:
1. Determine the bearing length by choosing a ratio of L to d from
Table 21-1.
2. Check the bearing pressure p = P/Ld from Table 21-1 for probable
satisfactory value.
3. Assume a clearance ratio c/d from Table 21-1.
4. Assume a lubricant from Fig. 21-16, and an operating temperature
t0. This temperature should in general be between 80 and 140 F with
180 F as a maximum for high-temperature installations such as steam
turbines. The lower temperatures result in high operating viscosity and
resulting high power loss. Higher temperatures may cause film break¬
down, rapid vaporization and oxidation of the lubricant, varnishing of
the bearing, excessive side leakage, as well as the possibility of burning
the hands of the attendant.
5. Determine the operating value of ZN/p for the assumed bearing
temperature, and check this value with corresponding values in Table 21-1
to determine the possibility of maintaining fluid-film operation.
6. Determine the coefficient of friction by Eq. (21-5).
7. Determine the heat generated by Eq. (21-9).
8. Determine the heat dissipated by Eq. (21-10).
9. If thermal equilibrium is indicated by comparing the preceding two
steps, then the assumed bearing temperature is established, and the
operating conditions may be regarded as satisfactory.
10. If approximate equilibrium is not indicated, the designer must
assume a different operating temperature, a different oil, or changed
values of L and d, or artificial cooling if it is impossible to otherwise
obtain equilibrium.
The above analysis assumes that the bearing load is essentially constant
in magnitude and direction and that the speed is constant. If these con¬
ditions do not exist, the designer must consider the variations carefully
and arrive at a prediction of probable conditions of service in the light of
his own experience and that of others.1
Properly designed, installed, and operated journal bearings function
very satisfactorily; however, their specifications, including that of the
lubricant, require a careful and mature consideration of the numerous
general and special considerations involved in any service.
1 J. T. Burwell, The Calculated Performance of Dynamically Loaded Sleeve Bear¬
ings, J. Appl. Mechanics, vol. 14, no. 3, p. A-231, September, 1947.
Example 21-1. A journal bearing is proposed for a centrifugal pump. The data
are: load on journal = 8,770 lb, diameter = 6 in., length = 9% in., speed = 900 rpm,
diametral clearance ratio = 0.0015, ambient temperature = 60 F. Assume oil No. 9
from Fig. 21-16. For an assumed temperature of the oil film, t0 = 130 deg, deter¬
mine the heat generated (friction loss) and the heat dissipated.
Solution: For 130 deg, the specific gravity is
p = 0.8638
From Fig. 21-16,
S = 97 sec
From Eq. (21-7),
Z = 16.8 centipoises
8,770
Pressure p = = 150 psi
6 X 9.75
and
ZN 16.8 X 900
= 100.8
V 150
Item 27 in Table 21-1 shows that the pressure of 150 psi is within usual limits,
and the operating value of ZN/p given in the table indicates that the point of film
breakdown (c in Fig. 21-11) is of the order of one-fifth of the tabular value, i.e.,
X 200 = 40. Hence the bearing being investigated will be expected to operate
under hydrodynamic conditions and Eq. (21-5) may be used to estimate the coefficient
of friction. Hence
473 (ZN 473

/ =
1010 V V )© + K =
10 10 <100-8) (mm) + 0 002
= 0.00317 + 0.002 = 0.00517
7rd rpm 7r X 6 X 900
V = = 1,415 fpm
12 12
The heat generated, Eq. (21-9), is
Hl = fPV = 0.00517 X 8,770 X 1,415 = 6,300 ft-lb per min
The heat dissipated is

, , to - ta 130 - 60 Q_ ,
h — ta — —o— = -o- = 35 deg
From Fig. 21-19,
C(tb — ta) = 90 ft-lb per sq in. per min

Hi = CA(tb — ta) = 6 X 9.75 X 90 = 5,260 ft-lb per min
Thus the heat generated (6,300 ft-lb per min) is greater than the heat dissipated,
which indicates that the bearing is warming up. As its temperature rises, the heat
dissipated becomes greater, and the heat generated becomes less on account of the
decrease in viscosity (and friction) as the temperature becomes higher.
In Fig. 21-20 is shown a graph drawn for heat generated and heat dis¬
sipated for other temperatures than the assumed value of 130 F. The
intersection of the two curves shows that the equilibrium temperature is
approximately 137 F. From item 4 in Art. 21-13, this temperature
336 MACHINE DESIGN
would be reasonable and the choice of lubricant would be regarded as

satisfactory.
The chart shows also that the friction loss at the equilibrium tempera¬
ture is approximately 5,900 Tt-lb per min.
21-14 Systems of lubrication. There are various means employed to
supply the lubricant to bearings and to gears. It is important that the
proper method be selected. A brief description of some systems is given
here
1. Hand oiling. Suitable for slow speed and lightly loaded bearings.
2. Drop-feed or wick-feed oilers. Better than hand oiling, since the
supply of lubricant is more uniform.
Fig. 21-20. Thermal equilibrium for a 6 X 9%-in. journal bearing.
3. Oil rings or chains. Suitable for horizontal bearings. The ring or

chain is slipped over the shaft at the mid-point of the bearing, runs with
the shaft, and carries oil to the bearing from the reservoir into which it
dips. This type of lubrication has been very widely used for motors,
generators, and line shafts.
4. Splash system. In this system a rotating part, such as a gear, or a
rotating disk called an oil slinger, runs in an oil reservoir and causes a
shower of oil in the enclosed casing. The oil runs into channels and
ducts and is led to the bearings. A continuous supply of lubricant may
be secured by this method.
5. Pressure feeding. Bearings may be positively lubricated by intro¬
ducing the lubricant under pressure at the proper points of the bearing.
This method ensures an adequate supply of lubricant at a pressure suffi-
I
cient to prevent suction of air into the region of low pressure in bearings
operating under conditions of fluid-film lubrication. The flow of lubri¬
cant may be in such quantities that most of the heat generated in the bear¬
ing will be carried away by the lubricant itself, so that the viscosity of
the lubricant may thus be controlled. This may be necessary in heavily
loaded bearings.
6. Grease lubrication. In parts loaded heavily and with low speeds so
that fluid-film lubrication is not possible, grease may be used. Grease
may be used also where dripping or spattering of oil is not permissible,
as in food- or chemical-processing equipment.
7. Shaft seals. In order to prevent leakage of the lubricant from a
bearing and to prevent foreign matter from entering, shaft seals may be
necessary. Seals are available in many forms, such as felt or leather
inserted in grooves in the bearing housings, or they may be purchased sep¬
arately as a self-contained seal to mount in the housing.
CHAPTER 22
ROLLING CONTACT BEARINGS
22-1 General considerations. Ball and roller bearings are frequently

referred to as “antifriction bearings.” This is a misnomer, however,
since friction is always present in such bearings owing to rolling resistance
between the balls or rollers and the races and also to friction between the
retainers and contacting parts and to churning of any excessive lubricant.
For starting conditions and at moderate speeds, the friction of a ball or
roller bearing is lower than that of an equivalent journal bearing, but at
high speeds a properly designed and lubricated journal bearing has less
Fig. 22-1. Comparison of friction of bearings.
friction than an equivalent ball or roller bearing. The friction-speed rela¬

tions are shown qualitatively by Fig. 22-1.
Because of the kinematic line contact of rollers, the roller bearing is
generally used for large bearing loads in preference to ball bearings, which
have kinematic point contact.
Some advantages of ball and roller bearings as compared with journal
bearings are:
1. Friction is low except at high speeds.
2. Relatively accurate shaft alignment can be maintained.
3. Heavy momentary overloads can be carried.
4. Lubrication is simple.
338
ROLLING CONTACT BEARINGS 339
5. Both radial and axial loads can be carried by some types.
6. Replacement is easy.
7. Selection of bearing from manufacturers’ information is relatively
simple.
Some disadvantages are:
1. The expense is generally greater because of the cost of the bearing
and the necessary provision for mounting.
2. Failure of bearing can occur without warning and cause damage to
machinery.
22-2 Load ratings of ball and roller bearings.1 Because of the elastic
deformations of the surfaces of the regions of contact between a ball or
roller and the races, the contact
pressure is distributed over the
areas in contact. The extent of
the areas depends on the elastic
properties of the materials, on the
size of the balls or rollers, and on
the form of the raceways. In
order to provide a large area and
in turn to reduce the pressure in a
ball bearing, the raceways are Fig. 22-2. Contact areas in ball and roller
formed as shown schematically in bearings.
Fig. 22-2(a). The radii of curva¬
tures for the curves ab are made slightly greater than the radius of the
ball. The contact area in a roller bearing is shown in Fig. 22-2(6).
In a ball or roller bearing, the load-carrying capacity is based on the
surface endurance limit of the material. The rated load of a bearing is
the load it will carry for a specified life of the bearing when operating at a
definite speed. For other expected periods of life and for other speeds,
the load that the bearing should be required to carry must be altered
accordingly.
The theoretical basis for the design of ball and roller bearings is found
in the works of Hertz and Stribeck, but the commercial ratings and the
factors for speed, life, and shock are based mainly on performance data.
For satisfactory bearings, it is necessary to make the parts of high-grade
material properly heat-treated, accurately formed, and with finely finished
surfaces. Failure of bearings generally occurs because of pitting of the
balls or rollers and the raceways.
High-speed bearings, however, usually fail at the cage or retainer used
to space the rolling elements. The friction at this point is of sliding
1 See R. K. Allan, “Rolling Bearings,” Sir Isaac Pitman & Sons, Ltd., London, 1945,
for history, theory, design, and application of ball and roller bearings; A. Palmgren,
Ball and Roller Bearing Engineering, SKF Industries, Philadelphia, 1945.
340 MACHINE DESIGN
nature and failure is caused by relatively high loads at high speeds and
by inadequate lubrication.1 High-speed bearings in aircraft units may
operate at speeds over 20,000 rpm. It is customary to refer to the opera¬
tion of such bearings in terms of DN values, where D is the diameter of
the bore in millimeters and N is the rpm. DN values up to 3.5 X 106
are in development, for a bearing life of 1,000 hr.
22-3 Types of radial ball bearings. As shown in Fig. 22-3(a), the
parts of a radial ball bearing are the outer race c, the inner race d, the balls
e, and the retainer /. The retainer is usually in two parts which are
assembled after the balls have been properly spaced. The grooves which
form the path for the rolling elements are known as raceways. When a
ball bearing supports only a radial load R, as shown in Fig. 22-3(a), the
Fig. 22-3. Radial and thrust loads on ball bearings. Section lines are omitted in con¬
formity with general practice.
plane of rotation of the ball is normal to the center line of the bearing.
As shown at (6) in the figure, the action of a thrust load T is to shift the
plane of rotation of the balls. Both radial and thrust loads may be
carried simultaneously.
Many types of radial ball bearings have been developed. The general
features of some commonly used industrial types are discussed here.
Single-row deep groove bearings. In the assembly of this bearing, the
inner race is placed eccentric with the outer race, and then the balls are
inserted in the crescent-shaped space, as shown in Fig. 22-4. The balls
are then evenly spaced as the races are brought into concentric relation.
The halves of the retainer are then assembled so that the balls will remain
evenly spaced but can rotate freely. Even spacing of the balls is desir¬
able from the standpoints of distribution of the load and balance of the
bearing.
1 See Trends of Rolling-contact Bearings as Applied to Aircraft Gas Turbine

Engines, NACA TN 3110, April, 1954.
Radial ball bearing with filling notches. By employing a filling notch

in each of the races, as shown in Fig. 22-5, additional balls may be inserted
after the races are brought into concentric relation. The bottom of each
filling notch does not reach to the bottom of its raceway and therefore
the balls inserted through the notches must be forced in position. Some
bearings of this type, particularly small ones, have as many balls as the
Fig. 22-4. Assembly of deep-groove ball Fig. 22-5. Filling-notch type of radial ball
bearing. bearing.
annular space will contain and thus a retainer is not necessary to space the
balls. Generally less than the maximum number of balls are used and
therefore a retainer is necessary.
Since a bearing of this type contains a larger number of balls than a
corresponding unnotched one, it has a larger rated radial load capacity
but the filling notches reduce the allowable thrust load.
light Light Medium Heavy
Fig. 22-6. Ball-bearing series.
Proportions of radial ball bearings. Ball bearings are available in four

series for each standard bore as shown in Fig. 22-6. The extra-light and
light series are used where the loads are moderate and shaft sizes compara¬
tively large and also where available space is limited. The medium
series has a capacity 30 to 40 per cent over the light series. The heavy
series has 20 to 30 per cent capacity over the medium series. This series
is not used extensively in industrial applications.
Other types of radial ball bearings. Of the many types of radial ball
bearings made by the various manufacturers, the single-row type that
342 MACHINE DESIGN
Table 22-1. Rated Radial Loads for Single-row Deep-groove
Ball Bearings*
Bear- Revolutions per minute

mg
No. 50 100 300 500 720 1,000 1,200 1,800 2,500 3,600 5,000
3200 570 430 365 320 295 270 260 235 215 195 180
3300 760 640 485 430 390 360 345 310 285 260 240
3201 570 480 365 320 295 270 260 235 215 195 180
3301 985 825 630 555 505 465 445 400 370 340 310
3202 615 515 390 345 315 290 275 250 230 210 195
3302 1,230 1,040 785 695 630 580 555 500 460 420 390
3203 1,080 905 685 605 555 510 485 440 405 370 340
3303 1,500 1,260 960 845 770 710 680 615 565 515 475
3204 1,350 1,130 860 755 690 635 605 550 505 460 425
3304 2,150 1,800 1,370 1,210 1,100 1,010 965 870 805 735 680
3205 1,460 1,220 930 820 750 690 660 595 550 500 460
3305 2,350 1,970 1,500 1,320 1,210 1,110 1,060 960 885 805 740
3206 2,150 1,810 1,3-70 1,210 1,110 1,020 975 880 810 740 680
3306 3,110 2,620 1,990 1,750 1,600 1,470 1,400 1,270 1,170 1,070 985
3207 2,940 2,480 1,880 1,660 1,510 1,390 1,330 1,200 1,110 1,010 930
3307 3,850 3,240 2,460 2,170 1,980 1,820 1,740 1,570 1,450 1,320 1,220
3208 3,370 2,830 2,150 1,890 1,730 1,590 1,520 1,370 1,260 1,150 1,060
3308 4,650 3,910 2,970 2,610 2,390 2,200 2,100 1,900 1,750 1,600 1,470
3209 3,610 3,040 2,310 2,030 1,860 1,710 1,630 1,480 1,360 1,240 1,140
3309 5,440 4,580 3,480 3,060 2,790 2,570 2,460 2,220 2,040 1,870 1,720
3210 3,850 3,240 2,460 2,170 1,980 1,820 1,740 1,570 1,450 1,320
3310 6,350 5,340 4,060 3,570 3,260 3,000 2,870 2,590 2,390 2,180
3211 4,760 4,000 3,040 2,680 2,440 2,250 2,150 1,940 1,790 1,630
3311 7,170 6,030 4,580 4,040 3,680 3,390 3,240 2,930 2,700 2,460
3212 5,390 4,540 3,450 3,030 2,770 2,550 2,440 2,200 2,030 1,850
3312 7,990 6,720 5,100 4,490 4,110 3,780 3,610 3,260 3,010 2,740
3213 6,320 5,310 4,040 3,550 3,250 2,990 2,860 2,580 2,380 2,170
3313 8,860 7,450 5,660 4,980 4,550 4,190 4,000 3,620 3,330 3,040
* Based on average life of 3,800 hr. Type 3000 New Departure. (Abridged table.)
Table 22-1. Rated Radial Loads for Single-row Deep-groove Ball
Bearings (Continued)
Bear- Revolutions per minute

mg
No. 50 100 300 500 720 1,000 1,200 1,800 2,500 3,600 5,000
3214 6,730 5,660 4,300 3,790 3,450 3,180 3,040 2,750 2,530 2,310
3314 9,760 8,210 6,240 5,490 5,020 4,620 4,410 3,990 3,670 3,350
3215 6,730 5,660 4,300 3,790 3,450 3,180 3,040 2,750 2,530 2,310
3315 10,175 8,560 6,510 5,730 5,220 4,810 4,600 4,150 3,830 3,490
3216 7,250 6,100 4,640 4,080 3,720 3,430 3,280 2,960 2,730 2,490
3316 11,125 9,360 7,110 6,260 5,710 5,260 5,030 4,540 4,180
3217 8,870 7,460 5,670 4,990 4,550 4,190 4,000 3,620 3,330
3317 12,050 10,125 7,700 6,780 6,190 5,700 5,450 4,920 4,530
3218 9,870 8,300 6,310 5,550 5,070 4,670 4,460 4,030 3,710
3318 13,050 10,975 8,340 7,340 6,700 6,170 5,890 5,330 4,910
has been described is in most common industrial use. Other types are
shown in section in Fig. 22-7.
22-4 Selection of radial ball bearings. In determining the type and
size of radial ball bearings for a particular installation, the data and pro¬
cedure used should be that recommended by the company from whom the
bearings are to be purchased. The selection procedures used by the var¬
ious companies are similar in general but vary in detail. For the purpose
of illustration, a typical procedure is outlined here.
*- <b—1
J L
Single
\_
cx
Snap
-J
row ring row ring
ta
v__/
Shielded
—...
s..
a
Self
aligning aligning
Fig. 22-7. Types of radial ball bearings.
The rated radial-load capacities for light (200) and medium (300) series
of the deep-groove type of bearing for various speeds are given in Table
22-1. The table applies to bearings in which the inner race rotates.
344 MACHINE DESIGN
The data that should be known for the selection of the bearing are as
follows:
Radial load, lb
Thrust load, lb
Speed of bearing, rpm
Desired life of bearing, hr
Conditions of loading
The rated, radial load of a bearing to comply with these conditions may
be found from the following equation; then a bearing may be selected
from Table 22-1.
C = RFLS (22-1)
where R = radial load on bearing, lb
T = thrust load on bearing, lb
F = thrust factor (Table 22-2)
L = life factor (Fig. 22-8)
S = service factor (Table 22-3)
C = rated radial load, lb (Table 22-1); if rating in Table 22-1 is
not given for specified operating speed, Table 22-4 may be
Average life, hours

Fig. 22-8. Life factor L.
used for interpolation by multiplying the load rating in

Table 22-1 at 1,000 rpm by the speed factor Fs, from Table
22-4, for the operating speed
Table 22-2. Thrust Factor F
T/R F T/R F T/R F T/R F T/R F
0.05 1.01 0.30 1.12 0.60 1.37 1.25 2.02 4.00 4.76
0.10 1.02 0.35 1.16 0.70 1.46 1.50 2.27 5.00 5.77
0.15 1.04 0.40 1.20 0.80 1.56 1.75 2.52 7.50 8.27
0.20 1.06 0.45 1.24 0.90 1.67 2.00 2.77 10.00 10.77
0.25 1.09 0.50 1.28 1.00 1.77 3.00 3.77
Table 22-3. Service Factors S
Service S
Uniform and steady load. 1.0
Light shock load. 1.5
Moderate shock load. 2.0
Heavy shock load. 2.5
Extreme and indeterminate load. 3.0
Table 22-4. Speed Factors Fs
Rpm Fs Rpm Fs Rpm Fs Rpm Fs Rpm Fs Rpm Fs Rpm Fs
270 1.387 825 1.049 1,725 0.8726 3,250 0.7448 5,100 0.6654 8,700 0.5823
10 3.162 280 1.375 850 1.041 1,750 0.8695 3,300 0.7419 5,200 0.6622 8,800 0.5806
15 2.858 290 1.363 875 1.034 1,775 0.8664 3,350 0.7392 5,300 0.6591 8,900 0.5790
20 2.659 300 1.351 900 1.027 1,800 0.8633 3,400 0.7364 5,400 0.6560 9,000 0.5774
25 2.515 310 1.340 925 1.020 1,825 0.8604 3,450 0.7337 5,500 0.6530 9,100 0.5758
30 2.403 320 1.330 950 1.013 1,850 0.8575 3,500 0.7311 5,600 0.6501 9,200 0.5742
35 2.312 330 1.320 975 1.006 1,875 0.8546 3,550 0.7285 5,700 0.6472 9,300 0.5726
40 2.236 340 1.310 1,000 1.000 1,900 0.8518 3,600 0.7260 5,800 0.6444 9,400 0.5711
45 2.171 350 1.300 1,025 0.9938 1,925 0.8490 3,650 0.7235 5,900 0.6416 9,500 0.5696
50 2.115 360 1.291 1,050 0.9878 1,950 0.8462 3,700 0.7210 6,000 0.6389 9,600 0.5681
55 2.065 370 1.282 1,075 0.9821 1,975 0.8436 3,750 0.7186 6,100 0.6363 9,700 0.5666
60 2.021 380 1.274 1,100 0.9765 2,000 0.8409 3,800 0.7162 6,200 0.6337 9,800 0.5652
65 1.981 390 1.265 1,125 0.9710 2,050 0.8357 3,850 0.7139 6,300 0.6312 9,900 0.5637
70 1.944 400 1.257 1,150 0.9657 2,100 0.8307 3,900 0.7116 6,400 0.6287 10,000 0.5624
75 1.911 410 1.250 1,175 0.9605 2,150 0.8258 3,950 0.7093 6,500 0.6263
80 1.880 420 1.242 1,200 0.9554 2,200 0.8211 4,000 0.7071 6,600 0.6239
85 1.852 430 1.235 1,225 0.9506 2,250 0.8165 4,050 0.7049 6,700 0.6215
90 1.826 440 1.228 1,250 0.9457 2,300 0.8120 4,100 0.7027 6,800 0.6193
95 1.801 450 1.221 1,275 0.9411 2,350 0.8077 4,150 0.7006 6,900 0.6170
100 1.778 460 1.214 1,300 0.9365 2,400 0.8034 4,200 0.6985 7,000 0.6148
110 1.736 470 1.208 1,325 0.9321 2,450 0.7993 4,250 0.6965 7,100 0.6126
120 1.699 480 1.201 1,350 0.9277 2,500 0.7953 4,300 0.6944 7,200 0.6105
130 1.665 490 1.195 1,375 0.9235 2,550 0.7914 4,350 0.6924 7,300 0.6084
140 1.635 500 1.189 1,400 0.9193 2,600 0.7875 4,400 0.6905 7,400 0.6063
150 1.607 525 1.175 1,425 0.9153 2,650 0.7838 4,450 0.6885 7,500 0.6043
160 1.581 550 1.161 1,450 0.9113 2,700 0.7801 4,500 0.6866 7,600 0.6023
170 1.557 575 1.149 1,475 0.9074 2,750 0.7765 4,550 0.6847 7,700 0.6003
180 1.535 600 1.136 1,500 0.9036 2,800 0.7731 4,600 0.6828 7,800 0.5984
190 1.515 625 1.125 1,525 0.8999 2,850 0.7696 4,650 0.6810 7,900 0.5965
200 1.495 650 1.114 1,550 0.8962 2,900 0.7663 4,700 0.6792 8,000 0.5946
210 1.477 675 1.103 1,575 0.8926 2,950 0.7630 4,750 0.6774 8,100 0.5928
220 1.460 700 1.093 1,600 0.8891 3,000 0.7598 4,800 0.6756 8,200 0.5910
230 1.444 725 1.084 1,625 0.8857 3,050 0.7567 4,850 0.6738 8,300 0.5892
240 1.429 750 1.075 1,650 0.8823 3,100 0.7536 4,900 0.6721 5,400 0.5874
250 1.414 775 1.066 1,675 0.8790 3,150 0.7506 4,950 0.6704 8,500 0.5856
260 1.400 800 1.057 1,700 0.8758 3,200 0.7477 5,000 0.6687 8,600 0.5840
346 MACHINE DESIGN
Table 22-5. Principal Dimensions for Radial Ball Bearings
Bore Diameter Width Balls

Brg. Radi-
No. ins
mm in. mm *. in. mm in. Diam. No.
plain r
3200 30 1.1811 9 0.3543 7

10 0.3937 0.025
3300 35 1.3780 11 0.4331 H 7
3201 32 1.2598 10 0.3937 0.210 8 0.025

12 0.4724 1.4567 12 0.4724 7 0.04
3301 37 A2
3202 35 1.3780 11 0.4331 0.210 9 0.025

3302 15 0.5906 42 1.6535 13 0.5118 7 0.04
He
3203 40 1.5748 12 0.4724 %2 8 0.025

17 0.6693 47 1.8504 14 0.5512 7 0.04
3303 %
3204 47 1.8504 14 0.5512 He 8 0.04
20 0.7874 52 2.0472 15 0.5906 7
3304 1%2
3205 52 2.0472 15 0.5906 He 9 0.04
25 0.9843 62 2.4409 17 0.6693 8
3305
3206 30 1.1811
62 2.4409 16 0.6299 % 9 0.04
3306 72 2.8346 19 0.7480 Y%2 8
3207 72 2.8346 17 0.6693 He 9 0.04

35 1.3780 80 3.1496 21 0.8268 8 0.06
3307 ll/32
3208 80 3.1496 18 0.7087 1/'3 2 9 0.04

40 1.5748 90 3.5433 23 0.9055 8 0.06
3308 19A2
3209 85 3.3465 19 0.7480 15/32 10 0.04

45 1.7717 100 3.9370 25 0.9843 8 0.06
3309 21/S2
3210 90 3.5433 20 0.7874 1A 2 11 0.04

50 1.9685 27 1.0630
3310 110 4.3307 23A 2 8 0.08
3211
55 2.1654
100 3.9370 21 0.8268 % 11 0.06
3311 120 4.7244 29 1.1417 2%2 8 0.08
3212 110 4.3307 22 0.8661 1%2 10 0.06

60 2.3622
3312 130 5.1181 31 1.2205 % 8 0.08
3213 120 4.7244 23 0.9055 21/32 10 0.06

65 2.5591 140 33 1.2992
3313 5.5118 2%2 8 0.08
3214 70 2.7559
125 4.9213 24 0.9449 % 11 0.06
3314 150 5.9055 35 1.3780 % 8 0.08
3215 75 130 5.1181 25 0.9843 % 11 0.06

2.9528 160
3315 6.2992 37 1.4567 1 8 0.08
3216 140 5.5118 26 1.0236 % 11

80 3.1496 0.08
3316 170 6.6929 39 1.5354 1 He 8
3217 150 5.9055 28 1.1024 25A 2 11 0.08

