a ^ ^

C*)
;\ : v- v Q L lU

a rq. \\ \■'
l ~\. lV
u
\c L\ :
 NUNC COGNOSCO EX PARTE

THOMAS J. BATA LIBRARY

TRENT UNIVERSITY
 Digitized by the Internet Archive
in 2019 with funding from
Kahle/Austin Foundation

https://archive.org/details/servomechanismsdOOOOmill
SERVOMECHANISMS
SERVOMECHANISMS
Devices and Fundamentals

RICHARD W. MILLER
State University of New York
Canton Agricultural and Technical College

RESTON PUBLISHING COMPANY, Inc., Reston, Virginia

A Prentice-Hall Company
Library of Congress Cataloging in Publication Data

Miller, Richard W
Servomechanisms : devices & fundamentals.

Bibliography* p.
1. Servomechanisms.
TJ214.M48 629.8’32 76-58013
ISBN 0-87909-760-4

© 1977 by Richard W. Miller

All rights reserved. No part of this book

may be reproduced in any way, or by any means,
without permission in writing from the author.

10 987654321

Printed in the United States of America

 This book is dedicated to my wife,
Irene, for her patience and understanding
and to all those students who participated
and contributed in one way or another to
the development of this text and laboratory
experiments over the past few years.
Contents

PREFACE, xi

1 INTRODUCTION TO SYSTEMS, 1
1.1. General, 1 1.2. Control, 2 1.3. Block diagrams, 4 1.4. Closed-loop appli¬
cations, 6 Questions and problems, 9

2 PRECISION POTENTIOMETERS, 11

2.1. Introduction, 11 2.2. Definition and schematic, 11 2.3. Mechanical con¬

struction, 12 2.4. Rotation, 13 2.5. Mathematical relations, 14 2.6. Load
resistance, 15 2.7. Electrical specifications—wire-wound potentiometers, 16
2.8. Nonlinear functions, 24 2.9. Non-wire-wound potentiometers, 24 2.10. Po¬
tentiometer connections, 26 2.11. Operation in a system, 26 2.12. Zeroing a
potentiometer, 28 Problems, 29

3 VARIABLE RELUCTANCE TRANSDUCERS, 32

3.1. Introduction, 32 3.2. Principle of operation, 32 3.3. Use in servo systems, 36

3.4. Zeroing an LVDT, 38 3.5. The Microsyn, 38

4 RESOLVERS, 41
4.1. General, 41 4.2. Four-winding resolver, 42 4.3. Electrical zero, 44
4.4. Resolution, 45 4.5. Composition, 45 4.6. Rotation of coordinate axes, 46
4.7. Data transmission—angular addition/subtraction, 51 4.8. Performance
characteristics, 54 4.9. Zeroing a resolver, 54 Problems, 55

VII
viii CONTENTS

5 SYNCHROS, 58

5.1. General, 58 5.2. Basic synchro systems, 59 5.3. Basic synchro theory, 61
5.4. Electrical zero, 61 5.5. Summary, 63 5.6. How a receiver follows a trans¬
mitter, 64 5.7. Synchro torque, 65 5.8. Wiring errors, 68 5.9. Differential
synchro, 67 5.10 Synchro control transformer, 69 5.11. Geared synchro
systems, 71 5.12. Synchronism, 73 5.13. The need for adjustment, 77 5.14.
Zeroing procedures, 77 5.15. Zeroing a transmitter-—TX and CX with a volt¬
meter, 77 5.16. Zeroing a receiver—TR and CR by electrical lock, 78 5.17. Zero¬
ing a control transformer—CT using a voltmeter, 79 5.18. Zeroing a differential
transmitter—TDX using a voltmeter, 79 Problems, 80

6 MATHEMATICS OF BLOCK DIAGRAMS, 83

6.1. General, 83 6.2. Diagram mathematics, 84 Problems, 91

7 MECHANICAL ELECTRICAL CHARACTERISTICS, 96

7.1. General, 96 7.2. Rotary mechanical motion, 96 7.3. Gear Trains, 99

7.4. Units and conversion factors, 103 7.5. Servo characteristics, 104 7.6. Elec¬
trical circuit analogy, 105 7.7. Test commands and servo responses, 106
7.8. Response of a servo, 107 Problems, 109

