Principles of Direct Current Machines

Ij
ECTRICAL
NGINEERING TEXTS
PRINCIPLES
t
^S
OF
DIRECT CURRENT MACHINES
BY
ALEXANDER
S.
LANGSDORF, M. M. E.
PROFESSOR OF ELECTRICAL ENGINEERING AND DEAN OF THE SCHOOLS OF ENGINEERING AND ARCHITECTURE, WASHINGTON UNIVERSITY; FELLOW, AMERICAN INSTITUTE OF ELECTRICAL ENGINEERS
FIRST EDITION
McGRAW-HILL BOOK COMPANY, INC. 239 WEST 39TH STREET, NEW YORK
6
BOUVERIE STREET, LONDON,
E. C.
1915
COPYRIGHT, 1915, BY THE
McGRAW-HiLL BOOK COMPANY,
INC.
THE MAFZiE PRESS YORK PA
Dedicates
TO
MY MOTHER
SARAH SUSS LANGSDORF
321585
PREFACE
This book has been prepared with the object of placing before junior and senior students of electrical engineering a reasonably
complete treatment of the fundamental principles that underly

the design and operation of all types of direct-current machinery. Instead of attempting to touch the "high spots" in the whole
field of
trated
direct-current engineering, attention has been concenupon certain important features that are ordinarily dis-
missed with little more than passing mention, but which, in the. opinion of the author, are vital to a thorough grasp of the subject. For example, the book will be found to contain in Chapter III
derivation of the rules covering armature windings (followProfessor Arnold), in addition to the usual description of ing typical windings; Chapters VI and VII include a considerable
full
amount
istics
of
of
material concerning the operating charactergenerators and motors, the treatment being largely
new
for depicting the and in Chapters
graphical and including the use of three-dimensional diagrams mutual relationships among all of the variables
;
VIII and IX there has been developed a much more extensive treatment of the important subject of commutation than has been heretofore easily accessible to students of the type for whom the book is intended. In the selection and
arrangement of the material dealing with commutation, care
has been exercised to eliminate those minute details and excessive refinements that are more likely to confuse than to clarify.
Although the methods of the calculus have been freely used throughout the book, a conscious effort has been made to give special prominence to the physical concepts of which the equations are merely the short-hand expressions; to this end, the mathematical analysis has been preceded, wherever possible, by
a
and copiously illustrated discussion of the physical facts problem and their relations to one another. This has been done to counteract the tendency, manifested by many
full
of the
students, to look
upon a mathematical
-in itself,
an end complete
solution of a problem as apparently without a due realization

vii
viii
PREFACE
first essential is
a clearly thought out analysis of physical of this procedure, attention is directed to the new material of Article 210 of Chapter XI. The illustrative problems at the end of each of the first ten
that the
realities.
As an example
chapters have, for the most part, been designed to prevent the practice of feeding figures into one end of a formula and ex-
No attempt tracting the result (painlessly) from the other end. has been made to include as complete a set of problems as is desirable in studying the subject, for the reason that each instructor will naturally prepare a set to meet his own needs. Some of the problems at the end of Chapters VI and VII will be
found to tax the reasoning powers of the best students, but all of them have been successfully solved in the author's classes. Answers have not been given in the text, but will be supplied upon request to those instructors who ask for them. It is not to be expected that a new book on direct currents can avoid including much material common to the large number of existing texts on the subject. Such originality as has been brought to bear, aside from that represented by the new matter already referred to, has been exercised in selecting from the vast amount of available material those parts that seem most essential to an orderly presentation of the subject. Numerous wellknown texts have been freely drawn upon, with suitable ac-
knowledgment in all essential cases. That part of Chapter IV which deals with details of the calculation of the magnetization curve and of magnetic leakage, and the part of Chapter VIII in which the formulas for
armature inductance are developed, may be omitted without interfering with the continuity of treatment, in case design is taught as a separate course. In conclusion, the author desires to express his sincere thanks to Professor H. E. Clifford, of Harvard University, who made helpful criticisms and suggestions after reading the original manuscript, and who also assisted in the proof reading; and to the various manufacturers who have kindly contributed illustrations.
ALEXANDER
WASHINGTON UNIVERSITY, ST. Louis, Mo.
August, 1915.
S.
LANGSDORF.
CONTENTS
PAGE
PREFACE TABLE OF SYMBOLS
vii
xv
CHAPTER
GENERAL LAWS AND DEFINITIONS
1.
1
1
.
Introductory
2.
3.
4.
5.
Magnetic Field. Paramagnetic and Diamagnetic Substances Unit Magnet Pole -,*
'
2
.
.
Field Intensity
'.
..
Lines and Tubes of Force
3 5 6
6.
7. 8.
9.
10. 11.
12.
13.
14.
15.
Flux Issuing from a Magnet Pole Magnetic Potential Equipotential Lines and Surfaces * Induced Currents and E.M.F. Direction of Induced E.M.F Force Due to a Current in a Magnetic Field Unit Current. Unit Quantity Direction of the Force on a Conducting Wire Magnitude of Induced E.M.F
. .
8 8 9 10
11
12 12
Lenz's
Law
14
15 15 16 17 19
16. Practical
17.
Units of Current and E.M.F Heating Due to a Current
18. Field Intensity 19. Field Intensity
Due
to a Circular Coil
on the Axis of a Solenoid 20. Magnetic Potential on the Axis of a Circular Coil 21. General Expression for the Magnetic Potential Due to a Coil of any Shape at any Point 22. Magnetomotive Force
23. Permeability 24. The Law of the
20
21
Magnetic Circuit 25. Applications of Law of Magnetic Circuit

26. Kirchhoff's
Laws
27. Self-induction 28.
29.
Mutual Induction Energy Stored in a Magnetic Field

Electromagnets
30. Tractive Effort of
23 26 27 30 32 34 37 38
CHAPTER
THE DYNAMO
31.
II
Dynamo, Generator and Motor

ix
43 43
CONTENTS
PAGE
32. 33.
34.
35.
E.M.F. of Elementary Alternator Induced and Generated E.M.F General Case of the E.M.F. of an Alternator Rectification of an Alternating E.M.F
36. Effect of Distributed
37. 38. 39.
Winding Magnitude of E.M.F. Pulsations Average E.M.F. of an Armature Resistance of Armature Winding
40. Construction of
Dynamos
Machines
41. Bipolar and Multipolar 42. The Commutator 43.

44.
45.
The Armature Core The Pole Cores and Pole Shoes The Yoke
Ring
.
46. Brushes, Brush Holders and Rocker 47. Motor-generator. Dynamotor

48.
Turbo-generators
46 48 50 52 52 55 56 57 58 59 60 62 65 66 66 68 70
71
49. Interpole Machines 50. The Unipolar or Homopolar

51. Field Excitation of
Machine
Dynamos
52. Separate Excitation
53. Self-excitation
54. Series Excitation 55.
56.
........
...... ....
Windings
.
.
Shunt Excitation
Compound
Excitation
57. Construction of Field 58. Field
Rheostats
'I
59. Polarity of Generators 60. Direction of Rotation of
72 73 73 74 74 76 77 79 81 83
Motors
.....
III
85
CHAPTER
ARMATURE WINDINGS.
61.
.
...
62. 63.
64. 65.
66. 67.
Types of Armatures. Types of Windings Ring and Drum Windings Winding Element Ring, Lap and Wave Windings Number of Brush Sets Required Simplex and Multiplex Windings.
88 88 89 89 90
91
of
Degree
Reentrancy
...
.
68. General Considerations 69. 70.
Number
of Conductors, Elements and Commutator Segments Winding Pitch, Commutator Pitch and Slot Pitch
71. Field
Displacement
of
72.
Number
Armature Paths
73. General Rules
74. General
Rule for the Degree of Reentrancy
93 94 95 96 97 98 98 99 103
CONTENTS
75. 76.
xi
PAGE
Two-layer Windings Examples of Drum Windings 77. Equipotential Connections
104 107 107
CHAPTER
THE MAGNETIZATION CURVE.
78.
79.
IV
113 113 115 116
117 119
121
MAGNETIC LEAKAGE The Magnetization Curve

Experimental Determination of Magnetization Curve
80. Calculation of the Magnetization 81.
Curve
Magnetic Leakage
Magnetization Curve
82. Details of Calculation of

83. Correction to Pole
Arc
for the for the
84. Corrected Axial

85.
Ampere-turns 86. Ampere-turns 87. Ampere-turns 88. Ampere-turns

89.
Length Required Required Required Required
123
for
Teeth Armature Core the Pole Cores and Pole Shoes
124
128
.
.
.128
129 129
for the
Yoke
The
Coefficient of Dispersion
CHAPTER V
ARMATURE REACTION
90. 91.
92.
93.
133
Magnetizing Action of Armature
Commutation Components of Armature Reaction

Cross-magnetizing and Demagnetizing Ampere-turns Cross-magnetizing and Demagnetizing Effect in Multipolar
133 135 137 138

139
94.
Machines
95. Corrected
Expression for Demagnetizing Effect of Back Ampere'
turns
96. 97.
98.
99.
100. 101.
Shape of Magnetic Field Produced by Armature Current Approximate Distribution of the Resultant Field Demagnetizing Component of Cross Magnetization Excitation Required under Load Conditions Experimental Determination of Flux Distribution Potential Curve
143 145 147 148 149

152
102. Predetermination of Flux Distribution in the Air-gap
154 155
CHAPTER
103. Service
104.
VI
159 159 161
161
OPERATING CHARACTERISTICS OF GENERATORS

Requirements Characteristic Curves
105. Regulation 106. Characteristic
Curves of Separately Excited Generator 107. Effect of Speed of Rotation on the External Characteristic
108.
109.
162
.
. .
Load Characteristic The Armature Characteristic
165 166 167
xii
CONTENTS
PAGE
110. Characteristic
111.
Curves
of the Series
Generator
168
Dependence
of the
Form
of the Characteristic
upon Speed
'
.170
170
171 173
112. Condition for Stable Operation 113. Regulation for Constant Current
114. Characteristics of the
115.
Shunt Generator
.
.
116.
Dependence Dependence
of the of
Form of the Characteristic upon Speed .177 Form of Characteristic upon Resistance of Shunt
178
Analysis
of
Field Circuit
117.
Approximate
Mathematical
of the
Shunt
Generator
179
181
Characteristics
118. Characteristic
119.
The
Series
Curve Shunt
Compound Generator
120. 121.
Connection of Generators for Combined Output
The Thury System

.
122. Parallel Operation of Generators 123. Three-wire Generators. . . . . 124. Tirrill Regulator
184 185 185 186

191
195
CHAPTER
MOTORS
125. Service
126. 127.
VII
-.
.
Requirements Counter E.M.F., Torque and Power The Starting of Motors

. .
X
.
.
\
. .
199 199
'.
-.
199
128. Characteristics of the Separately Excited 129. Characteristics of the Shunt Motor
130. Characteristics of the Series
Motor
Motor
Compound Wound Motor
132.
Counter E.M.F.
The Reversing Motor
133. Starting of Differentially Wound Motors 134. Regulation of Speed of Shunt Motors 135. Applications of the Series Motor 136. Cycle of Operation of Railway Motors
..........
202 204 208 209 211 214 215 217

221 225 227 229 235
137. Series-parallel Control 138. Railway Controllers

139. Division of
Load between Motors
CHAPTER
COMMUTATION
140.
141.
VIII
237 237 239
241 243 245 246 248
Fundamental Considerations Physical Basis of the Theory
of
Commutation
142. General 143.
144.
145. 146. 147.
Equation, Case of Simple Ring Winding Elementary Mathematical Relations Discussion of the General Equation Modified Form of Sparking Criterion Linear Commutation The Current Density at a Commutator Segment.
General Case. 250
CONTENTS
148. Variation of Local Current Density at the
149. Further
150.
xiii
PAGE
Brush
Examples Simultaneous Commutation
of
Adjacent Coils
.
151. Successive Phases of Short-circuit in Coils of a Slot 152. Selective 153.
Commutation
in
Wave Windings
of Several Coils.
Duration of Short-circuit 154. Simultaneous Commutation Brushes
254 254 256 259 263 263
Effect of
Wide
.
. .
155. Calculation of the Self-inductance, L, in Slotted 156. Calculation of the Mutual Inductance,
Armatures
.
157. 158.
159.
The Commutating E.M.F Pulsations of Commutating
.,.
'*
.
.....
v
. .
.
.
Field
...
264 268 273 276 280

281 282 283
Sparking Constants 160. Reaction of Short-circuit Current upon Main Field 161. The Armature Flux Theory
CHAPTER IX
COMPENSATION OF ARMATURE REACTION AND IMPROVEMENT OF COM286 MUTATION
162. Principle of Compensation 163. Compensating Devices 164. 165. 166. 167. 168.
169.
..-'...
Commutating Devices Commutation in Machines having no Auxiliary Devices .... ..... Commutating Poles ,, . Winding of Commutating Poles Effect of Commutating Poles upon Coil Inductance Compounding Effect of Commutating Poles
.
:
286 287 290 292 297 298 300 301
CHAPTER X
EFFICIENCY, RATING AND HEATING 170. Sources of Loss
171.
172. 173.
The Ohmic Losses The Core Losses

Mechanical Losses
174. Additional Losses 175. 176.
Summary
True
of Losses
304 304 304 306 314 315 316
Efficiency, Efficiency of Conversion, Electrical
and Me317 318
chanical Efficiency
177.
The Stray Power Loss

of Efficiency with
178. Variation
Load.
Condition for
Maximum
321 324 325 326 327 332
Efficiency
179. Location of Point of 180. All-day Efficiency 181. Rating and Capacity
182. Allowable Operating Temperatures 183. Heating of Railway Motors
Maximum
Efficiency
xiv
CONTENTS
PAGE
184. 185.
186.
187. 188. 189.
Temperature Specifications of Electric Power Club Output Equation Heating and Cooling Curves Heating of the Armature
.
Heating of the Field Coils Heating of the Commutator 190. Rating of Enclosed Motors
334 334 336 340 343 345 345
CHAPTER XI
BOOSTERS AND BALANCERS.
191. Boosters
TRAIN LIGHTING SYSTEMS

. .
'.
..
192. 193. 194.
The Series Booster The Shunt Booster The Constant-current

.
..'...
or Non-reversible Booster
;......
195. Reversible Booster
...*.....
System
.^.
196. Auxiliary Control of Boosters 197. The Hubbard Counter E.M.F.
198.
199.
The Entz System The Bijur System

.
".
200. Balancers
........... ........... -357 358 ,.,/... \

.
347 347 347 348 350 353 354 354 355
-.
201. Train Lighting .. . 202. Voltage Regulation in Train Lighting Systems. 203. Resistance Regulation
204. Generator Field Regulation 205. Field and Line Regulation 206. Regulation by Means of Armature Reaction 207. The Rosenberg Train Lighting Generator
. .
208. Operation of Rosenberg Machine as a Motor 209. Modification of the Rosenberg Type of Generator
210.
The Wagner Auto mobile Lighting Generator

.
359 360 362 362 364 366 366 373 374 376 395
INDEX
TABLE OF SYMBOLS
(The figures refer to the page on which the symbol is first introduced; symbols formed from those given below by the mere addition of subscripts or primes for the purpose of distinguishing between quantities having the same general meaning, are not separately listed. Unless otherwise indiInch units are distinguished in the text cated, metric units are implied. by the use of the double prime ("). In a few instances the same symbol has been used to represent more than one quantity, though such cases are widely separated in the text, and the meaning may readily be determined from the lettering of the accompanying illustrations or from the context.)
A
PAGE
a constant length number of armature circuits constant in Froelich's equation

radiating surface a x cross-section of tube of flux
at
at a
at c
at g
at 8
att
at y
amp-turns amp-turns amp-turns amp-turns amp-turns amp-turns amp-turns
per cm per cm., armature core per cm., pole core per cm., air-gap per cm., pole shoes per cm., teeth
per cm., yoke
A Aa
Ab
area, sq. cm cross-section of
armature core
total area of brush contacts
A Ag
c
cross-section of pole cores cross-section of air-gap

cross-section of pole shoe
At cross-section of teeth under pole A y cross-section of yoke AT amp-turns per pair of poles ATa amp-turns required for armature core ATarm armature amp-turns ATC amp-turns required for pole core (A T) d demagnetizing amp-turns per pair of poles ATg amp-turns required for air-gap
39 57 179 340 156 117 116 116 116 116 116 116 4 116 315 116 116 116 116 116 116 116 300 116 138 116
XV
xvi
TABLE OF SYMBOLS
PAGE
A Ti amp-turns
c
required for two interpole cores
ATig amp-turns A Tn amp-turns ATa amp-turns A T amp-turns ATy amp-turns

t
required for two interpole air-gaps required for two sets of teeth opposite interpoles required for pole shoes required for teeth
,
required for yoke
300 300 300 116 116 116
B
b pole arc
119 179 241 119

131
.
b'
constant in Froelich's equation brush width corrected pole arc
'.
b c peripheral
b'i
width of pole core

.
corrected arc of commutating pole
......
299
122 125
ba
width of
slot
308 123 270 B flux density 24 B' amplitude of flux pulsations at pole face 313 Ba flux density in armature core 116 field intensity in axis of commutation due to armature current 280 Bc flux density in pole core 116 field intensity at pole tip due to armature cross-field 141 B g flux density in air-gap 116 299 Bi g flux density in air-gap due to interpoles
bv
bt width of tooth at tip b' t width of tooth at root
width b width
of ventilating duct of slot opening
flux density in teeth B't apparent flux density in teeth

t
B B
flux density in pole shoes

-.
-.
,,<.
.116
....
116 125 116
By
Bo
flux density in
field
yoke
intensity in axis of commutation, no-load
279
C
c
constant constant
'.
309 34
D
d diameter diameter of armature d c diameter of pole core dcom diameter of commutator
47 146 131 281

17
distance or length
TABLE OF SYMBOLS
xvii
E
PAGE
e e.m.f. in volts
commutating
e max
e.m.f. at beginning of period of
commutation
,
,
maximum
reactance voltage
er
drop in shunt field rheostat average e.m.f. of self-induction

e.m.f. in short-circuited coil e.m.f. in volts
es
c.
36 242 248 79 247 389

15
E E Ea
ECO
e.m.f. in abvolts
12
counter e.m.f Eb e.m.f. generated in EC commutating e.m.f

ci
field axis,
Rosenberg generator
....
.-"
>
-.
...
no-load
,,.
;
(
E E ET
t
commutating commutating
e.m.f., e.m.f.,
......
N
.
...
load
,
terminal or line voltage
.,
...
. .
commutating
e.m.f. at
end
of period of
commutation
77 369 242 279 280 77 245
F
f force in dynes field step in terms of single pole pitch frequency of magnetic reversals
coefficient of friction
.
.................
.....
.
force in dynes
3 100 306 315 39
H
h radial depth of armature core under teeth circular mils per ampere
128 335
131
h c radial length of pole core h s radial length of pole shoe
130
depth ht depth h' 2 depth h 3 depth h* depth

hi
of
armature
coil
of straight part of slot
above lower
upper
coil
coil
H H
e
of straight part of slot above of inclined part of tooth-tip of tooth overhang
269 269 270 270 270

3
18
field intensity
field intensity at
field intensity at
end
of solenoid
Ho
center of coil or solenoid
17
I
i
current in amperes current in line

current in short-circuited coil
total
31
ia
armature current
77 243 57
XVill
TABLE OF SYMBOLS
PAGE
(ia) ib if
ii
armature current, no-load exciting current, Rosenberg generator field current, separately excited machine
linear
319 369
162
component, short-circuit current
io
field current extra component of short-circuit current ia/a current per armature path / current in amperes 7 current in abamperes
is
ix
shunt
250 77 250 237

11
10
K
k constant
...'.
'
lamination factor
correction factor, pole face loss
constant
ratio of iron section to air section
under a pole
125 313 34 125
L
I
I'
length axial length of core corrected length of core

.
3
*
.
.
la
lc
length of magnetic path in armature total length of wire on armature length of magnetic circuit in pole core axial length of pole core
axial length of commutator total length of end-connections of element
.
119 119 116 305 116

131
Icom
I/
I'i la
It
corrected axial length of interpoles length of magnetic path in pole shoes

.
mean
length of turn length of magnetic path in teeth

length of magnetic path in yoke distance between adjacent pole shoes
distance between inner surfaces of pole cores self -inductance in henries
.
ly
li
lz
L Lb L
self-inductance of coil edge at bottom of slot self-inductance of shunt field winding
Lt self -inductance of coil edge at top of slot LI self-inductance of element due to slot leakage L 2 self-inductance of element due to tooth-tip leakage
....
.
self-inductance of element due to end-connection leakage
345 271 299 116 79 116 116 130 131 33 269 82 270 270 270 271
M
m
strength of magnet pole 3
TABLE OF SYMBOLS
field
xix
PAGE
... 98 displacement in terms of commutator segments mutual-induction ...... 36 '_ ....... M\z coefficient of mutual-induction between coils in adjacent slots. 275 Mi 3 coefficient of mutual-induction between coils separated by 276 one slot
coefficient of
...
..../
N
n revolutions per minute number of coil edges per element number of coil edges carrying reversed current in neutral zone .... n/ turns per pair of poles, separately excited machine "... turns per pair of poles, series machine na turns per pair of poles, shunt machine nsc number of coils simultaneously short-circuited by brush n v number of ventilating ducts n ideal no-load speed of motor
: . . .
. . .
47 98 144 162 168 79 280 125 206

14
number
of turns of coil or circuit
............
P
p constant distance
11
; -
number
of poles
sq. in.
p e brush pressure, Ibs. per P power permeance summation of losses
50 315
-
13 123
copper loss, commutator PC/ copper loss, field P C m,i constant part of power loss P ca eddy current loss, armature core
.
P bf P ca P cc
brush friction loss copper loss, armature

.
...
.
.
.
.-.;..'
..'..'.
.
.
...
.'
,.;,....
Pet
eddy current
loss,
teeth
Pha hysteresis
loss,
armature core
loss
.
PIU hysteresis loss, teeth
Ph
combined hysteresis and eddy current

loss in pole face stray power loss
Pp
PS Po rated output
317 315 305 306 305 322 310 312 306 308 322 313 319 324
Q
q degree of reentrancy
amp-conductors per cm.
of
armature periphery
.
103 146
Q Q
heat generated per sec., kg-cal quantity of electricity, abcoulombs
336
12
XX
TABLE OF SYMBOLS
R
PAGE
r distance,
radius
resistance of shunt around series field

r'
resistance of resistance of
resistance of
ra resistance of
77
r s resistance of
3 172 armature and starting rheostat 201 armature including brushes and brush contacts. 57 series field winding 79 162 separately excited field winding shunt field winding 77
:
R Ra R
c
resistance in
ohms
on armature
15
total resistance of all wire
57
Rb resistance of entire brush contact

resistance of short-circuited coil
...
Ri resistance of
commutator lead
244 244 244
S
s
distance
13
of
number
specific
sa
winding sections
coil section
diagonal of rectangular heat

of
55 271
cross-section of armature conductor
S number
commutator segments
336 305 96
T
t
time in seconds tooth pitch
13
tv
working temperature of armature, deg. C thickness of laminations distance center to center of ventilating ducts period of commutation, seconds
torque
122 305 309 123 136 201
V
v velocity
13
vc
peripheral velocity of armature peripheral velocity of commutator
281 264
6
energy, work, potential

velocity
volume
of core
47 306
Vt volume of tooth
308
W
w
watts radiated
.
340
TABLE OF SYMBOLS
work, energy ..................... ..... weight of core ............. W loss at brush contact ..... ........... WR loss in rheostat ..................
xxi
PAGE
W
c
12
306 250 228
X
x variable distance
.......... ..... amp-turns required for double air-gap, two sets of teeth and armature core ...................
'.
..
.'
129
Y
y commutator pitch
yi
2/ 2
back pitch
front pitch
................... ............ .... ..................

.........
Z
.
97 97 97
number
of
Z number
a 60
of
conductors per coil edge armature conductors

.
...........
.
.
;.
......
269 47
201 101
10
a
a.
angle angle of brush displacement from neutral

coefficient
.......... ....... ......... ......... of cooling ................. constant ...................

;
.
138
336 383
/3
angle subtended by pole arc
width
0'
of
commutator segment
of double angle of
constant
.............. ............. .......................

brush lead
.
supplement
........
119 241 383 139
specific resistance of core material
constant
.......................
5
.........
310 383
length of single air-gap
....... ..........
116
xxii
TABLE OF SYMBOLS
PAGE
5'
5'i
5X
corrected length of single air-gap ; 122 corrected length of single air-gap under commutating pole 301 length of tube of flux in air-gap 156 relative shift of segments with respect to brushes ...... 257
.
........
...
Ae brush drop
............. ..... ................

e
164
eddy current constant
.......
.....
312
77
hysteresis constant
efficiency
...........
......
..
.....
306 317
variable angle
rise of
-.
..."
.-
temperature, deg. C.
X
...'./ ...... ...........
331
X number of flux linkages A Carter sparking criterion
.... ....
. .
' .
.,
20 267
AI
permeability
...... ..... ...... ....

.
24
v coefficient of dispersion or
vi
leakage coefficient
.
......
coefficient of dispersion of interpoles
..........
118 299
output coefficient
.............
p
...*...
335
p specific resistance of copper.
.......
79
a intensity of magnetization
correction factor for fringing flux
............... ............
39 122
TABLE OF SYMBOLS
T
xxiii
PAGE
T pole pitch
......................

<P]_
3>i
................. inner surfaces of pole cores ......... p leakage ............ end-connection leakage flux lateral surfaces of pole cores ......... leakage flux ...... ................... $ flux ......................... useful flux per pole .......... ........ Rosenberg generator ............. armature Rosenberg generator .... ....... useful flux due to interpole ............... total flux due to interpole ..... .... ....... total flux per pole ..................
tooth-tip, leakage flux
flux,
3
.
f> 2
angle leakage flux per pole leakage flux, inner surfaces of pole shoes slot leakage flux leakage flux, lateral surfaces of pole shoes
........................ ..................
.................
.
.
.........
......
10 118 130 268 130
268
131
268
131 14 4
?4
flux,
>6
field flux,
>#
flux,
>;*
4>
47 369 369 299 299 118
ratio of pole arc to pole pitch
..........
...
142
solid angle
19
PRINCIPLES OF
DIRECT CURRENT MACHINES

CHAPTER
I

A clear conception of the theory underlying 1. Introductory. the design and the operating characteristics of electrical machinery depends upon a thorough understanding of a few fundamental physical facts concerning the properties of electricity and magnetism and of the formulation of these facts as laws or definitions. The object of this chapter is to present in condensed form those facts, laws and definitions which are immediately For a more applicable to the theory of direct-current machines. extended treatment of these basic principles the student is referred to the numerous texts in which the subject is treated in
detail.
Field. Paramagnetic and Diamagnetic SubThe space surrounding a magnet is called a magnetic The existence of a magnetic field is manifested by the field. measurable mechanical forces which act upon magnetic substances or upon electric currents in the field. In the case of a magnet of elongated form, the force due to it is greatest near its
2.
Magnetic
stances.
and these are referred to as the poles of the magnet. such a magnet is freely suspended at its center of gravity, it will turn until its polar axis coincides with the magnetic meridian at the point of support provided it is subjected to the influence of the
extremities,
If
earth's field only. The north-seeking end of the the north or positive pole, and the south-seeking
magnet
end
is
is
called
called the
south or negative pole. Experiment shows that if two such magnets are placed near each other poles of like sign will repel each other and poles of unlike sign will attract each other.
The magnetic
fields
produced by natural magnets

1
like the
PRINCIPLES OF DIRECT-CURRENT MACHINES

and by
artificially
lodestone,
made permanent magnets,
are not
sufficiently powerful for practical purposes except that permanent magnets are used in small magneto-generators and in some types
of measuring instruments. The powerful magnetic fields required in generators and motors are always produced by the magnetizing action of an electric current.
net
steel are attracted by a magwhen placed in the field of the latter. This property is also possessed by nickel and cobalt but to a less extent than in the
Experiment 'shows that iron and
case of iron. Other substances, of which bismuth is the most prominent example, are repelled by a magnet. Materials of the former class are called paramagnetic substances, those of the latter
class are called diamagnetic substances.
The
classification of
material as paramagnetic or diamagnetic depends upon the nature of the medium in which it is immersed. The medium used as the
standard of reference
is
air,
which
is
assumed to be neutral or
Since paramagnetic substances, iron and steel in particular, are used in practice to the virtual exclusion of all others, they are usually called simply magnetic
non-magnetic in
its
properties.
substances.
If
into a magnetic
a magnetic substance, originally unmagnetized, is brought field, it will become magnetized by induction in such
way that
will
the induced pole adjacent to the nearest inducing pole have a polarity opposite in sign to that of the inducing pole. 3. Unit Magnet Pole. Every magnetized body exhibits the
of polarity, that
is,
phenomenon
the simultaneous existence of
One polarity cannot exist without the poles of opposite sign. other. The magnetized condition obtains throughout the entire
mass
to point.
of the magnet, but its intensity generally varies from point In speaking of the pole of a magnet it should be under-
is no one point at which the magnetism is actually but the conception of concentrated point poles is concentrated, useful for purposes of computation even though the idea is In the case of a long, slim magnet, like a knitting artificial. the needle, magnetism acts as though it were mostly concentrated at or near the ends, so that such a magnet approximates fairly well the condition of concentrated point poles. In particular, if one
stood that there
pole of such a magnet is placed in a magnetic field, its other pole being so far removed as to be acted upon with little or no force, the
magnet will behave as though it consisted of a single isolated pole, and the forces acting upon it can then be studied. For purposes of quantitative measurement, a unit magnet pole is
defined as a point pole of such strength that it will exert a force of 1 at a distance of 1 cm., both poles being The force will be a repulsion if the two unit poles are of in air.
dyne upon an equal pole

the
be an attraction if they are of opposite sign. If a unit pole is placed 1 cm. away from a pole of unknown strength, the surrounding medium being air, and the force between them is found to be ra dynes, it is assumed that the second
same
sign;
it
will
In other words, the strength of a magnet pole is measured by the force in dynes with which it acts upon (or is acted upon by) a unit pole at a distance of 1 cm., and ra', respectively, in air. Two magnet poles of strength placed 1 cm. apart, will then act upon each other with a force
pole has a strength of
m units.
of
mm' dynes, in accordance with this definition. In 1800 Coulomb discovered the fact that the force of attrac-
tion or repulsion between two magnet poles is inversely proporIn general, tional to the square of the distance between them. a and m! separated by distance r the force between two poles
is
then
/-fc
If force is
?
and
a)
pole-
measured
in dynes, distance in centimeters,
strength in
r are all
terms of the unit defined above, then when m, m', and equal to unity, / is likewise unity, hence k = 1, or
/
4.
= mm' 3- dynes
intensity of a
(2)
Field Intensity.
The
magnetic
field
at a
given point is measured by the force in dynes which acts upon a unit magnet pole placed at that point. It is represented by the symbol H. A field is of unit intensity at a particular point when
it
acts
upon a unit magnet pole

This unit
at that point with a force of one
dyne.
field intensity (in air) is called
the gauss.
field intensity of
H gausses then means a field which will act upon a unit pole with a force of H dynes, or with a force of mH dynes upon a point pole of strength m units.
5.
Lines and Tubes of Force.

in a
If
a unit magnet pole

it will
is
moved
about
magnetic
field,
the force acting upon
in general

in
vary
magnitude and direction from point to

field
point.
At each
point in the
length
is
the force can be represented by a line whose proportional to the magnitude of the force and whose
If
direction coincides with that of the force.
curves are
now
their tangents are at each point in the direction of the force at that point, such curves are lines
drawn
in such a
manner that
Obviously there will be an infinite number of of magnetic force. such lines in any magnetic field, since there is an infinite number of points which do not lie on one and the same line of force. It is also clear that lines of force cannot intersect, for if they did, each of the intersecting curves would have a different tangent at
the point of intersection, therefore implying that a magnetic pole placed at that point would simultaneously experience more than
one force
The
a condition that is clearly impossible. positive direction of a field or of a line of force

free north pole
is
that in
which a
would move.
Although the number

field is infinite, it is
ized
is
by a
finite
of actual lines of force in any magnetic convenient to picture the field as characternumber of conventionalized lines of force. This
accomplished by representing a field of intensity H, in air, by lines per sq. cm. of cross-section taken at right angles to the In a uniform magnetic field the force is direction of the field. everywhere the same and in the same direction, in which case the total number of lines, or the magnetic flux, crossing an area A is
$ =
If
AH
field,
the area
is
not at right angles to the
the flux crossing
it is
where a
tion.
is
AH cos a (3) the angle between H and the normal to the plane sec$ =
field,
In the case of a non-uniform

'$
f H cos a.dA
H H
(4)
intensity at the element, normal to the element.
is the field a differential element of the surface, and the and a is the angle between The total flux across an area is usually expressed in lines; the International Electrical Congress of 1900 adopted the name maxwell to represent unit flux, but this is seldom
is
where dA
used.
GENERAL LAWK AND DEFINITIONS
bundle of
lines of force
threading through a given area will
converge or diverge as the field intensity increases or decreases, The outer lines of such a bundle constitute the respectively. elements of a tubular surface, Fig.
called a tube of force.
of force
is 1,
and the
entire bundle
is
The
total flux across all sections of a tube
the same, or
fH'cosa'.dA' = fH"cosa".dA"
for
of the
by hypothesis the longitudinal walls tube are made up of lines of force,

FIG.
1.
and since lines of force cannot intersect, no flux can cross the walls of the tube.
the walls of a tube of force
6.
is
Tube
of force.
This fact can be otherwise stated by saying that the flux across
zero.
is
Flux Issuing from a Magnet Pole. If a unit magnet pole placed at a distance of r cm. from a pole of strength m units, will be repelled with a force of
it
f
This
is
= m dynes
i
equivalent to saying that the

r
field intensity at
a distance of
cm. from a pole of
m m
units
is
-^
gausses.
Since the locus of
all points distant r cm. from a sphere having as center and a radius r, as in Fig. 2, the field intensity will be the same at all points on its surface, and the direction of
is
Fio. 2.
intensity
Field
the
field will
around
magnet
pole.
that point. the gphere
be at each point along the radius to The total flux across the surface of j g then
TYl
$ =
The
1
AH
4?rr
-3
47rm lines
is,
(>or
maxwells)
4?r lines.
1
(5)
flux
from a unit magnet pole

4?r
therefore,
many of the equations that apply to magnetic presence in certain of these equations has been objected to by some scientists because of its incommensurable nature, and attempts to eliminate it from these equations have been proposed. This can be done by suitable changes in the definitions of some of the funrecurs in
Its
The quantity
electric
and
problems.
damental
1904.)
units.
is
It
(Trans. International Electrical Congress, Vol. I, p. 130, interesting, however, to note that'47r appears in the results
6
7.

Magnetic Potential. Let Fig. 3 represent two magnet poles m and m units, respectively, separated by a distance
r
of strengths
777 777
x cm.
Each
will repel
the other with a force of
3-
dynes.
Let
this force,
one of the poles, as m', move a distance dx under the influence of then work will be done to the extent of
- dx dV = mm'. ergs 2
.
dx
FIG.
3.
Determination of potential energy of two magnet poles.
The
an
entire
amount
of the
infinite distance,
work done in separating the two poles from an initial separation of r cm., is
r
,
to
V= Cmm ~^dx
j
r
(6)
Since no work has been done upon the system by any outside agency during this process, the work represented by the expression
must have come from the system
itself,
and therefore
represents the stored or potential energy of the two poles in each other's presence. It represents also the amount of work or
energy required to bring one pole from an infinite distance into the presence of the other, with a separation of r cm. for this is given
;
by
or the
same
as before.
If
m'
171
1,
the expression becomes
>
which represents the potential energy of a unit magnet pole placed cm. from a pole m; or it is the magnetic potential due to a pole m at a distance of r cm. from the pole.
because it is an inherent function of the geometry of space, as is clearly If the evident from the above evaluation of the flux issuing from a pole. quantity 4ir is eliminated from some expressions, it will inevitably reappear
in others.
GENERAL LAWK AND DEFINITIONS

The work required
to
move a
unit
magnet pole from one point
to another in a magnetic field can be calculated as follows: Let Pi, Fig. 4, be the initial, and Pz the final position of the unit
pole,
and let the path between them be any curve whatsoever. At any general point on the curve, distant r cm. from m, the
77?
force will be
-^,
and the direction
of this force will be displaced

9.
from the elementary path ds by an angle

the distance ds will be
The work done over
dV =
and the
total
-s 2
ds cos 6
r2
-5
dr
will
work
in going
from PI to
2
ri*
But
7*1
n-*
mm
P2
be
(7)
is
the magnetic potential at PI, and

7*2
is
the magnetic
the magnetic
Hence the work done by the agency producing upon a unit magnet which moves from one point to pole
potential at P%.
field
another in the
field is
simply the
differ-
ence of magnetic potential between the points, and is independent of the path
followed in the travel from one point to the other.
the magnetic potential Ot the terminal point of the travel is higher than that of the starting point, work
Tr
Determination of FIG. 4. difference of magnetic potential between two points in a
must be done by an external agency to produce the motion of the unit testing pole, and the work so performed reappears as into
If such a system is left creased potential energy of the system. will the stored be itself, energy dissipated by the separation of
the poles under the influence of their mutually repelling forces, provided the poles are free to move.
above discussion, the pole -f is replaced by an of or all of the equal pole opposite sign, m, expressions for the forces and potentials will be reversed in sign. Repulsions become
If,
in the
attractions and work done by the system becomes work done upon the system. This case is entirely analogous to that of two heavy particles of ordinary matter which attract each other with
a force proportional to the product of the masses of the particles and inversely proportional to the square of the distance between
them.
tential
finite
energy
In the case of attracting magnet poles the magnetic pois the work required to move one pole to an in-
gravitational
from the other; in the case of attracting masses the gravitational potential energy is also the work required to move one of them to an infinite distance from the other. On the one hand the work is done by moving magnet poles in a magnetic field; on the other hand by moving gravitating matter in a gravitational field. 8. Equipotential Lines and Surfaces. The locus of all points in a magnetic field which have the same magnetic potential is called an equipotential surface. Linear (or curvilinear) elements of such a surface, connecting points of equal potential, are called
distance
No work is required to carry a magnet pole equipotential lines. from point to point in an equipotential surface or line. It follows, therefore, that the lines of force must intersect the equipotential surfaces at right angles, for if they did not there would be a component of force acting along the tangent to the equipotential surface at the point of intersection, consequently work would be required to move the pole along the surface, which is
contrary to the assumption. 9. Induced Currents and
E.M.F.
It
was discovered by
is
Faraday
in
1831 that a closed conductor which
threaded
by, or linked with, a magnetic field will have a current induced in it when the
strength of the
field
is
altered.
This
phenomenon
duction,
and
called electromagnetic inmay be demonstrated in a

is
number
of different ways.
Thus,
if
the
stationary closed conducting ring of Fig. 5 is threaded by lines of magnetic force

FIG.
5.
Conducting
cir-
cuit in magnetic field.
whose number changes from instant to

instant, a current will flow in the ring;
again, there will be a current flow if the field is steady and the ring is rotated around a diameter so as alternately to include and
exclude the magnetic flux; and again if the field is steady and the ring is given a motion of translation parallel to itself from
a region where the
field
has a certain intensity to a region
where the intensity is different; but if the ring be given a simple motion parallel to itself in a field of uniform intensity there will be no induced current. It should be borne in mind that the flow of current in the various cases mentioned above is dependent upon the condition that The the circuit be closed. H II II II induced current is a secondeffect of the
ary phenomenon, the primary changing magnetic field being to induce an
e.m.f.
which
1
in turn
produces
FIG.
G.
the current.
let
For instance,
the wire ab
move
to the
Development
of
E.M.F.
represent magnetic right along the rails SS', Fig. 6, and let H, There lines of force at right angles to the plane of the rails. will result a displacement of electricity along the wire ab, a positive charge appearing at b
and a negative charge means that there will be a

if
at a.
This
difference of
potential (Art. 14) between a then the rails are joined by a b; conductor cd, a current will flow around the closed circuit abed, but the difference
electrical
and
of
electrical
potential
may
exist inde-
pendently of the current, as for instance, when the circuit is open.

FIG.
7.
hand
tion.
rule.
Fleming's rightGenerator ac-
10. Direction
of
Induced E.M.F.
Fleming's Right-hand Rule.

ient
conven-
method
of determining the direction

itself is
If
an
electrical circuit that
is
not closed upon
"cut" by
lines
force, as described above, what actually occurs is a displacement of electricity along the conductor in a direction mutually perpendicular to the direction of the field and to the direction of motion of the conductor.
of
magnetic
its
is in process of displacement, constitutes a true current, but currents of this sort are called displacement currents to distinguish them from the dynamic currents ordinarily dealt with in direct-current circuits. Strictly, therefore, the
While the
electricity (consisting of electrons)
movement
primary
circuit
is
effect of
a changing magnetic flux upon a conducting electrical
to produce a displacement current, which in turn gives rise to a difference of electrical potential between the terminals of the circuit, and a
(dynamic) current
if
the circuit be closed.
10

and of the resulting current if the circuit as Fleming's rule, which is as follows: Hold the forefinger and middle finger of the right hand mutually
is
of the induced e.m.f.

is
closed
known
thumb,
perpendicular to one another, like the three axes of space coordinates, as illustrated in Fig. 7 point the forefinger in the direc;
tion of the lines of force, the thumb in the direction of the motion of the wire, then the middle finger will point in the direction of
the induced e.m.f.

11.
Force
Due
to
a Current in a Magnetic Field.
The
defi-
nition of unit current in the absolute electromagnetic system of units is based upon Oersted's discovery that a current of electricity will deflect
a compass needle in
its
neighborhood.
This
:
fact
form by Laplace as follows Let dl, Fig. 8a, represent an element of a wire ab, which is carrying a current of / absolute units (abamperes, see Art. 12), and let
into mathematical
was put
df
00
FIG.
8.
Force duo to current in a magnetic
field.
m be a magnet pole of strength m units.

element
dl will
The
force acting
on the
then be
df
-j / dl sin
(p
dynes
(8)
where
gent
is
<p
is
the angle between the radius vector r and the tan-
to the wire at the element.
The
direction of this force

7
,
perpendicular to the plane through r and T and under the Conconditions shown in the figure -will be directed upward.
versely, the pole
m will be acted upon

also
by an equal
force, directed
downward.
1
This relation
is
known
as the law of Biot-Savart.

In equation
(8),
11
-y
is
the
field intensity at
the element due to
m.
Hence
df
If r is
perpendicular to
dl,
= HI dl sin = 1, in which sin

<p <p
case
It
follows,
therefore,
that
if
straight wire I cm. long carrying a current of 7 abamperes is placed in

in a uniform field of intensity such a manner that its length is
perpendicular to the lines of force (Fig. 9a), it will be acted upon by

a force
Force due to a current FIG. 9. in a magnetic field.

(9)
/
If
= Hll dynes
<p
the axis of the wire makes an angle
with the direction of the
flux (Fig. 96), the force
becomes
/
= Hll
sin
<p
(10)
Equation
(8) serves to
determine the force exerted on a mag-
net pole m, Fig. 86, by a current I flowing in a wire of infinite length, the pole being at a distance p cm. from the axis of the
wire.
Thus,
-.r
if
mldl
sin
<p
m =
1,
Jo / becomes the
straight wire of infinite
P #, at a distance p from a length carrying 7 abamperes; or

field intensity,
:
= 2mp I
dl
2ml
y
<p
2j
_ :=
27 To^
(H)
where I
If
is
the current in amperes (see Art. 16).

is
the wire
of finite length
and
m =
P
1,
-Zi
"
Tv
sin
(sin ai -f sin
2)
(12)
12.
(9)
Unit Current. Unit Quantity. It follows from equation that the absolute unit of current may be defined as a current of
such strength that if it flows in a straight wire 1 cm. long placed perpendicular to the lines of force of a uniform magnetic field of
12
unit intensity, the wire will experience a side thrust of 1 dyne. This absolute unit of current is called the abampere. Unit quantity of electricity (in the same absolute electromagnetic system of units) may then be defined as that amount of electricity which will pass a given cross-section of a conductor in
one second when the current strength is one abampere. This unit of quantity is called the abcoulomb. 13. Direction of the Force on a Conducting Wire. Fleming's Left-hand Rule. Equation (10) shows that whenever a wire carrying a current lies in a magnetic field there is a force exerted upon it amounting to HIl$m<p dynes. This action is utilized in the electric motor, which consists essentially of a
number
of wires carrying currents
and
arranged to rotate in a powerful magnetic field. It is important to be able to determine the direction of the force acting on a wire for any given set of
This can be done by means of Fleming's left-hand rule: Hold the thumb, forefinger and midconditions.
dle finger of the left hand mutually perpendicular to one another, as shown in Fig. 10. Point the forefinger in the direction of the lines of force and the middle finger in the direction of the current,
will point in the direction of the force on the be noted that this rule is the same as Fleming's rule for the direction of the induced e.m.f. (generator action) except that the left hand is used instead of the right hand. 14. Magnitude of Induced E.M.F. Unit difference of electrical potential in the absolute electromagnetic system is said to exist between two points in an electrical field or in an elec-
then the
wire.
thumb
It will
trical circuit
when
unit
work
(the erg)
is
expended
in
mov-
ing unit quantity of electricity (the abcoulomb) from the one This unit is called the abvolt. If, then, Q point to the other. abcoulombs are moved from one point in a circuit to another
by
point whose electrical potential differs from that of the E abvolts, the work done is
first
point
= EQ
ergs

If this
13
work
is
is
done
in a
time
seconds, the power, or rate of doing
work,
P =
since Q/t
W
r
t
O = = #7 El
t
ergs per second
(13)
I (Art. 12).
Consider now the circuit of Fig. 11, and let the wire ab of = ds/dt cm. per length I cm. move to the right with a velocity v an e.m.f. in wire the will be There second. of, say, E generated
14

it
In the above discussion

netic field
was
tacitly
assumed that the mag-
swept across by the moving wire was uniform, and by supposition the velocity was constant. But if the field is not
if the velocity is variable, equation (14) still holds e.m.f. will simply vary from instant to instant the rigidly true; in such a manner that the equation is continuously satisfied. Let the wire start from position aibi and move to position
uniform, and
a 2 &2.
The
flux originally linked with circuit aibicd is 0i
His i
and
in the final position the flux enclosed is
02
Hl(si
+
=
s) is
The change
and
if
of flux during the
movement
01
=
this
02
t
His
is
change occurs in
~T
seconds the average rate of change

n. t/Vaverage
Hi
S ~
& average
which, in words, states that the average induced e.m.f. is equal to If the the average rate of change of flux linked with the circuit.
rate of change of flux is not uniform, the resulting variable e.m.f. will at any instant be given by
is a general form of the equation expressing Faraday's law. the circuit linked with the flux has turns, the absolute e.m.f. induced by a change in the flux will be at any instant
which
If
E = N
15.
abvolts
(15a)
Lenz's Law.
is
wire ab
field
moved to the right by the
Inspection of Fig. 11 shows that when the external force /', the induced
its reaction against the magnetic = /'. Moreover, the current is an force/ produces opposing so directed that it tends to produce a magnetic field directed upward, in opposition to the inducing field H, and as the wire moves to the right in such a way as to enclose more and more flux
current has such a direction that
in the circuit, the effect of the current is to oppose this increase. In other words, the induced current opposes the action which pro-
duces
it.
This
is
known
as Lenz's law.
15
e.m.f.
Since the induced current and, therefore, also the induced oppose the inducing action, it follows that a positive increof flux should be regarded as
producing a negative e.m.f. be written should Consequently equation (15a) d E = abvolts (16)
ment
-N ^
in
Units of Current and E.M.F. The practical unit which the strength of a current is expressed is the ampere, which has such magnitude that 10 amperes are equivalent to 1
16. Practical
abampere.
8 practical unit of e.m.f., called the volt, is taken 10 times as large as the absolute unit, the abvolt, since the latter is incon-
The
veniently small; that
is,
10 8 abvolts are equivalent to
volt.
Hence, from equation (16)
E =
-N~X
(it
10- 8 volts
is
(17)
Since the power in a direct-current circuit
P = El
when
ergs per second
tively, this expression
and 7 are expressed in abvolts and abamperes, respecbecomes

(E
P =
10 8 )
(^)
= El X
10 7 ergs per second
when E and / are expressed in volts and amperes, respectively. But 10 7 ergs per second are equivalent to 1 watt, so that
P = El
17.
watts
of / of
Heating Due to a Current. The Joule. When a current amperes flows through a resistance of R ohms, the difference electrical potential between the terminals of the resistor is E
such that I
IR, in accordance with Ohm's Multiplying both sides of the last equation by 7, there
or
volts,
= E/R,
E =
law.
results
P = El =
I 2 R watts, or
PR X
10 7 ergs per second
which represents the power absorbed in the resistance R. time t seconds the energy supplied to the circuit -is
In a
W
and
this
Pt
= PRt X
10 7 ergs
resistor.
energy reappears as heat in the
Since 10 7 ergs
16

is
are equivalent to 1 joule, and 1 joule per second to 1 watt, the above expressions can be written
equivalent
W
The heating
readily calculated
= I Rt p = El =
2
joules
I 2 R watts
(18)
effect of a current flowing
through a resistor can be
when
it
is
remembered that the mechanical
is 4.19 joules in the metric system; 4.19 joules, or 4.19 X 10 7 ergs of work are required to raise the temperature of 1 gram of water 1C.
(Joule's) equivalent of heat

is,
that
18. Field Intensity Due to a Circular Coil. Let P, Fig. 12, represent a unit magnet pole on the axis of a plane circular coil of turns and radius r cm. Let P be x cm. from the plane of the Then the coil, and let the current in the latter be 7 abamperes.
force acting on
an element
dl of the coil
is,
by equation
(8),
acting in the direction indicated in the figure; and the pole is acted upon by an equal force in the opposite direction. 1 Resolv-
dF
ing
this
force
into
spectively parallel ular to, the axis of the

clear that the
components reto, and perpendiccoil, it will
be
of all the perpendicular components is zero since for
sum
each element of the

rise to
coil
which gives
a perpendicular component of force in one direction, there is a diaFIG, 12. Field intensity axis of coil.
on
metrically
opposite element which an produces equal and opposite com-
ponent.
The
axial
component
CLtL
of df
is
= Nidi 2
-~^-~:
+x
COS
a.
Nrldl
7-, 2
(r
r
+ x ^z )^
2
positive
magnet pole
in the field
to
move along the

The
lines of force that are
current.
rule:
direction of these lines
produced by a current will tend produced by, and link with, the can be easily found by the following
in the direction
Grasp the wire with the right hand, with the thumb
of the current flow.

lines of force.
The
ringers will then point in the
direction of the

Hence
17
(19)
-f-
At the center
of the coil,
where x
0,
HQ
from which
it
2irNI _2TT_N1
r
~10
(20)
follows that the absolute unit of current
(abam-
pere) may be denned as that current which, when flowing in a circular coil of one turn and 1 cm. radius, will act upon a unit
magnet pole
19. Field Intensity
at the center with a force of 2w dynes. on the Axis of a Solenoid.
Let Fig. 13
turns uniformly distributed over the represent a solenoid of It is desired to find the field intensity at a point length I cm. P on the axis, distant cm. from the center of the solenoid.
Consider an elementary section of the solenoid dx, distant x cm. from P. The element may be considered as a plane circular
]V"
coil of
-rdx turns; the
field intensity
due to
this
elementary ring
at the point
is,
by equation
(19),
JTJ
_
'
and the
total field intensity
is
then
ZrNIf*
'-~-
C^ +D
_dx
x*)*
ro
~r
./-&~(i-
27T^/
10
2+ D
-'2-
2- D
(21)
^
where
v^ +
D =
27TJV/
l2 becomes
At the center
of the solenoid,
0, this
18
which reduces to
i~o
if I is
(22)
large
compared with
r.
JVTur

The
tion of
19
physical interpretation of these facts concerning the variaalong the axis is as follows For some distance on either
middle section of the solenoid the lines of force inside the winding are nearly parallel, hence the field is nearly uniform will be practically constant; near the ends of the solenoid and the lines diverge in the manner indicated in Fig. 18, and the
Bide of the
decrease. greater the divergence the more rapidly will 20. Magnetic Potential on the Axis of a Circular Coil.
It has
.been
shown
in Art. 18 that the field intensity
on the
axis of a
circular coil, at
the plane of the

jfZ
any distance x from coil, is given by

4>
|
n\ 3Z
this being the force in
dynes that would act upon a unit magnet pole With the curplaced at the point.
FIG. 15.
rent flowing as indicated in Fig. 15, the unit (positive) pole would be
Magnetic potential on
axis of circular coil.
urged to the
coil.
left,
or
toward the
To move
the pole to the right over a distance dx there
must be expended
T7 dV
,
Hdx
--
--
(r
/2
ergs
of
infinity
work, and the total work required to move the unit pole out to from a point distant D cm. from the coil is
V=
2<irNIr 2
(
**
. 2
(r
+
=
2irNl(l
~j^= / Vr + D=)
2 2
cos a)
(23)
where a
is
P by
the
coil.
the semi-angle of the right cone subtended at the point But 2ir(l cos a) = a> is the solid angle at the
vertex of the cone, hence
V =
If
a>NI
(24)
would have been
the test pole had been of strength units, the work done times as great as the above amount, or
V m = wmNI
(25)
20
V, equation (24), is the magnetic potential at a on the axis of the coil; it represents the work regeneral point to a move unit quired pole from the point out to an infinite
distance,
The expression
when
the current flow
is
as indicated.
If
the current
is
reversed,
V becomes the work required to bring the unit pole from

up to the point
in question.
an
infinite distance
General Expression for the Magnetic Potential Due to a Coil of any Shape at any Point. Equation (25) can be put into a more convenient form, as follows: The total flux emanating from a pole of strength m units is
21.
maxwells
hence
cow
CO
4rr
that passes represents that part of the total flux due to the pole or links the turns of the coil. through, with, Therefore,
Vm =
<f,NI
\I ergs
(26)
where X = <pN, the product of the flux and the number of turns with which it links, is called the number of flux linkages. Or,
in other words, the potential
field
energy of a current in a magnetic
produced by some other agency is the product of the current (in abamperes) and the number of flux linkages. The above expression for the potential energy of a magnet pole in the presence of a current was derived by assuming a circular coil and
allowing the magnet pole to move along the axis of the coil. It can be
shown, however, that the resulting

equations (24) and (25) are general. For let Fig. 16 represent a coil of any shape which may or may not be plane,
+m
FIG. 16. Magnetic potential due to coil of any shape.
and
let
co
tended by
in
represent the solid angle subit at a point pole placed
any general position, linking the coil is
Then
the flux emanating from
and
com maxwells
Now allow the pole to move in such
manner that the solid angle

changes
the coil
21
magnitude by will change by

in
dco in a
time
dt.
The
flux linked with
=
and there
will
m
coil
be induced in the
an
e.m.f.
E = - JV~ abvolts
If this
change of
flux occurs while a current of /
abamperes
is
flowing in the coil, the
work done
is
given by
dV = -EJdt =
and the
tance
total
(GO
NIdtp
mN I do) ergs
infinite dis(co
work required to bring the pole from an
0) to a point near the coil
co') is
= mNI
or the
/**
'
f'
mta'NI
<pNI
\I
same
as equations (25)
and
(26)
Let Fig. 17 represent the side and 22. Magnetomotive Force. front elevations of a plane coil of any configuration, carrying a current of I abamperes, and let a
unit
magnet pole be placed
at a
point outside
in the plane of the coil

its
but
of
boundary.
None
the flux emanating from the pole will pass through the coil, and
therefore
at
the
zero.
magnetic potential
is
Now
let
the unit
pole be carried along any path to a point Q infinitely close to the The solid angle plane of the coil. subtended at Q by the coil is
27r,
FIG. 17.
Closed path linking with coil.
Q is then 2irNJ, and the difference P and Q, that is, the amount of between magnetic potential work required to carry the pole from P to Q, is lirNI ergs. If the pole is now carried back to P along any path which requires the pole to be threaded completely through the coil, as Q'P, a further amount of work equal to 'ZirNI ergs must be expended, making a total of 4irNI ergs to carry the pole once around a closed path linking
the magnetic potential at
of
ivith the coil.
coil.
This
is
The
unit of m.m.f.
called the magnetomotive force is called the gilbert.
(m.m.f .) of the
22
It

should be particularly noted that m.m.f.
is riot
a/orce;it
is
of the nature of
work per unit magnet
to e.m.f ., which
trical
If
It is exactly analogous is likewise not a force, but work per unit of elecpole.
quantity (see Art. 14).

coil of Fig.
the
Fig. 18, the
12 is pulled out into the form of the solenoid of work required to carry a unit magnet pole around a
all
closed path threading through
of the
N turns
is
47rNl
= fev7
FIG. 18.
Magnetic path linking solenoid.
But the
noid
is
field intensity at
the middle p'oint of the axis of the sole47T
NI
so that
H
which means that
axis,
TNI
m.m.f.
if the field intensity were constant all along the the work required to carry the unit pole from end to end of the solenoid would be
force
distance
^ NI
axis, so
However, the
the total
is
field intensity is
not constant along the
that
amount
of
work to describe the complete path
of Fig. 18
INI =
which states that the m.m.f.
force.
Hdl
(27)
is the line integral of the magnetic,
23
It is interesting to note that the area under the curve of Fig. 14 represents the work required to carry a unit magnet pole through the solenoid from one end to the other, while the area of the rec-
tangle enclosing the curve represents the work required to carry the unit pole once around a closed path linking all the turns of the solenoid. It follows, therefore, that the area above the curve, but
inside the rectangle, represents to the same scale the work required to carry the magnet pole along the path from end to end of the
solenoid, but outside of its windings.
23. Permeability. solenoid of Fig. 13 is
The
field
intensity at the center of the
47T
16
NI f gausses
and
lel
if the lines of force passed straight through the solenoid paralto the axis, and were uniformly distributed over the area of cross-section A, the flux across any section would be
(f)
AH
A
4?r
NI
Yt\ _Lu
maxwells ~i~A I
Now
I
let
the solenoid be provided with an iron core of length

sq. cm., as in Fig. 19;
cm. and cross-section
induced poles of
FIG. 19.
Solenoid
will
wound on
iron core.
strengths
bar,
+ m and
then be developed at the ends of the
and each
of these poles will give rise to a field intensity at the
center of the solenoid equal to 777^2 an(^ opposite in direction to

(I/ Z)
H.
The
resultant field intensity at the center
is
then
4T
10
NI
I
"
_2ro_
<Z/2)
_ ==
47T
NI
/
Sm
Z
10
24
In other words, the induced poles exert a demagnetizing effect
upon the field which produces them. This demagnetizing or "end effect " becomes negligibly small if the solenoid and core are long, and will be neglected in the remainder of this discussion. From the pole + m there will emanate 4?rm lines of force, all of m through the surroundwhich find their way back to the pole These lines of force may be assumed to be continued ing air. through the iron core back to the starting point, not as lines of
force,
but as
lines of induction, since
the former exist only in the
external non-magnetic medium. total flux consists of the original

lines of induction, or
HA
Inside the iron, therefore, the lines of force and the 4?rm
$ =
Assuming the
flux
3>
AH +
to be uniformly distributed over the crossis
section A, the flux density
B
where
==
+
M
=
=
ffl+4,r= MH
+
4T _
(29)
(30) It
is
called the permeability of the material of the core.
is
the ratio of the flux density in the material to the intensity of the inducing field, and is therefore a measure of the specific magnetic conductance of the material.
Its
magnitude
is
dependent
upon the
ratio
-rjf
or
77,
that
is,
the ratio of the strength of the
induced pole to the intensity of the inducing field. The better the material from a magnetic standpoint, or the more it is susceptible to magnetization, the greater will be the strength of the induced
*???
pole for a given inducing

susceptibility.
field,
hence the ratio
-,
is
called the
between m and H, so that it is 5 in terms of H. The relation must be found experimentally for each material. Curves showing the relation between B and H are called normal B-H curves, or magnetization curves. Fig. 20 shows a number of such curves as used in practice, for several different kinds of iron and steel.
There
is
no known
relation
/z
impossible to express either
or
25
In certain kinds of iron the permeability ju reaches very large values 2000 to 3000 but its value varies as varies, even in
one and the same material. In paramagnetic substances like is always greater than unity; in air and iron, nickel and cobalt, other non-magnetic substances, ji = 1; and in diamagnetic substances, n is less than unity.
//,
In the equation
5
100
4-jrNI
Jn ~y~>
500 600
700 00
tne term
NI represents the number

)
Ampere Turns
200
300 400
per Inch (Upper Curves
900 1000 1100 1200 1300 1400
2'0
30
40
50
60
100
//mC.G.S.Uaits
FIG. 20.
Magnetization curves.
of ampere-turns of the exciting winding, and NI/l expresses the number of ampere-turns per cm. It follows that
ampere-turns per cm.

In practical calculations
it is
=
-y-
-^-H
= O.SH
to deal with
more convenient
am-
pere-turns per cm. than with H, for if the value of the former quantity, corresponding to a given value of B, can be found, the total excitation (in ampere-turns) is simply the product of ampere turns per cm. and the length of the circuit. magnetization curves are commonly drawn with
Consequently, (lines per sq.
26
cm.) as ordinates and with 0.8#

abscissas.
If
(= ampere-turns per cm.) as English units are employed, the curves are drawn with ordinates B X (2.54) 2 (= lines per sq. in.), and with abscissas
2.54
O.SH (= ampere-turns per

24.
cuit.
inch).
The Law
of the
Magnetic Cir-
Magnetic Reluctance. The demagnetizing effect of the ends of the

core
of Fig.
19,
referred to in the
preceding section, can be eliminated by bending the core into a closed ring
The value of form, as in Fig. 21. will then be uniform around the entire
circular axis of the coil,
FIG. 21.
and
will
be
Ring
core.
H=
will
TT^JI
where
flux
is
the
mean
length of the core in centimeters.
The
total
through the core
then be
4?r
$ =
or
AB
A r>
NI
The numerator
solenoid, netic circuit.
is
of
equation
(31)
is
is
the
m.m.f.
of
the
and the denominator
the reluctance of the magIt will be noted that the expression for reluctance
for the resistance of
of the
same form as that

it is
cuit, for
proportional to the length
an electrical cirand inversely proportional
to the cross-section; moreover, the permeability p appears in the expression for the reluctance in exactly the same manner as does the specific conductance in the expression for electrical resistance,
hence the reference to permeability (Art. 23) as "specific magnetic conductance." The unit in which reluctance is measured is
called the oersted.
Equation
(31)
is
of the
form
flux
m.m.f.
reluctance

or, in
27
terms of the units themselves,
maxwells
- gilberts
-
oersteds
3-
which corresponds term by term with Ohm's law of the

circuit
electric
current
or
e.m.f.
resistance
volts
^r
amperes
ohms
is
The
reciprocal of reluctance, or -y,
called permeance.
Circuit.
25. Applications of
Law
of
Magnetic
Magnetic
cir-
be joined in series, in parallel, or in series-parallel, and the solution of problems involving any of these combinations is in every case carried out by methods that are the exact analogues of those used in the corresponding eleccuits, like electric circuits,
may
trical circuits.
magnetic
is
circuits
indicated.
Thus, in Fig. 22, parts a and b represent typical and in each case the analogous electrical circuit The following examples will serve to illustrate the
in the solution of ordinary problems. circuit consisting of a number of
methods to be employed
1.
SERIES CIRCUITS.
reluctances in series will have a total reluctance given
by
and the resultant
flux
through the
circuit will
be
UlAi
V-lA 2
HnA n
where
NI
is
the total
circuit as a whole.
number of ampere-turns acting upon the This equation can be written
and in this form each term of the right-hand member represents the m.m.f. (in gilberts) required to maintain the flux through the corresponding portion of the circuit. The total m.m.f. is then
28
merely the sum of the m.m.fs. required by the individual parts. In the analogous electrical circuit
/
Rl
+ R% +
2
Rn
.
and
E = IR 1+ IR +
+IR n
or the total e.m.f. required to maintain the current through the circuit is the sum of the potential drops in each part of the circuit.
In most cases arising in practice it is not necessary to compute the reluctances of the several parts of a circuit in accordance with the above equations. A more direct solution is possible through
the use of curves like those of Fig. 20. Thus, in Fig. 22a, let it be required to find the number of ampere-turns necessary to
the core
produce a total flux of 160,000 maxwells on the assumption that is made of cast iron. Assume that the mean path of the
lines of force follows the center of gravity of the cross-section
and
that at the corners the

/. Zi
mean path
follows quadrants of circles.
= =
length of path in cast iron 26.15 in.

length of path in air
flux density in iron
lines per sq. in.
2(6
in.
4)
2*
0.125
la
0.125
B =
and
in air-gap
40,000
the curve for cast iron in Fig. 20 it is found that a flux of density 40,000 lines per sq. in. corresponds to an excitation Hence the number of of 79 ampere-turns per inch length of core.
From
ampere-turns required by the core is 79 X 26.15 = 2060. The number of ampere-turns required to maintain the flux through the
air-gap,
4?r
where
/*
1,
may
be found from the relation

,
B = H =
NI
-T- (where
all
Y~
quantities are in metric units)
or
(32)
NI =
0.8BI
0.3133
lines per sq. in.
air-gap in inches
Hence the ampere-turns
for the air-gap are
NI =
and the
3627 ampere-turns.
0.3133
40,000
1567
total excitation for the entire circuit is
2060
1567

2.
29
two magnetic
In Fig. 226 each of the type illustrated in Fig. 22a, Just as in the corresponding electrical are connected in parallel. circuit the entire battery e.m.f. acts equally on each of the parcircuits,
allel electric
PARALLEL AND SERIES-PARALLEL CIRCUITS.
circuits, so
does the entire m.m.f. of the exciting
circuit act
on each
of the
magnetic computed as though the other part were not present; are exactly alike the flux will be the same in each.
In the case of the parallel magnetic circuits. circuit of Fig. 226, the flux in each part is to be
if
the parts
Suppose, for example, that the left-hand circuit of Fig. 226 is exactly the same as that of Fig. 22a, but that the right-hand
circuit,
though having identical
dimensions, is made of cast steel instead of cast iron. Assuming that the flux through the left-
hand branch
maxwells, the
turns.
It
is
again
160,000
coil
must supply
an excitation of 3627 amperedoes not follow that

the flux through the cast-steel
be 160,000 maxwells, FIG. 22. Typical magnetic and it does not have that electric circuits. value because the reluctance of The problem is the cast steel is less than that of the cast iron. then to find that value of flux through the cast-steel circuit which will require 3627 ampere-turns for its maintenance. This can be done by trial, as follows: Assume a series of values of the total flux, and for each value
circuit will
and indeed
compute the corresponding total number of ampere-turns. Plot flux and ampere-turns, extending the computations far enough so that a curve may be drawn that will include within its range the
given number of ampere-turns; the actual flux corresponding to the latter can then be read from the curve. In the case under
discussion the total flux through the cast-steel circuit way found to be approximately 303,200 maxwells.
is
in this
In such a case as that illustrated by this particular problem, the reluctance of the cast-steel part of the circuit is so small compared with that of the air-gap that a first approximation to the final
result
may
be found by assuming that the reluctance of the path
30

steel is negligible,
and that the entire excitation is maintaining the flux through the air-gap. On this basis the flux density in lines per sq. in. would be given by
through the
consumed
in
3627
0.3133
X B" X
or B" = 92,800 lines per sq. in. and $ = 371,200. Reference to the curves of Fig. 20 shows that at this flux density the steel
circuit would 3627 + 1635 = 5262 ampere-turns, or considerably more than the available number. It is then necessary to select a
would require 1635 ampere-turns, so that the entire
require
smaller value of B", repeating the calculations until the given number of ampere-turns is included in the range of trial values.
KirchhofFs Laws. The solution of problems involving the flow of current in networks of conductors depends upon two experimental facts, known as Kirchhoff's Laws:
26.
1.
The
algebraic
sum
of the currents at
is
any junction
of the
conductors in the
2.
network
zero.
The
algebraic
sum
is
of the potential drops
around any closed
loop in the
network
these
zero.
is
The
first of
two laws
a statement of the fact that the
sum of all the currents entering a junction point is equal to the sum If this were not so, the of all the currents leaving that point.
charge of electricity at the junction would steadily change and its potential would change correspondingly; no effect of this kind has ever been observed. In applying the second law it is convenient to make a diagram
of the network and to give to each active e.m.f. (such as that from a battery or dynamo) an appropriate symbol to indicate its magnitude and an arrow to indicate the direction in which it acts; and each conductor is to be given a symbol to indicate the magnitude of the current flowing in it, and an arrow to indicate the assumed direction of the current flow. Let it be agreed that the clockwise direction around any closed loop of the network
be considered to be the positive direction through the circuit (though cases may arise when it might be more convenient to select the counter-clockwise direction as the positive one).
shall
active e.m.f. that is directed positively around the loop to be given the positive sign; it may in that case be considered A curto produce a rise of potential in the positive direction.
Then any
is

rent of
i
31
amperes flowing in the positive direction through a ohms will produce a fall or drop of potential of ir volts, which must, accordingly, be treated as negative. Example. Fig. 23 represents a storage battery connected to " a three- wire system, the neutral" wire c being connected to the middle point of the battery. Each half of the battery has an A and B are the loads e.m.f. of 115 volts, directed as shown. which are supposed to con^â Ohm
resistor of r
0.1
sist
of lamps,
heaters,
or
_T
other devices which do not
oimiiiit 113 y
lts
develop a counter e.m.f.,
~y~
.
0.2
ohm
and have resistances of 6 and 8 ohms, respectively. The outer wires, a and 6, have each a resistance of 0.1 ohm, and the neutral
_=.
|jB|80hros
Ohm-=-[115Volts
I<IG.
23.
.Net-work ot conductors.
wire c has a resistance of 0.2 ohm.
Each
a virtual internal resistance of 0.5 ohm.
half of the battery has It is required to find
the current in each of the supply lines. From the first law, the sum of the currents entering point P must equal the sum of the currents leaving it; hence, from the
assumed directions
of current flow
From
diagram,
the second law, referring to the upper loop of the
+
And
115
4(0.5
+
-
0.1
+ +
6)
+ +
0.24
referring to the lower loop
+
ia
115
0.24
4(8
0.1
0.5)
From these three independent equations it is easily found that = 17.31, 4 = 13.47 and 4 = - 3.84. The meaning of the negative sign of 4 is that that current actually flows in a direction
opposite to the assumed direction. KirchhofFs laws are applicable to the magnetic circuit as well as the electric circuit. Thus at any junction in a magnetic
circuit,
tion
the number of lines of induction coming up to the juncmust be equal to the number leaving it, for the reason that
32

always closed loops. This is equivalent law of the electric circuit. Again, in any
lines of induction are
to KirchhofTs
first
closed magnetic circuit, the algebraic sum of the drops of magnetic If in any part of the closed magnetic potential must be zero. circuit the flux is and the reluctance R, the drop of magnetic
<
potential is &R, and the summation of all such drops must then be equal to the summation of all the active m.m.fs., with due attention to the sign of each term.
27. Self-induction.
When
a straight wire. Fig. 24a, carries
a current in the direction indicated, the wire will be surrounded by magnetic lines of force as shown. As the current increases
from zero to any arbitrary value, the flux will increase proportionally from zero, and may be thought of as issuing from the center of the wire and expanding outward, like spreading ripples
FIG. 24.
(6) (a) Lines of force surrounding a conductor.
The lines of force thus expanding cut across the wire manner indicated in Fig. 246, which represents a crosssection of the wire in (a) when viewed from the left. The txon a pond.
in the
about to cut across the longitudinal at P, the motion of the line of force at this point being radially outward. Relatively, the effect is the same as though the filament were moving radially inward so that if Fleming's (right-hand) rule is applied, it is found that the induced e.m.f. is directed outward from the plane of the paper, or in opposition to the direction of the current flow. The whole effect is in accord with Lenz's law; here, however, the original change in the current strength which produced the change in the flux immediately calls into existence an opposing e.m.f. which tends to retard the change in current. Conversely, the same line of reasoning will show that an initial decrease of current induces an e.m.f. of reversed direction, which tends to main-
panding
line of force
is
filament of the wire
shown

tain the current at
its
33
self-
original strength.
This e.m.f. being
induced,
called the e.m.f. of self-induction. Let Fig. 25 represent a coil of wire wound constant permeability ^, a cross-section of
is
on a core having sq. cm., and a
mean length of magnetic path of I cm. The mean path is to be taken as passing through the center of gravity of the cross-section On passing a current of i amperes through the coil of the core. there will be produced a flux
N'
*
=
10
"T
4*Ni A To~T"
and a change
of current di will
produce a change of flux of

N_di
4rr
This change of flux will then induce an e.m.f.
where
T
4ir
N nA
2
TO
r~
The quantity L is called the coefficient of self-induction or the self-inductance of the circuit, and in the practical system of units is measured in terms of a unit called the
It is evident henry. the self-inductance
from equation
is
(34) that
square of the number the flux, and is dependent upon the shape, size and material of the magnetic circuit. Its magnitude is of very great importance in all electrical circuits in which the current is
proportional to the of turns linked with
changing in strength, as for instance, in those coils of a direct-current generator or motor which are undergoing commutation (Chap. VIII). From equation (33) it is seen that the self-inductance L of a circuit is numerically equal to the e.m,f. induced in it by a current
34

is
which
that
is,
changing at the rate of
ampere per second

1
(~r.
=
Ij
;
a circuit has a self-inductance of

1
henry
if
a change of
of
1
current of
ampere per second induces an
e.m.f.
volt.
The
self-inductance
(33)
may
also be defined in another
way.
Thus
from equation
or
L =
in this
N^ X 10number
(35)
equation -p
is
numerically equal to the rate of change

it is
of flux with current, or

1
the
of lines of force
produced
ampere. Equation (35) says that the product of the numby ber of lines produced by 1 ampere, multiplied by the number
of turns
which
this flux links,
and divided by 10 8
is
equal to the
coefficient of self-induction.
The product
the
number
of turns with
which
of flux per ampere by this flux links is called the num-
ber of flux linkages per ampere, so that, briefly, the self-inductance 8 is equal to the number of flux linkages per ampere, divided by 10
.
turns, respectively, are so placed with respect to each other that the magnetic field due to a current in one of these links in whole
28.
Mutual Induction.
If
two
circuits of
Ni and N%
the
or in part with the other, a change in the current strength in first circuit will induce an e.m.f. of mutual induction in
the second circuit.
It is clear that the
magnitude of
this will
depend upon the geometrical shapes and relative positions of the two circuits, as well as upon the rate of change of current in
the inducing circuit. Let a current of ii amperes in the
first circuit
produce a flux
3>i
such that
47r
A7
,
-
*i
Jgtfrfl
n
Mi^-i
= Ci# 1*1
(36)
part of this flux, or
= Xi*i =
KiCiNiii
(37)
35
the total
(where K\ ^1) will link with the second circuit of 2 turns, so that number of linkages with the second circuit is
X 21
= NWI = K&iNiN&i
is
(38)
i2
and
if
the second circuit
traversed by a circuit of
amperes
its
potential energy in the presence of the
first circuit is,
by
Art. 21,
V 21 = X,i= Ho
Similarly, the current
total flux
4?r
i2
K&NiNMi
circuit
will
ergs
(39)
in the second
produce a
3>2=r-- =
of
10
N*
C 2 N zi z
(40)
which a part
 2
2i2
(41)
(where
total
1) will
link with the first circuit of
Ni turns, so that the

is
number
of linkages with the first circuit

Xi 2
Nw
= KiCtNiN&a
(42)
The
potential energy of the

is
first circuit in
the presence of the
second
Fi2
X 12
~=
Ho KtC^Niite
ergs
(43)
But Vzi must be equal to V\, sijice the potential energy of the system can have but one value;
.'.
KiCiNiNs =
C 2N
(44)
From
(38)
or
it is
the
number
of flux linkages with the second circuit

first circuit;
due
to
unit current in the
and from
(42)
which represents the number of flux linkages with the first circuit due to unit current in the second. Hence, from (44), it follows
36
that unit current in one circuit will produce the same number of linkages in the other, as unit current in the latter will produce
in the former.
When
the current in circuit No.
changes, the e.m.f. induced
in circuit
No. 2
is
10
~8
= ~ KiCiNiN
10
~8
and when the current in circuit No. 2 changes, there duced in circuit No. 1 an e.m.f.
e,
will
be
in-
N^ X
10-
= - KtCtNtN,-* X
be written
ID-"
From
(44), these
equations
62
may
= -
di,
and
(45)
where
M
is
= KiCiNiNi X
10- 8
= KtCtNiNt X
.
10~ 8
(46)
the number of flux linkages with one circuit due to unit current 8 This is called the co(the ampere) in the other, divided by 10
efficient of mutual induction, or the mutual inductance, of the two circuits. It is obviously of the same nature as self-inductance, and is measured in henries. From (45) it follows also that the mutual inductance of two circuits is numerically equal to the e.m.f. induced in one of them when the current in the other changes at the
rate of 1
It
is
circuit
ampere per second. from equation No. 1 is

clear
(36)
that the self-inductance of
Li
^^ X 10î
= Ci#i 2 X
10- 8
(47)
and from equation
(40) that
L2 =
Hence, from
^X
and
10~ 8
(48)
- C 2 AV X
10- 8
(48)
(46), (47)
= KiKsLiLi
(49)
37
If the circuits are so related that there is no leakage of flux between them, that is, if all of the flux produced by one circuit links with all of the turns of the other,
K, =
and
M
or the
mutual inductance
of
two perfectly coupled
circuits is a
mean proportional between their respective self-inductances. The factor \/KiK 2 is sometimes called the coefficient of coupling. The phenomenon of mutual induction is utilized in the inducand in the alternating-current transformer, both of which consist of an iron core upon which are wound two coils, the primary and the secondary, insulated from the core and from each An interrupted or alternating current in one winding sets other. a periodically varying flux which in turn induces an alternatup in the other winding. Mutual Induction is also of ime.m.f. ing
tion coil
portance as a factor in the commutation process in direct-current machines.