85 3.3465
3317 180 7.0866 41 1.6142 1 H 8 0.10
3218 90 160 6.2992 30 1.1811 2VS2 11 0.08

3318 3.5433 190 7.4803 43 1.6929 1 He 8 0.10
In many cases it is found that a bearing selected on the basis of load¬

carrying capacity has a bore too small to allow it to be mounted on its
shaft so that it is necessary to select a bearing on the basis of dimensions,
even though it has excessive load-carrying capacity. The dimensions of
bearings referred to in Table 22-1 are given in Table 22-5. Fortunately,
the dimensions of rolling-contact bearings are standardized. Inter¬
changeability tables are available for various makes of bearings.
22-5 Installation of radial ball bearings. A common type of ball¬
bearing installation is that in which the bearing is used to support a
rotating shaft. The inner race is fitted to the shaft with a light press
fit and may be held firmly on the shaft by a retaining nut or other positive
means. All fits and clearances should
follow standards. The inner-race fit on
the shaft should be sufficient to give a firm
mounting but not too tight a fit because
this may deform the inner race and destroy
* the clearance between the rolling elements
and the raceways. The outer race should
be mounted with a firm fit in the housing.
If it is not tight enough, the outer race may Fig. 22-9. Shoulder heights and
tend to “ creep.” This is undesirable in fillet radii.
that it may cause “fretting.” Failure
records of ball bearings with restrained outer races1 show that 60 per cent
are inner-race failures; 20 per cent are ball failures; 10 per cent are
retainer and outer-race breakdowns.
The reason that inner-race failures are more prevalent than failures
of the outer race is that the contact pressure is greater at the inner race¬
way where it curves away from the ball than at the outer raceway where
it curves with the ball.
In mounting the bearing, the housing shoulder and the shaft shoulder
(Fig. 22-9) should follow standards.
A typical mounting with cap and shaft seal is shown in Fig. 22-10.
The bearing at the left-hand end of the shaft at (a) will take thrust in
either direction in addition to a radial load. The shaft should have end
play of 0.002 to 0.01 in., depending on the size of the bearing.
The outer race of the bearing on the right-hand end of the shaft at (b)
should be free to move axially to avoid cramping of the bearings because
of axial deformation of the shaft produced by thrust loads or expansion
caused by temperature changes.
22-6 Ball thrust bearings. Where heavy thrust loads are to be carried,
for example, in worm drives and vertical shafts, ball thrust bearings are
1 Irving Kalikow, Creep of Ball Bearing Races, Product Eng., vol. 20, no. 4, April,
1949.
348 MACHINE DESIGN
applicable. Sections of three types of ball thrust bearings are shown in

Fig. 22-11. The mounting of the worm shaft shown in Fig. 22-12 is an
application of a double-direction ball thrust bearing to carry the axial
load and of radial roller bearings to carry the radial loads.
Fig. 22-10. Ball-bearing mounting.
den
(c)- Double
direction
Fig. 22-11. Ball thrust bearings.
Fig. 22-12. Mounting for worm shaft.
22-7 Radial roller bearings. Norma Hoffman and SKF type. This
bearing has cylindrical rollers that run in cylindrical raceways. In a
common construction (see Fig. 22-13) the inner race is provided with
retaining shoulders, but the outer raceway is straight. This construction
permits axial movement of the shaft. The bearing is interchangeable

with ball bearings of corresponding size, and has a radial load capacity
100 per cent higher than the ball bearing; but it will not carry a thrust
load.
A common type of shaft mounting where a load is carried by a shaft
extension is shown in Fig. 22-14. The roller bearing carries the heavy
1 t
Fig. 22-13. Radial roller bearing. Fig. 22-14. Mounting for shaft with
extension.
radial load while the ball bearing carries the lighter radial load and serves
to locate the shaft axially.
The Hyatt roller bearing has hollow cylindrical rollers, each made by
winding a flat strip of steel in the form of a helix. The flexibility pro¬
vided by this construction allows the rollers to adjust themselves to
slight irregularities in the raceways. It is thus possible to run the rollers
directly on the hardened and ground shaft, thus saving radial space.
(a)-Torrington (b)~ Bantam

Fig. 22-15. Hyatt roller bearing. Fig. 22-16. Needle, or quill, bearings.
The rollers are wound alternately right- and left-hand in order to aid in
the distribution of the lubricant along the rollers.
Needle bearings. These are roller bearings in which the rollers are small
in diameter and relatively long. They run without retainers, as shown
in Fig. 22-16. The rollers, or “ needles,” may run directly on the shaft.
This type of bearing is used where space in a radial direction is limited.
Two types are available, one with a drawn shell in which the rollers
usually run directly on the hardened shaft, and the other for greater
350 MACHINE DESIGN
capacities has a heavy outer race and usually an inner race [see Fig.
22-16(a) and (6)].
22-8 Angular roller bearings. The angular, or tapered, roller bearing
combines the high radial load-carrying capacity of a roller bearing with a
thrust capacity in one direction. Since at least two bearings are neces¬
sary to support most shafts, angular bearings can be
so mounted that one will support the thrust load in
one direction and the other in the opposite direction.
As shown in Fig. 22-17, this bearing has four
parts: (a) the inner race or cone, (b) the outer race
or cup, (c) the rollers, and (d) the retainer. The
principle of operation of the bearing, as shown in
Fig. 22-18, is that the elements of the conical surfaces
Fig. 22-17. Angular
of the rollers and of the raceways intersect at a
roller bearing. common apex on the center line of the bearing, so
that pure rolling motion will exist between the rollers
and the raceways. The taper of the rollers is only a few degrees.
Since the conical surface of each roller is subjected to pressures acting
normal to the surfaces, as shown in Fig. 22-19, there will be a resultant
force tending to force the roller from its place between the races. To
retain the roller in place it is necessary to provide a rib on the cone as
shown at e in Fig. 22-17. The contact between the large end of the roller
and the rib introduces sliding friction in the bearing, which in well-
designed bearings is usually small. A typical mounting using tapered
roller bearings is shown in Fig. 22-20.
rrrn
Fig. 22-18. Elements of conical sur¬ Fig. 22-19. Forces on tapered roller.
faces intersect axis of bearing.
22-9 Lubrication of ball and roller bearings. The pressures existing

between the balls or rollers and the raceways are so high that a fluid film
of lubricant will not generally exist in loaded bearings. In order to assist
in the maintenance of an adsorbed film between the load-carrying parts
and thus prevent undue friction and wear, the surfaces of the balls or
rollers and the raceways must be finely finished and the finish must be
maintained throughout the life of the bearing.
The functions of the lubricant in ball and roller bearings are to (a) pre¬
vent rust and corrosion; (b) aid in preventing entrance of foreign matter,
such as water, dust, etc.; (c) reduce friction and wear between rubbing
parts of the bearing; (d) dissipate heat.
Oils and greases are the most common lubricants for ball and roller
bearings. Oil may be used where the lubricating service is dependable
and the loads are relatively light, and also where the unit operates con¬
tinuously and the loads and speeds are such that it is necessary to make
use of a circulating oil-feed system to carry away the heat generated.
In order to prevent excessive churning of the oil, especially at high speeds,
the level of the oil for a bearing of a horizontal shaft should be near the
top of the lowest ball.
Fig. 22-20. Worm-gear shaft mounted on Timken tapered-roller bearings.
Greases are used where the loads are high or where it is desired to
extend the lubrication periods because of service conditions or because of
the inaccessibility of bearings. In many cases regreasing of bearings is
required every 6 to 24 months, depending on the service and the type of
grease. To allow space for grease expansion due to temperature change
and to prevent churning, it is recommended that the bearing be packed
half full of grease.
22-10 Shaft seals and bearing shields. To prevent loss of the lubri¬
cant from a ball or roller bearing or from the housing in which it is
mounted and to aid in preventing the entrance of foreign matter as water
and dust, a seal or shield should be used. The seal or shield may be an
integral part of the bearing, it may be incorporated in the bearing cover,
or it may be a separate self-contained seal. Examples of seals and shields
are shown in Fig. 22-21. Many other types and arrangements of shaft
352 MACHINE DESIGN
and bearing seals are available or may be devised to suit the individual
installation.1
-
-\
^-Shielded (fJ-Sea\ed
seal bearing bearing
Fig. 22-21. Shaft seals and bearing shields.
Example 22-1. Select a single-row, radial-load ball bearing (Type 3000) for a
radial load of 800 lb, a thrust load of 1,000 lb, operating at a speed of 1,600 rpm for an
average life of 5 years at 10 hr per day. Assume steady load with no shock.
Solution: Try a light-series (200 series) bearing:
R 800 lb
T 1,000 lb
T 1,000
1.25
R 800
F 2.02 (Table 22-2)
L 1.5 for 18,250 hr (Fig. 22-8)
S 1 (Table 22-3)
c RFLS = 800 X 2.02 X 1.5 X 1 = 2,425 lb
An inspection of Table 22-1 for approximately 1,600 rpm shows that a No. 3213
bearing should be satisfactory. This bearing has a rating at 2,990 lb at 1,000 rpm.
By using Table 22-4, the speed factor Fs for 1,600 rpm is 0.8891; hence the load rating
of the No. 3213 bearing at 1,600 rpm is 2,990 X 0.8891 = 2,658 lb. (A check on the
next smaller bearing, No. 3212, shows that its load rating at 1,600 rpm is 2,270 lb,
which is under the required rating.) Hence, the No. 3212 bearing would be a good
choice.
The bore of bearing No. 3213 (from Table 22-5) is 2.5591 in., which should be
checked against the shaft size at the bearing location.
1 Nordenholt, Kerr, and Sasso, “Handbook of Mechanical Design,” McGraw-Hill

CHAPTER 23
METAL FITS AND TOLERANCES
23-1 Dimensional control. Two important facts in the control of

dimensions of parts that come in contact or near contact are that (a) the
more exact the dimensions of the parts, the more satisfactory will be the
characteristics of the machine, and (6) the more exact the dimensions of
the parts, the higher will be the production cost of the machine.
The exactness of dimensions may be fixed by such requirements as
space, leakage, lubrication, assembly or interchangeability, limits to
impact loads, expansion or contraction, deflection, and pressure required
in press or shrink fits.
In addition to establishing correct dimensions for the parts of a machine
being designed, it is necessary for the mechanical-design engineer to select
tolerances for the dimensions of the various parts.1 It has been aptly said
that tolerances make or break a machine. Too large tolerances may
affect performance and life adversely. Too close tolerances may mean
excessive cost (see Fig. 23-1).
Imagine the difficulty in assemblying the parts of an automobile engine
if each part was made in a separate department which used for measure¬
ment the sixteenth-century rod, which was defined as “the total length
of the left feet of the first sixteen men who came out of church on Sunday
morning.” Imagine the performance of such an engine if it were assem¬
bled. Tolerances today in critical engine elements are less than 0.0001 in.
Some idea of the relative sizes involved is indicated in the comparison
that one ten-thousandth of an inch is to an inch as the diameter of a
human hair is to the diameter of an oil drum.
23-2 Metal fits. If the diameter of a member is specified as in.
by the designer, there may be some uncertainty as to the meaning of the
size. The degree to which the diameter of the member should conform
to exactly V/i in. depends on the use of the member. If it is to be a rod
for holding up a shelf, then, obviously, a high degree of accuracy is not
required. If, however, the member is a journal that must run smoothly
1 R. W. Bolz, Design Considerations for Manufacturing Economy, Trans. ASME

Payer 49-A-53, 1949.
353
354 MACHINE DESIGN
in a bearing, then it is necessary to hold the dimension within close

limits.
The basic dimension is the nominal value of a dimension from which
variations are permitted. vTolerance is the maximum permissible amount
of variation in the size of the part. For instance, the size may be desig¬
nated as 1.500 ± 0.010, which indicates a tolerance of 0.020 in. The
tolerance may be bilateral, in which case variation is permitted so that
the part may be either larger or smaller, as in the previous example, or
it may be unilateral, in which case the part may be either larger or smaller
Fig. 23-1. Relative cost of accuracy. (From R. W.Bolz, Design Considerations for
Manufacturing Economy, Trans. ASME Payer 49-A-53, 1949.)
than the basic dimension, but not both. An example of unilateral toler¬
ance is 1.500 Unilateral tolerances are usually used when a
metal fit is to be obtained.1

In order to indicate the snugness or looseness of a fit, it is necessary
to determine the clearance between the parts. Clearance is the difference
in size between mating parts. It is apparent that because of the tolerance
of the parts, the clearance can vary from a maximum to a minimum. The
minimum clearance is called the allowance. The various dimensions are
indicated in Fig. 23-2.
1 E. Buckingham, “Production Engineering/’ John Wiley & Sons, Inc., New York,
1942.
METAL FITS AND TOLERANCES 355
It is convenient to refer to various kinds of fits. The ASA1 has

approved the following classification:
Loose fit (class 1). Large allowance. Intended for use where accu¬
racy is not essential, such as in some agricultural machinery and road¬
building equipment.
Free fit (class 2). Liberal allowance. Suitable for use on rotating
journals where the speeds are 600 rpm or greater.
Medium fit (class 3). Medium allowance. For use with running fits
under 600 rpm and for sliding fits.
-Size for basic dimension

-Size with tote rat nee
A = Basic diameter of shaft
B = Basic diameter of hole
B-A -Allowance
B~A'-Maximum clearance
A~ATolerance on shaft
B~B - Tolerance on hole
Fig. 23-2. Shaft and hole dimensions.
Snug fit (class 4). Zero allowance. Closest fit that can be assembled
by hand to be used where very small play is permissible.
Wringing fit (class 5). Perceptible negative allowance. Selective
assembly with small pressure required for assembly.
Tight fit (class 6). Slight negative allowance. Small pressure
required for assembly, where the assembly is usually considered semi¬
permanent. Suitable for drive fits or shrink fits on light sections.
Medium force fit (class 7). Negative allowance. Pressure is required
for assembly, where the assembly is considered permanent. Used for
shrink fits on medium sections or long shafts and is the tightest fit that
it is safe to use with cast-iron external members. Suitable for use in
fastening locomotive wheels, car wheels, and motor armatures.
1 ASME Engineering and Industrial Standards, ASME B4a, 1925; International
Federation of the National Standardizing Associations, ISA Bull. 25, 1941; John
Gaillard, ASA Tolerances for Cylindrical Fits, 1941; Ordnance Manual on Dimension¬
ing and Tolerancing, Office of Chief of Ordnance, 1945; Kent’s “ Mechanical Engineers’
Handbook, Design and Production,” 12th ed., sec. 24, John Wiley & Sons, Inc.,
New York, 1952.
356 MACHINE DESIGN
Heavy force fit (class 8). Considerable negative allowance. Used as
force or shrink fits for steel external members where considerable bond is
required as on locomotive wheel tires.
The recommended allowances and tolerances for the various classes of
fits are given in Table 23-1.
Table 23-1. Allowances and Tolerances
(d = Nominal Diameter of Hole)
Selected
Method of average Hole Shaft
Class of fit Allowance
assembly interference tolerance tolerance
of metal
1. Loose Strictly inter¬ 0.0025d* + 0.0025d* - 0.0025d*

changeable
2. Free Strictly inter¬ 0.0014d3 + 0.0013d* - 0.0013d3
changeable
3. Medium Strictly inter¬ 0.0009d* + 0.0008d* - 0.0008d*
changeable
4. Snug Strictly inter¬ 0.0000 + 0.0006d* - 0.0004d*
changeable
5. Wringing Selective 0.0000 + 0.0006d* + 0.0004d*
assembly
6. Tight Selective 0.00025d + 0.0006d* + 0.0006d*
assembly
7. Medium Selective 0.0005d + 0.0006d* + 0.0006d*
force assembly
8. Heavy Selective O.OOld + 0.0006d* + 0.0006d*
force or assembly
shrink
Example 23-1. Determine the allowance, tolerance, and size of a U+in. shaft and
hole for a free fit.
Assuming that the basic diameter of the hole is 1.500 in., the allowance, from Table
23-1, is
Allowance = 0.0014^ = 0.0014 X (1.5)*
= 0.00184 say 0.002 in.
The tolerance for hole and shaft is
Tolerance = 0.0013d* = 0.0013 X (1.5)*
= 0.00149 say 0.001 in.
Maximum diameter of shaft = nominal size — allowance
= 1.500 - 0.002 = 1.498 in.
+0.000
Using unilateral tolerance, the shaft diameter is specified as 1.498 or
-O.OOl’
between 1.498 and 1.497 in. The diameter of the hole is 1.500 q qoq’ °r ê^ween
1.500 and 1.501 in. The maximum clearance is therefore 1.501 minus 1.497, or
0.004 in., whereas the minimum clearance is 0.002 in.
Selective assembly.1 When the difference between allowance and max¬

imum clearance must be small, the tolerances on assembled parts as a
shaft in a hole may become very small with corresponding high cost of
production. Saving may be realized by using larger tolerances and
assemblying the parts selectively by hand or otherwise. This method
has the disadvantage, however, of high cost of sorting and loss of complete
interchangeability.
23-3 Force fits and shrink fits. Tight fits, medium force fits, and
heavy force, or shrink, fits all have negative allowances. In these cases
the diameter of the hole is less than the diameter of the shaft. Thus there
is interference of the metal so that assembly must be accomplished by
forcing one part over the other, or by shrinking the external member on
the internal member. These fits are usually used to prevent slipping of
the two members relative to each other. For large interference, the
stresses induced in the members may be very high; in some cases the
stresses may be sufficiently high to cause failure of the external member.
The maximum stresses induced in a thick cylinder by an internal pres¬
sure are given by the Lame equations, Eqs. (11-3) and (11-4). In a
force-fit assembly, the pressure between the parts depends on the amount
of interference. The maximum tangential stress and radial stress at the
,
inside surface of the external member are
s, (max)
v
P
[1 + (d?/dl)\ /OQ
(23-1)
^
sr = —p (23-2)
where di = nominal internal diameter of hub, in.

d0 = external diameter of hub, in.
p = contact pressure, psi
The contact pressure between the external member and a solid shaft
of the same material is shown to be2
V = (23-3)
where 5 = interference, in.

By substituting the value of p given by Eq. (23-3) into Eqs. (23-1)
and (23-2), the principal stresses st and sr become equal to
1 See M. F. Spotts, “Design of Machine Elements,” 2d ed., p. 42b, Prentice-Hall,

2 S. Timoshenko, “Strength of Materials,” pt. II, D. Van Nostrand Company,
Inc., New York, 1941; J. H. Faupel, Designing for Shrink Fits, Machine Design,
January, 1954.
358 MACHINE DESIGN
(23-4)
(23-5)
The maximum shear stress is obtained by combining the foregoing

principal stresses
Ed
s8 (max) = (23-6)
2 di
The allowable stress depends on the material used. As before, for

brittle material, the maximum stress should not exceed the ultimate ten¬
sile strength of the material. Equating the allowable stress for brittle
material to the maximum induced tensile stress in the hub, Eq. (23-4),
gives
_&u_ _ Ed^ / d%\
(23-7)
f.s. 2di V1 + d2J
where f.s. = factor of safety

su = ultimate tensile strength of material, psi
The limiting value of the interference for any fit can therefore be
determined.
The allowable shear stress for ductile material based on the maximum-
shear theory of failure is given by Eq. (5-1). By equating this stress to
the maximum shear stress given in Eq. (23-6), the conditions for yielding
of a ductile hub are indicated:
sy Ed
(23-8)
f.s. di
where sy = yield point, psi.
Example 23-2. Determine the maximum allowable interference between a steel
shaft and hub if yielding is to be prevented. Assume steel with a yield point of 30,000
psi. Using Eq. (23-8) with a factor of safety of 1, since the yield conditions are to be
found, then
Sydi _ 30,000d{
0.001d»
~W ~ 30,000,000
It should be noted that this interference is the selected average interference for a
heavy force or shrink fit (class 8), as indicated in Table 23-1. If this class of fit is used
with a ductile material with a yield point less than 30,000 psi, yielding will occur.
Class 8 fit is usually not recommended for brittle material because of the severe stress
conditions.
23-4 Force fits—steel shaft with cast-iron hub. The use of unlike
materials with different moduli of rigidity will yield a different pressure
relation than indicated by the previous article. A safe equation for the
contact pressure can be obtained if the ratio of the moduli of elasticity of
steel and cast iron is assumed to be 3. A reduction of the Timoshenko1
equation for a fit of two materials on this assumption gives
EA1 - 0i]/dl)]
(23-9)
di[1.53 + 0.47(d2/d2)]
where Ec is the modulus of elasticity of cast iron. The maximum stress

is found by substituting this value for p in Eq. (23-1). Equating the
results to the allowable stress for brittle material gives
^ = Ecd[ 1 + (<#/<£)]
(23-10)
f.s. dil 1.53 + 0.47(d2/d2)]
Example 23-3. Determine the maximum average interference between a steel shaft
and a cast-iron hub if the outer hub diameter is twice the inner diameter. Assume a
grade of cast iron with an ultimate tensile strength of 30,000 psi and a modulus of
elasticity of 14,000,000 psi.
From Eq. (23-10), using the factor of safety as unity,
= M4E53 + 0.47 (d]/dp]
Ec[ 1 + (dl/dl)]
30,000 X d;[1.53 + 0.47 X (0.5)2]
14,000,000 X [1 + (0.5)2]
= 0.000283d;
This interference is less than the selected average for a medium force fit (class 7).
23-5 Holding ability of force and shrink fits. The torque that can be
transmitted by a shrink or force fit without slipping between the hub and
shaft can be estimated. Assuming that the contact pressure p is uni¬
formly distributed, the total radial force between the surfaces of length
L in contact is pivdiL. The tangential force due to friction is the coeffi¬
cient of friction times the radial force. The torque is the tangential force
times the radius, or
= fpTd%
2
where T = torque transmitted, lb-in.
p — contact pressure, psi
di = diameter of shaft, in.
L = length of hub, in.
/ = coefficient of friction, usually from 0.1 to 0.05
23-6 Assembly of shrink fits. Whenever fits are used with large inter¬
ference or large shaft sizes, the force required to press the hub on the
1 Timoshenko, op. cit.
360 MACHINE DESIGN
shaft is too large for practical assembly. To obviate this difficulty,

assembly is frequently accomplished by shrinking the hub on the shaft.
The hub is heated until it can be slipped on the shaft, whereupon it
shrinks as it cools until a'fit is obtained according to the interference
specified on the cold parts. The inside diameter of the hub must be
expanded by an amount at least as much as the interference before it is
slipped on the shaft. Since a uniform heating of the hub increases all
its dimensions, the increase in inside diameter for a temperature change
AT is equal to adi AT, where a is the coefficient of thermal expansion.
Therefore the minimum change in temperature for assembly is
(23-11)
where 8 = diametral interference, in.

a = coefficient of expansion, in. per in. per deg F
AT = change in temperature, deg F
In assembly, the inside part may be cooled either as an adjunct to or
a substitute for heating of the external part. “Dry ice” (solid carbon
dioxide) is widely used for this purpose, and temperature differences of
100 F are easily obtainable by cooling alone.
CHAPTER 24
VIBRATION AND VIBRATION CONTROL
24-1 Introduction. The cause of vibration in machinery is the disturb¬

ing forces of reciprocating masses, unbalanced rotating masses, and vari¬
able fluid pressure. Frequently the disturbing forces can be removed or
balanced; if not, their harmful effects can be reduced by proper design
of the parts and their mountings.
The simplest vibrating system has one degree of freedom, i.e., the
system’s position can be specified by one number. This is illustrated by a
weight suspended by a spring so that it can move only vertically; the
weight can be located at any instant by its distance (one number) from a
reference point. If the weight could rotate about a vertical axis at the
same time that it moves up and down, it would require an additional
number to specify its angular position at any instant, and would represent
two degrees of freedom. In general, if it requires n numbers to specify
the position of a vibrating system, the system has n degrees of freedom.
The single-degree-of-freedom system will be considered here. Many
mechanical systems are or can be reduced to this system, and the prin¬
ciples for the simple system can be extended to apply to more complicated
ones.1
In our discussion, two frequencies are involved; one is the frequency
of the disturbing force. This frequency is equal to or related to the oper¬
ating speed of the machine, which is generally specified and which cannot
be altered by the designer.
Another frequency, independent of the operating speed, is the natural
frequency of vibration of the system. This frequency depends on the
mass of the system and the stiffness of its support.
When the frequency of the disturbing force on a system coincides with
its natural frequency, resonance occurs, and the system may vibrate with
large amplitudes and develop forces larger than the disturbing forces.
These forces may cause annoyance to personnel and failure of parts.
Vibration control consists of removing or balancing the disturbing
forces, as mentioned, or of changing the natural frequency of vibration so
1 J. P. Den Hartog, “Mechanical Vibrations/’ 3d ed., McGraw-Hill Book Com¬
pany, Inc., New York, 1947.
361
362 MACHINE DESIGN
that it is sufficiently remote from the operating frequency to avoid unsat¬

isfactory operating conditions.
24-2 Equation of motion for a system having a single degree of free¬
dom. Consider a spring-supported weight arranged so that the weight
can move only in the vertical direc¬
V/////////A tion, as shown in Fig. 24-1. The
disturbing force is assumed to be
P0 sin oot, which is impressed ver¬
tically at a time t by an eccentric
mass rotating1 with an angular
velocity co. The spring constant k
is equal to the force required to
extend the spring a unit distance.
The motion of the weight is re¬
Fig. 24-1. Weight supported by spring. strained by damping, which is
assumed to vary directly with the
velocity of the weight so that the damping force is the damping constant c
multiplied by the velocity of the weight. This is known as “viscous’’
damping.
Let m = mass of the weight
x = displacement of weight measured from its position of rest, posi¬
tive downward
~ = velocity and represented by x
d2x
= acceleration and represented by x
dt2
The forces on the weight are
Spring force = —kx

Damping force = — cx
Impressed force = P0 sin cot
If the weight is displaced downward, the force on the weight equals
1 A convenient way of producing such a force is to rotate two eccentric masses at the
same speed but in opposite directions, as shown in the sketch. The masses may be
mounted on mating gears. The vertical components of the centrifugal forces equal
P0 sin cot. The horizontal components are balanced. This system is used in an
experimental vibrator.
VIBRATION AND VIBRATION CONTROL 363
— kx — cx -\- Po sin cot. This force must equal the accelerating force.
Therefore
mx — —kx — cx + Pq sin cot
or
mx + cx + kx = Pq sin cot (24-1)
The above equation is the differential equation of motion for a system

having a single degree of freedom.
24-3 Free vibration without damping. For this case, Eq. (24-1)
becomes
mx + kx = 0
or
x = - —x (24-2)
m
The general solution of this equation is
x = C i sin t + C2 cos t (24-3)
To evaluate the constants C\ and C2, let
x = Xq when t = 0 (a)
and
x = 0 when t = 0 (b)
By substituting condition (a) in Eq. (24-3),
xq = Ci X 0 + d2 X 1
Differentiating the general solution, Eq. (24-3)
x = xl— Ci cos t +/— — C2 sin t A/—

\m \m \m \m
By substituting condition (b) in the last equation,
0 = .I— Ci X 1 — c2 X 0
\m \m
or
Ci = 0
Therefore Eq. (24-3) becomes
X = Xq cos t (24-4)
364 MACHINE DESIGN
and is the equation of motion for a system having a single degree of free¬
dom vibrating freely without damping. Figure 24-2 is a graphical rep¬
resentation of the motion.
c
<D
£
<D
O
_c
a.
• w—
o
Fig. 24-2. Free undamped vibration.
In Eq. (24-4), one cycle of vibration occurs as t \/k/m goes through

27t radians. If T represents the period of one cycle,
or
The natural frequency of the system is
_ J. _ _1_ jk_ in cycles per sec (24-5o)