8 ANALYSIS OF SERVOMOTORS AND GENERATORS, 112

8.1. General, 112 8.2. Application factors, 112 8.3. Fundamental equation, 113
8.4. AC servomotors, 115 8.5. DC servomotors, 117 8.6. Evaluation factors, 123
8.7. Servo generators, 126 Problems, 129

9 STABILITY IN LINEAR SERVO SYSTEMS, 131

9.1. General, 131 9.2. System responses, 131 9.3. Servo damping, 137 9.4.
Damping methods, 139 9.5. Mathematical analysis, 143 Questions and prob¬
lems, 148

10 FREQUENCY RESPONSE IN SERVO SYSTEMS, 150

10.1. Analysis using Nyquist plots, 150 10.2. Nyquist plot correction, 153
10.3. Analysis using Bode plots, 153 10.4. Bode plot construction, 157 10.5.
Bode plot summary, 161 Problems, 165

11 SERVO TRAINERS, 168

11.1. General, 168 11.2. Theory coverage, 168 11.3. Motor characteristics and
operation, 169 11.4. Transient response, 170 11.5. Speed control, 171 11.6. The
operational amplifier, 172 11.7. Reversible speed control, 174 11.8. The error
channel, 178 11.9. General performance, 179 11.10. Instability and stabilization,
180 11.11. AC control systems, 183 11.12. Relay-controlled systems, 183
11.13. Digiac Corporation Servo Trainer, 189 11.14. Electro-Craft Corporation
Servo Trainer, 190
CONTENTS IX

12 OTHER ELECTRO-MECHANICAL DEVICES, 193

SHAFT ENCODERS

12.1. Introduction, 193 12.2. Incremental encoder, 194 12.3. Absolute encoder,
194 12.4. Basic techniques, 195 12.5. Logic and codes, 196

STEPPER MOTORS

12.6. General, 199 12.7. Types, 199 12.8. Switching modes—permanent magnet
type, 200 12.9. A comparison of switching modes, 204 12.10. Developed torque
of stepper motors, 205 12.11. Use in control systems, 207

13 MICROPROCESSORS, COMPUTERS, AND SERVOMECHANISMS, 209

13.1. General, 209 13.2. The minicomputer, 210 13.3. The microprocessor, 211
13.4. Conclusion, 216

14 LABORATORY EXPERIMENTS, 217

14.1. General, 217 14.2. DC servo motor characteristics, 217 14.3. DC servo
motor speed-control system, 221 14.4. DC servo speed- and position-control
system, 226 14.5. Testing of a DC servo position-control system, 230 14.6. A DC
position-control servo using AC error detectors, 234 14.7. A relay-controlled
position-control system, 237

APPENDICES

A SI (METRIC) UNITS WITH ENGLISH AND CGS EQUIVALENTS, 241

B GLOSSARY OF TERMS, 243
C TRANSFER FUNCTIONS, 248
D UNIVERSAL TRANSIENT RESPONSE CURVES, 250
E REFERENCES, 251
F ANSWERS TO SELECTED PROBLEMS, 253
INDEX, 257
Preface

The purpose of this textbook is to provide material for a course in

servomechanisms given to electrical or mechanical technology students at
the two-year college level. It is more hardware oriented than mathematic¬
ally oriented so that it might be used in some engineering schools. It will
also give a technician or a practicing engineer in industry an up-to-date
refresher course. Approximately 60 percent of the material concentrates
on the different devices that are used to make up a complete servo system,
especially those devices that are used as transducers and/or error detectors
and correctors.
Basic theory in electricity/electronics, machines, amplifiers, power
supplies, and the like has not been covered because most curricula in
electrical or mechanical technology have done this adequately in other
courses. For the same reasons, only enough physics has been presented to
make the reader aware of the effects of velocity, acceleration, torque,
inertia, and friction in a servo system. Analysis of simple linear servo
systems is done without a rigorous mathematical approach. The differen¬
tial equation for such systems is derived only to emphasize the effects of
inertia and friction on system performance under change of input signal
or when an external disturbance occurs. The conversion from the time
domain to frequency domain is done algebraically by using the “s”
operator as a tool only, and then by converting to the “jW” operation.
Finally, the Bode plot is used to analyze and to predict system performance
of such systems.
One chapter deals with some of the different servo trainers now on the
market for laboratory use. These trainers are basically of a modular type