29.
Energy Stored
in a
Magnetic Field.
coil
or circuit of
self-inductance
henries carrying a variable current will have
induced in
it
an e.m.f.
e
di
at
-yr
VOUS
If
the current
is i
amperes at the moment when the rate of change

effect
of current
is -j-
amperes per second, the power required to

is
the change of current

(
e)i
= Li
time
-T7
watts
and the work done
in the
(
dt is
dW =
The
total
e}idt
Lidi joules
raise the current
amount
i
of
is,
work required to
therefore,
from
zero to a value
=
j
Lidi
= ^Li 2
joules
(50)
This energy is not lost, but is stored in the magnetic field, and may be recovered by allowing the magnetic field to collapse to
38
It is this energy which appears in the spark or arc zero value. formed on opening an inductive circuit. It is instructive to compare equation (50) with the equation This is of the form for the kinetic energy of a moving body.
W
where
y%
mv 2
In the case ra is the mass of the body and v its velocity. of the electric circuit the current i is the quantity of electricity that passes a given point in a second, and is analogous to velocself -inductance L represents a sort of electrical inertia, operates to resist any change in the current flow, or electrical velocity; it is therefore analogous to the mass of a ity.
The
it
since
mechanical system.
The energy J^Li 2 may,
therefore, be con-
sidered as the kinetic energy of electricity in motion. When two circuits of self-inductances LI and L 2 have a mutual
inductance
equal to
M,
there
is
stored in the system an
amount
of
energy
= JLii! 2
+ KL
2t 2
Miiit joules
(51)
provided the two currents magnetize in the same direction. The first two terms of equation (51) is obvious from (50) the last term, Miii Zj can be derived as follows The potential energy of one circuit in the presence of the other
derivation of the
;
:
is
from (39) and
(43)
W
and by
= Ho
tfiCitfitf 2*1*2
= Ho
#2Catfitf 2*1*2 ergs
(46) this
becomes
10 8)*i* a
= Ho(M X
= Miiia X 10 7 = Mi iiz joules
ergs
If the two circuits magnetize in opposite mutual potential energy is evidently reversed the stored energy of the system is
directions, their in sign, so that
= %Liii 2
+ JL
2
2 *2
~ Mitt
(52)
30. Tractive Effort of Electromagnets. Let a unit magnet pole be placed at the point P, Fig. 26, on the axis of a cylindrical bar magnet of radius r cm., and distant a cm. from the end of the bar magnet; let the strength of the pole of the magnet be m units, assumed to be uniformly distributed over the end surface of the

cylinder.
39
The
pole strength per unit area, or the intensity of

is
magnetization,
then
Considering an annular element of radius x and width dx on the end surface of the magnet, the force which it will exert upon the
unit pole at
is
and the
total force
F =
where
is
due to the entire pole C r xdx

j
of the
magnet
cos
0)
is
27rcm
^p^p =
27r<7(l
(54)
P by the end of the magnet.
the semi-angle of the right cone subtended at the point If the distance a is made very small,
FIG. 26.
Field intensity on axis of bar magnet.
if a is small relatively to the dimensions of the end area of the magnet, cos0 approaches zero as a limit, in which case
or
F =
If
27TO-
(55)
two bar magnets are placed end to end with a very small separation, and if the intensities of magnetization of the adjacent surfaces are + a- and a, respectively, the attraction of one of the magnets upon an elementary magnet pole of area dA on the other will be dF = 2iro- X vdA and the total attraction between
;
adjacent poles will be
F =
From equation
of
"77?
2w(7
dynes
(56)
(53),
<r
=-r> and since the flux issuing from a pole

it
units
is
$ =
47rm,
follows that a
= T~X\
<j>
whence, from
(56),
2
<J>
BA
2
-8T
40
This
is

the fundamental equation underlying the design of tract-
ive or lifting electromagnets.
PROBLEMS
20 cm. in diam. is placed in a vertical plane that makes 45 deg. with the plane of the magnetic meridian. If the total intensity of the earth's magnetic field is 0.42 gauss and the angle of dip is 62 deg., what is the total flux that passes through the coil? 2. Two identical slender bar magnets, A and B, each 20 cm. long and having concentrated poles of 200 c.g.s. units at their ends, are placed with their axes in the same horizontal straight line, the south pole of A being adjacent to, and 5 cm. distant from, the north pole of B. Find the force exerted by one magnet on the other, and indicate its direction. 3. How much work must be done to turn magnet B of Problem 2 through 90 deg. about a vertical axis passing through its middle point? How much work must be done to turn magnet B through 180 deg. ? 4. A trolley feeder running in an east and west direction, and carried on poles 100 ft. apart, carries a continuous current of 300 amp., flowing from east to west. If the intensity of the earth's magnetic field has the
1.
circular coil of
an angle
value given in Problem

poles?
6.
-
1,
what
is
the force on a length of the feeder between
What
is its
direction?
A
is
amp.
storage battery having a short-time discharge rating of 10,000 connected to the switchboard by copper bus-bars which have
;
a cross-section of 1 in. by 10 in. the bus-bars are spaced 6 in. center to center, the 10-in. faces being placed in parallel vertical planes. Assuming that the current may be considered to be concentrated at the center of crosssection, what must be the distance between supporting brackets in order that the bus-bars may not deflect more than in.? 6. Solve the preceding problem on the assumption that the current in each bus-bar flows uniformly along the vertical plane through the central
axis.
7.
A
lie
that
horizontal wire 30 cm. long falls freely between frictionless guides in a plane perpendicular to the magnetic meridian plane. The
acceleration of gravity is 980 cm. per sec. per sec. and the intensity and What direction of the earth's magnetic field are as given in Problem 1. is the difference of potential between the ends of the wire at the end ,of the
third second of
8.
its fall?
10 turns and radius 10 cm. is revolved and west at the rate of 10 rev. per sec. (a) What is the average e.m.f generated in the coil if the intensity and direction of the earth's magnetic field have the values given in Problem 1? (b) What is the maximum e.m.f. generated in the coil, and what is the posi-
concentrated circular
coil of
on a horizontal
axis pointing east

.
tion of the coil

9.
when the
e.m.f. has its
maximum
value?
coil of insulated
is
a current
mounted
wire having a resistance of 200 ohms and carrying in a glass tube through which is passed a stream of

ing air
41
water at the rate of 500 cu. cm. per min. The temperature of the surroundThe is 30 C., and the initial temperature of the water is 20 C. current is then adjusted until the temperature of the outflowing water is constant at 40 C. What is the strength of the current? 10. A rectangular coil 30 cm. by 60 cm. and having a single turn is placed Parallel in the magnetic meridian plane with its 60 cm. sides horizontal. to it, and 40 cm. distant, there is placed a circular coil of one turn having a radius of 25 cm., the centers of both coils lying on a line perpendicular to
the line joining their centers a current of 10 amp. flows in the must flow through the circular coil in not be deflected? What must be the the two coils? long having concentrated poles of 200 c.g.s. units at its ends is placed on the axis of a one turn circular coil of The north pole of the magnet lies nearest to the plane of the radius 25 cm. The coil carries a current of 20 coil and is originally 10 cm. away from it. amp. flowing in a counter-clockwise direction when viewed from the magnet. How much work must be done to carry the bar magnet along the axis to a symmetrical position on the other side of the coil? 12. A circular coil of 10 turns and radius of 20 cm., carrying a current of The 10 amp., is placed so that its plane lies in the magnetic meridian.
their planes.
is
Midway between them and on

If
placed a small compass needle. rectangular coil, how much current order that the compass needle may relative directions of the currents in 11. A slender bar magnet 20 cm.
and direction of the earth's magnetic field being as specified in Problem 1, compute the amount of work required to turn the coil through 90 deg. about a vertical axis through the center of the coil. 13. A cast-iron ring has a circular cross-section of 1 inch diameter and a
intensity
of 10 in. Haw many ampere-turns are required to produce a flux of 30,000 maxwells? Compute the permeability and reluctance of the ring. 14. A magnetic circuit made of sheet steel punchings is built up to the
mean diameter
dimensions of Fig. 22a. The net thickness of the core is only 90 per cent, of the gross thickness because of scale and air spaces between the punchings. Find (a) the number of ampere-turns to produce a flux of 320,000 maxwells;
(6)
the reluctance of the sheet steel;
(c)
the reluctance of the air-gap;
(d)
the permeability of the steel. 15. Three sections of a storage battery are connected in parallel with each other and supply current to a circuit whose resistance is 2 ohms.
Each
section of the battery contains 24
cells,
but owing to different con-
ditions of charging, one of them h,as an e.m.f. of 50 volts, the second 48 volts and the third 45 volts. Their internal resistances are, respectively, 0.1,
0.12 and 0.15 ohm.
battery.
Find the current supplied by each section of the current would flow through each section if the external circuit were disconnected?
What
16. Compute the self -inductance of the cast-iron ring of Problem 13 assuming that the winding has 500 turns. If the current has such a value that it produces a total flux of 30,000 maxwells, what e.m.f. would be induced if the current and flux are reduced to zero, at a uniform rate, in 0.01 sec.?
42
17.

Two
circular coils, current in
one inside links with B; and a current in B produces a flux of which 90 per cent, links with A. When the two coils are connected in series so that they magnetize in the same direction the self -inductance is found to be 0.3 henry; when they magnetize in opposite directions the self-inductance is 0.05 henry. Find (a) the self-inductance of A and B] (b) the mutual inductance of A and B', (c) the amount of work required to turn coil B through 180 deg., starting from the position in which the two coils magnetize in the same direction, assuming that they are connected in series and are carrying a current of 50
the other.
A and B, are mounted concentrically, A produces a flux of which 70 per cent,
amp.
parts,
18. If the cast-iron ring of Problem 13 is split into two semicircular what is the pull, in pounds, required to separate the two halves?
is
What
they have been separated

before?
the magnitude of the pull between the two parts of the ring when He -> the excitation remaining the same as
CHAPTER
II
THE DYNAMO
Generator and Motor. A dynamo-electric 31. Dynamo, machine, or a dynamo, may be defined as a machine for the conversion of mechanical energy into electrical energy, or inversely, for the conversion of electrical energy into mechanical energy.
When
used for the
first
it is
purpose
it is
called a generator,
and when
In other words, the word dynamo is a generic term which includes the other two; a dynamo is a reversible machine, being capable of operation either as generator or motor.
used for the second
called a motor.
Every generator consists of a conductor, or set of conductors, subjected to the influence of a varying magnetic field, so that Current will be produced when e.m.fs. are induced in them.
the circuit of these active conductors
external receiver circuit.
sults
is
completed through an
when
the other hand, motor action recurrent from some external source is sent through
is
On
a set of conductors located in a magnetic field. In the case of generator action, each conductor an induced e.m.f. of
the seat of
# =
where
I
Hlv ab volts
(1)
is
the length of the wire in centimeters,
H is the intensity
of the field
through which it is moving, and v is its velocity in centimeters per second in a direction perpendicular to that of the
field
and to
flow a current
upon Ohm's law.
On closing the circuit there will length. say, I abamperes, the value of which will depend the resistance of the circuit as a whole, in accordance with
its
own
of,
The conductor
will
then be acted upon by a force of
F =
HU
dynes
(2)
in a direction opposite to its motion; hence, to maintain the action, a driving force must be applied to the conductor and work
must be done
at the rate of
Fv
43
Hllv
= El
ergs per second.
44
PRINCIPLES OF DIRECT -CURRENT MACHINES,
In the case of motor action, each conductor is caused to carry a current of / abamperes, so that it IF acted upon by a lateral
thrust of
F = Hll
motion of the wire
intensity
results,
v
dynes
of the field
and under the influence

there
is
and velocity
induced in the wire an e.m.f.
E =
flow there
Hlv abvolts
in a direction opposite to the current.
To maintain
work
the current
must be impressed an
e.m.f. of sufficient
to balance this counter-generated e.m.f., and the electrical source of supply at the rate of
is
El =
magnitude done by Hlvl = Fv
ergs per second. In this discussion ideal conditions have been tacitly assumed, namely, that all of the energy supplied reappears as useful energy
after the conversion process has
fact this condition
been completed. As a matter of never realized in practice; the energy supplied must be greater than that usefully converted by an amount equal to the loss of energy inevitable in the conversion. The armature of a dynamo is the part in which the e.m.f. is generated in the case of a generator, or the part which carries the working current in the case of a motor. The field member is the The relative motion of part which produces the magnetic field. the one structure with respect to the other is most easily obtained by making one or the other rotate, so that in general the two have concentric cylindrical forms. Either may be the rotating member if the armature rotates, the machine is called a revolving armature machine, while if the field rotates it is called a revolving field
is
;
machine.
There are two distinct types of dynamo-electric machines, according to the nature of the e.m.f. and current produced they are (1) alternating -current machines, and (2) direct-current machines.
;
The
first type, when used as a generator, is called an alternator and produces an e.m.f. which acts alternately in opposite directions, so that when the armature circuit is completed the current in the cirThe second cuit flows first in one direction and then in the other. type produces a current through the external circuit which flows in one direction only. A direct current, though characterized by
constancy of direction, may, however, vary in magnitude from
THE DYNAMO
45
instant to instant, that is, it may be pulsating; or it may be conIn the former case, stant in magnitude as well as in direction. the current is said to be a direct current; in the latter case the cur-
rent
is
said to be a continuous current.
form
alternating-current generator or motor is the simplest dynamo. Reduced to the most elementary type, it consists of a loop of wire, abed, Fig. 27, rotating in a magnetic
The
of
It is understood to pole S. field that passes across from pole and S are the extremities of the field structhat the pole pieces ture, and that the excitation of the magnets is effected by a direct
the
current from some suitable source circulating in coils wound on The ends of the armature coil are attached field structure.
.
In the position shown in the to the insulated slip-rings n, r 2 in it an e.m.f. directed from will wire ab have generated figure,
FIG. 27. Elementary alternator.
FIG. 28.
Multipolar revolving
field alternator.
front to back, while the e.m.f. in cd will be directed from back to front; the collecting brush touching ring r\ will therefore be positive,
tion
r 2 will be negative. After half a revolube seen that the polarity of the terminals reverses, so that each terminal is alternately of opposite polarity. In practice, alternating-current machines usually have more
it
and that touching

will
than the two poles shown in Fig. 27; in other words, they are The winding consists of a number of coils connected multipolar. in series in such manner that the e.m.fs. of the individual coils
add together.
Fig. 28 represents diagrammatically
an 8-pole
46
revolving field machine with the winding of the stationary armature arranged in eight slots. Fig. 29 is a development of this particular type of winding as it would appear if the cylindrical
surface of the armature were rolled out into a plane. With the exception of the homopolar machine described in
Art. 50, all
standard forms of direct-current generators and
FIG. 29.
Developed armature winding

of
of alternator of Fig. 28.
motors consist
a wire- or bar-wound armature arranged to
rotate between inwardly projecting poles of alternate polarity, in the manner illustrated in Fig. 44. Each of the armature
conductors
is,
changes
its
therefore, the seat of an alternating e.m.f. which direction each time the conductor moves from the
influence of one pole to that of the adjacent pole. It is the function of the commutator to convert this internal alternating e.m.f.
into a uni-directional e.m.f. in the external circuit; but so far as
FIG. 30. Elementary alternator.
FJG. 31.
Alternator armature
in a general position.
the armature winding itself is concerned, every direct-current machine (with the exception of the homopolar machine) is essentially
an alternating-current machine, hence it is important to analyze the development of the e.m.f. in an alternator in order to understand thoroughly what is happening in the case of the direct-current machine.
32. E.M.F. of Elementary Alternator. Consider first the elementary alternator of Fig. 30, whose armature winding consists
THE DYNAMO
of a concentrated coil
47
having
conductors (or
N=
v
-^
on z turns)
the external periphery of the amature core A. Let it be assumed that the poles N, S are so shaped that the magnetic flux issuing
from them passes straight across the air-gap between the pole shoes and armature core, as indicated by the light horizontal lines, and that the field in the gap is uniform and of intensity H Further, let the armature rotate with a speed of n revogausses. The generated e.m.f. may then be found lutions per minute.
in either of the following
(a)
ways
7?
The
peripheral velocity of the conductors
is
V =
ird
cm. /sec.
(3)
and at any general .instant, when the coil has moved 6. degrees from the vertical (Fig. 31), the component of velocity perpendicular to the direction of the flux
v
is
71
= V
sin
ird 7.- sin
(4)
FIG. 32.
Time
variation of E.M.F. of elementary alternator.
The instantaneous
e
e.m.f. per
conductor
is
then
Hlv
10- 8
irdlH -^sinfl bU
it is
10~ 8
and
for the entire
Z
e
conductors
TrdlHZ
i sin
=
10~ 8
(5)
But dlH
e
is
the entire
flux,
<l>,
that passes from pole to pole, so that

27rJV<J>~sin
t)U
7rZ<J>^sin uU
10~ 8
10~ 8
(6)
This equation shows that the e.m.f. varies from instant to
48

The graph
is
instant according to a sine law.

sine curve
of this equation
is
whose maximum value
Emax =
as
irZ^^Q
10- 8
2irN3>
~X
10~ 8
(7)
to 6 = IT, period from 9 = corresponding to a half revolution of the armature, is found by dividing the area of one loop of the curve of Fig. 32 by the base;
shown in Fig. 32. The average e.m.f. during the
or
Eaver -
ede
Z3>
10~ 8
j
sin OdO
7?
X
x
10~ 8
(8)
10- 8
>
(6)
In the general position of Fig. 31,
tn*b flux
linked with the
coil is
9
hence,
= $
cos
(9)
is
by equation
e
(17),
Chap.
8
I,
the instantaneous e.m.f.
= -
N~ X 10~
at
e
= N$sin6>^ X at
is
10~ 8
But
-rr is
the angular velocity, and this
equal to 2x
hence
= 2wN3>
)
sme X
1Q- 8
which
(c)
is
the same as equation (6) above. By definition, the average e.m.f. induced in a conductor
is
.
equal to the number of lines of force cut per second, divided 8 Now, in one revolution each conductor cuts 2$ lines, by 10
hence the entire

71
conductors cut
second.
2$Z
lines per revolution
and
2$Z^
Eaver =
33.
lines
per
The average
e.m.f.
is
therefore
identical
2$>Z
~X
(8).
10~ 8
= 4N
^Q Q X
10- 8
which
is
with equation
In the preceding article the direction of the flux in the air-gap was assumed to be everywhere parallel to the axis of the pole pieces. This is equivcilent
Induced and Generated E.M.F.
to the assumption that the intensity of the field at each point on the periphery of the armature, in the direction of
THE DYNAMO
49
the radius, is proportional to the sine of the angle between the particular radius considered and the axis of reference (the vertical Such a distribution of flux is said to be sinuline in Fig. 31).
soidal.
If
the armature surface
is
developed into a plane, as
indicated in Fig. 33, the strength of field at each point of the periphery will be represented by the ordinate of the sine curve
whose
maximum
the
coil is
When
each
value, opposite the center of the pole, is H. in the position shown by the full lines (Fig. 33),
cutting the field at the greatest rate, and the When the e.m.f. generated in the conductors is a maximum.
coil
edge
is
FIG. 33.
Distribution of field intensity at armature periphery.
has reached the dotted position, the e.m.f. is zero. It is important to note that in the first instance the total flux linked with the coil is zero, as shown by the equality of the hatched
coil
positive
and negative areas;
so that
if
the armature coil were
stationary in this position, a change in the magnitude of the polar flux would not induce e.m.f. in the coil, since half of the
one direction and the other half in the opposite direcit is advantageous .to distinguish between and induced e.m.f. e.m.f., the former being due to the generated motion of a conductor across magnetic lines of force, the latter As a to a change in the total flux linked with a closed coil. further instance of the distinction between these two views of the matter, it will be noted that in the dotted position of the coil the generated e.m.f. is zero, whereas if the coil were stationary in this position the induced e.m.f., due to a given change in polar flux, would be a maximum, since all of the flux through the coil is then in the same direction.
flux acts in
tion.
For
this reason
50
34.
General Case of the E.M.F. of an Alternator. The discuswas based upon the assumption of a bipolar field structure, a sinusoidal flux distribution, and a full-pitch armature coil, that is, a coil spanning the arc from center to center of poles. Generally, however, there is more than a single pair of poles, the flux distribution may depart considerably from the sinusoidal, and the coil spread may be greater or less than the
sion of Art. 32
pole pitch.
Let Fig. 34 represent a partial development of an alternator
having p poles
(like Fig. 28),
and
let
the distribution of flux at
FIG. 34.
Multipolar alternator, non-sinusoidal flux distribution.
the armature surface be represented by curve further, let the armature have a diameter of d cm., the conductors have an active
;
length of I cm. in a direction parallel to the shaft, and let the speed of rotation be n revolutions per minute. The instantaneous
e.m.f. generated in each
conductor
10-
is
HIV X
Hlird
6Q
10"
(10)
and the graph of this e.m.f. will then be a curve which is the same as that showing the flux distribution, except for a change in scale.
The average
e.m.f. per
conductor
is
w
<
if"?, edx
Jo
= C Hldx
I
= - V.>f d\ Hldx X Jo
10~ 8
PQQ& X
10- 8
(ID
T
where
is
the flux per pole.
The
last result
might
Jo
have been anticipated Trom the fact that the average

equal to the
10 8
;
<
e.m.f. is
number of lines of force cut per second, divided by thus each conductor in one revolution cuts lines per pole,
THE DYNAMO
77
51
or
p$
lines per revolution,
hence
P^QQ
lines per second.
It is
interesting to note that
conversely,
If
is
the
is
first
$ is the integral of the function; derivative of the flux function.
the armature
series as in Fig. 29, in
wound with Z conductors, all connected in such a manner that the coils are of full pitch,
is
the total average e.m.f.
Eaver =
pQZX 10-
(12)
which becomes identical with equation (8) if p = 2. As pointed out above, each conductor is the seat of an e.m.f. whose variation from instant to instant is represented graphically by a curve identical (except for a change of scale) with the curve If the conductors are arranged as in of flux distribution, Fig. 34. Fig. 29 so that the coil spread is the same as the pole pitch, the e.m.f. in all conductors will be simultaneously in the same phase
FIG. 35.
E.M.F.
in coil of fractional pitch.
and the total instantaneous e.m.f. will be simply times that of a single conductor. But if the coil spread differs from the pole pitch, as indicated by coil ab, Fig. 34, the instantaneof the variation,
ous e.m.fs. of the two sides of the coil will differ in phase, that of coil-edge a following curve a, Fig. 35, and that of coil-edge b following curve b. The total instantaneous e.m.f. of the coil is obtained by adding the ordinates of the individual e.m.f. curves. It is evident from Fig. 35 that the maximum e.m.f. of such a
52
PRINCIPLES OF DIREC T-C URREN T

is less
MA CHINES
"short-chord" winding
the same a and
than that of a full-pitch winding of
number
of conductors.
If the terminals the elementary alternator of Fig. 27 are connected, respectively, to the two insulated segments of a
35. Rectification of
c of
an Alternating E.M.F.
coil of
the armature
commutator
as to
tionary brushes,
C, as in Fig. 36, and staare mounted so 61, 6 2

,
make
sliding contact with the re-
volving commutator segments, the plane

of the brushes being coincident with that
through the shaft and the polar axis, brushes 6 2 and 61 will always be of
positive
ively.
coil
and negative
The
polarity, respectreversal of the e.m.f. of the
takes place coincidently with the passage of the brushes across the gaps between the segments. If, then, the FIG. 36. Elementary flux distribution is sinusoidal, the brush direct-current generator. voltage will vary in the manner shown in Fig. 37, that is, it will be uniform in direction, but pulsating between zero and a maximum value. Let the Z peripheral conductors of the coil of Fig. 36 be wound
as in Fig. 38.
Here there are

coils,
still
17
peripheral conductors, but
there are
two
each having
-^
turns, or
turns in
all.
The
part of the winding inside the ring core plays no part in generating The windings of Figs. 36 e.m.f., since it cuts no lines of force. and 38 are therefore electri-
and the latter the same give pulsating e.m.f. as the former.
cally identical,
will
36. Effect
of
Distributed
-
Winding. An e.m.f. or curFlG 37 Rectification of alternating rent varying as in Fig. 37 is not desirable, and means must be found to make it more nearly continuous. The large amplitude of the pulsation of the e.m.f. in Fig. 37 is due to the fact that the entire armature winding of Figs. 36 or 38 is inactive twice during each revolution; if the winding can be so disposed that
THE DYNAMO
small sections of
sations will
sections
is
53
it undergo commutation successively, the pulbecome insignificant when the number of such winding
Let Fig. 39 be a diagrammatic of sketch of the armature Fig. 38, but with the Z conductors in four equidistant groups of concentrated coils conarranged
sufficiently large.
nected to a four part commutator. A study of the directions of the e.m.fs. generated in the coils shows that the brushes must
FIG. 38.
Elementary
ring-
wound armature.
FIG. 39. Ring winding with four sections.
now be
placed in a plane perpendicular to that through the shaft and polar axis, and that the entire winding is now equivalent to two equal halves connected in parallel. Each half, in turn, is
made up
of a pair of
winding sections connected in
series.
Since
the winding consists of two equal halves in parallel, the voltage at the brushes is equal to that of either half alone. Considering the particular half-winding made up of sections A and B, it will
be observed that section
generates a
wave
of e.m.f. similar to
FIG. 40.
E.M.F. generated
in four-coil winding.
that of Fig. 37, but of only one-fourth the amplitude since it has but one-fourth as many conductors as the coil of Fig. 38. (A
sinusoidal flux distribution
is
assumed.)
Similarly, section
,
generates a
wave exactly like that of section A but the two waves differ in phase by 90 deg., as shown in Fig. 40. The resultant brush voltage will be obtained by adding the ordinates of these two component curves, as shown. There are now four pulsations
54
instead of the original two, but the range from minimum to maxiis much reduced. The average voltage at the brushes will
mum
be
Eaver = $Z
or one-half as great as given
10-'
(8) since
(13)
by equation
only Z/2 con-
ductors are in
Similarly,
if
series.
the entire
conductors are grouped in eight equidistant concentrated sections, connected to an eight-part commutator,
made up
in
as in Fig. 41, the winding is of two halves in parallel,
each half containing four sections

series.
Each
of
these
four
A, B, C, D, generates a wave differing in phase by 45 deg. from its neighbor, and the maximum e.m.f. of any one section is
sections,
X
FIG. 41.
10~ 8
The resultant
and
their
Ring winding with
eight sections.
is
brush voltage, as shown in Fig. 42,

in place of the original two,
made up of eight pulsations

is still less
than in the case of the four-section winding. It follows, then, that by subdividing the winding and correspondingly increasing the number of commutator segments, the amplitude of the pulsations may be reduced to any desired extent.
amplitude
27T
FIG. 42.
E.M.F. generated
in eight-coil winding.
The
physical explanation of this fact may be traced to the circumstance that the winding section cut out from the armature circuit during its passage under a brush plays a smaller and
sections
smaller part in its effect is increased.
upon the
total e.m.f. as the
number
of
THE DYNAMO
37.
55
of E.M.F. Pulsations. Assume a ring winding peripheral conductors divided into s sections having Z/s turns each, Fig. 41, and assume also that the magnetic As before, let be the total field has a sinusoidal distribution.
Magnitude
consisting of
and let n be the number of revolutions per minute. Each winding section will then generate a sinusoidal e.m.f. whose maximum value is - $Z X 10~ 8 volts. There are s/2
flux per pole
sections in series in each of the parallel paths through the
arma-
ture winding, and the e.m.fs. in these sections are, respectively,
--
3>Z
sin
10~ 8
E'max sin
-f
es
= # wax
'
sin(0
+ +
2-
e s/2
E' max sin
-- l
E' max sin
Investigation of Figs.
4.0
and 42
of
will
show that a complete
pul-
sation occurs in an interval

e.m.f.
deg.; a -7 s
minimum
6
value of
occurs
when
and a maximum when
=
s
deg.
Substituting these values of 6 in the equation
E=
ei
+ e + e +.
2
3
+ es/2 = #'max
I
sin
+ + sin (0 + -H o
\ /
it
follows that
Emin
E max
r
sin
\
sin
-o
sin(7r \
o / J
= 5
3>Z
~ cotan OLf
S
TT
X
STT
10~ 8
(14)
--
'4sin
sin
-+.
X
10- 8
TT \
"1
m*(,
) J
cosec
(15)
56

variation from
value,
is
7T 7T cotan -
The percentage
of the
minimum
to
maximum
in
terms
minimum
cosec
s
s^
7T
100
cotan -
and the magnitude of shown in the following
this quantity, for various values of
s, is
table:
THE DYNAMO
inside the
57
armature winding. But in general the number of poles, p, may be any even integer, and the number of parallel paths through the armature may be any other even integer, a, as will be explained in detail in the chapter on armature windings (Chap. III). The average value of the generated e.m.f. in such a case may be easily calculated as follows: Each conductor cuts
p$
lines per revolution, or

is
p$ Alines per second, so that the aver-
age e.m.f.per conductor
X 10~ 8 volts; the entire number oU of conductors is divided into a groups connected in parallel, so that there are Z/a conductors in series per group; the e.m.f. per group of conductors, and, therefore, of the armature as a whole, is
X x
a 60
10 s
(17)
This is the general equation for the generated e.m.f. of a direct-current machine, provided the brushes are so placed that the winding sec-
any one group are simultaneunder the influence of one pole. ously if the brushes of the armature Thus,
tions of
of Fig.
43 are so placed that

of
commu-
tation takes place in coils opposite the
middle
tial
the pole shoes, the poten-
difference
zero.
ture
from proper position. ... which are subjected to the inductive action of one pole and the other half to the influence of a pole
.
is
between them will be Each path through the armamade up of conductors half of
...
FlG : 43.
Brushes displaced
of opposite polarity. 39. Resistance of Armature
Winding.
In an armature having
a paths, the total armature current i a will divide equally between them, provided all paths have the same resistance. If the total
resistance of
all the wire on the armature is R a ohms, the resistance per path will then be R a /a ohms, and since all of these a paths are connected parallel, the actual resistance of the arma= ra ohms. ture, as measured between brushes, will be R a /a*
The drop
of potential due to the entire current i a flowing through the resistance ra or ia ra volts, is, of course, equal to the drop of
,
58

any one
of the paths, or
ia /a
potential through
volts.
X Ra /a =
ia ra
40. Construction of
Dynamos.
The dynamo
consists essen-
tially of an electrical circuit and a magnetic circuit placed in inductive relation to each other. The electric circuit consists of the armature winding and the commutator. The magnetic
THE DYNAMO
circuit is
59
made up
of the yoke, pole cores
and pole
shoes,
and
the armature core.
The annular space between the revolving
armature and the stationary field structure is called the air-gap. Other parts of the machine are the field winding, the brushes, brush-holders and the rocker-arm, the armature spider and the bearings. Fig. 44 shows a common arrangement of these parts in the open type jpf construction. Fig. 45 shows a semi-enclosed
type,
and Fig. 46 a totally enclosed motor. The principal structural features of the various parts of the machine, with the exception of the armature winding, are described in the following articles.
The
subject of armature windings
is
taken up
in detail
in
Chap.
III.
FIG. 45.
Semi-enclosed motor
(Sprague).
FIG. 46.
Totally enclosed motor (Sprague).
41. Bipolar
of simplicity
and Multipolar Machines.
Although for the sake
much
is
on the assumption seldom used except in machines
of the preceding discussion has been based of a bipolar field structure, this type of field
of the smallest size.
The
actual
number
of poles generally varies
from four to a
maximum
machines) of twenty to twenty-four, the number the with The capacity, though not at all regularly. increasing of the the of choice the number explanation principles underlying of poles in any given case must be deferred to a later section; in
(in direct-current
general, however, the choice of the
number
of poles
depends upon
the consideration that the magnetic reaction of the armature,
when carrying
current, cannot exceed definite limits without im-
pairing the operating characteristics of the machine.
Further,
60
with an armature core of given dimensions, and with pole pieces that cover a definite percentage of the armature surface, the field frame becomes more compact, up to a certain limit, as the number of poles is increased beyond two. The optimum limit occurs
peripheral spread of the pole faces is approximately to the axial length of the pole face. A compact field frame equal is advantageous in that comparatively little of the field flux leaks
when the
from pole to pole without entering the armature core. 42. The Commutator. The commutator is built up of wedgeof shaped segments drop-forged or hard-drawn copper insulated from one another by accurately gauged thin sheets of insulating
FIG. 47.
Construction of commutator.
The process of assembling a large number material, such as mica. of segments into a rigid structure is an interesting one. The segments, separated from one another by the mica insulation, are placed around the inner periphery of a sectored steel ring, as in Fig. 47, and the copper segments are then wedged together to form a rigid circular arch by means of cap-screws tapped radially through the outer steel ring. The V-shaped grooves are then turned out and the commutator spider bolted into place, after which the auxiliary steel clamping rings are removed and the
external surface turned to true cylindrical form. The insulation between the commutator and the supporting hub consists of molded mica cones and cylinders. The completed
commutator must be giVen a high voltage test to insure the thorough insulation of each segment from the others and from the
THE DYNAMO
spider. to be as
61
The insulation between adjacent segments does not have heavy as that between the segments and the commutator spider, for the latter must withstand the full terminal voltage of the machine while the former is only called upon to withstand The average voltage the smaller voltage between segments. between adjacent segments should not exceed 10 to 15 volts in lighting and railway generators, and from 20 to 25 volts in the These limiting values of average voltage case of railway motors. between segments are imposed by the requirements of sparkless commutation, and they determine the minimum number of segments in the completed commutator. For example, if a 6-pole, 600-volt railway generator is to have not more than 10 volts between adjacent segments, there must be at least 60 segments between adjacent brushes of opposite polarity, or not less than 360 segments in the entire commutator. The minimum diameter of the commutator is then determined if the minimum peripheral width of a segment is known; this minimum width is rarely less than %g in. for two reasons: first, because the taper of the segments would result in too thin a section at the inner periphery if a smaller external width were used; second, because some allowance in the radial depth of the segments must be made to
permit turning down the surface in case of pitting, blistering or wear. The thickness of the insulation between segments varies from 0.02 in. in low voltage machines up to about 0.06 in. in high The material must be so selected that its voltage machines.
is
Amber mica is the same as that of the copper bars. Commutators this used because it meets requirement. largely are sometimes built in such manner that the insulation does not come quite flush with the surface, thereby obviating the necessity of selecting the material for a definite rate of wear. Commutators must be designed to have a sufficient amount of exposed peripheral surface to radiate the heat caused by brush
rate of
wear
friction
The perloss due to brush contact resistance. working temperature of the commutator is from 95 C. 1 The to 130 C., depending upon the current per brush arm. design must provide sufficient mechanical strength to withstand
and the
missible
the centrifugal force.

1
In the case of turbo-generators running at
See Standardization Rules, American Institute of Electrical Engineers,
1914.

high speed the diameter
peripheral velocity shall
is
limited
by the consideration that the
ft. per minute, hence, to secure sufficient radiating surface the commutator must have a considerable axial length. To prevent springing of the segments are in held steel they place by rings shrunk over the segments, and insulated thoroughly therefrom, as shown in Fig. 57.
not exceed 8000
43.
The Armature
Core.
EDDY CURRENTS.
The armature
core not only carries the magnetic flux from pole to pole, but revolves through it in exactly the same manner as the conductors of
the armature winding.
If
the core were solid

of as
might be thought very large number of metallic filaments running parallel to the armature conductors and all connected together; in such a case each filament would
it
made up
of a
be the seat of a generated

FIG. 48.
e.m.f.,
Eddy
current paths in
core.
sketched in Fig. 48. The e.m.f. will obviously be greatest near the surface where the peripheral velocity and the active component of the flux are likewise To minimize these eddy or Foucault currents, which, greatest.
if
solid
armature
and currents would circulate in the mass of the core in the manner
.
..._,.
unchecked, would result in excessive heating and loss of power, the core must be laminated in such a manner as to preserve the continuity of the flux path and to break up the The plane of the laminations must be at all current paths.
points perpendicular to the direction of the generated e.m.f. at those points; or, by Fleming's rule, parallel to the direction of the flux and to the direction of motion. Accordingly, in machines of the usual radial pole type, Fig. 44, the armature
built up of thin sheet steel punchings insulated from each sometimes the insulation consists of a coating of varnish other; on one side of each disk, but generally the oxide, or scale that forms on the sheets, is relied upon to provide the necessary insula-
core
is
tion; in
some designs a layer
of
paper
is
inserted at intervals of
Laminating the core does not completely eliminate eddy currents, but the loss due to them decreases as the square of the thickness of the sheets; the sheet steel ordinarily
used in armature cores
is
an inch or two.
0.014
in. thick.
Armatures
of the
now
THE DYNAMO
63
obsolete disk type, Fig. 49, with active conductors arranged radially, had cores built up of concentric hoops, or, more practically,
of thin strap iron
wound
as a flat spiral.
in.
first
Core punchings up to a diameter of about 16 made in one piece, as in .Fig. 50. The disks are
are generally blanked out
Disk Machine
Spirally
Laminated Armature
FIG. 49.
Lamination of disk armature.
then punched by a special punch press which Core punchings of this sort are generally keyed directly to the shaft, and are sometimes provided with holes near the shaft to form longitudinal ventilating passages. Cores of large diameter are built up of segments which are at-
and the
slots are
cuts one or
more
slots at a time.
tached to the spider by means of a dove-tail joint, as in Fig. 44;

the joints between segments are
staggered from layer to layer in order to preserve the continuity of the magnetic circuit. The
core punchings are held together
by end
flanges which, in the case of small machines, are supported by lock nuts screwed
directly to the shaft; in larger
machines the end plates are held

together by bolts passing through the laminations, but insulated
for the
FIG. 50.
One-piece armature punching.
therefrom, and the end plates are shaped to provide a support end connections of the armature winding (see Fig. 44).
Ventilating ducts through the core are formed by spacing pieces placed at intervals of from 2 to 4
the axis of the core.
means
in.
of
along
The spacing
pieces are generally
made by
64
riveting brass strips, on edge, to a punching of heavy sheet steel, as illustrated in Fig. 51; or they may be made by pressing spherical depressions into a thick punching. The ventilating ducts
to in. The spacing pieces should be so vary in width from as to the teeth as well as the body of the core, designed support in order to prevent vibration and humming.
FIG. 51.
Armature core assembly, showing spacing
pieces.
Fig. 52 illustrates typical forms of teeth
current machines.
special machines. used, except in the case of very small machines, for the reason that they permit the use of insulated, formed coils that can be readily
and slots for directSmooth core armatures are used only in Open slots with parallel walls are generally
slipped into place.
Where
5
semi-closed slots are used, the coils

?
FIG. 52.
Typical shapes of teeth and

jig,
slots.
may
coil
be formed on a winding
but the wires of each side of a
must be slipped
into the slot one at a time.
The
coils are
held in place in open slots either by steel or bronze bandingwires, or by wooden or fiber wedges driven into the recesses at the
tips of the teeth.
The embedding
of the
armature winding
in the slots serves a
THE DYNAMO
65
double function; the air-gap, or distance from the pole face to the iron of the armature core, is less than it would be in a smooth-core construction having the same amount of armature copper, and so
reduces the
amount
of field copper necessary to
produce the
flux;
and the armature conductors are supported by the teeth when subjected to the tangential forces caused by the reaction of the armaWhen the armature conductors ture current upon the field flux. in slots the thus embedded are they are apparently shielded from
the inductive effect of the field flux, since the latter in large measure passes around the slots by way of the teeth. At first
sight, therefore, it seems surprising that the fundamental equation for the generated e.m.f. is the same for a slotted armature as for a smooth-core armature. It must be remembered, however, that
a line of force which at a given instant crosses the air-gap from the pole face to a given tooth tip, must later, by reason of the motion of the armature, be transferred from this tooth-tip to the The line of force holds on, as it were, to the first following tooth.
tooth in the manner of a stretched elastic thread, until the increasing tension causes it to snap back suddenly to the next tooth.
increased velocity of cutting of the lines of force by the conductors exactly compensates for the reduced value of the field intensity in the slot.
44.
erally
The
The Pole Cores and Pole Shoes. The made of cast steel. When cast steel
pole cores are genis
used, the poles
usually have a circular cross-section because this results in minimum length and weight of the copper wire in the field winding. Laminated poles of course require a rectangular cross-section. Solid poles are commonly bolted to the yoke. Laminated poles in be secured either a dovetail may place by joint or may be cast
into the yoke.
The
flux density in the
body
of the pole core,
110,000 lines per sq. in., is considerably greater
running as high as than can be eco-
nomically produced in the air-gap. The average flux density in the air-gap should not exceed 62,000 lines per sq. in., hence the This pole faces must have greater area than the pole cores. increased area is secured by means of pole shoes bolted or dove-
by means of projecting tips or horns punched integrally with the sheets composing a laminated pole. The pole faces or shoes are almost always lamitailed to the core in the case of solid poles, or
66

when
solid poles are used, in order to reduce the loss
nated, even
and heating due to eddy currents set up in the pole faces by the armature teeth; for, as shown in Fig. 53, the flux passing between the pole face and armature core tends to tuft opposite the teeth, and as the teeth move across the pole-face these tufts are drawn
tangentially in the direction of rotation until the increasing tension along the lines of force causes them to drop back to the
next following tooth.
The
tufts of flux are therefore continu-
ously swaying back and forth, and if the pole face is considered as built up of thin filaments, as at P in Fig. 53, each of the filaments
will
tufts first in
be cut by these swaying one direction, then
in the other, thereby inducing an alternating e.m.f. directed
parallel to the shaft.

FIG. 53.
To
mini-
mize the flow of current the pole tips of * face must therefore be laminated in planes parallel to those of the armature laminations, though the laminae of the pole shoes do not have to be made as thin as those of the armature core. These pole-face eddy current losses will obviously be reduced by so proportioning the dimensions
Tufting of flux at
and slots as to prevent appreciable lack of uniformity in the distribution of the flux along the pole face. The determining factors in this proportioning are the ratio of slot opening to airof teeth
gap and the length of the air-gap itself. 45. The Yoke. The yoke is that part of the field structure which carries the flux from pole to pole, and at the same time serves as a mechanical support for the pole cores. It is made of cast iron in small machines and of cast steel in larger sizes, or whenever saving in weight is important. The yoke is usually split on a horizontal diameter for convenience in assembling and In machines of moderate size the yoke is cast as an repairing. integral part of the bed plate in larger sizes it is cast separately, but with lugs for bolting to the bed plate. 46. Brushes, Brush Holders and Rocker Ring. The connection between the revolving armature and the external circuit is made through the brushes^ which are usually made of graphitic carbon, except in the case of low-voltage machines when they may consist of copper or copper gauze. Carbon brushes are
;
THE DYNAMO
made
67
of varying degrees of hardness to suit the requirements of commutation, as discussed in a later chapter. The graphite in
the brush serves to partially lubricate the commutator, which, when fitted with brushes of the proper composition, takes on a polished surface of dark brown color. The width of the brush
from three to five times that several armature a so width of commutator the segment, short-circuited. The carbon brush must coils are simultaneously in limit the current the short-circuited resistance to have sufficient in sparking when result below that which would a value coils to
in the tangential direction is generally
the short-circuits are opened. The brushes are commonly set at a trailing angle with respect to the direction of rotation, though in machines designed to run
in
both directions, such as railway motors, they are set radially. When the tangential width of the brush has been decided upon,
the total axial length of the brushes constituting a set is determined by the consideration that there must be a contact area
every 30 to 50 amperes to be carried by the set, though this current density may be exceeded in the case of interpole machines. The individual brushes of a set must not be too large in cross-section, otherwise there would be
of 1
sq. in. for
brush
difficulty in
making and maintaining a good contact over
its
entire contact surface.
The
subdivision of the set offers the
additional advantage of allowing the individual brushes to be trimmed one at a time without interfering with the operation of
the machine
when under load. Single brushes are used only in the case of machines of small current output. The individual brushes are supported in metal brush holders
which are
in turn
supported by studs attached
to,
but insulated
from, the rocker ring, as illustrated in Fig. 54. The brush holders serve as guides for the brushes, and should allow the brush to slide freely in order that the brush may follow irregularities in
the commutator surface.
must be such that there
construction of the brush holders be no vibration of the brushes, this being a common cause of sparking. Tfye brushes are held against the commutator surface by adjustable springs attached to the holder, but in such a manner that the springs do not carry any current. The tension of the springs is adjusted until the brush presses against the commutator with a force of from 1.5 to 2 pounds
will
The
68
per sq.
PRINCIPLES OF DIRECT -CURRENT MACHINES

in. of
contact area.
this limit does
Increasing the brush pressure above not materially lower the contact resistance, but
increases the sliding friction and, therefore, results in increasing loss of power and heating of the commutator. The connection
between the brush and the brush holder
is
made through
flexiis
ble lead of braided copper wire, called a pig-tail, which tached to the outer end of the brush by means of a metal
at-
band
clamped tightly around the carbon. The carbon is generally copper-plated at its outer end to insure good contact.
FIG. 54.
Brush holder and rocker
ring.
47. Motor-generator. Dynamotor. It is frequently necessary to convert direct current at one voltage into direct current
For this voltage, higher or lower than the first. a As ordinarily constructed, a purpose motor-generator is used. motor-generator set consists of two separate machines, a motor
at
some other
and a generator, direct connected to each other, and mounted on a common bed plate, as illustrated in Fig. 55. Motor-generators are also used to convert direct current into alternating curThis type of machine has the advantage rent, or vice versa.
that the voltage of the generator end may be controlled independently of that of the motor end of the outfit. The over-all
efficiency of the set
is
equal to the product of the efficiencies of
THE DYNAMO
the motor and generator.
in general
69
The power
rating of the motor
must
be sufficiently greater than that of the generator to allow for the losses that occur in the double transformation of
the energy. Instead of using two separate machines, as in a motor-generator set, to convert the current from one voltage to another, it is
combine the two into a single unit, called a dynamotor, a By having single field structure and a single armature core. providing the armature of such a machine with two separate
possible to
FIG. 55.
Motor-generator
set.
windings, each with its own commutator and brushes, current may be introduced into one of the windings, thereby causing motor action, while the other winding will then generate an e.m.f.
This type of machine is built in small sizes only. It is open to the objection that the voltage at the generator terminals cannot be independently regulated, but is fixed by the voltage impressed
upon the motor terminals. The truth of this statement can be seen from the following reasoning: If the e.m.f. impressed upon the motor terminals is E mj the rotation of the armature through the field flux will generate in the motor armature winding an approximately equal and opposite voltage (Art. 31); if there were
<
70
no losses in the motor, this counter e.m.f. would be equal to hence by equation (17)
Em
Em =
am
60
10 8
Since the generator winding rotates through the same field as the motor winding and at the same speed, the generator e.m.f. is
Eg
or
=
a~ g 60
10 8
E
(18)
The disadvantage of the fixed ratio of voltage transformation is by the reduced cost of construction made possible by the armature and field structure. The dynamotor has in addisingle
offset
free
tion a higher efficiency than a motor-generator, from trouble due to armature reaction.
and
is
practically
MB*
FIG. 56.
Turbo-generator
set.
48. Turbo -generators. Generators for direct connection to steam turbines must be designed for high speed of rotation since the steam turbine develops its maximum efficiency under this
condition.
The high
rotative speed calls for special design to
THE DYNAMO
71
mutation.
withstand the centrifugal forces and to provide sparkless comThe end-connections of the armature winding are held in place by metal end-shells in place of the usual banding
ing
and the commutator segments are prevented from springby a steel ring or rings shrunk over them. To provide for satisfactory commutation these machines are provided with
wires,
whose function it is to generate in the coils undergoing commutation an e.m.f. of the proper magnitude and direction to reverse the current in the short time required for the segments to pass across the brush. Fig. 56 represents a 300-kw.,
interpoles (Art. 49)
FIG. 57.
Commutator
of high-speed generator.
manufactured by the General Electric This machine has a double commutator, and the interpoles are clearly seen between the main poles. Fig. 57 shows the steel rings around the commutator of a 125-volt, 125-kw. machine.
125-volt, 1500-r.p.m. set
Company.
Turbo-generators require a high grade of brushes to insure The brushes wear down quite rapsatisfactory commutation.
idly and must be adjusted with great care. 49. Interpole Machines. A full discussion of the function of
The interpoles, interpoles must be deferred to a later chapter. also called auxiliary poles or commutating poles, are small poles
placed midway between the main poles they are wound with coils through which the armature current, or a fractional part thereof, is made to flow. Interpoles are used in machines where sparkless
;
72
commutation would otherwise be difficult or impossible of attainment, as in turbo-generators, variable speed motors, etc. 50. The Unipolar or Homopolar Machine. In the type of armature described in the preceding sections, the individual coils have generated in them alternating e.m.fs. which are then rectified by the commutator; the latter plays much the same part as the valves of a double-acting reciprocating pump. In the centrifugal pump, on the other hand, the developed pressure acts continuously in one direction, thereby obviating the necessity for the rectifying valves, and the electrical analogue of the centrifugal pump is found in the so-called unipolar, or homopolar, or In principle, this acyclic generator, shown in section in Fig. 58.
Load
FIG. 58.
Homopolar or
acyclic generator.
machine consists of a conductor so disposed in a magnetic field that the cutting of the lines of force is continuously in one direcThe armature tion; it is a true continuous-current machine.
consists of a metal cylinder A of low resistance, insulated from the shaft, and upon whose edges the two sets of brushes BI and B 2 make sliding contact. The armature rotates in a magnetic
,
produced by the exciting winding F, the path of the flux being indicated by the dashed line. The lines of force pass radially across the air-gap all around its periphery. If the intensity of the magnetic field in the air-gap is B lines
field
per sq. cm., the axial length of the active part of the cylinder I cm., and its peripheral velocity v cm. per second, the generated e.m.f. will be e = Blv X 10~ 8 volts. The maximum e.m.f. obtainable
with this type of machine is determined mainly by the consideraB and v may not exceed definite limits; the length I is likewise limited by such mechanical features as rigidity, freedom
tion that
THE DYNAMO
73
from vibration, etc. At the high rate of rotation required for any reasonable value of e.m.f., difficulty is experienced in securThus if B = 15,500 (100,000 lines per ing good brush contact. = cm. 2 ft.), and v = 5000 cm. per second I 60 (about sq. in.), (about 10,000 ft. per minute), e = 46.5 volts. Because of the fact that the armature consists of a single conductor of large crosssection, the machine is adapted for heavy currents at relatively low voltage. Unfortunately, however, the magnetizing action of the large armature current when the machine is under load so weakens the field produced by the exciting coil F that the voltage
drops considerably.
The analogy between the homopolar machine and the
centrif-
ugal pump suggests the idea that, just as high pressures may be obtained with the latter by using several stages, higher voltages
may
be obtained with the former machine by using several inseries. Such a machine has been built by the General Electric Co. 1 for 300 kw. at 500 volts and 3000 r.p.m.; and the 2 Westinghouse Electric and Manufacturing Company has built one for 2000 kw. and 260 volts, running at 1200 r.p.m. 51. Field Excitation of Dynamos. In every dynamo-electric machine the generation of the armature e.m.f. depends upon the motion of the armature inductors through a magnetic field. In the earliest types of machines this magnetic field was produced by permanent magnets; such machines are called Their use is magneto-electric machines or, briefly, magnetos. now confined to small machines intended for ringing call-bells
ductors in
in small telephone systems, for gas-engine igniters and for testing The field excitation of all other generators and motors purposes.
is
types of
accomplished by means of electromagnets. field excitation may be recognized:
The
following
Separate excitation
f
Series excitation
Self excitation
Shunt excitation
Compound
excitation
field excitation
52. Separate Excitation.

field
1
In this type of
the
winding
is
traversed
by a current supplied from a source
J.
B. G.
E. Noeggerath, Trans. A.I.E.E., Jan., 1905. Lamme, Trans. A.I.E.E., June, 1912.
74

itself,
such as a storage battery or another of this type are altergenerator. and of lowvoltage directcertain kinds nating-current generators current generators used for electroplating. Fig. 59 represents the connections of such a machine. diagrammatically
external to the machine
The most prominent examples
53. Self-excitation. of
The
use
electromagnets, separately excited for the production of

field, was introduced by Wilde in 1862. A great step in advance was made in 1867 when Werner Siemens
the magnetic
discovered the principle of selfexcitation, whereby the armature

FIG. 59.
current,
in
whole or
in
Diagram
of separately
part,
field
was made to traverse the

winding, thus causing the
excited machine.
machine to develop its own magnetic field. Self-excited machines may be divided into three classes, depending upon
the connections of the
field winding to the other parts of the classes series excitation, shunt excitation and these are circuit;
compound
excitation.
FIG. 60.
Connections of
series generator.
In Fig. 60, A represents the armature All of the of a four-pole machine. field structure S, N, S, the current taken by the external circuit passes through the field
54. Series Excitation.
and N,
in series.
winding and the armature, since all these parts of the circuit are The arrows indicate the direction of the current in the
THE DYNAMO
75
case of generator action and for clockwise direction of rotation of the armature. If the field structure of such a machine is originally unmagnetized, rotation of the armature cannot generate e.m.f., hence there can be no current in the circuit. In order that the machine
self-excite, it is necessary that there be some residual magnetism in the field poles due to previous operation, or, in the case of a new machine, produced by sending current through the field winding from some suitable external source. Assuming, then, that residual magnetism is present, a small e.m.f. will be generated when the armature is rotated, and, upon closing the external circuit through the load, a small current will flow. This current will further excite the field structure, thereby devel-
may
oping more e.m.f. and a still greater current, and so on. This gradual increase of both e.m.f. and current will continue until a condition of equilibrium is reached, this being determined by the degree of saturation of the field magnet and by the resist-
ance of the
circuit, in
a manner that will be discussed fully in

field
the chapter on operating characteristics. It is important to note, however, that if the

reversed the machine will refuse to
"
terminals are
build
up"
as described above.
For
in this case the generated e.m.f. will send a current through the circuit in such a direction as to neutralize the remanent
magnetism.
critical value,
Further, if the resistance of the circuit exceeds a the resultant flow of current may be insufficient to
produce the requisite magnetizing force. From the above description of the process of building up of a series generator, it is obvious that such a machine when running on open circuit (the receiving circuit disconnected) will develop only the small e.m.f. caused by residual magnetism; and that with increasing current, as the external resistance is lowered,
the generated e.m.f. likewise increases, though not in general
proportionally.
The
field
winding of
series
machines consists of relatively few
turns of coarse wire.
Since the entire current, i, delivered by the machine to the receiver circuit must flow through the resistance, 2 r f of the field winding, there occurs a loss of power equal to i r f
,
watts in this part of the circuit. Obviously, this loss must be kept as small as possible in order that the efficiency of the dynamo
76

not be seriously impaired, and since the magnitude of the i is fixed by considerations of the load to be supplied, it
may
current
follows that r/
cross-section
55.
field
must be kept
as small as possible; hence the
conclusion that the wire of the field winding must have large
Fig.
and moderate length. Shunt Excitation. Fig. 61 shows the same armature and frame as Fig. 60, but provided with a shunt field winding. 61a represents the connections in a simple diagrammatic
It is evident that the exciting current
manner.
now depends
FIG. 61.
Connections of a shunt generator.
upon the
e.m.f. at the brushes

it is
and upon the ohmic resistance
of
not dependent upon the resistance of the receiver circuit in the same sense as in the previous case, but only to the extent that variations of the external resistance affect the brush voltage. If the external circuit is entirely disconnected, the
the field winding;
remaining connection between armature, shunt winding and

field
regulating rheostat
is
precisely the
same
as that of a series
On open circuit, therefore, a shunt generator will generator. build up just as a series generator does under load conditions; if
it fails
to do so, it is usually because of one or the other of the reasons discussed in the preceding section. It is clear, therefore, that a shunt generator, unlike one of
the series type, develops full terminal voltage at open circuit, that Supis, when no current is being supplied to the receiver circuit.
pose, now, that the external circuit is closed through a considerable resistance so that a small load current, i, is drawn from the The armature current, which was originally equal to generator.
THE DYNAMO
is
77
i), and the effect of this increased curalone, now becomes .(i rent through the ohmic resistance of the armature is to cause a
drop of terminal voltage; this in turn results in a decrease of the exciting current, i a and consequently also of the magnetic flux
,
and the generated e.m.f. As the load current becomes greater and greater the terminal voltage therefore becomes less and less. It is clear that the drop of voltage will be minimized if the resistance of the armature is kept low. The drop of voltage under load conditions is also affected by armature reaction and by the
degree of saturation of the magnetic circuit. A complete discussion is given in Chap. VI. The field winding of a shunt machine consists of many turns of
fine wire, for
the following reason

t
If
the terminal voltage of the
machine
power
is
volts, the
shunt
will
field current, i a , will
be
E
is
t
,
and the
Ts
loss in the
winding
be
2 is r s
E = -
2
t ;
since
rs
fixed
by
other considerations, it follows that r s must be as large as is feasible (in order to keep down i s and the loss of power) hence the
use of wire of small cross-section and great length. The relation between the armature current i a the line current
,
and the shunt
field
current
i8 ,
in the case of generator action, is
given by the equation

ia
+
+
is
(19)
is
In the case of motor action the relation

i
obviously
(20)
ia
is
It should be remembered that the armature and field currents of a shunt motor do not divide in the inverse ratio of their respective
resistances, for the reason that the armature, when running, is the seat of a counter-generated e.m.f. The field current is given by
is
=
56.
Tji
ET
,
ET
but the armature current
is i a
Ta
,
where
Ea
is
7*8
the counter e.m.f.
Compound
Excitation.
In some of the most important
applications of direct-current machinery, such as incandescent lighting, street-railway operation, and the like, it is necessary to
maintain a constant difference of potential between the supply mains no matter what the load may be. Since the center of the
78
load
is
usually at a distance from the generator, it follows that the potential difference between the generator terminals should rise as
the external current increases, in order to compensate for the drop Field windings adapted to give of potential in the supply mains.
this characteristic are called
compound windings,
illustrated in
They are Fig. 62 and diagrammatic ally in Figs. 63a and 636. combinations of shunt and
series
field
nections
made
Conwindings. in accordance
with Fig. 63a result in a shortshunt winding, those of Fig.
636 in a long-shunt winding. The shunt winding of itself would produce a " drooping"
characteristic, that
is,
one
in
which
falls
the
terminal
e.m.f.
FIG. 62.
Connections of a compound
generator.
with increasing current, as explained in the preceding
section; but the series wind-
ing contributes field excitation which increases with increasing current, hence the resultant effect will depend upon the relative
magnitudes and directions of the magnetizing actions of the two windings. By properly proportioning them, the voltagecurrent curve may rise, in which case the machine is said to be over-compounded; or the voltage may remain very nearly confield
(a) Short Shunt FIG. 63.
()
Long Shunt
Diagrammatic scheme
of connections of
compound machines.
stant for
;
all
machine or
permissible values of current, as in a flat-compounded it may fall at a greater or lesser rate than with the
rise to
shunt winding alone, giving compounded machines.
the classification of under-
In the short-shunt compound-wound generator the relation
THE DYNAMO
between armature current
rent
i, is
79
i,
ia ,
line current
and shunt-field cur-
given by
ia
is
The terminal
e.m.f.,
t,
and the generated

t
e.m.f.,
Ea
are related
by the equation
Ea = E +
and the shunt-field current
i,
*>/
tr
(21)
is
given by
(22)
^-^ = il+i*
Ts
fs
In the long-shunt
compound-wound generator
these relations
become
= i + is E a = E + ir,
ia
t
,
z>
(23)
f;
(24)
In designing the field 57. Construction of Field Windings. windings of shunt, series and compound machines, the selection pf the correct number of turns and the cross-section of the conductors follows from a knowledge of the number of ampere-turns per pole required to produce the flux <f>, and from the dimensions
of
the pole core.
The
calculation of these latter quantities

.
depends upon principles that are discussed in detail in Chap. IV. Assuming that the number of ampere-turns per pole and the dimensions of the pole core are known, the determination of the size of wire to be used is as follows Let = current in shunt winding is E = terminal voltage at no load
:
er
voltage consumed in regulating rheostat, varying from 10 per cent, to 20 per cent, of E
t.
The object of the field rheostat is to permit an increase of i 8 by cutting out a part or all of the regulating resistance, thereby raising the generated e.m.f.
The
resistance of the winding per pole
is
then
80
where
p
specific resistance of
copper at the working temperaC.)
ture of the winding (75
na
It
= number
shunt turns per pair of poles mean length of a turn

of
A =
If
area of cross-section of conductor.
lengths are expressed in feet and cross-sections in circular = 12.6 at 75 C. Hence mils, p
A =
The mean length
of
.,
circular mils.
is
(26)
of a turn,
1 to
lt
lt,
found by assuming a depth

of
winding of from
is
in.
If the cross-section
the pole
core
rectangular,
will
be approximately equal to the perim-
()
(6)
FIG. 64.
Ventilated field
coils.
eter of the core plus four times the winding depth; if the pole core is circular, of diameter d c It ir(d c winding depth).
,
The winding depth must not exceed a
definite limit, otherwise the
heat generated in the interior of the core cannot be readily conducted to the surface. As a check on the calculations, it must
be ascertained that the power lost in the coil (t*r/) does not exceed approximately two-thirds of a watt per sq. in. of exposed
radiating surface. Shunt coils are usually made of cotton-covered wires, of either round or rectangular section. Sometimes they are wound on
THE DYNAMO
;
81
are
metal frames arranged to slip onto the pole cores sometimes they wound on removable winding forms, the coils being held in
shape by suitable insulating materials and dipped in, or painted When metal frames are used, they with, moisture-proof varnish. are frequently made with a double wall to allow the circulation of
between pole core and winding, as shown in Fig. 64. The coils motors are usually impregnated with then taped and varnished. The series coils insulating compound,
air
of series-wound railway
FIG. 65.
Interpole machine, edge-wound copper strap
coils.
of
compound and
interpole machines are often
made
of copper
strap, wound on edge, the turns being separated of insulating material, as shown in Fig. 65.
by distance
pieces
In order that connections may be easily made between the coils of adjoining poles, the terminals of the coils are brought out on
opposite sides, so that the
plus one-half. 58. Field Rheostats.
number
of turns per coil
is
an integer,
To permit regulation of the voltage of shunt and compound generators, the current in the shunt-field winding must be under control. To this end a variable resist-
82
ance, QY field rheostat, is inserted in series with the shunt winding, as shown diagrammatically in Figs. 61 and 62. This resistance
arranged in the manner shown in brought out from the high resistance wire or ribbon composing the resistor to a series of insulated studs over which moves an adis
Fig. 66, taps being
i i f that clockwise rotation of the regulathandle increases the resistance in circuit and so throttles the ing current in the manner of an ordinary valve.
, .
FIG. 66. Diagram of connections of field rheostat.
The terminals justable contact arm. are always brought out in such a way
,,
i
Field rheostats are generally arranged to be mounted
on the back of the switchboard, with the regulating handle on the front of the board.
field
Fig. 67 represents a
made by the General Electric Company.

rheostat
Field rheostats for machines
of large capacity are made of cast-iron grids, as shown
in
Fi
fiS
FIG. 67.
Field rheostat, back connection.
In shunt and compound generators of large capacity the energy stored in the magnetic
field
is
very considera2
t
amounting to %L a i, where L s is the inducble,
tance of the shunt winding.
The inductance
have a value of several hundred henries. For instance, if L s = 600 and i s = 4, the energy
may
FIG. 68.
Large
field rheostat.
stored
in
the
field
is
4800 joules. If the field circuit is abruptly broken, this energy will have to be dissipated in the arc formed on breaking the circuit; if, for example, the
THE DYNAMO
current were
83
made
to disappear in one-half second, the average
rate of energy
dissipation would be 4800 -f- J/ = 9600 watts, and the average voltage induced by the collapse of the magnetic
field
would be
600
4800
volts.
In this case
the arc would be very destructive, and the high induced voltage would be likely to puncture the insulation of the winding. To
obviate this danger the field current must be gradually reduced In large machines the reduction of before breaking the circuit.
field
current is accomplished by allowing the field windings to discharge through a field discharge resistance, connected in the
manner shown
in Fig. 69.
1
Exciting Circuit
Discharge Eesistance
Carbon Break Contacts
Field
OOOOOOQO
FIG. 69.
Diagram
of connections of field discharge resistance.
59. Polarity of Generators. In order that a self-exciting generator of any of the types already described may be operative, it is necessary that there be some remanent magnetism in the field
system further, that the initial flow of current through the exciting winding have such a direction that it will strengthen the remanent field. In other words, the polarity of the machine is determined by that of the remanent magnetism. For example, consider the conditions existing in the shunt;
wound generators illustrated in Figs. 70a and 706, respectively. The machines are identical except that the remanent magnetism of the second is reversed with respect to that of the first. In both cases the machine will build if the direction of rotation
plockwise, but with the polarity of the terminals of the one opposite to that of the other. With the connections as shown,
is
84
PRINCIPLES OF DIRECT-CURRENT MACPIINES
counter-clockwise rotation would set up a field current which would wipe out the remanent magnetism; but with counterclockwise rotation the machines would again become self-excitif the terminals of the field winding are interchanged. In Figs. 70a and 706, the armature winding is right-handed, that is, it is wound around the core in the manner of a right-handed
ing
FIG. 70.
Effect of reversal of residual magnetism.
armature windings had been used remaining the same, annulment of the remanent magnetism would again be the result. Finally, it is clear that the direction of the winding around the poles plays a
screw thread.
If left-handed
(Fig. 71), other conditions
similar role.
There are therefore four elements which

such a machine
:
The
effect the polarity of sense of the windings of armature and pole pieces, respectively; the direction of rotation;
and the order
of
connections
of the
field
winding terminals to the armature terminals. With a given remanent magnetism, the machine will operate only
relation
FIG.
Left-
when
there
is
a definite
between them.
Assuming that the
71.
in
the machine to lose its residual magnetism, but a change in any two of them will not affect the operation. Thus, a right-handed armature rotating clockwise in a given field flux will yield the same brush polarity es a left-handed armature rotating counter-clockwise in the same held. In general, a change in an odd number of the four elements
armature.
handed ring wound
conditions for operation are satisfied, a change any one of these four elements will cause
THE DYNAMO
will disturb
85
they were previously correct, while a change in an even number of them will not affect the operation. The same types of field 60. Direction of Rotation of Motors. windings and connections as are used for generators find equal
conditions
if
Series motors, when supplied with constant terminal e.m.f., fall off rapidly in speed as the load " increases, or, to put it in another way, race" as the load is re-
application in the case of motors.
moved;
e.m.f.
is
this characteristic of variable speed at constant terminal
mirror" image of the series generator charThe acteristic, namely, variable voltage at constant speed. speed characteristic of the series motor adapts this machine to street railway and hoisting service. Again, the shunt motor, when supplied with constant terminal e.m.f., operates at praca sort of
"
Flux
FIG. 72.
Diagrammatic sketch
FIG. 73.
Diagrammatic sketch
of series generator.
of series motor.
tically
constant speed at
all
delivers a nearly constant
machine capacity) when

creasing load,
if
loads, just as the shunt generator terminal voltage (within limits of driven at constant speed. An over-
compounded generator used
as a
motor
will rise in
speed with in-
supplied from constant potential mains (see
Chap. VII). Let Fig. 72 represent diagrammatically a series machine used as a generator, the shaded half of the armature circle representing a belt of current flowing into the plane of the paper and the unshaded half representing current of opposite direction. If this machine is now connected to mains of the polarity indicated in Fig. 73, and is operated as a motor, its direction of rotation will be reversed as may be seen by applying (the left-handed) Fleming's rule.
This means that a series generator supplying a network fed by other generators may, if overloaded, tend to reverse its direction of rotation and so buckle the connecting rod of the
86
The fundamental reason for the reversal of driving engine. direction of rotation is that both armature current and field current reverse simultaneously. If only one of these were reversed, the direction of rotation would remain unaltered. If, however, the motor, Fig. 73, is supplied with current from mains of reversed
To
polarity, the direction of rotation will be the same as before. reverse the direction of rotation of a motor it is necessary to
reverse the connections of either the armature or the field wind-
not those of both. In the case of two identical shunt machines, one used as a generator and the other as a motor, as in Fig. 74, the direction of rotaing,
FIG. 74.
Showing direction
of rotation of
shunt generator and motor.
A shunt generator supplying a netnot reverse its direction of rotation if its will, therefore, mover is or disconnected shut prime down, but will continue to run as a motor. if the Further, polarity of the line supplya shunt motor is direction of rotation will not the ing reversed,
tion
is
the same in both.
work
be affected since both field and armature current reverse simultaneously. The direction of rotation of a shunt motor can be reversed by reversing the connections of either the armature
winding or the
field
winding separately.
PROBLEMS
ing 50 cm. square.
100 turns is wound on a wooden frame measurrotated at a uniform speed of 1000 r.p.m. about an axis passing through a diagonal of the square. If the coil is in a uniform magnetic field of intensity 200 gausses, whose direction is at right angles to the axis of rotation, what are the maximum and average values of the generated e.m.f.? What are the positions of the coil with respect to the direction of the field when the instantaneous e.m.f. has (a)
1.
concentrated
coil of
The
coil is
its
value, (6) a value equal to the average e.m.f.? the square coil of Problem 1 is replaced by a circular coil of the same number of turns and a diameter such that it encloses the same area, what
2. If
maximum
THE DYNAMO
will
87
unit length of wire in
be the average e.m.f.? Problems 1 and 2.
Compare the voltage per
10 6 lines per 3. The 8-pole alternator of Fig. 28 has a field flux of 4 pole, distributed sinusoidally around the periphery of the stationary armaEach of the 8 slots contains 20 conductors, all conductors being conture.
nected in
series.
If
and maximum values
the speed of rotation is 900 r.p.m., what are the average of the generated e.m.f.?
4. The alternator of Fig. 28 is provided with pole shoes that cover twothirds of the armature surface, and they are so shaped that the flux of 4 10 6 lines per pole crosses the air-gap along uniformly distributed radial
lines.
If
the
as in
Problem
number of armature conductors and the speed are the same 3, what are the average and maximum values of the generated
e.m.f.?
Construct a curve showing the variation of the e.m.f. from instant
to instant.
5. A ring-wound armature like Fig. 41 has 400 conductors distributed uniformly on its periphery, and rotates in a 4-pole field structure that 6 produces a flux of 1.5 X 10 lines per pole. At what speed must the armature rotate to develop an e.m.f. of 120 volts? 6. If the total amount of wire on the armature of Problem 5 consists of 500 ft. of No. 16 B. and S. wire, which has a resistance of 4.085 ohms per thousand feet at 75 F., what is the resistance of the armature measured between brushes? 7. The commutator of a machine which runs at 650 r.p.m. has a diameter There are four sets of brushes, each set consisting of three carbon of 18 in. brushes; each individual brush has a width of 1.5 in. and a contact arc of 0.25 in. The contact pressure is 1.5 Ib. per sq. in. of brush contact area. If the coefficient of friction of carbon on copper is 0.3,. what is the brush friction loss, expressed in watts? 8. A small series- wound generator has a normal rating of 110 volts and 10 amp. and its field winding has a resistance of 1.4 ohms. If the machine is to be used as a separately excited generator, the field current
being supplied from 115- volt mains, how much resistance must be put in series with the field winding to produce normal field excitation? 9. A 220-volt shunt motor takes a field current of 1.4 amp. and, when
running without load, an armature current of 1.5 amp. It is found by experiment that the full-load armature current of 25 amp. can be made to flow through the armature, under standstill conditions, by imWhat are the resistances pressing 11 volts upon the armature terminals. of the field and armature windings, and what is the counter-generated e.m.f. when the machine is running under no-load conditions? 10. The field structure of the motor of the preceding problem has 4 poles and the winding is made of No. 18 wire which has a resistance of 7.76 ohms per thousand feet at the working temperature. Each field coil is wound on a cylindrical bobbin and has a mean diameter of 7 in. Find
the
number
of turns per coil.
CHAPTER
III
ARMATURE WINDINGS
may
Types of Armatures. Armatures, considered as a whole, be divided into three classes according to the shape of the core and the disposition of the winding upon it. These three
61. classes are:
1.
RING ARMATURES.
2.
DRUM ARMATURES.
DISK ARMATURES.
3.
The ring armature is one in which a ring-shaped core is wound with a number of coils, or elements, each of which winds in and out around the core in helical fashion, as in Figs. 38, 39, etc. In these windings the coils are usually connected successively to
each other so as to form a continuous circuit, but this feature is not essential to the definition. The characteristic feature of
ring windings
is
which do not cut
that there are conducting wires inside the ring lines of force, and which do not, therefore, con-
tribute to the e.m.f.
dead wire
The drum armature was developed to reduce the amount of in a ring winding. Each active wire, wound on the
outer surface in a direction parallel to the shaft, is connected to another active wire by means of connecting wires which do not
thread through the core. The only reason for having any openIn ing in the core at all is to permit ventilation and cooling. bipolar machines the end connections run across the flat ends of the core and join conductors which are nearly diametrically In multipolar machines they join conductors separated opposite.
by an
interval
approximate^ equal
e.m.fs. in the conductors thus
connected
to the pole pitch, so that the may be additive.
The drum armature may be thought of as evolved from the ring type by moving the inner connections of the winding elements to the outer surface, at the same time stretching the coil circumferentially until the spread of the coil
pitch.
is
approximately a pole
88
ARMATURE WINDINGS
The
disk armature differs
89
from the other two types in that the of instead active conductors, lying on the outer cylindrical surof are face the core, disposed radially on the flat sides of a disk.
number of pairs of poles of opposite on both faces of the disk are active (see the wires signs, so that used in modern practice. Disk are seldom armatures Fig. 49).
The
disk revolves between a
of armatures described, the drum armature used practically to the exclusion of the others, because it obviates the hand winding required in ring armatures; also, since the coils are wholly outside the core, they may be wound on formers, or winding jigs, and can be thoroughly insulated before being
Of the three types
is
slipped into place.