~ T ~ 2tt > m
or
a; n in radians per sec (24-56)
24-4 Free vibration with damping. For this case, Eq. (24-1) becomes
mx + cx + kx = 0
When this equation is integrated, its solution for small damping, as

found in machines, may be plotted as shown in Fig. 24-3. The figure
shows how the amplitude of the vibrations is decreased in the presence
of damping. The smaller the damping, the flatter will be the enveloping
curve ABC and the longer will be the time required for the vibrations to
die down. The natural frequency of vibration for this case is
co n (24-5 c)
Values of c so large that (c2/4m2) > k/m eliminates vibration. Such is

the case for a pendulum placed in a very viscous fluid.
24-5 Forced vibration without damping. For this case, Eq. (24-1)
becomes
mx + kx = Pq sin cot
The solution of this equation is
Po/k
x sin cot
1 - (co/<On)2
The quantity Po/k is the deflection of the mass owing to a statically

applied force P0; hence, xs = (Po/k), and the above equation may be
written as
x 1 _
sill cot (24-6)
xs 1 — (C0/C0n)2
The maximum value of x occurs when the value of sin cot is a maximum,
i.e.j unity. Hence, letting xq represent the maximum value of x, Eq.
(24-6) becomes
Xo_1
(24-6 a)
Xs 1 — (co/cOn)2
The graphical representation of Eq. (24-6a) is shown in Fig. 24-4 and

is useful in interpreting characteristics of vibrating systems. In this fig¬
ure, the ordinate xo/xs is the ratio of the maximum deflection of the mass
from its position of rest to the deflection due to a statically applied force
Po. The abscissa co/con is the ratio of the frequency of the applied force
Po to the natural frequency of the system. Since xs and con are constants
of the system, it may be seen that Fig. 24-4 is a representation of the var¬
iation of the maximum deflection xo of a vibrating body with the fre¬
quency of application co of the applied force. The figure shows that when
co is zero, the deflection of the body is equal to the static deflection xs,
and when co is equal to con, the deflection in the absence of damping is
infinite. For operating frequencies greater than y/2(0n, the deflection
366 MACHINE DESIGN
of the body becomes less than xs. In using Eq. (24-6a) in the region
where (oo/oon) > 1, it is necessary to change the sign of the right-hand
side of the equation. The physical interpretation of this change is that
the deflection and the impressed force are out of phase when co/con is
greater than unity.
24-6 Forced vibration with damping. When damping is present in the
system in the form of viscous resistance or of friction, the effect of the
damping force is to decrease the amplitude of motion of the vibrating
mass. In viscous damping, it is assumed that the resistance to motion

of the vibrating mass is proportional to the velocity of the mass. Since
high velocities accompany large displacements, viscous damping will
affect the curve of Fig. 24-4 as shown in Fig. 24-5.
The equation for the maximum deflection with damping is
Po
X0 = - /-
V (coo)2 + (k — moo2)2
In mechanical systems, there is always some inherent damping present;

hence the zero-damping curve is never realized. Curves d and e represent
degrees of damping which are higher than those found in most machinery.
The damping in most machinery falls in the intermediate stages, as shown
by curves b and c. These curves show the reduction in amplitude to finite
values as the critical speed is passed.
24-7 Lateral and torsional vibrations. The vibration of a spring-

supported weight, as discussed in the preceding article, was used as an
illustration of lateral vibration. Other examples are the vibration of the
string of a musical instrument, as shown in Fig. 24-6(a), and a beam or
shaft vibrating laterally, as shown at (6). Another type of vibration in
mechanical systems is torsional vibration, illustrated by the motion of a
torsional pendulum, as shown at (c), in which the mass oscillates through
an angle as indicated. In machinery, a shaft may rotate uniformly, in
Fig. 24-5. Effect of damping on forced vibration.
which case there is no torsional vibration, or the shaft may have super¬
imposed on its uniform motion one of vibratory character which gives a
motion similar to that represented by the curve in Fig. 24-6(d) in which co
represents the angular velocity of the shaft. Torsional vibration is quite
common and also troublesome in machinery having rotating parts. The
frequency of torsional vibration is usually greater than the speed of rota¬
tion of the shaft.
The natural frequency in torsional vibration of the system, as shown
in Fig. 24-6(c), depends on the weight and radius of gyration of the sus¬
pended mass, and on the stiffness of the supporting rod. The natural
frequency may be determined from the following equation:
1 GJg
Natural frequency, cps
2?r ^ LWp2
368 MACHINE DESIGN
where G = torsional modulus of elasticity, psi

J = polar moment of inertia of rod, in.4
g = acceleration due to gravity = 386 in. per sec2
L = length of supporting rod, in.
W = weight suspended, lb
p = radius of gyration of weight, in.
Some common examples of forced vibration in machinery are the lateral
vibration of a mass, such as an engine resting on its foundation, or the
lateral vibration of a shaft subjected to variable loads, or the torsional
//ZZ
(a)-Free,lateral vibration
(b)-Forced lateral
vibration
^-Torsional
vibration
frfJ-Torsional vibration of rotating shaft

Fig. 24-6. Lateral and torsional vibrations.
vibration of a shaft subjected to variable torque. Figure 24-5 was devel¬

oped by considering a spring-supported weight under forced and damped
vibration. However, the quantitative aspects of the vibration phenom¬
ena illustrated by the figure apply to all types of vibration, torsional as
well as lateral. These vibrations are common in machinery and may be
destructive because of the large forces and stresses involved, and because
vibration induces repeated stresses that may lead to fatigue failure. In
addition to the effects of vibration on the machine producing them, the
vibrations may have annoying and destructive effects on surrounding
equipment.1
1 Paul C. Roche, Vibration Isolation in War Machines, Machine Design, July, 1943,
p. 124.
24-8 Critical speeds of shafts. In Eq. (24-5a), the natural frequency

of vibration of a weight was given as
f = —
fn 1 /* CPS
By neglecting the weight of a shaft to which, for instance, a flywheel

is attached, Eq. (24-5) may be used to determine the natural frequency
o ° s OOOOOO oo ooo
o m o
— Cvl
o OOO o o oo ooo
ro tj- toio oo o too ooo
_ — cni m tj-in
Natural frequenoy,cpm
Fig. 24-7. Lowest natural frequency versus static deflection.
of vibration, or critical speed, where m is the mass of the flywheel and k

is the number of pounds required to deflect the mass 1 in.
By substituting in the above equation the static deflection 8 in inches
of the shaft at the point where the flywheel is located, its equivalent,
12mg/k, the undamped natural frequency becomes
/n = 187.8 ^ cpm (24-7)
It is evident that a stiff shaft will have a small 8 and that the natural
frequency will be high, while a flexible shaft will have a large 8 and the
corresponding natural frequency will be low (see Fig. 24-7).
370 MACHINE DESIGN
As shown by Eq. (24-5c), one effect of damping is to lower the natural

frequency of vibration. For the degrees of damping in most practical
systems, the difference between the damped and the undamped natural
frequencies will be small, and thus Eq. (24-7) may be used for a close
approximation to the critical speed of a shaft where the mass can be con¬
sidered concentrated at one point as a heavy flywheel attached to a light
shaft.
For a uniformly loaded shaft in transverse vibration, the mass cannot
be considered concentrated at one point, so that it is necessary to con¬
sider the deflection of each unit mass of the shaft. For a uniformly
loaded shaft in transverse vibration, the critical speed is given by the
following equation:1
fn = 211.4^ cpm (24-8)
where 5 = maximum deflection of the shaft.
Pasin cot sin cot

\P0su
sin cot
\p0s"
V/77/////////77/Z/}
faTUnsprung ft>J-Spnnq fcTSpring mounted,
mounted with damping
Fig. 24-8. Vibration mountings.
The above equation indicates the lowest, or fundamental, frequency of

vibration at which it vibrates, as shown in Fig. 24-6(a). The vibrating
shaft may have nodes at the mid-point, the third points, etc., when it
vibrates at frequencies 4 and 9, etc., respectively, times the fundamental
frequency.
24-9 Force transmitted to supports. Since a rigid mounting transmits
all the forces impressed on it, an “unsprung” machine subjected to vibra¬
tion, as represented by Fig. 24-8(a), transmits all the impressed force
Pq sin cot directly to the foundation. The introduction of springs between
the machine base and foundation alters the transmitted force, and the
introduction of damping, such as a dashpot, further alters the trans¬
mitted force [see Fig. 24-8(6) and (c)].
The ratio of the force transmitted to the foundation and the impressed
force is termed “ transmissibility.”
1 Thomas Bevin, “Theory of Machines,” p. 492, Longmans, Green & Co., Inc., New
York, 1939.
transmitted force
Transmissibility = T =
impressed force
T = (24-9)
In Fig. 24-9, transmissibility is plotted for various degrees of damping

as a function of co/con. In the figure, it may be noted that for operating
Fig. 24-9. Transmissibility and the effect of damping.
speeds exceeding y/2 con, damping is somewhat detrimental. A compar¬

ison of the magnitude of transmissibility for the two ranges cited shows
that as a machine passes through the critical speed the beneficial effects of
damping are pronounced, while for speeds higher than the critical speed,
the detrimental effects of damping are not marked. The foregoing illus¬
trates that a vibration-absorbing mounting should have damping prop¬
erties. For speeds several times the critical speed, the transmitted force
is a small fraction of the impressed force, and the machine may be said
to “run smoothly.”
The effect of damping on transmissibility is usually small for the region
(co/con) > y/2. Equation (24-9) may be reduced to the following form
by making the damping coefficient c zero and changing the sign of
372 MACHINE DESIGN
the right-hand side of the equation for the reason given in Art. 24-5.
Therefore
1
(24-10)
(///»)2 - 1
24-10 Vibration control. In order to control, or to eliminate vibration
the following means are available:
1. Remove the disturbing force.
2. Introduce an equal and opposite disturbance.
3. Reduce the effect of the disturbance transmitted to the foundation
or support by introducing flexibility and damping.
An example of method 1 is balancing rotors by removing material from
the heavy part of a rotor. A simple example of method 2 is when a bal¬
ancing weight is placed on the light side of an unbalanced rotor. There
are other examples of this method which amount to introducing in a
machine or instrument a balancing system which vibrates automatically
in such a manner that the effects of the forced vibration are neutralized.
In machines, method 3 is generally used when vibration must be con¬
trolled. It may be noted from Fig. 24-9 that if the normal operating
speed / coincides with the natural frequency of the machine fn, the condi¬
tion of operation is undesirable. However, by introducing a flexible
mounting, the natural frequency may be altered, so that any point along
the co/con axis of Fig. 24-9 can be fixed as the normal operating condition.
In torsional vibration, the natural frequency may be increased by mak¬
ing the shaft more rigid by increasing its diameter, or the natural fre¬
quency may be decreased by making the shaft less rigid by decreasing its
diameter, by increasing its length, or by introducing a flexible coupling.
Example 24-1. As an illustration of design procedure, consider a single-cylinder
vertical engine that operates at 1,200 rpm and weighs 800 lb. If the maximum unbal¬
anced force in the vertical direction, P0, equals 30 lb and the engine is mounted directly
to the floor, the total force P0 will be transmitted. If it is desired to reduce the trans¬
mitted force to 3 lb (that is, T = 0.10), Eq. (24-10) may be used as
f = l = 1.200 _ rÔO _
V(1 + T)/T V(1 + 0.D/0.1 3.32
Hence, the mounting should be selected so that the 800-lb compressor as a mass
vibrates with a natural frequency of 362 cpm.
By solving Eq. (24-7) for the static deflection 5 and substituting values,
, _ /187.7V _ /187.7V _ . orQ .

5 V fn ) \ 362 ) 0,269 in'
If the engine is mounted on four springs, the spring rate for each spring becomes
800
= 745 lb per in.
4 X 0.269
The maximum amplitude of vibration, from Eq. (24-6a), is
xs 0.269
Xo 0.116 in.
(.f/UY - 1 (l,200/362)2 - 1
The connections for the compressor should allow for this movement.
24-11 Vibration-absorbing mountings. In machines subjected to lat¬

eral vibrations, such as reciprocating engines, pumps, compressors, etc.,
it is usually impractical to control vibration by methods 1 and 2 of Art.
24-10. Method 3 is commonly used by employing vibration-absorbing
mountings. As previously discussed, the mountings should have flex¬
ibility and damping. In addition, the mounting should be so constructed
that the machine would not be freed if the resilient part of the mounting
Fig. 24-10. Vibration mounting. (Courtesy of Vibrashock Division of Robinson Avia¬

tion, Inc.)
should fail. In some mountings, flexibility is secured by mechanical

springs, and damping is secured by dashpots or by rubber or cork buffers
(see Fig. 24-10).
Other types of mountings employ rubber, cork, or felt for the resilient
part. Rubber has a high ratio of deflection to load, especially when it is
used in shear (see Fig. 24-11), and has in addition good damping charac¬
teristics. The tube form of mounting shown in Fig. 24-12 has the rubber
bonded to the steel tube in the center, which is bolted to the supported
member, and to the outer sleeve, which is attached to the supporting
member. The plate form in Fig. 24-13 is used particularly for light and
medium loads.
The three-angle type of rubber mounting, as shown in Fig. 24-14, is
suitable for medium and heavy loads.
Many other types of vibration mountings are available in a large vari¬
ety of forms and sizes and with load capacities that vary from a few
ounces for mounting delicate apparatus and instruments, to 32,000 lb or
more for power-plant installations.
374 MACHINE DESIGN
Fig. 24-11. Load-deflection characteristics of cylindrical form of rubber mounting.
Fig. 24-12. Vibration mounting, tube Fig. 24-13. Vibration mounting, plate
form. form.
Fig. 24-14. Three-angle type of mounting using rubber in shear.

24-12 Torsional vibration dampers. It has already been noted that

the critical speed in torsion of a shaft can be lowered by making the shaft
more flexible in torsion by reducing its diameter, by increasing its length,
or by using a flexible coupling, such as a rubber coupling or a Bibby cou¬
pling. By lowering the critical speed, the normal operating point as
shown in Fig. 24-9 is moved to the right into the region of low transmis-
sibility, thus reducing the harmful effects of torsional vibration. To
supply damping in addition to that of the material of the rotating parts
and their supports, a viscous damper or a Coulomb-friction damper may
be used.
There are many other types of torsional-vibration dampers, such as the
energy-dissipation type, rotating-pendulum type, and fluid type.1
1 W. K. Wilson, “ Practical Solution of Torsional Vibration Problems,” vols. I and
II, John Wiley & Sons, Inc., New York, 1942.
CHAPTER 25
MOTOR SELECTION
25-1 Motor rating. The horsepower rating of an electric motor is

designated on the basis of its output at rated speed and voltage, with the
manufacturer’s guarantee that the maximum temperature rise in the
motor will not exceed a specified value. The temperature rise for general-
purpose motors is 40 C for dripproof motors, 50 C for splashproof motors,
and 55 C for totally enclosed motors. It is assumed that the ambient
temperature does not exceed 20 C. These are continuous ratings.
Intermittent ratings are special ratings for time periods, as for 1 hr,
hr, 15 min, etc. This kind of rating is common for motors having
intermittent loads such as crane motors, elevators, and some types of
machine tools, so that they may be operated for a short time at relatively
high loads followed by a cooling period. The rating and corresponding
temperature rise is stamped on the name plate.
25-2 Motor types and applications. The selection of the proper type,
horsepower, voltage, and speed, as well as the proper controls, is a most
important matter and the choice may spell the difference between a long
life of satisfactory operation or on the other hand a continuous round of
trouble.
Many machine-design problems involve the selection of a motor drive,
and the relation between the motor characteristics and the performance
of the machine requires careful consideration.1
This chapter does not pretend to do more than to bring to the attention
of the designer some important factors involved in motor selection, in
particular for the squirrel-cage induction motor.
In general, direct-current (d-c) motors are used where special operating
characteristics such as good speed control are required, or where direct
current is the only kind available as aboard most ships.2
1 The retail value of motor-operated machines produced annually in the United

States is of the order of 3 billion dollars for domestic use and of 200 million dollars for
industrial applications. See T. J. Woodson, Motors for Integral Mechanisms, Mech.
Eng., August, 1950.
2 A comprehensive table for selection of type of motor is given by F. H. Pumphrey,
“Fundamentals of Electrical Engineering/’ Prentice-Hall, Inc., New York, 1951.
376
MOTOR SELECTION 377
By far the bulk of industrial motors are of the alternating-current (a-c)

type. Single-phase current is usually satisfactory up to Y hp for across-
the-line starting and up to 1 hp with resistors or autotransformer starting.
25-3 Induction motors. The most common type of motor for indus¬
trial use is the squirrel-cage induction motor. This motor is simple,
rugged, and foolproof, and the current is induced in the rotor bars by
Per cent synchronous speed

Fig. 25-1. Speed-torque characteristics for induction motors.
induction, so that slip rings or commutator are not required. In Table

25-1 is given information to assist in selecting National Electrical Man¬
ufacturers Association (NEMA) class A, B, C, and D squirrel-cage induc¬
tion motors for various applications. Figure 25-1 shows curves for
approximate speed-torque characteristics for these motors. Class D
motor, known as a high-torque high-slip induction motor is suited for
use with a flywheel drive, for example, in a punch press. Since a flywheel
cannot transfer kinetic energy to the load without slowing down in the
process, it is necessary to use a motor that is capable of speed variation.
The motor must be able to slow down sufficiently for the flywheel to
transfer enough energy to the system to carry the load through the peak
in the duty cycle. The motor must also be capable of accelerating the
378 MACHINE DESIGN
system back to its original speed, thereby restoring the kinetic energy to
the flywheel before the next peak in the duty cycle occurs. The high-
torque high-slip induction motor is well suited to this purpose.
Table 25-1. Characteristics and Applications of Squirrel-cage

Induction Motors*
Start¬
Start¬ Break¬
NEMA ing On such
ing down Slip Application
class cur¬ machines as:
torque torque
rent
A Normal High Nor¬ Low For moderately easy-to-start loads; for Generators
mal loads requiring slightly more than full¬ Pumps
load starting torque and low slip; rela¬ Lathes
tively high breakdown torque to sus¬ Drill presses
tain occasional overloads, where higher Grinders
starting current can be tolerated Machine tools
Conveyors
Compressors
B Normal Normal Low Low For moderately easy-to-start loads; for Generators
loads requiring slightly more than full¬ Pumps
load starting torque and low slip; rela¬ Lathes
tively high breakdown torque to sus¬ Drill presses
tain occasional overloads Grinders
Machine tools
Conveyors
Compressors
C High Normal Low Low For hard-to-start loads; for loads requir¬ Electric stair¬
ing high starting torque but not requir¬ ways
ing high overload demands after run¬ Pulverizers
ning speed is reached. Not suitable Compressors
for loads requiring long accelerating Conveyors
time
D High Low High For hard-to-start intermittent loads and Hoists

shock loads; no sharply defined break¬ Elevators
down torque point. Surge load causes Punch presses
appreciable decrease in speed with Machines with
high torque developed to recover speed large flywheels
rapidly. Available in intermittent and Centrifuges
continuous ratings
* From Westinghouse Electric Corporation.
The wound-rotor induction motor is used to drive variable-speed equip¬

ment such as hoists, cranes, centrifugal pumps, and fans. The three-
phase windings on the rotor are brought out to three slip rings which are
connected to variable external resistors.
25-4 Controls and mountings. Small induction motors may be started
across the line, but larger motors require starters employing resistors or
autotransformers. Motors may be bolted directly to the bedplate, or
adjustable bases may be provided for chain, belt, or gear drives. Many
MOTOR SELECTION 379
motors are made with special bases or flanges when they are to form an
integral part of a machine or appliance.
25-5 Standard sizes and dimensions. Motors are built into standard
size frames so that one frame size may be used for more than one horse¬
power rated at different speeds. For instance, Frame 213 is used for the
following motors:
5 hp at 3,600 rpm
3 hp at 1,800 rpm
2 hp at 1,200 rpm
1^2 hp at 900 prm
The price of each of these motors will not vary more than 10 per cent
from the average of the four.
In Table 25-2 are given standard horsepowers, speeds, frame numbers,
outside diameters of motors, and diameters of shaft stub ends. The
frame numbers are according to NEMA Standards approved in 1954.
Table 25-2. Standard Horsepowers, Speeds, Frame Numbers, and Dimensions

for Dripproof, Polyphase Squirrel-cage Induction Motors
Frame numbers Dimensions
Rpm Frame
No
A* u\
±ip
3,600 1,800 1,200 900
182 9 /8
182 184 9 Vs
182 184 213 ioh iH
i 182 184 213 215 io H iH
IH 182 184 184 213 254U 12 M i%
2 184 184 213 215 256U 12 M IVs
3 184 213 215 254U 284U 14 iy
5 213 215 254U 256U 286U 14 iy
7H 215 254U 256U 284U 324U 16 i%
iO\
\oo
r-H
10 254U 256U 284U 286U 324S 16

15 256U 284U 324U 326U 326U 16 1%
20 284U 286U 326U 326S 16 iy
25 286U 324U
* A= outside diameter of motor,
30 324S 326U
40 326S f U — shaft stub-end diameter,
- in.
25-6 Selection of motor size. Considerable care should be used in

specifying the size of motor. If it is too small, it will overheat, which
will shorten its life; if it is larger than required, its efficiency will be low
and, for an induction motor, the power factor will be low thus unduly
increasing the cost of operation.
380 MACHINE DESIGN
Open 40C integral-horsepower motors are good for continuous service

greater than the rated horsepower by the following factors: 1 hp, 1.25;
and 2 hp, 1.20; 3. to 200 hp, 1.15.1
Power factor. On account of the reduction in power factor at part
load, which is undesirable in that it increases power loss in the lines and
transformers and reduces the kilowatt output of the power-supply gen¬
erator, it is better to select motors to avoid, if possible, sustained part¬
load operation. This is particularly desirable in a plant in which a large
number of small motors are used. As an example, a 1-hp induction
motor, 1,750 rpm, has power factors as follows: full load, 73 per cent;
% load, 65 per cent, and half load, 50 per cent.
Efficiency. The efficiency of motors increases with the size of motor,
speed, and load up to full load. Thus large, high-speed motors operating
at full load are most efficient.
Intermittent rating. Since the horsepower rating of a motor is based
on temperature rise, advantage may be taken of cooling periods if the
duty cycle requires partial load and rest periods in regular cycles. Since
most of the heating of the motor is caused by copper losses due to the load
current, which (losses) depend on the square of the load current, it is
customary to calculate an average horsepower on the basis of the root-
mean-square (rms) value. The motor may be selected on the rms horse¬
power rather than on the maximum-load horsepower.
The rms horsepower for repetitive duty may be determined by mul¬
tiplying the square of the horsepower for each part of the cycle by its
duration in seconds. Divide the sum of these results by the total time
in seconds for a cycle. The square root of this last result is the rms value.
If the motor is stopped during the cycle, only one-third (for open motors)
of the standstill time should be added in the time for one cycle. This is
due to the reduction in cooling when the motor is at rest.
In using the rms horsepower to choose a size of motor, the duty cycle
should be short as compared with the time required for the motor to
reach a steady temperature. Also if there are extremely high peaks in
the duty cycle, such as in drop-forge and punch-press applications, the
motor may stall at the peak load. In this case a flywheel should be used.
Example. Assume a machining operation in which a motor operates at 8-hp load

for 4 min, 6-hp for 50 sec, 10-hp for 3 min, and standstill for 6 min. Select a motor
size for this repeated cycle of operation (see Fig. 25-2).
, /(82 X 240) + (62 X 50) + (102 X 180)

rms np y 240 + 50 + 180 + 3 6%
= V54.6 = 7.7 Select a 7^-hp motor
1 See NEMA Specification, pt. 4, p. 9.

MOTOR SELECTION 381
25-7 Connected rotating masses. For fast, repeating cycles which

require rapid acceleration of rotating masses, it may be necessary to cal¬
culate the time required for acceleration. In this case it is usually
convenient to make calculations on the basis of torque rather than
horsepower.
From the equation for the rate of change of angular momentum, the
equation for the time required to accelerate uniformly a rotating body
subjected to a constant torque may be found:
T = la
where T is the torque, I is the polar moment of inertia of mass, and a is

the angular acceleration. By substituting for / its equal (W/g)R2,
10
Repeat
8
6
RMS hp
2
0 1 1 1 1 iiii i t
0 100 200 300 400 500 600 700 800 900 1000
Time, sec
Fig. 25-2. Motor duty cycle and corresponding rms horsepower.
where W is the weight of the mass and R its radius of gyration, the follow¬
ing equation may be derived for the time required to accelerate uniformly
the rotating mass for constant applied torque:
WR2 X (change in rpm)

(25-1)
308 T
where t = time for acceleration, sec

W = weight of rotating mass, lb
R = radius of gyration, ft
T ~ constant torque, lb-ft
Equations for determining WR2 for various bodies are given in Appen¬
dix XI. Values for WR2 for standard motor rotors may be found in
manufacturers’ catalogues.
In case the accelerating torque is not uniform, a more extensive anal¬
ysis will be necessary.
■ x
.
Appendix I: Abbreviations*
Abbreviations should be used sparingly in text material and only when preceded by
numerals. Use lower case for abbreviations except where the word indicated requires
a capital. Abbreviations representing compound adjectives should be hyphenated.
Periods are not used in abbreviations with few exceptions, as in. for inches.
Absolute. abs Feet per second. fps

Alternating-current. a-c Figure. Fig.
(as compound adjective) Foot. ft
American Society of Mechan¬ Foot-pound. ft-lb
ical Engineers. ASME Gallon. gal
American Standards Associ¬ Gallons per minute. gpm
ation . ASA Gallons per second. gps
Ampere. amp Gram. g
Atmosphere. atm High pressure. h-p
Average. avg Horsepower. hp
Barometer. bar Horsepowerhour.’.. hphr
Boiler pressure. bp Hour. hr
Brake horsepower. bhp Inch or inches. in.
Brinell hardness number. Bhn Inches per second. ips
British thermal unit. Btu Indicated horsepower. ihp
Calorie. cal Intermediate-pressure. i-p
Centimeter. cm Internal. int
Coefficient. coef Kilogram.v. . . . kg
Constant. const Kilovolt. kv
Cubic feet per minute. cfm Kilovolt-ampere. kva
Cubic inches. cu in. Kilowatt. kw
Cycles per second. cps Kilowatthour. kwhr
Cylinder. cyl Liquid. liq
Degrees. deg Logarithm (common). log
Degree centigrade. C Logarithm (natural). In
Degree Fahrenheit. F Maximum. max
Diameter. diam Mean effective pressure. mep
Direct-current. d-c Meters. m
(as compound adjective) Miles per hour. mph
Dozen. doz Millimeter. mm
Efficiency. eff Millivolt. mv
Electromotive force. emf Minimum. min
Equation. Eq. Minute. min
External. ext Ounce. oz
Feet. ft Per cent. spell out
Feet per minute. fpm Pint. pt
* Based largely on ASA list.

383
384 MACHIN DESIGN
Pound or pounds. lb Square centimeter. sq cm
Pounds per square inch. psi Square foot. sq ft
Quart. qt Square inch. sq in.
Radians.•.jspell out Standard. std
Revolutions per minute. rpm
Temperature. temp
Second. sec
Volt. spell out
Shaft horsepower. shp
Watt. spell out
Society of Automotive Engi¬
Watthour. whr
neers . SAE
Specific gravity. sp gr Weight. wt
Specific heat. sp ht Yard. yd
Square. sq Year. yr
Appendix II: Elements of Sections
Rectangular Polar
Section
Moment of / Moment of J
inertia I c inertia J r
bh3 bh2
12 6
7tD4 irD3 7rD4 irD3

64 32 32 16
7r(D4 - d4) 7r(D4 — d4) 7r(D4 — d4) 7r(Z)4 — d4)

64 32 D 32 16D
Appendix III: Bending Moment and Deflection Formulas
M = maximum bending moment, lb-in.
d = maximum deflection, in.
P = concentrated load, lb
W = uniformly distributed load, lb
L = length of beam, in.
/ = moment of inertia of section, in.4
Beam loading and support M 8
A
PL3
PL
3 El
TV
WL WL* 3
2 8 El
L
\P
PL PL3
4 48£7
L/2 - \
TV
i i i rri i 1 i r
WL WL3
ZX -2x 8 76.8 El
/
// PL PL3
f; 8 192 El
'A —1 i i i i W
f/
i i i i 1 WL WL3
'A v. 12 384P7
k-L-^
385
^ I OHWfO^^CONOOOJOHWCOîO ^ lOHweo^^cosooooHiNcoîc
r-< r rH rH rH rH -1 ^ 1 »H tH *-H *-H t-H »—1
00 05
to
H 00 N to O it 05 ©
CNNfflMO® oo i—• oo 05 oq © oo
rt ^ Ol ffl ® H «5 ONifqoO©iN
d i-i it i-i co h id o d cd >d i-l © cd
HINTfC *• i-l <M CO
© 00
rH Tt CO © n co cq
moitocoiQNOa ®if«iHOOHO©coeo
HN©Tt(NOOM»«rHlOOOHlO«!D o©MN«®q®®©coif©cqHoo
© *-i it i-i cq" oi co id td os <n oo © it © © ddcdidi-iosi-ltdooitedit’oicdcdcd
HMC' ®OlM 5 00 ©NH(N'>fM HHMit©ai(M©o®iMO>
HHiNnîocN h h cq cq eo eo
co
(x
lO O) 00 cq © ^
to
O © © 00 00 © 00 to
© 00 00 If if 00 if
ri
H H co 00 00 © o o © h it © cq o
o i-l Tji 6 h od ci © © cd id © 05 i-l
H H CO
H (N M ©
00 it
Oq Cq eo i—< eo cq eo
oo©oo©itcooo© 2cqifNcq©®ON©
OOOCOOMtOM©OffliOO^HH ©iq©i-Hio©i-ct^t^©©coiOccqi-<©
di-tifd'-iNOHHcicdooNdON ©dcdiddoodididdcddedididcd
HNCO©fflCO«^HOHCO© HHMlt ©Cl«©O©H00
i-i h cq cq co co
00 ©
to
CO 1^ H ©
to h cq eo co
© Iji H rH N (M N co b- o cq rH ep eo
cq © oq oq oq i-h *>. © it 05 05 rH © co
r*
© © ed © d ed oo o © i-l ui © oo os
in eo © h h cq
05 ©
N M H M co i—i © eo cq
Ttif©©Cqit©if cic)NNHcq©cq®if
005©C ^ 0 01000 H©©H©N 3 ©itti;eqtîtioio©iooocq©co’—i»o
docdddittdtdedddooidedif© ©di-iito5tdo6cdcdtdi>Tttdo6tilit
(d = diameter of shaft)
HCO©OOONC ®H^ 00 05 i-i<Mit©OOi-iiO©it©Ix

HHNCOMi#©N h h h cq co co
05 A03 lO
it 05
it b-
QQ
to
b- © Cl © CO
to CO it O tn © © © h s © co cq ® ©
o n co oo © n ^ ©cqeqitcooqr-
© © ed oo oo
H
cd ©
CO ©
0> © © i-l it 05 ed t>
i-i oq
© 0) co
cq i-i it oo eo
if eo oo ©
N©©HO!©(MM Hcocooifeoocqos
O©Oif®® l ®«C 0 00 0©©H©CI oeq©cq©eo©itoqrHeocooo©eo©
ddcdodtdciedoqdodtdodcdedodi-i o o h d oo ed ed i-i d d cd d h h o) ©
i-<eo©oooq©oq©oooo©eo Hcqif ®»Hif ®it®®
HH(MIMC if©N 0 h h h cq cq co
it oq
© co co © oo ®q cq oo ©
to H 00
If N © © 00 to © © cq oo oo co ©
o © q q h © m © cq it a © n h
© © oq td td rd cq’ © o' i-i cd oo id ed
H CO © i-i cq
(N
Values of 7rd3/16
It CO
cq
O O O 00 lO © © it ©
0©HNcqif if ©no
HHCOrfifON© eo
©©eqiqîtoqtêqoqeoOrHoqeqe© ©cqcqb^cqcqiteo©©©©o©©oo
docdtdcdddodudi-iososcdoicdcd
00 00
5 © O »-l co oo id id © td © © it ed it i-i ed
HCO©NH©H N® H Hcqco©oooôoco©©
HHiMcqcoif©N rH i-< i—i cq cq co
eo it co eo Tj co cq it oo to
05 ocqH©Ncqo
to
O It
o It
<N
It
CO ©
1-1t-
It
It CO £ to
o cq cq © oo n n
o o ci n © oi oi *3 © © i-l ed td it it
1-1 (N It
> it
rH cq
rH ©
CO it © o h cq oo io
ococo-ttî-icocq O©HNC0ONH«O
ocodNOifqqMififoqqeqif OHHM©cq®ifHNN00©ifHO
6
ddcdeddo tdTtddi-lo i-iedo ed 5 6 ©©rHcdtditcdfdidtdidddoditoo
HNifNH©HN©©eD® iHcqco©NOcooocqooif
HHIMlMCOif ©O HHHcqcqco
©
CO ©
HO©® © It 00
to o N © © N Hi-(
to
o © cq oo h h ©
o co o co if if © o h o h cq n cq
© © ci ed it td ed d d i-l cd td cd cd
i-i (N it rH cq
00 ©
rH
eo
O O it CM © cq © cq
©COOO©OOCOOqoq Oitit©©©q©rHeo©>
ocqoo®Nifi-cqMcqoo«o®coit OrH0505oqcqi©iqeo©c35Cq©©rHt^
ddr-iideded©i-l©©’odddeded© © © o ci ed cd oq © ci it rd © © cq’ ed ©
HN^NOifON©if©N Hcqeo©NOcoNcqNcc
HHNeqcOit©© h h h cq cq co
00 it
cq
© ©
to O eo CO 05 to © oo cq
o co cq co © n © o rH © cq oo it oo
©®qt>qi-;itt> O rH 00 00 ©1 N OO
d © rd ©j cd © cd © o © ci ed cq r-i
i-i oq it rH cq
CO rH rH 00 © rH co
O N O N ^ H »c ©00©00t^rH00t~t^00
i-iiqeqiqiqitcqiqi-iiteocoitoot^ Oti.eqoqcqoqeqcqiOrHt'-t-.r-Htcq
o o
o o i-i d cd it cd td o cd ed i-l d i-i oo cd ©dddedcii-icddi-ioodaiiddi-l
Hfqif®Oif®©cocoeo© rH(MCO©tii©CO©rH©CO
HHHNCOlt©© h h cq cq co
o^cêo^ôcotôocio^c^co^kc T3 OHNCC^»OCDN000)OHNW^»O
- rH rH rH rH rH rH rH rH rH rH rH rH
386
Appendix V: Properties of Materials
Tension Elonga¬
Endurance Modulus of
tion in
Material No. Description limit elasticity
Ultimate Elastic 2 in.,
(bending)* in tension*
strength* limit* per cent
Cast iron. 20 ASTM /4-in. section 20,000 6,200 6,500 10,000,000

20 ASTM < Kin- 24,000 7,500 7,500 12,000,000
25 ASTM /4-in. section 25,000 10,000 9,500 13,000,000
25 ASTM < /4in. 30,000 12,000 11,000 14,000,000
50 ASTM /4-in. section 50,000 30,000 18,000 20,000,000
Malleable cast iron. 50,000 20,000 24,000 25,000,000
Nickel cast iron . . Grade II 35,000 17,000 13,000 16,000,000
Cast steel. 0.20% C Annealed 60,000 30,000 24,000 29,000,000 22

0.30% C Annealed 70,000 35,000 28,000 29,000,000 18
0.40% C Annealed 80,000 40,000 32,000 29,000,000 15
Wrought steel .... SAE 1010 Hot rolled 54,000 31,000 30,000,000 36
SAE 1020 Hot rolled 62,000 35,000 30,000,000 30
>—1
SAE 1030 Annealed 75,000 42,000 30,000,000 26
SAE 1040 Annealed 90,000 50,000 > 30,000,000 22
SAE 1050 Annealed 95,000 52,000 "d 30,000,000 20
Annealed 120,000 60,000 d
SAE 1095 0) 30,000,000 20
ft
SAE 1095 Drawn at 900 F 150,000 80,000 ft 30,000,000 16
Nickel steel. SAE 2320 Annealed 70,000 45,000 O 30,000,000 29

o
SAE 2320 Drawn at 1000 F 120,000 80,000 m 30,000,000 25
SAE 2340 Annealed 95,000 55,000 30,000,000 26
SAE 2340 Drawn at 1000 F 120,000 95,000 30,000,000 22
Aluminum bronze. SAE 68 65,000 14,000 14,000 17,000,000 20
Brass. SAE 41 Cast 25,000 10,000 7,000 4,500,000 20

SAE 41 Rolled 45,000 25,000 12,000 6,000,000
SAE 41 Drawn 75,000 31,000 20,000 6,000,000
Bronze. SAE 64 25,000 12,000 8,000 14,000,000 8
Copper. Soft 32,000 2,800 12,500 15,500,000

Drawn 50,000 12,000 17,000 17,000,000
Monel metal .... Cast 72,000 25,000 20,000 25,000,000

Hot rolled 85,000 30,000 30,000 25,000,000
Drawn 120,000 70,000 50,000 25,000,000
Cast aluminum . . . SAE 30 22,000 9,000 7,500 9,700,000 2

SAE 33 22,000 9,000 8,500 9,700,000 2
SAE 35 19,000 7,500 6,500 9,700,000 4
SAE 38 31,000 8,000 7,500 9,700,000 8
Wrought aluminum SAE 26 Soft 26,000 7,000 11,000 10,300,000 20

SAE 26 Age hardened 58,000 23,000 15,000 10,300,000 20
SAE 24 Age hardened 65,000 28,000 14,000 10,300,000 20
Aluminum mag-
nesium * . AM 241 26,000 11,000 7,000 6,500,000 5
AM 53S 37,000 22,000 12,000 6,500,000 12
AM 57S 43,000 30,000 17,000 6,500,000 17
* Values in psi.
387
Appendix VI: Tensile Strength and Yield Point Versus Brinell Hardness*
100 150 200 250 300 350 400 450 500 550
Brinell hardness
* From Noll and Lipson, Allowable Working Stresses, Proc. Soc. Exp. Stress Anal., vol. Ill, no. II
1946.
Appendix VII: Endurance Limit Versus Tensile Strength*
* From Noll and Lipson, Allowable Working Stresses, Proc. Soc. Exp. Stress Anal., vol. Ill, no. II
1946.
388
Appendix VIII: Yield Points for Steel
point in 1,000 psi
Yield
Fig. VIII-1. Yield points for some oil-quenched steels.
389
390 MACHINE DESIGN
Tempering temperature, deg F

Fig. VIII-2. Yield points for some water-quenched steels.
Appendix IX: Hardness Numbers Conversion Chart*
Shore hardness
Tensile strength, thousand lb per sq in
* Courtesy of International Nickel Company, Inc.
391
Appendix X: Curves for Stress-concentration Factors*
Stress-Concentration Factor
* Curves taken by permission of the author and publisher of “Stress-concentration Design Factors,”
by R. E. Peterson, John Wiley & Sons, Inc., New York, 1953.
392
Stre*s-Conc«ntrotion Foctor,
APPENDIX X
393
Str&is — Co.-canlrction Factor, Kt
394
MACHINE DESIGN
O 30
Stress—Concentration Factor,
APPENDIX X
395
0.30
Stress — Concentration Factor,
396
MACHINE DESIGN
APPENDIX X 397
Stress-Concentrotion Factor,
O 0 05 0 10 0.15 0.20 0 25 0.30

r/d
Fig. X-6. Stress-concentration factor Kt for a grooved shaft in tension.

398 MACHINE DESIGN
Combined Stress-Concentrotion and Shear-Energy Factor,
Fig. X-7. Combined factor Kt' for a grooved shaft in tension.

APPENDIX X 399
Stress-Concentrotion Factor,
O 0.05 010 0.15 020 0 25 0.30

r/d
Fig. X-8. Stress-concentration factor Kt for a grooved shaft in bending.
400 MACHINE DESIGN
Combined Stress-Concentration and Shear-Energy Factor, K(
I *d*
c 32
Fig. X-9. Combined factor Kt> for a grooved shaft in bending.
Stress-Goncentrotion Foctor, KfS
APPENDIX X
401
Stress-Concentration Factor,
402
MACHINE DESIGN
APPENDIX X 403
Combined 'Stress-Concentrotion ond Shear-Energy Factor, Kf
Fig. X-12. Combined factor Kt> for a stepped shaft in tension.

4CM- MACHINE DESIGN
Stress— Concenfrofion Foctor,
r/d
Fig. X-13. Stress-concentration factor Kt for a stepped shaft in bending.
APPENDIX X 405
Combined Stress-Concentration ond Shear-Energy Factor, Kf/
r/d
I = Trd3
c 32
Fig. X-14. Combined factor Kt> for a stepped shaft in bending.
Stress—Concentration Factor, Kf
406
MACHINE DESIGN
APPENDIX XI 407
Stress-Concentration Factor,
Fig. X-16. Stress-concentration factor Kt and Kts for a shaft with transverse hole.
Appendix XI: WR2 for Rotating Bodies
W = weight of body, lb
R = radius of gyration, ft
D = outer diameter of cylinder, ft
d = inside diameter of hollow cylinder, f+
Steel weighs 0.283 lb per in.3
Cast iron weighs 0.260 lb per in.3
For rotating solid cylinder,

WD2
WR2
8
For rotating hollow cylinder,
W (D2 + d2)
WR2 =
8
Equivalent WR2 referred to shaft at different speed is
<™*>* - WR2)' [Sil

Equivalent WR2 for mass in translation at velocity (fpm) is
(fpm)
WR2 = W
27r(rpm)
Appendix XII: Method for Determining Lewis Factors
On a drawing of the tooth, extend the normal to the tooth profile at the corner A
until it intersects the center line at B. Draw BC perpendicular to the center line of the
tooth. By means of a scale, determine by trial a line tangent to the profile at D and
such that DC equals CE. It can be shown then that point D locates the weak section
of the tooth and the dimensions for determining the Lewis factor may be scaled.*
* M. A. Durland, Machinery, vol. 29, no. 12, p. 958, August, 1923.
408
Appendix XIII: Graphical Integration
Bending moments. As an example of the graphical determination of bending

moments, assume that the shaft or beam in Fig. XIII-1 is supported on two bearings
at its ends and loaded as shown. By placing letters A, B, C, and D, the force between
A and B may be labeled ab, that between B and C may be labeled be, the right-hand
reaction may be labeled cd, and the left-hand may be labeled reaction da. In Fig.
XIII-2, the forces ab and be were laid off to a scale, and a pole o placed to the right of
409
410 MACHINE DESIGN
ac at a convenient location. The “rays” oa, ob, and oc were then drawn. The
polygon thus formed is a force polygon. In this polygon, the vectors oa, ab, and bo
represent forces in equilibrium since they form the sides of a closed polygon. By
laying off in Fig. XIII-1 the lines a and b parallel, respectively, to the rays oa and ob
intersecting the line of action AB, these three lines will then represent concurrent
forces. The diagram being formed is known as a “funicular polygon” and the lines
a, b, etc., are known as “strings.”
Now if the string b is extended until it intersects the line of action of BC, and
through that intersection the string c is drawn, the directions of the three concurrent
forces, BC, co, and ob will be introduced into the funicular polygon.
So far nothing new has been determined, but if the point where the string a inter¬
sects the line of action of DA be connected with the point where c intersects the line
of action of CD, then the string d thus drawn will represent the direction of the force
which will be in equilibrium with DA and CD. In Fig. XIII-2, the ray od is drawn
parallel to the string d and by so doing, the magnitude of the reactions cd and da are
determined graphically.
In the above procedure, the funicular polygon was drawn solely to determine the
direction of the string d, but as will be shown below, the funicular polygon is the bend¬
ing-moment diagram.
In Fig. XIII-3 the funicular polygon was redrawn. The bending moment at a
section as nm is equal to the sum of the moments of the forces to one side or the other
of the section. To the left, the moment
M = (da)g — (ab)f
The triangle Ijn is similar to oda in the force polygon, or
In _ da
or (■da)g = (ln)h
g ~ h
where h is called the “pole distance” and is measured from the line of action of the
forces to the pole o in the force polygon.
Now, extend the ray a until it intersects nm at l, and from the similar triangles Ikm
and oab,
Y = ^ or (ab)f = (ln)h
Therefore,
M = (da)g — (ab)f = (lri)h — (lm)h
= h(ln — Im) = h(mn)
Thus, the vertical distance mn is a measure of the bending moment at the section
of the shaft.
The scale of the bending-moment diagram is sph, where
s = space scale; 1 in. on drawing = s in. on shaft

p = force scale; 1 in. = p lb
h = pole distance, in.
Graphical integration. Assume that the integral of the curve AX in Fig. XIII-4 is
to be determined graphically.
Divide the curve into a convenient number of strips OB, BC, etc. Label the points
corresponding to the mean ordinates of the strips pi, p2, etc., and project these points
horizontally to give the points 1 , 2, etc., on a line parallel to the Y axis. Choose a
APPENDIX XIII 411
pole P on OX extended and draw the rays Pi, P2, etc. These rays are the chords of
the first-integral curve and may be drawn through the points Oi, bi, etc., as shown
in Figs. XIII-5.
In Fig. XIII-4, let y be the mean ordinate of the strip OB and bx its width so that
y bx is closely its area. By construction,
bbl = -
— y or y bx = h(bbi)
bx h
Hence, the ordinate bbi is a measure of the area of the original curve to the point B.
Similarly, it may be shown that the intercept CiC2 is a measure of the strip BC, and
when cic2 is added to bb i, the ordinate cci is a measure of the area of the original curve
to the point C.
^ 4
1
\<—a-A
Fig. XIII-6. Deflection of shaft.
Deflection of shafts. As shown in Fig. XIII-6, the bending-moment diagram may

be determined graphically; then a modified bending-moment diagram may be drawn
by dividing the actual bending moment at each section of the beam by the product
El for that section. The modified bending-moment diagram is M/EI which equals
d2y/dx2, where y is the deflection of the shaft. By integrating this curve twice, the
deflection curve y/x is determined.
Appendix XIV: Table of Wire Sizes
Washburn Washburn
Fraction Decimal Fraction Decimal
& Moen & Moen
of of of of
wire gauge wire gauge
inch inch inch inch
No. No.
XA* 0.5000 Yz2 0.09375

7-0 0.490 13 0.092
% 0.46875 14 0.080
6-0 0.462 15 0.072
Ke 0.4375 16 H6 0.0625
5-0 0.431 17 0.054
l%2 0.40625 18 0.047
4-0 0.394 19 0.041
% 0.3750 20 0.035
3-0 0.3629 21 0.032
% 0.34375 M2 0.03125
2-0 0.331 22 0.0286
Me 0.3125 23 0.0258
0 0.307 24 0.0230
1 0.283 25 0.0204
%2 0.28125 26 0.0181
2 0.263 27 0.0173
H 0.250 28 0.0162
3 0.244 0.0156
4 0.225 29 0.0150
V$2 0.21875 30 0.0140

5 0.207 31 0.0132
6 0.192 32 0.0128
Me 0.1875 33 0.0118
7 0.177 34 0.0104
8 0.162 35 0.0095
M2 0.15625 36 0.009
.' \
9 0.148 37 0.0085
10 0.135 38 0.008
H 0.125 39 0.0075
11 0.120 40 0.007
12 0.105 41 0.0066
42 0.0062
♦Between H and 1 in., wire sizes vary by 1^2-in. intervals.
412
Appendix XV: Decimal Equivalents of Fractions
(Advancing by sixty-fourths)
Ye± = 0.015625 3%4 = 0.515625

H2 = 0.03125 XVZ2 = 0.53125
= 0.046875 3^4 = 0.546875
He = 0.0625 /K 6 = 0.5625
^4 = 0.078125 3%4 = 0.578125

^2 = 0.09375 1%2 = 0.59375
= 0.109375 3%4 = 0.609375
H = 0.125 5A = 0.625
%4 = 0.140625 4^4 = 0.640625

%2 = 0.15625 234J2 = 0.65625
llA4 = 0.171875 4%4 = 0.671875
%6 = 0.1875 % = 0.6875
l^4 = 0.203125 4%4 = 0.703125

%2 = 0.21875 % = 0.71875
= 0.234375 4^4 = 0.734375
M = 0.25 3A = 0.75
X%4 = 0.265625 4%4 = 0.765625

%2 = 0.28125 % = 0.78125
L%4 = 0.296875 5^4 = 0.796875
Ke = 0.3125 LHe = 0.8125
21Aa = 0.328125 5%4 = 0.828125

% = 0.34375 % = 0.84375
2%4 = 0.359375 5^4 = 0.859375
% = 0.375 7A = 0.875
2^4 = 0.390625 5%4 = 0.890625

1%2 = 0.40625 2%2 = 0.90625
2%4 = 0.421875 5^4 = 0.921875
Vie = 0.4375 15Ae = 0.9375
2%4 = 0.453125 6H4 = 0.953125

15A2 = 0.46875 % = 0.96875
3K 4 = 0.484375 6V64 = 0.984375
y2 = 0.50
413
QUESTIONS AND PROBLEMS
In many of the following problems, assumptions are stated which would ordinarily
be made by the designer. They are given here in order to limit the number of solu¬
tions for convenience in checking results. The student should consider whether the
assumptions are reasonable, and he may wish to make other solutions with
altered assumptions.
The majority of the problems are of the design type (not analysis) with data taken
from actual installations. In some cases over-all dimensions, capacities, and weights
are stated which may not have direct use in the solution of the problem; in these
cases they are included to give the student a comprehension of the size of the unit or
component he is designing.
Chapter 2. Machine-design Computations
2-1. Determine the relative error in percentage due to an absolute error of 1 oz in

weighing (a) 1 lb, (6) 7\^ lb, and (c) 10 tons.
2-2. A rectangular bulkhead is found by measurement to be 12 ft wide and 8 ft high
with errors not exceeding 1 in. Determine the possible relative error in percentage in
the area of the bulkhead.
2-3. The seconds pendulum of a clock expands 0.05 per cent per degree centigrade.
How many seconds would the clock lose per day if the temperature is 1^ C above that
for which the pendulum beats true seconds?
Note:
t = 2tt
2-4. A straight cantilever is made of a bar with rectangular section. If the load on
the beam is increased 5 per cent, and if it is desired that the deflection of the beam
remains unchanged, determine the required percentage change in (a) the width of the
beam, or (b) the depth of the beam.
2-6. The natural frequency of vibration of a mass on the end of a cantilever is
fn = 188 \/1/8, where 8 = (PL3/3EI). If the length of the cantilever L is shortened
by 1 per cent, what will be the percentage of change in the natural frequency?
2-6. The indicated horsepower of an engine is given by the expression
PLAN
Ihp
33,000
Assume that the approximate values of the quantities are as follows:
P = mean effective pressure = 50 psi

L = length of stroke = 2 ft
A = area of piston, diameter D = 16 in.
N = working strokes per minute = 100
How precisely should P, L, D, and N be determined in order that the computed

horsepower may be reliable to 2 per cent?
415
416 MACHINE DESIGN
Chapter 3. Loading, Induced Stresses, and Failure
3-1. A machine member is subjected to an impact load P produced by a weight W

falling through a height h in. and striking the member so that it is stressed within the
elastic limit of the material. Derive an equation for the impact load P in terms of W,
h, and the flexibility constant C equal to pounds per inch of deflection, (P/8). The
final equation should be reduced to its simplest terms.
3-2. A 12-in. steel I beam resting on supports 15 ft apart is impulsively loaded by a
weight of 1,000 lb falling through a height of h in. and striking the beam at its mid¬
point. The moment of inertia I for the beam section is 246 in.4 Assuming that the
supports are rigid, determine the value of h for a maximum induced stress in the beam
equal to 20,000 psi.
3-3. Determine the maximum bending stress in the cast-iron beam with a hollow
section shown in the figure. Fi = 2,000 lb, and F2 = 3,000 lb.
Fig. P 3-3.
3-4. The spindle shown in the sketch is used to connect the brake shoe of an
industrial spring-set brake to the brake arm. The loads P are each equal to 825 lb.
Assuming an allowable stress equal to 17,500 psi, determine the diameter of the spindle
that you would recommend.
P p
1i 'r
i i
3" 3"
4 ^-S'
4
Fig. P 3-4.
3-6. A simply supported steel beam, rectangular in section and 8 ft long, is to carry
a concentrated load of 4,000 lb at a point 3 ft from one end. The depth of the section
is to be twice the width. Using an allowable working stress in tension and compres¬
sion of 10,000 psi, determine the recommended dimensions of the section.
3-6. A hollow nickel-steel shaft, whose outside diameter is twice the inside diameter,
is to transmit 2,000 hp at 220 rpm. The shaft is subjected essentially to torsion.
(a) Assuming an allowable stress of 32,500 psi, determine the size of hollow shaft
required. (b) Determine the size of solid shaft for the above conditions, using the
allowable stress as 25,000 psi. (c) Determine the percentage of saving in weight of
the hollow shaft over the solid one.
3-7. A standard railway-car axle carries a total load of 15,000 lb. On account of
the rounding of curves and swaying of the car, it may be assumed that one bearing
QUESTIONS AND PROBLEMS 417
can take % of the load. Draw the bending-moment diagram, and determine the
maximum stress in the axle.
Fig. P 3-7.
3-8. The load P on the C clamp shown in the sketch is 5,000 lb.
Assuming that the clamp is made of steel casting, h = 2b, and
e = 6 in., and that there is an allowable stress of 14,000 psi, deter¬
mine the recommended dimensions b and h.
Fig. P 3-8.
3-9. A steel link, as shown at (a), has a

cross section in- square. It is necessary
to replace this link by an offset one to pro¬ (a)
vide clearance, as shown at (b).
Assuming the depth of the replacement =^5.
link h = 2 in., determine the width required
so that the strength of the offset link will be Fig. P 3-9.
equal to that of the straight link.
3-10. A 28-ft lifeboat to carry 50 persons is supported from

two davits. The weight of the boat fully equipped is 3,000 lb
and an estimate of the weight of the occupants is 7,000 lb.
Assuming that the overhang L = 76 in. and that the davits
share the load equally, determine the diameter d for an allow¬
able nominal stress of 15,000 psi.
418 MACHINE DESIGN
Chapter 4. Stress Concentration in Machine Members
IP
4-1. A bar as shown in the sketch carries an axial load P =