62. Types of Windings. All armature windings, for bothdirectand alternating-current machines, belong to one or the other of the two types, open-coil and closed-coil windings. An open-coil is in one conductor and tracing with winding which, starting any
;
" " progressively through the winding, a dead-end is finally reached whereas, in a closed-coil winding, the starting point will finally be reached after having passed through all, or some sub-multiple,
of the conductors.
The use of open-coil windings is at present confined to alternating-current machines and need not, therefore, be considered here. Open-coil
windings were at one time used to a large extent in direct-current

series arc-lighting generators,
such
as the
Brush and Thomson-Hous-
ton machines. 1
CLOSED-COIL WINDINGS
Ring and Drum Windings. In designing the armature of a F^. 75.-Ring wound armature. generator or motor, the number of armature conductors is determined by the fundamental equation
63.
for the e.m.f. (equation 17,
will
Chap.
II).
The problem
is
then so
to connect the various conductors that their individual e.m.fs.
way
1
add together to produce the desired total e.m.f., and in such that the winding as a whole will be at all times symmetri-
cal
with respect to the brushes.

See
Dynamo
Electric Machinery, S. P.
Thompson.
90
closed-coil windings for
In Figs. 75, 76, and 77 .there are shown three distinct types of a 4-pole machine having twenty-two
Fig. 75 is
active conductors.
a ring armature, while Figs. 76 and

in the
of the
77 are
case
drum armatures. The arrangement of ^he winding of the last two is made clearer by resorting to the use
FIG. 76.
Drum
armature, lap winding.
FIG. 77.
Drum
armature, wave winding.
developed diagrams of Figs. 78

Figs. 76
of
and 79, which are derived from and 77, respectively, by rolling out the cylindrical surface the armature core into a plane.
64. Winding Element. It will be seen that in each case the winding consists of a number of identical elements which are
Developedjtap winding.
'
\ An element Figs. 76 to 79, inclusive. may be defined as that portion of a winding, which, beginning at a commutator segment, ends at the next commutator segment encountered in tracing through the winding. It will be evident
n J8.
^
ARMATURE WINDINGS
at once that
of
91
turn,
for instance, Fig. 80 represents elements of windings similar to those of Figs. 75, 76 and 77,
an element may consist more than two active conductors;
of
more than one
i.e.,
but with two turns each, instead of one. Small machines for relatively high voltage, railway motors for instance, frequently have as many as five turns per element; but
FIG. 79.
Developed wave winding.
in
machines of large capacity there is, as a rule, only one turn per element for the purpose of improving commutation. Every time an element passes through the neutral zone of the magnetic field the current which it has been carrying must be reversed in direction; hence its self -inductance must be kept as small as possible in order that the reversal of the current may not be im-
FIG. 80.
Winding elements having more than one turn.
peded, and as the coefficient of self-induction increases as the square of the number of turns, the number of turns should be a
minimum,
or unity, for best results.
Lap and Wave Windings. The three windings of and 77 belong, respectively, to the ring, lap, and wave types of closed-coil windings. The derivation of the terms lap and wave winding will be evident from an inspection of Figs. 78 and 79 in the former, the successive elements lap back over each
65. Ring,
Figs. 75, 76;
other, while in the latter they progress continuously in fashion around the periphery of the armature.
wave
92
be noted that in both lap and wave windings the two element are subjected to the influence of adjacent poles of opposite polarity, so that the e.m.fs. generated on the two sides add together. In a simple lap winding, the end of any element, say the xih, connects to the beginning of the
It will
sides of a coil or
l)st element, and the beginning of the (x + l)st element under the same pole as the beginning of the xih element; in a wave winding, however, although the end of the xih con(x
-f-
lies
nects to the beginning of the (x -\- l)st element, the latter is not under the same pole as the beginning of the xih element, but is
separated from it by a double pole-pitch. An examination of the directions of the current flow in Figs. 75, 76, and 77 will show that in the case of the first two diagrams
there are four separate and distinct paths for the current through the winding (a = 4); each of these paths will carry one-fourth of the entire current supplied to the external circuit in the case of generator action, or supplied from the line in the case of motor
In Fig. 77, on the other hand, though there are four poles as in the other machines, there are only two paths through
action.
the winding (a
Other things being equal, therefore, the diagram will generate twice the e.m.f. of either of the other two in accordance with the fundamental equation
2).
wave winding shown
in the
}v
3>Zn
8
a60X10
or,
what amounts
to
the
same
thing,
the
same
e.m.f.
will
be generated with only half the number of conductors required by a ring or lap winding. Furthermore, the diagrams show that four brushes are required in the cases of the ring and lap windings, while two will suffice in the case of the wave winding.
These two facts
in conjunction explain the reason for the
use of
wave windings
in the case of direct-current railways motors,
where the combination of the cramped space and the moderately high voltage require a minimum number of conductors; and no less important, considerations of accessibility for inspection and
repairs limit the
number of brush sets to two. Lap and wave windings are often referred
to as parallel
and
series windings, respectively.
ARMATURE WINDINGS
66.
93
Number
of
rent in an element
Brush Sets Required. Inasmuch as the curmust undergo commutation once for each pas-
element
sage of the element through a neutral zone, it follows that the may be short-circuited by a brush at each such reversal.
Since the
number of permissible brush In sets may in all cases be the same as the number of poles. lap windings and in simple ring windings of the type shown in Fig. 75, the number of brush sets must be equal to the number of But in wave windings, though p brushes may be used, two poles. brushes will suffice irrespective of the number of poles. Thus, in Fig. 81, which shows a wave winding for a 6-pole machine having
equal to the
of poles, the
number of number
neutral zones and consequent reversals
is
FIG. 81.
Six-pole
wave winding, showing elements short-circuited by brush.
thirty-two armature conductors, any two of the three positive brushes a, b, and c may be omitted, as b and c (provided that a corresponding pair of negative brushes are removed at the same
time), in which case the remaining brush, for example, brush a
in Fig. 81, will short-circuit three
elements in series when
it
spans
two adjacent commutator segments. The three elements thus short-circuited by brush a are shown in heavy lines. Fig. 81 also makes it clear why two brushes instead of six, will suffice to collect the current, for it will be observed that brushes a, b and c are connected not only by the external conductor A but also by the short-circuited elements shown in heavy lines; these elements are in the neutral zone, consequently have little or no e.m.f. generated in them and are, therefore, equivalent to additional dead conductors joining the three brushes; hence conductor A and any two of the brushes a, b and c may be omitted. But if brushes b and c are retained, it will be observed that brushes
94
a, 6,

and
which are connected together to form one terminal machine, operate in pairs to short-circuit single elements.
c,
of the
This reduces the e.m.f. of self-induction to one-third of the value that would otherwise have to be handled, thereby improving
commutation
67.
If
conditions.
Simplex and Multiplex Windings. Degree of Reentrancy. identical ring-wound machines are connected in parallel as indicated inFig.82,the combined current output will bedouble that of either machine separately. The same result may be
two
attained, together with
economy in the use of material, by placing two independent windings on the same armature core, subjected
to the magnetizing action of a single field structure, as indicated
FIG. 82.
Armatures
in parallel.
in Fig. 83.
segments
bricated.
of the
Here both the winding elements and the commutator " independent windings are sandwiched" or imThe same result might also be secured by using two
independent commutators, one at each end. Windings of the type of Fig. 83 are called duplex windings as distinguished from Obvithe simplex windings of Fig. 82 and those preceding it. of to the there is independnothing multiplication ously, prevent ent windings so as to form triplex, quadruplex, etc., windings. In general, windings of this sort are called multiplex, to distinguish them from a simplex winding, in which each conductor must be passed through once, and only once, in tracing through its entire
closed course.
Drum windings, both of the lap and wave varieties, may be treated in the same way as has here been described for the case of
ring windings.
mutator segments
be noted that the interleaving of the comcomponent windings requires the use of brushes of sufficient width to collect the current from each pair of circuits at a neutral point. The arrangement illustrated in Fig. 83 shows two independent
It will
of the
ARMATURE WINDINGS
95
windings each containing twelve elements, or twenty-four in all. Suppose, now, that one of the twenty-four elements is omitted, and that the remaining twenty-three elements, uniformly spaced, Instead of having two are connected alternately, as in Fig. 84.
independent windings, each closed upon itself, as in Fig. 83. there is now but a single closure; but a study of the direction of current flow, indicated by the arrow heads, will reveal the interesting fact that there are still four paths through the armaIn other words, ture from brush to brush, just as in Fig. 83.
13
12
FIG. 83.
FIG. 84.
84.
FIGS. 83
and
Duplex armature windings.
is
both Figs. 83 and 84 illustrate duplex windings, but the former The doubly reentrant while the latter is singly reentrant. degree of reentrancy of a winding is, therefore, numerically equal to the number of independent, separately closed windings on the armature. Thus, it is possible to design windings as triplex,
reentrant; triplex, reentrant, etc.
triply
singly reentrant;
quintuplex, singly
It should be understood that all of these conclusions apply with equal force to lap and wave drum windings, the ring type having been used in the above discussion solely for the sake of
simplicity. 68. General Considerations.
The first systematic analysis of the relations to be satisfied in order that a symmetrical closed
winding might result was the work of Professor E. Arnold of
The Karlsruhe, who published the result of his studies in 1891. following derivation of the fundamental formulas is based upon that of Professor Arnold.
96
Probably the first questions that will present themselves to the student examining diagrams like those of Figs. 78, 79, and 81 are: How does one know in advance the number of coil edges to be stepped over in joining the end of one bundle of wires to the beginning of the next? Thus, in Fig. 81, the order is 1-6-11-16,
of connection do equally well? the effect of changing the total number of coil edges from 32 to some other number? The answer to these and related questions is implicitly involved in a general equation
etc.;
would not some other order
And what would be
covering
69.
all
kinds of closed windings this equation

;
is
derived in a
succeeding
article.
ments.
of Conductors, Elements, and Commutator SegWithout regard to the number of turns per element, ring windings usually have only one active coil edge per element, while drum windings have as a rule two active coil edges per element. Further, in accordance with the definition of an element, there must be as many commutator segments, S as there are elements. Consequently, in ring windings the number of commuta}
Number
equal to the number of active coil edges, while in the number of commutator segments is usually equal to half the number of coil edges. S must of course be an
tor segments
is
drum windings
(a)
FIG. 85.
(6) coil edges.
Elements having four active

;
ings which have one turn per element
be either even or odd therefore, in ring windand in which S is odd, the number of peripheral conductors may also be odd; but the number of elements in ring windings is usually made even, and more particularly in simplex windings a multiple of the number of poles in order that each branch path of the armature may be at all' times identical with all of the others, in which case the number of conductors will be even. In drum windings, no matter whether S is even or odd, and irrespective of the number of turns per element, the number of conductors must be even.
integer,
but
it
may
ARMATURE WINDINGS
97
Since the 'number of active conductors, Z, must be a simple multiple of the number of commutator segments, 8, it follows that the study of the arrangement of conductors may be reduced to one involving the order of connections of the elements to the com-
mutator, so that the quantity S is the factor of importance. In certain drum windings it is desirable to reduce the number of commutator segments to a value smaller than that which corresponds to one segment for each pair of active coil sides. This
can be accomplished in the manner indicated in Fig. 85, where each element has four active edges. In 70. Winding Pitch, Commutator Pitch and Slot Pitch.
Fig. 78
be observed that the back, or pulley, end of coil edge No. 1 is connected to the corresponding end of coil edge No. 8, and that the front, or commutator, end of No. 8 is connected The number of coil edges passed over to the front end of No. 3.
it
will
in this
pitch,
way
is
called the winding pitch; thus, in Fig. 78, the back
which
5.
will
or
2/2, is
be designated by yi, is 7, while the front pitch, In Fig. 79 both front and back pitches are positive
and equal to
5.
Again, in Fig. 78, the beginning and end of each element are connected to adjacent commutator segments, whose numbers
by unity. Similarly, in Fig. 79, the terminals of the elements are connected to segments which differ numerically by 5. This numerical difference between the terminal segments of an
differ
element is called the cojnmutator pitch, y. In slotted armatures the number of slots spanned by a element is called the slot pitch.
Coil or
Lap windings
are right-handed or left-handed, respectively,
depending upon whether y\ is numerically greater or less than 7/2In other words, if one faces the armature at the commutator end, the winding is right-handed if it progresses clockwise from seg-
ment
to
segment
of the
On
the other hand,
commutator in tracing through the circuit. wave windings are right- or left-handed ac-
cording to whether one arrives at a segment to the right or left, respectively, of the starting point after tracing through p/2 elements, where p is the number of poles. Thus, in Fig. 81, the
winding
is
left-handed.
The
algebraic
sum
of the front
and back pitches
is
a measure
98

advance or retreat per element in tracing through the In the case of lap windings,
of the total
winding.
2y =
while in that of
yi-yz =
yz
2y
(1)
wave windings 2y = yi +
2y
(2)
the factor 2 arising from the circumstance that each element has two coil sides. In cases where there are more than two coil
edges per element, say n,
2y = ny
or
3)
^
'
show that the terminals
Reference to Figs. 78 and 79 will 71. Field Displacement. of each element of a winding are cont
nected to segments which do not occupy corresponding positions with respect to the polar axes. There is a field displacement between them which may be expressed in terms of the number of commutator segments, ra, by which the terminals fail to occupy
homologous positions. Thus, in Fig. 78, m = y = 1, while in = y = 2. In wave windings, as Figs. 79 Figs. 83 and 84, m and 81 plainly show, the terminals of an element are separated by an interval approximately equal to a double pole pitch, but differing from the latter by an amount which is again a measure of the field displacement. Expressed in terms of commutator
segments, the double pole pitch
case
is
-*-
~>
&
or 2S/p, so that in this
y+m
m
2Sf
(5)
In lap windings and in ordinary ring windings, m is necessarily an may be fractional (H, M> J4> integer, but in wave windings determines whether the winding is right is positive the winding is rightor left-handed; obviously, if handed, while if it is negative the winding is left-handed. It has already been noted 72. Number of Armature Paths.
etc.).
The
sign of
that simplex lap and ring (or parallel) windings always have as many paths through the armature as there are poles, while simplex
wave windings
(series)
have but two paths, irrespective of
ARMATURE WINDINGS
the
99
exist-
number
of poles.
It
remains to determine the relation
ing between the number of paths and the winding and commutator pitches, since it will have been observed from Figs. 83 and
84 that the change of y from 1 to 2 changed the winding from simplex to duplex, that is, doubled the number of paths. In the first place, it will be noted that in tracing through the
winding from any arbitrary starting point, say a segment in contact with a negative brush, one complete path will have been passed over when the successive increments of the field displacement ( m) have brought the total displacement to a value equal to a pole pitch, S/p. This may easily be seen by referring to a simple ring winding like Fig. 75, but it is equally true for lap and wave windings. In the process of tracing through one path
there will have been encountered a certain number, S'
of commutator segments (not necessarily an integral number), to each of which there corresponds a displacement m. The total dist
placement
is
then
mS' = P
or
o
7
= mp
(6)
Since S' segments have been encountered per path, the total
number
of paths
must be
which
or
is
necessarily integral, hence
= mp
(8)
m=
73.
P
General Rules.
has
In the case of ordinary lap windings
it
now been shown

v
that
-*-it is
(9)
while in the case of
y=
wave windings 2/i -+ -?y 2 28
2Sa
-
do)
100
It will

be noticed that these two expressions for y
differ
by the
2/S/p, which represents a double pole pitch and expresses the fact that the terminals of a wave element having two active
term
edges are separated by approximately that amount. In a winding like that of Fig. 856 the term 2S/p would be replaced by 4S/p, and in general the numerical coefficient of S/p represents the field step, /, of the elements in terms of the number of single The end of a pole pitches included between their terminals. lap-winding element always occupies the same field zone as the
coil
beginning, so that /
therefore,
for lap
windings.
Quite generally,
-**--
<>
From this general equation there may be deduced certain convenient rules for determining the order of connections of the coil edges, thereby fixing the design
OQQQQQQOOO
I
TN
135
r^
5
(a)
TN
(6)
winding elements. has been pointed out in a previous section that the number
It
of the
of
7
coil
edges (nS)
is
of
drum
winding
then,
bered,
HO
DD DC
79
1
necessarily even. If, the coil sides are numhalf of
them
will
bear
even numbers and the other half odd numbers. Since each coil side proceeding outwardly from
a
ODD ODD
commutator
coil
a another
have
return
segment must path through edge, the number-
ing may be so arranged that the even numbers will constitute the
outgoing group while the odd numbers will comprise all of the return group. This means that even-numbered coil edges will
be connected to odd-numbered coil edges at both ends, therefore, This is a general rule for front and back pitches must be odd.
drum windings provided the numbering is carried out in accordance with the system indicated in Fig. 86. 1. LAP OR PARALLEL WINDINGS. An examination of the
all
ARMATURE WINDINGS
number
of elements,
>,
101
which may, accordingly, be even or odd. In the great majority of commercial windings there are only two coil sides per element (n = 2), so that
yi
y*
= 2y=
CL
2m
it follows that the pitches must differ by twice the There redegree of multiplicity in addition to their being odd.
from which
mains the further condition that both y\ and

2S>
y%
must not
differ
too greatly from the pole pitch,
as otherwise the e.m.fs. of

It is
the connected sides will not be effectively additive.
not
2S
essential that the average pitch
approximate
so far as
mere
closure
is
concerned, and in certain so-called chord windings or

is
fractional pitch windings the average pitch larger or smaller than this value.
purposely
made
As an example
of these rules, it
may
2y
be observed that in Fig. 78
Z = 22 = m =
S =
a -
U
y2
9
it
p = 4
=
2/2
yl
yl
Had
the pitches been

close,
made
but
winding would
chord winding. Since m = ~
and 7, respectively, or 5 and 3, the would be an exaggerated form of
y, it
follows that in an m-plex lap winding the
commutator pitch equals the degree of multiplicity. Thus, in a simplex lap winding the ends of an element are connected to adjacent segments; in a duplex winding they are separated by one segment, etc. 2. WAVE OR SERIES WINDINGS. The general formula
P
reduces to y
of this type.
- for
most commercial windings
number
It is clear that the choice of S, and therefore of the of active conductors, is not unlimited as in lap windings. In Fig. 81, for instance, which represents a simplex wave winding
for a 6-pole
machine, a =
2,
6,
2S =
32, hence y
32
+ -2
102

5%.
The
latter value of y being impossible,
5 or
take y
select
5.
Since the pitches
must approximate
of 7
we must 2S
5J/3,
2/i
?/2
5,
though values
and 3 would
result in a
closed chord winding.
The
restriction
upon the number
of elements in
wave windings
frequently causes the use of "dummy coils." Suppose, for example, it is necessary to design a simplex 4-pole wave winding to
be placed on an armature core having 65

sufficient size to
slots,
each
slot
being of
of
accommodate
four coil sides, in the
manner
Fig. 86c.
This means that Z = 260, assuming each coil edge to consist of a single conductor, and of course this value of Z must accord with the fundamental equation (17, Chap. II). Summar= 260, a = 2, p = 4; whence izing, 2S 9fiO 4- 2 -^-y =
But
since y
must be an
integer, the value of

is
2S nearest
is
to 260 that
impracticable because the maximum number of coil edges that can be placed in the slots is 260). Taking 2S = 258, it follows that there must
will satisfy the
equation
258 (2S
262
be one element, consisting of two conductors, that is not a part of the winding; it is put in simply to fill up the space in the two slots
which contain only three active conductors each. Therefore, 258 -)- 2 y = Z^ = 64 or 65. Since 2/1 and 2/2 must be odd, and
further
sible:
f
--
y,
the following pairs of pitch values are pos-
2/i
1 2/2
= =
65 65
etc.
ARMATURE WINDINGS
103
conductors 1 and 3 may then be joined to conductors 66 and 68, respectively, thereby allowing the conductors to be insulated together in pairs and facilitating the placing of the bundles in the
slots.
In wave windings the
field
displacement
is
given by
m=
->
so that after tracing through

circuit of
p/2 elements, corresponding to one the periphery, the total displacement is X- = s Zi Z p
Therefore, in simplex windings (a = 2) the end of the p/2ih element connects to a segment adjacent to the starting segment; in duplex windings it connects to the next
commutator segments.
but one, etc. The ordinary wave wind3. SERIES -PARALLEL WINDINGS. = 2) through the armature irreing results in but two paths (a
spective of the
number
of poles.
But
it is
possible to secure
any
multiple of this number of paths by a suitable choice of S in the Wave windings having more than two paths general formula.
are called series-parallel windings. Thus, if a 6-pole armature has 154 conductors wound to form 77 elements, it may be arranged as a 4-circuit (duplex) wave winding; substituting
2,
S =
77, a
=
25.
4,
and p = 6
in the equation y
>
there results y
74.
General Rule for the Degree of Reentrancy.
If,
in the
general formula,
P
the two sides of the equation have a
common
factor
q,
we have
y
q
.
p
or y
.
p
is
i
really
which means that the original winding
made up
of q
independent windings, each of which has S'

,
- elements and a
f
commutator pitch of y' the latter counted with respect to the S segments. That is, the winding will be multiplex and multiply reentrant of the qth degree in the event that y and S have a
104

factor
</;
common
it
will
be singly reentrant
(/
if
y and
are prime
to each other.
In ordinary duplex wave windings
2)
2)
yfrom which
factor,
it
2S
2 (S
~-
follows that
also be
if
is
S must
even, that o _j_ 2
is,
contains 2 as a
even because
must be an integer and

is even;
is
always an even number.
This leads to the simple rule that a
duplex wave winding is doubly reentrant if y corollary that it is singly reentrant if y is odd.
and
to the
In triplex wave windings in which/
2,
2S6 . 2JS_
P p
3)
Suppose now that y contains 3 as a factor, where x is an integer; then from (13)
3z-
in
which case y
3z,
= S
Therefore, since ^- xis integral,
S must be
a multiple of
triplex
3,
hence
the winding
reentrant
if
is
triply reentrant.
Hence a
3,
3.
will be triply reentrant if y is
a multiple of
and
it will
wave winding be singly
is
not a multiple of
In the case of quadruplex wave windings, however, such simSuch windings plifications of the general rule are not possible.
may be singly, doubly, or quadruply reentrant. Thus, if / = 2, a = 8, and p = 6, S = 79 leads to a singly reentrant winding in which y = 25; S = 82 results in a doubly reentrant winding, y = 26;and>S = 80 gives quadruple reentrancy, y = 28. An inspection of any of the wind75. Two-layer Windings.
ings of Figs. 78, 79, 81, etc., will show that the end connections of successive conductors proceed alternately in opposite directions.
If all of the conductors lay in the same cylindrical surface, as in the case of smooth core armatures, the crossing of the end connections would make the actual winding process difficult of exe-
cution.
But where
slotted armatures are used,
if
the conductors
ARMATURE WINDINGS
are arranged in
all
105
two layers, the end connections of the upper layer proceed in one direction while the end connections of the lower layer, at the same end of the armature, may all proceed in Since the upper and lower the opposite direction, as in Fig. 88,
may
FIG. 88.
Partially
wound drum armature
(lap winding).
FIG. 89.
Samples
of
winding elements.
layers include
tively,
all the odd and even numbered coil sides, respecconductors in the top layer must connect to others in the lower layer, the transition being effected by the peculiar bend in
the coil
shown
at B, Fig. 89.
106
It is easy to recognize an armature as lap or wave wound, when the conductors are made of bars or strips of copper, by observing the relative directions of the top end-connections at the two ends
a) Lap Wound
Armature of
I ) Wave Wound
Armature
in lap
FIG. 90.
Showing direction
end connections
and wave windings.
FIG. 91.
Duplex lap winding, singly reentrant. = +17, 7/2 = -13. 2/i
Z =
62,
S =
31, y
2,
Thus, if the top end-connections, when produced, meet at or near the center of the core, as in Fig. 90a, the winding is a lap winding; whereas if the top end-connections are parallel, as in Fig. 906, the winding is a wave winding.
of the armature.
ARMATURE WINDINGS
76.
107
Examples
clearly
of
Drum
Windings.
show
how
a slight change in the
Figs. 91 to 94, inclusive, number of coil edges will
change the winding from single reentrancy to multiple reentrancy. Thus, in Figs. 91 and 92, although both windings (duplex lap) have identical pitches, the former, with 62 coil edges, is singly reentrant, while the latter, with 64 coil edges, is doubly reentrant.
Electrically, however, these windings
have identical properties,
FIG. 92.
Duplex lap winding, doubly reentrant. y = +17, y z = -13.

l
Z =
64,
8 =
32, y
2,
except for the slight difference in e.m.f. due to the different number of active conductors. Similar remarks apply to Figs. 93
and
which show singly- and doubly-reentrant duplex wave Note that in Fig. 93 the pitches (y, yi and y z ) equal windings. but that 17, pitches of 15 would also work out correctly; and that in Fig. 94 it would also be possible to design the winding with
94,
pitches of y
= 14, y = 15, y 2 = 13. 77. Equipotential Connections. Consider

like
parallel-wound
armature
Fig. 95, which represents diagrammatically a
108
winding for an 8-pole machine. The eight parallel paths through the armature from terminal to terminal are shown somewhat more clearly in their relations to one another in the diagram of Fig. 96. It will at once appear that if each path is to carry its proportionate share of the total armature current, each path must at all times generate the same e.m.f. and have the same
resistance as
the other paths. In any case, the current will divide between the eight paths in accordance with KirchhofTs
all
FIG. 93.
Duplex wave winding, singly reentrant.
Z =
y*
64,
S =
32, y
yi
17.
laws for divided
circuits,
namely:
(1)
the
summation
of all the
potential differences in each closed circuit must be zero; (2) the sum of all currents meeting at a point must be zero. If for any
reason the e.m.f. generated in one path is greater than in another, for instance, if that of circuit 3-2' is greater than that of 3-3', the brushes 2' and 3' will not have the same potential and an
equalizing current will flow in the lead joining brushes 2' and 3'. Even very small differences of potential may give rise to internal
ARMATURE WINDINGS
109
equalizing currents of large magnitude, owing to the low resistance of the circuits, so that excessive heating of the winding and sparking at the brushes may result if preventive measures are not
employed.
The causes
as follows
1.
:
of possible unequal e.m.fs. in the various paths are
The armature may not be
the pole shoes, due to natural
exactly centered with respect to irregularities in construction or to
FIG. 94.
Duplex wave winding,

y
16, yi
doubly reentrant. = 15. 17, y t
Z =
60,
S =
30,
wear of the bearings. The air-gap is consequently not uniform, and some of the poles therefore carry more flux than others,
thereby generating more e.m.f. in the coils subject to their influence than is generated in coils under. the weaker poles. This
cause
armature
importance in lap and ring windings, where each is at any one time under the influence of one pole only- in wave windings each path is simultaneously acted
is
of
circuit
110

all of
upon by
the poles, hence this type of winding
is
free
from
the disturbing effect of non-uniform polar flux. 2. The poles may not all be identical in construction, so that
their fluxes
may
+
differ
even
if
joints
is uniform. Thus, the between the poles and the yoke, or between the pole cores and
the air-gap
the shoes, may not all be equally good, or the magnetizing effect of
rH
2 (TJOOOOOOOOOOOOOOOOOOO
1
OOOOOOOQOOOQOOOOOOQ 2/
^
2'
j)OQJ)QQ^!QOOOOOa(!_0(!
3'
ti
FIG. 95.
Parallel-wound armature
with equalizer connections.
FIG. 96. Diagrammatic scheme of connections of armature of Fig. 95.
the
field
coils are
3.
windings may differ, especially in cases where the field connected in parallel instead of in series. The armature circuits may be unsymmetrical, due to a
is
choice of
the
number of elements which number of paths. In this

it
not an exact multiple of
connection
should be ob-
served that in multiplex windings there is always dissym-
metry between the circuits, due to the fact that the
FIG. 97.
ments
Short-circuiting of eleof multiplex winding.
FIG. 98. Equalizer connections in parallel winding.
brushes short-circuit an unequal number of elements of the component windings (Fig. 97). This kind of asymmetry will give rise
to equalizing currents even if the individually generated e.m.fs. are equal, because the circuits have slightly different resistances.
The
equalizing currents are a source of loss because of the
ARMATURE WINDINGS
extra heating caused
111
by them.
This effect can be minimized by
striving for the greatest possible degree of magnetic and electrical symmetry. To obviate the remaining difficulty of sparking at
tions,
the commutator, resort is had to the use of equipotential connecwhich are low-resistance conductors joining points in the
winding which, under ideal conditions, would at all times have the f Thus, in Fig. 98, the points a, b, c, d, and a', b', c potential. d', etc., are always under the influence of corresponding parts of Should irregularities exist, equalizing current poles of like sign.
same
FIG. 99.
Equalizing rings of large lap-wound armature.
will flow
extra current
of course
through these connections, relieving the brushes of the and preventing sparking. The equalizing current is
an alternating one. The equipotential connections, occasionally referred to as the equalizing rings, are sometimes placed between the commutator and the armature core under the end
connections in large drum armatures they are generally mounted on the exposed side of the core as in Fig. 99. When the equipotential connections were first introduced by
;
Mordey they were intended

sets.
to reduce the usual
number
of brush
Thus,
it is
clear
from Fig. 98 that two brushes might be
112
expected to take care of the entire current, since the commutator segments that would be touched by three of the brushes of one polarity are already connected to the fourth brush by the equalizdifficulty arising from the use of only two the armature circuits are not identically situated with respect to the line terminals, the extra resistance o the equalizing connections between the remote armature paths
ing connections.
The
brushes
is
that
all of
and
being sufficient to introduce an unbalancing of the In present practice the number of brush sets is not duced when equalizing connections are used.
line
cuits.
cir-
re-
PROBLEMS
of a simplex lap winding for a 4-pole drum arma ture having 48 coil edges, using the kind of diagram shown in Fig. 91 Use pitches of 13 and 11. Show position of brushes and indicate
1.
Maka a drawing
direction of current flow in the winding. 2. Make a drawing of a simplex wave winding for a 4-pole drum armature having 46 coil edges, using the kind of diagram shown in Fig. 93. Use
front
and back pitches of +11. Show position of brushes direction of current flow in the winding.
and
indicate
3. Make a table showing the order of connections' of the coil edges of th( winding of Problem 2 if the front pitch is + 11 and the back pitch + 13 4. A 6-pole drum armature has 450 coil edges. Find all possible lap
and wave windings, up to and including triplex windings, that are possible give front, back and commutator pitches and degree of reentrancy ii each case. Assume that each element has two active coil edges. 6. The armature core of a 4-pole 110-volt generator has 63 slots each of>
accommodate 4 conductors. If the flux per pole is ap6 proximately 10 lines and the speed of the machine is 1300 r.p.m., how many conductors are necessary, and how must they be wound? State pitches. 6. A 4-pole generator has a rating of 15 kw. at 125 volts and 1200 r.p.m. The diameter of the armature is 12 in., the length of armature core is 5 in., and the pole arc is 70 per cent, of the pole pitch. The armature has 47 j slots, each containing 4 conductors, and there are 47 commutator seg ments. Knowing that the flux density under the pole faces does not*exceed 55,000 lines per sq. in., how many parallel paths are there through the armature? Find the type of the winding-, and the winding and commutatoi
sufficient size to
.
pitches.
7.
A 6-pole duplex wave winding has
127 elements each having two turns,
of one complete turn is 37.5 in. and the conductors have a cross-section of 60,000 circular mils (a circular mil is the area of a circle
1 mil, a mil being the thousandth part of an inch). the specific resistance of copper at the running temperature is 12 ohms per circular mil-ft., what is the resistance of the armature measured betweer brushes?
The mean length
which has a diameter of

If
CHAPTER
IV
THE MAGNETIZATION CURVE. MAGNETIC LEAKAGE

78. The Magnetization Curve. Every dynamo consists of an electrical circuit interlinked with a magnetic circuit. The armature winding is the electrical circuit in which the e.m.f. is produced in the case of a generator and in which the workIn either ing current produces the torque in the case of a motor. case, the activity of the armature is dependent upon the magnitude of the magnetic flux, and the latter, in turn, is dependent upon the magnetizing effect of the field winding and the reluctance of the magnetic circuit in accordance with the relation
Flux
Since
= $=
p a 60
m.m.f.
reluctance
E =
it
10 8
m.m.f.
follows that
Zn
a 60
10 8
reluctance
which means that

of this function
is
curve, or the If the magnetic circuit

reluctance,
is a function of the field excitation; the graph called the magnetization curve, or the saturation no-load characteristic.
had constant
characteristic
the
no-load
line
would be a straight
origin,
through the but since the permeability of the iron of the magnetic circuit falls off as the flux increases, the flux does not bear a constant ratio to the m.m.f.;
the result
teristic
is
Field
Ampere-Turns
that the no-load charac-
FIG. 100.
droops below the straight line
Magnetization curve of a dynamo.
form, as indicated in Fig. 100.
The dashed lines in Figs. 101 and 102 represent the mean paths of the flux in typical forms of bipolar and multipolar ma8
113
114
chines.
These lines are so drawn that they pass through the centers of gravity of the sections of the tubes of induction. It will be observed that a complete path or magnetic circuit, such as
C', Fig. 102,
air-gaps,
two pole
includes the armature core, two sets of teeth, two shoes, two pole cores, and the connecting yoke.
magnetizing winding P on one pole will set up the same flux each of the paths C and C" (assuming perfect symmetry of A similar windconstruction) since these paths are in parallel. ing on every alternate pole would then magnetize all the poles equally, hence the excitation required to drive the flux through
in
a complete magnetic circuit
is
the excitation per pair of poles.
FIG. 101.
Magnetic circuit of bipolar machine.
FIG. 102.
Magnetic
circuits of a
multipolar machine.
The
is, therefore, that required to maintain a magnetic circuit such as abed, Fig. 102, consisting of a single air-gap, one set of teeth, one pole shoe and core, and half of the connecting circuit through the armature core and the yoke. Field excitation is generally expressed in terms of the
excitation per pole
the flux in
number
of ampere-turns per pole; or in terms of ampere-turns per pair of poles. The magnetization curve is of great importance. Whether the
machine
teristics.
is
to be used as a generator or as a motor, the

will largely
form
of the
magnetization curve
determine
its
operating charac-
measure
Conversely, a given set of specifications will in large It is, therefore, the form of the magnetization curve. fundamental is of this curve of that the determination apparent the In the a machine case of importance. magnetizacompleted
fix
THE MAGNETIZATION CURVE

tion curve can be determined experimentally; culated when the dimensions of the machine
it
115 can also be

cal-
and the nature are known. construction in its used the materials 79. Experimental Determination of Magnetization Curve.
Since
of
E
it is
p -
Zn
^7rr;~r
'
m.m.f. reluctance
\
a 60
10
kn
function of field ampere(1)
,.
turns
clear that
it is
only necessary to run the machine at a constant speed n (driving it with a motor or other suitable prime mover) and to observe a series of simultaneous pairs of values of
and ampere-turns. '<- In the above equation generated in the armature by rotation through the flux produced by the
therefore,
rectly,
field
(1)
is
the e.m.f.
current;
measure E dithe armature must be

to
without current, that is, it must be on open circuit. The machine must then be separately excited during this test, as in Fig. 103, current being .. .. supphed to the field winding from, a suitable external
-.-,.
i
.
Source of Current
.
FIG. 103. Experimental determmation of magnetization curve.
.,-,..
source, controlled
an ammeter A.
current
by a variable resistance R and measured by The procedure then consists of varying the by means of R, and taking a reading of voltmeter V for
each setting of A, the speed being kept constant throughout. Inspection of equation (1) indicates that with a fixed value of field excitation the generated e.m.f. E would be directly prois
portional to the speed. possible that current
This
may
is, however, not quite true; for it flow in those elements of the arma-
ture winding which are short-circuited by the brushes, and these short-circuit currents may easily reach values of sufficient mag-
nitude to react upon the flux and so affect the generated e.m.f. To reduce the disturbing effect of these short-circuit currents to a minimum, it is necessary to cut down the e.m.f. which gives
rise to
them, and, with a given field excitation, this can be done by reducing the speed. It is best, therefore, to determine the
116
magnetization curve at a speed considerably below the rated speed, and then to multiply the observed voltage by the ratio of rated speed to the speed actually used. The effect of the current in the short-circuited coils of the armature is discussed in more
detail in
Chap. VIII.
of the magnetization curve obtained experimentally if the exciting current in the field winding is first
The form
is
not the same
gradually increased from zero to a max-
imum and then gradually this maximum back again

form
Field Ampere-Turns
reduced from
to zero.
The
observed readings when plotted take the
The
is
of Fig. 104. difference between the
two curves
due to the
hysteresis of the iron part
FIG. 104.
teresis
Effect of hys-
on magnetization
of the
curve.
the
circuit, hysteresis being to that property of iron (or other magnetic substance) by virtue
magnetic
name given
of which the induced magnetism lags behind changes in the magnetizing force. 80. Calculation of the Magnetization Curve. To analyze the method of computing the coordinates of points on the magnetization curve, the following symbols will be used:
Part of circuit

(all
117
the requisite
number
quantities being expressed in c.g.s. of ampere-turns for the length l x
units),
is
But
10 j
AT X = TT H x =
7
dtx
ampere-turns per cm.
.'.AT
ata.L
2att.lt
2at a .S
+
B
2at s .l s
of
at
2at c .l c
at y l y (3)
The magnetization (B-H) curves

usually plotted in
magnetic materials are

(the latter being simply
terms of
and
of
Q.8H), so that
circuit
when the values
for the various parts of the
have been determined, the corresponding values of at read from the curve and substituted in the expression for be may AT. Fig. 20 shows magnetization curves for the usual commercial materials, the coordinates being plotted in terms of metric
and
also of English units. becomes pression for
If
English units are used, the ex-
AT
The general method of calculating the coordinates of points on the magnetization curve then consists of assuming a number of values of the generated e.m.f. E, ranging from about a quarter of
full-load value to, say, 25 per cent, in excess of that value. To each value of e.m.f. thus selected there will correspond a value of
TTT
flux,
<, as
determined by the equation $

<
p 6n
,
- =
in
60
10 8
this value of
then fixing the flux density,
Bx
each part,
x, of
the circuit.
Having determined
X)
at x is
found from Fig.
20,
whence
AT =
the
Sok.Z, c
circuit,
summation being extended over a complete closed
such as C, Fig. 102. 81. Magnetic Leakage. The flux per pole, $, may be designated the useful flux, since it is this flux which is involved in the production of the generated e.m.f. But the entire flux produced
by the magnetizing action
of the field
the armature, an appreciable part of
winding does not penetrate " it leaking" across from
118
and in general between all points which have between them a difference of magnetic potential. This "leakage flux," <p, increases the total flux from $ to
pole to pole,
The ratio
called the leakage coefficient, or preferably, the coefficient oj It is always greater than unity, and in machines of dispersion.
is
larger values corresponding to small machines.
the usual radial multipolar type ranges from about 1.1 to 1.25, the Since the leak-
age flux must traverse the poles and yokes, the cross-section of these parts must be sufficiently large to carry it as well as the
useful flux, hence the necessity of keeping down leakage as much as possible. The conditions to be satisfied to attain this end are,
accordingly, minimum reluctance of the main magnetic circuit and maximum reluctance of leakage paths; this means, practically, a
compact magnetic
circuit
made up
of short poles, the interpolar
spaces being wide and of small section. The magnitude of the coefficient of dispersion
for a given
is
not constant
flux,
<p,
machine under
all
conditions.
The leakage
being mainly in air, is very nearly proportional to the m.m.f., while $ is less and less proportional to the m.m.f. as the saturation of the iron
is
increased.
In general, therefore,
increases
more or less with increasing excitation. Methods for the calculation of the coefficient v will be given
section.
in a
v
subsequent
is
For the present
it will suffice
to state that
This introduces a difficulty in a new design because the flux densities, etc., cannot be determined until the dimensions have been decided upon, and the dimensions are themselves dependent upon << and, conseIt is therefore necessary in such a case to quently, also upon v.
of v in accordance with previous experience, prothen to the calculation of $ and the dimensions. The ceeding true value of v can then be calculated and the tentative computations modified in case the discrepancy is sufficiently large to warrant a readjustment.
t
a function of the dimensions of the machine.
assume a value

82. Details
1.
119
Curve.
of
Calculation
of
Magnetization
great permeability of iron as compared with air is responsible for the fact that the reluctance of the air-gap often constitutes from 70 to
AMPERE-TURNS REQUIRED FOR THE AIR-GAP.
The
00 per cent, of the entire reluctance of the magnetic circuit. accurate determination of the excitation consumed in the
The
air-
therefore, of predominant importance. Two different cases arise in practice (a) smooth core armatures, and (6) slotted armatures.
gap
is,
(a)
Smooth Core Armatures.
In a machine having p poles, the

is
.
angle subtended
by the
is
pole-pitch
The angle
2-jr
;
subtended
the quantity
If
by the pole shoe

3
usually between 0.55 and 0.7 of
X 100 2~7
g called
the per cent, of polar embrace.
the flux
crossed the air-gap along radial lines, the determination of
Bg
FIG. 105.
Fringing flux at pole tips.
FIG. 106. Spread of flux at flanks of pole shoes.
g would be very simple; actually, however, the flux out spreads beyond the pole tips, as indicated in Fig. 105, and there is a further spreading at the flanks, as illustrated in the side
and
AT
The flux always distributes itself in such a elevation, Fig. 106. that the total reluctance is a minimum. The spreading of the way
flux
is
from
b to b
equivalent to an increase in the length of the polar arc and an increase in the axial length from I to I'. The
f
,
mean
flux density in the
gap
a
is
then
b'l'
B -
(6)
120

in air
and since
B = H
it
follows that
g
AT =
all
0.8B g 5
If
(7)
dimensions being in centimeters.
the inch
is
taken as the
unit of length,
AT =
g
0.8
(2%^" X
2.54)
= 0.3133B
fl
"5"
(8)
For practical purposes it is sufficiently accurate to take b' as the average of the polar arc 6, and of the arc on the armature subtended by the angle and increased by 25 on each side; that is,
d
/
_l
5)
25
(9)
Similarly,
/'
may
be taken as V =
25
(10)
and armature core (between If these lengths are not equal, let them heads) are the same. be represented by l a and I, respectively; then b'l' in the above equation for B g should be replaced by an area A g such that
in case the axial lengths of pole shoes
Ag =
(11)
where A' g is the area of the pole shoe and A" g is the area on the armature core threaded by the flux. Obviously A' g
bl.
(12)
J
i
and
/
Ck
(13)
(b)
Slotted Armatures.
In this case the calculation of
Bg
is
complicated by
J,
the fact that the flux tends to tuft at the tips of the teeth, and that more or less of it enters
the teeth
Fig. 107.
by way
of the slots, as indicated in
It is clear that a given difference of
magnetic potential between the pole face and the armature core will produce less flux when slots are present than when the armature sursmooth, the clearance (5) being the both cases. In other words, the slots increase the gap reluctance, and this effect may be allowed for either by assuming d to have been increased to a larger value, or
FIG. 107. Fringing flux at tooth tip.
face
is
game
in
by assuming a contraction
of the pole arc b to a smaller value, b'

The problem
gap
is
121
To
is further complicated by the fact that the airfrequently not of uniform length over the entire pole face. improve commutation it is common to chamfer the pole tips,
Fig. 108a, or to
make the cylindrical surfaces of the armature and pole face eccentric, Fig. 1086. The effect of the increased gap length at the pole tips is to produce a fringing flux in the inter-
(a)
FIG. 108.
Chamfered and eccentric pole
shoes.
polar space, as
shown by the
flux distribution curve of Fig. 109.
component of flux armature on the The periphery. density at corresponding points ripples at the crest of the curve are caused by the slots and teeth.
Ordinates of this curve represent the radial
shown
Similarly, there is a fringing field at the ends of the core, as in Fig. 110, and if ventilating ducts are provided, there will be dips in the curve of axial flux distribution corresponding
FIG. 109.
FIGS. 109
FIG. 110.
field
and
110.
Peripheral and axial distribution of
intensity.
to the depressions opposite the slots, just as in Fig. 109. The extra reluctance due to the ventilating ducts is equivalent to a
reduction in the axial length equivalent to an increase in

neutralize each other.
I, /,
and the fringing so that the two

It
at the flanks
effects
is
tend to
83. Correction to Pole Arc.
has been shown by F.
W.
122
Carter 1 that the presence of slots increases the effective length

of the air-gap
from
5 to d
f
,
where
and where
and
123
ratio
Values of the factor a are plotted in Fig. 112 in terms of the bs = slot opening
-
clearance
fringing field at the pole tips
6,
is equivalent to an increase but this effect is generally offset by the increased 1 gap space at the tips. Arnold has given a method for computing the increased length of pole due to fringing, but it is generally unnecessary to introduce such refined calculations. The reluctance due to the ven84. Corrected Axial Length. tilating ducts may be considered as reducing the axial length I to
The
in the
value of
:
-
'
l '>
-~t~
-r-
w
The length
l'\
where t v is the distance between centers of the ventilating ducts and b v is the width of the duct (Fig. 113), and where a is to be
found from Fig. 112 using as argument
can
be further corrected to take account of the flux which enters the

sides of the core
from the flanks

l' z
by the dotted
lines in Fig. 113; the correction takes the

,
of the pole shoes, as indicated form of

is
an additional length,
so that the equivalent axial length
l'i
J'
(17)
(T
hH
Pole Shoe
rp
FIG. 113.
FIGS. 113
FIG. 114.
flux.
and
114.
Correction of axial length due to fringing of
The value
of
1' 2
may be
estimated as follows:
lines of force of the fringing flux are
circles and of straight lines, as in r an elementary tube of width dx and breadth b
Assume that the made up of quadrants of The permeance of Fig. 114.

is
Die Clleiohstrommasehine, Vol.
I,
p. 274,
2nd
ed.
124

entire
end the
x
and x =
r,
permeance of all the tubes between the on both sides of the core, is
limits
C
P =
But the permeance
of force of length 5
I
b'dx
-f-
TTX
2,, -b' log e

,
6 + 5
?rr
TT
is
to be
made
equivalent to that of a tube

b'l' 2 ,
and cross-section
hence
^
and
2j
ZY =
For values
of r
1.56logi
to 5 times
(l
6,
+y)
/' 2
(18)
from
varies from 0.96 to 1.86.
Generally sufficiently accurate to take /' 2 = 1.56. b' and I', the corrected value of flux found Having density in the air-gap is
it is
$ B =
and therefore
VV
Q.8B
5
AT =
g
if
gap
metric units are used (flux density in lines per sq. cm. and airin centimeters), or
AT =
g
0.3133
B" g 8"
(B"
)
if
flux density is given in lines per sq. in.
and air-gap
in
inches (6").
85. Ampere-turns Required for the Teeth. The same difference of magnetic potential that maintains the flux through the teeth also produces a certain amount of flux through the slots,
since the
two paths
are in parallel.
When
the teeth are not
highly saturated their permeance is so considerable that the flux passing down the slots is relatively insignificant and may be
neglected; but in
many machines the iron of the teeth is purposely worked at high flux density in order to limit the effect of armature reaction (see Chap. V), and in such cases the permeance
of the teeth is decreased to
permeance becomes comparable with
such an extent that the slot it. If, then, it were as-
sumed that the
entire flux per pole passed through the teeth
125
immediately under the pole (with an allowance for the spread of the flux at the pole tips), the resultant tooth density would be higher than it is in reality, and the ampere-turns per unit length corresponding to this apparent density might be greatly in excess
because of the flatness of the magnetization The actual tooth density, B must therefore be distinguished from the apparent density, B' The conditions that determine
of the true value
curve at high saturation.
t,
t.
the relation of the actual to the
apparent density are (1) that the total flux per pole is equal to the sum of the flux in the iron
of the teeth
slots,
and
in the air of the
and insuFIG. 115. Dimensions of teeth between laminae; space and (2) that the magnitudes of the flux in the iron and in the air are proportional to the permeventilating ducts,
lation
ances of the respective paths.
Therefore
(19)
__
$ = $,- + ^ air $ iron - = M - X cross-section of iron = f $a cross-section of air

-.
-.
u,K
(ZO)
ir
where
/x
is
the permeability of the iron corresponding to the

t .
actual tooth density B Referring to Fig. 115,

cross-section of iron
b t (l
n v b v )k
(21)
where k
is
the lamination factor (usually about 0.9), and
cross-section of air
bs l
+
+
b t .n v b v
b t (l
n v b v )(l
k)
(22)
equations (21) given set of dimensions.
From
and
(22)
K
=
may
be determined for a
It follows that
3>air
$
$ir on
Qiron
$> ^ron
_t*K
= #J
~
x^v
compute from this equation and B' by assuming values for B finding the corresponding values of ju from the magnetization curve of the core material, and substituting in equation (23). shows the relation between B and at Thus, in Fig. 116, curve
For a given value
t ,
of
it is
possible to
a series of simultaneous values of
t,
as
determined by
test, for
commercial sheet
steel.
The remaining
126

t
curves give B' for various values of K. B corresponding to B' = PR, when
t t
Thus
0.5.
QR is
t
the value of
K=
Bt when B' is known has the disadvantage that differ from may any of the values for which curves have been prepared, and the labor of preparing such curves as those of Fig. 116 is in itself very tedious. It is
The above method
of determining
150
M
140
K=
Q
Magnetizat on Ciir
a 130
120
110
100 100
200
300
400
500
600
700
800
900
1000
1100
1200
Ampere-Turns per Inch
FIG. 116.
B from B' for any value of K, directly from the magnetization curve (M, Fig. 116), as follows: From the relation
possible to find
t
t,
B,
we have
R' B>
R Bt
'
H
(24)
provided
is
expressed in lines per sq. cm.
127
In Fig. 117, let C represent the magnetization curve plotted terms of B and H, and assume for the present that B and To the left of the origin lay off are plotted to the same scale. any convenient scale to represent values of K, and lay off is any value ON equal to unity to the scale of K. Then, if
in
OM
of
K,
OM
-r
=
will
tan a, and
if
OR
is
drawn
parallel to
MN the
There-
intercept
fore,
QR
equal
t
H A
corresponding to
= OQ.
PR = OS =
That
B't,
is,
Bf
value of
B' since PQ for a given
lay off
OS on
the
axis of ordinates equal to this
given value and through draw a line parallel to

until it intersects curve
S
in
MN
a point P. Ordinate PQ is then the actual tooth density
(B
t)
and
OQ is
the correspond-
ing value of H. Since B and

plotted to the
are never
scale,
same
and
since magnetization curves are usually drawn in terms of
and at, suitable modifications must be made in the construction.

If
FIG.
117.
Graphical
relation
bein-
tween apparent and actual tooth

duction.
is
plotted to represent lines per sq. cm. and at in
am-
pere-turns per cm., the length

to the scale of
ON
must be made equal

cm. per unit length
to
-77^ IU.DO
K, where
of ampere-turns per
of hori-
A = number
zontal axis
B = number of gausses per unit length of vertical axis. If B is in lines per sq. in. and at in ampere-turns per in., ON
must be made equal to 2.54
K, where
A' Q
^4r IU D
o
3.19
to the scale of
-DO
= number
of ampere-turns per inch per unit length of hori-
zontal axis
128
B'Q
= number
axis.
of lines per sq. in. per unit length of vertical
Otherwise the construction is the same as before. It does not immediately follow that AT = at -l t) because the tapering of the teeth results in an increasing density from the tip
t t
from point to point along the length
to the root, consequently at changes in value of the

t
tooth; nor does it follow that the value of at to be used is that corresponding to the flux
If values density at the middle of the tooth. of B t are computed for a number of points along the length of the tooth, and the cor-
responding values of at found from Fig. ]16, are plotted, a curve like Fig. 118 will
t,
118. Variae xci tation along tooth axis.

ti
FIG.
on
of
in general result. Evidently the true value of at t is the mean ordinate of the curve;
'
.
assuming that the curve mean ordinate, by Simpson's rule, is

at,
is
parabolic, the
+ 4a< + at,
2
whence
AT =
t
(at t ) mean -l t
(26)
It is
86.
clear
Ampere-turns Required from Figs. 101 and 102 that the iron
for
the Armature Core.

of the core
below the
roots of the teeth carries half of the useful flux per pole.
Therefore,
*--2z:
If
the radial depth
ofr
the iron under the teeth
is h,
(27)
Aa =
To
the value of
kh(l
nv bv )
(28)
of ata
thus determined there corresponds a ampere-turns per unit length, whence
Ba
value
ATa =
87.
ata
la
(29)
Ampere-turns Required
for
the
Pole Cores and Pole
from section to
pole cores and pole shoes varies section, being greatest near the yokes and gradually decreasing toward the gap as more and more of the leakage
Shoes.
The
flux carried
by the

flux is
129
shunted
is
off;
that the flux
but it may be assumed without sensible error uniform and equal to v$. We have then,
B =
c
V *&
and -jAC
B =^~ "
s
V*&
(30)
to which values of flux density there correspond unit excitations

of atc
and
at s
respectively; hence
AT =
C
at e 'l c
and
AT =
S
at a -ls
(31)
The flux carried 88. Ampere-turns Required for the Yoke. by the yoke is either equal to v$ or Y^v <, depending upon the type of machine. Fig. 101 illustrates the first case, and Fig. 102 the second case; the latter is representative of most modern
machines.
Then, usually,
(32)
y
to which value there corresponds at y ampere-turns per unit length,
and
AT =
y
aty-ly
(33)
89.
The
Coefficient of Dispersion.
The leakage
flux
<p
that
enters into the equation
includes the flux in
all
and originating
through
in the exciting winding,
paths associated with the main flux $ but which do riot close
mapped
the armature. If the leakage paths are correctly out, the stray flux in each of them is equal to the m.m.f. divided by the reluctance. The calculation can be simplified, and a fair degree of accuracy attained, by assigning simple
geometrical forms to the leakage paths. Since the greater part of the leakage flux takes place through air, the reluctance of the path in the iron may be neglected. It must be remembered also
that
all of
the leakage paths are not acted upon
by the same
difference of magnetic potential. For instance, the m.m.f. acting between the tips of adjacent poles is that required to drive the useful flux across the double air-gap, two sets of teeth, and
the armature, or
it is
a
equivalent to
g
t
AT + 2AT + 2AT
ampere-turns;
(34)
130
then points on adjacent pole cores that are each half way between the yoke and the shoe will have between them a difference
of
magnetic potential approximately equivalent to
^X ampere-
turns.
Let Fig. 119 represent a development of a portion of a multiThe leakage flux in any one pole P is represented polar machine. the dashed lines 2 <î, by etc., and the total leakage flux per
(f>
,
pole
is
<p
fyi
+
\r
4<p 2
+
~
2^3
4^4
-~:
(35)
)\
==
^\\
(r~=-~=

Leakage between Inner Surfaces of Pole Cores,
<ps
<p^.
131
f -f- (lengths expressed in centimeters) 1U Z Is

(38).
or
<ps
= 1.6X -y-^
(lengths expressed in inches)

,
the pole cores are round, of diameter d c they may be asto have been replaced by a square pole of equal crossIn that case section.
If
sumed
bc
lc
= ~ \/^ =
0.89d c
(39)
The above expression for p 3 is derived on the assumption that the axes ot the pole cores are parallel. This is approximately the case when the machine has numerous poles. If the poles
are considerably inclined to each other, let (Is) min and (ls)m ax represent their minimum and maximum separations, at the pole shoes and yoke, respectively; then it is readily shown that
_47T te
lc
hc
-I
(I'sjmax
('s/mt'n
-,
[\^)max
\^)max
or
_ O
<ps
(I'SJmin
(lengths expressed in centimeters)

.
(40)
o.
OV
vc ' vc
I
"V^max
|_
(I
)max
1 "j
(Is) min
Leakage between Lateral Faces of Pole Cores, <p. The leakage paths may be assumed to be made up of straight lines of length Is and of quadrants of circles. The average m.m.f. acting on each elementary tube of force is
47T
(lengths expressed in centimeters)

or
(41)
I
I*
The
coefficient of dispersion is
f
then
l+_
1
=
-
(function of frame dimensions)
(42)
132
If

the flux
<
the coefficient
were directly proportional to the excitation X, would be constant; but since X includes the excitation required to drive the flux through the teeth, and these
v
j-
are frequently highly saturated, the ratio

v is
is
not constant, hence
more or
less variable.
It generally increases as the load
on
the machine increases.
PROBLEMS
arid
The shunt-field winding of a 10-pole machine has 550 turns per pole 6 produces a flux of 9 X lT) lines per pole when the exciting current ia Find the inductance of the shunt circuit and the energy stored 18 amp. in the magnetic field produced by a skunt-field current of 18 amp.
1.
there is no magnetic leakage. Construct the magnetization curve of a 6-pole machine which has the following dimensions, assuming that the' coefficient of dispersion is 1.17, and that the rated speed is 375 r.p.m.
2.
Assume that
ARMATURE
;
(of
sheet steel)
External diameter of core Gross length of core No. of ventilating ducts Width of each duct Radial depth of core' below teeth No. of slots
46
12
in.
in.
%
6%
150 0. 56 1 80
.
in.
in.
Width Depth
of slot
of slot
in.
in.
AIR-GAP (clearance)
0.375
:
in.
POLE CORES
(cast steel)
Ratio of pole arc to pole pitch

of pole cores Radial length of core
0.7
12
in.
Diameter
13 in.
YOKE
(cast steel)
Radial depth Axial width
4
.
in.
15 5 in.
ARMATURE WINDING:
Type
Total number of conductors
Simplex lap 900

6
Conductors per slot No. of winding elements
450
COMMUTATOR:
Diameter
3.
Compute
2.
35 in. the coefficient of dispersion of the machine specified in
Problem
CHAPTER V
ARMATURE REACTION
90.
Magnetizing Action of Armature.
In the foregoing dis-
dynamo, it was tacitly assumed that Under this no-load condition the the armature was currentless. magnitude and distribution of the magnetic flux are dependent only upon the excitation due to the field winding and upon the shape and materials of the frame. But, under load conditions,
cussion of the behavior of a
the current in the armature conductors gives rise to an independent excitation which alters both the magnitude and dis-
This tribution of the flux produced by the field winding alone. magnetizing action of the armature is called armature reaction.
For the sake of simplicity, let us examine first the conditions If the armature is currentless, the flux in a bipolar machine. due to the field excitation will be symmetrically distributed in
the manner illustrated in Fig. 120. The line ab, drawn through the center of the shaft at right angles to the polar axis, is the geometrical neutral axis; it is an axis of symmetry of the flux
under no-load conditions.

sitely directed e.m.fs.
Armature conductors on opposite
sides of the geometrical neutral axis will then

If,
be the seat of oppounder no-load conditions, the brush axis coincides with the geometrical neutral, the winding elements lying in the neutral axis will be short-circuited by the brushes at the moment when they are not cutting lines of force, hence the short-circuit is harmless. If the field excitation is now removed and the armature is supplied with current from some external source, there will result a magnetic field whose distribution is approximately as shown in Fig. 121. Magnetic poles will be developed in the line of the brush axis. Most of the flux will be concentrated in the region
covered by the pole shoes, since the reluctance there than in the interpolar gap.
is
much
less
Under load conditions, the armature current and the field excitation exist simultaneously, and the resultant flux can then
133
134

fields
be thought of as compounded of the two
shown separately
in Figs. 120 and 121, at least as a first approximation. 1 The form of the resultant field is shown in Fig. 122, which serves
equally well for the cases of generator and motor action. It will be observed that in the case of the generator the field is strengthf ened at the trailing tips, A and A and weakened at the leading tips,
,
and B''; whereas
in the case of the
motor the exact reverse
is
Moreover, the neutral axis (that in which the winding elements are not cutting lines of force) has been shifted to the
true.
FIG. 120. netic field,
Distribution of mag-
armature currentless.
FIG. 121. Magnetic field due to armature current, field magnets not excited.
position a'b''; the effect is the same as though the flux had been twisted or skewed in the direction of rotation in the case of the
generator,
and
in the opposite direction in the case of the
motor.
As a
to be
result of the shift of the neutral axis, the brushes
(assumed
in the axis ab) short-circuit elements which are cutting lines of force and in which an active e.m.f. is being generated.
still
Currents of large magnitude
may
therefore flow in such elements
1 NOTE. It is not exactly true that the resultant field is made up of the What actually happens separate fields of Figs. 120 and 121 as components. is that the windings of the field structure and of the armature each produce a definite m.m.f., and that these m.m.fs. then combine to form a resultant The composition of the m.m.f., which in turn produces the resultant flux.
separate fields would give correct results only if the flux were at all points proportional to the m.m.f., and this condition is, of course, not satisfied in the presence of iron cores, especially if the iron is worked at a flux
density at or near the knee of the magnetization curve.
(See
98.)
ARMATURE REACTION
135
because of the low resistance of the circuit which includes the

short-circuited winding elements and the brush contacts; and the rupture of this circuit, as the commutator segments pass from
under the brush, will cause sparking and perhaps blistering of the commutator. Furthermore, the machine will not develop its full e.m.f.; for of the Z/2 conductors in series, say on the lefthand side of the armature of Fig. 122, those between b and b will generate an e.m.f. opposite in sign to that of the e.m.f. due to conductors between b and a. Both of these effects are
f
f
Indicates Current
Flowing from Observer

Indicates Current
Flowing Toward Observer
FIG. 122.
Distribution of magnetic field under load conditions.
objectionable, the former because it reduces the life of the commutator and lowers the efficiency, the latter because it unneces-
output of the machine. Obviously, the remedy for both troubles is to shift the brushes until they are in (or near) the neutral axis; but when the brushes are shifted,
sarily reduces the available
the armature
axis.
field (Fig. 121) moves with them in such a way that the resultant polarization of the armature coincides with the brush
The net result is that the resultant field tends to skew more and more as the brushes are moved toward the neutral axis.
Fortunately, however, the piling up of the flux in the pole tips A and A' ultimately results in their saturation, so that
further skewing becomes insignificant even pass the neutral.
91.
and the brush
axis
may
examine the commutation somewhat more in phenomena occurring during