6,500 lb. Determine the following: (a) the maximum nominal
stress in the bar; (b) the maximum stress in the bar for the
following fillet radii r: 34 in., 34 in-> 34
3"—-
Fig. P 4-1.
4-2. A flat bar as shown in the sketch carries an axial load P = 8,000 lb. Deter¬
mine the minimum radius r required so that the maximum stress in the member will
not exceed 16,000 psi.
VL_
^p n
r
\L _r
Fig. P 4-2.
P
4-3. A stepped shaft of circular cross X///////////////A
section carries a load P = 2,500 lb and is
supported as shown in the sketch. Deter¬
f#
0
7 >• y
1
T
Z*A//
mine (a) the maximum nominal stress in
the shaft and (b) the maximum stress in
the shaft. ^ V///////////////, l- -^ 4"
Fig. P 4-3.
4-4. A flat bar as shown in the
sketch is subjected to an axial tensile
load P equal to 50 tons. Assuming
that the stress in the bar is limited to
30,000 psi, determine the thickness t
required.
4-6. A shaft transmitting 50 hp at 870

rpm rests in a bearing as shown in the
sketch and has attached to it a pinion 3 in.
in pitch diameter. The shaft fillet radius
r = in. Assuming that the shaft is
rigidly supported by the bearing and that
the tooth load is uniformly distributed
Fig. P 4-5.
over the face of the gear, determine the
following: (a) the torsional moment in inch-pounds transmitted by the shaft; (6) the
maximum shearing stress in the shaft due to torsion; (c) the maximum bending stress
in the shaft due to bending.
4-6. The spindle shown in the

sketch is part of an industrial brake
and is loaded as indicated. Each load
P = 825 lb and is applied at the mid¬
point of its bearing area as shown.
Assuming that d\ = % in. and that
d2 = % in., determine the radius of
fillet r required in order that the maxi¬
mum stress in the part of the spindle
of length equal to % in. will not exceed the stress in the part of diameter d\.
r
4-7. A tension member, as shown at (a), supports an
axial load equal to P lb. It is necessary to replace this
member by one having a %-in. hole, as shown at (6).
Assuming that the width of the replacement member is
equal to 2 in., determine the thickness t and the fillet
radius so that the maximum stress will not exceed that
in (a).
V-t
P
1"-A v'h
2 *-+
(a) (b)
Fig. P 4-7.
420 MACHINE DESIGN
4-8. A bolt with NF threads, as shown in the sketch, is subjected to shock
loading in tension. The diameter of the body of the bolt in (a) is % in. and in (6) is
0.435 in. The maximum stress in the
(- _ Jc=2.S
bolt is equal to the stress at tjie root
diameter multiplied by the stress-con¬ f II 11
centration factor k shown in the sketch. \ II IT
k_ (cl)
Determine the percentage of im¬
provement in energy-absorbing capac¬
ity for the bolt in (b) over that in (a),
assuming that the maximum stress of
both bolts is of the same value, that the
material of the bolts is steel, and that
the energy is absorbed in the 2/^-in. (b)
portion of the bolts. Fig. P 4-8.
4-9. The spindle in the sketch is

repeatedly loaded by a weight W =
600 lb which falls through a distance
equivalent to h = 0.05 in. The
length L = 5 in. The material of
the spindle is SAE 2345 nickel steel.
Assuming that the support for the
spindle is rigid, determine the maxi¬
mum stress induced in the spindle.
4-10. The cutter bar of a hand-operated

embosser shown in the sketch is operated
by pressing at C. The cutting edge is AB.
The bar is made of hardened steel and
broke along section BD after about 100
operations. What do you feel was the
cause of the fracture and how could the
part be improved? Note that the sharp
corner near B was produced by a bevel-
edge grinding wheel.
4-11. A business machine has 9,999

punchings in its mechanism, as the one
shown in Fig. P 4-11. A part slides
along AB and comes in contact with the
cheek BC, thus producing an impact
load P. The corner at B was made
square to accommodate the sliding part n
which cannot be rounded so that the •I

punching has a very small radius at B.
Trouble developed when the nose broke
off owing to a fatigue crack originating
at B,
Assuming that the material (SAE 1025) and the outline dimensions cannot be
changed, what is your recommendation for improving the design? What tooling
changes would be required?
4-12. In a card-notching machine, the cutter is made by stamping a steel plate,
as shown in the sketch. The cutter is bent upward (before heat-treatment) as shown,
the card inserted at A, and the free end of the cutter is pressed down by the ram,
thereby cutting the notch in the edge of the card.
Allowing a stress-concentration factor at the fillets at the end of the slots equal to
1.3, determine the minimum length L to limit the bending stress to 20,000 psi. The
cards are 0.015 in. thick, and to notch the card the deflection of the cutter at A is
0.020 in. The plate thickness l is % 2 in.
Card
r~
1
n
i
1
<========^!
_ D
1
1
1
L
Chapter 5. Allowable Stresses
6-1. A flat bar, as shown in the

sketch, is to carry an axial load P =
//
12,000 lb. The material is to be SAE

p
T k£ii
1040 steel. *-S-i-
Assuming a factor of safety of 2, de¬ V
termine the thickness t of the bar for
± r
the following conditions: (a) assuming ■ — t
that the load P is a static load; (6)

Fig. P 5-1.
assuming that the load P is completely /
reversed and applied repeatedly.
6-2. A stepped shaft of circular cross section,

as shown in the sketch, is made of SAE 1045 steel. KWWWWNI
The load P is repeated and completely reversed
with a value of 2,500 lb. Assuming that r/d is /.5d
approximately }£, determine the diameter d and
the fillet radius r so that the maximum stress will
be limited to a value corresponding to a factor of Fig. P 5-2.
safety of 2.
422 MACHINE DESIGN
J"
t © —*rm*
6-3. The machine part shown in the
S" l1" >
sketch is to be loaded repeatedly by a 's '4
©
force P = 360 lb. Determine the re¬ \ -*-f
vvr** o7" vL
h* ^r>" ^T <p
quired thickness b for the part, assuming
SAE 1045 steel and a factor of safety of 2. r—4*--—■-,- L
~T
Fig. P 5-3.
6-4. A bar, as shown in the sketch, is - l F

P
to be subjected to a completely re¬ •*—
© .2d P
versed, repeated load P = 5,000 lb.
T_ _ T
Assume SAE 1020 steel. Determine
_L
T
the thickness t of bar required, using
_±_ .
a factor of safety of 2.
Fig. P 5-4.
6-6. A camshaft % in. in diameter is

mounted on ball bearings and carries a cam
midway between the bearings, as shown by
the sketch. The follower, which weighs 625
h
lb, is out of adjustment so that impact is pro¬ 1-1-
duced when it is contacted by the cam. The 1X1 1 1 E3
1 |
height of fall h equals 0.01 in. The shaft is Lj
made of SAE 1045 steel, (a) Determine the
p . O *4* O *
maximum bending stress in the shaft. (5)
Would the shaft be expected to fail, and if so, Fig. P 5-5.
under what conditions?
Chapter 6. Members That Fail by Buckling
6-1. The piston rod of a steam engine may be considered to be a column. Deter¬
mine the recommended diameter of the rod, assuming the following:
Diameter of cylinder. 18 in.

Maximum net steam pressure on piston. 85 psi
Distance from piston to crosshead. 64 in.
Material. SAE 1030 steel
Factor of safety. 4
End-restraint coefficient n. 3.5
6-2. A compression link of a valve mechanism may be considered as a column. The

link is to be circular in cross section and 18 in. long. The estimated axial load on the
link is 800 lb. Determine the recommended diameter for the link, assuming the
following data:
Material. SAE 1020 steel

Factor of safety. 2
End-fixity coefficient n. 3.0
6-3. A steam-engine piston rod is to be made of nickel steel. The rod is to be 56 in.
long and the maximum steam load carried by the rod is 45,000 lb. Assuming that the
yield point for the steel is 65,000 psi, the end-fixity coefficient is 3.5, and the factor of
safety is 4, determine the diameter for the rod.
T-T
6-4. A part of a nail-heading machine is to be circular

in cross section and 6 in. long, as shown in the sketch, and
is to carry a compressive load of 5,000 lb. Assuming
6"
SAE 1020 steel and a factor of safety of 2, determine the
diameter required.
Fig. P 6-4.
6-6. A steel strut, as shown in the sketch, is rec¬

tangular in cross section and is to be part of the oper¬
ating mechanism of a bascule bridge. The compres¬
sive load to be carried is 40,000 lb. Determine the
following: (a) the relation between b and w required
for equal strengths of the strut in failure by buckling
in either plane; (b) the required dimensions b and w,
assuming a factor of safety of 4 and SAE 1020 steel.
P 6-5.
6-6. Assume a reasonable material, and recommend the diameter of the piston rod
for the cylinder in Prob. 11-2.
Note: Assume that the length of the piston rod is 30 in., that its right-hand end is
pin connected to the operating lever, that the piston end of the rod is fixed ended, and
that the factor of safety is 2.5.
Chapter 8. Mechanical Fabrication and Processes
8-1. A static load P = 70,000 lb is to be supported by steel plates welded as shown

in Fig. P 8-1 by shielded-arc welds, (a) Assuming that ^-in. plates are to be used
and that each weld carries one-half of the load, determine the width of plate to be
specified. (6) Determine the maximum stress in the plates, and compare it with
the allowable value.
Fig. P 8-1.
424 MACHINE DESIGN
8-2. A 6- by 4- by 3^-in. angle is to be welded to a steel plate by fillet welds along
the edges of the 6-in. leg. The angle is to support a tension load of 60,000 lb. Deter¬
mine the lengths of shielded-arc welds to be specified.
8-3. A gear is to be made by welding a

steel hub, ring, and rim as shown in Fig.
P 8-3. The torque on the gear is steady
at 2,000 lb-ft. Determine the size of
welds to be recommended at the hub and
the rim.
8-4. Determine the thickness of the

plate and size of weld to be specified for
the bracket in Fig. P 8-4, where
F = 8,000 lb.
Chapter 9. Detachable Fastenings
9-1. For a National Standard bolt and nut made of steel, determine the relation
between the thickness of the nut and the outside diameter of the thread for equal
strengths of the bolt in tension and the threads in shear. Assume that the root
diameter of the threads is 0.85 X the outside diameter and that the allowable stress
in shear, sSf equals 0.55 X the allowable stress in tension, st.
9-2. A pulley bracket is supported by four bolts,

two at A and two at B, as shown in the sketch. The
weight of the pulley and bracket equals W = 175 lb,
and the load on the rope P is 6,000 lb. Assuming that
the bracket is held against the wall and prevented
from tipping about 0 by the two bolts at A (i.e., neg¬
lecting the bolts at B), and using an allowable tensile
stress in the bolts as 5,000 psi, determine the size of
NC bolts required.
9-3. A bolt is subjected to axial shock loading so that the maximum applied load
equals 2,400 lb. (a) Assuming an allowable stress of 8,000 psi, determine the size of
bolt required using NC threads. Neglect stress
concentration. (6) It is desired to increase the r \
* - <
shock-absorbing capacity of the above bolt. -\ |
Determine the diameter of the hole required to > -
(
>
i
reduce the cross-sectional area of the shank to s_ 4
that at the root of the threads, (c) Determine L-<?"-.
the percentage of improvement in shock-absorb¬ Fig. P 9-3.

ing capacity by using the drilled bolt.
9-4. Assuming that the load on a bolt due to tightening is equal to 16,000 times the
nominal bolt diameter d, derive an equation for the stress due to tightening. The
result should be in terms of d. Assume the root diameter is 0.8 X the nominal diam¬
eter. What should be the minimum size of the bolts used on the basis of the above?
9-6. A safety stop is composed of steel tension bolts ar¬

ranged to catch a platform in case of failure of a support. In
this case each bolt may be subjected to the impact of W =
2,500 lb, dropping through a distance h = 0.05 in., as shown
schematically by the sketch. The bolts are turned down to
% in. in diameter, which is less than the root diameter of the
threads. It is expected that the safety stop will be required
to function infrequently. Using a factor of safety equal to
1.25, what mechanical property of the material and what
value of that property would you specify?
9-6. For a sliding hub connected to a shaft by means of feather keys, show that the
friction of sliding when two keys are used is one-half the friction when one key is used.
9-7. Draw a neat sketch of a shaft fitted with a key, and derive equations for (a) the
bearing pressure on the key and (6) the shearing stress in the key. The results should
be in terms of the transmitted torque T, the shaft diameter D, length of key L, width of
key w, and depth h. (c) Assuming that the key is equally strong in bearing and shear
and that the yield point in compression is twice the yield point in shear, determine
the proper proportions of the cross section of the key. (d) Assuming that there is a
square key whose sides are one-quarter the shaft diameter and that the shaft, key, and
hub are made of the same material, determine the length of key in terms of the shaft
diameter so that the key will be as strong as the solid shaft.
9-8. The low-speed shaft of a 20-hp speed reducer rotates at 384 rpm and is 1% in.
in diameter. Determine the size of key that should be recommended and compare the
shearing stress in the key with the shearing stress in the solid shaft.
9-9. Determine the size and length of the key you would specify for a 5-hp 1,800-
rpm motor (see Table 25-2).
9-10. A line shaft 2% 6 in. in diame¬
ter is to be driven by a belt running over
a 36-in. pulley. The belt tensions are
Fi = 1,500 lb and F2 = 1,000 lb. The
hub of the pulley and its key are both 6
in. long, (a) Using a standard square
key, determine the shearing stress and
the bearing pressure for the key. (b)
Determine the loads coming on the
bearings A and B.
426 MACHINE DESIGN
9-11. Make a neat sketch showing two views of a knuckle joint, and write equations
showing the strength of the joint for the most probable methods of failure.
Note: Assign notation to represent the various dimensions of the joint and express
the equations for strength in terms of this notation.
9-12. Make a neat sketch showing two views of a cotter joint with key, and write
equations showing the strength of the joint for the most probable methods of failure.
Note: Assign notation to represent the various dimensions of the rod end, socket,
and cotter, and express the equations for strength in terms of this notation.
9-13. For the cottered connection

shown in the sketch, determine the
maximum load P that should be
applied for the following allowable
stresses: = 5,000 psi, s« = 2,500 psi,
and Sb = 12,000 psi. Neglect stress
concentration.
Fig. P 9-13.
9-14. The torque capacity of a spline

fitting is based on the allowable pressure on
the active surfaces of the splines. Assuming
a value of this pressure as 1,000 psi, deter¬
mine for the spline in Fig. P 9-14 the follow¬
ing: (a) the length of spline required for a
torque capacity of 500 lb-ft; (b) the force
required to slide the member axially under
load, assuming the coefficient of friction
as 0.1.
Chapter 10. Springs
10-1. a. Derive the following equation for the shearing stress in a helical spring in
which P is the axial load in pounds, D is the mean coil diameter in inches, d is the wire
diameter in inches, and C is the spring index:
8 PD = 8 PC
7T d3 Trd2
b. How may the above equation be modified to take into account the effect on the
stress due to direct shear, wire curvature, etc. ?
10-2. Derive the following equation for the deflection 8 of a coil spring having n
active coils of mean diameter D, made of wire of diameter d and subjected to an axial
load P.
8PDsn 8 PC*n
6 ~ Gd4 Gd
Write an expression for the spring rate in terms of physical constants of the spring.
10-3. A helical steel compression spring is 4 in. in outer diameter and has eight
active coils of 34-in. round wire. Assuming average service, determine the following:
(a) spring index C; (6) stress factor K; (c) permissible axial load; (d) deflection; (e)
spring rate.
10-4. A helical coil spring is to be subjected in service to loads ranging from 520 to
650 lb. The axial compression of the spring over the above load range is to be approxi¬
mately 34 in. Assuming a spring index of approximately 5, determine the following:
(a) the size of wire; (b) the outside diameter of the spring; (c) the number of active
coils.
10-6. A loaded narrow-gauge car weighing 4,000 lb and moving at a velocity of
240 fpm is brought to rest by a bumper consisting of two helical steel springs. In
bringing the car to rest the springs are to be compressed 8 in. Assume the following:
Allowable stress. 52,000 psi

Spring index. 6
Determine the following: (a) maximum load on each spring; (b) diameter of wire;
(c) mean diameter of coils; (d) number of effective coils.
10-6. A weight of 160 lb located at a position 234 ft above the center of a platen is
to drop on the platen. The platen is to be supported on four helical springs. Assume
the following:
Deflection of each spring. 3 in.

Spring index. 5
Maximum stress. 60,000 psi
Determine the diameter of wire, outside diameter of the coils, and the number of
active coils.
10-7. A valve for a hydraulic pump is to have a lift of % in. and is to be closed
by a helical compression spring having a load range of approximately 100 to 175 lb
from valve-closed to valve-open conditions. Assuming SAE 6150 steel and a spring
index of 6, determine the wire diameter, outer diameter of coils, and number of active
coils to be recommended.
10-8. A circular cam 8 in. in diameter rotates
off center with an eccentricity e — 1 in. and
operates the roller follower that is carried by the
arm as shown by the sketch. The roller fol¬
lower is held against the cam by means of an
extension spring. The device is part of a book¬
binder. Assuming that the force between the
follower and the cam is to be approximately
50 lb at the low position and 80 lb at the high
position and that there is a spring index of 7,
determine the diameter of wire, outside diameter of spring, and number of active coils
you would recommend. The spring may be assumed to be attached at the third point
on the arm.
428 MACHINE DESIGN
10-9. In a free-rolling conveyor, a crate loaded with nonfragile material, 60 lb tare,
reaches station A at a velocity v = 10 fps and is normally unloaded at station B.
In case it “gets away” from B, a pair of helical buffer springs at C are expected to
arrest the motion of the crate. Assume that the springs are set up with an initial
compression of 2 in., and that the compression of the spring due to impact is to be
6 in. Determine the following: (a) maximum load on each spring; (6) diameter of
wire and diameter of coils; (c) number of coils.
10-10. A helical coil spring for an industrial spring-set brake is to be made of SAE
6150 wire and is to be loaded in compression by an axial load varying from a minimum
of 650 lb to a maximum. The deflection of the spring over the service load range is
to be approximately % in.
a. Assuming average service, a spring index of 5, and that the range of stress during
the service loading will be from 80 to 100 per cent of the maximum, determine the
diameter of wire and the mean diameter of coils.
b. Determine the number of active coils required and the actual spring rate in
pounds per inch.
c. Determine the yield point in torsion for the wire, and compute the corresponding
spring load and deflection.
d. Assuming that the spring ends are squared and ground, and using the deflection
found in (c), determine the pitch of the coils and the actual free length of the spring
so that the closure stress will not exceed the yield point.
e. Make a copy of the table shown below and enter values.
/. Check the possibility of buckling.
g. Make a neat freehand sketch approximately to scale of the spring in its free
condition.
Specifications Characteristics
Outside diameter, in. Load Stress Deflection Length
Wire diameter, in. Free
Total number of coils 650
Pitch of coils, in.
Free length, in. Solid

10-11. A Westinghouse-Nuttall cou¬

pling, as shown in Fig. 16-8, is to have a
rated torque of 2,800 lb-in. There are to
be four springs located as shown in Fig.
P 10-11 in which the radius of the spring
center line R = 4 in. (a) Assuming
that the angular displacement between
the coupling halves at twice rated torque
is to be 6 deg, determine the spring rate.
(6) Assuming that the springs are to have
an initial compression of 0.1 in. when the
coupling is unloaded, determine the load
on the springs at twice rated torque.
10-12. A torsion-bar spring is made of a solid, circular steel bar 2.27 in. in diameter
with upset ends 2.75 in. in diameter for spline fittings. The splines are composed of
55 serrations 0.037 in. deep with sides having 100-deg included angle. One end of the
bar is anchored to the piece of ordnance equipment of which it is a part, and the other
end carries a torque arm 14 in. long. The length of the bar between splines is 72.7 in.
The maximum force on the torque arm is 23,800 lb. (a) Make a sketch of the bar.
(6) Determine the maximum torque on the bar in inch-pounds. For maximum
torque, determine the following: (c) shearing stress in the body of the bar; (d) pressure
on the flanks of the splines; (e) wind-up of the bar in degrees; (/) motion of the free
end of the arm in inches.
10-13. In Fig. 10-18 is shown a Neg’ator spring counterbalance in which the Neg’ator
spring in tending to rewind itself on the storage bushing rotates the output bushing
and the cable drum to support the weights on the cables A, B, and C. The total
weight on the three cables is 36 lb. Determine the constant spring force required to
produce the counterbalancing. The diameter of the cable drum is 1.74 in., the
diameter of the output bushing is 4 in., and the gear ratio between the output bushing
and the cable drum is 4:1.
Chapter 11. Pressure Cylinders
11-1. The drum of a steam-generating

unit is 18 ft long and 48 in. in diameter. The
steam pressure is 500 psi gauge. Assuming
an allowable tensile stress of 10,000 psi and
that the efficiencies of the circumferential
and longitudinal joints are 65 and 90 per
cent, respectively, recommend the thickness of shell.
11-2. The force analysis of a 7^-ton air-operated arbor press shows that the piston
rod for the operating cylinder must exert a maximum force F = 4,500 lb. The air
n
"p -
3*-
\
u u
Fig. P 11-2.
430 MACHINE DESIGN
pressure p in the cylinder is 100 psi. (a) Determine the diameter of cylinder bore
required, assuming that over-all friction due to stuffing box and piston packing is
equivalent to 10 per cent of F. The piston bore should be selected on the basis of
3^-in. increments. (6) Determine the thickness of cylinder, assuming that it is
seamless-steel tubing. The allowable tensile stress is 3,000 psi. (c) Determine the
thickness of the cylinder head, assuming that it is made of cast iron and that the allow¬
able stress is 3,000 psi.
11-3. An hydraulic control for a straight-line motion utilizes a spherical pressure
tank A that is connected to a work cylinder B as shown. A pump maintains pressure
in the tank equal to 400 psi.
Fig. P 11-3.
a. Assuming that the tank A is 32 in. in diameter, that it is welded with joints
having strength equal to that of the plate, and that the tank is made of steel plates
having an allowable tensile strength of 7,500 psi, determine the thickness of the plate
required for the tank.
b. Assuming a pressure drop of 30 psi between the tank and the cylinder, determine
the diameter of piston required to produce an operating force F of 5,000 lb. Assume
an allowance for friction in the cylinder and packing equal to 10 per cent of F.
c. Determine the thickness of the cylinder wall, assuming that it is made of cast
iron having an allowable tensile stress of 4,000 psi.
d. Determine the horsepower output of the cylinder during a working stroke, assum¬
ing the piston stroke is 18 in. and that 5 sec is required for a work stroke.
e. Assuming that the work cycle of the piston rod occurs once every 30 sec, that the
over-all efficiency of the hydraulic control is 80 per cent, and that the pump efficiency
is 60 per cent, determine the horsepower required for the motor continuously operating
the pump.
Chapter 12. Translation Screws
12-1. a. Make a neat sketch showing one turn of a right-hand square-threaded

screw, and indicate thereon the following notation:
Q = axial load, lb
d = mean diameter of the thread, in.
a = lead angle at mean diameter
<f) = friction angle
L = lead of thread, in.
b. Derive an equation for the torque T required to raise the load and to overcome
thread friction in terms of Q, d, f, and L.
c. Write the expression for the torque Tc required to overcome collar friction.
12-2. A sluice gate weighing 60 tons is raised and lowered by means of two 2^-in.
Sellers standard square-threaded screws. The screws are operated by a 570-rpm
motor. A ball thrust bearing is used at the collar, reducing the coefficient of collar
friction to 0.03 at a diameter of 4 in. The coefficient of thread friction may be
assumed to be 0.14.
If the gate is to be raised at the rate of 2 fpm, determine the following: (a) the rpm
of the screws; (b) the horsepower of motor required to raise the gate, assuming a me¬
chanical efficiency of 0.85 for the speed-reducing equipment.
12-3. A screw having a double square thread is to raise a load of 20,000 lb at a speed
of 3 fpm. The following data apply:
Outside diameter of thread. 2% in.

Pitch. 1 in.
Mean collar diameter. 5 in.
Coefficient of thread friction. 0.10
Coefficient of collar friction. 0.12
Determine the following: (a) the horsepower required to drive the screw; (6) the
efficiency of the screw.
12-4. The screw of a 30-ton toggle press rotates at 72 rpm and is driven by means of
the gear g shown in the sketch. The crossheads n move axially along the screw and in
*9
MtWliWh'liWWMiWirS?!
n n
Fig. P 12-4.
opposite directions since one head has left-handed threads and the other right-handed
threads. They move against the axial forces S in operating the press. Assuming that
S = 23,000 lb, that the screw is a square-threaded type 2^4 in. in outside diameter and
has two threads per inch, and that there is coefficient of friction of 0.10, determine the
horsepower required to drive the gear g. The friction at the supporting bearings may
be neglected.
12-6. The screw of a shaft straightener, as shown in

the sketch, exerts a load P = 6,000 lb. The screw is
3 in. in outside diameter and has four square threads
per inch, (a) Determine the force required at the rim
of the 12-in.-diameter hand wheel, assuming that the
coefficients of thread and collar friction are 0.125 and
that the mean diameter of the collar is 2^ in. (6)
Determine the maximum compressive stress in the
screw, the bearing pressure on the threads, and the
shearing stress in the threads, (c) Determine the
efficiency of the straightener. (d) Determine the size
of the two bolts required to fasten the straightener to
the base, assuming the allowable stress is 8,000 psi.
Neglect tightening-up stresses.
Fig. P 12-5.
432 MACHINE DESIGN
12-6. The sketch shows the arrange¬

ment for a type of linear actuator for air¬
craft controls, for example, landing gear
and bomb-bay doors. These units weigh
from 4 to 15 lb. The screw for such an
actuator is to be designed and the motor
selected.
Fig. P 12-6.
Data:
Operating force on actuator 1,000 lb
Maximum static force 2,500 lb
Extension of actuator. . . . 5 in.
Time for extension. .. 6 sec
Motor speed. . . . 7,200 rpm
a. Determine the root area of the screw for the maximum static force, using an
allowable stress in direct tension or compression equal to 10,000 psi. Select a Sellers
standard square thread.
b. Using a coefficient of thread friction of 0.08 and neglecting collar friction since a
ball thrust bearing is used, determine the torque on the screw in pound-inches, cor¬
responding to the operating force on the screw.
c. Assuming an over-all efficiency for the triple-reduction gear between the motor
and screw such as 94 per cent, determine the horsepower of motor required.
d. Determine the over-all gear ratio from the motor to the screw.
12-7. A portable conveyer is to be raised

and lowered by means of the double-screw
arrangement shown in the sketch. The two
screws are supported by collars A and are
connected by the roller chain B and rotated
by the bevel gears C, operated by the hand
crank as shown. The movable frame E is
raised and lowered by the screws against a
load W = 1,000 lb, but when the frame is
in position the load may be increased to
2,000 lb.
Assume the following data:
Allowable compressive stress in the screw. . 1,000 psi

Mean diameter of collar. . outside diameter of screw
Coefficient of friction. . 0.15
Chain efficiency. . 90 per cent
Bevel-gear efficiency. . 90 per cent
Bevel-gear ratio. 2:1 (pinion on crankshaft as shown)
Maximum tangential force on crank. . 20 lb
Use Sellers’ square threads.
Determine the size of the screws to be used and the length of crank arm required.
12-8. In a shipbuilding crane, as shown in Fig. P 12-8(a), the travel of the load is
controlled by three operations: (1) raising and lowering the load vertically by reeling
in or out of the cable on the main hoisting drum, (2) raising or lowering (luffing) the
boom, and (3) rotating the crane on a vertical axis. Each of these operations is
controlled by a separate drive.
w
Fig. P 12-8.
In one of the largest cranes in the world (see Mech. Eng., September, 1944, p. 569),
luffing of the boom is accomplished by rotating two screws located side by side, as the
one shown in (a) in the figure. Each of the two screws is driven by a motor through
a gear keyed to the screw at its mid-point. The screw has left- and right-hand
threads, each running with a bronze nut. One nut is trunnioned to the base of the
crane and the other trunnioned to the boom so that when the screw is rotated the
boom will be raised or lowered. The dimensions of the screw are shown in (b), and
at (c) is shown a detail of the buttress thread. Other data are as follows:
Capacity of crane = 75 tons
Maximum tensile load on screw = 327,000 lb
Material of screw is SAE 2340 steel with yield point in tension = 70,000 psi
Material of nut is bronze alloy with yield point in tension = 55,000 psi
Pitch of threads = 2}^ in.
Determine the following:

a. Maximum torque in inch-pounds on each screw, assuming a coefficient of friction
equal to 0.07. (The threads are very accurately cut, finely finished and EP grease is
used as a lubricant.)
b. The factor of safety for the screw.
c. The factor of safety for strength of the threads.
434 MACHINE DESIGN
d. The time required to raise the boom (same as for the nut to traverse the threaded
portion of the screw), assuming that the screw rotates at an average speed of 30 rpm.
e. The horsepower required for hoisting at the lowest position of the boom at which
time the screw rotates at 20 fpm. This is the maximum horsepower and the selection
of motor would depend on the duty cycle for the hoist.
Chapter 1 3. Shafting
13-1. A shaft transmitting 50 hp at 870 rpm is supported in a bearing as shown in

the sketch and has attached to it a pinion 3 in. in pitch diameter. Assuming that the
shaft is rigidly supported in the bearing and
that the tooth load is uniformly distributed
over the face of the gear, determine the follow¬
ing: (a) the torsional moment in inch-pounds
transmitted by the shaft; (6) the maximum
bending moment on the shaft, assuming 20-
deg involute teeth; (c) the shaft diameter
required, assuming an allowable stress in shear
equal to 6,600 psi.
13-2. An overhung crank, as shown in the sketch,

supports a load P = 1,000 lb. Assuming an allowable
stress in shear equal to 6,600 psi, determine the diam¬
eter of shaft required.
13-3. A steam engine that has a

stroke of 12 in. has an overhung crank
as shown by the sketch. The maxi¬
mum tangential force P on the crank
may be assumed as 7,500 lb. Assuming
an allowable stress in shear as 4,400 psi,
determine the recommended diameter d
for the crankshaft.
13-4. A mild-carbon steel shaft, transmitting 20 hp at 210 rpm, is supported on two

bearings that are 30 in. apart and has two gears keyed to it. The pinion, having
24 teeth of 4 diametral pitch, is located 4 in. to the left of the right-hand bearing and
delivers the power horizontally to the right. The gear, having 80 teeth of 4 diametral
pitch, is located 6 in. to the right of the left-hand bearing and receives power in a
vertical direction from below. Using an allowable working stress in shear as 7,750 psi,
determine the diameter of the shaft.
13-5. A centrifugal circulating pump for a steam-turbine condenser is to be direct-
connected to a motor through a short shaft and a pair of flexible couplings. The pump
delivers 5,500 gpm at 865 rpm against a dynamic head of 25 ft. The pump efficiency
is 79 per cent, (a) Determine the horsepower of motor required. (6) Assuming hot-
rolled SAE 1020 steel, determine the diameter of shaft required using a factor of safety
equal to 4. (c) Compare the diameter of the connecting shaft with the diameter of the
motor shaft, and state reasons for any difference.
13-6. The solid shaft of a high-speed boat is 32 ft long from sprocket to propeller,
as shown in the sketch. At maximum torque, the horsepower transmitted to the shaft
is 300 at 130 rpm of the shaft. The sprocket is 12 in. in diameter. Assuming the
maximum shear theory for the design, with an allowable stress in shear of 8,000 psi,
determine the diameter of the shaft to be specified.
^Sprocket
-32--
Fig. P 13-6.
13-7. The shaft of a hand-operated wire-rope

hoist is supported on the bearings A and B shown
in the sketch and has keyed to it the drum, which
is 6 in. in diameter. The tangential load on the
drum at rated capacity of the hoist is 125 lb. As¬
suming 100 per cent overload capacity and an
allowable shear stress of 5,000 psi, determine the
shaft diameter to be recommended.
13-8. A 5-hp 570-rpm motor is to drive a

countershaft, as shown in the sketch. The
motor pinion has 30 teeth of 12 diametral
pitch and meshes with a 120-tooth gear. The
countershaft has keyed to it a pinion having
24 teeth of 4 diametral pitch meshing with a
72-tooth gear that drives a rotary kiln. As¬
suming a starting overload of 30 per cent and
an allowable stress in shear of 5,500 psi, deter¬
mine the diameter of shaft you would
recommend.
13-9. The pinion of a water pump is at¬
tached, as shown in Fig. P 13-9, to the shaft
which is supported on ball bearings A and B.
The data are as follows: pinion has 15 teeth;
diametral pitch = 5; rpm = 1,200; horse¬
power = 50. (a) Using an allowable stress
ss = 8,000 psi, determine the diameter
of shaft required. (6) Using a standard
square key for the gear, determine the shear
stress in the key.
436 MACHINE DESIGN
13-10. Refer to the data and sketch in Prob. 3-7. The detail of the ends of the
axle is shown in Fig. P 13-10. (a) Determine the maximum bending stress at the
section A in the figure. (6) Determine the radius of the fillet r so that the maximum
stress at the fillet will not exceed vthe stress at the section A.
13-11. A horizontal belt conveyor arranged as shown in the sketch operates at a

speed of 270 fpm and requires 21 hp at the 30-in. diameter driving drum. The drum
receives power from a countershaft through a pair of spur gears. A motor operating
at 860 rpm delivers power to the countershaft through a roller chain. The mechanical
efficiency of the chain drive is 94 per cent and of the gears is 96 per cent, (a) Deter¬
mine the horsepower of the motor you would recommend. (b) Calculate the number
of teeth on the large sprocket required to run the conveyor belt at the speed specified,
(c) Determine the torque in inch-pounds on the countershaft corresponding to the
24T 4P /24T
5"
A
Z I
\
42" 1
c;
270 fpm
£
T“ ' ) y k -\
H 5"
\
1_
Chain, )"piich
rated horsepower of the motor. (d) Assuming an allowable stress in shear equal to
5,500 psi, determine the diameter of countershaft you would recommend.
Chapter 14. Belt Drives and Hoists

14-1. a. Derive the equation for the ratio of belt tensions in a flat belt transmitting
power. The following notation should be used:
Fi = tension in tight side of belt, lb

F2 = tension in slack side of belt, lb
/ = coefficient of friction between belt and pulley
6 = angle of contact of belt with the pulley
b = width of belt, in.
t = thickness of belt, in.
w = weight of belt material, lb per in.3
v = belt speed, fps
b. Assuming that the ratio of belt tensions is (Fi — FC)/(F2 — Fe) = efd, derive the
equation for the horsepower transmitted per square inch of belt section. The final
equation should be in terms of /i, fr and of the belt velocity v, where/i is the maximum
stress in pounds per square inch and/c is equal to 12(wv2/g).
14-2. A 100-hp Corliss engine, running at 95 rpm and having an 8-ft fly wheel, is
connected to a 42-in. pulley by means of an open leather belt. The distance between
the shafts is 233+ ft. Assume that cemented joints are used. Determine the follow¬
ing: (a) the minimum angle of contact, (6) the cross-sectional area of belt required, and
(c) the thickness, width, and length of belt.
14-3. Two pulleys which are 24 in. in diameter and which run at 370 rpm are con¬
nected by a leather belt % in. thick. If the belt transmits 30 hp and the distance be¬
tween the shafts is 10 ft, determine the width and length of belt that is required.
14-4. A 40-hp 1,200-rpm motor is to drive an air compressor through the medium
of V belting. The mean diameters of the motor pulley and the compressor pulley are
9 in. and 36 in., respectively. The shaft center distance is 55+ in. Determine the
size and number of V belts to be recommended.
14-5. Select a V-belt drive for connecting a 2-hp 1,170-rpm motor to a 300-rpm
blower which is part of an air-conditioning unit. The center distance may be taken
as approximately 13+ times the diameter of the large sprocket.
14-6. A block and tackle, having two sheaves at the top block and two at the hook,
with the rope anchored at the top block, is reefed with wire rope. Determine the fol¬
lowing: (a) an equation for the effort P required to raise a load Q, in terms of Q and
the coefficient C; (6) the force P required to raise a load Q = 2 tons, assuming C =
1.076; (c) the efficiency of the hoist.
14-7. A hoist is arranged as shown in the sketch, in

which a hydraulic cylinder applies a force P in order to
raise a load Q. The ratio of rope tensions is equal to C,
where C is greater than unity. Derive an equation for P
in terms of Q and C.
Fig. P 14-7.
438 MACHINE DESIGN
14-8. A hoist is arranged as shown in the sketch, in which a hy¬

draulic cylinder applies a force P in order to raise the load W = 7,500
lb. Determine the diameter d of the rope required, assuming 6X19
steel rope, a factor of safety of 4, sheaves 45d in diameter, and a
coefficient C of 1.10. Neglect weight of rope.
////////////.
14-9. The pulley system shown in the sketch is used to raise a

load of 10,000 lb with a maximum acceleration of 10 fps.2 Deter¬
mine the diameter, d, of rope required, assuming 6 X 19 steel rope
with a factor of safety of 5, sheaves 45d in diameter, and a coeffi¬ p
\
cient C of 1.10. Neglect weight of rope. Effort u

Load i
Fig. P 14-9.
Z///////////X
14-10. The pulley system shown in the sketch is to be used to

raise a weight of 10,000 lb with a maximum acceleration of 10 fps.2
Determine the diameter d of the wire rope, assuming a factor of
safety equal to 5, a coefficient C of 1.10, and sheaves of the follow¬ <T>
ing diameters, 20d, 30d, 40d, and 50d.
w
Fig. P 14-10.
A
14-11. A hoisting winch for a Maritime Commission vessel is arranged as shown in
the sketch. The following data apply:
Rated load on drum cable. 3,800 lb

Maximum load on cable. 10,000 lb
Drum diameter for first layer of 6 X 19, %-in. rope. 16% in.
Motor speed (at rating). 520 rpm
Mechanical efficiency, motor to drum. 85 per cent
Determine the following: (a) speed of drum cable at rating; (6) horsepower of motor
for rated load and speed; (c) factor of safety for cable at maximum load; (d) the load
V.
Fig. P 14-11.
that may be hoisted and the speed of hoisting, using a three-part line with 16%-in.
sheaves (assume that C = 1.076); (e) the required rating of the brake in pound-feet
for rated load on the winch.
14-12. A hand-operated wire-rope hoist is to be

arranged as shown in the sketch to raise a load
Q = 250 lb. The force F on the operating lever
is limited to 35 lb. The lever arm L = 12 in., the
diameter of the drum A is 8 in., and the diameter
of the sheaves is 10 in. Determine the following:
(a) the number of strands of rope leading to the
hook block; (6) the expression for the tension in
the rope T in terms of Q and C; (c) the factor of
safety of the hoist, assuming that it is reefed with
in. 6 X 19 wire rope.
Chapter 1 5. Power Transmission Chains

15- 1. A 25-hp squirrel-cage induction motor running at 1,150 rpm is to be connected
to a centrifugal pump by means of a roller chain. The pump operates at a speed of
approximately 400 rpm and is located 30+ in., center to center, from the motor. A
service factor of 1.2 may be assumed. Determine the pitch, width, and length of the
chain you would recommend, and also the pitch diameters of the sprockets.
16- 2. A 10-hp motor running at 1,150 rpm is to drive a compressor through the
medium of a roller chain. The compressor runs at approximately 300 rpm and is to be
located 20+ in. from the motor shaft. Determine the following: (a) pitch of chain;
(b) number of strands; (c) number of links; (d) pitch diameters of sprockets.
16-3. It is desired to transmit 40 hp by means of a roller chain at a motor-sprocket
speed of 720 rpm. The pitch diameter of the motor sprocket must not exceed 6% in.
Assuming uniform service 24 hr per day, determine the pitch and number of strands
required for the chain.
440 MACHINE DESIGN
16-4. A 10-hp motor running at 1,750 rpm is to drive a blower through the medium
of a silent chain. The blower runs at approximately 600 rpm and is to be located
28+ in. from the motor shaft. Determine the following: (a) pitch of chain; (6) width
of chain; (c) number of links; (d+pitch diameters of sprockets.
16-6. A ventilating fan having a delivery of 18,000 cfm when operating at 298 rpm
requires 2.84 hp under these rated conditions. The fan is to be installed in a hospital
where quietness of operation is essential.
The fan is to be connected to a squirrel-cage induction motor by means of a chain
drive.
Assumptions:
Motor speed. 1,160 rpm

Minimum number of teeth on small sprocket. 19
Efficiency of chain drive. 95 per cent
Determine the following: (a) horsepower of motor required; (b) type of chain; (c)
pitch and width of chain; (d) number of teeth and diameters for sprockets; (e) length
of chain.
16-6. A silent chain drive, as specified below, is to be replaced by a V-belt drive.
Assuming that 6-in. sheaves are to be used, determine the number of B-section V belts
required for the replacement drive.
Pitch. . % in.
Width. 13+ in.
Sprockets. 21 teeth
Speed. 1,120 rpm
Velocity ratio. 1:1
16-7. An 1,160-rpm motor is to be used to drive an oil-field pumping unit that re¬
quires 280 lb-ft torque at approximately 320 rpm. Assuming that a roller chain drive
is to be used and that there is an efficiency of 95 per cent and a service factor of 1.5,
determine the following: size of motor, number of sprocket teeth, and pitch and num¬
ber of strands of chain you would recommend.
16-8. In a roller-chain drive arranged as shown in Fig. 15-1 (c), a chemical processing
unit is connected to each of the two sprockets B. Each of the units requires 250 in.-lb
torque. The driving sprocket at A rotates at 100 rpm and has 11 teeth, while the
sprockets at B have 33 teeth each. The chain is one-strand No. 41 chain. Deter¬
mine the service factor for the drive.
Chapter 16. Shaft Couplings

16-1. (a) Make a neat sketch of a flanged shaft coupling showing a face view
and a cross-sectional view. (6) Derive expressions in terms of the torque T in inch-
pounds for the following: the shearing stress in the shaft; the shearing stress in the
coupling bolts; the bearing pressure on the bolts.
Use the following notation:
d = shaft diameter
R = radius of bolt circle
n = number of bolts
d\ = bolt diameter
t = thickness of radial flange
16-2. In a flanged shaft coupling having a l^-in. bore, it is desired that the tor¬
sional stress in the shaft will not exceed 3,500 psi. The outside diameter of the
coupling is 7^ in. There are three %-\n. bolts on a bolt circle 2% in. in radius.
The radial flange thickness is % in. Determine the following: (a) the horsepower
that may be transmitted at 600 rpm; (6) the shearing stress in the bolts; (c) the bearing
pressure on the bolts.
Chapter 17. Clutches and Brakes
17-1. Assuming that the pressure is uniformly distributed over the friction surfaces
in contact, derive an expression for the torque T transmitted by a plate friction clutch
in terms of the coefficient of friction/, the axial force P, and the outer and inner diam¬
eters of the plates D and d, respectively.
17-2. Assuming that the normal wear of the surfaces in contact is proportional to
the work of friction, derive an expression for the torque T transmitted by a plate
friction clutch in terms of the coefficient of friction /, the axial force P, and the outer
and inner diameters of the plates D and d, respectively.
the work of friction, derive an expression for the torque T transmitted by a cone clutch
in terms of the coefficient of friction /, the axial force P, outer and inner diameters D
and d, and the cone angle a.
17-4. A multiple-disk clutch with 17 disks, 8 disks turning with one shaft and 9 with
the other shaft, is rated at 25 hp at 500 rpm. The outside and inside diameters of the
disks are 5/£ and 2% in., respectively. The coefficient of friction is 0.2. (a) Deter¬
mine the axial force required to transmit the rated torque; (6) determine the pressure
between the disks.
17-5. An automotive engine develops maximum torque at 1,000 rpm, at which speed
the horsepower equals 34.6. The engine is to be equipped with a single plate clutch
having two pairs of friction surfaces, (a) Using the equation for torque derived on
the basis of “normal wear is proportional to the work of friction,” and assuming that
the coefficient of friction is equal to 0.4 and that the mean diameter of the clutch
disks is 7% in., determine the axial force required to apply the clutch. (b) Assuming
that the allowable average normal pressure on the clutch faces is 18 psi, determine the
outer and inner diameters required for the disks.
17-6. The large diameter and face of the disks of a multiple-disk clutch are 10 in.
and 1 in., respectively. The helical compression spring used to engage the clutch has
9.5 effective coils of %-in. steel wire, the outer diameter of the coils being 3^ in. The
free length of the spring is 734 in. and when in place with the clutch engaged, its length
is 534* in. Assuming that there are 10 pairs of friction surfaces in contact, that the
motor runs at 1,200 rpm, and that the coefficient of friction is 0.15, determine the
following: (a) the axial force produced by the spring, and (6) the horsepower that
the clutch will transmit.
the work of friction, derive an expression for the torque T that can be resisted by a
single block of a flat-rim block brake for a given radial force on the block P.
Let b = width of block
D = diameter of wheel
6 = half the angle of contact of block and wheel
442 MACHINE DESIGN
17-8. Determine the horsepower transmitted

by the double-block clutch shown in the figure,
assuming the following data: The drum is 10 in.
in diameter and rotates at 300 rpm. The con¬
tact angle for each shoe is 90 deg and the co¬
efficient of friction is 0.35. Assume the force on
each block P = 285 lb.
17-9. Determine the torque that may be re¬

sisted by the single-block brake shown in the
figure, assuming the following: The brake
drum is 10 in. in diameter, a = 4 in.; R =
140 lb. The contact angle is 90 deg. Coeffi¬
cient of friction between drum and lining is
0.35.
17-10. Solve Prob. 17-9, assuming a = 6 in.
17-11. The block brake shown in the sketch is set by

a spring that produces a force S on each arm equal to
682 lb. The wheel diameter is 14 in. and the angle of
contact for each block is 110 deg. Using 0.40 as a
value of the coefficient of friction between the drum
and lining, determine the maximum torque in foot¬
pounds that the brake is capable of absorbing.
17-12. Solve Prob. 17-11, assuming counterclockwise rotation of the wheel.

17-13. Solve Prob. 17-11, assuming that the 1}^ in dimensions are changed to 7 in.
17-14. The wheel of the double-block spring-set
brake shown in the sketch is 12 in. in diameter and
the dimensions are a = 8 in. and b = 10 in. (a)
Assuming that there is a coefficient of friction of 0.4
and that the angle of contact for each block is 110
deg, determine the spring force that is required for
the brake to resist a torque of 600 lb-ft. (b) Assum¬
ing that the value of the design factor pV is 28,000
ft-lb per min per sq in. of projected bearing area and
that the normal operating speed of the drum is 300
rpm, determine the width of brake shoes required.
17-15. The layout and dimensions of a block

brake are shown in the figure, (a) Assuming that
the coefficient of friction for the brake lining and
wheel is 0.4 and that the contact angle for each block
is 90 deg, determine the force P on the operating arm
required to set the brake for counterclockwise rota¬
tion of the wheel, which is 12 in. in diameter. The
torque on the wheel is 667 lb-ft. (6) What type and
size of chain would you select?
17-16. The arrangement of a transmission brake is

shown in the sketch. The arms are pivoted at 0 and,
when force is applied at the end of the hand lever,
the screw AB will rotate. The left- and right-hand
threads working in nuts on the ends of the arms will
move the arms together and thus apply the brake.
The force on the hand lever is applied 15 in. from
the axis of the screw. The drum is 7 in. in diameter,
and the angle subtended by each block is 90 deg.
The screw has six square threads with a mean diam¬
eter of % in. and a lead of 234 in.
Assuming a coefficient of friction for the braking
surfaces that is equal to 0.30 and for the threads
equal to 0.15, determine the force on the hand lever
required to set the brake when the torque on the
drum is 150 lb-ft.
Fig. P 17-16.
17-17. Determine the horsepower that the block

brake shown in the sketch will absorb when the drum
is rotating clockwise at 60 rpm and the operating
force S is 100 lb. The coefficient of friction for the
materials in contact is 0.3 and the angle of contact
for each shoe is 90 deg.
444 MACHINE DESIGN
17-18. The sketch shows the arrangement and
dimensions for the preliminary design of a double
block brake to be set by a spring providing a force
S on the bell crank a. The brake is to have a
torque rating of 2,400 ft-lb at 250 rpm. Assuming
that the angle of contact for each shoe is 120 deg
and that the coefficient of friction for the materials
in contact is 0.35, determine the following: (a) the
direction of rotation that requires the largest spring
force for the rated torque and the value of the
spring force S for that direction; (6) the width of
shoes required, assuming pV = 40,000 ft-lb per
min per sq in. of projected area; (c) the ratio of shoe width to drum diameter. Is
this a satisfactory ratio?
17-19. A 5-hp 850-rpm a-c motor is to be fitted with a
centrifugal starting clutch. This type of clutch as shown
schematically by the sketch is composed of the arms a,
connected to the rotating armature of the motor, and two
rotating shoes b. These shoes, which are identical, are
faced with friction lining and they move outward by cen¬
trifugal force when the arm rotates, and, when a certain
speed is reached, the force between the shoes and the
ring c will produce sufficient friction to rotate the ring c
against a resisting torque.
Determine the weight of each shoe required for picking
up the load at the following conditions:
Torque at load pickup. 80 per cent of torque at rated hp and speed

Speed of motor at pickup. 600 rpm
Coefficient of friction. 0.3
Inner diameter of ring, d. 8 in.
Radius to center of gravity of shoe, r... 3 in.
Arc of contact of each shoe. 90 deg
17-20. A slip coupling, as shown in Fig. 16-11, is designed to slip when the load on
the driven member becomes excessive. The pressure on both faces of the flange, which
is part of the hub a, is produced by a number of springs, such as the one shown at c.
Using the data given below, determine the deflection with which each spring must be
set up in order that the coupling will slip at a torque corresponding to 20 per cent
overload on a 250-hp motor running at 300 rpm. Use the equation based on the
assumption that the normal wear is proportional to the work of friction.
Outer diameter of friction disks. 31 in.

Inner diameter of friction disks. 19 in.
Number of bolts and springs. 12
Coefficient of friction (assumed). 0.14
Spring rate for each spring per inch. 2,500 lb
17-21. The backstop shown schematically in Fig. P 17-21 is used to prevent back¬
ward rotation of the shaft. A sector is pivoted at 0, and one end of the band is
attached to it and operates at a radius r2 = 2)4 in. The other end of the band is
attached at point a so that Oa = rq = 1 in. The diameter of the wheel is 834 in.,
the angle of wrap is 270 deg, and the width of the band is 2)4 in. The torque on the
wheel is 300 lb-ft. Assuming a coefficient of friction between the band and wheel
equal to 0.2, determine the following: (a) the maximum band tension, (b) the maxi¬
mum pressure between the band and wheel, and (c) whether the backstop is self¬
locking.
Fig. P 17-21.
17-22. For a band brake similar to that shown in Fig. 17-15, determine (a) the
torque in foot-pounds on the wheel for a force A on the operating lever equal to 20 lb
(assume that the wheel is 9 in. in diameter, a — 8 in., b = 10 in., angle of contact
6 = 210 deg, and the coefficient of friction/ = 0.2), and (b) the time in seconds required
to bring to rest a drum having WR2 equal to 225 lb-ft2 rotating initially at 200 rpm.
Chapter 18. Spur and Parallel Helical Gears
18-1. The layout of a drum hoist driven by an

1,160-rpm motor through double-reduction gears
is shown by the sketch. The speed reduction be¬
tween the motor shaft A and the intermediate
shaft B is 4:1, and between shaft B and the drum
shaft C is 5:1. The drum is 24 in. in diameter
and the maximum cable load is 5 tons.
a. Determine the torque in foot-pounds and
the speed in rpm for shafts A, B, and C, and enter
the results in the table below. The efficiency of
each gear reduction is 95 per cent. Fig. P 18-1.
446 MACHINE DESIGN
Shaft A B C
Torque, foot-pounds
Speed, rpm v
b. Complete the table below.

A to B (4 : 1) B to C (5 : 1)
Pinion Gear Pinion Gear
Diametral pitch 5
Number of teeth 24
Pitch diameter 4 30
c. What horsepower of motor would you recommend?

d. If a spring-set brake is used to control the load, on which shaft, A, B, or C, would
you attach the brake?
18-2. The two-speed hoisting winch shown by the sketch is to be rated at 3.5 tons
on the drum cable at 100 fpm with the low-speed gears engaged. Data regarding the
gears are shown in the table, (a) Copy the table, and enter values to complete it;
(6) determine the speed of the motor, assuming that the pitch diameter of the drum
is 21^ in.; (c) determine the horsepower of the motor required, assuming a 50 per cent
overload capacity and a 97 per cent mechanical efficiency for each pair of mating gears.
Gear a b c d e /
Number of teeth 15 75 32 13 71
Diametral pitch 23^ 2 2
Pitch diameter
18-3. (a) Make a neat sketch of a loaded gear tooth showing the tangential load W,
the inscribed parabola, and the dimensions of the tooth necessary for the derivation
of the Lewis equation. (6) Assuming that a gear tooth is equivalent to a cantilever
beam of uniform strength, derive an equation for the tangential load W in terms of
the face width F, the circular pitch p, maximum stress s, and the Lewis form factor,
(c) How is the allowable stress related to the static stress s0?
18-4. It is required to determine the proportions of a pair of spur gears to transmit
40 hp from a shaft running at 250 rpm to a parallel shaft with a speed reduction of 3:1.
The center-to-center distance of the shafts is 15 in. (a) Determine the pitch diam¬
eters of the pinion and the gear. (6) Assuming that the pinion is made of untreated
SAE 1030 steel, the gear is cast steel, the teeth are full-depth 20-deg involute teeth,
determine by the use of the Lewis equation the diametral pitch, the face width, and
the number of teeth for the gears.
18-5. Two parallel shafts are to be connected by means of spur gears to transmit
7.5 hp from 1,150 rpm of the high-speed shaft to the low-speed shaft with a reduction
in speed of 5:1. The following assumptions may be made:
Pinion. SAE 1045 (hardened, 21 teeth)

Gear. cast steel
Teeth. AGMA 20-deg involute stub
By the use of the Lewis equation, determine the diametral pitch, pitch diameters,
and face width.
18-6. A cargo winch is driven by a 35-hp

480-rpm motor through the medium of a single re¬
duction of spur gearing so that the speed of the
drum is approximately 57 rpm.
Using the Lewis equation, determine the diam¬
etral pitch, pitch diameters, and face width.
Assume the following:
Teeth. . . AGMA 20-deg stub

Pinion. SAE 1040, forged steel
Gear. . cast steel
Number of pinion teeth.. . 22
Maximum load on motor . . 150 per cent rating
18-7. A straight spur pinion made of SAE 1040 steel is to be keyed to the shaft of a
15-hp 1,140-rpm motor, and is to drive a gear made of ordinary cast iron. The speed
reduction is to be 6:1. Assuming 20-deg involute stub teeth and 24 teeth on the
pinion, determine the following: (a) which of the gears has the weaker teeth; (6) the
value of the constant in the equation for the induced stress, i.e., = ( )P3; (c) the
diametral pitch and pitch diameters; (d) the face width.
18-8. Using the data and results for Prob. 18-7, determine the following: (a) the
dynamic load on the teeth, assuming class 2 cut; (b) the maximum stress corresponding
to the dynamic load and the required Bhn for the gears, assuming steady load con¬
ditions; (c) the value of the load-stress factor and the required Bhn for the gears.
18-9. A 15-hp 1,140-rpm motor is to be connected to a drive shaft by means of a pair
of spur gears having a speed ratio of pinion to gear equal to 6. The gear teeth are to be
of 20-deg involute stub form. The pinion is to have 24 teeth made of SAE 1040 steel
and the gear of medium-grade cast iron, (a) Determine the diametral pitch, the pitch
448 MACHINE DESIGN
diameters, and face width of the gears by the use of the Lewis equation. (b) Assum¬
ing steady-load conditions of service, determine the class of gears, and state if the
cast iron as assumed will be of satisfactory grade, (c) Determine the Bhn required
for the gears. v
18-10. Values may be assigned or adopted for A, B, C, and D in this problem.
The following data apply to a single-reduction speed reducer having a pair of
straight spur gears:
Rated horsepower A_
Pinion speed. B_rpm
Velocity ratio C_to 1
Form of teeth.... 20-deg stub
a. Assuming SAE 2320 pinion, SAE 1040 untreated gear, and a minimum of
D_teeth on the pinion, determine by the use of the Lewis equation the following:
(1) diametral pitch of the teeth; (2) pitch diameters of the gears; (3) face width.
b. Assuming steady load conditions and class 2 gears, determine the Bhn required
for the pinion and gear on the basis of dynamic loading.
c. Determine the Bhn required for the pinion and gear on the basis of wear.
d. Enter recommended values for the speed reducer in the table below.
Diametral Pitch
Face BHN
pitch diameter
Pinion
Gear
18-11. Values may be assigned or adopted for A, B, C, D, E, and F in this problem.