Commutation.
It is desirable at this point to
136
detail than has yet been done, in order to settle in a general way the conditions that must be satisfied by the brush position. Fig. 123 represents three elements, a, b and c, of a ring winding operat-
ing as a generator. It is evident that element a will occupy successively the positions of b and c, and that during the b position its current
must change from the value existing in conductors to the left of the brush, to the equal and opposite value existing in conductors on the right. This change cannot occur instantaneously; in the ideal case the current
would change uniformly from
the initial to the final value in exactly the time required for the element to pass under the brush, during which time the element
FIG.
123.
armature current commutation.
Reversal of during
FIG. 124. Ideal variation of current in coil undergoing com-
mutation.
is
short-circuited in the
manner
of coil
b.
This
is
represented
are, respect-
diagrammatically in Fig.
ively, the initial
124, where
and
and final values of the current in the element, and T is the duration of the short-circuit, or the period of commutation.
Now, the
self-inductance of the element undergoing
commuta-
tion tends to keep the current at its original value and in the original direction, and in order to counteract this tendency, the
must be balanced by an opposing e.m.f. the were brushes If, then, exactly in the neutral axis, no e.m.f. would be generated in the short-circuited coil, the effect of selfinduction would not be opposed, and the reversal of the current
e.m.f. of self-induction
could not be completed in the time T.
Since the direction of the
e.m.f. required to cancel the e.m.f. of self-induction
must be the
as the final direction of the current (see coil c, Fig. 123), it follows that the short-circuited coil must be under the influence of the pole in advance of the neutral axis, in the direction of rotation.
same
ARMATURE REACTION
The
tral axis, is
137
position of the axis of commutation, with respect to the neushown in Fig. 125 for both generator and motor. It is
,
is
seen from this figure that the brush displacement, a m of a motor slightly less than the displacement, a g of an identical generator.
,
the angular displacement of the brushes is in the direction a generator, the angle is called the angle of brush lead; when in the direction opposite to the rotation, as in a
of rotation, as in
When
motor,
92.
of the
it is
Imagine the brushes armature of Fig. 121 supplied with constant current from some external source, the field being unexcited. If the brushes are rocked forward or backward, the armature m.m.f. will It may, follow, and it will remain constant in magnitude.
Components
called the angle of backward lead. of Armature Reaction.
l<--y
|
/ '
Generator
/^Brush Axis
Brush Axis
FIG. 125.
in generators
Axis of commutation and motors.
FIG. 126.
armature M.M.F.
Components
of
therefore, be represented by a line of constant length, , Fig. If the fields are now excited, their 126, in line with the brushes.
OA
be represented by a line OF (the direcand field windings being taken the same as in Fig. 122). Resolving OA into the components OC and CA, it is seen that the armature magnetizing action is equivalent to a cross-magnetization due to OC (so called because it acts across the main m.m.f. OF), and a demagnetization due to CA, which directly opposes the main excitation, OF. The demagnetizing action of the armature is a direct consequence of the rocking of the brushes to the position most favorable for commutation.
magnetomotive force
may
tion of current flow in armature
138
It will be clear from Fig. 126 that if the brushes of a generator have a backward lead, the armature will assist in magnetizing
the
field,
that
is,
izing.
Similarly
the demagnetizing component becomes magnetif the brushes of a motor are given a forward
lead; but in both cases the commutator will spark viciously. Were it not for the fact that a negative brush lead affects
commutation unfavorably, the armature reaction might be purposely exaggerated to such an extent as to self-excite the fields. This feature is taken advantage of in the Rosenberg type of genlighting (see Chap. XI), but in general it requires special auxiliary devices to take care of the commutation
erator for train
difficulties.
93.
Cross-magnetizing and Demagnetizing Ampere-turns.
Tho
resolution of the armature m.m.f.,
OA,
is
in Fig. 126, into
com-
ponents
not a strictly accurate

it is
proceeding;
than quantitative.
qualitative rather But it leads
directly to the conclusion that the entire armature winding of Fig. 127 may be considered as made up of two distinct " belts" of conductors,
namely,
those between CB, and those between
AD
CA
and and
BD. The former

Cross-magnetizing and demagnetizing belts of conductors. FIG.
127.
conductors,
when
grouped in pairs in the manner

indicated
constitute a
by the horizontal lines, number of turns whose magnetizing effect is directly across that of the main exciting The rewinding; they are called the cross-magnetizing turns.
maining conductors, grouped in vertical pairs, constitute the demagnetizing turns, since their effect is in direct opposition to the main exciting winding. It follows, therefore, that in a bipolar
machine the demagnetizing turns per pair of poles are equal to the number of armature conductors within the double angle of lead, 2a; and the demagnetizing (or back) ampere-turns per pair of
poles,
AT
d,
equal this number multiplied by the current per
conductor.
360
360
ARMATURE REACTION
Similarly, the cross-magnetizing ampere-turns are given
139
by
(2)
AT =
C
360
180
720
where
'
2a
It is important to realize that the demagnetizing turns are a consequence of the cross-magnetizing action of the armature; for if the brushes are originally in the geometrical neutral axis, the entire magnetizing action of the armature is across the main field,
thereby causing a distorted resultant field. The shifting of the brushes to a position near the resultant neutral axis then brings
the demagnetizing turns into existence.
polar Machines.
Cross-magnetizing and Demagnetizing Effect in MultiIn the foregoing discussion of the case of bipolar machines, a ring-wound armature was tacitly assumed.
94.
FIG. -128.
Reduction
of
demagnetizing action caused pitch of winding.
by
fractional
Accordingly, the brush axis and the axis of commutation coin-
and no distinction was made between them. It must be remembered, however, that the end connections of lap and wave
cided,
windings are generally so shaped that the brushes are opposite the middle of the poles when the sides of the coil undergoing commutation are in the geometrical neutral.
An extension of the principles developed for the case of the bipolar machine leads to the generalization that all of the conductors lying within the double angle of lead have a demagnetizing
140
effect

upon the
field,
while the remaining conductors produce a
This conclusion holds accurately for all cross-magnetization. windings of the ring type, and for lap and wave windings of full But in short-chord windings it will be found that the pitch.
conductors occupying the space between pole tips carry currents which are partly in one direction and partly in the other, thereby For example, partially neutralizing the demagnetizing effect. assume a 4-pole simplex lap winding having 80 conductors; that
is
= = Z = ra =
p
a
4 4
80
1
15 and y 2 = 13. On tracing through the winding a of which is shown in Fig. 128, it will be portion diagram, found that the current in the interpolar region is alternately in
Take
y\
opposite directions.
FIG. 129.
Demagnetizing belt of conductors
in multipolar
machine.
Omitting from consideration the special case of short-chord windings, the number of demagnetizing ampere-turns per pair of poles can be determined as follows (see Fig. 129)
:
ARMATURE REACTION
The
belts
141
total
number
of
conductors lying within the demagnetizing
is
~- 2a.p
and, therefore, the
pair of poles
is
number
i' /2
of demagnetizing ampere-turns per
AT
'
360
"
?p
p/2
it
'
180a
(3)
The cross-magnetizing ampere-turns produtee-disttfrtion of the field by strengthening the field at one tip of the pole and weakening it at the other, as in Fig. 130. But though all of the
main
conductors outside of the double angle of lead contribute to this effect, those
which are not in the angle
ft
subtended
by the pole exert their m.m.f. upon a path so largely in air that the flux due
to
them
is
negligible; attention
n
r?
may,
therefore, be confined to the ^p. con-
ductors under the pole.
The m.m.f.
due to these conductors

gilberts,
is
^
FIG. 130.
upon a path C whose reluctance is mainly due to the double air-gap and two sets of teeth; the teeth consume a m.m.f. equal to
and
this acts
1 J.V/
Cross field in multipolar machine.
n (2A T
gilberts,
hence the remainder, or
^
J-V/
l_OO\y
0-7.
- 2A T
CL
\,
I
will
produce a cross-field whose intensity at the tips
is
=
where
5'
Tn oTfri a 10L360
I
"
2A L
7777
(4)
d
t
is
'equal to the
slots.
gap length corrected to take
account of the effect of the
The
resultant pole tip densities will then be
/D (B g
D B
\ c)
at the
leading
.
) \
\ (
.,.
pole h tip of a
* [generator)
\
trailing
motor
142
and
(Eg
+B
c)
at the
[
i
trailing
,
]
\
,.
pole *
tip of a
f
\ [
generator e
1 \
.
leading
motor
since commutation takes place at the weakened pole tip, and since the direction of the "commutating field" must be that
Now,
of the original field,
Bg
it
follows that
<B
,in order to generate
in the short-circuited coil
an
e.m.f. of the proper direction to
balance the e.m.f. of self-induction.

c
Generally,
B g - B = 2000 to 3000 (lines per sq. cm.) and since B g is usually between 6000 and 10,000, it follows that B g = (1.25 to 2)B Substituting this relation in the expression for B and transposing
C c,
.
1
it is
(1.26 to 2)
, s S
1.6
B,
from which
possible to compute the length of air-gap necesreversal of the field at the commutating tip. lf\ to sary prevent
the clearance, 5, has been fixed, the formula gives an idea of the extent of chamfer to be given to the pole tips. The above formula also leads to a relation which serves as an
approximate criterion for a successful machine. ing the term AT we may write
t ,
Thus, neglect-
(1.25 to 2)
~^=
1.65,3'
= 2ATg
(6)
or
(7)
Now,
/?
360\i/
- where ^, the ratio of pole arc to pole pitch,

is
is
usually about 0.7 in direct-current machines; and Z

total
the
from
armature ampere-turns. Further, A T g will vary 0.7 to 0.9 of the ampere-turns per pole. Substituting these it will be found that relations, Armature ampere-turns per pole ^1.1 field ampere-turns per pole
number
of
(8)
The factor
ordinarily
1.1 is
it will
an upper limit that is seldom found have a value of from 0.8 to 0.9.
in practice;
ARMATURE REACTION
95. Corrected Expression for
143
Demagnetizing Effect of Back Ampere-turns. (a) Lap Windings. Fig. 131 is a development " chorded" (short-chord) lap winding emof that portion of a It is rebraced in a span slightly greater than the pole-pitch.
=yo7r^ due to the fact that the coils in the neutral zone carry currents which are not all in the same direction. In the first place, it will be evident that if the winding were of
quired to determine the reduction in the value of
ATd
2S
full-pitch, all of the
coil sides lying to
the
left of
would carry
current in the
same
direction (vertically
upward
in the figure).
FIG. 131.
Fractional pitch lap winding.
In the second place, however short the chording may be, the current in the coil sides immediately to the right 1 of, and including, AZ will all be in the same direction, or downward in the figure. It follows, therefore, that the reversed currents all lie in a zone to the left of A 2 the extent ot this zone depending upon the differ,
ence between y\ and the pole pitch. Similarly, there will be zones of reversed currents to the left of all the coil edges, which, like A i and A 2, are connected to commutator segments touched
by the brushes.
If coil
edge
is
numbered
the
first coil
is
edge carrying
re-
versed current
is
(y\ -f 1),
the second
(yi
3),
etc.
The
number corresponding 2S
,
to the last one in the group will evidently
be-~-'=yi
1
This
is
If the winding were left-handed, 7/2right-handed, i.e., y\ the words "right" and "left" would have to be interchanged.
is
a consequence of the fact that the winding sketched in Fig. 131 ?/i y,
>
<
144

results,
Summarizing these
2/i
yi
T/i
?/i
2/i
+ + + + +
3 5
(2k
/2
Si
corresponds to the 1st corresponds to the 2nd

1)
corresponds to the 3rd corresponds to the kih

\
(~
2/1)
corresponds to the nth
In other words, there are n of such reversed bundles, where
2n-l=
or
25
2/1
Since the current in each of these n coil edges balances the demagnetizing effect of the current in n bundles whose direction is
normal, the total reduction will be that due to 2n bundles.
Bi(nce
each element contains
â
turns,
ATd
will
be
less
than the com-
puted value by
Zb
It
(
tt
2/1+1) ampere-turns
(10)
should be noted that extreme chording

fall
may
cause some of
the n reversed coil sides to

lead,
outside of the double angle of
But such and, therefore vitiate the above correction. extreme chording would not be used in practice, hence the correction may be safely used. It should be noted, further, that neither the formula for A Td nor for the correction due to chording takes account of the number of coil edges in the neutral zone which are short-circuited by the brushes during commutation. If 2/1 is the back pitch of the (6) Wave Windings. (Fig. 132).
winding (at the pulley end), and mutator end),
2
2/2 is
the front pitch (at the com-
_ 2/i_+J/2
_ 2S
+a
p
the positive sign of a indicating that the winding is right-handed. The extent to which 2/1 falls short of the pole pitch is then a measure of the chording; obviously, then,
would
result in
Full-pitch winding 2/2 > 2/iuniform opposition of direction of current on
ARMATURE REACTION
either side of coil sides like AI,
145
in contact
A2
Fig. 132,
which are
with the brushes.
Due
to chording, however, the first conductor

is
which
carries
etc.
reversed current
(y t
1),
the second
is
(yi
3),
If
there are n such conductors, the number of the nth conductor will All of these conductors bear even numbers, be yi 1). (2n
~n
FIG. 132.
Fractional pitch
wave winding.
since y\
is
necessarily odd.
2/S
Now,
y
-
the a
-,
number
of conductors
(coil sides)
per pole pitch
is
which is a mixed number,

If
but which
may
number
or
in the
be taken as equal to y. 1, whence group is y

y
is
odd, the last even
2/1
(2n
1)
(11)
if
is
even,
= =
T/
(2n
(12)
is
2n
y *~~
hence the number of ampere-turns to be deducted from (ATd)

if
is
odd,
(13)
96.
Shape
of
Magnetic Field Produced by Armature Current.
The current in the armature conductors lying to one side of the commutated coil has a direction opposite to that of the current
in the
conductors on the other
side.
In other words, the current
10
146
is
distributed around the periphery in a series of alternately directed bands or belts, equal in number to the number of poles.
indicated in Fig. 133 which represents a development of a The peripheral distribution of m.m.f. due 4-pole generator. to the armature current will then be represented by the ordinates
This
is
of the
broken
line.
The m.m.f.
it is
the current reverses, and
is a maximum at the points where zero at the points midway between
Rotation
FIG. 133.
Peripheral distribution of armature m.m.f.
them.
be at
If
and the surface
the pole shoes completely surrounded the armature of the latter were smooth, the armature flux would
all points directly proportional to the m.m.f., since in such a case the reluctance would be constant all around the gap (neglect-
ing the reluctance of the flux paths in the iron in comparison with those in the air).
The number
of conductors
rr
surrounded by the elementary tube

is
of flux, P, Fig. 134,
is
-^ 2x where d
the diameter of the arma-
FIG. 134.
Calculation of armature
flux.
ture.
Since each conductor carries
amperes,
ia
being the total

~~
'
armature current, the m.m.f. acting on tube

4?r
jTj
P is^ r a 10 *d
2#
2x where q
= Z
i ,
is
the
number
of ampere-conductors
A RMA TVRE RE A CTION

per unit length of periphery.
147 x
The
is
flux density at a point distant
from the center of the pole
then
xq_
(14)
0.85
In the usual case of machines having pole shoes separated from each other by an intervening air-space, the flux distribution curve is not similar in form to the curve of m.m.f. Under the pole shoe
it will
closely follow the m.m.f. curve because of the practically uniform reluctance, but between the pole tips the reluctance increases at a much greater rate than the m.m.f., hence the armature flux density will be small at the point midway between
them, as shown in Fig. 135.
FIG. 135.
Distribution of armature flux.
97.
a, 6,
in
Approximate Distribution of the Resultant Field. Parts and c of Fig. 136 represent the effect of the armature field modifying the magnitude and distribution of the resultant
magnetic field for three positions of the brushes. In each diagram curve F shows the flux distribution due to the field excitation
is
is the flux curve due to the armature, and curve R alone; curve the resultant of F and A. The diagrams are drawn for the cases of commutation:
(a)
(6)
(c)
midway between pole tips, between the pole tips but nearer the leading under the middle of the poles.
tip,
In case (a) the distortion of the magnetic The symmetry of curve A with respect to
field is clearly
shown.
F means
that the flux
added to the trailing tip (generator action being assumed) is exactly equal to the flux removed from the leading tip, hence the flux per pole remains unaltered. In case (6) there is distortion
148

also a resultant demagnetization, as
it is
and
clear that the flux
under a pole is more subtractive than additive. In case (c) there is no distortion, but only demagnetization, as might be expected from the fact that the brushes have been shifted to such an extent as to eliminate all the cross ampere-turns and to raise the back ampere-turns to a maximum value.
FIG. 136.
Distribution of resultant flux for various brush positions.
In 98. Demagnetizing Component of Cross Magnetization. the preceding article the shape of the resultant field R was determined on the theory that it may be considered as made up of two
components: one, a field produced by the m.m.f. of the field winding acting alone; the other, a field produced by the armature m.m.f. As a matter of fact, this theory is not strictly acting alone.
correct, as the following illustrative analogy will
show:
ARMATURE REACTION
149
Imagine a rod of cast iron acted upon simultaneously by compressive and tensile stresses, and suppose that these stresses are If we assume that the stresses act independently in equal. deforming the rod, the elongation due to the tension would considerably exceed the shortening due to the compression, provided the tensile stress is beyond the elastic limit; on this basis there would be a resultant elongation. But it is quite clear from the assumed equality of the stresses that the resultant stress and, therefore, the resultant deformation are both zero, hence the
absurdity of the first method. In the case of the magnetic circuit, m.m.f. is analogous to Hence we must conclude that stress, and flux to deformation. the only correct procedure is first to combine the several m.m.fs.
and from the latter determine the and magnitude of the resultant flux. It will be clear from the above considerations, taken in connection with diagrams a and b of Fig. 136, that the increased
to
form a
single resultant,
distribution
of the pole shoe cannot raise the resultant on that side to the same extent as the diminished m.m.f. at lowers the flux on that side this because of the fact that the permeability of the iron of the pole shoe and armature teeth decreases with increasing magnetizing force. Therefore, even when commutation takes place midway between the poles, corresponding to zero brush lead and an entire absence of back am-
m.m.f. on the side

flux
pere-turns, there is still a resultant demagnetizing action
due to the cross ampere-turns;
though
not
99.
this effect is
usually
it is
very
pronounced,
nevertheless appreciable.
Excitation Required
Let
FIG. 137.
Amp. Turns per Pair

of Poles
under Load Conditions.

curve
137 represent the magnetization curve

of Fig. of a generator,
OM
Excitation required
under load conditions.
The
be the terminal e.m.f. at rated load. is then (AT) ampere-turns per pair of poles. When the armature delivers current to the load, the excitation required to maintain
and
let
OE
excitation required to generate this e.m.f. at no-load
150

in order to
constant terminal e.m.f. must be greater than (AT)
compensate
(a)
The ohmic
of the current
(b)
drop, or drop in potential, caused by the flow through the resistance of the armature.
effect of the
The demagnetizing
armature back ampereof
turns.
(c)
The demagnetizing component
the
armature cross
ampere-turns.
(a)
If
the terminal e.m.f. under load conditions
is
to remain the
same
as at
no load, the
total generated e.m.f., E\,
must -be greater

ia ra ,
than EQ by the drop

is
where
ra
the resistance of the armature.
Referring to Fig. 137, the excitation required to generate EI volts is
(AT),.
(b)
If
the armature demagnetiz-
ing ampere-turns per pair of poles (corrected, if necessary, to take

care of chording) amount to (AT) d the field excitation must be in,
FIG. 138.
of
Comparison of paths main and cross field.
creased over and above (AT), by v(AT) d that is, to the value (AT) 2
,
,
corresponding
e.m.f. of
It
to
an open-circuit
volts.
has been shown that the cross field is due to the con(c) ductors lying under a pole, and that the path of the cross field is The number of cross as shown by the line marked C, Fig. 138.
R7
7
'
ampere-turns acting upon this path
is
o^--
half of which oppose
B.
the main excitation at pole tip A, the other half reinforcing it. at These two excitations act on different magnetic paths, C and
F, but
be observed that the following parts are common to both paths: the double air-gap, two sets of teeth, the armature core The remaining parts of the two (in part), and the pole shoes. paths, namely, the pole cores and yoke in F, and the transverse path through the pole in C, constitute such relatively small percentages of the respective total reluctances that there is no
it will
appreciable error in assuming the two paths identical; and, there-
ARMATURE
fore, that the
REDACTION
151
two excitations may be combined to determine the (It resultant excitation, and therefore, also, the resultant flux. in cannot m.m.fs. more or two that understood will be readily act unless resultant a form to they upon general be combined one and the same magnetic circuit, just as forces cannot be combined to form a resultant force unless they act upon the same
body.)
It follows
from the fact that the magnetization curve (Fig. to the abscissas with per pair ampere-turns equal 137) is drawn the and one the at effect the tip, pole of poles that weakening to as due be taken must at the effect other, strengthening
@Z
i
-
360'fl
ampere-turns at each
of
tip.
In Fig. 137, therefore, this

d'
number
laid off to scale
ampere-turns is to be on either side of
(AT) 2, whence the resultant

diminution of the total flux per pole will be proportional to the excess of area abd over area bee.
In order that the total flux
still
may
have the magnitude which
will yield the desired e.m.f.,
the
excitation
must
,
be
inb'
f
FIG. 139,
f'(AT) 2
9 9'
creased to (AT) 3 on the curve
the point
Excitation under load

conditions.
being so selected that the following condition

area f'd'e'g'
OM
is
satisfied (see Fig. 139)
area 6
area a'b'd'
area/cfegr
(15)
The
same as though there were no due to the cross-magnetizing ampere-turns. demagnetizing The demagnetizing effect of the cross-turns is obviously represented by the difference between (AT) 3 and (AT) 2
total flux will then be the
effect
.
The curve
of resultant flux distribution 'in the air-gap
(R Fig.
t
136a) was obtained by assuming that the armature m.m.f.
produced a flux distribution as represented by curve A the ordinates of which were taken to be proportional to the m.m.f.
,
except at the tips of the poles. A greater degree of accuracy in the construction of curve A is possible through the application of the theory involved in Fig. 140. Thus, let curve M, Fig. 140,
152

let
be the magnetization characteristic of the machine, and
C be
that part of it required by the double air-gap, two sets of teeth, and the intervening portion of the armature core. When the
flux per pole is 0, curve abc will
show the
distribution of flux
under the pole due to the armature cross magnetization for the reason that curve abc relates to that part of the magnetic circuit acted upon by the cross-turns. The straight part of curve A, Fig. 136a, should then be replaced by a curve similar in shape to abc of Fig. 140 (see curve A', Fig. 136a).
Ampere Turns
per Pair of Poles
X
FIG. 141.
FIG. 140. Determination of flux distribution due to armature m.m.f.
Double
pilot brush.
100. Experimental
Determination
of
Flux
Distribution.
Since the instantaneous e.m.f. generated in an armature conductor is proportional to the radial component of the flux
density at the point occupied by the conductor at the moment in question (see Art. 32), the measurement of this e.m.f. will provide data for the calculation of the flux distribution.
Consider the case of a simplex ring- wound armature (Fig. 141) provided with such a large number of commutator segments that the turns of each element may be assumed to be concentrated in a radial plane, in other words, that all the turns of the element (if there be more than one turn per element) are simultaneously in a field of the same intensity. Take a narrow strip of tough paper
(sheet fiber or press-board) whose length is equal to the periphery of the commutator, and along its axis drill a series of small
whose spacing is the same as that of the commutator bars. Wrap the strip around the commutator and fasten it to the brush studs so that the commutator may rotate within it without
holes
binding.
The
free
ends of a pair of leads connected to a low-
ARMATURE REACTION
153
reading voltmeter are then to be provided with contact points made of moderately hard lead pencils. When the contact points are inserted into adjacent holes in the strip, the reading of the
voltmeter will be equal to the e.m.f. generated in the element minus the ohmic (ir) drop due to the current flowing through the resistance of the element, assuming that the experiment is made when the machine is running under load conditions.
Instead of the perforated strip described above, there may be " employed a pilot" brush made of two thin pieces of sheet brass
thickness
screwed on opposite sides of a strip of wood or ebonite, whose is such that the metal strips are separated by the dis-
tance from center to center of adjacent commutator segments. A similar arrangement will suffice in the case of a simplex
lap or a simplex
pitch;
wave winding, provided the elements are of full moreover, in the case of the simplex wave winding, the
voltmeter reading will be due to p/2 elements in series, instead If the winding is duplex or triplex, the of only one element. of the contact points must be two or three segments, spacing
if the winding is m-plex, the distance between contacts must correspond to m commutator segments. As stated above, the observed readings of the voltmeter are less than the true values of generated e.m.f. by an amount equal to the ohmic drop in the element (or elements) due to load In current if the machine is supplying an external circuit. order to eliminate this correction of the observed readings, an auxiliary "search" coil, of full pitch, may be wound on the armature, one of its terminals being grounded on the shaft and the other connected to an insulated stud on the end of the shaft. Connect one terminal of a voltmeter of the D'Arsonval type to the frame of the machine (or better, to a metal brush rubbing on the shaft) and the other terminal to a brush that makes contact
respectively; in general,
once per revolution with the insulated stud. If the moving coil of the voltmeter has sufficient inertia and is well damped, it will give a steady reading proportional to the e.m.f. generated in the
search coil at the instant
when the contact
is
established between
the brush and the rotating stud. If the brush is made capable of adjustment around the arc of a circle concentric with the shaft,
the contact can be
made
to occur
when the search
coil is in
any
desired position under the poles.
154

In Fig. 142 the ordinates
coils.
e\, e z , e 3 , etc.,
101. Potential Curve.
represent the e.m.fs. generated in individual therefore, that the expression
It is evident,
+
will
is
4-6,
be equal to the reading of a voltmeter one of whose terminals connected to the main brush (B) and the 'other to a single
FIG. 142.
Variation of voltage per element.
brush separated from B by an angle equivalent to the spread of n coils. If the pilot brush (P, Fig. 143) is moved around the periphery of the commutator and voltmeter readings (corrected for drop of potential if current is flowing) are taken at various points, a potential curve of the kind shown in full line in Fig. 144 will result.
pilot
FIG. 143.
Determination
of potential curve.
e, of this curve is the sum of the ordinates curve (which is the same as that of Fig. 142) lying to the left of e, it follows that the first derivative of the function which represents the potential curve will represent the curve of flux distribution, provided the winding is divided into a large
Since any ordinate,
of the dotted
ARMATURE REACTION
number
coil
155
of elements.
is
curve at any point
In other words, the slope of the potential proportional to the e.m.f. generated in the
corresponding to that point.
i
FIG. 144.
Relation between voltage per element and potential curve.
The change
cause of
102. Predetermination of Flux Distribution in the Air-gap. in the distribution of the air-gap flux due to the
is
magnetic reaction of the armature current

its effect
very important be-
upon the commutatmg characteristics of the as machine, explained in a preliminary manner in Art. 91 and in detail in Chap. VIII. It is therefore occasionally desiragreater
designing a new machine, to be able to predetermine the curve of flux distribution due to the field excitation alone (curve
ble, in
F, Fig. 136),
and also the curve of flux distribution due to the armature m.m.f. (A, Fig. 136). Several methods for determining these curves have been developed, but all of them, except that of Carter, are approximate; and Carter's method, though math1
ematically correct, is derived by assuming a simple shape of pole core and pole face that is not ordinarily used in practice. For determining the curve of field flux distribution, the method
J
W.
E. Goldsborough, Trans. A.I.E.E., Vol.
XV,
p. 515;
Vol.
XVI,
p.
461; Vol. XVII, p. 679.

S. P.
Thompson, "Dynamo
Electric Machinery," Vol. II, p. 206, 7th ed.
XXXVIII, p. 884 (1901). E. Arnold, "Die Gleichstrommaschine, Vol. I, p. 320, 2d ed. T. Lehmann, Elektrotechnische Zeitschrift, Vol. XXX, pp. 996
F.
W.
Carter, Electrical World, Vol.
and 1019
(1909). B. G.
Lamme, Trans. A.I.E.E., Vol. C. R. Moore, Trans. A.I.E.E., Vol.
XXX,
XXXI,
Part 3, p. 2362 (1911). Part I, p. 509 (1912).
156

of the
recommended by Arnold involves mapping out the paths

lines of force in
the air-gap and in the interpolar spaces in the manner illustrated in Fig. 145. This leaves much to one's judg-
ment, but some guidance is afforded by the consideration that the lines are substantially perpendicular to the surfaces of the pole faces and of the armature where they enter or leave the iron;
FIG. 145.
Distribution of
main
field.
it is
also true that the flux will distribute itself in such a

is
manner
a minimum, or, what is the same thing, for a given m.m.f. between the pole face and armature If, then, more than surface, the total flux will be a maximum.
that the total reluctance
one
trial is
made, that one
will
be most nearly correct which
yields the largest total flux.
FIG. 146.
Curve of
flux distribution.
central part of the pole, where the air-gap (5) is flux the uniform, density (B g) will also be uniform and inversely at any other point a tube of force of length d x to 5 proportional and mean section b x (taking a unit length along the core) will
;
Under the
have a permeance
surface will be
b x /8 x
hence the flux density at the armature

bx
8 6X
"
Br =
If
m.m.f.
d
6 X $x
ax
ax
(16)
B g is taken as
100 per cent., values of
B x can then be found
from
ARMATURE REACTION
the scaled values of
d x , bx
157
results
and ax and the computed

,
when
plotted along the developed armature surface will determine a curve like R, Fig. 146. Curves R' and R" represent portions of similar curves for adjoining poles (of opposite polarity), so that
the resultant of R, R' and
R" gives the desired
curve, F.
The area
FIG. 147.
Magnetic
lines of force
due to armature current.
of one loop of curve F, multiplied by V (the corrected length of armature core) must then equal 3>. The determination of the curves of flux distribution due to the armature m.m.f. (curve A, Fig. 136) is more difficult than in the case of the field flux, but the same general method is applicable.
Thus, in Fig. 147 are indicated the paths of the
lines of force
FIG. 148.
Calculation of flux distribution due to armature m.m.f.
emanating from the armature. At any distance x from the brush the permeance of the tube of force of unit depth parallel
x to the shaft will be equal to ,
and
if
Ox
the peripheral distribution
of
of
is plotted, as curve of Fig. 148, and the ordinates are multiplied by the corresponding ordinates of the curve of armature m.m.f. (M.M.F.), the resultant values will give
permeance
curve
158
the curve of armature flux (A). The field due to the armature m.m.f. has greatest influence in the axis of commutation, shown
at B in the figure; in the paper by Lamme (loc. tit.) it is recommended that the mean path of the flux issuing from the axis of commutation be taken as the arc of a circle extending to the middle point of the pole core, and intersecting the
surfaces of armature
and pole core
at right angles.
PROBLEMS
1.
six
The machine specified commutator segments.

amp. ?
in
Problem
2,
Chap. IV, has a brush lead of
How many ampere-turns per
to compensate the demagnetizing action of the armature

is
when
pole are necessary the machine
delivering 600
2.
field intensity at the pole tips due to the cross-field of the armature of the machine referred to in Problem 1, when it is delivering 600 amp. Make the calculations on the assumption that the field excita-
Compute the
tion
sufficient to give an average air-gap density of 50,000 lines per sq. in. Under the conditions assumed in Problem 2, what additional field excitation is necessary to compensate the demagnetizing component of crossis
3.
magnetizing action? 4. The armature resistance of the machine specified in Problem
2,
Chap.
When the machine is deliverIV, measured between brushes, is 0.016 ohm. 'ing an armature current of 600 amp., a voltmeter connected to a double
pilot
brush spanning adjacent commutator segments gives a reading of What is the field intensity in the region occupied by the winding element that is connected to the commutator segments touched by the pilot brushes?
10.67 volts.
CHAPTER
VI

103. Service
Requirements.
The lamps, motors,
or other
translating devices supplied with electrical energy from a distribution circuit may be connected to the supply mains in parallel,
in series, or in series-parallel, as
shown diagrammatically
in Fig.
149a,
b,
and
c.
Parallel connection (a)
is
used when the individual units con-
stituting the load on the system are designed to operate with a constant difference of potential between their terminals, and in
such a system any individual load unit can be disconnected
(6)
'oOOOO
FIG. 149.
Parallel, series
(O
and
series-parallel circuits.
without interfering with the operation of the remaining units; as examples of this class of service may be mentioned the use of incandescent lamps for interior lighting, and street railways
operating with constant difference of potential between trolley
and
rail.
Series connection (b) is used principally in arc-lighting and in series-incandescent flighting of streets and alleys, where each
lamp requires the same

this
current.
159
If
system
is
to be cut out of service,
an individual load unit in it cannot be disconnected
160
as in
the parallel system, but must be short-circuited by a "jumper" connection, as in Fig. 1496 above, in order to preserve
continuity of service in the remaining units. Series-parallel distribution (c) is merely a combination of the
other two, and requires no special consideration, once the principles underlying the series and the parallel systems are understood.
A common
example
of series-parallel connection is
found
in the lighting circuits of a trolley car, where several strings of five 110-volt lamps in series are each connected across the
550-volt supply circuit. If in a constant potential (parallel) system there are lamps (or other load units) each taking i amperes, the total current
supply is Ni amperes, and the power consumed, neglecting the loss Ni E watts, where E is the line voltage. In a constant current (series) circuit, if there are lamps each requiring e volts and i amperes, the total impressed e.m.f. must be Ne volts and the power consumed will be Ne i watts, again neglecting the line loss. In the first case (parallel system) the line conductor must have a cross-section at the supply end capable of carrying
in the line, is
off as the end of the line is approached. In the second case (series system) the line conductor will be of uniform cross-section from end to end, since the current is everywhere the same, but the difference of potential between the supply lines will be large at the generator end (G) and will gradually decrease as the distance frorn the generator increases. The parallel system requires a much greater weight of copper in the line than the series system, but this disadvantage is offset by the fact that the high voltage required in a series circuit
Ni amperes, and gradually tapering
any considerable power limits the use of series circuits to outdoor service. Thus, if 125 arc lamps, each consuming approximately 50 volts, are connected in series, the total voltage consumed by the lamps will be 6250 volts; adding to this the voltage consumed in overcoming the resistance of the line, the e.m.f. required at the generator will be of the order of 7000 to 8000 volts, or much too high for safety in indoor service. Although the parallel system of distribution is ordinarily called the constant-potential system, it will be readily apparent
of
that the difference of potential between the conductors will vary more or less from point to point, becoming less as the distance

from the generator increases, because
to the resistance of the line wires. to ohmic resistance
of the
161
drop of potential due This drop of potential due
may
be reduced to any desired amount by
increasing the cross-section of the conductor, but it is clear that a If limit is set by the rapidly increasing cost of the conductors.
the lamps (or other translating devices) are grouped at a distance from the generator, the voltage at the lamps may be kept constant, irrespective of the current in the feeder circuit, provided the voltage of the generator is raised, as the load increases, to a
compensate the drop of potential in the line. In view of the various types of service requirements described above, it becomes important to
sufficient extent to
104. Characteristic Curves.
investigate
the
characteristic
behavior
of
the usual types of
generators in order to determine the kind of service to which each is adapted. Probably the simplest way to study and compare the
several kinds of machines
is
to construct characteristic curves
which show the relations between the variables involved in the For example, the external characteroperation of the machine.
istic of
a generator is a curve showing the relation bet ween terminal voltage, plotted as dependent variable, and external (line) Other characteristic current, plotted as independent variable.
curves are discussed in following articles. In the case of a generator, the terminal vol105. Regulation.
tage at
load is generally different from that at no load. The between the two values is then a measure of the closeness with which the machine regulates for constant voltage; the
full
difference
difference
is
called the voltage regulation.
In order to
make
this
measure a perfectly definite one, so that machines of different makes and sizes may be compared, the Standardization Rules of the American Institute of Electrical Engineers define the percentage voltage regulation
(or
difference between the full-load
simply the regulation) as the and no-load voltages, divided by the

100. 1
full-load voltage,
1
and multiplied by
The new
(1914) Standardization Rules define percentage regulation as
change in the quantity occurring between the two loads to the value of the quantity at either one or the other load, taken as the normal value." Inasmuch as the full-load voltage is usually considered the normal voltage, it would he used as the divisor in obtaining the
''the percentage ratio of the
percentage regulation, hence the definition given above.

11
162
Similarly, if a machine is designed to regulate for constant current, as an arc-light generator, its regulation would be computed in like manner by dividing the difference between full-load
and no-load currents by the
full-load current.
The speed
regu-
lation of a motor, engine, turbine, etc., in terms of speeds at full-load and at no-load.
would be similarly defined
106. Characteristic Curves of Separately Excited Generator. The following symbols will be used:
E = E =
t
generated e.m.f. terminal voltage

resistance
of
ra
armature, including brushes and brush
contacts
= R =
i
if
resistance of field winding
resistance of external load circuit
= = = n/ n =
current taken
by load
current in field winding field turns per pair of poles
speed of rotation, r.p.m.
No-load Conditions.
being driven at
its
Under no-load
conditions, the armature
rated speed, the relation between the e.m.f. generated in the armature winding and the exciting current in the field winding is given by the magnetization curve discussed
in
Chap. IV
(see
Fig.
100).
is
Since the generated e.m.f. given by the equation
a
it
60
if
X
if
10 8
follows that
<S>
is
kept conconstant),
will
FIG. 150.
Effect of variation of speed
stant (by keeping

,,
upon magnetization curve.

speed.
If,
v^rmied
hp
the
directly proportional then, E, ifj and n are plotted along three axes of coordinates, there will result a surface of the kind illustrated in Sections of the surface cut by planes parallel to the Fig. 150.
to
(E,n) plane are straight lines whose slope increases as the distance of the section from the (E, n) reference plane increases
163
Sections cut by planes at first rapidly, then more slowly. are to the (E if} plane magnetization curves corparallel to the speed represented by the distance of the responding
t
secant plane from the origin of coordinates. External Characteristic. Load Conditions.
in Fig. 151, let the
With the connec-
machine be driven at its rated tions shown a current ?'/, and the brushes excited constant speed, the field by an lead most favorable for good commutation. set with a angular The line current (which is here the same as the armature current)
FIG. 151.
Connections for determining no-load characteristic.

excitation,
Separate
vary as the external load resistance R is varied, and the terminal voltage E will be less than the generated e.m.f. E by ira volts, the latter being consumed in the internal resistance of the
will
t
armature.
That
is,
E = Et
ira
(1)
In this expression
ra
armature winding itself, but the resistance
comprises not only the resistance of the of the brushes and their
connections, including the contact resistance between commutator and brushes. While the resistance of the armature is con-
when the steady working temperature has been reached, the contact resistance is not constant, but varies approximately inversely as the current; that is, the total drop of potential at the
stant
contact surface between commutator and brushes is approximately constant, and is of the order of 2 volts with ordinary grades of carbon brushes, provided the current per sq. in. of contact area does not exceed 45
sq. cm.).
amperes
(or 5 to 7
amperes per
164
In the case of copper brushes, which are used only with lowvoltage machines, the contact drop is of the order of 0.04 volt with current densities ranging from 65 to 160 amperes per sq. in.
(10 to 25 amperes per sq. cm.).
Strictly, therefore,
E = Et
ira
&e
(2)
where
is
ra
'
is
the. constant part of the
armature resistance, and Ae

In what follows
the drop over the variable contact resistance.
this correction is ignored for the sake of simplicity.

External ^Characteristic
FIG. 152.
Construction of external characteristic of separately excited

generator.
In Fig. 152 let O'M be the magnetization curve plotted with ampere-turns per pair of poles (n f if ) as abscissas and generated e.m.f. (E) as ordinates. If the excitation is adjusted by means of the field rheostat (Fig. 151) until it has the value represented by the abscissa OF the generated e.m.f. will be F G Q = OP and this will also be the terminal voltage when the machine is running
, ,
on open
circuit
(i
0,
R =
<).
of such reclosing the external circuit through a load sistance that a current of i amperes is drawn from the machine,
On
the terminal voltage drops below its no-load value; (1) because the demagnetizing action of the armature reduces the useful

flux
<f>,
165
and therefore
drop
in the
also the generated e.m.f. E', (2) because

.
In Fig. 152 let OA represent armature, ir a at an angle of 45 deg. and draw load current (i) any value of = = an such that at Draw i. OD OA so that OH angle
of the
AH
to the scale of the figure, then the intercept
HD
de,
magnetizing ampere-turns per pair of poles. Join H and F Then OF is the resultant field exand draw DF parallel to HF = OK = generated e.m.f. (E). Draw the line citation, and FG 0V through the origin at an angle <p such that
.
tan
to the scale of the figure. to the scale of volts, and if
<p
ra
The
intercept
AS
will
AS
is
deducted from
OK
KP
parallel to 0V), the difference, or .AP, will = i. If the load curvoltage corresponding to the current OA rent has a value represented by OA', the change being brought about by a change in the load resistance, a similar construction,
then equal ira (by drawing be the terminal
indicated in the figure, shows that the corresponding terminal The locus of all such points as P, P', etc., is voltage is A'P'. then the external characteristic of the machine. The line 0V
whose ordinates represent the ohmic drop in the armature is sometimes called the loss line. This method is subject to small errors because of the fact that it neglects a possible demagnetizing effect due to cross magnetization (Chap. V) and also due to the short-circuit currents The former may be taken in the coils undergoing commutation.
into account,
necessary, by slightly increasing the angle 0. It will be observed that the form of the external characteristic
if
P P is
dependent upon the form of the magnetization curve O'M, and <p. There will be a different as well as upon the angles
characteristic corresponding to each setting of the field excitation, OF The student will find it very instructive to run through
.
the construction using a value of field excitation such that the point GO falls on, or slightly below, the knee of the magnetization curve.
107. Effect of
istic.
Speed of Rotation on the External CharacterFor a given value of the excitation and, therefore, of the
166
flux,

the generated e.m.f. will be proportional to the speed. In is operated at a speed other than the rated speed,
case the machine
must be replaced by a new curve whose same relation to those of the original curve that the new speed bears to the rated speed. This is shown in the three-dimensional (diagram of Fig. 153. The latter figure also
curve
of Fig. 152
O'M
ordinates bear the
Po
FIG. 153.
Effect of variation of speed upon external characteristic of separately excited generator.
shows that the locus of the external characteristics for varying

values of speed (but with a fixed value of field-exciting current) a wedge-shaped surface, OP P.
is
108. Load Characteristic. By a load characteristic is meant a curve showing the relation between terminal voltage (as ordinates) and field excitation (as abscissas), subject to the condition that the current supplied to the load is constant. If this load current happens to be zero, the curve becomes the familiar no-
Now load characteristic or magnetization curve, OM, Fig. 154. circuit external suppose that the resistance R (Fig. 151) of the
is
varied, and that the field excitation is so adjusted that the curThe I. rent is maintained at its normal full-load value, i
167
terminal voltage will then be E t = IR, represented in Fig. 154 by the ordinate OA. At no-load this terminal e.m.f. would
OC, but under the assumed condiOA by an amount = e.m.f. DF Ir that the where to be generated a is, required DF, to yield a terminal voltage E = OA is DG, corresponding to an excitation OG. Finally, because of armature demagnetizing effect, the field excitation must be still further increased to OK, where
require an excitation tions the generated e.m.f.

,
AB =
must be greater than

t
GK =
An
ordinate through
v(AT) d
on the load characteristic corre-
K then intersects A B (extended) in the point

lies
P, which point accordingly

Since
sponding to full-load current.
DF
and
GK = FP
long as i = I, the load characteristic but no-load characteristic, shifted downward and to the
right
remain constant in magnitude as is the same in shape as the
DP,
as
by the constant length shown by the series of

is not with in-
ad Characteristic
shaded triangles. This construction

strictly accurate, for
creasing
excitation
the
de-
component of cross-magnetization becomes

magnetizing
greater, especially if the iron is saturated; in other words,
~o
eld
Amp-Turns
c G
-
FlG 154.
teristic,
should increase in magnitude as the curve rises. Further, remembering that the cient of dispersion, v, is itself not constant, but that
r,
n r
Const ruction of load characseparately excited generator.

coeffiit
in-
creases with increasing excitation (X increasing, equation (34), Chap. IV), it is clear that this feature will contribute to a further
increase in FP.
109.
The Armature
Characteristic.
It is evident
from
Fig.
154 that if the terminal voltage is to be maintained constant for all values of the load current, the excitation must be increased
as the load increases.
i
Thus, an increase in the current from
to
I requires an increase of excitation from
OC
to
OK
168
The curve in order to maintain a terminal voltage equal to OA. showing the relation between field excitation (as ordinate) and
load current (as abscissa), under the condition of constant terminal voltage is commonly called the armature characteristic " regulation curve." (Fig. 155) though a better name would be
Terminal Voltage = Constant
Load Current
vm/mwm
FIG. 156.
Armature characteristic FIG. 155. or regulation curve.
Connections of series
of
generator, determination ternal characteristic.
ex-
110. Characteristic
Characteristic.
Curves of the Series Generator.
External
Let
E = E =
t
generated e.m.f. terminal voltage

current in the circuit
resistance of armature winding, including brushes
ra
= =
and brush contacts
= = n/ R = =
77
resistance of series field winding

series field turns per pair of poles resistance of external circuit
speed in r.p.m.
(i)
Since the same current
winding, and the load
circuit
flows through the armature, the field (Fig. 156), it follows that an
increase of load causes an increase of excitation
and therefore
also of generated e.m.f., the speed of rotation being kept constant at its rated value. The external characteristic will have the form
of curve III, Fig. 157. If, then, to the ordinates of curve III there be added the ordinates of the loss line, curve II will result.
Curve
II
is
the internal characteristic,
showing the relation

between
current.
169
the
internally generated This follows beoausfi,, ir
e.m.f.
and the armature
E = E
If
(4)
there were no armature reaction, curve II would be the magnetization curve of the machine; but in the actual machine,
where armature reaction exists, the magnetization curve (I) is displaced from curve II in the manner indicated in the figure.
Thus, of the excitation OA required to produce terminal AP and generated volta.ge e.m.f. AG, a part, DG, is required to balance the demagnetizing component of armature reaction.
II
The remainder,
or OF, is then responsible for
the generated e.m.f., hence D is a point on the magnetization

curve.
this process to find other points, such as D', it will
Repeating
O
FIG. 157.
A'
External characteristic of
series generator.
be seen that the

angle
size of the tri-
PG
and
PGD must be altered from point DG are both proportional to the

of
to point.
current,
i,
But
since
their ratio,
and consequently the slope
DP,
will
remain constant.
This
leads to the following simple construction for obtaining the external characteristic from the given magnetization curve of
the machine. 1
= OA
The demagnetizing is given by
effect
AF
aZi 180a
corresponding to any current
1
"
''
nf
to the scale of current,

scale of volts.
and the length FB is equal to ira to the These two lengths, when laid off as in Fig. 155, To find a point P' locate the point B, therefore also the line OB. on the external characteristic corresponding to current OA'', draw
1
This
I.
is
the same construction given in Arnold's Die Gleichstrommaschine,
Vol.
170

f
A'B
at an angle
tan" 1
through B' draw B'D' verf

,
tically until it intersects
curve
M in D
to B'A' until
P'.
it
intersects the ordinate through
and draw D'P' parallel A' in the point
111. Dependence of the Form of the Characteristic upon Speed. Variation of the speed of a series generator affects the magnetization curve in exactly the same manner illustrated in Fig. 150 for the case of the separately excited generator. Thus, in Fig. 158, the surface bounded by OD'D is the locus of the
FIG. 158.
Effect of variation of speed upon external characteristic of series generator.
magnetization curves for various values of speed laid off along the speed axis n. Corresponding to each magnetization curve there will be an external characteristic constructed as in Fig. 157, and the locus of all such external characteristics will be a surface indicated
by the heavy
lines, as
OAoP.
The
intersection of this
,
surface with the base (n, i) plane is a curve OA which shows the relation between speed and current when the machine is shortcircuited
(E =
t
0).
112. Condition for Stable Operation.

it will
usually be found that
when the machine
Referring to Fig. 156, is driven at its
171
rated speed there is a critical value of the load resistance R above " which the machine will fail to generate, or to build up." When the load resistance has been lowered slightly below this critical
value, the terminal voltage and current will at first rise rapidly and then more slowly until a condition of equilibrium is reached,
but between the
initial
and
final
condition the machine
is
in a
state of unstable electrical equilibrium. Further reduction of will cause the current and e.m.f to change, but there is no further
.
evidence of instability. The reason for this behavior will be evident from a consideration of Fig. 159, in which the curve represents the external characteristic of the generator.
is
It
evident from Fig. 156 that Ohm's law must hold for the
external circuit, or
Et
R\
E =
t
iR
of
This
is
the
line
equation
straight
through
of
the
origin, the slope
the line
to
3,
being
proportional
R.
etc.,
Thus, ORi, correspond to
OR 2 OR
,
successively smaller values of R, and these

lines are characteristic of the
Condition for stable operaFIG. 159. tion of series generator.
external circuit.
values of
and
Since the points representing simultaneous must satisfy the characteristics of both gen-
their intersection.
erator and external circuit, the point of equilibrium will be at Thus, when the external resistance is high, as Ri, the terminal voltage and current will be represented by the
coordinates of point PI.
When R = R 2
the line
OR%
coincides
for a greater or lesser range with the external
characteristic,
P" 2
hence there is unstable equilibrium between the points P' 2 and Values of R such as R 3 and R^ will give stable equilibrium
.
at points P 3 and 4 respectively. 113. Regulation for Constant Current.

,
Inspection of the con-
struction of Fig. 157 will

series
generator will
action and internal
show that the external characteristic of a droop more and more as the armature reIn fact, by voltage drop become greater.
172
purposely exaggerating^, he magnitudes of these two quantities, the characteristic may be made to bend over to such an extent as
to
become nearly
r
vertical, in
which case the machine has an
in-
herent tendency to regulate for constant Such a characteristic is desirable current.

for series arc circuits, and some of the older types of Brush arc-light generators
were designed on this principle. Another method of regulating a

FIG.
160.
series
Diagram
showing connections of
constant current generator.
generator to make it deliver a constant current is to provide a variable resistance shunted around the series-field winding,
as in Fig. 160. It is clear that a portion of the exciting current is then by-passed
around the winding, thereby cutting down the generated e.m.f. Every variation of the resistance of the load must then be accompanied by a corresponding change of the regulating shunt. The relation between the resistance of the regulating shunt and
the corresponding terminal voltage of the machine is
shown
in Fig.
in a simple 161.
manner
Let
OM
be the magneti-
zation curve, with abscissas equal to the current in the

field
winding
(instead
let
of
ampere-turns), and
OA
is
be the constant curthat the

to
rent
machine
required develop; other words, the external

characteristic
vertical line
is
in
to be the
AY.
Lay
off
AB to
any convenient scale to the resistance (r/) equal
FIG. 161.
Regulation for constant
current, graphical solution.
of the series field winding, and to the same scale lay off equal to any arbitrarily selected value of the resistance (r) of
BC
BC
=
the regulating shunt.

will
AB
will
be fixed in magnitude while
be variable.
Vertically
upward from
set off
AD =
irf

constant, also
off
173
set
DN
i
AG = OA =
= ir a = constant, and downward from A = constant.

i,
The
total current,
will divide
between
r/
and
r in
such a man-
ner that
ir
in
=
TJ ?
(5)
and
^
ITS
(6)
Hence,
that
if
C and G
are joined
by a
parallel to
CG, point
ir
will divide
r/
.
is drawn straight line and = i into two parts, such
BH
AG
AH
parallel to
winding.
HG = i Joining G and 0, and drawing HK OK will be the current through the series-field = v ^7. or the equivalent deSetting off KL
and
GO,
,
magnetizing current of the armature, OL will be the net excitation of the machine, and the corresponding generated e.m.f.: if
LM
MP
is
the ohmic drop in the machine, or ilra

t
-\
--
),
the
terminal voltage will be E = LP = AQ. To find MP graphically, proceed as follows: Connect C and D and draw BF parallel to CD-, then DF = Thereand NF = i(r a + 7~r) i^r^r
fore,
point
parallel to
P is MN.
found by joining
Finally, therefore,
and
is
and drawing
Other
FP
AQ
the terminal voltage

points
corresponding to the shunt resistance r = BC. may be found by exactly similar construction.
It will
be evident from the above discussion that the voltage
of a series generator can also be controlled by varying the position of the brushes, thereby changing angle a and affecting the in Fig. 161. length
KL
114. Characteristics of
the Shunt Generator.
Open-circuit
Conditions.
Let
generated e.m.f. terminal voltage armature current
external or line current
shunt-field current
E = E =
t
ia i
is
ra
= = = =
armature resistance
174

re
shunt-neld resistance, including regulating rheostat
R =
resistance of external load circuit

of poles
n a = field turns per pair n = speed in r.p.m.
When the load or receiver circuit of a shunt generator is disconnected, as in Fig. 162, the armature and shunt field constitute a simple series circuit identical with that of Fig. 156. It is therefore easily seen that variation of the shunt-field rheostat will give rise to changes in and i a in the manner already discussed in the
case of the series generator.
There
is,
however, this difference,
FIG. 162.
Connections of shunt generator.

characteristic.
Determination of external
that the high resistance of the shunt-field winding will limit the flow of current (i a ) to values that are small compared with the
current carrying capacity of the armature, therefore the observed readings of E t under the conditions assumed will not differ appreciably from E, the total generated e.m.f. nal circuit open
Thus, with the exter-
E = E +
t
2> a
= E
(7)
rent in the armature
Moreover, the small flow of curnegligibly small armature reaction, hence the relation between E and i s will be closely represented by the magnetization curve of the machine.
ia
since both
and
ra are small.
means
Since
i8
-p
,
fa
the effect upon the terminal voltage of changing
the value of
r a will
be given by a construction identical with that
of Fig. 159; generally there will be a region of unstable equilibrium in the building up of the generated e.m.f.
The External
Characteristic.
The form
of the external char-

acteristic
line
t
175
current
showing the relation between terminal voltage E and i can be determined by the following method:
it is
From
Fig. 162
evident that
'
E = E +
t
iara
(8)
i.
(9)
(10)
between the generated e.m.f E, and the field excitan i s s tion, (ampere-turns per pair of poles), is given by the magIf the machine is running on netization curve, M, Fig. 163.
The
relation
.-,
open
circuit
(R
<),
let
the resistance of the field -rheostat be

,
so adjusted that the excitation is represented by OF the generated e.m.f. then being F QL] this will then be nearly equal to the
terminal voltage on open circuit, neglecting the small drop (i s r a ) in the armature. The line ON is then the "field resistance"
77f
line, its
slope being
lines
4 =
Tls^s
J
,
that
is,
proportional to r s
it
corre-
fl>s
sponds to the
ORi, ORz,
etc., of Fig. 159.
the external circuit be closed, R being so adjusted that a moderate current will flow. Then, even were the excita-
Now
let
tion to remain constant, as in the separately excited generator, the terminal e.m.f. would fall because of the ohmic drop in the
armature winding,
i.e.,
from equation
(8)
'
.'*, =
But, in the case
voltage
since
is is
-p --
-,>.
(12)
the shunt machine, a decrease of terminal accompanied by a proportionate decrease of excitation,

of. rs
hence,
when there is an appreciable load current
flow-
ing, the flux
and the generated e.m.f. are reduced, thereby causing It is clear, therefore, that the greater a further reduction of E the load current the less will be the terminal voltage, and that the
t
.
drop of terminal voltage will be greater in the shunt machine than in the separately excited machine, other things being equal. Suppose that the load has been increased to such a value that
1
Franklin and Esty, Elements of Electrical Engineering, Vol.
I.
176

GV,
Fig. 163; the
line
the terminal voltage has fallen to the value
problem is then to locate the point P on the horizontal such that P is a point on the external characteristic.
CVP
V draw the horizontal line VC intersecting 07V in C; If there were represents to scale the new value of n s i s no armature reaction, the ordinate FG would be the total generThrough
then
OF
ated e.m.f. corresponding to the excitation n s i s

fore, since
iar a
OF, and, there-
= E - E
/
t
)
.
CG would represent to scale the value of ia ra But since armature reaction does exist, the net excitation is less than OF by
an amount FD, where
aZi n
F F D
FIG. 163.
-BH
Construction of external characteristic of shunt generator.
Actually, therefore, the net excitation

e.m.f.
is
is
OD, and the generated
E = BD
ia r a
= E - E = BD t
HD
tan V
=
=
and
BH = BH r<ij ~^n
LtL
ar
ru
aiia 180a
180ar -T~~
vaL
constant.
In other words, when a point C on the field resistance line has been fixed, point B is found by drawing through C a line
ON
CB
making the fixed angle <p with the horizontal. Through B draw the horizontal line BK, and through
draw
111
KP
that
at an angle 6 with the horizontal, this angle being so chosen
tan 6
to the scale of the figure.
ra
It follows, then, that
KV
tan0
*> _
hence P is a point on a curve whose ordinates are terminal voltage (Et) and whose abscissas are total armature current (ia ). Corresponding values of line current (i) can then be found by
subtracting
OP
at
this can be done graphically by drawing the line an angle ^ such that tan ^ = rs to the scale of the figure;
is
-
for it
is
easily seen that
0_F*,
tan $
rs
Hence AP = E and XP =
t
are simultaneous values of terminal
voltage and line current. Similar construction will then serve to locate additional points, such as P', as illustrated in Fig. 163. It will be observed that the external current at first increases
is lowered, but that eventually a critical point is reached beyond which a further lowering of the external resistance causes the current to decrease rapidly. The terminal
as the load resistance
voltage falls steadily throughout the entire process, becoming zero when the machine is dead short-circuited (R = 0) under
;
this condition of
complete short-circuit the external current is not zero but has a small value OS due to the fact that residual
magnetism generates a small
e.m.f. that is entirely
consumed
in
It might be driving the current through the armature resistance. inferred from these facts that a shunt generator can be shortcircuited without danger, but this is not the case .except in very
small machines; for the critical point at which the line current begins to decrease is generally far beyond the current-carrying
capacity of the armature, and the winding will burn out before the current has had time to decrease to a safe value.
115.
Dependence
of
the
Form
Speed.
The diagram
of Fig. 163
of the Characteristic upon was drawn subject to the con-
dition that both the speed and the resistance of the shunt circuit remain constant. A change in speed (r s remaining the same)
12
178
will alter the form of^the characteristic, and the new relations between E ia and n can be most easily shown by a three-dimendonal diagram such as Fig. 164. In this figure the surf ace OOi M,
t,
drawn
to the left of the speed axis,
is
the locus of the magnetiza-
tion curves for various values of speed, and to each magnetization curve there will correspond a characteristic LI, L'i, etc., the
locus of which has the peculiar tubular form
shown
in
the
diagram. If the shunt field resistance has a constant value, the locus of the field resistance lines (ON) will be plane OOiJV, and the intersection of this plane with the magnetization surface OO^M
FIG. 164.
Effect of variation of speed upon external characteristic of shunt generator.
will
t,
be a curve OL'L.
will give
(E n) plane between terminal voltage and speed when the generator is operatIf there were no residual magnetism, curve ing on open circuit. OL'L would not pass through the origin, but would intersect the speed axis in a point Z\ that is, if there were no residual magnetism, the machine would fail to build up for any speed below a critical speed, OZ.
of
The projection of this curve on the curve OL'iLi, which shows the relation
Dependence of Form of Characteristic upon Resistance Shunt Field Circuit. If the speed of a shunt generator is kept constant and the resistance of the field circuit is varied bv moans
116.

of the regulating rheostat, the size istic will be affected in the manner
179
and shape of the charactershown in Fig. 165. OM is curve the magnetization corresponding to the speed at which the and machine is driven, ONi, ON 2 etc., are the field resistance lines
,
corresponding to the setting of the rheostat. The construction of the several characteristics is carried out in the manner described in connection with Fig. 163.
V o
FIG. 165.
Effect of variation of shunt regulating resistance characteristic of shunt generator.
upon external
Approximate Mathematical Analysis of Shunt Generator It will be evident from the preceding articles that the form of the external characteristic is in all cases dependent upon that of the magnetization curve, hence an equation representing the relations between the variables E i and n must
117.
Characteristics.
t,
be a function of the equation representing the magnetization curve. Since the latter would necessarily involve a relation between B and for the iron comprising part of the magnetic cir-
and since such a relation is entirely unknown, the best that can be done is to represent the magnetization curve by an empirical equation originally due to Froelich, as follows:
cuit,
E =
where a and
b are constants, and n is the speed. If the speed is held constant, this equation represents a hyperbola, with asymptotes as shown in Fig. 166. A suitable choice of the constants
(13)
a and b will make this hyperbola agree very well with the actual magnetization curve within the working range of the machine, but it cannot be made to follow the irregularities in the actual curve at low magnetizations, and it does not take account of
residual magnetism.
180
Using Froelich's equation, and ignoring armature reaction, we have the following relations (see equations 8, 9, 11 and 13):
am E=
E
whence
t
E =
t
brs
anE +E
t
(14)
t
FIG. 166.
Empirical form of magnetization curve.

is
Solving for E and simplifying by assuming that r a r a = r s ) tliere is obtained compared with r s (i.e., r a
t,
small
E =
t
an an
brs
ira
"
an
br8
ira
)a
ir ar,l
br s
ira
This
is
t,
(Vira
- Vbr
2
]
s)
[an
- (Vira + Vbr
2 s)
]
(15)
an equation of the second degree between the three variables E i and n, hence it represents a surface (Fig. 167) plane sections of which are conies or straight lines. Moreover, the
surface
is
symmetrical with respect to the plane
E
which
is
an
t
br s
ira
(16)
shown
in the figure as
ABC.

If in
181
tuted
the
the general equation of the surface, (15), there = 0, it is seen that
is
substi-
first of
these
br s and E = 0; (17) two equations represents open-circuit conditions,
E =
t
an
From the former it appears the second, short-circuit conditions. br = OB, hence OB is the critical speed when n = that E =
t
below which the machine would netism were not present.
fail
to build
up
if
residual
mag-
FIG. 167.
Idealized external characteristic surface of shunt generator.
the condition n t Inserting in the general equation for constant, there will result the equation of the external characterIt is obvious istic corresponding to the chosen value of speed.
that there are two values of the current
(i)
which
will
reduce the
radical to zero, hence the characteristic intersects the plane

in
ABC
two points; one

is
of these values of current
is
the other
(Vo^ +
Between these values
of current
the radical becomes imaginary, hence the theoretical external characteristics are hyperbolas.
118. Characteristic
Curve
of the
Compound Generator.
Shunt Connection.
The drop
in terminal voltage
Long between no-load
182

full-load inherent in a
shunt generator can be compensated, even overcompensated, by the addition of a series field winding excited by the armature current. As has been previously pointed out, the object of over-compounding
and
partially or wholly, or
is
to keep the line voltage constant at a distant point, at or near the center of distribution of the load, the increase in the voltage at the generator terminals being consumed by the resistance of
the
line.
(Fig. 168)
may
In a general way the compound-wound generator be considered as combining the rising character-
FIG. 168.
istic of
Connections of long-shunt compound wound generator.
the series generator with the drooping characteristic of the shunt generator, the slope of the resulting curve depending upon the relative slopes of the components.
Starting as before with the magnetization curve, O'M, Fig. 169, the external characteristic can be constructed in a simple
manner
Let
as follows:
ON
be the shunt-field resistance
line, its
equation being
(18)
=
J(n.i.);
mines a point
the intersection of this line with the magnetization curve deterL whose ordinate is (very nearly) the terminal
voltage at no-load.
Assuming that we are dealing with an overcompounded machine, let F&i = AP be the terminal voltage
corresponding to a value of ia = OA (the latter being supposed to be known). The field excitation due to the shunt turns will
then be given by OF 1} and the total field excitation will be OF 2 where FiF 2 = n fi a is the excitation supplied by the series turns. The net excitation, or OF, will be less than this by an amount
,
_-
= demagnetizing ampere-turns
per pair of poles,

hence the e.m.f. actually generated in the armature
difference
183
drop in
ing,
is FG. The between FG and FiGi, or GH, must, therefore, be the the armature and series field, or i a (r a + r/). Summariz-
FiF 2 = GiG 2 = n fia
GH =
M
N
(ra
FIG. 169.
Construction of external characteristic of
compound
generator.
It follows, therefore, that all three sides of the triangle

,
GGiG 2
are
proportional to i a hence their ratios remain fixed no matter what the value of ^ a may be, and the angles at the vertices of the triis
In particular, the slope of the side GGi constant, and its length is proportional to i a Through the point G draw GJ parallel to ON, J being on the axis of ordinates (prolonged downward); and join J with A. If, then, it is desired to find the terminal voltage, A'P', correspondangle are also constant.
.
ing to
any other value of armature current, as OA', draw A'J' AJ, draw J'G' parallel to JG until it intersects curve in G', and through G' draw <7'(/' parallel to GGi, the point G'i being on the line ON. Draw G\P' horizontally until it interparallel to
T,
sects ordinate
A'P* in the point P then the latter on the curve showing the relation between E and ia
f
,
is
a point
184

desired to
If it is
show the
relation
i,
sary to subtract the value of

ia .
To
this
end draw the
line
between E and i, it is necesfrom each corresponding value of OX at an angle with the horit
\f/
= rs to the scale of the drawing. Then zontal, so that tan \f/ and P'X' are the values of i corresponding to ia equal to OA
PX
and
OA'j respectively. It will be observed that this method presupposes a knowledge The of the coordinates of at least one point on the characteristic. chief value of the construction lies in the clearness with which it
shows the intimate relation between the magnetization curve and the external characteristic. Thus, it becomes evident from the diagram that the external characteristic will approach a linear form more and more nearly as the magnetization curve
flattens out (that
is,
when the
figure
LGiG approaches
it is
triangular
shape).
If
the point
L is so
placed that
curve, the external characteristic will convex (from above), the curvature being considerable for small
below the knee of the become more and more
values of the load and less pronounced as the load increases. Considerations of this kind become important when the specifications of a machine call for a compounding that shall not depart from a linear relationship by more than a limited amount. A study of Fig. 169 shows that a reduction in the number of series turns will shorten GiG 2 and will make the characteristic more
nearly horizontal.
If it is desired to
make
is
at full-load equal to that at no-load, that
the terminal voltage to say, to make the

full
machine flat-compounded, GGiG z (assumed to correspond to load conditions) will degenerate to Lgg 2 where g%gh = G 2 GH.
,
It
is
clear frorn the construction that the characteristic of a flar-
generator, like that of an over-compounded genetcannot be exactly a straight line because of the curvature of the magnetization curve. Short-shunt Connection. In this case the current through the series winding is i = ia is hence the construction of Fig. 169 is not strictly applicable. But at or near full-load the difference between i and i a will be relatively small, especially in large
compounded
ator,
machines, so that the above method will give a very close approximation to correct results. 119. The Series Shunt. In practice it is quite common to design the series field windings of compound generators with a

sufficient
185
produce the maximum per cent, of may reasonably be specified. If a lesser degree is of compounding required, the magnetizing effect of the series reduced is then by connecting a shunt across the terwinding of the series minals winding, as indicated in Fig. 168. This shunt of German-silver is made strip, and serves to by-pass a portion
number
of turns to
compounding that
of
the main current.
The
its
total current will
divide between
the series
winding and
shunt in the inverse ratio of their
resistances.
When 120. Connection of Generators for Combined Output. the load on a circuit exceeds the capacity of a single generator, one or more additional units must be connected to supply the excess.
in proportion to the load, additional generators
Thus, in a constant-current system in which the voltage varies must be connected in series when the voltage limits of the machine or machines already in service have been reached. Similarly, in constant potential systems, additional generators must be put in parallel with those already in service when the safe currentcarrying capacity of the latter has been reached. The series system, in which series121. The Thury System. 1 wound generators, regulated to give constant current, are connected in series, has thus far found no application in the United States, save in those now obsolete plants in which constant-current motors were supplied from high voltage arc circuits. But in been this has to a state of Europe system developed high perfection through the work of M. Thury, who has installed a number of
plants operating on this principle, 2 Hungary and Russia.
most
of
them
in Switzerland,
In the Thury system the series-wound generators are driven at constant speed and the current is kept constant by a regulating device which shifts the brushes (though it may be arranged to vary the speed). The regulating device is actuated by a solenoid through which the main line current flows. A sufficient number of generators are connected in series to develop the voltage required by the load.
also connected in series,
1 2
The load
consists of series motors,
which
in turn drive generators (generally
See also Chap. VII.

Jour. Institution of Electrical Engineers, Vol. XXXVIII, p. 471. Electrical World, Vol. LXIII, No. 11, p. 583 (1914).
186
alternators) for the supply of current at the receiving or disIn other words, the system is generally tributing end of the line.
used for the transmission of power over considerable distances, as
The individual distinguished from merely local distribution. generators are grouped in pairs, each pair being driven by a water wheel (or other prime mover). In plants now operating, the
maximum
voltage per commutator
is
about 3600
volts,
though
it is
possible to design machines of this type to give 5000 volts at 500 amperes, or 5000 kw. per pair of generators. The maximum
line
voltage
in
use
at
the
Moutiers-Lyon plant is 57,600 volts, though a new installation projected for transmission from
Trollhatten (Sweden) to Copenhagen a distance of 200 miles
contemplates
the
use of a
of
line voltage of 90,000 volts. 1
The
the
FIG. 170.
starting
is
and stopping
in
generators
is
the
system
Diagram of .connections, Thury system.
generator
very simple. equipped with an
Thury Each
ammeter, a voltmeter, and a

switch, as in Fig. 170, the switch being so arranged that when it is in the "off" position the generator is short-circuited, and when in the "on", or running position, the machine is in series
with the
line.
To
start the machine, the switch being in the
mover is brought up to normal speed, and the switch thrown to the running position when the ammeter reads normal current. To shut down the machine this process is reversed.
off position,
the prime
when under load, the from generator to generator, hence the machines must be carefully insulated from earth to prevent breakdown of the insulation.
Since
all of
the machines are in series
potential of the circuit rises
122. Parallel Operation of Generators. Series generators connected in parallel (a) Series Generators. as in Fig. 171 will not operate satisfactorily, for the reason that
1
Electrical World, Vol.
LXI,
p. 294, 1913.
187
if one of them suffers a momentary reduction of its output (as from a momentary drop in speed), both its voltage and current will be reduced, as may be seen from the form of the characterThe other machine will then assume the part of istic, Fig. 157. the load dropped by its mate and its current and voltage will
accordingly
rise;
the increased voltage will cause a further in-
crease of current, hence an additional increment of load is thrown on the second machine and the load, current and voltage of the
be still further reduced. This process will tend to continue until the first machine is driven as a motor by the second machine; moreover, the direction of rotation of the former will
first will
^T
/
Equalizer
g<
s eric
,
reverse
when
it
becomes a motor,
FIG. 171.
Series generators in
so that the connecting rod of its driving engine will tend to
buckle.
Series generators conare, therefore, in unstable equilibrium, there
nected in parallel being no inherent the load between a consequence of
tendency to bring about a proper division of This is the two units under consideration. the rising form of the external characteristic
of the series generator.
The natural instability of series generators in parallel can be overcome by the equalizing connection shown in Fig. 171 as a dashed line. The effect of this connection is to put 'the series field windings in parallel with each other. If then one machine assumes more than its proper proportion of the total load, the excess current will divide between the two field windings, thereby raising the excitation and voltage of the machine which has
momentarily dropped
(6)
its load,
hence automatically readjusting

of
if
the division of the load.
Shunt Generators.
its load, its
The drooping form

shows that
the external
characteristic of the shunt generator
such a machine
drops
if
voltage will automatically
rise.
Consequently
two shunt generators are connected
in parallel, as in Fig. 172,
their operation will be stable. Any tendency which causes one machine to lose its proper share of current, thereby shifting an equal amount of current to the other, will result in a rise of vol-
188

first
tage of the
machine and a drop
in the voltage of the second.
The
original conditions will be restored,
assuming that the prime
movers are properly governed.