The following data apply to a single-reduction speed reducer having a pair of
straight spur gears:
Rated horsepower. A_
Pinion speed. B_rpm
Velocity ratio. C_to 1
Form of teeth. 20-deg stub
(a) On the basis of 150 per cent rating and using SAE numbers for pinion as D_
and gear as E_and a minimum of F_pinion teeth, determine by the use of
the Lewis equation the following: (1) diametral pitch of the teeth, (2) pitch diameters
of pinion and gear, and (3) face width of the gears. (b) On the basis of 150 per cent
rating and for steady-load conditions and class 2 cut, determine the Bhn required for
the pinion and gear on the basis of dynamic loading, (c) On the basis of 100 per cent
rating, determine the Bhn required for the pinion and gear on the basis of wear,
(d) Enter recommended values for the speed reducer in a table similar to that given
below:
Diametral Pitch
Face BHN
pitch diameter
Pinion
Gear
18-12. A parallel helical gear 10 in. in diameter having 20-deg involute stub teeth
and a helix angle of 30 deg carries a torque of 3,000 lb-in. Determine the tangential
and radial components of the load on the teeth in the plane of rotation, the axial com¬
ponent of the load on the teeth, the angle of obliquity in a plane normal to the pitch
element of a tooth, and the resultant load on the teeth.
18-13. A 15-hp motor runs at 1,170 rpm and drives a ventilating fan through a
gear reduction having a velocity ratio of 4:1. The gears are to have 20-deg involute
stub teeth. The pinion is to be made of Micarta and the gear is to be made of cast
iron. Assuming 24 teeth on the pinion, determine the diametral pitch, pitch diam¬
eters, and face width of the gears using the Lewis equation.
18-14. The shaft of a wire-pulling machine rotates at 1,800 rpm and is required to
transmit a torque of 800 lb-in. from a gear located at one end of the shaft to a flexible
coupling at the other end. The bearings that support the shaft are 30 in. apart and
the gear is located 4 in. outside its adjacent bearing, center to center. The gear is
driven by a driving gear from a motor with a 1:1 velocity ratio, (a) Determine the
horsepower of motor to be specified. (6) Assuming that the gears have 18 teeth of
20-deg full-depth form and made of SAE 1035 steel, determine, by the use of the Lewis
equation, the diametral pitch, pitch diameters, and face width of the gears you would
recommend, (c) Using an allowable shearing stress of 5,000 psi, determine the shaft
diameter you would recommend.
18-16. A hotel-restaurant belt-conveyor system is to be driven by a 2-hp 870-rpm
motor through a gear train having a velocity ratio of 4:1. Assuming a starting over¬
load of 20 per cent, a Micarta pinion having at least 26 teeth, cast-iron gear, and 20-deg
full-depth teeth, determine by the use of the Lewis equation the diametral pitch, pitch
diameters, and face width you would recommend.
18-16. A reciprocating compressor system is to be driven by an 870-rpm motor
through a pair of straight spur gears. The compressor operates at approximately
220 rpm at a running torque of 1,300 lb-in. Assume a starting overload of 20 per cent,
20-deg full-depth teeth, 30 teeth on the pinion, and SAE 1045 and 1030 for the pinion
and gear, respectively, and determine your recommendations for the horsepower of the
motor, diametral pitch, pitch diameters, and face width for the gears.
18-17. In the line shaft shown in Fig.
P 18-17, a belt runs over the pulley A which r—'
is 36 in. in diameter and rotates at 320 B
rpm. The belt tensions are Fi = 1,500 lb
andF2 = 1,0001b. (a) Assuming the pin¬
M
ion B is to be made of SAE 1020 steel and
with eighteen 14^-deg teeth, determine
the diametral pitch, face width, and diam¬
eter required by the Lewis equation. (6)
4 7" 21"~ 5"
Fig. P 18-17.
Assuming an allowable shearing stress
equal to 6,000 psi, determine the diameter of shaft to be specified.
18-18. Refer to data for Prob. 14-11.
(a) Assuming that the motor pinion and
its mating gear have a face width equal
to 3^ in. and are made of SAE 1025
steel, would you regard the gears as satis¬
factory, based on the Lewis equation?
(b) Determine the diameter of the inter¬
mediate-speed shaft that is required (see
sketch Fig. P 18-18). (c) How would
you anchor the cable to the drum? Use
sketch.
450 MACHINE DESIGN
18-19. Assume that you are planning the design of a drive to connect the rotating
drum of a cement kiln to a motor. The drum is to rotate at approximately 20 rpm
and the speed of the motor is 1,160 rpm. Assume that it is reasonable to consider
V-belt drive, roller chains, and/or gears in any combination of any one, two, or all
three for the drive, (a) Make a neat sketch showing the layout of the drive you
would propose. (6) Specify the speed ratios, belt-sheave diameters, numbers of
sprocket teeth, and/or numbers of gear teeth in the combinations you propose in the
drive, (c) Assume that the loaded drum has a high WR2, how would you provide
for starting?
Chapter 19. Gears for Nonparallel Shafts
19-1. Make a neat sketch of a bevel-gear tooth having large pitch radius R, pitch-
cone element L, and face width F. Derive an equation for the tangential load W in
terms of the maximum stress s, circular pitch p, Lewis form factor y, and F and L.
19-2. Derive an equation for the resultant tooth load TFo and the radius at which
it acts.
19-3. Design the full-depth 14^-deg teeth of a pair of straight bevel gears to trans¬
mit 20 hp at 1,250 rpm of the pinion. The velocity ratio should be about 3.5 and the
pinion should have 18 teeth. Determine the diametral pitch, face width, pitch
diameters, and pitch-cone angles for the gears.
Chapter 20. Surface Finish, Friction, and Wear
20-1. For a roller supporting a load W and rolling on a surface, derive the equation
for the coefficient of rolling friction, a = Fr/W, where r is the roller radius and F is the
rolling force applied to the axis of the roller.
Note: Draw a free-body diagram for the roller. What are the units of the coeffi¬
cient of rolling friction?
20-2. A load of 4,000 lb rests on cast-iron rollers 8 in. in diameter that roll over a
wooden floor. The load is supported by steel rails that rest on the rollers. Assuming
the coefficients of rolling friction between the cast iron and steel and the cast iron and
wood as 0.02 and 0.10, respectively, determine the force required to move the load.
Note: Draw a free-body diagram for a roller. Determine the work expended for
each revolution of the rollers.
Chapter 21. Sliding Bearings and Lubrication
21-1. The following data apply to a lubricated bearing:
Load on bearing. 785 lb

Speed of journal. 360 rpm
Diameter of journal. 3 in.
Length of journal. 6 in.
Assume a clearance of 0.001 in. per in. of diameter, oil No. 9 in Fig. 21-16, oil tem¬
perature _deg F, and ambient temperature of 60 F. Determine the following:
(a) the actual value ZN/p] (b) coefficient of friction; (c) heat generated in foot-pounds
per minute; (d) heat dissipated in foot-pounds per minute; (e) assumed temperature to
give thermal equilibrium.
21-2. A 3-in. shaft rotates at 1,000 rpm. One of the bearings carries a load of 1,000
lb. (a) If the bearing is an average industrial journal bearing 4 in. long, lubricated
with SAE 40 oil, maintained at a film temperature of 160 F, and having a diametral
clearance of 0.003 in., determine the friction horsepower. (6) If the bearing is a ball
bearing, with a coefficient of friction of 0.002, and if it is assumed that the frictional
force acts at the radius of the shaft, determine the friction horsepower.
21-3. The following data apply to an average industrial unventilated journal bear¬
ing for a generator:
Load on journal. 500 1b

Speed of journal. 870 rpm
Diameter of journal. 1)4 in.
Length of journal. 2)4 in.
Assume a clearance ratio of 0.001, oil No. 6 (Fig. 21-16 in text), and a room tem¬
perature of 60 F. The maximum operating temperature to of the oil film is expected to
be 140 F. (Note that the value for absolute viscosity Z at 140 F has been determined
as 23.4 centipoises.) Determine whether fluid friction may be expected and whether
artificial cooling is necessary, and if so, the heat to be removed in units of foot-pounds
per minute.
21-4. A 15-hp 300-rpm centrifugal pump has two identical journal bearings, de¬
scribed below:
Journal diameter. 3 in.

Bearing length. 3)4 in.
Load on journal. 250 lb
Clearance ratio c/d. 0.001
Ambient temperature. 60 F
Average industrial unventilated bearing
Assuming oil No. 9 (Fig. 21-16), determine whether the bearings may be expected
to operate under conditions of thick-film lubrication. If so, determine the probable
operating temperature and the power loss for the two bearings expressed in per cent
of the total horsepower.
Note: To obtain the approximate equilibrium temperature, the heat-generation
and heat-dissipation curves may be assumed to be straight lines.
21- 5. The following data apply to an average industrial unventilated journal
bearing for a generator: load on bearing = 800 lb, speed = 870 rpm, diameter =
1)4 in., length = 2)4 in.
Assume a clearance ratio of 0.001 in. per in., oil No. 6 in Fig. 21-16, and a room
temperature of 60 F. The maximum operating temperature to of the oil film is ex¬
pected to be 140 F. The value for the absolute viscosity Z at 140 F has been deter¬
mined as 23.4 centipoises. Determine whether fluid lubrication may be expected
in the bearing, whether artificial cooling is necessary, and, if so, the heat to be removed
in units of foot-pounds per minute.
Chapter 22. Rollins-contact Bearings
22- 1. A bearing is to carry a radial load of 500 lb and a thrust of 300 lb. The
service imposes light shock and the bearing will be in use 40 hr per week for 3 years.
452 MACHINE DESIGN
The speed of the shaft is 1,000 rpm. Determine the size of medium-series ball bearing
that should be used.
Chapter 23. Metal Fits and Tolerances

\
23-1. A cast-steel crank having a hub whose outer diameter and length are 7 in.
and 6 in., respectively, is to be shrunk on the end of a steel shaft. The bore in the
crank is 4 in. and the shaft is 4.004 in. in diameter. Neglecting the effect of the crank
arm on the stresses, determine the following: (a) the radial pressure that is acting
between the shaft and hub; (6) the tangential stress that is induced in the crank;
(c) the temperature to which the crank must be heated in order that it may be slipped
on the shaft. Assume a room temperature of 70 F, a coefficient of expansion for steel
of 0.0000063, and a diametral clearance of 0.002 in.; (d) the maximum force required to
press the crank off the shaft and the corresponding work in foot-pounds. Assume a
coefficient of friction of 0.15.
Chapter 24. Vibration and Vibration Control
24-1. A 2-hp 1,200-rpm motor weighing 108 lb is mounted on four springs, each of
which has a spring constant of 1,000 lb per in. The motor is mounted so that it can
move only in a vertical direction. Determine the natural frequency of the system.
24-2. A 1,750-rpm motor and its support weighs 640 lb and is supported on four
springs, each having a spring rate of 15,000 lb per in. A damping device is used that
has a damping constant of 200 lb sec per in. The rotor has become unbalanced to the
extent that a periodic force for which P0 = 150 lb caused the motor to vibrate.
Assuming that the unbalanced force acts through the center of gravity of the motor
and that the motor vibrates only in the vertical direction, determine the following:
con, xo, and F for undamped and damped vibration, and enter the results in a table as
shown below.
Units Undamped Damped
CO rad per sec
COn rad per sec
Xo in.
F = kx0 lb
24-3. A 5-hp motor weighing 228 lb operates at 1,760 rpm. The motor is to be
mounted on four springs so that 90 per cent reduction in transmitted force (i.e., 10 per
cent transmissibility) will be secured. Assuming that the effect of damping at the
operating speed may be neglected, determine (a) the required natural frequency of the
system in cycles per minute, (6) the static deflection of the motor in inches, and (c) the
spring rate of each spring in pounds per inch. If an equivalent unbalanced weight of
1 oz is located centrally on the rotor at a radius of 8 in., determine (d) the maximum
force transmitted to the base at operating speed.
24-4. A motor which weighs 30 lb and rotates at 1,150 rpm has a vertical shaking
force equal to 5 lb due to an unbalanced rotor. The motor is mounted on two steel
cantilevers placed side by side of depth in. and width 3^ in. The distance between
the motor center line and the support is 24 in. Determine the maximum stress in
the cantilevers.
Chapter 26. Design-room Problems
26-1. Cotter joint. In Fig. P 26-1 is shown a common form of cotter joint used to
connect rods that are subjected to axial loads. This joint is rugged, is dependable,
may be easily connected and disconnected, and on reconnection the rods assume their
original alignment. The joint is used where ruggedness is essential and where the
bulkiness of the joint is not objectionable.
Specifications: 1. Materials. The materials of the rod end, the socket, and the
cotter and the allowable stresses (pounds per square inch) are as follows:
Rod end, Socket, Cotter,

mild steel wrought iron machinery steel
Tension or compression. 5,000 4,000 6,000

Shear. 2,500 2,000 4,000
Bearing. 10,000 10,000 12,000
2. Load. The joint is to be designed to support a tensile or compressive load

equal to_lb. (Assign 2,000, 3,000 or 4,000 lb.)
3. Assumptions. The thickness of the cotter may be taken as one-fourth the
diameter of the rod end. The taper of the cotter should be % in. per ft. Allow
clearances at the cotter equal to in., and clearance between the rod end and the
bottom of the hole in the socket equal to in.
454 MACHINE DESIGN
4. Procedure. Determine the dimensions as follows:
1. Steel rod, di 7. Socket in tension across slot, d4

2. Wrought-iron rod, d2 8. Socket and cotter in bearing, d5
3. (a) Rod end in tension, d3 9. Socket end in shear, b
(6) Rod end in bearing, d3 10. Socket in shear, c
4. Thickness of cotter, w 11. Collar in bearing, d6
5. Width of cotter, h 12. Collar in shear, t
6. Rod end in shear, a 13. Length of key (from sketch)
5. Sketch. Make a neat undimensioned freehand sketch approximately to scale

(full size) of a side view of the joint in section and a plan view showing the outside.
26-2. Flanged shaft coupling. Specifica¬

tions : This problem will consist of the design
of a two-piece, flanged shaft coupling of cast
iron for a given diameter of shaft. Starting
with the diameter of the shaft, the design
will be developed by working from the inside
outward.
1. Size. The diameter of the shaft is 2)4,
2)4, 2%, 3, 3)4, 3)4, 3%, or 4 in. (Assign
one.)
2. Proportions of hub and key. It is
impossible to proportion the hubs of gears,
pulleys, clutches, and couplings rationally.
They are invariably proportioned empiri¬
cally. The following equation is frequently
used in determining the outside diameter of
hubs of gears, pulleys, couplings, etc. (see
Fig. P 26-2).
Di = 13AD + )4
where Di = outside diameter of hub, in.

D = diameter of shaft, in.
The length of the hub is fixed by the length
of key. Select the size of key required and
specify its dimensions.
The key is to be made without taper, and a cup-pointed setscrew is to be used.
The diameter of the setscrew should be somewhat less than the width of the key. To
reduce hazard and to comply with safety requirements, a socket-head setscrew should
be used. Give specifications for the setscrew and the tap drill size for the setscrew
hole.
3. Coupling bolts. The number of bolts may be fixed empirically by the following
formula:
n = HD + 3
where n is the number of bolts.

So that each bolt may take its share of the load, finished bolts fitted to reamed holes
must be used. The size of the bolts will first be obtained by assuming that the shaft
and bolts are of the same material and that the bolts are as strong in direct shear as
the shaft in torsion, the friction between the halves of the coupling being neglected.
To determine the size of the bolts, it is necessary to have the radius R of the bolt
circle. This radius depends on the diameter of the head of the socket wrench, which,
in turn, depends on the unknown diameter of the bolts. The diameter Z)2 of the
head of the socket wrench in terms of the diameter d of the bolt is
Z)2 =: 1.85d +
and the radius of the bolt circle is
R = ^ ^ + clearance
The clearance between the head of the socket wrench and the hub may be taken
as %6 in. On substituting R in terms of d in the expression for the direct shear of
the bolts and placing this in terms of the torque on the shaft, the diameter of the bolts
may be determined by trial. Having the diameter of the bolts, the radius of the bolt
circle may be determined.
The bolts must also be of such size that they will not be overstressed when setting
up the nuts. The equation in Art. 9-6 may be used with a value of k = 10,000 to
12,000 lb per in.
It may be assumed that the bolts are of SAE 1025 to 1030 steel. The setting-up
stress may be assumed as three-fourths the yield strength. If the diameter found
should exceed that obtained by the first method, the radius of the bolt circle must be
altered accordingly.
Before reaming, a hole is drilled undersize to provide a small amount of material
for sizing and hole alignment. Specify the drill size, reamer size, and spot-face diam¬
eter. In order that the holes may register perfectly, they should be reamed with the
halves of the coupling assembled.
4. Coupling flanges. For good casting design, the change in thickness from the
hub to the radial flange should not be too great. A satisfactory finished thickness £2
for the radial flange in terms of thickness £1 of the hub is
For safety in operation and to comply with the safety code, the halves of the
coupling are to be provided with circumferential flanges to function as guards for the
boltheads and nuts. The thickness of the circumferential flange should be not less
than one-half that of the radial flange. The inside diameter of the circumferential
flange should be determined so that the boltholes are located midway between the
flange and the hub. Determine the outside diameter of the coupling.
To provide for automatic alignment, one-half of the coupling is to be counterbored
and the other turned to fit. The depth of the counterbore should be sufficient for the
machinist to establish a fit without difficulty, and therefore should not be less than
%6 in. The height of the male boss should be 3di2 in. more than the depth of the
counterbore to ensure contact of the faces. Provision should be made for “ one-
position” assembly of the coupling halves.
Before beginning the drawing, the following dimensions should be computed:
(a) spot-face thickness, (b) length of bolts, and (c) length of thread.
5. Drawing. Title: “(Size) Flanged Shaft Coupling.” A complete detail of the
halves of the coupling is to be made on 17 X 22 in. sheet. A bill of material is to be
placed above the title.
456 MACHINE DESIGN
26-3. Bench-model press. Hand-operated presses of this type are used for general-
purpose work, including pressing for force fits of gears, pulleys, sprockets, couplings,
bushings, and dowels and also for straightening or bending shafts, rods, and tubes.
The press should be rugged and trouble free.
Specifications:
Maximum load = 2^ tons

Clearance, ram to base plate, maximum = 20 in., minimum = 6 in.
Throat = 10 in.
Diameter of handwheel = 10 to 18 in.
Maximum tangential force on handwheel at % load = 60 lb (30 lb for each hand)
Materials are as follows: frame, cast iron; screw, SAE 1030 steel; bushing, bronze:
base plate, 1015 steel.
CT D
Procedure. 1. Screw, (a) Determine the root area of the screw, select type of
thread, and determine its size. A compressive stress of 10,000 psi may be used for
the screw, (b) Design the collar, using removable collet, (c) Check the force on
the handwheel. The coefficient of thread and collar friction may be assumed as 0.1.
2. Determine the length of the nut, considering an allowable bearing pressure of
1,500 psi and a shearing stress for bronze of 3,000 psi. Provide for lubrication.
3. Make a detail sketch of the head including the screw, bushing, collet, and
handwheel.
4. Frame, (a) It will be sufficient to calculate the dimensions of the frame at
section AA in the figure. Assume dimensions for the frame at this section, and check
the induced tensile and compressive stresses against allowable values in tension of
8,000 psi and in compression 32,000 psi. (6) Determine the size of bolts required at
B to hold the frame against the base plate. For rebound, use two small bolts at the
back of frame.
5. Base plate. Determine the thickness required for bending and provide for
hold-down bolts.
6. Sketch. Make a neat, undimensioned, freehand sketch to scale of the assembly
on an 8}^ X 11 in. sheet, showing a side view and a front view of the press.
7. Determine the approximate weight of the press.
26-4. Capstan. A motor-driven capstan, as shown in Fig. P 26-4, is composed of a
drum driven by a motor through a worm drive. A rope is wound around the drum
so that when the drum rotates, a load Fi may be moved by the application of a com¬
paratively small force F2 produced manually on the free end of the rope. The con¬
nected end of the rope may be used to spot a railroad car, haul an anchor, dock a ship,
for dragging operations in the field, etc.
The operation of the rope around the drum is similar to a belt drive where contact
is several turns around the drum.
In the capstan shown, the maximum load FT is to be 3,000 lb and the manually
applied force F2 is limited to 40 lb.
a. Assuming a coefficient of friction for the rope against the drum equal to 0.3,
determine the number of turns or wraps of the rope required.
b. Assuming that the capstan is rated at a load of 2,500 lb at a velocity of 50 fpm,
determine the horsepower of motor to be specified. Assume a duty cycle.
c. If the diameter of the drum is 10 in. and it is desired to use an 1,160-rpm motor,
determine the over-all speed reduction between the motor and drum.
d. What approximate percentage of the work to move the load is provided by the
operator?
26-5. Gears and brake. A motor-driven, geared drum hoist, as shown in Fig.
P 26-5(a), has a spring-set, double-block brake attached to the motor shaft. The
arrangement of the brake and its dimensions are shown at (6).
458 MACHINE DESIGN
(«) (b)
Fig. P 26-5.
Data:
Rated load on hoisting cable.. 1,000 lb

Diameter of hoisting drum.... 12 in.
Number of teeth on pinion. .. . 24
Number of teeth on gear. 72
Form of teeth. 20-deg stub
Material of pinion. SAE 1030
Material for gear. Untreated 20
points C cast steel
Motor speed. 570 rpm
Brake-drum diameter. 10 in.
Angle of contact for each shoe. 120 deg
Coefficient of friction for brake 0.40

a. The size of motor required using class D, squirrel-cage induction motor (see
Fig. 25-1). For duty cycle, full load for 6 sec, rest for 1 min.
b. The diametral pitch, pitch diameters, and face width for the gears.
c. The spring force required to set the brake for a torque on the drum of 1 lb-in.
d. The spring force required to set the brake for rated load on the cable.
e. The width of brake shoes required for a value of 30,000 ft-lb per sq in. of
projected area.
26-6. Belt conveyor. A 20-in. belt conveyor in which the belt is supported by
2-in.-diameter steel rollers is 100 ft long and runs at 200 fpm of the belt. The con¬
veyor is horizontal except for an inclined run (20-ft horizontal projection) to raise
the belt from the loading level 2 ft above the floor to the 7-ft unloading level. At the
unloading end is a metal platform with a 45-deg baffle to unload the material sideways.
The capacity of the hoist is to be six thousand 100-lb bags or boxes per hour. The
friction for the rollers may be assumed as 10 per cent of the weight on the belt.
a. Make a sketch of the layout of the conveyor.
b. Assuming an 1,160-rpm induction motor and 24-in. belt sheaves, determine the
over-all speed reduction for the drive and indicate the various reductions, i.e., belt,
chain, gears, etc., and speed ratio of each.
c. Assuming the efficiency of each reduction is 95 per cent, determine the size of
motor to be recommended using class C squirrel-cage induction motor for continuous
rating.
26-7. Traveling stairs. In Fig. P 26-7 is shown a sketch of the elevation.
Fig. P 26-7.
Data:
Capacity = 7,500 persons per hour

Width of stairs = 42 in.
Speed of stairs — 125 fpm
Rise H = 20 ft
Angle of incline = 30 deg
Number of steps = 44 on incline and 2 at each landing
Tread = 11 in.
Diameter of sprocket = 36 in.
Speed reductions:
1. V belt (efficiency 97 per cent)
2. Helical gears (efficiency 98 per cent)
3. Spur gears (efficiency 97 per cent)
Brake is located on shaft for large V-belt pulley.
Assumptions:
Average weight of each passenger including luggage is 175 lb.

One person per step at full load
Coefficient of friction for stairs = 0.05
Motor speed = 1,170 rpm
Horsepowers
a. For raising passengers (potential energy)
b. For accelerating passengers (kinetic energy)
c. For friction
d. For motor selection at continuous rating
Speed reductions
e. Over-all, motor to sprocket
/. Recommended for belt and gears
Required torque capacity for brake or backstop in foot-pounds at 50 per cent
overload
460 MACHINE DESIGN
26-8. Single-reduction parallel helical-gear speed reducer. 1. Specifications. It
is required to design a single-reduction, parallel helical-gear speed reducer having a
3:1 speed ratio and capable of transmitting the full-load rating of a __-hp 1,150-rpm
motor. (Assign 10, 20, or.30.)
a. Gearing. The gears are to be of the single helical type and made of the following
materials:
Pinion. SAE 3140, chrome-nickel steel

Gear. SAE 1040, mild steel
b. Shafting. The shafts are to be made of SAE 1045 forged steel in the as-rolled
condition.
c. Bearings. The bearings must be of the rolling-contact type and must be pro¬
tected against dust and leakage of oil.
d. Lubrication. Adequate lubrication of the gears and bearings must be provided
by a simple and positive splash system which will circulate the oil from a reservoir
located in the housing.
e. Housing. The housing is to be made of a good grade of cast iron and designed
to enclose all moving parts properly. The upper part of the housing must be remov¬
able without disturbing the bearing alignment, and provision must be made for
interchanging the shaft extensions from left to right, or vice versa. A suitable cover
plate must be provided to facilitate casual inspection of the gears and for replenishing
the lubricant in the reservoir. An air vent in the top of the housing and an oil drain
and level indicator at the bottom must be provided.
f. General. In designing the speed reducer, the following factors should be given
consideration: rigidity, weight, accessibility, compactness, appearance, and cost of
manufacture. The problem will consist of a set of computations and a 17- by 22-in.
drawing.
2. Preliminary sketch. Designers find that their thinking is greatly stimulated
if when beginning a new design a freehand sketch is made of the proposed machine.
The sketch will frequently indicate the procedure to be followed in making the compu¬
tations as well as provide a place to record any ideas that may come to mind. Many
helpful suggestions may be secured by studying the literature of manufacturing
companies of similar equipment. With the aid of any available literature and any
original ideas, draw a neat freehand sketch of your proposed design.
3. Gears. The following data may be assumed:
Helix angle. 23 deg

Form of teeth. 20-deg involute stub
Minimum number
of pinion teeth. 24
Face width. 3.5 times circular pitch or AGMA
minimum, whichever is greater
Electric motors will deliver under starting conditions around 200 per cent rated
torque. In the design of the gears for strength the overload torque may be used.
For wear of the gears, the rated torque may be used because for the major portion
of the operating life the motor operates at rated conditions rather than at starting
conditions.
a. Determine the diametral pitch, pitch diameters, and face width, using the Lewis
equation.
b. Assuming steady load conditions, determine the minimum required Brinell
hardness numbers for pinion and gear.
c. Determine the load-stress factor and the minimum surface hardness for the
pinion and gear based on wear of the teeth.
d. Specify the core and surface hardness of pinion and gear and required heat
treatment.
e. Check the AGMA thermal rating for the reducer.
4. High-speed shaft and bearings. The shaft and bearings are subjected to thrust
and radial loads caused by the gears, to overhanging loads caused by the driven
pulleys or gears, and to torsion due to the power being transmitted.
It is reasonable to make the tentative assumption that the maximum bending
moment on the shaft due to the overhanging load is 150 per cent of the maximum
torque on the shaft.
The bearings should be selected with an expected life of 5 years operating 7 days
each week at 10 hr per day.
a. Make a neat sketch of the shaft showing the gear, bearing mountings, bearing
cover, and oil seals.
b. Compute the minimum size of the shaft at the bearing adjacent to the shaft
extension.
c. Tentatively select bearings.
d. Determine the bearing span. Allow generous clearance between the pinion and
the housing, and locate the bearing so that it is fully supported in the housing.
e. Compute the length of the shaft extension measured from the bearing centers,
allowing 2 in. for the cover, seals, etc., and record all known dimensions on the sketch.
/. Compute the tangential, radial, and axial forces exerted by the pinion and the
overhanging load.
g. Determine the force on the most heavily loaded bearing due to the pinion tooth
loads and the overhanging load. Assume that the reaction due to the latter is in the
same direction as the resultant reaction due to pinion tooth loads.
h. Determine the required bearing capacity, and check whether the tentative bear¬
ing selection is satisfactory.
i. Make a sketch of the bearing showing its principal dimensions.
j. Select an oil seal and record the catalogue number, manufacturer, and its prin¬
cipal dimensions.
k. Assume the bore of the pinion to be the nearest nominal dimension above
the bearing bore, determine the size of key.
5. Low-speed shaft and bearings. Because the torque and bending moments vary
inversely as the speed, the low-speed shaft may be designed by simple proportion.
Develop the equation for this proportion in terms of the speed ratio, and determine
the minimum shaft size. Follow the same procedure for the low-speed shaft, as out¬
lined for the high-speed shaft.
6. Housing, a. Design the housing, observing the various factors and limitations
referred to in the specifications.
6. Suggestions. Thickness of housing wall % to % in. Provide ample clearance
inside of housing for all moving parts, not less than % in. around the gears. Provide
suitable floor flanges. Provide suitable stiffening ribs on the housing walls. Use
3^-in. to %-in. screws for fastening housing top. Use %-in. to ^-in. screws for
fastening bearing retainers.
7. Drawing. Make a one-half size undimensioned assembly drawing of the speed
reducer. Show two views, one side view and a top view. The top view should show
the cover removed with the gears and one bearing and mounting on each shaft in
section. The title of the drawing should be
462 MACHINE DESIGN
_Horsepower
Single-reduction Speed Reducer
Assembly
26-9. Shaft deflection. Lay out the shaft and loading force vectors for the shaft
shown in Fig. P 26-9 and, by graphical methods, determine the following: (a) bending-
moment diagram, (6) M/EI curve, (c) dy/dx curve, (d) y/x curve, (e) maximum
deflection, (/) critical speed, and (g) deflection within the bearings.
Fig. P 26-9.
AUTHOR INDEX
Albert, C. D., 18, 208, 226, 268 Feng, I-Ming, 29

Allan, R. K., 182, 339 Frocht, M. M., 46
Almen, J. 0., 133, 163 Frye, J. H., 42
Archbutt, L., 310, 315
Atwood, G. H., 175
Gaillard, J., 355
Gemant, A., 309
Barr, J. H., 131, 310 Glassco, R. B., 44
Barton, M. V., 41 Goodier, J. N., 19, 31, 32, 124, 125, 131,
Baud, R. V., 35 169
Berlin, D. R., 98 Gough, H. J., 29
Bevin, T., 370 Griffith, W. M., 158
Bolz, R. W., 353
Bowden, F. P., 306, 309
Bradford, L. J., 233 Hall, G. S., 132
Brewer, G. A., 44 Ham, C. W., 178
Hartenberg, R. S., 43, 44
Briggs, C. W., 120
Broghamer, E. I., 269 Hartman, J. B., 109, 160, 247
Haushalter, F. L., 165
Bruce, C. S., 308
Henwood, P. E., 132
Bryan, G. H., 69
Buckingham, E., 3, 262, 269, 354 Hersey, M. D., 321, 333
Hertz, H., 339
Burwell, J. F., 309, 334
Hesse, H. C., 118
Hetenyi, M., 43, 172
Chase, H., 3 Heydenkampf, G. S. von, 72
Clark, D. S., 79, 85 Hinkle, R. T., 59
Clark, H. H., 159 Hollister, S. C., 115
Clark, P. C., 164 Horenburger, W. E., 125
Clayton, D., 329 Horger, O. J., 42, 74
Colbert, T. P., 289 Howarth, H. A. S., 322
Cook, W. J., 164 Hoyt, S. L., 85
Hunsaker, J. C., 322
Hunter, M. C., 311
Davis, F. W., 183
Deeley, R. M., 310, 315
Den Hartog, J. P., 138 Illmer, L., 331
Dolan, T. J., 27, 44, 269 Inglis, C. E., 33
Drucker, D. C., 212
Duck, T. J., 308
Jacobsen, L. S., 33
Dudley, D. W., 269
James, R. V., 44
Durland, M. A., 408
Jennings, C. H., 110
Edgerton, C. T., 157

Kalikow, I., 347
Karpov, A. V., 39
Faires, V. M., 291 Kehl, G. L., 42
Faupel, J. H., 357 Kelvin, Lord, 82
463
464 MACHINE DESIGN
Kerr, J., 352 Sadowski, M. A., 46

Kimball, D. S., 310 Samans, C. H., 85
Kingsbury, A., 322 Sasso, J., 352
Kommers, J. B., 42, 55 Saunders, H. F., 70, 169
Krouse, G. N., 40, 54 Scarborough, J. B., 7
Kuhn, J. J., Sr., 2 Schmitter, W. P., 268
Seely, F. B., 17, 18, 44
Shaw, M. C., 311, 324
Langmuir, I., 306 Smith, G. V., 85
Lasche, O., 333 Smith, J. F. D., 165
Laszlo, A., 163 Smith, J. 0., 18, 55
Lee, G. H., 43 Soderberg, C. R., 59
Lewis, W., 254 Southwell, R. V., 69
Speller, F. N., 42
Spotts, M. F., 12, 357
Mabie, H. H., 226 Sterne, B., 161
McAdam, D. J., Jr., 166 Straub, F. G., 28
McClellan, A., 29 Stribeck, R., 339
McCorkle, I. B., 42 Sturm, R. G., 169
MacCullough, G. H., 66 Sutton, A. D., 257
McKee, S. A., 324 Sweeney, R. J., 124, 125
McKee, T. R., 324 Swigert, A. M., 299
Macks, E. F., 311, 324
Maleev, V. L., 109, 160, 247
Maulbetsch, J. L., 172 Tabor, D., 306, 309
Maurer, E. R., 67 Tachau, H., 212
Merritt, H. E., 259, 291 Thorpe, P. L., 29
Mikelson, W., 304 Tierney, W. D., 29
Mindlin, R. D., 46 Timoshenko, S., 12, 19, 32, 45, 66-68,
Mises, R. von, 49 169, 357
Moore, H. F., 40, 54, 55, 129, 132 Tomlinson, G. A., 29
Mumma, P. F., 42
Murray, W. M., 14, 54
Uhlig, H. H., 29
Nadai, A., 46, 49

Nordenholt, G. F., 352 Van Doren, H., 121
Norton, A. E., 322
Varney, W. R., 79, 85
von Heydenkampf, G. S., 72
von Mises, R., 49
O’Neill, H., 306
Votta, F. A., 164
Palmgren, A., 339

Parker, E. R., 72 Wahl, A. M., 41, 44, 148, 149, 151, 153,
Pernick, A., 257 154
Peterson, R. E., 41, 44, 48, 392 Walker, H., 291
Poritsky, H., 257 Way, S., 28
Pumphrey, F. H., 376 Weibel, E. E., 40
Wescott, B. B., 27
White, A. E., 28
Reynolds, O., 321 Wilson, W. K., 375
Rightmire, B. G., 29, 322 Windenburg, D. F., 70
Roark, R. J., 43, 44, 66, 69 Withey, M. O., 67
Roche, P., 368 Woodson, T. J., 376
Rogers, F. S., 226
Rossman, P. F., 98, 145
Ryan, D. G., 178 Zimmerli, F. P., 42, 152, 166
SUBJECT INDEX
Abbreviations, 383 Bearings, journal, thermal equilibrium

Absolute errors, 6 332
Acme threads, 175 Kingsbury, 319
Aero threads, 41, 127, 129 lubrication of, 310
Air springs, 164 materials for, 327
Allowable stresses, 52-60 Michell, 319
in cyclic loading, 53 needle, 349
for gears, 257 oilless, 329
in rivets, 93 pivoted-segment, 319
in shafts, 188, 191 roller, 348
for springs, 152 angular, 350
in static loading, 52 compared with ball bearings, 338
in welding, 108, 110 seals for, 352
Allowance, 354 Timken, 350
Arc welding, 105 ZNIp number for, 322
Axles, 184 Belleville springs, 163
Belt drives, American, 206
angle of contact in, 193
Ball-bearing screw, 182 initial tension in, 197
Ball bearings (see Bearings) materials for, 192
Band brakes, 243 power transmitted by, 196
Beam columns, 65 ratio of tensions in, 194
Bearings, ball, 339-348 Rockwood-type, 205
comparison with roller bearings, 338 service factors for, 200
dimensions of, 346 Belts, coefficient of friction for, 197
installation of, 347 creep in, 192
radial, 340 flat, 194
ratings of, 342 timing, 204
selection of, 343 V (see V belts)
series of, 341 Bending of shafts, 24, 56
service factor, 345 Bending moments, equations for, 385
thrust, 347 Bevel gears (see Gears)
types of, 340 Block brakes, 238
collar, 314 Block chains, 216
heat dissipation in, 333 Block clutches, 238
Hyatt, 349 Bolts, 122
journal, 319-336 effect of rigidity on, 132
bearing pressure in, 325 resilience of, 134
design of, 334 shock loading on, 133
friction in, 322, 323 tap, 123
heat generated in, 332 tightening-up loads on, 132
imperfect lubrication in, 330 (See also Fastenings)
lubrication of, 336 Brakes, 231-250
pressure in, 321-326 band, 243
stiffness of, 262 block, force analysis of, 238
strength of, 262 differential band, 244
465
466 MACHINE DESIGN
Brakes, disk, 232 Cylinders, of brittle material, 171
double-block, 241-248 Clavarino’s equation for, 171
energy in, 245 heads for, 172
heat dissipation of, 247 Lame’s equation for, 170
multidisk, 236 v stresses in, 170
single-block, 240 thick-walled, 170
torque resisted by, 240 thin-walled, 169
Buckingham equations for gears, 261, 262
Buckling, of columns, 61
of springs, 149, 150 Damping, in metal, 72
Buttress thread, 173 in vibration, 362, 364
Decimal equivalents, 413
Deflections, beam, equations for, 385
Cast iron, 74 by graphics, 409
Casting design, 88, 119 of shafts, 411
Castings, distortion in, 90 of springs, 148, 151, 160
effect on, of cooling rates, 88 Design stresses (see Allowable stresses)
of mass, 89 Diametral pitch, definition of, 352
fluidity in, 73 standard values for, 260
Caustic embrittlement, 27 Die casting, 118
Chain couplings, 226 Differential screw, 181
Chains, power (see Power chains) Disk springs, 163
tension linkage, 223 Ductility and stress concentration, 37
Clavarino’s equation, 171 Dynamic loads in gears, 262
Clutches, 231
band, 243
block, 238 Eccentric loading, 20
cone, 236 Efficiency, of hoists, 208, 209
plate, 231, 232 of riveted joints, 102
Columns, 61—68 of translation screws, 175, 177
beam, 65 of worm gearing, 293
buckling of, 61 Elastic stability of columns, 61
cast-iron, 64 Electric-resistance welding, 104
Euler’s formula, 61 Elements of sections, equations for, 384
factor of safety for, 65 Endurance limit, 54
J. B. Johnson equation for, 61 chart for, 388
stability of, elastic and plastic, 61 effect on, of finish, 388
Combined stresses, 23 of surface, 39
Computation errors, in addition, 7 of working, 42
in division, 8 for range of stress, 54
in multiplication, 8 Errors, absolute, 6
for powers, 9 computation (see Computation errors)
for roots, 9 relative, 6
in subtraction, 7, 8 Euler’s formula, 61
Computations, form of, 5
in revision, 9
Corrosion fatigue, 27 Factors of safety, 57
Cotter joint, 142 in columns, 65
Couplings (see Shaft couplings) for machine parts, 57, 58
Creep, in belts, 192 for wire ropes, 212
in metal, 85 Failure, in gears, 271
Critical speeds, 369 types of, 24
Curved beams, bending stresses in, 16 Fastenings, bolts, 122
stresses in, 18 cap screws, 123
Cyclic loading, 12, 26, 54 locking devices for, 124, 125
stress concentration in, 38 machine screws, 123, 124
stress-concentration factors in, 53 set screws, 123
SUBJECT INDEX 467
Fastenings, studs, 123 Gears, spur, dynamic loads on, 262
tap bolts, 123 form factors for, 408
threads for (see Screw threads) Lewis equation for, 253
Fatigue loading (cyclic loading), 26, 54 maximum bore in, 260
Feather keys, 134 strength of, 253
Finish (see Surface finish) stress concentration in, 269
Fits, allowance in, 354, 356 terminology for, 251
classes of, 355 velocity factors for, 258, 271
force, 357 wear in, 264
steel and cast-iron, 283 worm, 286-288
holding power of, 359 design of, 288
metal, 353 efficiency of, 293
shrink, 357 heat dissipation of, 292
assembly of, 359 Hindley, 287
tolerance in, 354, 356 mounting of, 293
Flange coupling, 224 strength of teeth in, 291
Flat keys, 136 terminology for, 286
Flat plates, 68 Gleason system of bevel gears, 284
Fluid coupling, 229 Graphical integration, 409
Fluid lubrication (see Lubrication)
Fretting, 29
Friction, clinging, 310 Hardness, 73, 80
coefficients of, 310 Hardness numbers conversion chart, 391
Coulomb, 310 Heat dissipation, in bearings, 333
effect of surface finish on, 305 in brakes, 247
fluid, 312 in worm gear boxes, 291
rolling, 311, 312 Heli-coil thread, 41, 129
solid, 309 Helical gears (see Gears)
Functional design, 3 Helical springs (see Springs)
Fusion welding, 104 Hoists, efficiency of, 208
Manila rope for, 209
relation of rope tensions in, 207
Galling, 308 safety factors for, 212
in gears, 271 wire rope for, 210
in screw threads, 129 Hooke’s coupling, 226
Gas welding, 105 Hydraulic couplings, 229
Gears, bevel, Gleason system of, 284 Hydrodynamic theory of lubrication, 320
mounting of, 284 Hypoid gears, 278
resultant tooth force in, 282
strength of, 279
terminology for, 279 Impact loads, 12
types of, 278 Imperfect lubrication, 330
failure of, 271 Induction motors, 377
galling in, 271
helical, crossed, 285
parallel, 266, 274 Johnson, J. B., formula, 61
velocity factor for, 268 Joints, cotter, 142
herringbone, 267 knuckle, 140
hypoid, 278 riveted (see Riveted joints)
lubrication of, 274 single-weld lap, 112
nonmetallic, 271 Jominy hardenability specification, 80
service factor, 277 Journal bearings (see Bearings)
silent, 271
spur, 251
allowable stresses for, 257 Keys, feather, 134
Buckingham equation for, 261 flat, 136
design of, 259 Kennedy, 135
468 MACHINE DESIGN
Keys, length of, 139 Motors, types of, 376, 377
proportions of, 138 WR2 in acceleration of, 381
round, 135 Mountings, rubber, 165, 373
square, 135
standard sizes of, 140
stresses in, 137, 138 Needle bearings, 349
Woodruff, 135 Nonmetallic gears, 271
Key ways, 136 Nonmetallic springs, 165
Kingsbury bearing, 319 Notch factor, 50
Knuckle joints, 140 Notch sensitivity, 50
Lame’s equation, 170

Offset links, 220
Lapping, 302
Oiliness, 329
Lay in surface finish, 297, 305 Oilless bearings, 329
Leaf springs (see Springs) Oldham coupling, 225
Lewis equation, 253
Overhanging shaft loads, 273
Loading, cyclic, 12, 26, 54
eccentric, 20
shock, 12, 133
Parallel helical gears, 266, 274
static, 12
Permissible stresses (see Allowable
allowable stresses in, 52
stresses)
stress concentration in, 36
Photoelastic method, 46
types of, 12
Pins, 143
Lock nuts, 124, 125
Pipe threads, 129
Lubrication, 313-331
Pitting, 28
boundary, 323, 330
of gear teeth, 272
converging film in, 315
Plastic stability of columns, 61
of gears, 274
Power chains, 213
hydrodynamic theory of, 320
block, 216
imperfect, 330
cases for, 221
mechanism of film, 315
inverted-tooth, 218
of rolling-contact bearings, 350
horsepower for, 219
(See also Bearings)
length of, 220
lubrication of, 221
offset links in, 220
Machinability, 74, 81
roller, 216, 217
Machine screws, 123
horsepower for, 218
Machine welding, 106
service factors for, 216
Machining, built-up edge in, 301, 302
silent, 218, 219
Manila rope, 209
speed of, 213
Materials, nonferrous, 76
sprocket bore for, 220
properties of, 387
Mechanical pitting, 28 Pressure, bearing, 325
Meehanite, 257 Production design, 3
Metal fits, 353 Profilograph, 297
Michell bearing, 319 Profilometer, 299
Mohr’s circle, 24 Properties of materials, 387
Motors, application of, 376
controls for, 378
dimensions of, 379 Range of stress, 55
duty cycle for, 380 in springs, 153
induction, 377 Relative errors, 6
rating of, 376 Repeated loads, 26, 54
intermittent, 380 Resilience, 72
selection of, 376, 379 Rigidity of shafts, 189
sizes of, 379 Ring springs, 162
SUBJECT INDEX 469
Riveted joints, allowable stresses in, 93 Service factors, for speed reducers, 275,
ASME Boiler Code for, 100 277
eccentrically loaded, 95 for V belts, 200
efficiency of, 102 Set screws, 124
failure of, 91, 100 Shaft couplings, 224
machine-member, 93 Ajax, 227
pressure vessel, 99 American, 225
structural, 93 Amerigear, 225
thin plates, 98 chain, 226
Riveting, dimpling in, 98 compression, 225
flush, 98 Falk, 227
thin plates, 98 flange, 224
Rivets, Cherry, 97 flexible disk, 229
draw, 97 fluid, 229
explosive, 98 Hooke’s, 226
stress concentration in, 92 leather-link, 228
Rolled threads, 130 Lord, 229
Roller bearings (see Bearings) Oldham, 225
Roller chains, 216-218 slip, 229
Rolling of metal, 118 Westinghouse-Nuttall, 228
Rolling resistance, 311 Shaft seals, 352
coefficients of, 312 Shafts, allowable stresses for, 188, 191
Rope, Manila, 209 bending of, 24, 56
wire (see Wire rope) in bending, 186
Round key, 135 and torsion, 186
Rubber mountings, 165, 373 deflection of, by graphics, 411
Rubber springs, 165 gear loads on, 272
rigidity of, 189
seals for, 351
SAE numbers for steel, 77 in torsion, 185
Safety factors (see Factors of safety) and compression, 186
Screw fastenings, types of, 123, 124 transmission, 190
Screw threads, acme, 173-175 Shock factor, 13
Aero, 127 Shock loads, effect of, on bolts, 133
American National, 126 Silent gears, 271
American Unified, 126 Sintering, 119
buttress, 173, 174 Snap rings, 143, 145
designation of, 126 Solid friction, 309
endurance limits for, 129 Speed reducers, 274
form of, 127 service factors for, 275, 277
galling in, 129 thermal rating of, 275
Heli-coil, 41, 129 Spiral springs, 162
proportions of nut for, 130 Splines, 135
rolled, 130 Spring rate, 149
Sellers, 174 Springs, 146
sizes of, 128 air, 164
stress-concentration factors for, 132 allowable stresses for, 152
thread forms for translation screws, 173 Belleville, 163
Unified Standard, 126 conical, 161
Screws (see Fastenings; Translation disk, 163
screws) helical, buckling of, 149, 150
Section modulus, equations for, 384 concentric, 160
values for, 386 deflection of, 148, 151, 160
Sensivity index, 50 design stresses for, 154
Service factors, for ball bearings, 345 end connections for, 156
for gears, 277 index for, 147
for power chains, 216 natural frequency of, 158
470 MACHINE DESIGN
Springs, helical, noncircular wire in, 151 Stress-concentration factors, determina¬

stresses in, 147 tion of, 43
surge in, 157 for gear teeth, 269
tension, 158 for screw threads, 132
torsional elastic limit, 153 in static loading, 53
leaf, 159 values for, 392-407
deflection of, 160 Stresscoat, 47
stresses in, 160 Stresses, allowable {see Allowable
material for, 151, 166 stresses)
Neg’ator, 163 combined, 23
nonmetallic, 164 in keys, 137, 138
range of stress in, 153 in springs, 147, 160
ring, 162 Structural welding, 106
rubber, 165 Struts, 62
shot peening of, 166 Superfinishing, 302
spiral, 162 Surface finish, cost of, 303
torsion, 162 damage to, 299
torsion-bar, 164 effect of, on friction, 305
volute, 161 effect of wear on, 305
Wahl factors for, 148 irregularities in, 297
Spur gears {see Gears) lay in, 297, 305
Square keys, 135 measurement of, 231
Steel, hardness conversion chart for, 391 roughness in, 303
impact resistance of, 87 roughness-width cutoff, 304
low temperature applications of, 86 specification of, 304
machinability of, 81
properties of, 77
SAE numbers for, 77 Thermit welding, 104
wrought, applications of, 75 Threads, pipe, 129
yield points for, 389, 390 screw {see Screw threads)
Stiffness of members, 56 Thrust bearings, collar, 314
Stress, range of, 55 Kingsbury, 319
in springs, 153 Michell, 319
Stress concentration, 31 Timing belt, 204
brittle coatings for, 47 Timken bearing, 350
causes of, 31 Tolerances, 354, 356
in cyclic loading, 38 Torrington bearing, 349
effect on, of holes, 32 Torsion, of rectangular bars, 19
of notches, 33 of shafts, 185
electric strain gage for, 46 Torsion-bar springs, 164
membrane analogy for, 45 Torsional vibration, 367
methods for, brittle-material, 43 Translation screws, ball bearing, 182
experimental, 43 bearing pressure in, 182
magniflux, 47 coefficient of friction in, 178
photoelastic, 46 collar friction in, 178
rubber-model, 44 differential, 181
short-gage-length, 44 efficiency of, 175, 177
soap-film, 45 stresses in, 180
reduction of, 40 thread forms for, 173
in rivets, 92 Transmissibility, 370
seriousness of, 36 Transmission shafting, 190
in spur gears, 269 Tubes, design of, 69
in static loading, 36 in external pressure, 69
in tension, 32
visualization of, 33
Stress-concentration factors, 47 V belts* 198
in cyclic loading, 53 horsepower rating of, 201
SUBJECT INDEX 471
V belts, service factors for, 200 Welding, fusion, 104
sheaves for, 200, 201 gas, 105
sizes of, 199 machine, 106
Vibration, 361 parallel welds, 109
control of, 372 precise calculations in, 110
critical speeds in, 369 stress relieving in, 115
dampers for, 373, 375 structural, 106
equation of motion for, 362 Thermit, 104
forced, 365, 366 Welding design, allowable loads, 108, 110
free, damped, 364 Welds, eccentrically loaded, 112
undamped, 363 inspection of, 116, 117
lateral, 367 moment of inertia for patterns in, 114
mountings for absorbing, 373 single-lap joints, 112
natural frequency in, 364, 367, 369 stress distribution in, 111
torsional, 367 stress flow in joints, 109
transmissibility in, 370 stresses in, primary shear, 112
Viscosity, chart for values, 330 secondary shear, 113
definition of, 329 testing in, 116
Volute springs, 161 Westinghouse-Nuttall coupling, 228
Wire rope, bending stresses in, 211
lay of, 210
Wahl factor for springs, 148 safety factors for, 212
Waviness of surface, 297 sizes of, 210
Wear, measurement of, 306 wire sizes, table of, 412
mechanism of, 305 Woodruff keys, 135
types of, 306 Worm gearing (see Gears)
Welding, 102-118 WR2, equations for, 407
arc, 105 in motor acceleration, 381
allowable stresses in, 108, 110
cost comparison, 103, 104
electric-resistance, 104 Zerol bevel gears, 278
fillet welds in, 109 ZN/p number, 322
■

Black 2nd Ed Machine Design

Uploaded by

Copyright:

Available Formats

Black 2nd Ed Machine Design

Uploaded by

Document Information

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

Black 2nd Ed Machine Design

Uploaded by

Copyright:

Available Formats

ACPL ITEM

McGRAW-HILL BOOK COMPANY, INC.

New York Toronto London

Copyright © 1955 by the McGraw-Hill Book Company, Inc.

Copyright, 1948, by the McGraw-Hill Book Company, Inc. Printed in the

Library of Congress Catalog Card Number 55-7268

THE MAPLE PRESS COMPANY, YORK, PA.

This book is intended to serve as a textbook for courses in general

development of the book, to Dr. R. T. Hinkle of Michigan State College

Appendix I. ASA Abbreviations.383

VIII. Yield Points for Steel.389

Questions and Problems.415

As one stands before any machine which is operating—such as a plastic¬

2-1 Importance of computations. Design procedure in engineering

5. All pertinent assumptions should be stated and references given.

A + B — C the correct result

The absolute error in the result is equal to

The relative error in the result is equal to

A — B = C the correct result

The absolute error in the result is equal to

The relative error in the result is equal to

The last result is greater than m. Therefore, in subtracting two quan¬

Multiplication and division of two numbers when one contains an error.

A X B = C the correct result

The absolute error in the result is equal to

The relative error in the result is equal to

Therefore, in multiplication as well as in division of quantities, a rela¬

Multiplication of two numbers when both contain errors.

A X B = C the correct result

The absolute error in the result is

where s is the allowable shearing stress. Assume that a diameter d has

LOADING, INDUCED STRESSES, AND FAILURE

3-1 The design problem. In arriving at the most satisfactory design

load and the displacement, or yP8. By equating this energy to that

y2P8 = W(h + 8) (3-1)

But 8 = P/C where C is a constant, lb per in., for a system. Substi¬

If Eq. (3-2) is written as

W= 1 + yj1 +2W (3'3)

By decreasing the area A, or by decreasing the modulus of elasticity E

This equation indicates that the effect of applying a load suddenly is

In case the time interval of application of the load is appreciable as

I bh2 0.375 X 0.52

The moment of inertia is

bh3 _ 0.375 X 0.53

From the equation for deflection of a cantilever,

The shock factor, Eq. (3-3), is

P ! i a /i , 2Ch , , . L , 2 X 204 X 0.1875 .

1 See W. M. Murray, Effects of Shock Loading, Product Eng., December, 1954, p.

4.33 X 6,150 = 26,600 psi

3-3 Tension and compression. Tensile stress. The average tensile

Fig. 3-2. Tension Fig. 3-3. Distribution of bearing

Since the actual distribution of pin pressure is difficult to determine

where P = load on the pin, lb

Fig. 3-4. Distribution of bending stress in a straight beam.

beam made of homogeneous material with the same modulus of elasticity

where s = stress at a distance y from the neutral axis, psi

of curvature of the beam, causing a nonlinear distribution of stress, as

where M = bending moment, lb-in., positive when it increases curvature

Fig. 3-5. Distribution of bending stress in a curved beam.

The value of Z may be found analytically or graphically. The deter¬