It is, of course, not necessary that the two (or more) generators thus connected in parallel should have the same ratings. But it is essential to good operation that the machines should
divide the total load, whatever that
may happen to be, in proto their ratings. portion Sup+ pose, for instance, that two
shunt generators that are to be connected in parallel have external characteristics as
shown
repre-
FIG. 172.
Shunt generators
parallel.
.
in
in
Fig.
173,
curve
(a)
senting the characteristic of one machine, curve (6) that of the other. In this figure ordinates are plotted in volts and abscissas in per cent, of full-load current. Since the machines are in parallel, their terminal
voltages must necessarily be equal, hence if the load that the terminal voltage is OC, machine (a) will deliver
cent, of its rated current
is
such
and machine
that
(b)
OB
all
condition to be satisfied
is
OA = OB
at
per The per cent. loads, therefore it
OA
r"
189
This is a with the same equalizer connection shown in Fig. 171. consequence of the rising characteristic. But if the machines have drooping characteristics, that is, if they are under-compounded, the equalizer is not necessary.
The diagrammatic scheme of connections of two compound generators in parallel is shown

in Fig.
if
174.
It
is
clear that
ammeters were connected as at A'j they would not indicate

the true current actually deby the machines to
the external circuit,
for
livered
the
FIG. 174.
Compound
parallel.
generators in
readings would be affected by the equalizing current. Thus, a heavily loaded generator might be supplying an equalizing current of large magnitude to the
other lightly loaded machines and at the same time the ammeter of the loaded machine would read low while that of the other machines would For this read high.
reason
the
individual
in the
ammeters must be placed

as at A, that
is,
lead
that
connects
to
the
side
armature
on
the
to the opposite connection. equalizing For the same reason, if single-pole circuit-breakers are used, they should be placed in the same
lead as the
FIG. 175.
tions,
ammeters;
Diagram of switch-board conneccompound generators in parallel.
thus, if two machines in parallel are each delivering

full-load
cur-
rent and one of
them should
suffer a
momentary drop
in speed,
the heavy equalizing current might open its circuit-breaker, if incorrectly placed, with the result that the entire load would be
thrown on the other machine and so open
its
circuit-breaker also.
190

of
The complete switch-board connections
two compound
generators are shown in Fig. 175. The main switch and the equalizer switch are usually combined in a triple-pole switch.
The process of paralleling a compound generator with one or more that are already running is as follows: The main switch of the incoming machine being open and its circuit-breaker closed, the prime mover is brought up to speed and the voltage of the
incoming machine adjusted to equality with the bus-bar voltage by manipulation of the shunt-field rheostat; the main switch is closed, and proper division of the load is then secured, if necessary,
by further adjustment of the field rheostat. To shut down a machine running in parallel with others, its load is shifted to the others by weakening its shunt-field current, and the main switch is opened when the ammeter indicates a small, or zero, current. If two compound generators are to divide the load in propormust obviously be
tion to their ratings, their characteristics identical in the manner
explained in connection with shunt machines. Moreover, since the series fields
by virtue of the equalizer resistances of the series the connection, windings including the resistances of the
are in parallel
respective equalizing leads, must be inversely proportional to the rated currents

of
the two machines.
Neglect of this
feature will result in a disproportionate For example, if the division of load.

FIG. 176. Diagram of connections of railway
machines are at unequal distances from

the switch-board, the resistance of the series field of the more remote machine
will
generator
grounded
supplying
circuit.
equalizing connection, and

its full
be unduly high because of the longer this machine will, therefore, not take
share of the load.
T^he series field of a compound generator may be connected to either the positive or the negative terminal of the armature. In street-railway generators built by one well-known company
the series field
in this case it is
connected on the negative, or grounded side; not sufficient to use one single-pole circuit-breaker (A) on the positive or feeder terminal, but another circuitis
191
breaker (E) must be put in the lead to the grounded bus, as shown For if circuit-breaker B were not present and the in Fig. 176.
armature winding were to become grounded to the core, the shortcircuit current through the armature and series field would hold up the excitation and maintain the short-circuit without the possibility of protection by circuit-breaker A. 123. Three-wire Generators. Economy in the use of copper
and power dictates the selecand moderate current, rather than low voltage and large current; but in incandescent lighting, lamps designed for 110 to 115 volts are more efficient than those operating
in distributing circuits for lighting
tion of high voltage
<o
192
being connected in series in pairs and connected across the main wires. But if the load is not exactly balanced, the neutral will
carry a current equal to the difference between the currents supThe attempt is always made plied to the two sides of the system. to balance the system as completely as possible, but provision isusually made for an unbalancing of about 10 per cent., that is, 10 per cent, of full-load current in the neutral wire.
When
FIG. 178.
Three-wire system with balancer
set.
system employing a balancer set becomes unbalanced, the voltage on the more lightly loaded side tends to be higher than on the more heavily loaded side; in this case, the machine on the side having the lighter load operates as a motor and drives the other as a generator; the latter then supplies current for the excess load on its side of the system, and thus automatically tends to balance the system. With perfect balance of load both machines of the balancer set operate as motors running without load.
?
Outside Wire
Neutral Wire
Outside Wire
FIG. 179.
Three-wire generator with

of the kind
coil
mounted
inside
armature
core-
Systems
shown
in Figs. 177
and 178 are open
to the
objection that they involve the use of more than one piece of running machinery and so require extra attendance and main-
tenance, in addition to being higher in efficiency than a single machine of the

objections are
first cost and lower in same capacity. These overcome by a system originally devised by Do-
193
browolsky, and shown diagrammatically in Fig. 179. A coil of wire, CED, wound on an iron core, is tapped into the main armature winding of the generator at the points
trical
C and D,
180 elec-
degrees apart, that
is,
points that occupy homologous posi-
Y
1
I
slots.
FIG. 180.
Three-wire generator with auxiliary winding in
of potential
tions with respect to poles of opposite polarity. The difference between C and is alternating, so that the coil is
traversed
alternating current which goes through one cycle (two alternations) per revolution per pair of poles; this alternating current is small because of the large self-inductance due
to the iron core
by an
on which the
coil is
wound.
The middle
point
FIG. 181.
Three-wire generator with two winding.
coils
tapped into armature
of the coil, E, will have a potential tials of C and D, and therefore also
tials of
the brushes A and 5, since always symmetrically related to those of

13
midway between the potenmidway between the potenthe potentials of C and D are
and B, respectively.
194
tral wire of
tap brought out from the point E may then be used as the neua three-wire system. In machines of this kind built by the General Electric Company, the coil CD is wound on a core
that
is
mounted
from the middle point
inside the armature core, and the connection to the outside circuit is made through
FIG. 182.
Three-wire generator with two auxiliary coils mounted externally.
a single slip-ring
mounted on the main
The Burke
Electric
which the coil CD is armature winding, in the manner indicated in Fig. 180. The balance coil CD may also be placed outside of the genera-
shaft of the generator. builds a three-wire generator in Company wound in the same slots that carry the main
Equalizcrs
\ / ^ ^^ Series Field
Shunt Windings not shown
FIG. 183.
Diagram
of connections of
compound
three-wire generators in
parallel.
tor,
connection to the armature winding being made in that case through two slip-rings; or two balance coils, connected to the armature winding as in Fig. 181, may be used. The alternating
voltages between the points C,
and
C",
D' are 90
electrical

degrees apart, that
is
is,
195
zero,
and
vice versa.
one of them is a maximum when the other Fig. 182 shows the connections when two
balance
this
is
coils, mounted externally to the generator, are used; the standard construction used by the Westinghouse
and Manufacturing Company. be compounded, the series field in must be two equal parts, half of the turns being in winding series with one of the outer wires, the other half in series with the other outer wire, as in Fig. 183. If two or more three-wire generators are to be operated in parallel, two equalizer connections must be used, hence the main switch of a three-wire
Electric
If three-wire generators are to
generator
is
usually constructed with four blades.
It has been shown in preceding ar124. Tirrill Regulator. ticles how the voltage of shunt and compound generators may be
regulated either
shunt
is
field circuit or
by manual adjustment of the rheostat in the by the automatic compounding effect of the
In lighting circuits where steady voltage and automatic regulation
series field winding.
of the greatest importance, accurate
of voltage this device
be obtained by the use of the Tirrill regulator; it possible to maintain a steady voltage at the terminals generator irrespective of changes in the load or of fluctuations of speed, and also to compensate for line drop by in-
may
makes
creasing the generator voltage as the load increases. The regulator maintains the desired voltage by rapidly opening and closing a shunt circuit connected across the terminals of
The field rheostat is so adjusted that disconnected the generator voltage is about 35 per cent, below normal; on closing the regulator circuit the rheostat is short-circuited and the generator voltage rises. When
the exciter
field rheostat.
is
when the
regulator
the voltage reaches a predetermined value, the short-circuit around the rheostat is opened and the voltage again falls. The opening and closing of the short-circuit around the rheostat is
so rapid that the voltage does not actually follow the changes of the field circuit resistance, but merely tends to do so, with the
result that incipient changes of voltage are immediately checked. An elementary diagram of connections of the regulator is
shown
in Fig. 184. The opening and closing of the by-pass around the exciter rheostat is accomplished by means of contacts on the armature of a differentially wound relay magnet of U
196
PRINCIPLES^QF DIRECT-CURRENT MACHINES
One winding of the relay magnet is connected directly shape. across the main bus-bars, in series with a current-limiting resistor;
the other winding is also connected across the bus-bars, but through a pair of main contacts actuated by the main control
magnet.
The
latter is
wound with a
directly across the bus-bars,
and a current
potential coil connected coil (which may or
may
not be used) whose magnetizing action opposes that of the

coil.
potential
If the generator of the regulator is as follows: the coil of the main the current potential voltage falls, through control magnet is weakened and the spring closes the main con-
The operation
tacts.
Current then flows through both windings of the relay

Compensating Shunt
FIG. 184.
Diagram
of connections at Tirrill regulator.
magnet which is then demagnetized and the spring closes the reAs the lay contacts, thereby short-circuiting the field rheostat. voltage rises the armature of the main control magnet is again
is
main contacts are opened, and the relay magnet again energized, thus again inserting the rheostat in the field circuit. If the current coil of the main control magnet is used,
pulled down, the
action will cause the voltage to rise higher before the main contacts are opened than would otherwise be the case, thus giving a compounding action. The degree of compounding
its differential
may
be varied by means of the sliding contact on the compen-
The sating shunt with which the current coil is in parallel. condenser shown in the figure is for the purpose of reducing sparking at the relay contacts. A perspective view of a simple
regulator built
Fig. 185.
by the General
Electric
Company
is
shown
in
197
When several compound generators of moderate capacity are worked in parallel, a simple regulator may be connected to one of them and the others allowed to "trail." The generator with the will take the in load, and fluctuations regulator provided the load on the others will be equalized through the compound
windings. Regulators are also built for controlling the voltages two or more generators operating in parallel instead of using a single relay magnet, from two to ten are employed, part of them
of
;
FIG.
185.
Voltage regulator, General Electric Co.
serving to short-circuit sections of the field rheostat of one generator, the others performing a like function for the other machines.
Regulators of this kind are suitable for two-wire or
three-wire generators with shunt or compound windings, and will compensate for line drops up to 15 per cent.
rate excitation
In the case of very large machines, it is advisable to use sepaand to connect the regulators so that they act upon the exciter fields. Of course, in such a case the main control
line
magnet would be actuated by the main bus-bar voltage and

current.
198
PRINCIPLES dF DIRECT-CURRENT MACHINES
PROBLEMS
10-pole, 249-velt generator rated at 400 kw. at 200 r.p.m. has a magnetization curve represented by the expression
1.
~~
5^9 X amp.-turns per pole 10750 + amp.-turns per pole
The armature has a simplex lap winding of -800 conductors, each element having one turn, and the angle of brush lead is 15 deg. The armature (hot) The shunt field winding has 550 turns per pole resistance is 0.0033 ohm. and a (hot) resistance of 10.5 ohms. The coefficient of dispersion is 1.16. If the field winding is separately excited from 220-volt mains, how much resistance must be put in series with the field winding to develop an opencircuit voltage of 250 volts at 225 r.p.m.? 2. The machine of Problem 1 is operated as a separately excited generator at 200 r.p.m. and with a field excitation sufficient to develop an open-circuit What will be the terminal voltage when it is delivering voltage of 240 volts. 1000 amp.? Solve analytically and graphically. 3. The separately excited generator of Problem 1 is run at a speed of 200 r.p.m. Find the field current required to produce a terminal voltage of 240 volts when the armature current is (a) zero, (6) 800 amp., (c) 1600 amp. Plot a curve showing the relation between field current and armature current, the terminal voltage having a constant value of 240 volts. 4. A series generator has a resistance of 0.2 ohm, and the armature demagnetizing turns per pole at full-load amount to 6 per cent, of the field turns The open-circuit characteristic may be expressed by Froelich's per pole.
equation (13), such that at a speed of 1200 r.p.m. an exciting current of 10 amp. develops 120 volts, and an exciting current of 5 amp. develops 80 volts. Find the terminal voltage when the machine is operating as a series generator at 1100 r.p.m., the load current being 8 amp.
6. The machine of Problem 1 is operated as shunt generator at a speed of 200 r.p.m. and with a field resistance such that it gives an open-circuit Find its terminal voltage when delivering an armature voltage of 240 volts. current of 200 amp. How much current is supplied to the load? 6. The machine of Problem 1 has a series winding of 2.5 turns per pole with a total resistance of 0.001 ohm; a shunt around the series field winding reduces the series field current at full load to 1000 amp. If the generator, connected long-shunt, gives a terminal voltage of 240 volts at full load, what will be the open-circuit voltage? (NOTE. Neglect the difference between armature current and line cur-
rent at full load.)

7. Two shunt generators, A and B, have the same open-circuit voltages and the same full-load current ratings. Their external characteristics may be considered to be straight lines between no-load and full-load conditions,
but their percentage voltage regulations
(in terms of full-load voltage) are 4 and 6 per cent., respectively. If the current taken by the load is 125 per cent, of the full-load current of either machine, what currents are supplied
by A and B,
in per cent, of their respective full-load currents?
CHAPTER
VII
MOTORS
In the industrial application of 125. Service Requirements. the motor drive, there are three principal classes of service, characterized by constant speed, adjustable speed, and variable speed.
Constant-speed motors, of which the shunt motor is an example, maintain an approximately constant* speed at all loads when supplied from constant potential mains, and are used for such purposes as driving line shafting, fans, etc. In the case of adjustable speed motors, the speed can be fixed at any one of a
large
between a minimum and maximum value, remain substantially constant for all loads within the limits of the machine's capacity, the impressed voltage remaining constant throughout; motors of this kind are used, for example, in individual drives for machine tools. Variable-speed motors include those types in which the speed is inherently variable, changing as the load changes, with constant impressed
number
of values
and when so
set will
voltage; examples of this class are the series motor

lative
and the cumucharacteristics
compound-wound motor; make them especially suitable for that

railway and in hoisting service.
their
speed
class of service in
which
it
is
desirable to reduce the speed as the load increases, as in street
Intelligent operation of motors involves a knowledge of the relations between speed, torque (or turning moment), load (or
output), and the electrical and magnetic quantities involved. These relations determine the operating or mechanical characteristics,
which
will
be discussed for the different types of motors.
It has been shown 126. Counter E.M.F., Torque and Power. in Chap. II that when a current is sent into the winding of an
armature which
is
under the influence of a magnetic
field
the
individual conductors of the winding are subjected to a lateral thrust and that motion ensues. The immediate effect of this
motion
is
to generate in the conductors an e.m.f 199
whose direction
is
200
This counter-generated e.m.f. opposite to that of the current. is called the "back e.m.f." or the "counter e.m.f." The effective development of torque in the case of a motor is
dependent upon a proper space relation between the field flux and the armature current. If, for instance, the brushes are so set that
the axis of the armature current coincides with the axis of
field flux, as in Fig. 186a, there is no resultant tendency to rotation; but if the axes of armature current and field flux are at right angles to each other, as in Fig. 1866, the torque will be a maximum
for a given current in the winding.
FIG. 186.
Effect of brush position on the torque.
e.m.f.
in
In the case of a separately excited or of a shunt motor, the impressed upon the armature terminals must be consumed
overcoming the back
e.m.f.
resistance of the armature winding

.'.
and the ohmic drop due and the brush contacts.

arc
to the
E =E
t
(1)
where
p
a 60
10*
tional
In the series and long-shunt compound motor, there is an addidrop due to the resistance of the series field winding, hence
E = Ea +
t
ia (ra
+ +
r/)
(2)
In the case of the short-shunt

t
compound motor the

ia r a
relation
is
E = Ea +
irf
(3)
MOTORS
In general
201
E =
t
E,
ia r'
=
I
^^ +
and
iS
(4)
or
E - i a r' ~*Z^t
(5)
where
r'
resistance of armature
circuits in series
therewith
and
a 60
It is also seen that
10 8
E - Ea
t
r'
(7)
an equation which
is
of
importance in connection with the
starting of motors, as explained later.
Multiplying equation
(4)
t
by
ia ,
there results
za
E ia = E a ia +
The term
i a r'
2
t
(8)
E ia represents the power supplied to the armature, and the power dissipated as heat in the ohmic resistance of the armature circuit. It follows, therefore, that E a i a is the
is
of mechanical power developed by the armature. Not all developed power is useful power at the shaft or pulley, for some of it is lost in bearing and brush friction, windage, and iron
of this
losses.
If
amount
P =
T =
total
mechanical power, armature
in watts,
developed in the
torque in dyne-centimeters
-r-
10 7
(9)
/.
p = E a ia = 2ir~T
T =
watts
or
E a ia =
Z'$i a
(10)
The above
unit of torque is inconvenient for practical application expressing torque in kilogram-meters, pound-feet, and pound;
inches, respectively,
202

AO
in7
Q.$75Z'&i a kg-m.
60
2;r
1Q
980
453.6
30.48
la
7.05Z'<H a
lb-ft.
(11)
It is clear from these equations that the torque is dependent only upon the flux and the armature current, and is independent
of speed. 127. The Starting of
Motors.
If
a mo tor
is
called
upon
to start
a heavy load from rest, the starting torque may be as large as, or even larger than, the full-load running torque. If the flux at
starting has its normal full-load value, the starting current, by equation (11), will then have to be equal to, or perhaps somewhat
Other things equal, the starting greater than, its full-load value. current may be smaller the greater the flux. But since a =
when
at the
the armature
is
stationary,
ia
it is
clear
from equation
(7)
that
moment of
starting
-7, and, therefore, that the
Et
normal
rf) small running resistance of the armature circuit (r a or r a must be increased during the starting period by the insertion of
a starting rheostat in order to limit the flow of current to a reasonable value. Thus, a 10-h.p. 220- volt shunt motor would take an armature current of approximately 40 amperes when carrying
its
0.5
rated load, and would have an armature resistance of about ohm. If the full voltage were impressed directly upon the
armature, the initial current would be 440 amperes, or more than ten times normal full-load current. To limit the starting current to the -full-load value, the resistance that must be put in series
with the armature should be
220
0.5
^7j-
ohms
is
The
resistance of the starting rheostat

is
usually so adjusted that
the initial current

torque.
Fig. 187a
somewhat
greater than that giving full-load

of the start-
shows diagrammatically the connections
ing rheostat in the .case of a series motor, and Fig. 1876 those of a shunt motor. It should be carefully noted that in Fig. 1876
the rheostat
is
in series with the
armature only, so the shunt
field
MOTORS
winding receives the
full line
203
voltage at all times, including the Fig. 187c shows an incorrect set of connections, starting period. since here the shunt field current is seriously reduced at the
(a) Series
(6) Shunt Correct
(<?)
Shunt Incorrect
FIG. 187.
Connections of starting rheostats.
start,
thereby reducing the flux and also the torque, and, if the motor is unloaded, causing the speed to rise dangerously high. If an ordinary rheostat of the kind illustrated in Fig. 187 were used in commercial installations, there would be danger of burn-
FIG. 188. Diagram of connections of starting rheostat having no-voltage and over load release.
FIG. 189.
Motor
starting rheostat
with no-voltage and overload release.
ing out the armature if, after an interruption of the service and the consequent stopping of the motor, the voltage should again
be applied to the supply
line; for in
that case the
full line
voltage
204

of the ariria
would be thrown directly across the low resistance
ture (or armature and series field), resulting in a very heav; current. For this reason most motor starting rheostats are pro
vided with a "no-voltage release" which automatically restore the starting lever of the rheostat to the starting position when th line voltage is removed; quite frequently there is also an "over load release," which opens the circuit and automatically cuts ii
the starting resistance if the current becomes excessive for ani reason. The connections of such a rheostat are shown in Fig. 188, and Fig. 189 illustrates a starting rheostat of this type madi
by the Ward Leonard Company.

It is seen
small
from equation (5) that if the flux is reduced to value while the e.m.f. impressed upon the armatur*
l
remains constant, the speed will rise to a dangerously high value In other words, the motor will "run away" and may wreck itself This contingency may arise in the case of a shunt motor if tht field circuit is opened, as by a broken wire or loose connection and in the series motor by an accidental short circuiting of tht terminals of the series winding. This behavior is due to the tendency of any motor to run at such a speed that the back e.m.f shall be nearly equal to the impressed e.m.f., the lowering of th flux demanding an increased speed. 128. Characteristics of the Separately Excited Motor. Let it be assumed that both the (a) Speed Characteristics.
impressed e.m.f.,
stant.
It follows
E and the field exciting current are from the speed equation
t,
corl-
n
that were
it
E t
ia ra
~7
not for the demagnetizing action of the armature cur-
rent, the denominator of the fraction would be constant and the speed would decrease slightly and uniformly with increasing values of ia as in Fig. 190. This assumes that r a is constant
,
in other words, that the temperature of the armature is maintained at its normal running value. The separately excited
motor with constant excitation is, therefore, inherently selfBoth equation (5) and Fig. 190 regulating as regards speed.
indicate that
if ia
0,
n =
-^7.
Actually,
if ia
0,
there
is
no
torque and no rotation.
When
the motor
is
"running free"
MOTORS
that
is,
205
ture since sufficient
is still some current through its armamust be supplied by the line to overcome power nternal losses due to windage, friction, hysteresis and eddy curThe minimum value of armature current is indicated by ents.
unloaded), there
;he
point
in the figure.
The speed n =
pi
^ may be called the

f
;here
deal zero-load speed; it is the speed that would be reached were no losses, in which case also, a = and ia = 0.
if
It is
obvious that the speed
may
t )y varying either 3> or ?an be raised by reducing
or both of them.
be varied through wide limits Thus, the speed
or
t )y increasing However, she possible range of speed due to the adjustment of the excita.
S-r.
ion
-here
is
rather restricted, unless
special devices are used,
because
field
are
limits
to
the
r
FIG.
190.
Strength above or below which here are serious commutation

difficulties.
Variation
of
gives little or
is
no trouble so far
is
^0?^
Approximate
f
speed
separately excited
concerned, provided the flux is originally adnormal value unless, indeed, E t is raised to .bo great an extent. The fact that the field excitation and the armature impressed e.m.f. are independently variable in the separately excited motor gives to this type its chief advantage.
commutation
usted to about
its
The form of the speed (6) Effect of Armature Reaction. equation shows at a glance that the effect of armature reaction, since it reduces <, will be to raise the speed, thereby partially
neutralizing the slowing-down effect of armature resistance and mproving the speed regulation. The armature of a motor may
therefore
;Ure of
be designed magnetically more powerful than the armaan otherwise identical machine intended for use as a
relation
generator.
The curve showing the
between speed and armature
current can be constructed in the following manner: Let O'G, Fig. 191, be the magnetization curve of the machine, abscissas (drawn downward from O) representing ampere-turns
per pair of poles
n/if )
and ordinates (drawn
to the left of 0)
206
representing values of <J>Z'. Select any convenient scale of armature current along OA, and a scale to represent the impressed voltage along 0V. Assume that the field excitation is constant
and equal to OF and that the voltage impressed upon the armature is likewise constant and equal to OE = E v
,
t.
FIG. 191.
Construction of speed characteristic of separately excited motor.
ra
a straight line so that tan <p through the point to the scale of the figure; then for any value of i a} such as between and will OA, the intercept on the ordinate at
,
Draw
EC
EC
EB
be
is
BC =
then
ia ra
= ohmic
t
drop in the armature.
The back
e.m.f.
Ea = E - iara = AB - BC = AC Similarly, draw OD making an angle with OA, such that tan
dju V-IQK ; loUft
then when
ia
= OA,
AD =
CiZ
PtoA~& loUtt
demagnetizing
ampere-turns per pair of poles. If the armature were currentless (i a = 0) as in the ideal noload condition, the value of $Z' would be F G Q = OL and the ideal no-load speed is then
,
n =
OE
OL
MOTORS
207
At any other load, as when i a = OA, the demagnetizing effect is given by AD; joining D with F 0) and drawing AF parallel to DF the net excitation is reduced to OF and 3>Z' becomes FG = OL] at the same time the back e.m.f. is AC = OM, hence the speed is
E t
laTg
OM
OL
~~
S>Z'
In this
of
ia
way
and the
values of speed can be computed for various values results plotted to obtain the desired curve, Q P.
But the diagram lends

tion, as follows:
itself
readily to a complete graphical solu-
Select
parallel to
any convenient point L Q E; then
on the $Z'
Q
axis,
and draw QNo
n ~
Similarly, join
OL
OE ~ _ ON
OQ
L and
M and draw QN parallel to LM.

=
OM
OL
ON _ := OQ
ON
and are respectively it follows that be and made to n and equal to the speed by may proportional n, across to P, the latter a suitable choice of scale. Projecting
Since
OQ is
constant,
ON
being on the ordinate through A, P will be a point on the required curve. In precisely the same manner, P' is a point corresponding to i a = OA'. It is clear that the speed curve cannot be exactly
straight because of the curvature of the magnetization curve. From equation (11), the torque is (c) The Torque Curve.
T =
7.05 &Z'ia pound-feet
Referring to Fig. 191, this becomes
T = 7.05FG.OA =
when
ia
7.05 OL.
OA
= OA.
This
may
be written
OA
constant
which equation suggests the following construction for the curve showing the relation between the armature current and the torque: In Fig. 192 draw the axes of coordinates, the $Z' curve, and the line OD just as in Fig. 191. Proceed as before to locate Select any convenient constant length OR, points G and L.
208
draw LR, and then draw OP perpendicular to LR (using the semicircle on OR as a construction line) until it intersects the ordinate
through
in the point P.
By construction AP = OL ~ PL
OA
OR
constant
AP is proportional to the torque, or may be made equal to the torque by a suitable choice of the length OR, and P is then a The torque curve is slightly point on the desired torque curve. concave downward, but to all intents and purposes it is a straight
hence
line
of
ia
through the origin, since $ within the working range.
is
almost constant for
all
values
FIG. 192.
Construction of torque characteristic of separately excited motor.
Shunt Motor. SPEED AND TORQUE CHARACTERISTICS. A plain shunt motor operated on constant potential mains, and having fixed field, resistance, differs in no way from a separately excited machine with constant impressed e.m.f. and constant excitation. The
determination of
ceding
article.
its
carried out in exactly the
operating characteristics is, therefore, to be same manner as described in the pre-
MOTORS
The
209
chief point of difference between the shunt and the sepamotor is that in the former the field excitation excited rately and the impressed e.m.f. are not independently variable, as in
the latter.
less in
The
the shunt motor than
possible range of speed variation is therefore it is in the separately excited motor.
130. Characteristics of the Series Motor.

(a)
for the
Speed Characteristic. speed
Inspection of the general equation
series motor must decrease quite for increases with with load the reason that rapidly increasing of the fraci In other the numerator while increasing a words,
shows that the speed of the

.
<
the denominator increases. Theoretically, if hence at no load the speed would be infinite; 0, 0, the flux does not become zero because of residual while practically, still it becomes so small that the speed reaches a magnetism, dangerously high value, assuming that E t remains constant. For this reason, a series motor must always be so installed as to be positively connected to its load, by gearing or direct connection, never by belting, and the minimum load must be great enough to keep the speed within safe limits; such is the case, for
tion
decreases,
3>
ia
instance, in railway motors, hoists, rolling mills, etc. Assuming that the motor is to be operated on constant potential mains, its speed characteristic can be determined by a modi-
methods described in the case of the separately excited motor, as follows: Let O'G, Fig. 193, be the curve which gives the relation between $Z f and the exciting current (4). Let OE represent to scale
fication of the
the constant impressed voltage, E and draw EC so that tan tp = (fa -h T f ) to the scale of the figure then BC will represent to the same scale the internal ohmic drop corresponding to ia = OA,
t,
,
and
AC
will
be the back e.m.f.

is
Also,
draw
OD
so that tan
where n f
the
number
of
field
turns
per
pair
of poles; then will be the demagnetizing effect expressed in equivalent amperes instead of in ampere-turns per pair of poles. When the armature current is ia = OA, the field excitation (in
14
AD
210
amperes) is OF = OA, the point F being found by drawing AF The demagnetizing at an angle of 45 deg. with the horizontal. effect is AD = FH, hence the net excitation is OH and the
corresponding value of <bZ'
E t
HK ig (r +
is a
OL. ~
r/)
AC
$Z
HK
==
OM
constant, and
Selecting a point Q on the $Z' axis such that drawing QN parallel to LM,
_
OL OQ is
OM
OL
~~
~ ON
OQ
FIG. 193.
Construction of speed and torque characteristics of series motor.
hence is proportional to the speed. upon the Projecting ordinate at A, the resulting point P is a point on the speed-current
curve.
ON
Since
FH = AD =
and
tan
\f/
^
tan 6
1
OH = OF - FH = L - L
= .j =
OT-f
tan Q
constant
MOTORS
it
211
any other current, as ia = OA', it is only necessary to draw A'H' parallel to AH in the process of locating P As before, the torque is (b) The Torque Characteristic.
follows that for
r
.
T =
when
ia
7.05
&Zi a =
7.05
OL.OA
pound-feet
= OA.
Selecting a point R such that perpendicular to LR,
OR =
~
constant, and drawing
OP
APT OA
whence
OL
OR PT
is
AP T
is
proportional to the torque, and
a point
on the torque-current curve. It will be observed that the torque curve of a series motor deviates considerably from the linear form due to the fact that the If the magnetization curve were flux varies with the current. a straight line, that is, 3> proportional to i a the torque would be proportional to (i a ) 2 ,and the curve would be a parabola; actu,
ally it is a curve of higher order, lying
between the linear and
parabolic curves. Lines such as OP T can readily be drawn perpendicular to as a diameter and drawing by constructing a semicircle on
LR
OR
a line through
and the point where

discussed
LR
cuts the circle.
The torque curves
in connection with the sepa-
rately excited, shunt and series motors refer to the total developed torque, as given by equation (11). The actual torque
at the pulley that would be measured by a brake test is less than the total torque by an amount which corresponds to the torque
The required to overcome internal friction and iron losses. curve of useful torque may be obtained from that of total torque by subtracting from the ordinates of the latter the "lost torque";
the useful torque passes through zero value ciable value (see Fig. 209).
(a) General.
when
ia
has an appre-
Compound Wound Motor.
the shunt and series windings of a compound wound (long shunt) machine are so connected that their magnetizing effects cooperate, or are cumulative, when the machine is used
If
as a generator, then,
if the machine is used as a motor, the two windings will oppose each other, resulting in a differential effect. This is illustrated diagrammatically in Fig. 194. If the machine
212
is
designed to over-compound as a generator, the differential motor action will be considerable, resulting in a decided decrease of flux under load conditions, and hence a speed higher than would
In general, the case
is
obtain without the series winding.

lar to that of
simi-
a shunt motor with exaggerated armature demag-
netizing effect.
In the same way a differentially wound generator, having a drooping e.m.f. characteristic when driven at constant speed, becomes a cumulative-compound motor with a drooping speed characteristic when supplied with constant terminal voltage.
;(
jlfiM^n
Generator
FIG. 194.
Relative directions of shunt and series exciting current in com-
pound machine.
(b) (I)
Construction of Speed Characteristic.

Differential
The only
difference
Compounding. between this case and the one discussed

is
in connection with Fig. 191,
that
now
tan v
ra
rf
and
aZ
assuming that the resistance of the shunt field winding is constant. The construction has been carried out in Fig. 195, from which
appears that if n/ is sufficiently large, the speed rises with It is clear that there is a particular value of increasing load. for which the speed will be the same at full load as at no load,
it
but that it cannot be made absolutely constant at all loads (in the absence of special regulating devices) because of the curvature
of the
magnetization curve O'G.
The torque curve
being used in its due to the fact that the flux decreases with increasing current.
shown in Fig. 195, the fixed point R The curve is concave downward, construction.
is
also
MOTORS
213
FIG. 195.
Construction of speed and torque curves of differentially compound motor.
O'
\A i
FIG. 196.
Construction of speed and torque curves of cumulative compound motor.
214
(II)

Cumulative Compounding tan
(p
Here,
ra
r/
and
therefore, the line

it;
OD
must be drawn afowe
OA
instead of below
otherwise the construction shown in Fig. 196 is the same as in the previous cases for both the speed and torque curves. The
falls considerably with increasing load, and the torque concave upward. The cumulative compound motor has characteristics which are intermediate between those of the shunt and series motors. It differs from the latter especially
speed curve
now
is
in this, that its speed rises to a definite limit
when
full
load
is
suddenly thrown off, instead of running away. 132. Counter E.M.F. The Reversing Motor. The existence of the counter-generated or back e.m.f can be shown in a striking
.
FIG. 197.
Connections for the experiment of the reversing motor.
manner by the arrangement shown diagrammatically in Fig. 197. A series generator, G, is driven at constant speed by a suitable
motor, Mij and the generator is electrically connected through the switch S and the regulating resistance R to the armature of an unloaded and separately excited motor, The field of 2 from any con2 is excited by a constant current supplied venient source.
S the series generator will build up both and current, provided the resistance of the circuit is below the critical value. At the same time, the speed of 2 will rise and its back e.m.f. will rise nearly proportionally, so that the acclosing the switch
t
On
e.m.f.
tive e.m.f. in the circuit, available for producing current, is the difference between the e.m.fs. of generator and of motor Af 2
MOTORS
The current
it
215
therefore falls after rising to a certain value
and with
the generated e.m.f. of G decreases. Meanwhile, the speed of Mz continues to rise because of its acquired momentum, its back
overpowers the generated e.m.f. of G, the current in the may be shown by the two-way ammeter A), and MZ momentarily becomes a generator tending to drive G as a motor in opposition to motor Mi. But as Mz has no driving power other than its energy of rotation, it very quickly comes to rest. Since the current through the circuit has been reversed, the residual magnetism of G also reverses, .consequently as soon as MZ has come to rest G begins to build up again, but with polarMotor z then speeds ity opposite to that in the first instance. up again in the reverse direction until its back e.m.f. overpowers
e.m.f.
circuit reverses (as
the generator,
it
again stops, and the entire cycle of changes
is
repeated, over and over again. The function of the resistance
is
simply to prevent the cur-
rent from reaching excessive values, and its magnitude will depend upon the machines used in the experiment. The rate
up in speed depends upon the armature and upon the torque, the latter in turn depending upon the magnitude of the exciting current of The larger the excitation of Mz, the greater will be the 2 torque for a given armature current, and the more rapid will be
will build
at which
motor
M%
moment
of inertia of its
the process of picking up speed; moreover, the greater the excitation, the less will be the speed to produce a given back e.m.f. Finally, therefore, it will be seen that the reversals of Mz will be
more and more rapid, the greater the excitation of 2 The above reasoning will serve to explain why a series generator cannot be used to charge a storage battery. For as the charging
.
proceeds the counter e.m.f. of the battery rises, hence reducing the effective e.m.f. in the circuit; the current, therefore, falls
and as
zero,
it
The current
ing to
decreases, the e.m.f. of the generator also decreases. will therefore continue to fall off until it becomes
and then the battery discharges through the generator, tendmake it run backward as a motor.
wound compound motors

istic of
133. Starting of Differentially Wound Motors. Differentially are seldom used in practice for the
reason that in most cases the slightly drooping speed characterthe plain shunt motor meets the requirements of constant
216
speed to a sufficient extent. Moreover, the differentially wound motor is subject to " racing" in case of heavy overload, due to the considerable reduction of field flux caused by the large current
in the series field winding.
up
in the
wrong
direction
rheostat to the
first
Such motors are also liable to start on throwing the handle of the starting notch; for the high inductance of the shunt
winding, due to the large number of turns, may so impede the rise of the shunt field current that the current in the series winding,
inductance of that
which builds up much more rapidly because of the small circuit, may overpower the magnetizing effect of the shunt winding, and so reverse the flux and the direction of
rotation.
if
In that case the motor,
unloaded, will rapidly speed up in the wrong direction, thereby

ter e.m.f.
developing a considerable counand so reducing the cur-
rent flow and the torque; the acquired momentum of the armature
may even for a brief interval

TIG. 198. Differential effect in interpole motor due to backward shift of brushes.
cause
the machine to become a generator and send current back to the line.
In the meantime the shunt field current has been building up, and when the armature finally stops
because of reduced torque or the dissipation of its energy of rotation,

a heavy flow of current through the armature and series winding will result because there is now no counter e.m.f. The machine will then start up in the right direction, but if the initial current flow is sufficiently great the series excitation may overpower that
shunt winding, and so bring about another reversal of roThis process may go on indefinitely unless the design constants of the machine are such that the successive impulses are damped out, that is, do not synchronize with the natural period of oscillation of the armature. A similar state of affairs may arise in the case of shunt motors provided with interpoles if the brushes are not properly placed. Normally the axis of commutation coincides with the axis of the interpoles, but if the brushes are accidentally shifted backward,
of the
tation.
MOTORS
217
against the direction of rotation, as in Fig. 198, the interpoles will produce a component of flux in opposition to that of the main
poles,
and so convert the machine into one having the charactera differentially
istics of
wound motor.
Speed of Shunt Motors. Inspection of the fundamental equation for the speed of a motor
134. Regulation of
E t
ia r'
reveals the fact that there are three principal methods for regulating the speed, namely, rheostatic control, by varying the resistance voltage control, r', which includes the armature resistance r a
',
by varying the impressed voltage E and field control by varying $. A fourth method occasionally used involves changing Z' by using an armature having two windings and two commutators which may be connected either in series or in parallel.
t ]
(a) Rheostatic Control.
of the
armature
is
In this method the effective resistance increased by connecting in series with it (but
\\\
Baluucer,
Armatures Mechanically Connected
FIG. 199.
Speed regulation of motor by means of voltage control.
not in the main line or field circuit) a variable resistance. This has the effect of imparting a pronounced drofp to the speed characteristic (Figs. 190 and 191), the downward slope of the char-
combined resistance of the armature winding and external resistor. A motor used in this way has poor speed regulation, that is, the speed will fluctuate between rather wide limits as the load changes; moreover, the method is inefficient because of the loss of power due to the flow of the armature current through the external resistor. It is not to be recommended in commercial installations, but is frequently convenient in laboratory investigations and in special tests.
acteristic being proportional to the
218
(b)
Subdividing the voltage of the main Voltage Control. or bus-bars by means of a balancer set, as in Fig. 199, generator
makes it possible to impress upon the armature of the motor a number of different voltages, to each of which there will correspond a definite speed characteristic such as is illustrated in Fig. For any given impressed voltage the speed will be substan191. tially constant and will be approximately proportional to the impressed voltage. The variation in speed between full-load and no-load with normal voltage will usually be between 2 and 10
limit for small motors.
per cent., the smaller limit holding for large motors, the larger It should be understood that the motor
field
connections are such that the voltage impressed on the shunt winding is not changed when the armature is switched from
field flux
one circuit to another, in order that the

Supply
may remain sub-
FIG. 200.
Diagram of connections of Ward Leonard system
of speed control.
The armature connections are stantially constant. means of a special controller, somewhat resembling railway motor controller.
changed by an ordinary
With the arrangement indicated in Fig. 199 it is possible to impress six different voltages upon the motor, namely, 40, 70, Inter110, 150, 180 or 220 volts, giving six different speeds. mediate speeds may then be secured by adjusting the flux by
means of a rheostat in series with the shunt field winding. This method is extensively used for driving machine tools, such as
lathes, boring mills, etc.
It has the disadvantage of requiring a considerable investment in copper due to the extra wires of the
distributing circuits.
Where uniform gradation of speed in either direction is required, as in the operation of the turrets of battleships or in steering by
MOTORS
electrically controlled rudders, the
219
be used.
The motor
M Fig. 200, whose speed

,
Ward Leonard system may

is
to be regulated,
is separately excited from the main supply lines and its armature is supplied from an auxiliary generator G, the latter being driven which takes its power at constant speed by a shunt motor from the line; instead of driving the generator G by a motor, any The field of the generaother form of prime mover may be used. tor is excited from the constant voltage supply line, and may be adjusted from zero to a maximum, in either direction, by means of
a reversing field rheostat; in this
way
it is
possible to obtain a
FIG. 201.
Sectional view of Lincoln adjustable speed motor.
smooth variation of the voltage impressed upon the motor. This method is very effective, but is naturally expensive because
of the auxiliary
(c)
motor-generator
set.
simplest and cheapest method of reguthe of a shunt motor is that in which the flux is lating speed varied by means of a rheostat in the shunt field circuit. If the
Field Control.
The
machine operates normally with a nearly saturated magnetic

circuit, all resistance of the rheostat being cut out, the speed may be approximately doubled by weakening the field current; be-
yond this point the field intensity at the pole tips becomes so weakened by armature reaction, especially under load conditions,
220
that commutation is seriously interfered with. Consequently this method is limited to those cases in which a very moderate
range of speed
will suffice.
The
interpole
motor affords means whereby a wide range
of
speed is or 6 to
possible, a ratio of maximum to minimum speed of 5 The principle of the interpole 1 being fairly common.
made
motor involves the neutralization of the armature reaction of the motor by placing auxiliary poles in the axis of commutation and exciting them by the same current that flows through the arma-
FIG. 202.
Lincoln adjustable speed motor driving pipe cutting machine
ture, the winding of the auxiliary poles being so designed that the m.m.f. of the armature is either exactly balanced or else slightly overcompensated. In this way the main field may be varied
through a wide range without producing sparking, the interpoles always producing a field of the proper strength to reverse the current in the coils undergoing commutation. Interpole motors are used to a very large extent where variable speed is a necessity, as in machine tool operation. They are generally provided with a controller which serves not only to start the motor, and to reverse its direction, but also to vary its speed as desired.
MOTORS
The methods thus
221
far described effect the variation of speed
by
machine. But the adjustment flux and, therefore, the speed can be varied by mechanical In the Lincoln devices which change the length of the air-gap. in the armature in section shown 201, Fig. motor, speed adjustable core is conical, so that as the armature is moved sideways by means of the handwheel the effective length of air-gap may be A range of speed of 10 to 1 is increased or decreased at will.
of the electrical circuits of the
Commutation difficulties at readily obtained in the smaller sizes. are avoided weak by using interpoles. field) (and high speeds
Fig.
202 shows a similar machine made by the Reliance Electric
FIG. 203.
Stow multi-speed motor.
and Engineering Co., the motor here being geared to a pipe cutting machine. Speed variation
is obtained in the Stow multi-speed motor by which are moved in and out of the hollow pole cores by plungers and bevel gears, as shown in Fig. 203. a of handwheel means The speed is increased by drawing the plungers away from the
armature, thereby weakening the field; the thin shell of iron thus at the pole tips becomes saturated, and the commutating field is therefore sufficiently intense to prevent sparking at the
left
upper limit of the speed.
The rapid drop in 135. Applications of the Series Motor. as load is makes this type of the motor its increased series speed
222
of

for traction purposes, as in street and in rolling mills. In railway
machine especially valuable

j
railways and hoisting service
work, for example, motors having a constant speed character-
FIG. 204.
Box frame railway motor, forced

Electric Co.)
ventilation.
(General
FIG. 205.
Split
frame commutating pole railway motor, Westinghouse Elec. & Mfg. Co.
istic like
the shunt motor are seldom used for the reason that the
in going up a steep grade is excesspeed of such a motor will remain substantially the impressed voltage is constant, the additional power
current taken
by such a motor
sive; for since the
constant
if
MOTORS
223
FIG. 206.
Commutator and brushes
of
motor
of Fig. 204.
FIG. 207.
Brush holders
of
motor
of Fig. 204.
224
demands a proportionally increased on the other hand, will slow down as motor, the load increases, automatically preventing an excessive load, and, to a certain extent, tending to maintain a constant load on the" system; at the same time it develops a torque more than proportional to the current, while in the shunt motor the torque increases less than proportionately to the current. Series motors for railway, automobile, hoisting and rolling
required to climb the grade
current.
The
series
mill service are generally of the totally enclosed type.
In
rail-
100
FIG. 208.
Characteristic curves of motor of Fig. 204. Electric Co.)
(General
service, in particular, the motors must be waterproof and of rugged construction to withstand the rough usage to which they are subjected by reason of poor roadbed and improper handling of the starting controller. A too rapid cutting out of the starting resistance results in very heavy current, exces-
way and automobile
and a wracking of the armature winding. 204 illustrates a recent type of railway motor made by the General Electric Company. It is of the box frame, commutating pole type with forced ventilation, the inlet and outlet for
sive torque, Fig.
MOTORS
225
In the box the cooling air being at the pinion end of the frame. frame type the armature may be removed from the frame through the opening at the commutator end. Fig. 205 shows a split frame commutating pole motor made by the Westinghouse Electric and Manufacturing Company. Figs. 206 and ^07 show the commutator and brush rigging of the motor illustrated
20
40
100
120
140
1GO
180
200
Amperes
FIG. 20.9
Characteristic curves of motor of Fig. Elec. & Mfg. Co.)
205.
(Westinghouse
The characteristic curves of the motor of Fig. 204 in Fig. 208; those of Fig. 205 in Fig. 209. 136. Cycle of Operation of Railway Motors. The horse-power of a has little in motor rating railway significance determining its
in Fig. 204.
are
shown
suitability for a particular equipment; the nominal horse-power rating is defined as that load which the motor will carry for one
15
226
hour without exceeding a temperature rise of 90 C. at the commutator and 75 C. at any other normally accessible part, the motor being tested on a stand with an impressed e.m.f of rated value, and the motor covers being arranged to secure maximum
.
1 ventilation without external blower.
In any given case the
motors must be so selected that they will not overheat and the heating depends in part upon the average value of the square of the current taken throughout the whole of the working period, including stops. The current has its largest value during the starting or acceleration period, hence the heating is largely
dependent upon the number
of stops in
is
any given schedule,
When
the car
or train
started, the resistance in series
O
FIG. 210.
a\
l>i
Time
in
Seconds
Speed-time and current-time curves of railway motor.
with the motors should be cut out step by step in such a manner that the current through each motor remains practically constant until all the resistance is out of the circuit. The torque per
motor
then be constant, hence the draw-bar pull and the acceleration will also be constant, and the speed of the resulting car will increase uniformly, as indicated by the line Oa, Fig. 210.
will
When the resistance is all out, the speed will continue to increase, but at a steadily decreasing rate, as represented by the curved line ab, and during this interval the current will decrease from the initial constant value Oe in the manner indicated by curve fg.
The cause
e.m.f.
1
of the decreasing current is the increasing counter
due to the
rising speed.
After the time
Ofri,
the current
is
See 1914 Standardization Rules, A.I.E.E.
MOTORS
shut
off,
227
and the car allowed to coast, the speed accordingly falling manner shown by line be. The brakes are then applied and the speed rapidly falls from cci to zero. The broken line Oabcd is called a speed-time curve, and its area is proportional to the disin the
tance traveled by the car in the time Od. The slope of the line Oa is the acceleration of the car; the value ordinarily used varies from 1 to 2 miles per hour per second, and
turn determines the draw-bar pull, torque and current when the weight of the car, gear ratio and type of motor are known. In cars having a two-motor 137. Series-parallel Control. equipment the motors and starting resistance are at first all conthis in
nected in
series,
and
after the resistance has
been cut out, the
connections are quickly changed so that the motors themselves are in parallel with a resistance between them and the line; this
is then cut out, so that finally the motors are in parallel The elementary diadirectly across the full voltage of the line. gram of connections is shown in Fig. 211. In four-motor equip-
resistance
ments, the motors are usually connected in parallel in pairs, and the two pairs are then connected in series-parallel just as though
each pair were a single machine. The series- parallel control is a much more economical method than if each motor had its own starting rheostat, or than if the
starting purposes.
motors were permanently in parallel with a single resistance for For example, assume a two-motor equipment
with the following data:
E =
/
r
t
= = =
line or trolley voltage current per motor during acceleration period
resistance of each
motor
duration of acceleration period, in seconds.

of starting, the
At the moment
such that
motors being
in series (Fig.
21 la), the starting rheostat
must have a
resistance of
R\ ohms
E
or
Ri =
E -
2r
228

loss in the rheostat at the first instant is
The
then at the rate of
watts; but as the motor speeds up at a uniform rate under the assumption of constant current, the counter e.m.f. also increases uniformly, and in order to keep the current constant the
resistance
PRi
must be cut out
at a uniform rate.
All of the resistance
should be out of circuit in a time ~ seconds, and the motors then
switched to the parallel position (Fig. 2116). During the first half of the acceleration period the energy lost in the rheostat is then
watt-seconds.
At the instant when all of the resistance Ri is out of circuit, each motor receives half of the line voltage and this condition may be maintained efficiently
if it is
desired to continue run-
ning at reduced speed. But if the speed is to be increased at the original rate, as in Fig.
motors must be put and a new resistance R 2 inserted between them and the line. In order that there may be no break in the smoothness of the acceleration, each motor must
210, the
in parallel
FIG. 211.
Elementary
diagram
of
connections, series-parallel control.
and
continue to take / amperes, at the first instant after
the transition has been
made
the resistance #2 must consume
E/2
volts since the remaining E/2 volts are taken up by the motors. The resistance 2 must then have such a value that
ohms
and the energy
lost in the rheostat
during the second half of the
acceleration period will be
w,
I 2 R z t watt-seconds.
MOTORS
The
total loss in the rheostat is then
229
WRi + WR2
R
=
?*t(-T
+Rz)
~
HI*t(j
r
]
watt-seconds.
in parallel, as in Fig. If, now, the motors had been originally 2116, with a resistance of Rz ohms between them and the line, the value of 3 would have to be
r E R = 2I~2 hmS
*
in order to allow a current of /
motor.
The
loss in the rheostat
amperes to 'flow through each would then be
/E
Pi(j
watt-seconds
rj
control. riesor exactly twice as great as in the case of series-parallel 138. Railway Controllers. The successive changes in the
starting resistance and the change from sefies to parallel connection are accomplished by means of a controller. As the,
controller
varied
changed from notch to notch the resistance i^ amounts, so that the current does not remaî by constant throughout the, acceleration period, 4s absolutely curve line the ef, Fig. 210' but in reality this represented by assumes a saw-tooth form lying partly above and partly hfelow
is
finite
the desired constant value.

in
positions of the conyoller in full parallel, ipspecin these positions there because are called running points tively, is no loss in the rheostat; all other positions are called resistance points except in the interval in which the transition from series
The two
which the motors are
in full series
and
to parallel connection takes place. Controllers are commonly designated
by
characteristic letters
which indicate the type to which they belong. Thus, type R controllers are those in which rheostatic control is used, without the customary series- parallel arrangement; they are used for single motor railway equipments, mining locomotives with one controllers or two motors, and for cranes and hoists. Type
are designed for series-parallel operation of two or more series motors, and have the characteristic feature of not breaking
the power circuit during the transition from series to parallel
230
connection.
Type L
controllers are also designed for series-
parallel control of series motors,
and include the feature
of
opening the power circuit during the transition period; this type is now seldom used. Type B controllers have the usual power
circuit connections,
and
in addition allow the
motors to run as
generators for energizing magnetic brakes of the axle or track

type.
Fig. 212 represents a K-10 controller made by the Westinghouse Electric and Manufacturing Co. and Fig. 213 shows the succes-
FIG. 212.
K-10
controller,
Westinghouse Elec.
&
Mfg. Co.
sive stages of the connections. The oval shaped part near the middle of Fig. 212 is a solenoid connected in the main power
circuit; its function is to create a powerful magnetic field at the contacts between the stationary contact fingers and the segments on the controller spindle. This field is so directed as to blow
sents a
out the arcs that form on breaking the circuit. Fig. 214 repremore recent type of controller made by the General
Electric Co.; instead of a single magnetic blow-out coil, there are individual blow-out coils for each contact. The diagram of
MOTORS
231
connections of this controller are shown in Fig. 215. It will be observed that this diagram differs from that of Fig. 213 in that
during the transition period the latter involves short-circuiting and immediately thereafter open-circuiting one of the motors (or pair of motors) while in the latter both motors are continuNotch
Controllers
A' 10
and K\\
1
Res.
Motor
Motor 2
[EQ-o-AA/Wo-AAAAr-
FIG. 213.
Successive stages of connections, K-10 controller.
ously in circuit.
in Fig.
The system
of transitional connections
shown
215
is
called the bridge control.
two handles, one
All controllers, with the exception of certain types, have for the usual operation of accelerating the car,
in
These its motion. such a manner that the reversing handle cannot be moved unless the main handle is in the "off" position, and the main handle cannot be moved
the other for the reversal of the direction of
two handles are mechanically interlocked
232

is
unless the reversing handle
in either the forward or reverse

of rotation
field windterminals. the armature to with respect ings If the car is running and it is desired to reduce speed, the controller handle should be turned quickly to the off position and then brought back again to the proper notch before the speed
The reversing handle changes the direction position. of the motors by interchanging the connections of the
FIG. 214.
Railway motor controller with individual blow-out

Electric Co.
coils,
General
has fallen too low. A slow turning off is apt to draw destructive arcs at the contact fingers. A characteristic feature of all of the controllers described above is that the main current This is passes directly through them.
trailer,
perfectly feasible in the case of a single car, or motor car and but where several motor cars and trailers are to be
operated as a train, the multiple-unit type of control, called type M, must be used. The controller for this service carries only a
MOTORS
233
small auxiliary current supplied by the line, and this current actuates electromagnets which operate contactors that control
the main
The contactors are usually mounted in current. waterproof iron cases under the car bodies. In this system a
FIG. 215.
Bridge control, series-parallel system.
single controller, called a master controller, serves to operate the contactors of all the motor cars in the train, the auxiliary cir-
cuit being extended from end to A system similar to the type
end
is
of the train.
also used in large single cars,
234
Fuse and Switch
TPT
Kicking Coil
.
^rZ^
To Lights and Pu
MU Tripping Sw
Main Switch
Trolley Terminal
Auxiliary Contacts in Controller
T~
FuauBos
FIG. 216.
Railway motor controller with auxiliary
circuits.
FIG. 217.
Contactor with cover removed.
MOTORS
235
where very heavy current through the controller itself might be An elementary diagram of connection of such a objectionable.
controller
Fig. 217.
is
shown
in Fig. 216,
This type of control
which the trolley voltage is feeder at times of heavy load for in such a case it the current passing through the auxiliary circuit
;
and the contactors are shown in is not suitable on a system in to be low, as at the end of a long apt
is
possible that
may
be
insuffi-
Two or more shunt motors designed for the same voltage, when connected in parallel to the same supply circuit, and with their shafts rigidly
coupled, will divide the load in proportion to their capacities provided their speed-current curves (Fig. 191) are identical in
in connection with Fig. 173; that is, if the speed curves, plotted in terms of per cent, of full-load current, are identical. The same thing is true of series motors operating
cient to operate the contactors. 139. Division of Load between Motors.
the
manner discussed
in parallel. Series wound
motors when connected in
series in
a constant-
current circuit, will develop approximately constant torque if the brushes are kept in a fixed position; but if the torque were
the load.
constant the speed would have to vary in direct proportion to To obviate this variation of speed, and in particular
to keep the speed constant, the series motors in the Thury constant-current system (see Chap. VI) are provided with regulators which change the position of the brushes, thereby affecting
the torque instead of the speed. An interesting case of unequal division of load between series
wound motors
of the car is
is
afforded
rails.
grade on slippery
by the case of a car starting on an upAssume for example that the rear end
;
more heavily loaded than the forward end on turnthe controller handle to the first notch, the same current will ing flow through both motors (or both pairs of motors) since they are
in series with each other; therefore each will develop the
If
same
the weight on the forward trucks is fairly light, the torque. adhesion between the wheels and the rail may not be sufficient
to prevent slipping, in which case the forward motor will speed up and spin the wheels. The counter e.m.f. of the forward motor
will increase as its
speed
rises,
so that its impressed voltage
must
also increase; but
any
increase of the voltage on the forward
236
PRINCIPLES OF DIRECT-CURRENT MACHINES,
motor will be at the expense of that impressed on the already overworked rear motor, so that the result will be to stall the car unless the front wheels can be prevented from slipping.
PROBLEMS
A 220-volt compound wound motor has an armature resistance of ohm, a shunt field resistance of 169 ohms, and a series field resistance Find the line current, the shunt field current and the counter of 0.15 ohm. e.m.f. when the armature current is 25 amp. (a) if the connections are
1.
0.44
long-shunt,
2.
they are short-shunt. winding of the motor of Problem 1 is disconnected and the motor, when operated as a shunt motor from 220-volt mains, takes a no-load armature current of 1.5 amp. and develops a speed of 997*r.p.m. Find the value of &Z' and the ideal zero-load speed. 3. The shunt motor of Problem 2 has a magnetization characteristic that can be represented by Froelich's equation, such that an increase of the shunt field resistance to 338 ohms reduces the flux by one-third. What will be the ideal no-load speed if the resistance of the shunt circuit is 220 ohms? 4. When the armature of the motor of Problem 2 is carrying its fullload current of 25 amp., the armature demagnetizing ampere-turns per pole amount to 6 per cent, of the field ampere-turns per pole. Find the speed (a) at full load, (6) when the armature current is 25 per cent, greater than its
(6) if
The
series field
full-load value.
5.
The coefficient of dispersion is 1.2. the total torque developed by the motor current has the values specified in Problem 4?
What
is
when the armature
6.
full-load torque at the moment of starting, (6) a starting torque 25 per cent, greater than full-load torque? 7. The series field winding of the motor of Problem 1 is connected so
of
(a)
What resistance must be put Problem 2 to make it develop
in series
with the armature of the motor
a differential long-shunt compound motor. The series to 20 per cent, of the shunt excitation when the armature current is 25 amp. Find the speed of the motor when the armature current is 25 amp. 8. If the series field of Problem 7 is reversed, what will be the speed when the armature current is 25 amp.? 9. The motor specified in Problem 2 is connected as a shunt machine and is driven as a generator at a speed of 1000 r.p.m. The field rheostat is adjusted until the shunt current is 1.3 amp. Find the terminal voltage when the armature current is 25 amp. (a) when the direction of rotation results in a forward lead of the brushes, (6) when the direction of rotation
is
that the motor

field excitation
amounts
is
reversed.
10.
equipped with four motors which have characteristics is mounted on its own axle, the gear ratio being 25 73. The driving wheels on three of the axles have diameters of 33 in., while those of the fourth are 32 in. in diam. When the car is moving at a speed of 24 miles per hr., with all motors in parallel, what is the current taken by each motor and what is the total tractive effort ?
street-car
is
shown
in Fig. 208.
:
Each motor
CHAPTER
VIII
COMMUTATION
140.
Fundamental Considerations.
Each
of
the a parallel
paths comprising the entire armature winding consists of Z/2a turns in series in each of which the current is i a /a amperes. As
the
commutator segments
to which the terminals of the indi-
vidual winding elements are connected pass under the brushes, the elements are successively switched from a path or circuit
which the current has one direction to an adjoining circuit which the current has an opposite direction. During this transition period, or period of commutation, the current must be reduced from its original value to zero and then built up again to an equal value in the opposite direction. The period of commutation is of very brief duration, of the order of 0.0005 to 0.002 second, and it may easily happen that the reversal of
in
in
current is either retarded or unduly accelerated; in either case, the current at the end of the period will tend to have a value which differs from that of the circuit to which the commutated
about to be connected, and the result of the final equaliis a spark between the brush and the commutator segment. The study of the commutation process therefore has for its object the determination of the conditions which will
coil is
zation
winding element may be represented by a diagram such as Fig. 218, in which ordinates Immerepresent values of current and abscissas the time. of the before the commutation diately beginning period AB, the current in the coil under consideration has the value + i = /a /a; after the completion of the commutation it must have a
value
result in sparkless operation. The time variation of the current in a
assuming that the winding is symmetrical and, that the currents in all of the armature circuits are therefore, = T, the current may vary the same. In the time interval
i
,
AB
in the
manner shown by such
typical curves as a, 237
6,
c,
d, e, f,
238
tions.
each of which corresponds to a definite set of physical condiThese curves are called the short-circuit current curves. Curve a shows that the current has been reversed too rapidly,
its final value, such a condition being characterized as overcommutation. Here the current may reach its
overreaching proper
final
value without a spark, but
it
may
involve such
large localized current densities at the contact surface
between
as to lead to sparking and, perto haps, glowing (incandescence) of the brush and certainly to excessive loss and heating and deterioration of the brush.
commutator segment and brush
Curve
b represents a case in
which the current comes to
its
FIG. 218.
Types
of short-circuit current curves.
value smoothly, with a zero rate of change at the end of the commutation period. This will generally result in satisfinal
factory commutation. Curve c indicates a uniform transition of the current from

its initial
to its final value.
When
this occurs the
commuta-
tion
Linear commutation is very desiras will able, for, appear later, it gives rise to uniform current at the brush contact surface and the loss of power at density
is
said to be linear.
the contact surface

the curve
is
a minimum.
so-called
Curve d represents the

is
"
one-half of a sine curve.
sinusoidal" commutation; Such a short-circuit cur-
rent curve would generally result in satisfactory commutation.
COMMUTATION
Curve
e
239
change of current
vertical line
represents a limiting case in which the final rate of is infinite; that is, the curve is tangent to the
drawn through B.
Under such conditions sparkis,
ing would invariably result. "
Curve / shows
undercommutation," that
the current
is
not reversed with sufficient rapidity. Even though the final value of current may be correct, this condition may involve excessive current density under the brushes and hence possible
glowing, just as in the case of overcommutation. It should be understood that these curves represent only the more important cases. In practice, the short-circuit current
curves
may assume an
infinite variety of forms, subject
always
to the condition that the initial
and
final
values of current must
be equal in magnitude and opposite in sign, the armature winding being assumed to be symmetrical. No account has here been taken of the effect of mechanical irregularities such as vibration of the brushes, unevenness of
the
commutator
surface,
etc.
Such mechanical defects
will
invariably produce sparking even though the magnetic and electrical conditions are otherwise perfect. Vibration of the
brushes causes the short-circuit current curves to take on a
saw-tooth form.
141. Physical Basis of the Theory of Commutation. The theory of commutation is much less advanced than that of other parts of the theory of direct-current machines; that is to
say, the commutation characteristics cannot be predetermined with anything like the degree of accuracy that is possible in the calculation of the general performance characteristics.
fact, practice based upon more or less empirical rules has so far outstripped theory that manufacturers commonly guarantee sparkless operation between no-load and
Notwithstanding this
50 per cent, overload with a fixed setting of the brushes.'
The elementary theory of commutation is relatively simple and has been extensively discussed by numerous writers. It involves the -fact that the coil undergoing commutation has induced in it an e.m.f. of self-induction due to the changing
current in the coil, the self-induced e.m.f. acting always in such a direction as to oppose the change of current; and in case the short-circuited coil is in inductive relation to one or
240

coils
more
which are simultaneously undergoing commutation
mutual induction. For these reasons the theory may be designated the " inductance" theory. There is also to be considered the fact that the short-circuited coil may be situated in a magnetic
there will also be induced in it an e.m.f. of
field
the fringing field near a pole
tip,
or the reversing field
due to a commutating pole and that the rotation of the coil through this field produces a generated e.m.f. in the coil. As has been previously pointed out, this generated e.m.f. should in general be so directed as to neutralize the retarding effect of self-induction, in which case the process is referred to as voltage
however, the commutated coil is not acted field, that is, if there is no generated e.m.f. acting in it, the process is called resistance commutation
commutation.
If,
upon by any extraneous
inasmuch as the self-induced e.m.f. is controlled only by the ohmic drops in the coil and at the brush contact surface. Resistance commutation is largely relied upon in machines having a fixed brush position and no special commutating devices such as interpoles; to this end carbon brushes are employed, the
high contact resistance serving to keep the short-circuit current within reasonable limits. Examples of this type of machines
by railway and hoisting motors in which the brushes are permanently set at the geometrical neutral because of the frequent reversal of direction of rotation.
are afforded
The ohmic drops
at the transition surface between the
commu-
tator segments and the brushes, and in the short-circuited coils and their connecting leads are almost as important as the
e.m.fs.
due to
self-
and mutual-induction and
to rotation through
the magnetic field. The contact resistance between commutator and brushes is very much more complex in its nature than that of ordinary metallic conductors; it resembles that of the
electric arc in
many
of its properties, for
it is
dependent upon
such factors as the current density, the direction of the current, the temperature, material and chemical structure of the contact surfaces, and upon the nature of the current itself (continuous, it varies, moreover, with the conalternating, or pulsating)
;
tact pressure
1
and the
1 relative velocity of the surfaces.
found
very complete presentation of the effects of in Arnold's Die Gleichstrommaschine, Vol.
all of
I.
these factors will be
COMMUTATION
It
is
241
transition surface
probable that the passage of .the current across the between commutator and brush causes an
sets
ionization of the gaseous layer between them and that this up a counter e.m.f. similar to that encountered in the arc
stream.
From
contact surface
this standpoint the drop of potential across the is the sum of the counter e.m.f. and the true
ohmic drop; the quotient obtained by dividing the observed drop by the current is then not a true resistance, but what may be called an effective resistance, made up of the true resistance plus a fictitious resistance equivalent in its effects to the counter e.m.f. The transition layer between commutator and brush is the seat of an energy storage, and breakdown in the form of sparking may be expected when the amount of the stored
energy exceeds a critical value. On this basis, neither current density nor transition drop taken separately is a sufficient criterion of the sparking limit; this is confirmed by an experi-
ment
of Professor Arnold's in which the current density passing from a carbon brush to a metal surface was raised until the brush glowed, but without producing sparking.
THE INDUCTANCE THEORY

142. General Equation, Case of Simple Ring Winding. For the sake of simplicity there will first be considered a simple ring winding in which the brush width b is equal to the width a commutator segment. Under this condition only one |8 of winding element will be short-circuited at a time, as shown in Fig. 219, and the effect of the mutual induction of other coils is eliminated. It has been explained in Chap. VI that the axis of commutation must be slightly displaced from the neutral
axis (in the direction of rotation in the case of generators), in order that the fringing field at the leading pole tip may generate
in the short-circuited
element an e.m.f. of sufficient magnitude
to balance the retarding effect of the self-induced e.m.f.
But
during the commutation period the short-circuited coil moves through the fringing field from a position in which the generated e.m.f. has a certain value to another position in which the generated e.m.f.
is
not constant.
16
appreciably larger, so that the reversing e.m.f. If the distribution of the flux in the air-gap
is
is
determined experimentally, as by the pilot brush method de-
242

be found that, as a rule, the curve approximately linear for short distances therefore, the commutating e.m.f. may be
it
scribed in Chap. V,
will
of flux distribution is
between pole
tips;
closely represented
by the function
E =
c
kt
(1)
where
t
time counted from the beginning of the commutation

period
h
e
= constant = commutating when = 0.

t
e.m.f. at the beginning of the period,
FIG.
219.
Short-circuited
element,
FIG. 220.
Successive phases
simple ring winding.
of short-circuit of coil.
and c of Fig. 220 represent, respectively, the intermediate, and final stages of the commutation of a
Parts
a, b,
initial,
coil C.
a part of the right-hand branch of the = ia /a. All of the current and carries the current winding the brush from the two reaching adjoining paths must then pass lead 1. in final position c, coil C has the through Similarly, become an integral part of the left-hand branch, its current has
coil is
z'
In position a the
been fully reversed, and the combined current of the two paths must reach the brush by way of lead 2. The b position, in
which
coil
segment
A2
is short-circuited, shows that immediately after has reached the brush, the current from the right-
COMMUTATION
243
hand branch may reach the brush by way of both leads 1 and 2, and coil C therefore carries less current than before; as the contact area of segment AI diminishes and that of A 2 increases, the original current through C is diverted more and more from lead 1 to lead 2. At the same time that the right-hand branch current is being throttled in this way out of coil C, the left-hand branch current finds its way more and more readily through coil C and the increasing contact area A 2, and less and less
readily through the diminishing contact area AI. If the transfer of the brush current, 2?! from lead 1 to lead 2
,
occurs uniformly during the commutation period, linear commutation results. In that case the current in C will have zero value
insulation between segments A\ and A 2 is directly under the middle of the brush; and leads 1 and 2 will then each be carrying current io, from the left- and right-hand circuits, But if the axis of commutation is too near the respectively. leading pole tip (in the case of a generator) the e.m.f. generated in C by its motion through the field will act to accelerate the transfer of current from lead 1 to lead 2, therefore giving rise to abnormal current densities at the contact area A%j this is the case of overcommutation, Fig. 218a. On the other hand, a
when the
commutating
field
that
1
is
too
weak
will
delay the transfer of
current from lead
become
143.
to lead 2, so that the current density may excessive at contact area AI; this corresponds to the
curve of undercommutation, Fig. 218/.
same
Elementary Mathematical Relations. Fig. 221 is the as Fig. 2206 except that the currents in the various paths are indicated. It may be assumed that the currents from the
left-hand and right-hand paths pass to the commutator by way of leads 1 and 2, respectively, and that the current in the coil C
has a value which at any instant is i amperes, its path being completed through the brush. In the figure the current i is
represented as flowing in a clockwise direction through the short circuit but at a later instant during the commutation period it will have reversed. From the figure it follows that the currents
in leads 1
and 2
ii iz
are, respectively,
to
i
.'.
ii
z' 2
= = =
+
=
i i
\ }
(2)
2io
total current
from the brush,
(3)
244

let
Now
R = RI = Rb =
c
resistance of coil resistance of each
C
commutator lead
of coil
resistance of contact area of the entire brush.

t
Counting the time
from the beginning of the short-circuit
C, the contact resistances of areas A\ and
RI
T ,
R*
t
= Rb
A T
are, respectively,
2,
In the closed circuit consisting of coil C, commutator leads 1 and contact areas A\ and A 2 the two commutator segments and
the brush, the

to sign)
must
of all the potential drops (with due regard equal zero, in accordance with Kirchhoffs law.
sum
The
sign of any e.m.f. is to be taken positive or negative if it acts in, or in opposition to, some arbitrarily assumed positive
FIG. 221.
Current paths in short-circuited
coil.
direction, respectively; similarly, a drop of potential due to a current flowing in the positive direction is to be taken with the
negative sign, and vice versa.

direction as positive,
Taking the counter-clockwise
we have
L-jr=
e.m.f. of self-induction (negative)
= = iiRi iiRi = iiR'i = izRi = Ec =

iR c
ohmic ohmic ohmic ohmic ohmic

di
dt
in lead 2 (negative) commutating e.m.f. (positive).

.
drop drop drop drop drop
in coil C (positive) in lead 1 (positive) in contact area AI (positive) in contact area 2 (negative)
+iRc
>T
._
vJ
_T__
J.l/1)
~*
pjy
j,
1
c
T -C C*0
RI
+E
(4)
COMMUTATION
which
245
may
be written
L
where
S+
iR
f'-^t
<*''
i}
^T
l
(i
&
(5)
R = R + 2R
c
This equation involves the justifiable assumption that the resistances of the commutator segments and of the brush are negligible.
The complete integration of this differential equation has been worked out subject to certain conditions 1 and results in an equation of the form
subject to the terminal conditions that
when
0, i
i Q)
and
when
= T,i =
iQ.
The complete results of
the integration are
not essential in most cases, since it is generally the end of the commutation period that is most important so far as sparking is concerned.
144. Discussion of the
General Equation.
At the last moment

T, and
i
during the commutation process,
when
iQ
E =
c
7?
j
T7
O'o
hT = E T = 2Rbio i)
i)
V
The value
by
T R --m i(*o + i
b
t
o or indeterminate. n o
,
of the expression
(* _
can, however, be evaluated
differentiating numerator and denominator separately with respect to the independent variable t, giving
di
T The
di
i
i
t
dt
di
dt
di T ~TT
at
general equation then reduces to

T j j,
t'Qfl
J?
ac
-flfr-i
97?
'
^/t5?-o "f*
xaf
n U
or
A\
c?^
1
-^f RbT-L
die
Uber den Kurzschluss der Spulen und
Vorgange bei der KommutaRiebesell.

Kiel, 1905.
tion des Stromes eines Gleichstromankers,
by Paul
246
From the last equation (6) there tant conclusions, as follows:

1. If
may
be deduced several impor-
j~
=
oo,
1,
the final value of

i
-rr
will
be
infinite, i.e.,
(-77)
\(lt J
t
=
= T
provided
(R
274) differs from

di
L-rr,
ET
If this
were the
infinite,
case, the e.m.f. of self-induction,
would
also
be
and sparking would

7?
/77
result at the trailing edge of the

differ
brush.
2.
Therefore, -j
must in general
218a shows
from
in
unity.
Inspection of Fig.
that
case
of
overcoil is
commutation the
final rate of
change of current in the
positive in sign; since this condition of overreaching is to be avoided, the rate of change of current should always be negative,
hence the numerator and denominator of equation

the
(6)
must be
of
same
sign.
If,
then,
1,
iQ
^>
and
if
(R
274)
> ET
(7)
~r <
,
1,
(R
274)
< ET
(8)
Inasmuch as most machines are required to operate with the brushes in a fixed position, close to the neutral axis, the cornmutating e.m.f. E c and also its terminal value, E T will generally be small. Usually, therefore, E T
LJ
is
in
3.
most cases the
The
final
criterion for satisfactory commutation. rate of change of current will be zero,
i.e.,
/di\
ET =
in
(R
274)
will
(9)
which case the conditions for good commutation

This state of
affairs is
be favor-
able.
7?
shown
in curve
145. Modified V7
Form
of
Sparking
6, Fig. 218. Criterion. The condition
j-~
>
can be put into another form which has a simple physical
COMMUTATION
interpretation;
2i Q L/T,
247
the inequality
multiplying
both
sides of
by
we have
L=e
face
(10)
The term 2i Q R b is the drop of potential at the brush contact surand is usually of the order of 1 volt with carbon brushes. The term 2i /T is the average rate of change of current during
commutation, hence e r = 2i QL/T is the average reactance voltage, or average e.m.f. of self-induction. It follows, therefore, that e r < 1 is the condition to be satisfied.
The criterion R b T/L > Ishowsthat the brush contact resistance Rb and the time of commutation T must be large, and that L must be kept small. This accounts for the fact that carbon brushes are ordinarily superior to metal brushes, since the former have
the larger contact resistance.
248
It would appear at first glance that an increase in brush width would give good results because of the increased value of T.
this apparent advantage is offset because of the fact that the wide brush simultaneously short-circuits several additional coils whose mutual inductance is equivalent to an increase in the selfinductance of the original coil (see Art. 154).
But
The
criterion
R b T/L >
1,
or e r
2i QL/T
<
1 is
frequently ex-
H. M. Hobart has proposed a method which assumes that the commutation curve (Fig. 21Sd)
pressed in a
still
different form.
is
the
one-half of a sine curve of period 2T, as indicated in Fig. 222, maximum ordinate being ? whose equation, referred to
,
origin O,
is
I
2-irt
IQ
COS
om
is
The instantaneous value
of the self-induced e.m.f.
then
and
is
its
maximum
value, which
7T
Hobart
_ 2io j
calls the
reactance voltage,
7T
7T
Hence, emax< 2.
or e max
if
er
<
1,
e max
<
1.57.
As usually
stated,
however,
It follows
2,
>
from the previous discussion that if e r > 1, the commutating e.m.f. must be so adjusted that
E T >i
(R + 2R b ). 146. Linear Commutation. Equation (5) can be utilized to determine the conditions necessary for a uniform transition of the current from its initial to its final value. Thus, in case of linear commutation, curve c, Fig. 218, the current i at any instant
t
is
given by
.
2i
*0
'
4 I
T -2t
jT~
yT
di _ ~ dt
If)
and
2i Q
T
some
trans-
Substituting these values in (5), there results after
formation
"?
~ %
(T
2t) ]
(12)
COMMUTATION
It follows
249
from
(12) that
t
when
c
= =
0,
E =
c
e '=
(^r
IQ
f
T,
E = ET =
-=jr
Rj
In other words, the commutating e.m.f. must not only vary as a linear function of the time, as shown by equation (12), but it
must
to
it is directly proportional conditions for perfect linear commutation were satisfied for one particular load they would
iQ.
also
change with the load since

if
It follows, therefore, that
not be satisfied for other loads, without special corrective devices. An interesting consequence of linear commutation is that the
current density at the brush contact
to Fig. 221,
is
constant.
Thus, referring
iz
IQ
and
if
T-
2t
it
follows that
.
ii
=
=
9 2i
T-t
j,
j,
iz
2iv
But the contact areas
and
!
A
=
are given
t
by
rA T
t
T
where
is
the total brush area; hence the current density

i\
is
-r-
^1
It
jV2
2iin
r A
constant.
(13)
of the
can also be shown that the ohmic loss due to the resistance brush contact is a minimum in the case of linear commu-
For let it is, when the current density is uniform. be assumed that the short-circuit current in coil C, Fig. 221, is not linear; in this case the actual non-linear current can be
tation, that
250

t,
thought of as made up of a linear current, i and an extra It current, i xt where the latter may have any general form. follows then that
=
and
i\
ii
ix
?o -f- i
io -\- ij -|-
ix
The contact
resistances at the areas
AI and
A2
are, respectively,
T
and
# =
2
Rb
and the ohmic
loss at the contact areas is
Substituting for
rp
?'o
ii,
iz ,
Ri and
7? 2 ,
and remembering that
it
21
,
we
find that
=-
4i Q *R b
i f *R b
i x *R b
~-minimum
if i x
from which it follows that the if the commutation is linear.
loss is a
0, i.e.,
147. The Current Density at a Commutator Segment. General Case. The uniform current density over the brush width that is characteristic of linear commutation means that
the drop of potential across the contact surface is everywhere the same and, therefore, that there are no differences of potential along the brush contact (in the peripheral direction) to be
equalized by a flow of current.
It follows, then, that if such inequalities of potential do exist, or tend to exist, extra currents will flow along the brush and complete their paths through the
short-circuited coil or coils, thereby giving rise to non-linear short-circuit current curves and a non-uniform current density.
Of course, the potential differences which produce the extra currents are due to the fact that the e.m.f. generated in the
COMMUTATION
251
short-circuited coils differs in form, as a time function, from It is conthat which would produce linear short-circuit current.
sequently important to determine in what manner the distribution of current density is affected by a non-linear short-circuit For this purpose the following graphical method, due current.
to Professor Arnold, 1
may
be used.
Consider, for example, a case where the brush width is 3.5 times the width of a commutator segment, and let it be assumed that the current density at any given instant is the same over
the entire area of that part of the segment covered by the brush. Assume also that all the coils successively undergoing
commutation have
identical short-circuit current curves.
The
current density at a particular segment
will
then change from
JL
i\
iz
=io
*i
*2
ii=-o
*s
a
FIG. 223.
Successive phases of short-circuit, wide brush.
instant to instant as
it passes under the brush. phases of its motion may be recognized 1st, the segment approaches the brush,
:
Three distinct
segment is covered by the brush, segment emerges from the brush, as shown in Fig. 223, parts a, b, and c, respectively. Phase 1. The current crossing from segment S to the brush
2d, the 3d, the
iz
is
ti
io
i\
the
upward
2.
direction of current through
the segment being
taken as positive.
Phase
The current
crossing from segment

iz
to the brush
is
ii
Phase
3.
The current
iz
crossing from segment

ii
to the brush
is
iz ~\~ io>
since in this position coil
A
=
is
carrying current from the

i
.
left-
hand branch
1
circuit, i.e., i\
Die Gleichstrommaschine, Vol.
I,
p. 438,
2nd
ed.
252
In Fig. 224, curves C A and C B represent the short-circuit current curves of coils A and B, respectively. They are drawn in their correct time positions with respect to the edge 1 of the
brush, the line
tion
OO' being the
shown
is
in the figure the first
axis of their ordinates; in the posiphase of the motion of seg-
ment S
When
just beginning. the commutator has
moved
to the left a distance Xi
(during phase 1), across segment S is

iz
the current in coil
is
ab,
and the current
i\
io
i\
etc
ab
be
f_i
FIG. 224.
Current density at a segment.
The
current density
is
proportional to
io
ii
be
Xi
Me
it
Draw
vertical
the straight line Mb and produce through P in the point d, then

_
until
it
cuts the
'Me
MP
is a constant length, Pd is proportional to the current density at segment S. Projecting the point d across to e, ce is the current density corresponding to the abscissa XL
and since
MP
COMMUTATION
It is readily
initial reversal of
253
apparent from this construction that a too rapid the current in a coil (overcommutation) may result in excessive current density at a segment as it passes under
the brush.
During the second phase, or

represented by x%
is
}
after a travel of the
commutator
i.e.,
given directly by This intercept is also proportional to the current density to the
.
the current across segment S, the intercept fg between curves
iz
ii,
C A and C B
same
scale as
ce.
During the third phase of the motion, or after a travel indicated by x 3} the current across S is
iz
i\
iz
^'o
hk
o'
FIG. 225.
Curves of current density at segments.

of the
At the same time the length

with the brush
to
-,
segment
still is is
in contact
is
kQ, hence the current density

it
proportional
^,
kQ>
or,
to the scale previously adopted,
represented
by IN
line Qhl.
mk.
The point
is
found by drawing the straight
Continuing this method for several points in each of the three phases of the process, the current density curve afiyd of Fig.
225
For convenience, the ordinates representing is obtained. current density have been drawn upward with respect to the
254
Similar curves, a'0'y'd' and a"/3'Y'$", etc., show axis OCX. the variation of current density of segments S' and S", etc., respectively, and to the same time scale as that of curve afiyb.
Local Current Density at the Brush. Except commutation, the current density is not the same at the same instant all along the arc of con148. Variation
of in the case of linear
tact of the brush, nor does it remain constant at any given point. For instance, in Fig. 225 consider the point 7 of the brush, which is just about to make contact with segment S'" At this instant
.
the current density of segment S"' is ab, as read from curve a", and thereafter, until S'" has moved from under the point,
current density changes in accordance with curve bd; segment S" then comes under point 7, the density rises suddenly from
dc to
e
the
ec, the point e being on curve a", and again falls off to the value gf. The curve eg is, of course, the same as bd; in other
words, the current density at a given point in the brush varies

FIG. 226. Average local current periodically. density at points along brush.
In
precisely
line
the
same way the saw-tooth
with
cusps at
j8, /3', /3", etc., represents density at point 3 of the brush.
the variation of current
If the average current density is found for various points along the brush, the results when plotted give the curve 1'2'3'...8' of Fig. 226. This is the curve of average local current density
or PQ of that corresponds to the curve of commutation the local from 226 that 224. It is Fig. Fig. readily apparent
densities may differ considerably from the average current density of the brush as a whole. 149. Further Examples. Using the above methods, curves
MN
current
of current density
have been constructed

are
short-circuit
current curves.
228, since
is not shown, The case of obvious from the previous analytical discussion, as well as from the geometry of the construction, that the current
and 229.
it
shown They linear commutation
for several types of in Figs. 227,
is
COMMUTATION
Short-Circuit Current
255
Curve
Segment Density
Axis
of Ordiuatea
FIG. 227.
Short-Circuit Current
Curve
Segment
Density
FIG. 228.
Short-Circuit
Current Curve
Axis coordinates
FIG. 229.
FIGS. 227, 228, 229.
Curves of current density at segments and brushes.
256
density

is
the same at
all
times both at the segments and at

in
each point of the brush.

Figs. 227
of current
is
and 228 represent cases
which the rate of change
tion period. ties at the commutator segments as they
too great during the initial stages of the commutaThis results in correspondingly great current densi-
come under the brush, near the heel. current In Fig. brush and high average density in the the current coils causes a of reversal 228 the overreaching of direction of current at the receding segments. Fig. 229 shows a case of undercommutation, with consequent reversal of direction at the heel of the brush and excessive densities near the end
of the
commutation period. 150. Simultaneous Commutation

as the process of
of
Adjacent
is
Coils.
Inas-
much
commutation
in a coil
affected
by the
FIG. 230.
Simultaneous short-circuit of elements of lap winding.
mutual induction
of neighboring short-circuited coils, it is important to be able to predetermine the number and relative positions of those coils in the same neutral zone which are simultaneously short-circuited. In the case of the simple ring winding heretofore considered, all coils in the same zone are short-circuited by a single brush if b is the brush width and (3 the width of a commutator segment, the ratio b/0 fixes the number of coils short-circuited at the same time. This ratio is generally
;
a mixed number, and the actual number of coils short-circuited will vary alternately between the two integers lying on either side of it. In lap and wave windings, however, the conditions
are as a rule not so simple, since in a given neutral zone some of the conductors are short-circuited by a brush of one polarity, others by a neighboring brush of opposite polarity, as illustrated
COMMUTATION
in
257
The diagram represents a duplex lap winding Fig. 230. having the following constants:
Z = y =
It is clear
1
122
S =
-
61
p
2/2
6
19
a ^
= =
12
2.5
m=
2/1
23
= ~
from the figure that in the position shown, conductors and 4 are simultaneously short-circuited; a moment earlier conductors 1, 3 and 4 were short-circuited. The successive combinations of short-circuited coils can be conveniently studied by means of the following graphical method, due to Professor
Arnold. 1
1.
Lap Windings.
be observed that in the winding here selected brushes
It will
BI and B% are not identically situated with respect to the segments of the commutator in contact with them. This is a consequence of the fact that S/p is not an integer. Coil edges 1, 3, 5, etc., drawn in full lines to indicate that they occupy the tops of the slots, are connected to commutator segments which are correspondingly numbered in the top row of The other sides of the same coils, whose numbers are figures. l + 2/i, 3 + 2/i, 5 + 2/1, etc., are connected to segments which are numbered 1', 3', 5', etc. (i.e., dropping the term 2/1 and priming the numeral) in the bottom row of figures. A coil will then be short-circuited when the brush BI is in contact with any A similar pair of segments which bear the same numbers. arrangement
is
Now,
coil
indicated in the case of coil edges 2, 4, etc. edge 2 is connected to one on the left which
is
1 from 1. Segment 2 separated by a pitch 2/1 from 2, and by 2/1 But is therefore separated from 1 by J^ (2/1 1) segments. brushes B% and BI are separated by S/p segments, hence the relative shift of segments in the vicinity of B 2 with respect to
those at
is
A
and
it
is
=
|
M(j/i
1)
(15)
is
toward the
left
when
is
negative, toward the right
positive.
The simultaneous
1
In the case considered in Fig. 230, action of the two brushes can now be studied
I, p.
when %.

17
354,
2nd
ed.
258
of a diagram like Fig. 231; take a strip of paper cut to the width of the hatched area to represent the brush and slide it between the two commutators; when it touches segments simi-
by means
larly
numbered, the corresponding
coils will
be simultaneously
short-circuited
FIG. 231.
Diagram showing
coils
simultaneously short-circuited winding.
lap
2.
Wave
Fig.
Windings. 232 represents a portion of a duplex wave winding having

122
2/i
the following constants:
Z =
y
=
2/2
61
p
|
=
=
6
2.5
21
coil
In the position shown,

circuited.
edges 122,
1, 2, 3,
and 4 are short-
FIG. 232.
Simultaneous short-circuit of elements of wave winding.
the segments connected to coil edges 1, 3, 5, etc., with corresponding numbers, and the segments connected to the
Number
other sides of the same coils
1', 3', 5', etc.; similarly with respect to the other coils occupying the same neutral zone, as 2, 4, 6. Brushes AI and A 2 which are of the same polarity, are not similarly placed with respect to the segments in contact with
,
them.
The brushes
are separated
by 2S/p segments, while the
ends of an element are separated by y
2 Sf
segments,
COMMUTATION
where
259
m =
a/p; segments
1',
3',
5'
are therefore shifted with
respect to A 2 by an amount m(3, as compared with the relative The positions of segments 1, 3, 5 with respect to brush A\. is positive (as in the case illustrated), to shift is to the right if
the
The short-circuiting of these elements left negative. can then be shown by drawing two commutators one above the
if
is
other, as in the
upper part of Fig. 233. Obviously, brushes B\ and B 2 and the segments in contact with them, are related to each other in the same way as are AI and A 2 Now Bi is separated from AI by S/p segments, while
,
.
segment
2' is
separated from segment
by
^ (y\
1)
segments.
FIG. 233.
Diagram showing
coils
simultaneously short-circuited winding.
wave
The displacement
of
compared with that
of
segments 2', 4', etc., with respect to B\ as segments 1, 3, 5 to A i, is then

t
A
and
is
= |
is
if it is
y yi
2(
i)
A
are
is
=
If
to the right if %), to the left

in
positive (in the winding considered The complete relations negative.
shown
Fig. 233.
a strip of paper whose width is equal to that of -the brush moved across the fictitious commutators represented by AI,
,
A%, Bi
and B 2 it will touch a series of similarly numbered segments and the corresponding coils will be simultaneously shortcircuited.
151. Successive
Phases
of Short-circuit in Coils of a Slot
The method described above may be used
to investigate the order in which the coils occupying a given slot undergo commutation. Two distinct cases may be distinguished:
260
1.
Coil edges lying in the same radial plane (one above the other) enter and leave short-circuit simultaneously.
2.
Coil edges lying in the

1.
same
radial plane enter
and leave
short-circuit at different times.
If the coil edges are numbered in accordance with the in Art. 73 of Chap. Ill, and illustrated in described system and 2 of a two-layer winding will occupy the 1 coil sides Fig. 86, same radial plane, and so also will 3 and 4, 5 and 6, etc. Refer-
Case
ence to Figs. 231 and 233 shows, therefore, that if coil sides 1 and 2, 3 and 4, or, in general, any two in the same radial plane, are
and leave short-circuit simultaneously, there must be no displacement between the correspondingly numbered commutator segments; in other words, the condition to be satisfied
to enter
is
that
A =
or
2/i
X
1
(tfi
1)
~+
(16)
For example, consider the case of a simplex lap winding having coil edges per slot, a brush width of 2^ segments, and A = 0. With the help of a diagram like Fig. 231, but with A
six
equal to zero, it is readily shown that the successive phases of the short-circuiting of neighboring coils will follow the order shown in parts a, &, c, etc., of Fig. 234, where the shaded coils
indicate short-circuit conditions.
made
During a
brief interval the

is,
condition
shown
in
diagram
will exist,
that
all
the
coil
edges in a slot will be simultaneously short-circuited; a little later, as in diagram e, six coil edges are again short-circuited, but
four are in one slot and two in the next
slot.
study
of Fig.
234 shows that when
coil
edges
and 2 leave
short-circuit they are subject to the effect of mutual induction from the simultaneously short-circuited coils 3, 4, 5 and 6, all of which occupy the same slot. When coils 3 and 4 leave shortcircuit they are subject to the mutual induction of coils 5 and 6, which are in the same slot, and of coils 7 and 8, which are in the
next
slot; obviously, because of this separation of the short-circuited group of coils, the inductive effect upon coils 3 and 4 will be smaller than in the case of coils 1 and 2. Similarly, when
COMMUTATION
coils
261
5 and 6 leave short-circuit they are acted
upon by the mutual
induction due to the simultaneously short-circuited coils 7, 8, 9 and 10, all of which are in the slot adjacent to that occupied by 5 and 6, hence the inductive effect upon these two coils is still
less
than in the case of
coils
3 and 4.
The commutating
condi-
tions are, therefore, not the same in all of the winding elements, and their short-circuit current curves will have different forms.
An additional disturbing feature arises from the fact that when the successive coils of a slot, as 1-2, 3-4, 5-6, of Fig. 234 break contact with the brush, they are not identically situated with respect to ,the adjacent pole tip, consequently the e.m.fs. generated in
(a)
1
(b)
5
1
135
(d)
IDD
IDD
II
II
246
Dll Dll
Dll
IDD IDD
DDI
DDI
262
in 3
and 4, and much too great in coils 5 and 6. In the latter coils there will be a condition of overcommutation, and under these
circumstances every third commutator segment may become blackened because of the possible excessive current density. In order that there may be no marked difference between the field
intensities acting
upon the various coils of a slot while they are undergoing commutation, the angle subtended by a slot should be small. For this reason the number of slots per pole should not be less than 12, and preferably greater than 12, and the angle subtended between the edges of a brush should not exceed one-twelfth of the angle from center to center of the poles. 1
oo
1
(O
5
1
II
III
II
(/) 79
an
ii
135
IQD
1
CO 3
79
11
135
11
00 79
11
II
IDD
II
III 246
FIG. 235.
DM
8
10
12
IDD
8
10
12
246
246
coils,
10
12
Successive phases of short-circuit in adjacent
A =
1.
The order of commutation illustrated in Fig. 234 can occur only in full-pitch windings, since it is in such windings that the back pitch, y if is made nearly equal to 2S/p. Chord windings (fractional pitch) are, therefore, characterized by the condition
AÔ.
Case
occurs
2.
It follows
from the above analysis that the second

variation of this case
case arises
when A ^ 0. An interesting when A = 1, that is, when
->
(17)
Thus, if a winding has pitches that satisfy equation (17) and is arranged so that each slot contains six coil edges, the brush
1
Gray, Electrical Machine Design.
CO MM U TA TION
263
of adjacent coils will be as case, pairs of coils, like 2
width being 2j^ commutator segments, the order of commutation shown in Fig. 235. In this particular and 3, 4 and 5, etc., enter and leave
short-circuit simultaneously. 152. Selective Commutation in
Wave Windings.
study of
the simplex wave winding shown in Fig. 81 (p. 93) will show that the several brushes of one polarity are connected to each other
not only by an external conductor but also through the winding by way of the coils that they short circuit. The figure also shows
that the resistances of these internal paths are not equal because of the varying areas of brush contact; further, the short-circuited
coils are
fringing fields
not at any instant identically located with respect to the through which they are moving, hence the e.m.fs.
field,
generated in them by rotation through the are not the same in any two of
though small,
them.
Both
of these facts are
responsible for an unequal division of the total armature cur-
rent
between the several

The unequal comof
brushes.
ponents
,
the total
current
.
from brush to brush in cyclical order, in such a way that Kirchhoffs laws are continuously
shift
. 236. FIG. 00 Diagram showing duration of short-circuit.
satisfied.
This shifting of
the current values at the brushes in the case of

called selective
wave windings
is
In the case of a simple ring is simply the time required the duration of short-circuit winding for a given point on the commutator to move through an arc
equal to the width of a brush. But it will be seen from Figs. 231 and 233 that this simple relation does not hold in lap and wave windings, since a coil is short-circuited only when similarly numbered segments are simultaneously touched by the brush.
commutation. 153. Duration of Short-Circuit.
These segments being displaced with respect to each other, the time of short-circuit may be either greater or less than in a ring
winding. Consider the case of a multiplex lap winding, Fig. 236 (drawn to represent a duplex winding); the distance between corre= sponding edges of similarly numbered segments will be m0
264

where
ft
ft,
is
the width of a segment.
Short-circuit of coil
will
endure while the edge
of the
is
until edge
n reaches
q.
When n
brush moves from position at q, m will be at p', hence

p',
the short-circuit exists over the distance from p to equals

b
which
(mft
is
ft)
ft
(l
-}
The time
of short-circuit
then
(18)
where
v c is the peripheral velocity of the commutator. exactly similar result is obtained in the case of wave windings where the number of brush sets equals the number of poles. If one or more pairs of brush sets are omitted, the necessary
An
correction can be applied by remembering that mp, Fig. 233, is the displacement corresponding to the distance between a given brush and the next brush of the same polarity. Therefore, if
some
of the brushes are
removed, the term mft
ft
in
the
for T must be multiplied by the number of double pole pitches in the region from which the brushes have been omitted. In simplex lap windings a/p = 1, hence T = b/v c or the same T is less than this in multiplex lap windas in a ring winding. In wave windings, on the other hand, T is greater than ings.
above equation
in a ring winding, other things being equal, since
a/p
<
1.
154. Simultaneous
of
Commutation
of
Several
is
Coils.
Effect
less
Wide Brushes.
When
the brush width
equal to or
than that of a commutator segment, sparkless commutation

generally requires the satisfaction of the condition R b T/L > 1. F. W. Carter has shown 1 that the same condition, in somewhat
modified form, also holds when several coils are short-circuited simultaneously, that is, when the brush spans several segments, as in the preceding sections. Consider the case of a ring winding in which two coils are
simultaneously short-circuited, as in Fig. 237, the coefficients

1
Electrical World, Vol.
LV,
p. 804,
March, 1910.
COMMUTATION
of self-
265
and mutual-induction
respectively.
necting leads, currents in the coils

di*
J
of the two coils being L and M, Then, neglecting the ohmic drops in the conthe differential equations that determine the
become
(see Art. 143),
dt
^+E
Rb
f
.
ci
=
(19)
di.
Vo
_x _
Xl
Ec2 =
(20)
where Xi = (3/b and x is a linear function of the time, such that xb is the brush overlap on the receding segment. Evidently,
FIG. 237.
Current paths, several
coils
simultaneously
short-circuited.
dx
266

i\
the commutation period
IQ
and x =
finite.
0, so
that the ratio
becomes 0/0, and

the current
(^
this
i)
must remain
can be expressed as
Assume now that the sum of a series of

x
ascending powers of x (or of the time); then the ratio

will
there are no powers of x less than the first. If there are any powers less than the first, both sides o equation (21) will become infinite. Let the lowest power be x a so
remain
finite
if
that as x approaches zero, terms involving higher powers become 1 a a~ l = =

negligible
and
iQ
ii
Ax and
-r-
Aax
Also,
if
current iz is represented by the same short-circuit curve as i\ and therefore analytically by the same equation but with a
y
shift of the
x coordinates, -Tdx
'
= Bx a ~
for small values of x.
Substituting these values in (21), and (22) and transposing
-}-
ci
finite
quantity (23)
and
,-,
(L T
ff,
>
-\-
-fff
M Aa) X
A
I^Rc
r>
^&
r>
2 r^ ~
**
"
^
i
^ 2
1
#i_l
Xi
EcZ =
If
finite
quantity (24)
than unity, x would approach infinity as x hence to approaches zero, preserve the finite value of the lefthand terms we must have
less
a were
^ Aa +
and
^B
= ARb
(25)
Âa + ^B
Eliminating
(26)
and
between these two equations
R T
b
COMMUTATION
But
is
267
it
to be avoided,
was pointed out above that if infinite current density a must be greater than unity, hence
(27)
is
2
the condition to be satisfied.
In this inequality the term
j^f
takes the place of

of a single coil.
in the discussion covering the case
Proceeding in a similar manner, Carter has
shown that
in general
>i
where
Li Miz
(28)
Mi
L2
23
M
n3
Mi n
2n
M
M:
(29)
3n
When more than two coils are simultaneously short-circuited, the general differential equation takes the form
L
-jj
+ 2M X
is
-jj-
ohmic
drops
-f-
commutating
e.m.f.
where
dt
If
the rate of change of current in one of the

di
all
coils, x.
the commutation were linear in

(
the
coils,
-r.
would be con-
stant and equal to

(L
2i /T),
hence
+ SM
X)
-r
ohmic drops
commutating
e.m.f.
(30)
In other words, when a coil undergoes commutation in the presence of others that are simultaneously short-circuited, the effect is the same as though its self-inductance had been increased.
268
tures.
155. Calculation of the Self -inductance, L, in Slotted ArmaThe self-inductance of a coil has been shown to be equal
.
In to the number of flux linkages per ampere, divided by 10 8 the case of an armature coil embedded in a slot, the self-excited
may be separated into three parts: the slot from wall to wall of the teeth crossing and completing its path through the core, as indicated by
flux linking with the coil
1.
The
flux
<
Fig. 238.
2.
The The
between pole
3.
flux passing from tip to tip of the teeth within the space tips, as indicated by 3
linking with the end connections beyond the
edges of the core.
(">
FIG. 238.
^^
Paths of leakage
flux
(6)
coil.
surrounding
The number of linkages due to each of these fluxes will now be computed separately. 1. SLOT LEAKAGE FLUX. Practically without exception the windings of all direct-current machines are arranged in two
layers, in the
one side of each
coil
being in the top layer, the other side
The magnitude and distribution of the flux is therefore not the same on the two sides of a coil. a. Coil Edge Occupying the Bottom of a Slot, Fig. 239. The coil edge contains z = Z/2S conductors, whose total m.m.f. per unit
bottom
layer.
4ir
current
is
^.z gilberts.
The m.m.f.
acting
upon an elementary
tube dx
will
then be
4?r
COMMUTATION
assuming that the lines of force pass straight across the lux produced in this elementary path is then
slot.
269
The
47T^
bs
Wh?
I'dx
where
sions
I'
is
being in
the corrected length of the armature core, all dimencentimeters. The denominator of the above
expression represents the reluctance of the air part of the path, that of the iron part being negligible in comparison. This flux
inks with j-z conductors, hence the
It IS
number
of linkages
due to
and th
ries
by
dL'
.'.
270
In the same way the inductances due to the flux in the region A 3 and & 4 are
X
and
10- 8
= ~;
(33
<
* 7r
7'
60
(34
is
The
total inductance
due to
slot
leakage
L lb
L' 16
?-+.^rr + rJ
(a)
(35.
In the case of straight

this reduces to
slots,
Fig. 240a
(SG;
and with the shape comes

(6)
of slot Fig. 2406
it
be-
FIG. 240.
Coil occupying straight slots.
/Q7
tin
arc
6. Coil Edge Occupying the Top of a Slot, Fig. 241. Using same methods as in case a, the resulting expressions for L\
t
identical with those for Lib except that h% is replaced by h'z The total inductance of the coil due to slot leakage is then
T T
I
fQS
or less than twice Lib.

2.
force
6
,
TOOTH TIP LEAKAGE FLUX. Assume from tip to tip of teeth are made up of
that the lines o
a straight portion
242.
and of two quadrants of an elementary path dx is
circles, as in Fig.
The
flux
47T
I'dx
COMMUTATION
hence the total inductance for both sides of the
/"W(r-
271
coil is
dx
bo
'J JQ
+ TTX
l'
10 9
T.46
Iogio[l
+ ^^-
(39)
b) represents the distance between pole expression (r limit of the integral is taken to be half of and the superior tips, this amount on the assumption that the coils undergoing commutation are approximately midway between pole tips. Values of L 2 calculated by the above equation will be somewhat too
The
large because no account has been taken of the effect of neigh-
boring slot openings in reducing the
flux.
272
Arnold 1 gives
L =
3
zHf zHf
[o.46
logic
.(^p)
0.092J
10~ 8
[0.46
logio
(j~)
0.23]
10- 8
(41)
where ds is the diameter of a circle whose circumference equals the perimeter of the coil section including the insulation between turns (see Fig. 243); i.e.,
di
nab)
Hobart calculates the end-connection leakage on the basis of a flux of 0.8 lines per ampere-conductor per cm. of free length.
a-
ULi.
FIG. 243.
Cross section
FIG. 244.
of coil.
Relative slot dimensions, particular case.
The total inductance of a winding element is the sum of the inductances due to the several parts of the leakage, or
L =
Li
+L +L
2
i O..J
(42)
Parshall and Hobart have published 2 the results of measurements of the inductance of armature coils of commercial machines
from which they deduce

the coils
1
that,
is
at the rate of
on the average, the flux linked with 4 c.g.s. lines per ampere-conductor
I,

Electric Generators, 1900.
p. 376,
2nd
ed.
COMMUTATION
per cm. and 0.8
of
273
"embedded" length
lines per
of
wire (10 lines per inch),
ampere-conductor per cm. of "free" These values check fairly well with the length (2 lines per inch). results of the foregoing formulas when customary dimensions are inserted. Thus, consider a machine with straight open slots whose
c.g.s.
ratio of
depth to width
is
1 (Fig.
244) and in which
T
#0
is
approximately 10, which space between poles. Taking z = I and V =
is
equivalent to about five slots in the

cm.,
Lu=or
(M
+ M) X1010~ 8
z
8
,
an average
of
(% + 1%) X
coil,
= = =
1
3.05
10~ 8
The
value of
for
one side of the

lo glo [1
with
and V =
1, is
^ X 0.73
or a total of 4.16
15.7]
10- 8
1.11
10~ 8
10~ 8 henries, corresponding to 4.16 lines per ampere-conductor per cm. of length. Parshall and Hobart's method is rapid and simple, but it is open to the objection that the designer must exercise great discretion in selecting the proper unit value of flux to fit the dimensions of his machine.
156. Calculation of the Mutual Inductance, M. The mutual inductance of two coils is equal to the number of flux linkages with one of them when a current of 1 ampere flows through the other, divided by 10 8 The previous discussion of the simultaneous short-circuiting of several coils indicates that there are two cases to be considered one in which the coils in question occupy the same slot, the other
. :
in
1.
which they lie in different slots. COILS OCCUPYING THE SAME SLOT.
the coil edges
lie
In this case two

lie
dis-
tinct conditions are possible: (a) the coil edges

(b)
side
by
side;
one above the other, or one in the top layer,

lie
the other in the bottom layer. (a) It is clear that if the coil edges
18
side
by
side in the slots,
274

also throughout their entire lengths, there
is
and therefore
no
great error in writing
= L
(6) Since the coil edges are in different layers, the end connections run in opposite directions, hence the mutual inductance is due only to the slot and tooth-tip fluxes.
coil.
In Fig. 245 the cross-hatched areas represent the two sides of On the left-hand side the inducing coil is at the bottom
of the slot
and
4ir
exerts a m.m.f. of
on ^.z gilberts per ampere
the elementary tube dx] the slot flux linkages are
FIG. 245.
Flux leakage paths, mutual inductance.
On
the right-hand side, the slot flux linkages are
10
same as on the other side of the coil. The linkages due to the tooth-tip leakage flux are obviously the same as in the calculation of L. Finally, therefore,
or the
-}
2
2.
(43)
COILS OCCUPYING DIFFERENT SLOTS. (a) Both Coils in In this case the sides of the coils will be parallel Layer. to each other throughout their entire lengths, and the interthe
Same
COMMUTATION
275
linked flux will consist of tooth-tip leakage flux along the embedded portion and end-connection flux along- the free lengths. Considering the tooth-tip leakage first, coil edge 1, Fig. 246,
acts
4?r
Yp:
upon paths surrounding

per ampere.
coil
edge 2 with a m.m.f. of
z gilberts
In the tube dx the flux linkages are
then
47T
To*
op
I'dx
and the mutual inductance due to

coil is
this flux
on both sides of the
FIG. 246.
Mutual inductance,
in adjacent slots.
coils
FIG. 247.
coils
Mutual inductance, not in adjacent slots.
-b
-I
X)
(r
"
(44)
10 9
is somewhat too large since in carrying out the integration continuously to the pole tips the effect of the slot openings
This value
is
ignored.
cult to estimate.
The mutual inductance due to end-connection leakage is diffiArnold recommends taking it as one-half of
coil.
the self-inductance of a single
On
this basis
276

-
M"vt = zHf fo.2

L
logio ( ) \ S /
0.05] J
10~ 8
(45)
and
If 12
= M' l2
M"i2
slots,
(46)
When the coils considered are not in adjoining placed as in Fig. 247, the above equations become
but are
M'i 3
L
o
zH
60
+ 7r(2* +
1U
zH'
-
X)
^
and
7T
logio
jr~~T~o~T~~~ OQ -f- ZlTl
Mu"
= K^3 =
z 2 Z/ [o.l logio ( ) L \ s I
0.025] J
10~ 8
(48)
It is not necessary to carry the computation beyond the case shown in Fig. 247 for the reason that the numerical values become relatively small and the brushes are seldom so wide that coils are simultaneously short-circuited in more than three
consecutive
(6)
slots.
Coils not in the
Same
3t
ductances
M' 12 and M'i
In this case the mutual inLayer. due to tooth-tip leakage, remain the
as before; the end-connection leakage reduces to zero because the coils separate and run in opposite directions after Then leaving the slots.
6
same
M
arid
M'
+ ^ (T "
6)
zH'
lo
(50)
,, M
ls
,,, = M' = 13
^
is
4?r
zH'
4.6,
logio
2
b~~+2-jrt
(51)
These values are somewhat too large inasmuch as the

the slot openings has been neglected.
157.
effect of
The Commutating E.M.F.

7?
It
has been pointed out that

a sufficient criterion of
T7
the inequality
~rf~.
>1
not in
itself
COMMUTATION
277
sparkless commutation, because it is bound up with the relation E T <io(R 2/4), the latter condition holding when the brushes
In are in the neutral zone or only slightly displaced therefrom. other words, the above relation is of importance in the case of
resistance commutation.
The same thing
is
r>
therefore true in
regard to the relations which follow from
them ^ >1, among
the relation emax < 2. If, however, the brushes are given a decided displacement from the neutral so that E T > i Q (R 2Ri), it is possible to have the
7?
T7
condition
j L
<
1.
This last relation, taken alone, would lead
to the conclusion that large values of L and very small values of Rb and T are permissible, but this is naturally incorrect. Obviously,
therefore,
some other
relation
must
exist
which de(5),
termines the limiting conditions. An examination of the fundamental differential equation
L^ + iR +
-r
:
di
RbT y
..
(I Q
j
RbT
i)
..
(i
^)
+E =
rn
c
t\
shows that there are
in general three e.m.fs.
concerned in the com;
mutation process the reactance voltage,

e.m.f.,
L -r.
the commutating
',
latter are generally small in

larly
and the ohmic drops at the transition surfaces. The comparison with the others, particuthere
is
when
considerable brush lead, hence the conclusion
that the commutating e.m.f. should be at all times nearly equal and opposite to the reactance voltage. This is the condition
connection with while it is appears, desirable to keep the average reactance voltage small, it is even more important to keep the commutating e.m.f. within limits. For instance, consider a machine which is to operate with
that has already been discussed in Chap.
V in
armature
reaction.
It
therefore,
that
brushes fixed at a definite angle of lead at no-load the commutating field will have a certain value and will generate an e.m.f. in the short-circuited coils; this e.m.f. will set up currents in
;
these coils and sparking may result if the e.m.f. so generated is the sufficiently high to strike an arc between the brush and commutator segments as they break contact. It follows, therefore, that the
brush must be so placed that at no-load the com-
278

e.m.f. will be within the
sparking limit. When the operating under load conditions, the reactance voltage set up in the short-circuited coils will be opposite in direction to the commutating e.m.f., and the difference between
mutating machine
is
them must again be within the sparking limit. The field strength in the commutating zone
resultant
of
the m.m.fs.
of
the
which act in that region. If, commutating field would also be the resultant of the individual -fields produced by the field winding and by the armature
is
is due to the and armature windings there were no saturation the
field
winding acting separately. In actual calculations this last fact made use of because of its greater simplicity, and the disturbing effects of saturation must then be allowed for according to the
judgment
lar
of
his experience
the designer and with the particu-
sideration.
type of machine under con-The curves of flux
Armatur
Field
FIG. 248. Commutating field distribution.
and armature windings can be determined by the methods described in Chap. V, and in the commutating zone will have the forms indicated in Fig. 248, which represents the case of a generator with a forward lead of the brushes. It will be seen
distribution due to the field
that the strength of the commutating field decreases with increasing load, if the field excitation remains constant, the change being unfortunately in the wrong
direction inasmuch as the commutating e.m.f. should increase with the load. If the armature is magnetically too powerful, or if the brushes are not given a sufficient lead, the commutating e.m.f. may even reverse under load, assuming, as in all the pre-
ceding discussions, devices present.
that
there
are
no
special
commutating
In constant-speed generators and motors of the separately excited or shunt type, the commutating field decreases gradually
with increasing load, because of armature reaction, in the manner indicated in Fig. 248. The commutating e.m.f., therefore, de-
COMMUTATION
279
creases proportionately, so that at the maximum load (generally taken as 25 per cent, overload) it is appreciably less than at no-load it follows, therefore, that the reactance voltage at maximum load must be considerably less than double the sparking limit in order that the difference between it and the commutating
;
may remain within the sparking limitÎn over-compounded generators^ on the other hand, the commutating e.m.f.
e.m.f.
increases with increasing load, so that in machines of this type the reactance voltage at the maximum load may be more than
double the sparking
limit,
without introducing
difficulties.
In
either case/af the excess of the reactance voltage over the commutating e.m.f. at maximum (125 per cent.) load is equal to the
commutating
e.m.f.
at
no
load,
and the
\
latter
is
within the
sparking limit, the two e.m.fs. will exactly neutralize each other at ap-
"
|
proximately five-eighths of full load, and the commutation will then be

satisfactory.
/'\
"J
The above
discussion explains the

of
FIG.
249.
,
Series
circuit
coils.
through short-circuited
superior commutating properties
over-compounded or series-wound generators/^ Furthermore, if the armature cross-magnetizing effect in the commutating zone is neutralized, and a commutating field is set up such that it increases proportionally with the load, the reactance voltage can be exactly balanced at all loads and ideal commutating conditions will result. These effects can be secured by the use of interpoles or
commutating poles, which are placed midway between the main poles and excited by the armature current (see Chap. IX). In the case of a wide brush short-circuiting several coils, the
e.m.fs. evidently act in series through the brush. Thus, in Fig. 249, the brush short-circuits coils 1, 2 and 3 simultaneously, and the sum of the e.m.fs. generated in them by the commutating field acts through the brush as indicated by the dashed line. At no-load this sum must not exceed the sparking limit. Referring to the symbols used in Fig. 248, the magnitude of the commutating e.m.f. at no load, per element, is
commutating
E co
ttl'vB*
10- 8
(52)
280
where
B
is
is
Z/2S
the
the average strength of field in the commutating zone, number of turns per element, and v is the peripheral
This assumes that the coils are of full velocity in cm. /sec. pitch, so that both sides are at the same time cutting fields of equal intensity. In the case of fractional pitch windings it is
necessary to add algebraically the unequal e.m.fs. generated in each side of the coil. Under load conditions the commutating
e.m.f. per
element becomes
E = jg
cl
I'v
(B
Bo)
10- 8
windings.
(53)
If
with
similar
is
reservations
concerning
coils
chorded
now nac
the
by the brush,
number of we have the

17
simultaneously short-circuited relation
2E CO = n -^ Zo
sc
I'v
B X
10- 8
<
12 to 18 volts
(54)
It
is
evident from these considerations that the value of
BO, the flux density in the commutating zone, must be sufficiently under control to allow the brushes to be placed in a
position which will insure good
commutation under
all loads.
To
strength in the neighborhood of the commutating pole tip should shade off gradually instead of abruptly, a condition which can be realized fairly well by making the airthis
end the
field
gap at the pole tips longer than it is under the central part of the pole shoes. 158. Pulsations of Commutating Field. During the period of commutation the rotation of the armature periodically
changes the positions of the teeth and slots with respect to the pole shoes, thereby giving rise to peripheral oscillations of the armature flux in the interpolar space. The changing current in the short-circuited coils produces a further pulsation of the
flux in this region.
There
may
also
be pulsations of the flux
as a whole, due to periodic changes in the reluctance of the magnetic circuit if the surface of the teeth presented to the
poles does not remain constant.
All of these effects are of
high frequency and induce
in the short-circuited coils e.m.fs.
of rapidly changing direction which are superposed upon the main e.m.fs. considered in the preceding sections. They give
COMMUTATION
rise
281
to saw-tooth notches in the short-circuit current curves. Obviously these pulsations can be reduced by using numerous small slots with few coil edges per slot.
159. Sparking Constants.
er
The
expression
= -~ L
which has been shown to be of considerable importance can be put into another form involving the principal design constants of the machine, and may therefore be used as a check upon the
magnitudes of the constants selected before the design has proceeded too far. Thus, it has been shown that
b
rp
+
Vc
and
L
where F is a function have also
=^~
zH
XF
of the
dimensions of the machine.

Z^
We
=
and
where
v
d
dcom
= = =
=
peripheral velocity of armature
diameter of armature diameter of commutator.
Also,
~
ampere-conductors per cm. of periphery.

er ,
Substituting these values in the expression for

er
we have
= C
I'vq
where
27T/3F
&.
10- 9
+ /m -
282

r7
The quantity
-~
Vvq
may
be considered as characteristic of corn-
mutating conditions; it has an average value, using metric units, about 20 X 10 6 In English units (7 in inches, v in feet per and in minute, q ampere-conductors per inch of periphery) it becomes approximately 40 X 10 6 160. Reaction of Short-circuit Current upon Main Field. Let a and b, Fig. 250, represent the initial and final positions of
of
.
undergoing commutation in a generator. In the a posicoil clearly exerts a demagnetizing magnetomotive force upon the main magnetic circuit, while in the b position the action is a magnetizing one. If commutation takes place, on
a
coil
tion
the
the average, in the geometrical neutral, and if the short-circuit current curve is symmetrical with
respect to the point O (Fig. 218), the demagnetizing and magnetizif
ing effects annul each other. But the mid-point of the commutation period occurs when the coil is in the geometrical neutral, and
FIG.
250,-Magnetizing action
the short-circuit current curve
is
of short-circuited elements.
not symmetrical, one or the other of the two effects will preponderate.
Thus, in the case of overcommutation, the current will have the direction of the b position during the greater part of the time of short-circuit, hence there will be a resultant magnetizing action.
In case of undercommutation the resultant action
will
be demag-
netizing. Obviously, these statements are to be reversed in the case of a motor.
When
set
a generator
is
running without load, currents
will
be
by the brushes. The direction of the current flow will depend upon the direction of the field in which the coils are moving, and, therefore, upon the direction of displacement of the brushes. With a forward lead
up
in the coils short-circuited
of the brushes the current in the short-circuited coils will
have
the direction
in the b position, Fig. 250, and will then exert a magnetizing effect; a backward lead of the brushes will evidently result in a demagnetizing action. The magnitude of
shown
these no-load short-circuit currents
may
be sufficiently great
COMMUTATION
283
to materially influence the field flux, hence also the experimentally determined open-circuit characteristic.
The distorted air-gap armature current can action of the flux due to the magnetizing of the fluxes which would be of as be thought compounded m.m.fs. armature the field and acting separately. produced by
161.
The Armature Flux Theory.
Strictly speaking, this is not exactly true when saturation of the iron part of the magnetic circuits exists, but it will serve
as a first approximation to the truth. If, then, the brushes takes set so that commutation are place in the geometrical will not be acted upon by the coils the short-circuited neutral,
field
component
of the flux,
but
the armature flux, being stationin
ary in space, will generate e.m.fs. them in exactly the same way that the main flux acts upon the
conductors under the poles. The state of affairs in a generator will be as shown in Fig. 251. This
figure brings out the fact that the e.m.f. generated by the arma-
FIG. 251.
Production of armature flux.
ture
flux
in
the short-circuited
coil is in the same direction as the original, current flow and therefore tends to prevent the desired reversal of current. Evidently this e.m.f. must be
neutralized
by an
e.m.f. of opposite
direction
and somewhat
sufficient sur-
greater magnitude plus e.m.f. to effect the reversal.
in order that there
may
be a
This reversing e.m.f. may be obtained by shifting the brushes until commutation takes place in a sufficiently intense part of the field flux; or the brushes may be kept in the neutral axis and the armature m.m.f. wiped out by the opposing m.m.f. of an interpole or its
equivalent (see Chap. IX). A full treatment of the subject of commutation based on this 1 It has the adanalysis has been worked out by B. G. Lamme.
vantage that
in the
1
it
commutation
of
emphasizes the physical phenomena involved It does not involve the explicit process.
Its Application to Interpole
Theory
Commutation and
Machines,
Trans. A.I.E.E., Vol.
XXX,
Part
3,
1911, p. 2359.
284

self-
consideration of the e.m.fs. of
and mutual-induction, though
these are implicitly involved in the e.m.fs. generated by cutting For example, in the paper rethe stationary armature flux.
ferred to there occurs the following statement:
"According to the usual theory, during the commutation of the coil the local magnetic flux due to the coil is assumed to be reversed. However, in the zone in which the commutation occurs, certain of the
magnetic fluxes may remain practically constant in value and direction during the entire period of commutation. This is but one instance, of which there are several, to show where there is apparent contradiction of fact in the usual mathematical assumptions made in treating
this
problem."
reversal of the self-excited flux which
is responsible for the a mere mathematical however, abstraction in the inductance theory; for while this flux does reverse with respect to the coil, it does not reverse its direction in
The
e.m.f. of self-induction is not,
space as will be clear from Fig. 250, which shows a short-circuited coil in two successive positions, before and after the reversal of the current. This is a consequence of the motion of the coil.
If,
however, this self-excited flux
e.m.f. of self-induction, it should
tion of the
commutating
which
is
flux;
is considered as producing the be eliminated in the computain other words, the latter is then
to be taken as the resultant of the field flux
armature
flux
and that part due to the armature turns outside
of the
of the
commutating zone.
PROBLEMS
What is the ratio of the energy losses due to brush contact resistance two machines, which are identical in all respects except that in one of them the commutation is linear and in the other sinusoidal, as indicated in curves c and d of Fig. 218? NOTE. This problem may be solved most simply by plotting a curve showing the variation of contact loss throughout the period of commutation. The energy loss will then be proportional to the area under the curve. 2. Construct a curve showing the variation of current density at a commutator segment, and curves showing the variation of current density at five points equally spaced along the brush arc, for the case of a machine which has a brush covering 2^ commutator segments and in which the
1.
in
short-circuit current curves are sinusoidal.

3. The armature of a 6-pole machine has a simplex lap winding of 546 conductors arranged in 91 slots; the commutator has 273 segments and has a diameter of 42 cm., and the brush contact arc is 1.2 cm. Construct dia-
COMMUTATION
285
grams showing the sequence of commutation in the coils occupying two adjacent slots, (a) assuming full-pitch windings, (6) assuming a pitch that is less than full pitch by one slot. 4. The armature of a 10-pole machine has a quintuplex wave winding of 1230 conductors arranged in 205 slots; the commutator has 615 segments and a diameter of 131 cm.; the brush contact arc is 2 cm. Construct diagrams showing the sequence of commutation in the coils occupying two adjacent slots.
5. The machine of Problem 3 runs at a speed of 430 r.p.m. and that of Problem 4 runs at 94 r.p.m. Find the duration of the period of short-
circuit in
6.
each case.
of the armature of Problem 3 is 59.1 cm., the total length armature core is 32 cm., and the slots are 1 cm. wide by 2.9 cm. deep; each conductor has a cross-section measuring 0.18 by 1.2 cm. -The thickness of the insulation at the bottom of the slot is 0.1 cm., and the thickness of the insulation between the top and bottom layers is 0.12 cm. Assuming that the total length of the conductors of an element is twice the length of core plus three times the pole-pitch, and that the pole arc is 70 per cent, of the pole pitch, find the total inductance of an element, taking into account the mutual induction due to simultaneously short-circuited elements. 7. An armature has semi-closed circular slots of radius r cm., as shown in of
slot is ri
The diameter
the right-hand diagram of Fig. 52, p. 64. The opening at the top of the cm. wide and r 2 cm. deep, the center of the circular slot being If the slot contains z conductors, dis(r r-i) cm. below the periphery.
tributed uniformly over the cross-section of the circle, what is the inductance due to slot leakage flux, assuming that the lines of force pass Straight
across the slot (horizontally in Fig. 52)?
CHAPTER IX
COMPENSATION OF ARMATURE REACTION AND IMPROVEMENT OF COMMUTATION
The cross or transverse 162. Principle of Compensation. is the primary cause armature current of the action magnetizing of the field distortion which in turn necessitates the shifting
and thereby brings into existence the demagnetizthe armature. Clearly, then, if the transverse of action ing magnetomotive force of the armature were balanced by an equal
of the brushes
in space, the distortion of the field
and opposite magnetomotive force having the same distribution would be completely eliminated and brush displacement would be unnecessary except, possibly, to assist in commutation. Moreover, if the armature magnetomotive force is overcompensated, there will exist in the neutral
zone a component of flux having the proper direction to reverse coils, and the brushes could then be permanently fixed in the geometrical neutral axis. If the ratio of compensating ampere-turns to armature ampereturns is unity, nearly complete neutralization of the armature flux will result; if the ratio is slightly greater than unity, there will exist in the commutating zone a reversing flux which increases proportionally with the armature current (unless saturation of the iron of the magnetic circuit sets in) which is precisely the
the current in the short-circuited
condition necessary to secure good commutation at all loads. In either case, whether the above ratio is unity or greater than unity, the compensating winding must be traversed by the main
armature current or a fractional part thereof; consequently the compensating winding is connected in series with the armature, and may or may not be provided with a diverting shunt, as in the case of the series winding of a compound machine. The problem of compensating armature reaction then consists of two parts; one, the prevention of the field distortion and the consequent elimination of the demagnetizing action of the arma286
COMPENSATION OF ARMATURE REACTION
287
ture; the other, the production of a commutating e.m.f. for the purpose of neutralizing the reactance voltage of the short-cir-
the latter
and reversing the current in them. Of these two, the more important. In a German patent granted to 163. Compensating Devices. first is the there in 1884 exposition of the principle of Menges
cuited coils
is
compensating armature reaction.
The patent
specifications call
for the use of a stationary compensating winding wound around the armature at the sides of the poles and traversed by the arma-
ture current, or a part of it, in such a direction as to oppose the magnetizing action of the armature. Later, in 1892, H. J. Ryan and M. E. Thompson experimented extensively along similar
lines
and developed the method shown diagrammatically in 252 for the case of a bipolar Fig. The compensating windgenerator.
is
ing
faces.
embedded in slots in the pole The space distribution of the
compensating winding being very nearly the same as that of the armature winding, the neutralization of
N
tl
__
Jrp^
armature
complete.
in Fig.
reaction
is
practically
Since in the case
shown
FIG. 252.
252 the current in the armature conductors is half of the total

current, the
Diagram
of
com-
pensating winding.
number
of turns in the
compensating winding should be half of the effective number of cross-magnetizing armature turns. In larger multipolar machines the number of compensating turns per pole should be I/a times the armature turns per pole, where a is the number of current paths through the armature. Fig. 253 illustrates the construction of the field frame of a
Company.
type as built by the Ridgway Dynamo and Engine The entire magnetic circuit is built up of sheet steel stampings, the main ring or yoke being clamped between cast-
machine
of this
iron frames; the main pole pieces that carry the coils of the shunt winding are bolted to the yoke, and the cores that carry the com-
pensating winding are held by the bolts which pass through the main poles and by the wedges which hold the commutating
lugs in position.
These wedges also serve to reduce the cross-
288

from pole to
pole,
section of the magnetic path
and so keep
down magnetic
leakage.
FIG. 253.
Frame
of
Ridgway
generator, showing slotted pole face.
circuit of a
Fig. 254 illustrates diagrammatically a portion of the magnetic machine embodying the above device. Under load
FIG. 254.
Flux paths in poles of compensated machine.
conditions the magnetomotive force of the compensating winding acts in the directions shown by the dotted lines, assuming generator action and clockwise rotation. Section q of the bridge

is
289
acted upon by two m.m.fs. having the same direction, but the not materially increased on account of the initial saturation of the iron; and section p is acted upon by two m.m.fs.
flux is
The result is that the central tooth t is of opposite direction. acted upon by a resultant m.m.f. which makes it a north pole
under the assumed conditions, hence
it
produces a local
field of
FIG. 255.
Frame
of
Ridgway
generator, coils in place.
the proper direction to assist the reversal of the current in the short-circuited coils. This effect is accentuated by making slots
and n somewhat larger than the others (see also Fig. 253) and winding in them more than the normal number of conductors. The arrangement of the compensating winding and main field winding is shown in Fig. 255. Closely akin to the Thompson-Ryan device is an arrangement
19
290
due to Deri. Instead of using a field frame of the salient pole type with the addition of a slotted ring, the field structure consists of a slotted cylinder wound with two independent sets of
coils
in Fig. 256.
and concentrically surrounding the armature, as indicated The main winding, M, produces poles whose axes
are indicated
by the arrows; the compensating winding,
C, sets
FIG. 256.
Deri's arrangement of
main and compensating windings.
up a magnetomotive force acting along axes midway between the poles, and in opposition to the armature m.m.f. The field structure closely resembles the stator of an induction motor,
and since the reluctance
all
of the magnetic path is the same along diametral paths, the compensation can be made complete. 164. Commutating Devices. The entire prevention of field
is
distortion
not necessary for successful commutation.
The
FIG. 257.
Bayers' winding.
principal consideration
ing
coil
field of sufficient
.
to insure the presence of a commutatintensity to generate in the short-circuited

is
an e.m.f large enough to neutralize the reactance voltage. Thus, in the Sayers winding, Fig. 257, there is no compensation of armature reaction, but the reversing e.m.f. is introduced into the short-circuited coil by an auxiliary winding which is
so placed as to cut a part of the
main
flux during the
commuta-

tion period
coil
291
and which, during that undergoing commutation. At

not in
circuit.
interval, is in series with the

all
other times the auxiliary
The auxiliary coils are in reality merely extensions of the commutator leads which have been wound around the armature. The main coils are connected to auxiliary coils which lie behind them with respect to the direction of rotacoil is
tion so that
commutation
is
not dependent upon the flux density
at the leading pole tip, as in the ordinary machine, but upon that at the trailing pole tip; and since the field intensity at the latter increases with increasing current, the commutating e.m.f. increases
with
it.
limit is set to this automatic
adjustment of commu-
The tating conditions by the saturation of the trailing pole tip. is device of historical rather than as is Sayers practical interest,
also that of Swinburne. l
magnets, Fig. 258, excited
In the latter, small U-shaped electroby the main current, are placed in the
FIG. 258.
Swinburne's com-
mutating device.
FIG. 259. High resistance leads to commutator.
Another method that has been used for the prevention of commutation difficulties consists of the insertion in the circuit of the short-circuited coil of an auxiliary resistance
neutral zone.
must be handled
which serves to limit the current and, therefore, the energy that The simplest method at the brush contact.
that suggests itself for this purpose is the use of commutator leads of high specific resistance, as indicated in Fig. 259; this
arrangement
motors.
It
used in certain types of alternating-current however, open to the serious objection that the main current must flow through a set of these extra resistors with consequent heating and loss of efficiency. Instead of inserting resistance in the circuit in the above
is
is,
manner, the brushes may be so constructed as to interpose more resistance in the path of the short-circuit current than in that of
1
Journal Inst. of Electrical Engineers, 1890, p. 106.
292

by making them
of alternate longitudinal layers
the main current,

of
carbon and copper. The copper provides a path of high conductance so far as the main current is concerned, while the
short-circuit current
resistance of the carbon
must pass transversely through the higher and copper in series. It is possible,
however, that brushes of this type may fail to operate satisfactorily for the reason that the short-circuit current may pass from one copper layer to the next by way of the commutator A surface instead of through the intermediate layer of carbon. better design, due to Young and Dunn, 1 provides for the final rupture of the short-circuit at an auxiliary carbon brush insulated from the main brush, in the manner indicated in Fig. 260. This has the effect of considerably increasing the resistance of the
circuit at the last stage of the
commutation process. At the present time, however, the

is
device that
exclusion
|
of
all
used to the practical others, wherever

conditions
is
commutating
ITT
T.
FIG. 260.
Special brush with
special difficulties,
present the intervale or

.
auxiliary insulated section,
commutating pole, the principal features of which are discussed later.
structural arrangement of the interpole machine is shown in The magnetomotive force of the interpoles Fig. 65, Chap. II. is so adjusted that the m.m.f. of the armature is slightly over-
The
compensated. The final adjustment for sparkless operation made by varying the air-gap under the interpoles by means
shims, or
165.
It goes
is
of
by means Commutation
of a shunt
around the interpole winding.
in
Machines having no Auxiliary Devices.
without saying that the mechanical construction of the
commutator and brush rigging must be such as to insure perfect contact and absence of vibration. The natural period of vibration of the brushes and brush holders should differ from that of any vibration which may possibly be impressed upon them by the
motion of the commutator, in order that mechanical resonance may not occur. The copper of the commutator should be of uniform quality so that a true cylindrical surface may be preserved, and the mica insulation between segments should have a rate of wear
as nearly as possible the
1
same
as that of the copper.
Electrical World, 1905, p. 481.

Reference has already been
the
293
made
in the preceding chapter to
more important magnetic and electrical factors concerned in the commutation process, to which due regard must be paid to For the sake of completeness they insure sparkless operation. are here recapitulated and in addition there is given a summary of the methods adopted to secure them. 1. The summation of the commutating e.m.fs. generated in the coils under a brush must not be greater than 12 to 18 volts, or
S#c0 =
This equation
is
n sc
-^rvB Q
10~ 8
<
12 to 18 volts.
based upon the assumption of full-pitch windings; windings the two sides of a coil in fields of unequal strength, somecommutation are undergoing times of the same polarity, so that the e.m.fs. generated in them may tend to annul each other. This results in a decrease of the short-circuit e.m.f. At the same time there is a decrease in the
in the case of fractional pitch
demagnetizing action of the armature. Both effects indicate that chording the windings is advantageous, but a distinct limit is set by the fact that in such windings at least one side of the shortcircuited coil will lie close to a pole tip where the field intensity changes sharply; this feature greatly restricts the zone through which the brushes may be rocked, and is incompatible with the
requirement of a fringing field of gradual slope. 2. The average reactance voltage
er
=
~7p
L<
as
2i Q
Rb
should
T~)
be
less
than
volt,
determined by the relation

is
fjl
-j LI
> 1.
The
limiting value of e r
,
set
by the voltage drop
at
The the brush contact, 2i Q Rb which is of the order of 1 volt. harder the grade of carbon used in the brushes the greater will be
the drop.
resistance
Consequently hard carbon should be used where commutation is necessary. Fig. 261 (taken from Gray's Electrical Machine Design) shows the variation of brush
The contact drop with current density for an average carbon. tendency toward constant drop with increasing current density is obvious. The contact drop in the direction from commutator
294
to brush is generally somewhat higher than that in the direction from brush to commutator. 3. The brush width should not be too great, in order to cut down the mutual induction of the simultaneously short-circuited coils. In practice, the brush width seldom exceeds 3.5 times the width of a commutator segment. To secure sufficient contact
area to carry the current the axial length of the brushes is adjusted so that the average current density is in the neighbor-
hood
of 30 amperes per sq. in. (5 amperes per sq. cm.) for hard carbons and as high as 65 amperes per sq. in. (10 amperes per
i.o
0-8
0.6
0.2
10
20
30
40
50
60
70
80
90 100
Amperes per Sq.Inch
FIG. 261.
Relation between brush drop and current density, average carbon.
average of about 40 amperes per customary. Considerably higher densities are permissible with metal (copper) brushes, up to about 160 amperes per
sq.
cm.) for soft carbons.
An
sq. in. is
sq. in.
Since the
number
of simultaneously short-circuited
coils is
given by
sc
_b
p @' a
practice at from 2 to 3.5,
if
and since the ratioit
is
fixed
by usual
is
follows that p/a
must be kept small

ratio
encountered.
The
p/a
maximum
special difficulties are in series (two-
circuit) windings,
remedy
is
hence if SEco turns out to be too large the a series-parallel or full-parallel winding.

4.
295
The inductance of the coils can be kept within limits by number of turns per element; or conversely, for a given number of armature conductors, by increasing the number The value of L can also be made of commutator segments. small by selecting a relatively short axial length of armature.
reducing the
Both of these considerations point to the desirability of a design which the ratio of diameter to length of armature core is A large number of commutator segments relatively large. involves a moderate average value of volts per segment,
in
77!
-T-
--
= ^~ which
t
value, in simplex lap windings, should not

less
than 15 volts, if possible. The between adjacent segments in machines of the series types should never exceed 40 volts, otherwise there is the possibility that the machine will "flash over."
exceed 20 volts and should be
maximum
difference of potential
selection of a large diameter carries with it the possibility number of slots each containing but few coil sides, thereby reducing the mutually inductive action. The presence
of using a large
The
numerous armature slots means further that there will be a number of them between pole tips to cut down oscillations of the commutating flux and minimize pulsations of the
of
sufficient
flux as a whole.
5.
It
was shown
it is
reversal of the
Chap. V that in order to prevent the commutating field by the distorting action of the
in
armature
necessary to observe the relation
armature ampere-turns per pole
<
1.1 field
ampere-turns per pole,
though in most cases the factor is 0.8 to 0.9 instead of 1.1. " In other words, the main field should be powerful, or stiff," in comparison with the armature field. Since the armature magnetomotive force acts upon a different path from that subjected to the main field excitation, it is clear that the disturbing effect of armature reaction can be reduced by introducing
additional reluctance into the transverse magnetic circuit of the armature. If at the same time those parts of the magnetic
circuit
which are influenced by both the armature and
field
excitation are saturated at no-load, the load current will have little effect in producing distortion; the saturation acts, of course, as an increase of reluctance.
296

first of
is not as simple as it appears armature magnetomotive force acts upon a path which includes the most important part of the main magnetic circuit, namely the air-gap and the iron parts adjacent
The
the above methods
at first sight; the
FIG. 262.
FI.G.
263.
FIG. 264.
FIGS. 262, 263, 264.
Longitudinal slotting of pole cores.
thereto.
The
ture m.m.f. therefore adds
addition of reluctance to the path of the armamore or less to the reluctance of the
main
the
circuit,
cost
of
the
hence requires more field excitation and increases machine. Designs embodying this principle
FIG. 265.
FIG. 266.
FIGS. 265 and 266.
Chamfered and eccentric pole
faces.
in Figs. 262,
alone involve a longitudinal slotting of the pole cores, as shown 263 and 264. Their effectiveness is open to question
behind the
inasmuch as the armature flux will tend to pass around The most effective slots rather than across them.
FIG. 267.
FIG. 268.
FIGS. 267 and 268.
Construction of laminated pole.
designs are those which include both the features of additional reluctance and saturation of the iron; the extra reluctance
is
265,
usually obtained by chamfering the pole tip, as in Fig. or by making the bore of the pole faces eccentric,

as in Fig. 266.
297
at
The
saturation feature
is
most important
the trailing tip in the case of generators and at the leading tip in the case of motors; the desired saturation is obtained by using a long thin tip, or, in the case of laminated poles, by using
tions are built
a stamping of the form of Fig. 267, in which case the laminaup to the required thickness in such a manner
that the projecting tips are alternately on opposite sides; a plan view of the pole face as seen from the armature surface is
then like Fig. 268. It is, of course, not necessary that the entire pole be laminated to secure this construction, as the pole shoe alone may be built of stampings and bolted to a solid pole
core.
Fig. 269 illustrates the pole construction of a Lundell generator which embodies a number of these features; it is, however, only adapted to machines that run in one
direction.
Saturation of the armature teeth acts

in
the same
way
as
saturation
is
of the
The tooth density purposely made high, often

pole tips.
therefore
as high as 140,000 lines per sq. in. (21,000 per sq. This is particularly true in the cm.).
FIG. 269.
Pole of
Lundell generator.
of motors which are required to operate with the brushes fixed at the geometrical neutral, as in railway motors.
case
6. The requirement of a fringing commutating field of gradually changing intensity is met by properly shaping the tips of the pole shoes. The length of the air-gap at the pole tips can be com-
puted from the equation
r,
5'
(1.25 to 2) |_360a 1.6 E
- 2AT
as
166.
poles,
shown in Chap. V. Commutating Poles.
Commutating
poles,
or
inter-
have been extensively adopted for the improvement of commutation in generators and motors in which sparkless operation would be difficult or impossible of attainment under ordinary conditions. Examples of such machines are turbo-generators and shunt motors which are required to have a wide
298
range of speed, as discussed in Chaps. VI and VII. InterInterpoles are also extensively used in series railway motors. obviate the for the various poles necessity expedients commonly
employed
of
in ordinary machines. They are superior to forms construction involving compensating windings in the pole
faces because of their greater simplicity. Complete neutralization of the armature flux
is
not possible
by the use of interpoles for the reason that the space distributions of the m.m.fs. of armature and interpoles are different.
This
is
is
not objectionable, however, since the essential feature
the production, in the commutating zone only, of a reversing flux of proper intensity. The flux distribution outside of this
zone
is
of
minor importance.
of interpoles increases the magnetic leakage of the main poles, the calculation of the leakage factor being made in the manner outlined in Chap. IV. To reduce the leakage to a minimum, both the breadth and length of the interpoles
The presence
should be kept as small as possible, and the span of the main poles made smaller than in ordinary machines; the ratio of pole
arc to pole pitch
0.70.
is
The span
of the interpole should
usually between 0.60 and 0.65, instead of be equal to, or slightly
in
greater than, the distance moved over by a slot while the coils it are undergoing commutation. The axial length of the of can be made than that the main poles, for it is less interpole
immaterial in what portion of the coil the neutralizing e.m.f. is generated; but if the interpole is shortened, the intensity of the field under it must be greater than under a pole of full length in the ratio of full length to actual length, with due regard to
fringing of the flux. 167. Winding of
Commutating Poles.
The
calculation
of
the winding to be placed on the commutating poles presents no It is necessary to provide a sufficient number special difficulties.
of
ampere-turns to balance those of the armature and to supply
the m.m.f. required to drive the commutating flux through the transverse path n, s, Fig. 270, taking into account the m.m.f.
supplied by the main poles, N, S, in those parts of the path of the lines of induction which are common to both magnetic
circuits. The figure represents a bipolar generator revolving in the clockwise direction; if the machine were a motor revolv-
299
ing in the same direction, the polarity of the interpoles would have to be reversed, a condition that would be met automatically
the
upon reversing the current through the armature, since two windings are in series. Fig. 270 reveals the fact that the m.m.fs. of the main and commutating poles act in the same direction in two of the four quadrants of the armature core and of the yoke, and in opposite If there were no saturadirections in the other two quadrants.
would produce a proportional
<
tion each m.m.f.
flux, in
which
of the
case the flux in the armature core would be
$t
in
two
FIG. 270.
d>
Magnetic
5i>
*
circuits in interpole
machine.
quadrants, and
ing
flux
the other two, where <;

interpole.
is
the work-
produced
by the
Similarly,
in
two
'
of
the quadrants of the yoke the flux would be

^S
,
$t 4- ------&u
--~
an(^
_
o
^>
in the other two; here

of the
$ = v& where
t
v is
the coefficient
is
of dispersion
main
poles,
coefficient of dispersion of the interpoles.
and $ it = v&i where Vi The magnitude
the
of the
commutating
flux <{ is given
by
where B ig the flux density in the gap under the interpole, is determined by the value of the commutating e.m.f. to be generated,
,
300

b'i
and
and
l' t
are respectively, the corrected breadth
and length
,
of the interpole; these corrected lengths are greater
actual lengths
interpole. main flux
<
by from 3 to 4 times the
air-gap, d {
than the under the
Corresponding to this value of <f>; and to that of the there will be definite flux densities in each part of the closed magnetic circuit abcdefa, and to each flux density there will correspond a definite number of ampere-turns which may be determined from the appropriate B-H curves. Thus, let
AT =
ic
ampere-turns required by the two interpole cores
and shoes
AT =
ig
it
AT = A T' a A T" a A T' v A T" y

sum
ampere-turns for the two interpole air-gaps, 5; ampere-turns for the two sets of teeth opposite the
interpoles
= = = =
ampere-turns ampere-turns ampere-turns ampere-turns
for the for the
core, c to d for the yoke, e to f for the yoke, f to a
armature armature
core, b to c
Then, in the closed magnetic circuit, abcdefa, the algebraic of all the m.m.fs. must be zero, in accordance with KirchhofTs law. In this circuit there act, in addition to the m.m.fs. listed above, that due to the two interpole windings, AT and that due to the armature, A Tarm) where
t,
larm
p 2
_Zia ^ == Trd
.
ird
'
_=
''
^
be wound on each
not shunted.
It follows, then, that the

1
number
of turns to
interpole
is
ATi
r
2-a
*,
provided the interpole
coils are
larger number of turns may be used if a diverting shunt is placed across the interpole winding. The above discussion applies directly to the bipolar machine of
Fig. 270,
and with obvious modifications
applies also to multi-
polar machines.
168. Effect of
The presence
expression
of
Commutating Poles upon Coil Inductance. commutating poles causes an increase in the
inductance of the short-circuited
L =
LI
+L +L
2
under them. (Chap. VIII), the term

coils
In the
L 2 due
,

to tooth-tip leakage, as follows:
is
301
affected.
Its value
may
be computed
Suppose the center of the slot containing a coil edge to be a distance x cm. from the center of the commutating pole of corThe tooth-tip flux within the rected breadth b i} Fig. 271.
r
limits' of
the pole
is
10
10
Z
45'i
&'<
FIG.
271.
Tooth-tip flux under interpole.
and the average inductance, considering both

Yi
Pi)
sides of the coil,
is
fT^2
d%
f
(Vi
rO
10
t
In the above equations

all
b'i is
the corrected gap length,
5i r^_
If l'<l'i,
dimensions being expressed in centimeters. there must be added to the above expression a term
z
2
jgi
169.
it
(*'
I'd
1.46 logic
[l
Compounding Effect of Commutating Poles. In Chap. was shown that a forward lead of the brushes in the case of V, a generator produces a demagnetizing effect and consequently reduces the generated e.m.f while a backward lead causes a com.,
302

If
is provided with commutating brush displacement are accentuated, as may be seen from Fig. 272. Thus, Fig. 272a represents the conditions when the brushes are in the geometrical neutral axis; the armature winding between adjacent brushes of opposite polarity is then acted upon by the flux due to a main pole only,
pounding action.
the generator
poles, these effects of
Generator
(O
FIG. 272.
Compounding
effect of interpoles.
the effect of the oppositely directed fluxes due to the interpoles (shown by the hatched areas) being to annul each other. If the
brushes are displaced in the direction of rotation, Fig. 2726, the total flux is reduced by the difference between the hatched areas e f h and bed and in addition by the sum of areas a b c
and
g.
In the same manner a backward displacement of the
303
brushes, Fig. 272c, results in an increase of the flux linked with the armature winding. Similar considerations show that in the case of commutating
pole motors a forward lead of the brushes will increase the flux and therefore reduce the speed, while a backward lead will decrease
the flux and raise the speed. In Chap. VII it was shown that a considerable backward displacement of the brushes of a com-
mutating pole motor reversals of rotation.
may
If
result in a continuous succession of
actual reversal of direction does not
occur, pulsation of speed
may
to
result, for since
the effect of a
backward displacement
weaken the active field, there will be a corresponding decrease of counter e.m.f. and an increase of
is
armature current to produce the necessary acceleration of the

armature.
poles,
The
and so
still
increased current further strengthens the interfurther weakens the field and accelerates the
But the counter e.m.f. is proportional to the active and to the speed. The tendency to accelerate the armature will then continue until the decrease of counter e.m.f. due to reduced flux is offset by the increase due to greater speed. The decrease of field strength cannot, however, go on indefinitely, because the interpoles eventually become saturated, but up to the
armature.
flux
time that saturation of the interpoles sets in, the speed has been continuously increasing, and the momentum of the armature will cause the speed to continue to increase even after the flux has reached a practically constant value; especially will this be
if the rotating parts have large moment of inertia. The result will be a rapid increase of counter e.m.f., possibly to a value greater than the line voltage, in which case the machine would
true
become a generator drawing upon
its kinetic
energy of rotation
to send current back to the supply line. The speed under these conditions would rapidly fall, causing such a reduction of counter
e.m.f. that the
armature current again rises, thereby producing increased speed, and so repeating the above-described cycle of
changes.
CHAPTER X
EFFICIENCY, RATING
170. Sources
of
AND HEATING
Loss.
is
the net power output
less
In every dynamo-electric machine than the gross power input, the
difference being consumed in the internal losses of the machine. The losses may be grouped under the following heads:.
1.
THE OHMIC LOSSES, due

(a)
(6)
(c)
to the heating effect of the current
flowing through the resistances of
the armature winding, the field winding, the brushes and brush contacts.
2.
THE IRON OR CORE

(a)
LOSSES, due to
hysteresis in the armature core and teeth, (6) eddy currents in the armature core, teeth,
faces.
3.
and pole
THE MECHANICAL
(a)
(b)
LOSSES, due to
bearing friction,
friction
between the moving parts and
air,
or "wind-
age,"
(c)
brush
friction.
4.
ADDITIONAL MISCELLANEOUS LOSSES caused by

(a)
(6)
(c)
eddy currents in the armature conductors, the short-circuit currents in the commutated
pulsations of the flux set short-circuited coils,
coils,
up by the current
in the
(d)
eddy currents in the end plates of the armature, in non-insulated bolts through laminated core, etc.
171.
of
i
The Ohmic Losses.
amperes flows through a resistance of r 304
The power consumed when a current ohms is i 2 r watts.
EFFICIENCY, RATING
AND HEATING
armature and
"
field
305
wind-
The combined ohmic or

ings are
z i r
losses in the
commonly
referred to as the
"
(a) The armature copper motors is
loss in all
copper losses. types of generators and

(1)
Pea
*aV a Watts
at the
where the armature resistance is to be computed or measured normal working temperature. In general,
Ta
Po
sa
(1
~ 0.00420 a" ohms

at
(2)
where
p
la
sa
t
= specific resistance of the armature copper = total length of wire on the armature = cross-section of the armature conductors = working temperature of the armature in
Centigrade
C.
degrees
a
If
= number
of
armature
circuits in parallel.
mils, po
the length is expressed in feet and the cross-section in circular = 9.59 ohms; if these dimensions are expressed in centi-
meters and square millimeters, respectively, p

(6)
0.016 ohm.
is
The
field
copper
loss in
separately excited machines

if r f
2
P cf =
and
watts
is
(3)
in plain series generators
and motors
2
PC/
ia ?/
watts
field
(4)
where r/ represents the combined resistance of the and its regulating shunt if the latter is used. In shunt machines the field loss is
winding
Pcf =
where
is
i*
r8
#.2 - TS
= E is
t
watts
if
(5)
r a includes
the resistance of the regulating rheostat

field
one
used.
In long-shunt compound machines, the
loss (total) is
(6)
Pf =
C
z i a rf
2 is r s
watts
is
where
ia
ia
=
is
is
in
case
the
machine
a generator, and
in case it is
a motor.
20
306

compound machines
the total
field loss is
In short-shunt
P cf =
i rf
z -f i s r s
watts
.
(7)
the above relations again holding between i, i a and i 8 (c) The ohmic loss at the commutator depends upon the drop of potential at the transition surface between commutator and
brushes, as well as upon the amount of current flowing. drop of potential at each brush is Ae, the loss is
Pec
If
the
2i a
Ae
watts
(8)
volt
With the usual type of carbon brushes, Ae is approximately 1 when the brush current density has values common in
ordinary practice (see Fig. 261).
Values of
Ae
as determined
1 by Arnold are as follows:
Very soft carbon brushes, Soft carbon brushes,
0.4
to 0.6 volts
0.55 to 0.7 volts

0.9
1.2
Medium
carbon brushes,
to 1.1 volts
Very hard carbon brushes, Copper brushes (65 to 160 amperes per sq. in.)
It is
to 1.5 volts
0.017 to 0.03 volts.
common, however,
of the
mutator in that
sufficiently
ohmic loss at the comarmature winding, r a being increased to include the average brush contact resistance.
to include the
This method may lead to inaccuracy, especially at light loads. 172. The Core Losses.
(a)
Hysteresis Loss.
In the Armature Core. The relative motion between the armature core and the magnetic field of force produces a periodic
reversal of the
loss
magnetism of the core, thereby giving rise to a molecular friction in the mass of the through power armature core. This hysteresis loss can be represented by an empirical equation due to Steinmetz
of
Pha =
where
rjfVB a
watts
(9)
!Die Gleichstrommaschine, Vol.
I, p.
351, 2nd. ed., 1906.
EFFICIENCY, RATING
77
AND HEATING
307
of the core
a constant depending
T)Yl
upon the material
= ~T =
the
the
number
of
magnetic cycles per second
V = Ba =
If
the volume of the core
maximum
and
value of the flux density in the core.
metric units are used in the above equation

flux
(volume
in
cubic centimeters
rj= 0.0021
10~ 7
density in lines per sq. cm.), for ordinary sheet steel; if volume is ex-
77
pressed in cubic inches and flux density in lines per sq. in., = 0.0017 X 10~ 7 Since the weight, W, of the core is pro.
150000
50000
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.1-' 0.13 0.14 0.15
Watts per Lb.-
Cycle per Second
FIG. 273.
Curve
of hysteresis loss.
also
portional to its volume, the equation for the hysteresis loss can be written
Pn a
in
= rifWB a
7
watts
(10)
0.0062 X 10~ if British units are used (weight 77 pounds, flux density in lines per sq. in.). The curve of Fig. 273 shows the variation of the hysteresis loss, expressed in watts per pound per cycle per second, as a function of
which case
in
the flux density expressed in lines per sq. value of r.
in.,
using the above
308
In the Armature Teeth. The flux density varies from section to section because of the taper of the teeth, and it is not correct to compute the hysteresis loss by substituting an average value of flux density in the above equation.
Consider an element dx, Fig. 274, at a distance x below the
of the tooth; its
tip
volume
is
dV =
where k
is
blkdx
(b t
bi
b '*
Ikdx
that the total flux

flux density will
the lamination factor (from 0.85 to 0.90). Assuming is the same at all sections of the tooth, the
vary inversely as the width of the section, or
ri
tr dx
1
FIG. 274.
Computation
of hysteresis loss in teeth.
where
tooth.
is
The
the actual, or corrected, flux density at the top of the hysteresis loss in the element is then
dPht
.
-
dx
arid the total loss per tooth is
(11)
Since the volume of a tooth
is
the above expression can be written
EFFICIENCY, RATING
AND HEATING
1 -
309
'*
(12)
is
In other words, the expression for the hysteresis loss in the teeth similar to the general expression, but with the addition of the
factor
The
ordinates of Fig. 275 give the value of this factor for various
b' t /b t
.
values of
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
FIG. 275.
Correction factor, hysteresis loss in teeth.
(b)
Eddy Current
Losses.
That part
of the core loss
due to
eddy or Foucault currents can be approximately calculated by the formula derived below, but it is more usual to determine the
loss
under known experimental conditions for reasons that will appear later. Consider a radial element, Q, Fig. 276, of one of the armature core stampings. Let the thickness of the stamping be t, and let ct be the radial depth of the core, where c is a numeric. When the element is in the vertical position OA the flux passing through
,
its lateral
walls
is
maximum, and when

zero.
it is
in the horizontal
This change of flux occurs four times revolution in a per bipolar machine, or, in general, four times per magnetic cycle. The changing flux induces an alternating e.m.f.
axis
OB
the flux
is
310
and sets up a corresponding alternating current which may be assumed to flow in paths like those indicated in the lower part of the figure; an elementary current path is then bounded by similar rectangles of widths 2x and 2(x -j- dx), and lengths 2cx and
dx), respectively. elementary circuit will be
2c(x+
The change of flux through such an 4B a X 4cz 2 lines per magnetic cycle,
is
where
Ba
is
the
maximum
where /
volts.
lines per second,
flux density in the core, or 16B a cx 2f the number of magnetic cycles per
second.
The average
e.m.f. in the
elementary circuit
is
will
be
lQB a cfx 2
10~ 8
The
resistance of the path
r4cx 7 1 hdx
4x "[ hcdxl
FIG. 276.
Elementary paths
of
eddy currents.
where 7
is
the specific resistance of the material of the core.
The
loss in the
elementary path
is
1\
hdx
c)
and the
total loss
is
Pea =
~~
7x~io 16
c2
X * dX
==
ijo
7X10 16
c2
EFFICIENCY, RATING
AND HEATING
311
But
hct 2 is the
volume
7
of the element, hence the loss is

,.2
R. 2/2/2
10 16
c2
(volume of tooth) watts
(13)
This equation shows that the eddy current loss varies as the square of the flux density, the square of the frequency of the magnetic reversals, and the square of the thickness of the laminations;
and inversely
The equation cannot, however, be

sults,
as the specific resistance of the core material. relied upon for accurate re-
because the actual distribution of the current
may
differ
considerably from the assumed distribution, and the laminations

I5U.UUUI
312
the loss in watts per pound by the above formula is 0.22, while the observed value for these data in the case of annealed sheet steel is 0.44 watts per pound. Fig. 277 shows the variation of eddy
current loss with flux density at frequencies of 25 and 60 cycles per second and for laminations 14 mils thick. The loss at other frequencies and thicknesses can then be computed by observing
that the loss varies as the squares of these quantities. Eddy Current Loss in the Teeth. Referring to Fig. 274, the eddy current loss in an elementary section of a tooth is
dP et =
volume
fir
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
FIG. 278.
Correction factor, eddy current loss in teeth.
where
T-J
e is
the eddy current constant.

A.rt.n-I
rt.
Integrating,
let.
-.
*x-I
tv
*/
efH
B X volume
2
t
of tooth
J-^
(14)
(13), in that it
This equation differs from the original equation, contains the additional factor
log e bt/b't
the value of which
is
shown
as a function of
b' t /b t
in Fig. 278.
EFFICIENCY, RATING
AND HEATING
313
Eddy Current Loss made in Chap. II to

pole faces.
in
the Pole Faces.
Reference has been
the cause of the eddy current loss in. the This loss is confined to a relatively thin layer at
the face of the pole because the direction of the induced eddy currents is always such as to damp out the flux pulsations that
produce them (Lenz's law).
The flux pulsation at any given point in the pole face will pass through a complete cycle of changes in the time required for a point on the armature to move over a distance equal to
the tooth pitch, that
gives
is,
in a time
,.
seconds.
of teeth
This
a frequency of
ft
î
number
rev.
per sec.
Fig. 279 represents the variation of flux density at the pole face on the assumption that the curve of distribution is sinus-
FIG. 279.
Variation of flux density opposite teeth and

T>
slots.
T>
oidal.
The amplitude
if
of
the
pulsation
is
B =
r
^~
Then
= = =
peripheral velocity of the armature in centimeters per
second
fj,
permeability of the material of pole face specific resistance of material of pole face in absolute
electromagnetic units
1
the pole face loss in watts per sq. cm.
is
pp =
1
*ir STT \ HP
10
~7
fc2
x f STT \/~ \ HP
7
10
~7
(
15 )
Potier, L'Industrie Electrique, 1905, p. 35.
Rudenberg, Elektrotechnische
Zeitschrift, Vol.
XXVI,
p. 181, 1905.
314
where k 2
/B'\
(
is
a function of the relative dimensions of

slots.
the air-gap, teeth and
Adams
has worked out the curve
of Fig. 280 as giving fairly satisfactory values of k 2 in terms of the ratio b a /5. If British units are used (B g in lines per
sq. in., v in feet per second,
the loss in watts per sq.
in inches, and ^ and p as above) in. of pole face is given as

t
Pv =
1.65
10- 7 k 2 B a 2 ^l~ MP
(16)
k*
0.8
0.7
0.6
0.5
0.4 0.3
0.2
0.1
EFFICIENCY, RATING
AND HEATING
315
the peripheral velocity of the shaft in the bearings, up to velocities of about 1800ft. per minute; at higher velocities it varies The windage loss, as in the case directly with the velocity.
power of the speed. But in both cases these losses are independent of the load on the machine. (c) Commutator Friction Loss.
of fans, varies as the third
Let d com
Ab
= p = / =
c
diameter of the commutator, inches total area of brush contact, sq. in.
brush pressure
(Ib.
per sq.
in.)
coefficient of friction.
Cycles per Second 50 60
30
100
10
15
20
25 30 Watts per Lb.
35
40
45
50
FIG. 281.
Total core
is
loss.
Then the brush
friction loss, in watts,
Pbf =
746
0.0059 d com nA b p cf watts
(17)
is
Ordinarily the value of p c is from 1.5 to 2 Ib. per sq. in., and / about 0.3 for carbon brushes and 0.2 for metal brushes.
174. Additional Losses.
(a)
solid
of
Eddy Currents in the Armature Conductors. When large, armature conductors are used in open slots, different portions the same section of the conductor may be simultaneously in
Under these conditions
e.m.fs. of
fields of different strength.
316
magnitudes will be generated from point to point of the and eddy currents will result. The loss due to these eddy currents may be from 5 to 1 5 per cent, of the loss due to the ohmic resistance. This is equivalent to saying that, so far as the armature copper loss is concerned, the effective armature resistance is from 5 to 15 per cent, greater than the true resistance. This loss may be minimized by stranding the conductors or using
different
cross-section,
smaller conductors in parallel.

(b), (c)
and
(d)
Miscellaneous Losses
Due
to Short-circuited
Currents, Etc. These are minor losses, and cannot be computed. In testing, they are absorbed in the amount attributed to friction,
windage, and core

175.
losses.
Summary
of Losses.
COPPER Loss:
Armature ...............................
Field
:
i a *r a
2
separately excited ..................
if r/
................................. i a 2 r/ shunt .................... ............. is z rs compound, long shunt .................. i V/ + is zr9 2 compound, short shunt ................. i r/ + is 2 rs Commutator ............................ 2i a Ae
series
CORE Loss:
1 Hysteresis: armature core ................. rjfVB a
-
armature teeth ................ yfVtB
l
t
*-(rT X 5 -X
2
/p\Z~
Eddy
currents: armature core ............. efH 2 B a 2 V
armature teeth ............ tfWB^Vt

2 z pole faces .................. k Bg
ge&
^j
constant
MECHANICAL LOSSES:
Bearing friction and windage ..............
^ to 3
per cent, depending
upon
Brush
friction
size
> t
and speed.
........................... O.OQ59dcom nA, p .f
MISCELLANEOUS LOSSES:
Eddy currents in armature conductors ...... (0.05 to 0.15) Losses due to currents in short-circuited coils, etc.
2 i a ra
EFFICIENCY, RATING
AND HEATING
317
176. True Efficiency, Efficiency of Conversion, Electrical and Mechanical Efficiency. The true or commercial efficiency of a
machine, generally referred to simply as the efficiency, is the power output to the gross power input, both In the quantities being expressed in terms of the same unit. case of a generator the output is electrical and the input mechanratio of the net
ical; in
motors, the output In either case,

.
is
mechanical and the input
electrical.
efficiency
If
output T-T input
output " ri~~ losses output

-
input "
~~^
lossses "
(18)
input
are the terminal voltage and the line current, respectthe ively, expression for the efficiency of a generator becomes
t
and
Ei+P
t
(19)
where
is
the summation of
all
the losses.
That
of a
motor becomes
For the sake

into
of convenience, the total losses
(1)
may
be divided
the pure ohmic or copper losses, due to the flow of current through the resistance of the armature and fieldwindings, and (2) the stray power
loss,
two groups,
which include
all
the
remaining
losses.
Then
input
output
ohmic
losses
stray power loss
(21)
In the case of generators, the expression
input
stray power loss
output
ohmic
losses
(22)
represents the total electrical power actually developed in the armature, and the ratio of this total electrical power to the total
mechanical power input

that
77 C
is
defined as the efficiency of conversion;
is,
. .
~ = efficiency
input
ot
conversion
electrical
power developed -r r mechanical power input

; ;
stray power losses

-.
output
+
:
copper losses
electrical
input
input
The
ratio of the net electrical
output to the total
power
318

is
developed in the armature

rj e
called the electrical efficiency, or

electrical output -
electrical efficiency
=
(24)
electrical
power developed
out P ut
output
copper losses
Evidently,
f]
f\cf]e
(25)
In the case of motors, on the other hand, the expression
input
copper losses
output
stray power loss
(26)
represents the total mechanical power developed by the machine, and the ratio of this power to the input is again the efficiency of
conversion; that
_
is,
input
copper losses
input
output
stray power loss
input
the ratio of the net mechanical output
or
The mechanical
efficiency is
to the mechanical
power developed,
rj
mechanical efficiency
output
+ stray power loss

;
output
input
- copper
losses
It follows that
rj
Ticrim
(29)
loss includes all
stray power the losses except those due to the pure ohmic resistance of the various windings, and is, therefore, made up of the core loss,
friction
loss and the miscellaneous In the case of machines which operate at approximately constant speed and constant flux, as, for example, generators and motors of the separately excited and shunt
177.
The Stray Power
Loss.
The
and windage, brush contact
load losses.
power loss remains nearly constant at all loads within the working range. The distortion of the flux due to armature reaction tends to increase the core loss, but this
types, the stray
tendency
or less compensated, except in compound The stray the decrease of the flux as a whole. generators, by a simple loss in of a shunt be determined machine can power
is
more
manner by running
it
as a
motor without load
at its rated voltage
EFFICIENCY, RATING
AND HEATING
319
and normal excitation and measuring the current input to its armature (ia )o. The product of impressed voltage and armature current is then equal to the stray power loss plus the ohmic loss
in the
armature; or
P = E
s
(i a )
- (Oo
ra
(30)
This test should be made immediately after the machine has been running under load in order that the armature shall have reached
its
working temperature. In machines in which the speed or the
flux, or
both of them, are
inherently variable under operating conditions, as for example in The stray series motors, the stray power loss is also variable. a is found series motor loss of experimentally by running power
the machine separately excited and without load. Adjust the field current to about its full-load value and start the motor by gradually increasing the voltage impressed upon the armature.
After the motor starts, increase the excitation until the field current has the highest value it will have under load conditions, and adjust the armature voltage to a value at which a reading is
desired. Note the field current, armature current and voltage, and the speed. Keeping the excitation constant, adjust the armature voltage to two or three different values, and for each The armature voltages ordisetting take readings as before. narily used in this test are 250, 400 and 550 volts for 550-volt motors, and 300, 450 and 650 volts for 650-volt motors. Repeat
number of other values of field current, The to the lowest field current consistent with safe speed. stray power can then be computed by deducting the ohmic loss in the armature from the armature input.
this series of readings for a
down
into friction
to separate the stray power loss thus determined and windage and core losses, the machine should be run as a series motor without load and at reduced voltage. By
If it is desired
varying the impressed voltage through a sufficient range to cover the working range of speed, and taking simultaneous readings of impressed voltage, current and speed, the friction and windage
corresponding to any speed
may
be taken as equal to the power
supplied, less the ohmic losses in the armature and field windings, the core loss being negligible under these test conditions. The
core loss at any given speed
is
then equal to the difference be-
320

loss and the friction and windage loss In this manner curves like Figs. 282 and 283
tween the total stray power

at the
same speed.
14W
1200
1000
2800
^COO
400
200
EFFICIENCY, RATING
178. Variation
of
AND HEATING
321
Efficiency
Maximum
losses
:
Efficiency.
Let
it
with Load. Condition for be assumed that a certain 250-kw.,
550- volt, flat-compounded long-shunt generator has the following
COPPER Loss:
Armature and series field at full load Shunt field loss, including rheostat STRAY POWER Loss: Core loss Friction and windage
5500 watts. 1250 watts. 3500 watts. 2000 watts. 1000 watts.
Commutator
loss,
average
1000
200
250
350
400
450
Amperes Output
FIG. 284.
Losses and efficiency.
Assuming that all these losses, except those in the armature and series field, remain constant, there will be a constant loss of 7750 watts at all loads and a variable loss increasing according to a parabolic law from zero at no-load to 5500 watts at full-load.
21
322
These

losses
and the variation
ally in Fig. 284; the efficiency
77
550*
__
of the efficiency are
is
shown graphic-
computed from the equation
55(H'
7750
(r
+ r )i
f
where
i
2.27
The above example

consist of a constant
is typical of those cases in which the losses term and a term that varies as the square of
the load.
In such cases the
efficiency
has
its
maximum
value
when
the variable loss equals the constant loss.
Writing
t
E
it is
Pconst
,
(31)
,
-\
h (r a
clear that
77
will
have
is
its
maximum
t
value
when the denomi-
nator of the fraction
a minimum, E being assumed constant. the denominator with respect to i and equating to Differentiating
zero,
or
^
77)
(r.
r,)
i\ra
= Pconst =
i a *(r a
r/)
(32)
which agrees with the statement above. A second differentiation will verify the statement that the condition thus determined
is
for
minimum
it
value of the denominator of the fraction.
necessary to use greater refinement than in the preceding example, the following corrections may be applied: 1. The core loss (due to hysteresis and eddy currents) may be
is
When
taken to increase linearly with the load current because of the
compounding
where (Ph+e )o
2.
effect, or
Ph+e =
is
(Ph+e ) Q
+d
c is
(33)
the core loss at no-load and
a constant.
friction
will
The bearing friction, windage and commutator remain constant as before.
EFFICIENCY, RATING
3.
AND HEATING
be taken as
1
323
The
total brush contact
drop
may
20
so that the loss
total brush area in sq. in.

is
due to brush contact drop
Pb
4.
5.
ia
\20X
total brush area

2
+V
(35)
The armature
The shunt
loss is i a r a as before.
E
field loss is
2
t
in plain
shunt and in long-shunt
compound machines, and

(E
t
irj)*
ra
n short-shunt machines, the positive sign being used in case of generators, the negative sign in the case of motors. 6. The series field loss is i a 2 rf in long-shunt machines, and
<2
,
r/ in
short-shunt machines.
of all the losses therefore includes a constant
The summation
;erm, a
line current, and a ;erm that varies as the square of the current; hence the efficiency is given by an expression of the form
term that varies directly with the
77
____ E i E i+~Pconst + Cli + C E

t t
2Z
E + i-î
t
Ci
+ C*
(36)
Differentiating the
iition for
maximum
denominator and equating to efficiency is found to be
zero, the con-
or
Pconst
= C
2i
(37)
n this expression
lo-load
(i
Pcon
st
is
,
evidently equal to the total loss at
or that part of the load loss which aries as the square of the current, is nearly equal to i a 2 (7a ?"/)
0),
and C 2 ^ 2
Motor and Generator Testing, Westinghouse
Electric
and Manufactur-
ng Co.
324
Hence, for
field
maximum
efficiency, the loss in the
armature and
series
should equal the loss at no-load. 179. Location of Point of Maximum Efficiency. From the preceding article it is clear that by a proper choice of the rela-
tion between the fixed
and the variable

be
losses
the point
oi
maximum
efficiency
may
made
to
fall
at
any desired output
For example, assume that the total losses consist of a constant term, Pcons <, and a term variable with the square of the load Let the rated full-load output be P and let the constant loss
,
be
xP
where x
is
the variable loss
any fractional part of the output, and let at full load be yP 0) where y is any fractiona
part of the output.
Then the
r]
efficiency at full-load is
Y 1
+x+y
(38)
Let zPo be the output at which it is required that the efficiency be* a maximum. The variable loss will be z 2 (yPo), and for maxi-
mum
or
efficiency
z (2/Po)
2
= xP Q
(39:
be required to divide the total losses in such efficiency shall occur at three-fourths load and that the efficiency at rated full-load shall be 85 per cent.
let it
For example,
a
way
that the
maximum
Then
-
0.85
0.75
from which x
It is
6.35 per cent, and y = 11.3 per cent. if the fixed losses, represented by x, are relaand the variable copper losses, represented by y, tively large, are small, maximum efficiency will probably occur beyond full-
seen that
load.
To make
the
maximum
efficiency occur at a fractional
part of full-load, the copper losses should be large compared with the fixed losses. Thus, if it is known that a machine
to be operated for considerable periods at light loads
and only
EFFICIENCY, RATING
AND HEATING
325
as to
occasionally at full-load or overloads, it should be so designed have a relatively high armature resistance in order to make
the efficiency a maximum at or near the point of average load. The all-day efficiency of a machine 180. All-day Efficiency. is the ratio of the net energy output to the total energy input during a working day. Inasmuch as charges for electrical
service
are based largely on energy consumption

120
(kilowatt-
326

power input will vary as shown by curve the fixed and variable losses are, respectively, 10 and
of its rated output, the
B' while
y
if
5 per cent, of the rated output, the power input will be given by curve C. In the former case the all-day efficiency is 85.8 per The difference between the cent., in the latter 81.7 per cent.
two becomes greater and
greater, in favor of the machine with the lower fixed loss, as the period of light load increases; for example, if the machine runs for nine hours at 10 per cent, load,
and one hour at full-load, the all-day efficiency of the first machine is 75.7 per cent, as against 64.3 per cent, for the second. 181. Rating and Capacity. The rating or rated output of a machine is based on, but does not exceed, the maximum load which can be taken from it under prescribed conditions of test.
If these prescribed conditions are those of the A.I.E.E. Standardization Rules, the rating is said to be the Institute rating; if! the prescribed conditions are those of the International Electrical
Commission, the rating is said to be the I. E.G. rating. The capacity of a machine, expressed in terms of its output, is the load or duty it will carry for a specified time, or continuously, without exceeding certain temperature limitations, as described
in the next article.
The new (1914) Standardization Rules 1 two kinds of ratings, namely, continuous
rating, the latter
of the A.I.E.E. define

rating,
and
short-time
applying to machines designed for discontinuous or intermittent service. In determining the continuous rating,,
is
the machine
tions, for a sufficient length of
subjected to a heat run, or test under load conditime to bring about a constant
difference of temperature, of prescribed amount, between the machine and the surrounding air. If the load on the machine is normal full-load, it may take from six to eighteen hours to reach stationary temperature conditions, but the time may be reduced by overloading the machine to a reasonable extent during the preliminary period. By taking temperature readings at more or less regular intervals during the test, and plotting rise of
temperature against time, the shape of the curve so obtained will what extent the load should be increased or decreased. The short-time rating of machines intended to operate intermittently, that is, with more or less frequent stops of
indicate to
1
Proceedings A.I.E.E., August, 1914.
EFFICIENCY, RATING
AND HEATING
327
sufficient duration to allow cooling to occur, is the load that the
machine will carry for a specified, limited period, without exceeding prescribed conditions of test. Machines which operate on a cycle of duty that is repeated more or less regularly, as in elevator service, are rated in terms of an equivalent load which may be based either on a continuous
or short-time test, but selected to simulate as closely as possible
the thermal conditions of actual service.
The standard duraand 120
tions of short-time equivalent tests are 5, 10, 30, 60
minutes.
specify that the rated output both generators and motors shall be expressed in kilowatts, thus marking a departure from the practice of rating motors in terms of horse-power. For practical purposes, the horse-power rating, if used, may be taken as four-thirds of the kilowatt rating. 182. Allowable Operating Temperatures. Theoretically, the
of
1
The new Standardization Rules
output of a generator is limited only by the possibility of sufficiently reducing the resistance of the receiver circuit, at the same time maintaining the generated e.m.f. and supplying the driving
power; practically, however, the capacity of the machine is by the ability of the insulation to withstand without deterioration, and for long periods, the maximum temperature caused by the heating due to i z r and other losses, though in some cases the load limit may be determined. by commutating conditions. For each kind of insulating material there is a limiting temperature above which deterioration is very rapid, but so far as useful life of the insulation is concerned there seems to be no particular advantage in operating at temperatures below the safe limits. In case the machine is designed to operate at temperatures well within the safe limits, there will be a margin between its rating and its capacity, hence these terms are not synonymous.
limited
If
of the insulation is rapid, the
the safe limits of temperature are exceeded, the deterioration damage increasing with the dura-
tion
and extent of the excess temperature. In the Standardization Rules of the American Institute of
Electrical Engineers in force prior to the adoption of the

1
new
The material in this article is taken from the Standardization Rules of the A.I.E.E. adopted by the Board of Directors, July 10, 1914, and effective
December
1,
1914.
See Proceedings A.I.E.E., August, 1914.
328
(1914) rules, it was specified that the allowable rise of temperature of the parts of a machine (excepting railway motors) should be as follows: armature and field windings, 50 C.;
commutator, 55 C.; bearings, 40 C. These rises of temperature were based upon standard conditions of a room temperature of 25 C., a barometric pressure of 760 mm., and normal conditions
of ventilation.
It
perature differed
was further provided that if the room temfrom 25 C., the observed rise of temperature
per cent, for each degree difference between room temperature and 25 C., the correction to be added to the observed rise if the room temperature was below 25 C., and subtracted if it was higher.
should be corrected by
In the
new
rules (effective
December
1,
1914), emphasis
is
placed upon the highest permissible temperature of the hottest The rise spot as well as upon the maximum rise of temperature.
temperature in the case of air-cooled machines (excluding railway motors) is based upon an ambient temperature of 40 C., but it is particularly specified that the observed rise of temperature must never exceed the limits given in the following table:
of
TABLE OF HOTTEST-SPOT TEMPERATURES AND OF CORRESPONDING PERMISSIBLE TEMPERATURE RISES
Class
EFFICIENCY, RATING
whatever
test.
AND HEATING
329
may
be the ambient temperature at the time of the
It will be noted that the above temperature limits recognize the advances in the art of constructing insulating materials that have been made since the adoption of the superseded rules.
While it is known that insulating materials coming under the head of Class B can be supplied to withstand maximum temperatures of 150 C. and even higher, the limit has for the present been set at 125 C., pending the accumulation of more extensive data at higher temperatures; machines designed for maximum operating temperatures in excess of 125 C. must be subject to
special guarantees
by the manufacturer. The temperature limits of commutators

difficulties
Current per brush arm
so constructed that
no
from expansion can occur are as follows:

Max. permissible temperature 130 C.
200 amperes or less 200 to 900 amperes

900 amperes and over
130
C. less 5
for each 100
amperes increase above 200.

95
C.
These temperatures hold for the metallic parts only. The temperature of insulation used in the commutator, or of any insulation whose temperature would be affected by the heat of the commutator, must in no case exceed the limits prescribed in the
table of hottest-spot temperatures. The new rules abolish the requirement of a correction of the
rise of temperature due to a difference between the ambient temperature at the time of the test and the standard reference temperature (except in the case of air-blast transformers which are not considered in this text). This is due to the fact that numerous tests have shown that the effect of variations of the ambient temperature is small, obscure and of doubtful It is, however, recommended that tests be condirection. ducted at ambient temperatures not lower than 25 C. The
observed
high altitude in increasing the temperature rise of some machinery is recognized by specifying a reduction of the normal permissible temperature rise to the extent of 1 per cent, 1 for each 100 meters by which the altitude exceeds 1000 meters.
effect of
types of
In the case of machines intended for operation at an altitude of

1
Water-cooled
oil
transformers are exempt from this reduction.
330
is
less, a test at any altitude less than 1000 meters and no temperature correction is necessary. Three methods of determining temperatures of the various parts of a machine are specified, one or the other of these methods
1000 meters or
satisfactory
being adequate for commercial

1.
tests.
THERMOMETER METHOD,
including measurements
by mer-
cury or alcohol thermometers, by resistance thermometers, or by thermo-couples, any of these instruments being applied to the hottest accessible part of the completed machine. When this method is used, the hottest-spot temperature is estimated by adding a hottest-spot correction of 15 C. to the highest temperature observed, except that when the thermometer is applied directly to the surface of a bare winding, such as an edgewise
strip of 15
conductor or copper casting, the correction

C.
is
is
C. instead
of
The ambient temperature
to be measured
by means
several thermometers placed at different points around and halfway up the machine at a distance of 1 to 2 meters, and protected from drafts and abnormal heat radiation. To this end the
thermometers are to be immersed in oil in a suitable heavy metal cup, for example, a massive metal cylinder with a hole drilled partly through it. This hole is filled with oil, and must be sufficiently deep to, insure complete immersion of the bulb of the thermometer. The smallest size of oil cup permitted by the rules consists of a metal cylinder 25 mm. (1 inch) in diameter and 50 mm. (2 inches) high, but the size of the oil cup must be increased with that of the machine under test. The object of thus
increasing the size of the oil cup is to avoid errors in the calculations of temperature rise due to the time lag between changes of
temperature of the machine and the surrounding air, this time lag being greater the greater the size of the machine. Where machines are partly below the floor line in pits, the temperature of the armature is referred to a weighted mean of the pit and room temperatures, the weight assigned to each being based on the relative proportions of the machine in and above the
pit.
The temperature
pit.
stantly in the pit
must be
of the portion of the field structure conreferred to the ambient temperature in
the
EFFICIENCY, RATING
AND HEATING
331
2. RESISTANCE METHOD. This method consists in the determination of the temperature of windings by measurement of their
increase of resistance;
when this method is used, careful check readings must be taken by means of thermometers, but without disassembling the machine, in order to increase the probability
of
method
revealing the highest observable temperature. Whichever yields the highest temperature, that temperature shall
be taken as the highest observable temperature and a hottestspot correction of 10 C. added thereto. This method is not permitted in the case of low resistance field coils where the joints and
connections form a considerable part of the total resistance. In the case of resistance measurements the temperature
be computed from the formula temperature in degrees CenFrom this it follows that the rise of temperature of a tigrade. winding is given by the formula
coefficient
of
copper
is
to
I/ (234.5
-f-
t),
where
is
the
initial
=
where
Rt+e
(234.5
R
t)
^(i -ft
l)
(40)
resistance of winding at
6)
degrees
R =
t
resistance of winding at
degrees.
This IMBEDDED TEMPERATURE -DETECTOR METHOD. 3. method involves the use of thermocouples or resistance coils
located as nearly as possible at the estimated hottest spot, but is to be used only with coils placed in slots. The thermocouples
or resistance coils are built into the machine, and a sufficient number shall be employed to insure locating the hottest spot.
They should be placed

between the
coils
coils
and
in the case of
two sets of locations, one and one between the top and bottom two-layer windings, or between the coil
in at least
core;
and wedge in single-layer windings. Detectors of this kind will assume a temperature practically equal to that of the adjacent coils. A correction of 5 C. is to be added to the
highest reading in the case of two-layer windings with detectors between coils and between coils and slots; and a 10 correc-
C. for each tion in the case of single-layer windings, plus 1 1000 volts above 5000 volts terminal voltage (single-layer windings are commonly used in alternators, seldom or never in
direct-current machines).
332

Operating conditions in
183. Heating of Railway Motors.
the case of railway motors are much more severe than in ordinary motors because of restricted space and the nature of the
It is therefore good practice to permit higher workservice. ing temperatures for short periods than in other types of machines. Further, the variable nature of the load makes it more
difficult
to give a definite rating to railway motors.
The nominal
rating of a railway
therefore, arbitrarily defined as the mechanical output at the car or locomotive axle, measured in kilowatts, which causes a rise of temperature above the suris,
motor
rounding
C. at the air, by thermometer, not exceeding 90 commutator and 75 C. at any other normally accessible part,
run at its rated voltage on a stand with the covers arranged to secure maximum ventilation without external blower. The rise in temperature, as measured by C. The statement of the resistance, shall not exceed 100
after one hour's continuous
nominal rating must also include the corresponding voltage and armature speed. The continuous ratings of a railway motor are defined as the inputs in amperes at which it may be operated continuously at one-half, three-fourths and full voltage respectively, without exceeding the specified temperature rises tabulated below, when operated on stand test with the motor covers and
The system of cooling system, if any, arranged as in service. ventilation must be defined, and if cooling is by means of forced draft the volume of air on which the rating is based must be given.
TABLE OF MAXIMUM TEMPERATURES AND TEMPERATURE RISES
Class of insulating materials
EFFICIENCY, RATING
AND HEATING
333
334
ventilation due to the motion of the car or train.
be higher than in actual service because of the absence of the In general,
the temperature rise in actual service will be from 75 to 90 per cent, of the temperature rise on stand test in the case of enclosed motors, the losses being the same in both cases; and from 90 to 100 per cent, in ventilated motors. 184. Temperature Specifications of Electric Power Club. The Electric Power Club, composed of representative manufacturers, successor to the American Association of Motor
Manufacturers,
has
adopted
:
the
following
specification
for
direct-current generators
rise shall not be more than 40 C. on all and 45 C. on commutator after a run of normal windings, rated current and voltage for a time sufficient to give approximately stationary temperatures.
"The temperature
two hours immediately following the normal load
After a 25 per cent, overload in current at rated voltage for run, the
temperature rise on the winding should not exceed 55 C. and on the commutator 60 C. by thermometer. The machine should also be capable of standing a momentary overload of 50 per cent." Specifications for direct-current motors of the open, protected
and enclosed types, and for various classes of service, are summarized in the table on page 333. 185. Output Equation. A definite relation, originally derived G. between the rating, speed and dimensions of exists by Kapp, the armature. This relation, when expressed in algebraic form,
is
commonly
referred to as the output equation.
Thus,
let
E = â =
t
\l/
rated terminal voltage rated armature current

ratio of pole arc to pole pitch
= =
ampere-conductors per unit length of armature periphery.
Since
E =
>
eT^To*
^
(nearly)
and
3>
= BgU =
Bg
(nearly)
EFFICIENCY, RATING
AND HEATING
is
335
the power output of the machine in kilowatts
KW
where
"
Eia
*
^
60
10"
(42)
The numerical value of this coefficient. " depends upon the design constants" of the machine, tj Bg and q, but principally upon B g and q since the range of values of ^ is limited. Bg is clearly a measure of the degree
is
called
the output
coefficient
of utilization of the
magnetic material of the machine; similarly, in part a measure of the specific utilization of armature copper, for it is closely related to the thermal characteristics, as
is
has been shown by Adams. 1
Thus
let q
be expressed in ampere-
cir.mils.
FIG. 286.
Calculation of copper loss per sq.
in.
of armature surface.
conductors per inch of armature periphery, and let h be the current density in the armature conductors expressed in circular mils per ampere. Let Fig. 286 represent a portion of the armature
"
surface (shown as a smooth-core type for convenience) of
inch
square.
Each conductor
'
will carry
a current of
its
Qi
amperes,
and
its
cross-section will be
is
h circular mils
resistance per
inch of length
length =s
1 x Y~ =
a Y~h
a
1
Trans. A.I.E.E., Vol.
XXIV.,
p. 653, 1905.
336

working tempera-
since the specific resistance of copper at the
ture of the armature
The
i r
loss
very nearly per conductor is then

is
ohm
per circular mil-inch.
/%A
\a
I
a
-.--r
ijfi
T watts ah
ia
But the number

-.iia/
of conductors per inch of

i r
z
armature periphery
armature surface
in.
is
-, ia
hence the
loss per sq. in. of
is
dli
X
1a
=
fl
-jr
watts per sq.
(43)
The value
or
less,
up
of q varies from about 400 in machines of 20 kw. to about 800 to 850 in machines of 1000 kw. capacity.
The
ratio q/h (watts per sq. in. due to copper loss) is generally in the neighborhood of unity for ordinary peripheral velocities of 2500 feet per minute, but may be as high as 2.5 in large machines
running at high peripheral speeds (6000 feet per minute) where the ventilation is more effective. Values of B g range from about 40,000 lines per sq. in. in small machines up to 60,000 lines per The value of generally lies between sq. in. in large machines. 0.000015 (small machines) and 0.000056 (large machines). 186. Heating and Cooling Curves. The energy losses in any machine are converted into heat and cause a rise of temperature whose final value depends upon the heat capacity of the materials of the structure and upon the facility with which the heat may be
radiated or otherwise dissipated. The temperature will become stationary when the rate of heat generation becomes equal to the rate of dissipation.
It is of interest to derive the
homogeneous body
for the reason that
law of heating and cooling of a it throws light on the conof a generator
ditions obtaining in the or motor.
more complex structure
Let
Q = s =
heat generated per second, in kg-calories = amount of heat specific heat of the substance
required to raise 1 kg. 1 C. weight of the body in kg.
radiating surface in sq. cm.
coefficient of
= A = a =
cooling
= amount
of heat in kg-cal.
EFFICIENCY, RATING
AND HEATING
337
dissipated per second per sq. cm. of radiating surface per degree difference of temperature between body and
0i
= =
surrounding medium temperature of body in degrees Centigrade temperature of surrounding medium in degrees Centigrade.
1. Heating of the body. In a time dt the temperature will increase by dO degrees. During this interval the heat liberated amounts to Qdt kg-calories, and the body absorbs sWdd kg-calories. The remainder will be
dissipated, to the
amount Aa(0
Q\)dt kg-calories, so that
Qdt
Transposing,
= sWdB
~
Aa(B
Oi)dt
(44)
sWdB
Q- Aa(et
e,)
Assuming that
61
when
0,
J>
which gives
- sw
d6
Aa(0-0i)
When
=
,
(0
-0i)-> =
^
rise of
(46)
and
this is the limiting
equation (46)
may
temperature be written
the body.
The
last
Q = aA(B -
0i)*
co
(47)
which expresses the fact that when the temperature becomes stationary the rates of heat production and dissipation are
equal.
Cooling of the body. In this case no heat is being developed, consequently and the fundamental equation becomes
2.
()
Q =
(48)
= sWdB
is
Aa(0
t
0i)dt
If
the temperature
22
O degrees when
0,
338

dt
Jo
which gives
dd = - sW ( JQ Aa(9
=
that
0,
(0
sW
If
0i
(49)
as the equation of the cooling curve.

j,
is, if
(B
0j)*
= 00
is
the temperature at the beginning of cooling
equal to the limiting temperature at the end of the heating period,
20
10
12345G789
Time in Hours
FIG. 287.
Heating and cooling curves.
the equation of the cooling curve is the same as the variable part of the heating equation, but with a change of sign. Hence, in this case, the heating and cooling curves are of the same logarith-
mic shape, but one

other, as
is
turned upside down with respect to the
shown
in Fig. 287.
Differentiating the heating equation
_aA
-
~sW
and substituting t = origin is found to be
in the result, the slope of the curve at the
EFFICIENCY, RATING
that
is,
AND HEATING
339
dependent upon the mass and material of the body, but
not upon its cooling area or the nature of the radiating surface. In fact, at the first instant, all of the heat is absorbed and none of it is radiated; hence, the slope of the curve at the origin gives
the rate at which the temperature would rise if all the heat were If the temperature continued to rise at this rate, the absorbed.
limiting
temperature
is
rise,
would be
reached in a time
sW T = .seconds; T
cxA.
called the time constant of the body.
The heating equation can then be written
To
substitute
numerical values in the above equations, the
following relations obtain
Q = = s = s =
1
0.2386
0.527
X (loss in kw.) kg-cal. per sec. X (loss in kw.) Ib-cal. per sec.
0.11 for iron
0.09 for copper.
The value of a may be found from the experimentally determined fact that when air is blown across the bare (or thinly
varnished) surface of an iron core its surface temperature will rise 1 C. when the radiation is 0.0038 (1 0.25*;) watts per sq. cm.
the velocity of the air in meters per second this is equivalent to 0.0245 (1 0.00127?;) watts per sq. in. if v is in feet per minute. From this it follows that
of surface,
where
v is
0.906
0.250*0 sq. cm. per 1 C.

(1
10~ 6 kg.
cal.
per sec. per

(51)
where a
v is
given in meters per second; or

(1
12.89
sq. in.
+
1
per
0.00127*0 C.
10~ 6 Ib-cal. per
sec. per
(52)
where
.core is
1
v is
The experiments
Ludwig
Ott,
if the surface of the coated with a thick double layer of varnish the radiation
expressed in feet per minute. of Ott also showed that
London
Electrician, 1907, p. 805.
340
is

(1
watts per sq. cm. per 1 C. rise of surface meters temperature, ybeingin per second;or 0.0194(1 + 0.00054y) watts per sq. in. per 1 C, v being in feet per minute.
0.0030
0.107
v)
Temperature
rise
computed from the above formula
will not
observed rise in actual machines because it neglects the transfer of heat from the winding to the core, or vice versa; likewise, the irregular distribution of heat evolution and the thermal capacity of the insulation. But in general terms it will be true that the ultimate rise of temperature can be expressed by the equation watts dissipated B ,. 0i = constant X ,. (53) ^-j
agree in general with the
.
radiating surface
where the constant is
in each case to
be determined by experiment.
187. Heating of the Armature. The experimental results of Ott referred to above may be changed to a form applicable to the rotating part of the machine. Taking the value of the radiation
for bare or thinly varnished surfaces, the temperature rise
is
given by
61
_ w ~
460
1
+ O.Mv
where
w = a = v =
total watts dissipated total radiating surface, sq. cm. peripheral velocity of armature, meters per second
while for a heavily varnished surface
333
a
I
0.107v
1
(55)
Accordingly, the rise of temperature for a radiation of

per sq. cm.
is
watt
rise
found by putting a
w =
1.
The temperature
1
(in degrees Centigrade) for a radiation of expressing v in feet per minute, is
watt per
sq. in.,
71.3
1
7v + 0.00127z>
_
JL
rr for
a bare surface
52
.
~f~
U.UUuo4v
for a heavily varnished surface
Other writers give the value of this constant as follows:
EFFICIENCY, RATING
Metric units
AND HEATING
English units 85 _
1
341
55C
Kapp
O.lv
0.00051*;
Arnold
1
300
'+0.10 354
1
46.5_
O.OOOSlz;
Esson
O.OOOGy
Wilson
640
,
99
0.00091z;
+ 0.180
645
Thompson
100
1
+ 0.3 \/v
embodied
sq. in. as
0.0213 \/v
The above expressions

rise of
are
in Fig. 288,
which shows the
temperature per watt per
a function of the per-
1000
2000
3000
Feet per Minute
4000
5000
6000
FIG. 288.
Relation between temperature rise and peripheral velocity of armature.
It will be observed that the ipheral velocity in feet per second. curve represented by Arnold's formula lies nearly midway between the two corresponding to Ott's researches, at least for values of v within the usual limits of practice.
rise of
be noted that all of the above formulas for computing temperature are more or less uncertain, unless applied to a machine of the same type as that from which the constants were
It should
experimentally determined. tween the curves of Fig. 288
A
is
large part of the differences bedue to the fact that there is no
342
Some
absolute agreement as to what constitutes the radiating surface. writers specify the outer cylindrical surface only, but
including the surface of the end connections as well as of the core: others include the exposed sides of the core in addition to the
Evidently all exposed surfaces are but not to the same extent. The radiating heat, of the flow of heat from the interior to and direction magnitude
useful in
in different directions;
outer cylindrical surface.
the exterior of a mass will depend upon the heat conductivity and since the conductivity along the laminations is much greater than across them (Ott found it to be from
50 to 100 times greater) it follows that unless the core is very deep the greater part of the heat will be dissipated from the cylindrical surface. for the rise of
has been pointed out 1 that a rational equation temperature of an armature should be of the form
It
where
j^-; 2di -f c2a 2
+ bv
(56)
2i = sum
Sa2
c
of cylindrical cooling surfaces
= sum of end surfaces = a variable coefficient
less
than unity.
The value
of c will be smaller the greater the ratio of heat conductivity along the laminations to that across them.
The Arnold formula
for rise of
armature temperature
(40 to 70)
is
Ol
~ w
a
1
0.00051
where a and
respectively.
is
are expressed in square inches and feet per minute The numerical coefficient in the numerator is t(
be taken near the lower limit of its range when the ventilatior good. In using this formula, however, it should be notec that w does not include the watts dissipated in the end connec tions, nor does a include their surface; in other words, the rise* of temperature of the armature core is to be distinguished fron that of the end connections. Consequently, to estimate the rise* of temperature of the core, the value of w to be inserted in th(
formula
1
is
Ott,
London
Electrician, 1907, p. 805.
EFFICIENCY, RATING
AND HEATING
of winding
^~r-
343
w =
total core loss

of a
-\
embedded length r~n ~~c total of

length
winding
<58)
The value
recommended by Arnold
is
the cylindrical surface
of the core, plus the two end surfaces, plus half the lateral area of the walls of the ventilating ducts; or (Fig. 289)
irdl
ird aver h(2
+
X
nv )
(59)
In the case of the end connections,

<JM
.
free length of
,
total length of
winding winding
(60)
2irdL
344

it
heat conductor than the air

rise of
replaces.
Measurement
of the
temperature by the increase of resistance gives the average rise of temperature of the winding as a whole, the maximum rise at the hottest point being from 12 to 20 per cent, greater than the average rise. The average rise of temperature of the entire winding is from 40 to 60 per cent, greater than the average
rise of
temperature of the exposed surface; the latter

of
by taking the mean
is determined thermometer readings at the middle and
ends of the exposed cylindrical surface.

of the
Formulas for computing the form

9
rise of
temperature of
coil
field coils are
watts lost in
radiating surface of coil

Different writers assign various values to the constant C, depending upon the selection of what constitutes the radiating surface. Obviously, C means the rise of temperature due to a radiation of 1 watt per unit area. If the radiating surface is expressed in square inches and is taken to mean the outer cylindrical surface exclusive of the exposed end, the value of C for open type machines, and under standstill conditions, is from 70 to 80, with an average of 75. The value of C decreases with increasing peripheral velocity of the armature, due to fanning action, by approximately 5 per cent, per 1000 feet per
minute, or
C =
where
v is
75(1
0.00005?;)
(62)
peripheral velocity in feet per min. This is an average the decrease of C being somewhat greater if the coils are value, because of the cooling effect of the yoke; and somewhat short, less if the coils are long. In machines of the protected type C
is
in enclosed
approximately 50 per cent, greater than the above value, and machines from two to three times greater than given
(62).
by equation
Field coils of the ventilated type of construction are made of concentric parts with an open space of about J^ inch between them. The greater surface presented to the air by reason of this construction permits of increased radiation; however, the internal surfaces of the ducts are not as effective as an equal area on the
outside.
For a given temperature
rise
the ventilated
coil will
EFFICIENCY, RATING
radiate about 50 per cent,
AND HEATING
345
dinary coil; or, taken as equal to 50. 189. Heating of the Commutator. The commutator is heated to brush the due losses friction, P&/, and by the flow of the by current across the contact resistance between commutator and The rise of temperature can be computed from the brushes.
more watts per sq. in. than an orwhat amounts to the same thing, C may' be
formula
W W
2\.
where
= =
total loss at the

TTUcomvcom
commutator
peripheral velocity of
commutator
in
feet
per
minute.
190. Rating of Enclosed Motors. If a motor of the open type converted into one of the enclosed type, it is clear that its rating must be reduced to avoid excessive temperature rise.
is
Experience shows that a reduction in horse-power rating of about 30 per cent., accompanied by an increase of speed of 20 per cent.
The give a temperature rise within standard limits. reduction in horse-power rating decreases the current and con2 sequently the i r losses, and the increase of speed permits a
will
loss
reduction of the flux per pole, thereby lowering the excitation and the core loss. The core loss decreases notwithstanding
the increase of speed, for the effect of reduced flux density more effect of increased frequency of the magnetic reversals (see Fig. 281); the core loss varies nearly as the square
than outweighs the
of the flux density, and approximately as the first power of the speed since the hysteresis loss is always greater than the eddy current loss.
PROBLEMS
ohm and a 169 ohms. When running without load the armaFind (a) the stray ture current is 1.5 amp. and the speed is 997 r.p.m. power loss and the loss in the shunt winding; (6) the true efficiency, the efficiency of conversion and the electrical efficiency when the armature current has its full-load value of 25 amp., assuming that the stray power loss remains constant at its no-load value.
1.
220-volt shunt motor has an armature resistance of 0.44
shunt
field resistance of
346
2.
Find the efficiency of the above motor when the armature current has 5, 10, 15, 25 and 35 amp., and plot a curve showing the relation between efficiency and horse-power output. 3. At what value of armature current will the above motor develop its
values of
maximum
efficiency
efficiency, and what are the corresponding values of maximum and horse-power output? 4. If the rated horse-power output of the above motor is obtained when the armature current is 25 amp., what will be the armature current at ^, Plot a curve showing the relation between J^, 2ii and 1J times rated load? armature current and per cent, of rated load. 6. If it is required to design a shunt motor that shall develop a full-load efficiency of 92 per cent, and in which maximum efficiency shall occur at J of full-load, what must be the variable and fixed losses in terms of full-load
rating?
6.
What From the
will
results of
be the maximum efficiency under these conditions? Problem 4, compute the all-day efficiency of the
motor
load for 2 hr., at load for 3 hr., at load for if it operates at 3 hr., at full-load for 1^ hr. and at \Y times full-load for hr., the working day comprising 10 hr.
7. If the machine of Problem 1 is operated as a shunt generator with a terminal voltage of 220 volts and an armature current of 25 amp., what will be its true efficiency, its efficiency of conversion and its electrical efficiency, assuming that the stray power loss is the same as in Problem 1 ? 8. The total current output of the machine whose commutator and brushes have the dimensions specified in Problem 7, Chap. II (p. 87) is 90 amp. Find the total loss at the commutator. If the active length of the commu-
tator
is
5.5
in.,
what
will
be
its
probable
rise of
temperature under full-load
conditions?
9. A 220-volt shunt motor takes a field current of 1.3 amp. when the machine has been standing idle for several hours in a room which has a temperature of 30 C. After running under load for several hours the shunt field current is found to be 1.1 amp. Find the average rise of temperature
of the field winding. 10. In making the preliminary design of a 6-pole 100-kw. 250-r.p.m. generator, it is decided to use an air-gap density of 52,000 lines per sq. in.
and 600 amp.-conductors per in. of periphery. If the ratio of pole arc to pole pitch is to be 0.7, and if the pole faces are to be square, what must be the diameter and the length of the armature core?
CHAPTER XI
BOOSTERS AND BALANCERS. TRAIN LIGHTING SYSTEMS
A booster is a dynamo-electric machine whose connected in series with a circuit, its generated e.m.f. being added to or subtracted from that of the circuit, depending upon the polarity of its excitation. Boosters may be
191. Boosters.
is
armature
driven by any form of prime mover, but are generally directconnected to a shunt motor taking current from constant potential mains.
192.
The
Series Booster.
An
obvious use for a booster
is
to
raise the voltage of a generator, or of a section of the bus-bars of a central station, by an amount sufficient to compensate
the ohmic drop in a feeder supplying a distant load, in case the load is of such character as to require the same voltage Since as receiving devices at or near the source of supply. the line drop is directly proportional to the current, the voltage of the booster should also be proportional to the current; in other words, the booster should have an external characteristic
It is consisting of a straight line through the origin. impossible to exactly realize this form of characteristic without auxiliary devices, but it may be approximated sufficiently closely for practical purposes by designing the booster as a
series-wound
generator
with
flux
densities
well
within
the
point of magnetic saturation. The hysteresis effect illustrated in Fig. 104, p. 116, is especially objectionable in boosters, and
should be reduced to a minimum. Further, if the excitation is of such character that the main flux is subject to wide variations, the magnetic circuit must be laminated throughout in order
that eddy currents set up by a change in the flux may not be of sufficient magnitude to retard the change of flux and so make
the machine sluggish in
its
action.
of a series booster, of the
The compensation, by means
drop
347
348
of potential of a circuit due to its ohmic resistance is equivalent to a complete cancellation of the resistance of the circuit. Or-
dinarily,
if
the resistance of a circuit
duction would be
made by an
is to be reduced, the reincrease of the cross-section
and therefore
of conditions,
of the weight
and cost
of the line.
Up
to a
certain point, which

it
readily be computed for a given set will be cheaper to save energy by adding copper
may
than to
install
a booster equipment; beyond that point the

of a circuit
booster will be more economical.
The apparent cancellation of the resistance means of a series booster is sometimes utilized
of
by
in electric rail-
ways employing a ground return to mitigate the electrolysis underground structures such as water and gas mains, telephone cables, and the like. The return circuit of the ordinary street railway system consists of the track and the surrounding earth, the current dividing between these paths in the inverse
ratio of
tracks a take place through the earth along paths of low resistance afforded by underground metallic structures, resulting in damage wherever stray currents leave these paths to return through moist earth to the track or to the
their
resistances.
Even with well-bonded
considerable flow of current
may
grounded bus at the power house. It is becoming standard practice to minimize the danger of electrolysis in such systems by installing insulated negative feeders or cables which connect points along the track directly to the negative bus of the generating station, thereby draining the track current away from the stray paths. These negative feeders are clearly the more
If a series booster is now connected in such a feeder so that its e.m.f. acts in the direction from the track to the negative bus, the equivalent resistance of the feeder may be reduced nearly to zero, and most of the cur-
effective the lower their resistance.
rent will return to the station by way of the feeder. Boosters used in this way are called negative or track-return boosters. 193. The Shunt Booster. In constant-potential systems in which the load changes gradually, but covers a range from a very
small to a considerable value,

storage
may
battery in parallel then be used to carry the entire load at times of light load, and in parallel with the generator at the time of peak load. At
practice to use a with the bus-bars. The battery

it
is
common
BOOSTERS AND BALANCERS
349
other times the battery takes charging current from the generator, thus insuring a fairly uniform load on the generator during its working period, with consequent economy in cost of fuel: In
a system of this kind a so-called shuntbooster is used to force charging current into the battery against the latter's counter e.m.f.,
the connections being shown in Fig. 290. The field winding of is connected across the main bus-bars, never across its own armature, hence the machine is really separately excited.
the booster
The booster armature
is
charging period, and is called upon e.m.f., hence the above connection
booster voltage
is
in series with the battery during the to supply a relatively small

of the field winding. The means of the field by
manually controlled
rheostat, adjustment being
made when
the readings of the am-
FIG. 290.
Connections of shunt booster.
meters indicate that it is necessary. By inserting a reversing switch in the field circuit, or by using a reversing rheostat, the
booster e.m.f.
may
be added to that of the battery, thereby

is
assist-
ing the battery to discharge if its voltage for current is unusually heavy.
low, or
if
the
demand
lead storage battery
when discharged
to the permissible
limit gives 1.8 volts per cell, and when fully charged requires " an impressed e.m.f. of 2.65 volts per cell to give it the overcharge" that is periodically required to keep it in good condition.
therefore, the voltage of the system is E, the total number of cells required is E/l.S to provide for the contingency that the
If,
load.
battery alone, when nearly exhausted, may be used to carry the At the end of a prolonged charge the battery voltage will then have risen to 2.65 X (E/l.S), hence the booster must be
350
capable of generating
E (^rir
l)
= OA7E volts; thus, in a
110-
volt system, the maximum booster voltage will be 52 volts. The design of the booster will then be completely determined when the
maximum
discharge rate of the battery is known. of the motor that drives the booster need be only from two-thirds to three-fourths of the volt-ampere capacity of
The capacity
the booster for the reason that
mum current the voltage is
during the periods of erably reduced. The normal, (eight-hour) discharge rate of a lead battery is defined as that current which, flowing uniformly for
eight hours, will reduce the battery voltage to the minimum value of 1.8 volts per cell; the current during overcharge should
when the latter delivers its maxiand when the voltage is highest, overcharge, the current must be considlow,
be not greater than one-half of the eight-hour rate. In Fig. 290 the cells shown at the right-hand end of the battery are the end-cells which are cut in and out of circuit by means of an end-cell switch. Their purpose is to adjust the battery voltage to the line requirements to compensate for the changes in voltage due to varying conditions of charge and discharge. Thus, in a 110-volt system, the number of cells required will be 110/1.8
61 when fully discharged; but when a fully charged battery begins to discharge its terminal voltage is 2.15 volts per cell,
therefore requiring =-r 2. lo
.
51
cells.
Consequently in such a
system 61
cells
would be
installed, 10 of
them
as end-cells.
field is
The
pro-
number
of end-cells
may
be reduced
if
the booster
vided with a reversing switch, for in that case the booster e.m.f. can be made to oppose that of the battery to a sufficient extent to bring the terminal voltage to the proper value. In 194. The Constant-current or Non-reversible Booster.
isolated plants supplying a
lamp load and a
fluctuating
motor
load, as in hotels
and
office buildings, it is
necessary to maintain a
constant lamp voltage, and it is permissible or even desirable to allow the voltage of the power circuit to fall when there is a heavy rush of current, as on starting an elevator. Fig. 291 represents a
type of installation frequently used in such a case. The shunt field winding of the booster, /, is connected across the constant potential lighting bus and its magnetizing effect is opposed by that of
351
the series winding, S, as indicated by the arrows. The excitation due to/ is normally the greater of the two, and the resultant differ-
produces a booster voltage that acts in the same and which is from 10 to 15 .volts under At normal load the adjustments are such that the battery neither charges nor discharges, in other
ential excitation
direction as the generator, normal load conditions.
words, the sum of the voltages of generator and booster equals the open-circuit voltage of the battery. The entire lighting and power load is then carried by the main generator. If the motor
suddenly increased, there is an initial tendency to draw the increased current from the generator, but this results in an
load
is
increased excitation of the series winding of the booster and a reduction of its generated e.m.f the original condition of balanced
.
voltage at the battery terminals is therefore disturbed, and the battery discharges and .relieves the generator of the current in
FIG. 291.
Constant-current or non-reversible booster.
Conversely, a decrease of the a momentary weakening of the series excitation of the booster and a charging current therefore flows into the battery. The current through the armature and series field of the booster is therefore not constant, as the term constant-current booster might imply, but it is substantially so, the total variation of a few per cent, being no more than is
excess of the normal amount.
motor load
results in
sufficient to cause the battery to take
up the fluctuations
of
current above and below the average value. When a storage battery charges or discharges, its terminal voltage rises or falls, respectively, by an amount very nearly proportional to the
current; thus, if the current is equal to the eight-hour rate, the 1 change of voltage is 0.05 volts per cell, and at the one-hour rate
1
If
8-hr, rate (see p. 350), its capacity
the capacity of a storage battery is C 8 amp-hr. when discharged at the is greatly decreased if it is discharged at
352
(equivalent to four times the current at the eight-hour rate), the variation is 0.2 to 0.21 volts per cell, provided the battery is initially
fully charged.
The function of the booster is then to produce a change of voltage at the battery terminals corresponding to the charge or discharge rate demanded by the load. For example, asrequiring, say, 115 volts per cell. If
sume that the voltage of the motor circuit is 230, cells when charged to a normal voltage of 2
the load calls for a supply of current equivalent to the eight-hour discharge rate of the battery, over and above the normal supply
amperes, the booster voltage must be lowered by = 5.75 volts. This can be accomplished by so proportioning the series winding that an increase of the current through it from I aver to I avcr (1 -|- p) will produce the necesof
laver
115
0.05
sary change in field excitation, where pX 100 is the prescribed percentage variation of booster current. The linear variation
of booster voltage of course requires that the
magnetic
circuit
be worked on the straight part

(see Fig. 100, p. 113).
the magnetization curve Change of battery voltage with varying

of
conditions of charge can be compensated by manual regulation of a rheostat in the shunt field of the booster; but in the type of
service to
which the non-reversible (and other automatic) boosters
are adapted, the fluctuations of load causing alternate charge and discharge are so rapid that the general condition of the
greater rates. Thus, if the current is such that the voltage per cell falls to 1.8 volts in 1 hr., the current is said to be the 1-hr, rate, and the capacity falls to Cs amp-hr. The reduced capacity is due to the fact that the high
rates of discharge produce chemical changes of great velocity in a thin surface film of the active material, thereby preventing the electrolyte from
The relation between discharge rate (ri) and penetrating to fresh material. the corresponding capacity in amp-hr. (Cn} is given approximately by the
formula
'
(See data in Storage Battery Engineering, by Lamar Lyndon, 3d ed., p. 98, and Foster's Electrical Engineers' Pocketbook, 7th ed., 1913, p. 875.) If is is the current corresponding to the 8-hr, rate, and i n the current corresponding to the w-hr. rate, it is clear that C 8 = 8i 8 and C n = ni n whence, substituting the above approximate relation between Cn and Cs, it follows
;
that

battery
suited to
353
changes very little. The non-reversible booster is systems in which the average motor load is small
Booster.
and the fluctuations are considerable.

195. Reversible
In
systems in which
it
is
not
desirable that battery discharge shall be accompanied by a drop of voltage of the power circuit, as in a railway system having a
in Fig.
large average load, the reversible booster shown diagrammatically 292 is frequently used. It differs from the non-reversible
booster in that the current through its armature is not unidirectional, though in both types the shunt and series field windThe object of the booster is ings are differentially connected.
to hold the load on the generator at a constant value equal to the average load on the system, leaving the battery to take up the fluctuations. It is adapted to systems in which the average
load
is
large
compared with the range
of the fluctuations.
FIG. 292.
Differential or reversible booster.
Referring to Fig. 292, the battery is so designed that its open circuit voltage is equal to that of the system; consequently, when the load on the system has its normal (average) value, the
battery must neither charge nor discharge, and the current through the shunt field of the booster must be adjusted so that its magnetizing effect exactly neutralizes that of the series winding
With increased demand on the line there is a slight increase current of through S, and a resultant magnetization of the booster direction that the generated e.m.f. acts in the same a in such direction as the battery; discharge of the battery then takes place.
S.
e.m.f. generated in the booster armature must therefore be equal to the drop of battery voltage that corresponds to the disrate demanded by the load, plus the ohmic drop in the ar-
The
charge
mature
23
of the booster itself.
On
the other hand,
if
the load
354
falls

below normal the shunt winding overpowers the series windand the booster voltage is added to that of the generator,
ing,
with the result that a charging current flows into the battery. of the booster is determined by the fact that maximum current and maximum e.m.f. occur simultaneously. Although the open-circuit voltage of the battery is nominally equal to that of the generator and of the system, its actual voltage may vary over a considerable range, depending upon the state of the battery charge. To compensate these changes
The capacity
the excitation of the shunt field must be adjusted by hand regulation of a rheostat in series with the shunt winding. 196. Auxiliary Control of Boosters. Both the reversible and
the non-reversible boosters described in Articles 194 and 195 have

the disadvantage that a given change of current in the series coil of the differential winding produces a definite voltage, without re-
gard to the fact that the change of battery voltage corresponding to each rate of charge or discharge varies with the condition of the battery; that is to say, a given change of current in the series
winding of the booster will not always automatically result in the desired rate of charge or discharge. Moreover, the heavy current that must be handled by the series winding requires a conductor of large cross-section and a machine frame of excessive
dimensions and weight per kilowatt of capacity. To obviate these difficulties there have been developed several automatic
systems that regulate the battery by external means, and in which the booster has a simple shunt winding. These systems have practically superseded the types of differential boosters
described above.
197. The Hubbard Counter E.M.F. System (Controlled by the Gould Storage Battery Co.) is shown diagrammatically in The field coil / of the booster B is in series with Fig. 293.
latter being in turn excited
the armature of a small motor-driven exciter E, the field of the by the main generator current, or
its
fractional part thereof.
The adjustments are so made that when average value the exciter E produces an e.m.f. equal and opposite to that of the line. There is, therefore, no current through the booster field winding, no e.m.f. is generated in the booster armature, and the battery, which has a
the load has
normal voltage equal to that
of the line, neither charges
nor

discharges.
results in
355
An
increase
of
load
an increase
of the current
of the exciter, the
generated e.m.f.
above the average value through the series coil of the latter then exceeds
the line voltage, and a flow of current is established through the booster field winding in the proper direction to generate in the
booster armature an e.m.f. that assists the battery to discharge.
Shunt
FIG. 293.
Hubbard counter
e.m.f.
system of booster regulation.
Conversely, a decrease of load weakens the field of the exciter, the polarity of the booster reverses, and the battery then takes a charging current.
198. The Entz System (Electric Storage Battery Co.) of external control for installations of large capacity is illustrated in Fig. 294. The main output of the station passes through a coil
S consisting
of
a few turns of heavy strap copper, and produces an
FIG. 294.
Automatic booster regulation, Entz system.
electromagnetic pull on a core attached to one end of a pivoted lever L. When the station output has its average value the pull
in such a manner of the electromagnet is balanced by a spring that the lever presses upon the piles of carbon plates, R i and R%, with forces that make the resistances of the two piles equal to
each other.
The carbon
piles are
connected to each other at
356
the top and to one terminal of the field winding /' of a small motor-driven exciter E; at their lower ends the carbon piles are
connected to the terminals of a small auxiliary storage battery, the middle point of which is connected to the other terminal of The armature of the exciter E supplies current to the field /'.
So long as the resistances of RI /, of the booster B. and R% are equal to each other and all the cells of the auxiliary battery are equally charged, there will be no difference of potential between the terminals of /' consequently there will be no e.m.f. generated in either the exciter or the booster. Under these conditions the main battery, which has a normal voltage
winding,
;
equal to that of the line, will neither charge nor discharge. If the load current increases there will be a tendency to increase the
generator current through S, and the pressure on RI will be increased; this causes a reduction of the resistance of RI and the auxiliary battery will send a current through /', thereby
generating an e.m.f. in the exciter armature and energizing the field of the booster. The field windings of E and B are connected in such order that the booster voltage adds to that of
the main battery and a discharge results. In case the load falls below its average value, the spring overpowers the pull of S and the resistance of R 2 becomes less than that of Ri producing
a reversed flow of current through /' and, therefore, through / also, so that the booster voltage acts in the same direction as
the line voltage and forces charging current into the main battery. The combination of the resistances RI and R%, the auxiliary battery and the exciter field winding /' is entirely similar to the
a Wheatstone bridge. The two equal halves of the battery correspond to the ratio arms of the bridge, and RI and R 2 to the variable and unknown resistances; the field winding/'
circuits of
is
affected
the equivalent of the galvanometer. The polarity of /' is by the same causes that make the galvanometer in the
bridge circuit deflect one bridge arms are varied.
way
or the other as the resistances of the
In installations of small capacity the exciter and the auxiliary battery can be dispensed with; the use of the auxiliary battery is not absolutely necessary in any case, for the main battery, or a part of it, may be used directly. The purpose of the auxiliary battery is to avoid imposing unequal loads upon individual cells

of the
357
main battery. If the exciter is omitted, the connections shown leading to /' are transferred to /, but this can be done only when the capacity of the booster and the magnitude of its field current are small; the size of the carbon piles is limited by the
fact that the practically constant current through
S can produce of variation the a narrow on carbon plates, range pressure only hence the unbalancing of the bridge circuit can produce only moderate current through circuit/' (or/).
199.
The
Bijur System
external
control, Wheatstone principle of the or bridge, potentiometer circuit.
illustrated
(General Storage Battery Co.) of in Fig. 295, also utilizes the
The equal
ratio
arms Ri and
#2,
connected in series across the line, are each provided with a series of taps connected to a set
of contact points of
graduated
Spring
lengths, as at PI and Pz, which dip in or out of the small troughs
mercury, Hg, as the lever L is tipped one way or the other by

of
the control magnet

IS
restraining spring. designed to have a normal
S or by the The battery

FIG. 265.
open-circuit voltage equal to that of the line, consequently it
Automatic booster regu-
lation, Bijur system.
will neither charge nor discharge when the system is carrying its average load if the control apparatus is adjusted so that under these conditions the lever is horizontal for the booster field / is
;
then connected to points which have no difference of potential between them, with the result that the booster remains unexcited. An increase of load causes a slightly increased flow of
current through S and the resulting counter-clockwise movement of the lever short-circuits more and more of the resistance R\ as the movement proceeds; this disturbs the balanced condition
of the circuit
through/ and a current
will flow
through
it
in such a
direction that the generated e.m.f. of the booster causes the decrease of load will make the pull of battery to discharge. the spring overpower that of the coil S, the lever turns in the
358

,
clockwise direction, short-circuiting more or less of R z and the current through / reverses, causing a reversal of the booster
e.m.f.
and a flow
The Bijur system
of charging current into the battery. differs from other systems of external control
in that the latter involve a variation of generator current proportional to the battery charge or discharge, whereas the former
fixed variation of generator current.
causes a response from the battery to the desired extent with a This follows from the fact
that the magnet S and the restraining spring are so proportioned that with a given current through S the pulls due to them balance each other at all points within the range of motion of the lever. The result of this condition of neutral equilibrium is that a change of current that unbalances the forces by an amount just sufficient to overcome the friction of the moving
The parts will produce a continuous movement of the lever. excitation of the booster will then go on increasing in the proper
direction to relieve the generator of all but the initial variation. 200. Balancers. Fig. 296 (a, b and c) represents three methods of possible connecting a balancer set for the purpose of
maintaining equality, or approximate equality, between the voltages on the two sides of a three-wire system (see Art. 123, Chap. VI). If, with the connections shown in Fig. 296a, the load becomes unbalanced, the voltage on the more heavily loaded side will fall while that on the more lightly loaded side will rise. Under these conditions the unit on the heavily loaded side will act as a generator, thereby checking the extent of the voltage drop, while the other unit will act as a motor and so limit the rise of voltage on that side; but the drop in speed of the balancer, due to the load on the motor element, will prevent the generator element from assuming a sufficient part of the unbalanced load to maintain the potential of the neutral as nearly constant as would be
the case were the speed to remain constant. A partial compensation of this shift of the neutral may be effected by the system
of field connections in voltage
shown
on the heavily loaded side
in Fig. 2966; in this case the drop will weaken the field of the
motor, thus tending to increase its speed, while at the same time the rise in voltage on the lightly loaded side will strengthen the field of the generator element, thereby tending to still further balance the voltage on the two sides of the system; but the
359
balance cannot be perfect for the reason that the automatic response of the balancer depends for its inception upon an actual
unbalancing of the voltage. Perfect regulation is however posif the units comprising the balancer are compound-wound as in Fig. 296c, where the series windings are connected in such a manner that the current in the neutral excites the gensible
erator cumulatively, while in the
motor
it
acts
differentially.
voltage on the heavily loaded side is therefore kept up by the combined effect of increased excitation and increased speed; but whereas in the system of Fig. 2966, this automatic action
The
was dependent upon an unbalanced voltage, in the system of 296c it depends upon the unbalanced current, .and it is therefore possible to adjust the compounding to maintain perfect equality of voltage on both sides of the neutral.
Fig.
a
FIG. 296.
c
set,
Connections of balancer
three-wire system.
1 The condition to, be satisfied by Lighting. any system of train lighting is that the lamp voltage shall be maintained at a constant value independently of the number
201. Train
lamps in use and independently of the speed and direction motion of the train. In the case of steam railroads three methods of electric lighting are in use: 1. The straight storage system, in which each car is equipped
of
of
with
2.
single constant voltage generator placed in the baggage car or on the locomotive supplies current to the entire train. 3. The axle-lighting system, in which a small generator, mounted
its own storage battery. The head-end system, in which a
under each
car, is
driven directly from the axle.
iSee Trans A.I.E.E., Vol. XXI, 1903, pp. 129-227.
360
The straight storage system was used in the earliest installations of electric lighting on steam railroads. It has the disadvantage that the gradual exhaustion of the battery results in toward the end of long runs. The batteries must be charged at terminal or division points, or else be reinferior illumination
placed by fully charged batteries. In the head-end system a single compound- wound generator is driven by a turbine taking steam from the locomotive; in some
cases the turbo-generator unit is installed in the baggage car, in others it is mounted on the locomotive. The complete equip-
storage batteries, generally one for each car, may be operated when the train is parted, as during switching, and at low train speed. The standard (lead) for Pullman consists of 16 cells, corbattery equipment sleepers
in order that the lights
ment must include
responding to a nominal lamp voltage of 30 volts. The variation of battery voltage between the extremes of full discharge and full charge requires the use of an automatic regulator in order to maintain constant voltage at the lamps. In the axle-lighting system the maintenance of constant voltage is complicated by the fact that the speed of a generator
from the car axle will not only vary through wide the machine but must be capable of operating in either limits, direction. Generators of the ordinary types do not possess inherent operating characteristics suitable for such service, and to make a machine of ordinary type conform to the requirements, more or less elaborate regulating devices must be used. Naturally, axle-driven generators must be used in connection with storage batteries in order that the lights may not go out when the train is stationary or when the speed is so low that the generator voltage is less than the normal lamp voltage. The design of generators for automobile lighting is similar to that of axle-driven machines for train lighting except that there is no need to provide for reversal of the direction of rotation. This follows from the fact that in the former case the generator is driven from the engine, which always runs in the
positively driven
same
direction.
202. Voltage
Regulation
in
Train
of
prevent
objectionable lamps, automatic regulation must be provided to compensate
variation
Lighting Systems. the candle-power of
To
the

for the variation of battery voltage
full
361
between the extremes of
and, in the case of axle-driven generators, to overcome voltage variations due to change of speed. The various methods of regulation may be classified In some as either mechanical or electrical (or electromagnetic).
full discharge,
charge and
systems the maintenance of constant voltage also involves regulation for constant current output from the generator, hence the latter, when in use, delivers constant power; such
regulation is not entirely satisfactory, for it takes no account of the fact that the charging current of a lead battery should " taper," that is, become gradually less, as the battery ap-
proaches the fully charged condition. It is possible to arrange the regulatory devices in such a manner that the voltage and current output of the generator are controlled by the battery voltage, or else to make the generator control the line voltage and therefore also that of the battery.
Under the heading of mechanical methods of regulation may be included those axle-lighting systems in which the generator
voltage is controlled by the slipping of the driving belt, as in the Stone generator, or by a slipping clutch. In these systems the speed of the generator is maintained constant when the load
increases above a definite predetermined load which causes slipping to occur. In the Stone system the generator is provided
with an automatic device, consisting of a rocker arm on the shaft, for reversing the polarity of the generator terminals when the direction of rotation is reversed; there is also an automatic,
centrifugally operated switch arranged to establish the connection between the generatoV and the battery when the speed and
generator voltage have reached predetermined pick-up values, and to break the connection when the speed is below the assigned limit.
Under
electrical
all
or electromagnetic
methods
of regulation
is
may
(a)
be grouped which voltage control of resistance in the variation by the automatic

systems in
obtained
in the exciting circuit of the generator; or (6) of the armature reaction of the generator to secure the desired
characteristics.
lamp circuit or by the utilization
Examples
of
these methods are given in the
following 203. Resistance
articles.
Regulation.
Fig.
297 illustrates diagram-
362
matically a type of automatic regulator which operates by varying the resistance of a pile of carbon disks connected in the main lamp
the battery voltage E B rises above normal, as during the charging, lamp voltage EL tends to increase also. This .causes an increased flow of current through the solenoid Si, and
circuit.
If
of its plunger increases the pressure on the carbon thereby reducing its resistance and permitting an increased flow of current through solenoid $ 2 The motion of the plunger of $2 then releases the pressure on the carbon pile R, increasing its resistance to a sufficient extent to absorb the greater part of the
the
movement
r,
pile
increase of
EB
solenoids Si
as an ohmic drop in R. Since the response of the and $ 2 is dependent upon a variation of EL the
,
lamp voltage cannot be held absolutely constant, but the variation will be small; the use of the solenoid Si and pile r increases the
sensitiveness of the response of >S 2 to a change in L The lamp regulator is used in conjunction with a generator regulator described in Art. 204.
.
204. Generator
method
Field Regulation. Fig. 298 illustrates a of regulating the generator voltage by the variation of a
FIG. 297.
Voltage regulation by means of resistance in main line (Gould system).
resistance
in its field circuit.
by the two solenoids
The carbon pile R is acted upon V and B, the former responding to changes of
the generator voltage due to change of speed, and the latter to For example, assuming that variations of the battery current. the contact is closed and that the generator is charging the
any increase of speed will tend to increase both the generator voltage and the charging current. As the charging current increases, the upward pull of solenoid B relieves the pressure normally exerted upon R by the weight of the plunger of B, thus increasing the field resistance of the generator and lowerbattery,
ing
its
voltage.
To
prevent excessive overcharge of the battery

of solenoid
due to high generator speed, the plunger
is ar-
363
ranged so that the increased line voltage causes it to relieve the pressure on the right-hand side of R, thereby increasing the field resistance and lowering the generator voltage. is closed, and the connection between The automatic switch
FIG. 298.
Voltage regulation by means of (Gould system).

is
field
rheostat
generator and battery

erator
is
of the gensufficiently high to generate a voltage capable of actuat-
established
when the speed
ing the solenoid

series
the generator current, flowing through the winding of switch K, reinforces the pull of the shunt windS. When the speed falls below this pick-up speed, the battery ing
S',
FIG. 299.
Voltage regulation using vibrating contact.
voltage overpowers that of the generator and a reverse current flows through the series winding of K, with the result that the net
force acting upon the plunger is not sufficient to hold the contacts closed against the gravitational pull. The entire load is
364

The pole-changer
is
then carried by the battery alone. sented diagrammatically at PC.
repre-
Another method, somewhat similar to that
of Fig. 298,
but
involving a vibrating contact analogous to that of the Tirrill The automatic switch regulator, is illustrated in Fig. 299.
operates in the same manner as in Fig. 298, but the solenoids B and V act on the same magnetic circuit and open and close the
contact C.
the
field
Thus,
if
the battery charging current exceeds the
safe limit, coil
C and momentarily short-circuits winding of the generator, thus reducing the generator
closes contact
Coil V operates similarly if the generator voltage rises voltage. too high because of high rotative speed.
205. Field and Line Regulation. Fig. 300 is a diagram of connections of a system of train lighting which, like the system
FIG. 300.
Combined
field
and
line regulation (Safety Car Heating Lighting Co.).
&
described separately in Articles 203 and 204, includes independent regulation of generator voltage and of lamp voltage. The resistance of the carbon pile C, in series with the shunt
field
erted
Coil
winding of the generator, is controlled by the pressure exupon it by levers operated by the plungers of coils S and A.
carries the entire generator current Coil its full rated value.
and
is
adjusted to hold
the current at
line, is
A, shunted across the set to hold the generator voltage at 39 volts on equipments
having 16-cell lead batteries (2.45 volts per cell); and at 78 volts on "60-volt" equipments having 32 cells of lead batteries. If Edison batteries are used, these voltages are set at 43 volts and on part of 86 volts, respectively, by opening the short-circuit The object of the resistance in series with the voltage coil A.
XY
365
limiting the generator voltage to 2.45 volts per cell of lead battery is to prevent excessive overcharging of the battery when the
;
fully charged, the charging current will then automatThe limitation of generator ically taper down to a safe value. current imposed by coil S prevents overloading of the generator
battery
is
due to lamp load or to charging an exhausted battery. The lamp voltage is controlled by the pair of carbon piles, E, in series with the lamps, the two piles being connected in parallel with each other. The pressure upon these piles is due to a system of levers and a toggle joint actuated by a coil connected across the lamp mains. The pull of this electromagnet is opposed by a spring, the design being such that the armature of the electromagnet will remain in any position within the limits of its travel
is normal. In this system the armatures of the magnets controlling the generator and the lamp circuit are provided with air dash-pots having graphite plungers. The effect of variation of temperature upon the voltage coils of the generator and lamp regulators is compensated by means of resistors, having zero temperature coefficients, placed in series with these coils. The automatic switch for establishing the connection between
when the lamp voltage
generator and battery at train speeds above the pick-up speed is The shunt coil lifts the similar to others already described. pivoted armature when the generator voltage equals the battery
voltage, thus bringing into action the series coil, which assists the shunt coil in holding the switch tightly closed, and which accel-
erates the opening of the switch
when the generator voltage
falls
below battery voltage. The four brush arms of the generator are mounted on a rocker ring carried on ball bearings, the ring being free to rotate through 90 degrees between a pair of stops. When the machine is running in one direction, the friction of the brushes against the commutator holds the rocker ring against one of the stops and the brushes are then in the proper position for sparkless commutation. Reversal of the direction of rotation causes the rocker ring to be
turned through 90 degrees against the other stop, thus preserving
the original polarity of the generator.
Fig.
300a shows
cludes a
diagrammatic form another system which inlamp regulator and a generator field regulator, F. The
in
366
automatic switch
coil
K is closed in response to the pull of the voltage

their
when the generator speed and voltage have attained
proper values, thereby connecting the battery to the generator. The charging current, flowing through the series coil of the regulator F, tends to be maintained at constant value by the
action of the carbon pile rheostat in circuit with the generator field. At the same time the ampere-hour meter, is
AHM,
running in the direction of charge, and when the battery is has reached its point of charged and the contact needle
contact,
the resistance R is short-circuited;
thereupon the switch

pull of the shunt coil
is
energized, contact
is
closed,
and current flows through
the shunt coil of the regulator F.
The
4=^=3C

207.
367
Lighting Generator. The Rosendescribed 1 in 1905, embodies a number of interesting structural features and has operating characteristics that make it suitable for train-lighting service. Its distinctive
The Rosenberg Train

first
berg generator,
properties are
which
it
develops an e.m.f. the direction of and (2) that independent produces a current which, beyond a certain speed, remains
(1)
that
it
is
of the direction of rotation,
The diagram
practically constant no matter how much the speed of connections of a bipolar machine
is is
increased.
shown
in
Fig. 301, but it will be understood that with suitable modifications the principle is applicable to multipolar machines. The
FIG. 301.
Diagram
of
Rosenberg train-lighting generator.
battery E, which must be used in connection with the generator if the latter is to function properly, supplies current to
the lamps
train is at rest and also to the shunt field producing the polarity indicated by nn, ss. The axis of commutation of the brushes BB is in line with the axis of
when the
winding
//,
the poles, instead of being at right angles thereto as in the ordi-
nary type of machine.
The brushes
1
BB are connected to the battery terminals through
Elektrotechnische Zeitschrift, 1905, p. 393.
368
an aluminum cell C which offers a very high resistance to the flow from the battery to the armature and only a very small resistance in the direction from the armature to the battery;
of current
property of the aluminum cell prevents the discharge of the battery through the armature when the train is at rest or when running at a speed below that at which the generator picks up
this
its
load.
In addition to the main brushes there

auxiliary
is,
circuited
polar axis, that
is a pair of shortbrushes, bb, placed at right angles to the in the same position as the main brushes of an
ordinary generator. Rotation of the armature through the magnetic field set up by // will produce a flow of current through the short-circuited armature along axis bb, thereby creating a powerful cross-field,
field finding a path of low reluctance the shoes. As is clear from the figure, clockwise through pole rotation will result in a cross-field directed from left to right;
b,
the lines of force of this
the motion of the armature conductors through this cross-field then generates an e.m.f. and current along the BB axis in such a direction that the armature m.m.f., represented by the arrow B opposes the excitation due to the field winding //. In case
,
is reversed (that is, becomes counterthe of the cross-field Mb also reverses, and direction clockwise) the effect of this double reversal is to preserve the original polarity of the brushes BB. The fact that the armature m.m.f., B the due to the field winding means that the excitation opposes field flux parallel to the BB axis is small, and this in turn prevents
the direction of rotation
excessive current through the short-circuit bb. differs widely from the ordinary generator in that
The machine
what
is
usually
the main
secondary importance with respect to the cross-field. The weak field in the BB axis obviates commutation difficulties that would otherwise arise due to the short-circuiting of winding elements under the middle of a pole face; such diffifield is of
culty as might
still
exist
is
further overcome
by notching the
It is clear that there is a pole faces opposite the main brushes. definite limit beyond which the main current delivered by the
BB cannot increase, this limit being reached when the armature m.m.f., B neutralizes the field excitation due to //; for in that case there would be no e.m.f. and current in the bb It follows, thereaxis, hence no e.m.f. in the main brush axis.
brushes

fore,
369
that beyond a certain speed the machine will deliver a
Any desired limit to the current practically constant current. may be set by adjusting the rheostat in the field circuit //. The
generator
may
be driven either by a belt from the car axle or

directly
by mounting the armature
on the axle
itself.
the basis of the foregoing qualitative study of the physical phenomena occurring in the machine, Messrs. Kuhlman and
On
Hahnemann have developed

1
the quantitative relations between the speed and current output of the machine operating as a generator.
Thus
let
n = speed
if
of the
armature in r.p.m.
= constant exciting current in field winding // = main current output ia = short-circuit current in axis bb it n fif = ampere-turns due to // n ai a = effective ampere-turns of armature in axis BB n b ib = effective ampere-turns of armature in axis bb E = terminal voltage of line, assumed constant Eb = e.m.f. generated in short-circuit bb <# = field flux in axis BB = field flux in axis bb 3>& = ra armature resistance (including brushes)
t
If
flux
may
it,
saturation of the magnetic circuit is neglected, so that the be considered to be proportional to the m.m.f. that pro-
duces
the following relations will hold
Eb = c^ B n
$B = $6 = E =
t
(1) (2)
c z (n/if C 3 n6
na ia )
ia r a
(3)
i
there results
(4)
Ci$b n
(4),
(5)
Substituting (2) and (3) in

$&
1
CiC 2 c 3
i
(n/if
n a i a )n
1005, p. 525.
(6)
Elektrotechnisdie Zeitschrift, Vol.

24
XXVI,
370

this value of
2
and substituting
t
3?b
in (5)
E =
and
Ci c 2 c 3
-i
(nf if
n a i a )n 2
n f if ~ na
J.
i
i a ra
(7)
E
ra
-E
ra -f c^n a n
2
_1
2
JL^
(>
where
C4
Ci C 2 C 3
9 2
I
II.
From
(a)
If
equation
(8)
the following conclusions

TT
-
may
be drawn:
0, i a
~,
i
which means, simply, that were it not for the aluminum cell C the armature, at standstill, would be a dead short-circuit on the line
(or battery), the negative sign of into the armature from the line.
If
ia
indicating a flow of current
(6)
n =
oo
}
ia
= -Tl a
or
n a ia
= nf if
which means that at infinite speed the armature m.m.f. (M B ) would exactly neutralize the field excitation. This condition therefore determines the limiting current output of the machine
running as a generator, or
l (i ) \ 'ajmax
- --fl>a
This result also shows
how the rheostat in

.
the
field circuit
controls
(ia)max by fixing the value of if (c) If the machine is to act as a generator,
ia
must be
positive,
hence
_
or
E ^ =
\
~
c^ria
o
c4n
n2
-
Et
n2
But the term
a K TO. (ia)max - -
~^
is
the ratio of the
maximum ohmic
drop in
371
the armature to the line voltage, and since this ratio must be
small from considerations of efficiency,
still
it
follows that
5 is
speed
smaller in comparison with unity and that with increasing it rapidly approaches zero. Therefore the denominator of
ia
(8) may be considered equal to unity and the expression for becomes, with only slight error,
la
= Mmax ~
~
pi
-i
'~ 2
is
(9)
Equations
(8)
and
(9)
show that the current
zero
when
(10)
ftf\\
H =
and that
increases.
it
HQ
\l
-- -I^t
r-r-r
\c 4 na ^ a ) m ax
(ia } max
*/
Et -
\c4nt
rapidly approaches
as a limit as the speed
For example, suppose that the generator is to supply a maxi50 amperes at a terminal voltage of 50 volts and that it is to pick up its load at a speed of 300 r.p.m. From (10),
mum current of
'
300
50
c 4 na 1
50
and from
(9)
.
4.5
10 6
This
is
is
the equation of the curve
shown
in Fig. 302.
The manner
path
66,
of variation of 4, the current in the short-circuited
determined by combining equations
(2), (3)
and
(9), resulting in
the expression
(il)
which represents an equilateral hyperbola. It is seen that 4 is a function of c 3 and this is dependent upon the reluctance in the path of the cross-field <&. The curve showing 4 in Fig. 302 is based on the assumption that 4 = 60 amperes when n = 300, or
4 =
The dashed portions
of
the curves of Fig. 302
372
computed from equations (9) and (11), correspond to negative values of ia (indicating motor action), and while not entirely
accurate because of the neglect of terms involving ra in the pres100
90
80
70
60
&
*
50
40
30
20
10
2
000
800
1000
1200
1400
Uevoluti ns per
77
Min
-20
-40
-GO
-80
-100
FIG. 302.
Relation between current and speed, Rosenberg generator.
ence of small values of n, they depart only slightly from the correct curves within the range shown in the diagram. An examination of Fig. 301 will show that in two of the quadrants of the armature winding the currents i a and 4 flow in the same direction in the conductors, and in the other two quadrants

they flow in opposite directions.
case
is
.
373
former
The
Et
'
total current in the
_!_
2
Et
_
"
1.
\â)max
and reaches a
maximum
d(4
value
when
1
i b)
2E
-1 = o
or
when n =
~^*
Substituting the values used above, the
speed corresponding to this maximum current in the conductors is 500 r.p.m., and the currents themselves are i a = 32 and ib = 36. The 208. Operation of Rosenberg Machine as a Motor.
Rosenberg machine when supplied with current from an external
O)
FIG. 303.
Relation between currents and fluxes in Rosenberg machine operating as motor.
source will operate as a motor, but it has no torque at standstill. The reason for this absence of starting torque is clear from Fig. 301, for the armature current and the field flux due to // have
their axes in the
same
direction
and cannot, therefore, react upon
each other; and there can be no cross-field to react upon the armature current until rotation through $ B produces current and But if the armature is given a start in either flux in the bb axis. direction it will continue to run in that direction. Thus, let Fig. 303 represent the same machine shown in Fig. 301 but taking current from, instead of supplying it to, the line, and let the starting impulse be in the clockwise direction.
The
figure
is
drawn
in
374
two parts in order to show with greater clearness the effect of the two pairs of brushes; the arrows on the armature conductors of part (a) show the direction of flow of ia and those in (6) serve
,
similarly for
ib .
The
direction of
is
determined by applying
Fleming's right-hand rule for generator action; had the initial rotation been counter-clockwise, the direction of 4, and therefore also of $6, would have been opposite to that shown. In
either case the reaction
between <& and

initial
ia
produces a torque
in
the same direction as the
accelerate the armature; and between $B and 4 always opposes the rotation. The resultant torque is the difference between these two opposing torques. Analytically, the characteristics of this motor when supplied from constant potential mains are involved in the equations derived in the preceding article for the case of the generator.
rotation and, therefore, serves, to the torque due to the reaction
All that is necessary is to interpret negative values of i a in those equations as current input to the motor, but some care should be used in applying equations (9) and (11), especially at low
speeds, because the term involving r a in equation (8) is not then negligible as has been assumed. Thus, if r a = 0.10 ohm,
corresponding to a maximum armature drop of 10 per cent, when the constants are those used in the foregoing discussion, the
standstill current
value from
(8) is
whereas the true computed from (9) is i a = ia = The range of speed through which 500.
,
motor action occurs
is
from n
to
^ n = \ \ CÛf
I~~P
Without going into further particulars it will be clear Fig. 302). that the speed characteristic is similar to that of a cumulative
compound motor.
modification of the Rosenberg type of generator, together with a special method of voltage control, developed by Inthe Electric Storage Battery Co., is illustrated in Fig. 304. stead of connecting the shunt field winding across the machine
209.
it is connected between opposite points Wheatstone bridge circuit (marked " control bridge" in Fig. " series There is also a compensating winding, marked 304). reaction the for armature the of field," neutralizing purpose due to the main generator current. Two of the bridge arms, marked R, consist of ordinary resistors, while the other two,
terminals, as in Fig. 301,

of a

marked IR, have negative temperature
tion points of the bridge not
coefficients.
375
The
junc-
connected to the control
field are
connected directly across the line. A variation of generator voltage will then alter the difference of potential between the control field terminals, and the field current will change to a sufextent to readjust the generator voltage. If it is desired to give the battery an overcharge, the overcharge switch, which normally short-circuits the resistance R', is opened, the
ficient
lamp
having been previously disconnected. This has the reducing the voltage impressed on the bridge, in the same manner as though the generator voltage had itself decreased,
circuit
effect of
Auxiliary Switch
GOAmp.Fuse
FIG. 304.
Rosenberg generator and control Battery Co.
circuits,
Electric
Storage
hence the readjustment of bridge currents increases the main The extent of field excitation and raises the terminal voltage. the rise of voltage and, therefore, the magnitude of the charging
It will current, is determined by the value of the resistance R' be observed that the closing of the lamp circuit through the triple'.
pole switch short-circuits R', thereby reducing the generator voltage to the normal lamp voltage and preventing damage to
the lamps because of high voltage during charging. It follows, therefore, that overcharging of the batteries must be accomplished
during daylight runs. The automatic switch, in addition to the usual shunt and series and the coils, has a third coil connected between the generator
376
The pull due to the main shunt coil is insufficient to battery. close the switch, or to keep it closed without the pull due either to the auxiliary coil or the series coil. The auxiliary coil, therefore,
determines the closing of the switch by the difference be-
tween the voltages of generator and battery; and the switch will open when the current in the series coil drops to zero. 210. The Wagner Automobile Lighting Generator. 1 The principle of the utilization of armature reaction embodied in the Rosenberg generator is also used in the Wagner automobile The lighting generator, though in a quite different manner.
FIG. 305.
Diagram
of connections of
Wagner automobile lighting generator.
Fig. 305 (except for the omitted in recent models) are identical with those disclosed in an English patent (No. 9364) issued to W. B. Sayers in 1896. In this machine there is an auxiliary brush b
connections of the
Wagner machine,
series field winding,
the main brushes, and the shunt field connected between this auxiliary brush and one of the main brushes. The Sayers generator, intended for constant speed operation, was driven in such a direction that the portion of the armature winding included between the terminals of the shunt placed
midway between
is
winding
field
1
winding occupied the trailing half of the pole faces; in conII, Pt. I,
Langsdorf, Washington University Studies, Vol.
No.
1,
July,
1914.
377
sequence of armature reaction under load conditions, the total flux under the trailing half of the pole face increased, and the e.m.f. generated in 'that part of the armature winding included
between the terminals
of the
shunt
field
winding increased coraction.

is
respondingly, thereby giving rise to a compounding But in the Wagner machine the direction of rotation
such
that the shunt winding is connected across that part of the armature winding lying under the leading half of the pole face, as shown in Fig. 305, and moreover the machine is intended for
variable instead
directly
of
from the engine
constant-speed operation, as The of the automobile.
it
is
driven
Wagner ma-
chine presents a
number
of features of considerable technical
interest; accordingly, there is given below a discussion of the theory of its operation and a derivation of its characteristics.
As originally constructed, the machine had four poles and an armature wound with two distinct two-circuit windings, each provided with its own commutator, the two windings being connected in series. The field winding consisted of shunt and series coils connected differentially, the magnetizing effect of The series the shunt winding being the greater of the two. winding was connected in the main circuit in the usual manner, but the shunt winding, instead of being connected across the main brushes was connected between one of the main brushes (Bi, Fig. 305) and an auxiliary brush b placed midway between the main brushes and ahead of brush BI with respect to the direcIn other words, the shunt winding tapped tion of rotation. that part of the armature winding lying under the leading halves The diagram of connections shows the machine of the poles.
reduced to an equivalent two-pole model for the sake of simplicity. In the form here described, the machine was also used as a
for cranking the engine, current for this purpose being taken from the battery. In later models, however, the in a practice of combining both motor and generator functions
series
motor
single unit has
been discontinued, and the generator design has
been modified by the omission of the series field winding, experience having shown that its effects contributed little or nothing The analytical theory presented below has been of value. worked out on the assumption that the series winding is present, but it is interesting to note that the form of the equations indi-
378

minor
effects,
cates that this winding exerts only with the results of tests.
thus checking
tend to
Referring to Fig. 305, it will be clear that the connections will make the machine regulate for constant current without regard to change of speed, provided the battery voltage remains substantially constant, as is the case when lead batteries are
For if the machine is delivering current at some given value of speed, an increase of speed will tend to increase both the generated e.m.f. and the current; but the increased current
used.
will
weaken the
field
in
and, therefore, reduce the generated e.m.f. two ways: (1) by increasing the magforce
of
(2)
netomotive
the
decompounding
shifting the flux, because of increased cross-magnetizing acseries winding,
and
by
tion of the armature,
away from
the lead-
ing pole tip, thereby reducing the e.m.f. generated in the armature between brushes
FIG.
306.
netizmg
effect
-Demagdue to
BI and 6 and consequently weakening the shunt excitation. It is evident that this double demagnetizing effect will ultimately
it
short-circuited coil.
prevent a further increase of current, and is actually found that beyond a certain
speed the effect of rising speed is to cause the current to fall off from a maximum value. There is a third and most important
effect that arises
from the fact that the auxiliary brush b shortan element of the armature winding that lies opposite tha middle of the pole face and in which there is generated an active
circuits
e.m.f.; this e.m.f.
produces a considerable current in the short-
circuited element, and the direction of this current is such that it sets up an additional demagnetizing action in the air-gap
under the leading half of the pole face (see Fig. 306). An examination of the diagram of connections given in Fig. 305 will show that the operation of this machine is not independent
of the direction of rotation, for a reversal of the direction of rotation will cause the shunt winding to subtend the trailing half of the poles
and radically
is
This feature
not
it
alter the operating characteristics. objectionable in the case of automobile lightif the machine were to be used would be necessary to add to the equipment
ing for reasons stated above; but

for train lighting
379
an automatic switch so arranged that a reversal of rotation would simultaneously reverse the terminals of the series winding (if present) and change one of the shunt terminals from brush B\ to brush BZ.
ANALYTICAL THEORY
The
sis
following symbols recur frequently throughout the analyand are tabulated below for convenient reference:
t
E = Ea =
= i = n = r = = r/ ra =
i
8
a
constant line voltage

total e.m.f. generated in the line current
armature
shunt
field
current
speed in r.p.m. resistance of shunt field
winding
resistance of series field winding armature resistance measured between
main brushes and
= = series field turns per pair of poles Z = number of armature conductors $> = flux per pole d = diameter of armature = length of core T = pole-pitch = air-gap = ratio of pole arc to pole-pitch = number of poles p a = number of parallel paths through Bg = flux density in air-gap.
n,
including brush contact resistance shunt field turns per pair of poles
nf
I'
<5
\l/
armature
the line current
be clear from the diagram of connections, Fig. 305, that i shown entering brush B 2 can be thought of as dividing equally between the two paths leading to brush BI, and that the shunt field current entering brush b may be considered
It will
to divide into
two parts, Yi s and %>' s respectively, these two currents being inversely proportional to the resistances of the paths through which they flow. It follows, then, that the
,
currents
ii, i z
and
i 3)
indicated in Fig. 305, are, respectively,

ii
l = Ai
l Ais
(12)
380
i flowing through the armature winding a cross-magnetizing magnetomotive force distributed produces over the armature periphery in the manner shown by linearly
The main current
the sloping line aba' of Fig. 307. Under the pole faces this m.m.f. will produce a transverse field whose intensity at any
be taken to be proportional to the m.m.f. at that between the poles, because of the high reluctance of but point; the magnetic circuit in that region, the flux will be much less than proportional to the m.m.f., and will have a distribution
point
may
represented by the saddle-shaped curve. The resulting demagnetizing effect in the leading half of the pole face, between
brushes b and BI, will then be represented to a sufficient degree of approximation by the cross-hatched area A, Fig. 307. Con-
a'
FIG. 307.
Distribution of m.m.f. due to current
i.
line P,
sidering the closed magnetic circuit indicated by the dashed which is drawn so as to include all of the armature con-
ductors under a pole face, the m.m.f. acting upon this circuit
will
be
'o
>
na rf
of which, or 17;-^-^-, will be
consumed
in
the air-gap at each pole tip (assuming that the reluctance of the iron part of the circuit is negligible in comparison with that of the air-gap). If the length of the air-gap, corrected to take
account of the effect of the
slots,
cm., the field inten-
sity at the pole tip will be given
and the demagnetby ^7: oo 1U ^ ^-r,
izing flux represented
by the hatched area
is
2'
10
10 326
(15)

The
current
381
%i
flowing in the
j conductors between brushes
6 and BI results in a peripheral distribution of m.m.f. shown by the trapezoidal-shaped figure of Fig. 308. This fact will be made clear when it is remembered that each conductor in this particu-
lar belt of the
armature winding has a return conductor in the
opposite quadrant, as indicated in Fig. 309, this being a consequence of the fact that the winding has two layers. Under the leading half of the pole face between brushes b and BI, the effect
X'2
N
FIG. 308.
Distribution of m.m.f. due to current
it .
of this current
is,
on the whole, a magnetizing one.

current
at a
due to
1 47T o 77;
this
belt of
point
field
opposite
The m.m.f. brush b is
3
A
is >
anc*
^ ne
corresponding
intensity at that point
is
Tyr qnrjj.
The magnetizing
is
flux represented
by the hatched
area in Fig. 308
then given by
Ir47r
'
3Zis
'
r_
4ir3Zi s
2L10
3~26~
10 ^26^
47r
3Zi s rl
(16)
10 2566
Consider now the effect of the current Yi8 flowing from brush brush BI by way of the long path through conductors. Again bearing in mind that the actual space distribution of this belt of current is as shown in Fig. 310, it will be clear that the currents in the two layers of the first and third quadrants neu6 to
%Z
tralize so far as
magnetic
effects are concerned, leaving for con-
sideration
only those in
the
Comparing
this distribution
second and fourth quadrants. with that of Fig. 300, it becomes
evident that the distribution of transverse flux due to

identical' with that
Yi
&
is
due to
%i
s)
but that the total
effect is
only
382
one-third as great as that due to the latter. Consequently the total magnetization contributed by the currents %ia and %i t is
<t>
m = 3'TO
4?r
[1
(2^
I)
4rr
ZisTl^
To"
-l)*] (17)
The
face
is
net reduction of the flux in the leading half of the pole then, so far as armature reaction is concerned, given by
it is
the difference between (16) and (17), or
10 325
(18)
FIG. 309.
FIGS. 309
FIG. 310.
and 310.
0.7, (18)
Distribution of components of exciting current.
Taking ^
becomes (nearly)
(19)
In addition to the demagnetizing effect caused by the trans 1 verse armature reaction, there remains to be considered the effect of the short-circuit current in the armature element under
brush
b.
This brush short-circuits elements that are moving
through the strongest part of the field, so that the current set up in them may have considerable magnitude. Since the
is opposite the center of the pole face, the field intensity at that point is not appreciably affected by the cross-magnetizing action of the main current i and not to any great extent by
brush b
the currents
%4
and
}i
a.
It
may
be assumed, therefore, that
383
fore,
the e.m.f. generated in the short-circuited element and, therethe short-circuit current as well, is directly proportional
the speed. Since this short-circuit current is always so directed that it produces a demagnetizing effect in the leading
to
half of the pole face, it follows that the total demagnetization in this half of the pole face is given by an expression of the form
Pis (ai yn), where 7 is a constant depending upon the average flux density in the air-gap, the number of turns per armature winding element, and upon the resistance of the element and the
brush contact; its value is derived in equation (40). The values of a and p follow immediately from equation (19); but since (19) was derived on the assumption that the machine was a bipolar one, whereas the actual machine is multipolar, it becomes
necessary to divide the right-hand member of (19) by the number of pairs of poles in the machine in order that the expression may be perfectly general; therefore,^ *
^
8p
(*** p =
Since the total flux per pole of the pole face is
is
<l>,
(20)
the net flux in the leading half

pi,
Y^ ai
ai
- yn
The
and
e.m.f.
generated in the armature between brushes BI
b is
then
(%<
where Z'
Pi,
yn)Z'n
OU Xr^n
\/ ins' /\ JLU
P being the number of poles in the

of parallel paths through the
machine and a the number
arma-
This is the e.m.f. responsible for the production ture winding. of the shunt field current z a hence
;
Q^$ -
ai
Pis
yn) Z'n
i s ra
~^~
Substituting in (21) the value of
there
is
i3
from
(14)
and solving
for
itt
obtained (noting that
(22)
384

useful flux
3> is
The
produced by the
differential action of the
shunt and
series field windings,
neglected, the flux

excitation; that
is,
may
magnetic saturation is be taken as proportional to the net

if
and
$ =
and the
Ci(n a i8
is
n fi)
(23)
total generated e.m.f.
Ea =
where
3>Z'n
cn(n 8 i s
n f i)
(24)
(25)
Since the total generated e.m.f. must be equal to the sum of the terminal voltage and the ohmic drops in the series field winding and in the armature, there is the further relation that
Ea = E +
t
irf
i,(2r a }
= E
i(r a
rf )
+^
i.
(26)
the final form of (26) being obtained by substituting the value

of
ii
from equation
(12).
Substituting (22) first in (24) and then in (26), and reducing, there result, respectively,
and
E4 l
-o) = E>+
'+ r
'-MaZ>n+ %)]- jnr

+ %r a
z/nl (28)
where
R =
ra
pZ'n
i,
(29)
Dividing (27) by (28) and solving for
it is
found that
*+yZ'n* + E
a
+
(ra +r,)j8Z'
7
aZ']
+ (ra+r^ + y^}- ^
(30)
Equation (30) represents a .curve which shows the relation between the current output i and the speed n. The form of
385
this curve may be found by noting that i is the quotient of two functions of n, one of which (the numerator) is
(31)
and the other (the denominator)

yd
is
= n
(an s
&nf )cZ' +n\cn f (r s

T
(ra
+ rf )(3Z' -
a Z'
(r a
rf ) (r s
(32)
The function y nj since it is a cubic equation, represented by a curve of the form y n or y^,
will in general
be
Fig. 311,
and the
FIG. 311.
Curves representing equations
(31) arid (32).
function yd will be represented by a parabola with vertical axis, as shown by the curves y d and y d Fig. 311. The" shape of the curve corresponding to (31) will be that of
,
are all real; and will be that curve y n if the three roots of y n = The curves of curve y n if two of the three roots are imaginary. of Fig. 311 are not drawn to any particular scale, but are merely But whatever the roots of y n = 0, illustrative of typical forms.
386
the curves will cross the axis of ordinates at a distance below the
origin given
.
by
(rs
+ %r
a ).
Similarly, the intersections with the axis of
of the
parabola
represented by (32) can be found

case
by putting yd =
0, in
which
where
m=
q
s
(an s
cnf
(ra
= =
(r
|)
-^
)
(r
r,) (r,
+ %r if
r^2
-j-
The two
roots will be real
> 4ms,
in
which case the parabola
two points both of which lie to the left of the origin, as in curve y" dj Fig. 311; and the roots will both be 2 if # <4ws. to. curve y' It is very corresponding d imaginary,
will cut the axis of
in
shown, on inserting the actual values of the constants comprising m, q and s, that the latter condition is the only one that can arise in practice, which means that the parabola representing the function yd does not intersect the axis of abscissas.
easily
In any case,
o
it
intersects the axis of ordinates at a distance

r
2
~|
r,) (ra -f
fr a )
-^J
above the
origin.
The curve showing the relation between i and n can then be obtained by dividing the ordinates of the cubic curve by the
ordinates of the parabola. Obviously, if the to operate satisfactorily as a generator, the curves representing y n and y d must be so related that their ordinates
corresponding
is
machine
will
both be positive for a large range of speed (in the positive sense), which means that the function y n must have three real roots as in curve y' n and that point A (Fig. 311) must be as close Inas possible to the origin and point B as remote as possible.
,
spection of equation (31) will suffice to show that the constant n 7 is the controlling factor in determining the shape of curve y'
which it intersects the axis of n. For which would mean physically that there 0, is no demagnetizing effect due to the winding element shortcircuited by brush 6, or in other words, that there is no current
also the points in
and
suppose that 7

in
387
then
this
short-circuited
element; the numerator of (30)
reduces to y
'
- E
(r8
(34)
which represents a straight line shown as L in Fig. 311. In this case the curve obtained by dividing the ordinates of the straight line L by the corresponding ordinates of the parabola would
have the form shown
in Fig. 312 within the range of generator On the action; it is asymptotic to the axis of n at infinity. other hand, when 7 > 0, the curve showing the relation between
and n will have the form of Fig. 313. Since in practice 7 will always be greater than zero, it is important to investigate this case further, and in particular to determine the condition that
i
will give rise to will readily
be seen that
three real roots of the function y n = 0; for it if two of the roots are imaginary, as in
FIG. 312.
FIG. 313.
FIGS. 312
and 313.
Relation between current output and speed.
curve y'n, Fig. 311, the current through the machine will be negative for positive values of speed, which would mean that the machine would run as a motor and not as a generator.
Proceeding, then, to the analysis of equation (31), seen that
dy,
it is
easily
dn
Z'ri2
ra yZ'n
^1
If
= -[ScnsJ Z'n -
r aJ Z'}
dy,
'
is
put equal to zero,
it is
found that the condition which
determines the position of the points of of curves y n or y n is
maximum and minimum
=
Gen,
388
which means that one of these points lies to the right of the There is a point origin and the other to the left of the origin.
of inflection in the curve
where -~%
0,
that
is,
when n =
points of maximum and minimum. that the three roots of y n = shall be real, the function can be written
or
midway between the To find the condition
which, on substituting
.
.'
-x +
becomes
^
--} cnj
1 J
.:..-,,;
(36)
+ !2cV\ x - -( + cn yl l^^J^ + r Z' 6\2Z'

L
7 VIZ'
s
en)
~^ 108c n 8 3
3
=0
(37)
This equation
is
of the
form
X3
%X
-frj
=
3
which
will
have three
2
real roots
if
the absolute value of |= 27
is
greater than
j-.
In this particular case
and
(38)
t
cn s yl
Z'
6\2Z'
'
cn s /J
108c 3 n s 3
To reduce the values of and 77 to simpler and more manageable form, it is necessary to evaluate the characteristic constants that enter into their expressions. Thus, from (23)
flux per pole
Bg^/rl
\fsrl
amp-turns per pair of poles
1.65^5
1.6&5
where k
is the ratio of the excitation required for the complete magnetic circuit to that required by the double air-gap. Unless extremely high flux densities are used, k will be somewhere be-

tween 1.2 and 1.4, with an average value of about from (25) that
c
389
1.25.
It
follows
CiZ
'
'
~
8
(39)
The magnitude
of the constant
is
eration that the demagnetizing flux
cf>
determined by the considcaused by the current in
the element short-circuited

If
by brush
b is given
by yn,
.
or 7
the e.m.f. in the short-circuited element

is
,
is e s
c .,
the average
current in the element will be

resistance of the short-circuit,
47T
c>
where
r is the
apparent
and
'
_Z_
10 2S
25
ls c
-
4ir
Z trle
s.c.
10 2S 45
where S
is
the
number
of
commutator segments.
In practice,
^
o^
1>
as a general rule.
es
But
e,. e .
is
given by
c.
--
X X
1C 8
"
hence
=
n
2400
c
10 8
W V
in (38), it will
Substituting these values of

2
and 7
be found that
terms involving r a and negligibly small in comparison r in with the others. Dropping a comparison with r s the values
r a 3 are
,
of
and
17
become
12
dr /S\* /I 10 10 E Etdr (\* / ir,*Z'BT*l* \Z) \
+
I
1.09 kZ\
)
<n
38.4/c
10 10
rs 8 2 r
(41)
/S\"
I
Substituting these values in the criterion
l
1
27
The
factor
^
4
arises from the fact that in a series winding a single brush

series.
short-circuits
elements in
390

it
and reducing,
follows that
k'
2
10.85
rs 2 5B g ap
1.09/cZ\
\8
(42)
The
physical meaning of the last result
is
that unless the resist-
ance of the short-circuited element, including brush contact resistance, is greater than a definite limit, the demagnetizing effect of the short-circuit current will so far weaken the field
at the leading pole tip as to prevent the machine from building up; that is, an extremely small value of r would result in the
machine taking current from the

all
line, or running as a motor at In practice, the contact resistance between the carbon brush b and the commutator will be sufficiently high to insure that r will exceed the critical value, which
positive values of speed.
means that the characteristic curve of Fig. 313 will always cut the axis of n in two points lying to the right of the origin and in one
point to the left of the origin; the latter point has no practical significance, since it relates to the condition of running back-
ward. The three points of intersection with the axis of n can be located by finding the three roots of (37) and then substitutSince the coefficients of (37) are ing for x its value from (36).
3
necessarily so related that
^>
~T,
equation (37)
is
the irreducible
case of Cardan's rule, and the solution must be found onometric methods. The three roots of (37) are 1
by
trig-
sin
(60(43)
sin (60
where
sin
le
V27
1
Chauvenet's Plane and Spherical Trigonometry, pp. 99-100.
391
of Fig.
hence the abscissas of the three points of intersection of the curve 313 (only two of them are shown) are
= -^Qcn s
r,.
x 3 \/f
sin
ye
3
(60
- HO)
(44)
Substituting the values of
and
17
from
(41) in the expression
for sin 6 given in equation (43),

3.28/b
sin 8
rs 't!-'*i
a5B g p
i
no^y\
/^\
It
has been previously pointed out that satisfactory operation
and of Fig. 311 shall of the generator requires that the points be separated as widely as possible; and it is also important that
point
(corresponding to the speed n\) shall be as close to the

It follows, therefore, that sin 1^0,
origin as possible in order to secure the
pick-up -speed.
also sin
6, shall be small. Inspection of that the conditions to be satisfied to meet this requirement are
advantage of a low and therefore equation (45) shows
as follows:
1. Make the resistance r as large as possible, that is, make the brush b of hard carbon with the smallest practicable dimenThe desirability of this design is involved in the discussions.
sion of Fig. 312.

2.
Design the armature winding with one turn per element
-)
3.
Design the shunt

(
field
winding so as to have a low
resist-
ance per turn

4.
\n s
small
Use a simple two-circuit armature winding
(a
2).
5. Make the air-gap as small as possible. 6. Select a moderate value of flux density in the air-gap and make Z correspondingly large. In other words, the armature
should be magnetically powerful.
392

note that for moderate speeds the cubic 311 does not deviate greatly from the straight For the slope of the line L, as determined by (34), has the
It is interesting to
curve
line L.
y' n of Fig.
constant value
(-^
&Z
and that
of the cubic curve
is
given by
yZ'(2cn.n*
hence,
r.n),
when the speed n is low, as in the vicinity of the pick-up speed, these two slopes will not differ greatly, and will approach This fact may be used to deterequality as 7 approaches zero.
will give rise to the maxicurrent output of the machine. It is possible, of course, to determine this condition accurately by differentiating equation (30) with respect to n and equating the resulting expression
mine approximately the condition that
mum
to zero; but this procedure develops excessive complications for the reason that the differentiation gives an equation of the fourth
degree in n.
If, therefore, equation (30) is simplified by dropping terms containing 7, the approximate expression for i be-
comes
ran 2
+
=
qn
Putting in the condition that
0, it is
found that
=
CUs --
r__ ~
R7> + pZ
_L_
/.
CUs
V-g-
+
_1_
R7> / } pz
/.
m ( CU
\-g
Inspection of (46) will show that
is
the approximate value of the pick-up speed of the machine. Furthermore, the constants q and s are very small compared
ra,
with
so that the terms

ra
and
ra
in
(47) are
negligible.
It
follows, therefore, that the
maximum
current output will occur
BOLSTERS AND BALANCERS

at a speed
393
greater than twice the pick-up speed; this checked by test results. It has been previously pointed out that experience has shown that the series field winding can be omitted without materially The analytical reason affecting the operation of the machine. It will be for this fact is clear from the form of equation (30).
somewhat
conclusion
is
observed that the factors rf and n/, respectively the resistance and the number of turns per pair of poles of the series winding, appear only in the denominator, that is, in equation (32), which
is
represented graphically by the parabola of Fig. 311; since
= 2n(n. - W)cZ>
cn,r
it is
clear that within the

,
working range of speed
(to the right
the principal effect of r/ and n f is to alter the slope of the parabola, but only to a small extent since in any case
of the origin)
their numerical values are relatively small.
INDEX
(Numbers
refer to pages)
Armature, characteristic
circuits,
number
of
167 98
Abampere, Abcoulomb,
definition
of.... 11,
17
12
coils
105, 268, 295,
definition of
Abvolt, definition of Acceleration of trains

Acyclic generator Air ducts, armature
-gap, area
12
conductors, number of .... core construction

excitation
300 96 62
128
227 72
disk....;
63, 89
63 120 ampere-turns 119, 120 142 chamfering required by flux density 65

.
distribution
145, 147, 152, 155 121 fringing correction
All-day efficiency Alternating-current machines..

e.m.f. rectification
Alternator
generated e.m.f
Altitude, effect
rise
325 44 52 44 46, 50
.
88 62 e.m.f. generated in 56 145 field, shape of 283 flux 340 heating 133 magnetizing action of reaction 133, 137, 205, 286 361, 366 regulation by resistance 57, 305 \ 88 ring
drum
eddy currents
winding see Winding Automobile lighting generator

366, 376
Axis, of
on temperature
329 328
15
commutation
137, 241
Ambient temperature Ampere, definition of

Ampere-conductors Ampere-turns
air-gap
geometrical neutral neutral
armature core
armature, per pole cross-magnetizing demagnetizing field, per pole
interpole pole cores
146 25 119, 120 128 142

138, 141 138, 141
Axle lighting system
133 134 359
B
Back ampere-turns
e.m.f
138, 141
Balancer
191,
199 358
142 300
128 124 129 44 142
Battery ampere-hour capacity

350, 351
and shoes
discharge rate
350, 351
teeth
end
cells
yoke Armature
amp-turns per pole
train
equipment
350 360
349, 351
voltage range Bearing friction 395
314
396
Bijur system, booster control..
.
INDEX
357
10
Characteristics, no-load, 113, 115, 116
Biot-Savart, law of
Rosenberg machine..
tor
.369,
373
162
Bipolar machines
59
separately excited genera-
138 demagnetizing amp-turns. 138 347 Booster 354 auxiliary control 357 Bijur system of control.. constant current 350 differential 353 Entz control system 355 Hubbard control system. 354 motor capacity for shunt 350 348 negative non-reversible 350 reversible 353
.
cross-magnetizing turns
amp-
motor
series
204
168 209
173
generator
motor
shunt generator
motor
Chord windings
demagnetizing
208 101, 262, 293

effect of. ...
143
Coefficient of coupling 37 dispersion or leakage. .117, 129
273 335 self-induction, 33, 268, 295, 300 105 Coil, armature
36,
mutual-induction
output
series
shunt
track return Bridge control
Brushes'.
66,
347 348 348

231 292
dummy
.
.
102
.
mutual inductance of 273 100 numbering of order of commutation. 259

self -inductance, 268,
current density
67, 249, 254,
field
293
heating of
field intensity
295,300 79 343
. . .
contact resistance
240, 247
on
axis of
16
drop of potential at
163, 293, 306, 323
friction loss
holders
lead
315 66
137
pressure
sets required 'wide, effect of
67 93 264
width
Building up of generators
294
75, 76, 83, 171
Capacity of machines
Carter correction factor
criterion for
326
122
.
commutation
267
161
Characteristics
armature
compound generator
motor
load
. .
168 181 211
20 290 Commutating devices e.m.f 242, 244, 246, 249, 276, 293 field 142, 280, 295, 297 poles, 71, 279, 297, 298, 300, 301 Commutation 135, 237 256, 264 adjacent coils axis of 137, 241 criterion 246, 267 286 improvement of linear 238, 243, 248 over238, 243, 246, 262 237, 263 period of 240 resistance selective, in wave windings, 263 sinusoidal 238, 248 259 successive phases under-.... 239, 243, 256 240 voltage without auxiliary devices.. 292 262 Commutator blackening
19,
.
magnetic potential due to
.166
construction
60, 71
INDEX
Commutator diameter
friction loss
397
magnetizing ampere:
.
61
Cross-field,
315 'heating 61, 329, 334, 345 62 peripheral velocity 97 pitch segments, current density 250 number of 96 voltage betweeen 61, 295 287 Compensating devices Compensation of armature reaction 286 77 Compound excitation
.
turns
.138, 141
.
Cumulative compound motor.

Current, absolute unit of continuous
density in brushes
211
17
11,
45 293 44
67, 249, 254,
at
commutator segment 250

due
to.
.
.
direct
direction of force
12
generator characteristics.
generators in parallel
..
181
eddy or Foucault force due to heating due to

induced
62 10
15
motor
characteristics
effect
188 211
inter-
Compounding
poles
of
15 practical unit of. short-circuit 238, 282, 382
301
Conductors, number of
ture
arma96 350
D
Demagnetizing action, 137, 378, 382 ampere-turns 138, 141
corrected
Constant current booster
171 regulation Contact resistance of brushes 240, 247
143
of
Contactors Continuous current

rating of motors Control of boosters
bridge
series-parallel
shunt motor speed.

Controllers
Cooling curves Core, armature
233 45 326, 332 354 231 227 217 229 336

62 128 123
component
Deri winding
cross-mag-
netization
148
Diamagnetic substances
Difference
of
289 1, 25
elec-
potential,
trical
12
magnetic
Differential booster
compound motor
motors, starting of Direct current
ampere-turns for
corrected length heating of
machine
Discharge rate of battery Disk armature
.
339 losses, 304, 306, 314, 315, 345 65, 296 pole 128, 300 ampere-turns for 3 Coulomb, law of Counter e.m.f 199, 214 37 Coupling, coefficient of 246, 267 Criterion, sparking 368 Cross-field
magnetizing action of
137, 378,
Dispersion, coefficient of. Distribution of flux in air-gap
7 353 211 215 44 44 350 63, 89 .117, 129
147, 152, 155
Division of load, generators
188
motors
235
.89,
Drum
380
148
demagnetizing
nent..
compo.
107 102 Duration of short-circuit 237, 263 43 Dynamo 58 construction of 59 bipolar and multipolar ....
windings
coils
Dummy
398
Dynamo,
field excitation
INDEX
73 68
Elements, number of
Dynainotor
End
cells
96 350
105
connections, form of
E
62 66 351 loss, armature conductors core 309, 311 313 pole faces teeth 312 Efficiency 304, 317 325 all-day commercial or true 317 of conversion 317 electrical 318 maximum 321, 324 mechanical 318 321 variation with load Electric Power Club 334
Eddy
currents, armature core pole faces
343 leakage flux 268, 271 Energy stored in magnetic field, 37 Entz system of booster control, 355 187 Equalizer 107 Equipotential connections
heating
lines
and surfaces
of
8
see
Excitation
dynamos
Field Excitation.
Faraday's law of electromagnetic induction

8,
13
Feeders, negative Field, air-gap distribution
348
of
147, 152, 155
Electric
Storage Battery Co.,
system
of
train
374 lighting Electrical potential 9, 12 regulation of voltage, train

lighting Electricity, quantity of
361
11
Electrolysis
Electromagnetic induction Faraday's law of

Electromagnets, tractive effort
of
348 8
13
Electromotive force
of alternator
38 8 46, 50
142 ampere-turns per pole 145 armature, shape of coil heating 343 commutating, 142, 280, 295, 297 control of speed 219 cross368 displacement of windings, 98 44 dynamo 37 energy of magnetic excitation 73, 149 297 fringing 343 heating of coils
intensity due to circular coil
commutating
242, 244, 246, 249, 276, 293 of direct current generator, 56
solenoid
straight wire
at pole tips
"...
16 17
11
141
1
counter
direction of induced
199,
214
9
magnetic
pulsations of reaction
short-circuit
280
generated and induced .... magnitude of induced of mutual-induction

rectification of alternating of self-induction
48
12
34
115, 282 regulation, train lighting
currents
upon
pulsations of unit of
52 33 52, 55
12, 15
362, 364
reversal of
commutating
142,
295
81
'
rheostats
step of windings
Element
of
winding
90
100
INDEX
Field windings 75, 76, 79, 343 12 Fleming's left-hand rule 9, 32 right-hand rule. 4 Flux
399
70
light-
Generator, turbo-
Wagner automobile
density
air-gap pole cores
teeth
distribution, air-gap
24 65 65 297
366, 376 ing 21 Gilbert, definition Gould system of train lighting.. 362
Gravitational potential
II
145, 147, 152, 155, 283
Head-end system, train lighting, 359

Heat, mechanical equivalent
of,
leakage end-connection
slot
117, 129
16
268, 271
268
268, 270
tooth-tip
20 5 from magnet pole Force, magnetic, due to current. 10 direction of, due to current. 12
linkages
Foucault
currents
see
Eddy
Currents
Fractional pitch windings
101, 262, 293
Heating of armature 328, 340 commutator, 328, 329, 333, 345 curves 336 due to current.. 15 of end-connections 343 of field coils 343 of railway motors 332 326 Heat run 33 Henry, definition 72 Homopolar generator 328 Hottest-spot temperatures
.
demagnetizing
Friction loss
effect of. ...
143
Hubbard
Hysteresis
loss.
.
system
control
of
booster
314, 320
121, 123
Fringing of flux
Froelich's equation
179
354 116 306
Gauss, definition
Generated Generator
e.in.f
3 48 43
Imbedded-detectors Induced current

e.m.f.
8,
331 8 12, 48
acyclic,
homopolar or uni72
76,
83,
Inductance of armature
see Self-induction
coils
polar
building up.. .75,
171
Induction, electromagnetic
lines of
185 combined output compound, characteristics. 181

division of load
8 24
2
magnetization by
Insulation, classes of temperature limits
Intensity, field axis of coil
Lundell
parallel operation
polarity
Rosenberg
separately excited
series
shunt
stability of operation
three-wire
188 297 186 83 366, 374 162 75, 168 173 170 191
328 328, 332

3 16 17
'.
solenoid
of magnetization
39
71
Interpole machines
motors
Interpoles Iron loss..
71,
.
220 279, 297, 300 304
400
Iron
loss,
INDEX
eddy currents,
306,308,312,313 306, 308 314
.
Loss line Lundell generator.
165 297
hysteresis
total..
M
Magnetic circuit, law of examples of application..
field
. .
Joule, definition of
15
26 27
1
K
Kirchhoff's laws, 30, 108, 244, 263
distribution, air-gap. ...
uniform
flux
force, lines
145 4
4
3
L
Lap winding,
91, 100, 143, 257,
and tubes
of
due to current
263
10
direction of
10 12
117, 129
Laplace, law of Lead of brushes
leakage
potential
of coils
Leakage, magnetic
137 117 coefficient of.. 117, 129 flux, end-connections. .268, 271
pole cores shoes
slots
19,
reluctance
6 20 26
13
130
Magnetism, residual 75, 83 24 Magnetization curves of dynamos. .113, 115, 116

.
tooth-tip
268 268, 270

Fleming's
14,
by induction
intensity of Magnetizing action of armature
Left-hand
rule,
Lenz'slaw
12 32
Magnetomotive force
359 Lighting of trains Lincoln adjustable speed motor. 221

Linear commutation, 238, 243, 248 24 Lines of induction 8 equipotential
force
Magnet
flux
pole, unit
2 39 133 21 2
5
from
efficiency definition of
Maximum
322, 324
Linkages,
number
of flux
20
166 325 304
Load
characteristic
curve Losses
core or iron, armature
199 16 losses 304, 314, 315, 320 361 regulation of voltage Miscellaneous losses, 304, 315, 316
Maxwell, Mechanical characteristics, motors equivalent of heat
Motor
313
308, 312 314, 345
43, 199
306, 309, 320
pole face teeth

total
350 capacity, shunt booster compound, characteristics. 211

counter e.m.f
direction of rotation
division of load..
199, 214
copper, armature
commutator
field
mechanical
miscellaneous
stray power
305 306 305 314, 320 315

317,
enclosed, rating of
interpole
power
of
summary
of
318 317
railway heating of
reversing
85 235 345 220 199 224 332 214
INDEX
Motor, Rosenberg
separately excited
series
401
shunt
speed regulation, shunt..
pulsation, interpole
starting of
differential
.
compound.
373 204 209 208 217 303 202 215
Peripheral velocity of armature,

effect
on temperature
rise
Permeability
Permeance
Pig-tail connectors Pitch of windings
Polarity of generators Pole arc, corrected length
341 23 27 68 97 83
121
199 torque 68 Motor-generator 94 Multiplex windings Multipolar machines 45, 59 cross- and demagnetizing effect 139
.
changer
361, 364
71, 279, 297, 298,
commutating
cores,
ampere-turns
for.
construction of
flux density in
65,
300 128 296 65

131
Multi-voltage speed control..

coil
218
273 34 36
leakage
shoes, ampere-turns for
. . .
Mutual inductance, armature

induction
coefficient of
.
.
construction
eddy current
loss
X
Negative booster
feeder
leakage flux saturation unit magnet Poles, choice of number of

Potential curve.
electrical.
.
128 65 66, 313 130 296

,
.
2,
Neutral axis
geometrical Non-reversible booster
348 348 134

133
.
59 154
9,
12
6
energy
gravitational
8
6, 19,
350
Power
magnetic of motors
e.m.f.,
20
199
O
Oersted, definition
Pulsations,
commutating magnitude of
field.
280
55 303
26
10
speed, interpole motors.
discovery of
Ohmic
304 armature 305 commutator 306 field 305 335 Output coefficient 334 equation Overcommutation, 238, 243, 246, 262
losses
Q
Quantity of
electricity, unit..
. .
11
R
229 224 225 cycle of operations 332 heating 226, 332 rating 332 stand-test 326 Rating continuous and short-time 326 345 enclosed motors
Railway controllers motors
Parallel distribution ...
.159
.
operation of generators.
186
windings Paramagnetic substances Period of commutation
92, 100
1,
25
237, 263
402
Rating, railway motors.
. .
INDEX
332 293 205 286 137 regulation of voltage by, 361, 366 of short-circuit currents, 1 15, 282
.226,
Reactance voltage, 247, 248, 279, Reaction, armature 133, compensation of components of
Saturation of pole tips teeth

Sayers' winding
generator
296 297 290 376
Segments
of
commutator, num-
Rectification
of
alternating
e.m.f
52
94, 103
Reentrancy, degree of
Regulation, definition
161
by armature
curve
field
reaction. .361, 366
for constant current
171
168
362, 364
96 average voltage between, 61, 295 current density 250 Selective commutation 263 Self-excitation 74 Self -inductance 33 of armature coils.. 268, 295, 300 32 Self-induction coefficient of. .33, 268, 295, 300 e.m.f. of 33
Separate excitation Separately excited generator.
73 162
ber
resistance, train lighting.
speed, shunt motors voltage, train lighting
Regulator, lamp
Tirrill
362 217 360 362, 364

.
motor
Series booster
204 347
:
Reluctance
Resistance,
195 26
distribution
excitation
159 74
75, 168
armature winding
57,
generators, characteristics
in parallel
305 brush contact 240, 247 commutation 240 field, effect on speed. ..... 178 measurement of temperature rise by 331
regulation, train lighting
motors, applications of.

characteristics
...
core loss
shunt
windings
Series-parallel control
186 221 209 320 184
75, 92, 101
362, 364
227
159 103 259
Reversible booster
353
81
distribution
Rheostat,
field
windings
Short-circuit of adjacent coils.
.
83 203 Rheostatic control of speed. ... 217 9 Right-hand rule, Fleming's .... Ring winding 89, 91, 241, 256 Rocker ring 66 Rosenberg generator 366, 374 machine as motor 373 Rotation of motors, direction of 85
discharge
motor starting
current curves
demagnetizing effect
duration of reaction of Shunt booster
excitation
of..
238 382
237,263 115, 282 348 76

76, 173
generators
in parallel
187
motors
speed regulation windings Simplex windings Simpson's rule Sinusoidal commutation..
S
Safety Car Heating
208 217 76 94
.238,
&
Lighting
system of train lighting
364
128 248
INDEX
Slot,
403
. .
leakage flux
pitch
shape of Slotted armatures Smooth core armatures

Sparking constants
criterion..
268 97 64 120
119 17 281 246, 267
212, 214
Temperature correction
hottest-spot
rise
328, 329 330, 331
328
328, 329, 330,
Solenoid, field intensity on axis.
specifications,
332, 333, 339, 341 Electric
Power Club
time-constant-.
334 339
of
Speed characteristics, compound
Thermometer, measurement
resistance
motor
series
by
separately excited motor. 204
Thompson-Ryan winding
Three-wire generators
330 287
191
motor
shunt motor effect on of,
209 208
separately excited generator 165
Thury system 185, 235 Time constant, temperature rise, 339

Tirrill
217 227 Speed-time curves Stability of operation of generators 170 332 Stand-test, railway motors Starting of motors 202, 215 rheostats 203 Stone train lighting generator. 361
. .
on series generator on shunt generator pulsations, interpole motors regulation, shunt motor.
170 177 303
regulator Tooth-tip leakage flux
195
268, 270
199, 202 212, 214
.
Torque
compound motor
motor shunt motor Track return booster
series
separately excited motor
207 211 208 348

38 359
Tractive
effort
of
electromag-
nets
Train lighting
Tubes
of force
3,5
70
see
Storage battery see Battery system, straight Stow multi-speed motor Straight storage system Stray power loss .317,
.
359 221 359 318
Turbo-generators Two-circuit windings winding -layer windings
Wave
104
Surfaces,
characteristic
see
U
8
Characteristics.
equipotential
Susceptibility
24
.
Undercommutation 239, Uniform magnetic field

. . .
243, 256
Swinburne's commutating device..
291
Unipolar generator U. S. Lighting & Heating Co. system of train lighting
4 72
366
Teeth, ampere-turns for eddy current loss in

flux density in
hysteresis loss in
leakage flux
shape of Temperature, ambient
124 312 297 308 268, 270 64 328
Ventilating ducts
Volt, definition of
63,
344
15
Voltage commutation
240
between commutator segments 61, 295
INDEX
control of speed
limits of batteries
218
349, 351
lap
91, 257,
multiplex
263 94
98
9(:
reactance
regulation
247, 279, 293

161,
number
of circuits
360
conductors
open-coil
parallel pitch of
W
Wagner
automobile
generator
lighting
89 92
97.
reentrancy
resistance
94, 103
57,
366, 376
of speed
305
92
Ward Leonard system

control
ring
89,
91,241,256
108
218
15
in
.
series
Watt, definition of
series-parallel
Wave
windings.
.91, 101, 144,

. .
selective
commutation
Windage
Winding, armature chord or fractional pitch

closed-coil
258 263 314, 320 88
104, 143, 262, 293
89 52 89 drum 102 dummy 90 element equipotential connections 107 98 field displacement 100 step
distributed, effect of. ...
94 simplex 104 two-layer wave, 91, 101, 144, 258, 263, 264 field 75, 76, 79, 343 298 commutating poles 289 Deri 287 Thompson-Ryan
Yoke, ampere-turns for

construction
129
Young and Dunn brush
65 292
LIBRARY UNIVERSITY OF CALIFORNIA

IE
on the last date stamped
below^
LD
21-100m-l2,'46(A2012sl6)4l20
UNIVERSITY OF CALIFORNIA LIBRARY

Principles of Direct Current Machines

Uploaded by

Copyright:

Available Formats

Principles of Direct Current Machines

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

Principles of Direct Current Machines

Uploaded by

Copyright:

Available Formats

Ij

DIRECT CURRENT MACHINES

BOUVERIE STREET, LONDON,

COPYRIGHT, 1915, BY THE

McGRAW-HiLL BOOK COMPANY,

THE MAFZiE PRESS YORK PA

SARAH SUSS LANGSDORF

complete treatment of the fundamental principles that underly

for depicting the and in Chapters

solution of a problem as apparently without a due realization

PREFACE TABLE OF SYMBOLS

Lines and Tubes of Force

Units of Current and E.M.F Heating Due to a Current

18. Field Intensity 19. Field Intensity

Magnetic Circuit 25. Applications of Law of Magnetic Circuit

27. Self-induction 28.

Mutual Induction Energy Stored in a Magnetic Field

30. Tractive Effort of

Dynamo, Generator and Motor

36. Effect of Distributed

37. 38. 39.

41. Bipolar and Multipolar 42. The Commutator 43.

46. Brushes, Brush Holders and Rocker 47. Motor-generator. Dynamotor

49. Interpole Machines 50. The Unipolar or Homopolar

52. Separate Excitation

57. Construction of Field 58. Field

59. Polarity of Generators 60. Direction of Rotation of

68. General Considerations 69. 70.

73. General Rules

Rule for the Degree of Reentrancy

Two-layer Windings Examples of Drum Windings 77. Equipotential Connections

104 107 107

MAGNETIC LEAKAGE The Magnetization Curve

80. Calculation of the Magnetization 81.

82. Details of Calculation of

84. Corrected Axial

Ampere-turns 86. Ampere-turns 87. Ampere-turns 88. Ampere-turns

Length Required Required Required Required

Teeth Armature Core the Pole Cores and Pole Shoes

Magnetizing Action of Armature

Commutation Components of Armature Reaction

133 135 137 138

Expression for Demagnetizing Effect of Back Ampere'

143 145 147 148 149

102. Predetermination of Flux Distribution in the Air-gap

OPERATING CHARACTERISTICS OF GENERATORS

105. Regulation 106. Characteristic

Load Characteristic The Armature Characteristic

165 166 167

Connection of Generators for Combined Output

The Thury System

184 185 185 186

Requirements Counter E.M.F., Torque and Power The Starting of Motors

Compound Wound Motor

The Reversing Motor

202 204 208 209 211 214 215 217

137. Series-parallel Control 138. Railway Controllers

Load between Motors

Fundamental Considerations Physical Basis of the Theory

142. General 143.