XI
XII
PREFACE

with greater emphasis on devices and hardware than on the mathematical

equivalent. Their adaptability to a laboratory is only limited by the
imagination of the user. Several representative experiments follow at the
end of the text to illustrate what can be done with a trainer. The author
has available for perusal a complete, separate laboratory manual for this
textbook in which experimental procedures, recommended apparatus and
equipment, and analysis of results are presented. The combination of this
textbook and the laboratory manual comprise a four-credit-hour one-
semester course in basic servomechanism theory.
The author wishes to thank all the companies who sent him technical
information with permission to use what he could. The companies are:
Schaevitz Engineering; Bourns, Inc.; New England Instrument Co.;
Control Data Corporation; Feedback, Inc.; Reeves Instrument Corp.;
American Electronics Inc.; Helipot of Beckman Instruments, Inc.;
Automatic Timing and Controls Co.; Muirhead Instruments Ltd.; Astro-
systems, Inc.; Pro-Log Corp.; Chilton Technological Services, Digiac
Corp. In addition, publications from the United States Navy were also
used. The electrical specifications for the potentiometer section were taken
from the Variable Resistive Components Institute, Revision A 1974
Industry Standard.

Richard W. Miller
 ONE

Introduction to systems

1.1. GENERAL

Automatic control is used in many processes and applications, for

example, to control the temperature of a furnace, the rolling of a sheet of
steel, the machine tool in a milling process, the launching and maneu¬
vering of a space vehicle, as well as in a myriad of other industrial,
commercial and military applications. In many of these applications, the
term “servomechanism” is becoming more and more a part of the total
control package, whether it is an instrument servo, a control servo, or a
power servo. The term servomechanism is a combination of two words:
one from the Latin, servus, meaning a slave or servant, and the other
word being “mechanism.” Thus it is a mechanism that slaves or serves
to carry out the commands that it is given. Specifically, the term servo¬
mechanism means a control system in which the controlled quantity is
“mechanical position.”
One of the earliest applications of automatic control, and to a certain
extent a servomechanism, was the flyball governor invented by James
Watt in 1788 to control the speed of a steam engine. Although this
governor was not very accurate due to oscillations and instability, it still
did work, and it was gradually improved. As time passed, more and more
advances were made by many different people, though somewhat slowly.
The advent of World War II was really the turning point in the automatic
control field. This was especially true in the military field where the con¬
trol of ships, submarines, aircraft, radar systems, etc., greatly accelerated
the study, advancement, and practice of the servomechanism.

1
2 INTRODUCTION TO SYSTEMS

Automatic control systems may be open-loop or open-cycle in

nature, or they may be closed-loop or closed-cycle. The general term
feedback control system encompasses the latter; in such a system, feedback
is used to compare the controlled quantity against a reference quantity.
The comparison between these two determines whether or not the con¬
trolled variable (which might be, for example, temperature, pressure,
fluid flow, speed, torque, or mechanical position) is at the desired value.
This comparison is performed by a transducer, which is simply a device
that converts energy from one form to another; that is, mechanical
position is converted to electrical voltage via a potentiometer or motor
speed is converted to electrical voltage via a tachometer-generator.
Regardless of the type of variable, if the controlled output is not at its
desired value, then the device(s) used for comparison create(s) an error
signal which is fed to an amplifier, which in turn feeds an error-correcting
device of some nature, which then drives the system to restore the variable
to the desired value. In a servomechanism a motor would be used as the
error-correcting device. Once the system has been brought within its
tolerance band or accuracy limits, it will stay that way until some outside
disturbance again causes an error signal to be created telling the system to
correct itself once more.

1.2. CONTROL

In general, three types of control can be performed to position a load

or to control the rate at which a load is moved from one speed level to
another. These three types are (1) ON-OFF control, (2) stepwise control,
and (3) continuous control.
Figure 1-1 shows a switch controlling a motor. When the switch is
closed, “ON” control results. When the switch is open, “OFF” control
results.
Figure 1-2 shows a rheostat with definite taps controlling power to a
load. Each individual tap provides an increase of load voltage in a
definite step, fashion, hence the name stepwise control.
Figure 1-3 shows a precision potentiometer controlling a motor. The
wiper makes contact at practically every turn resulting in a continuous
flow of power; hence the name continuous control.
In any of these systems, if the output has no effect on the input, then
it is known as an open-loop or open-cycle system. When the output can
be fed back into the system so that it affects the input, it is known as a
closed-loop or closed-cycle system. Regardless of the system, whether it
be open-loop or closed-loop, a block diagram will simplify the under¬
standing and analysis of it, no matter how complex it is.
1.2 Control 3

On-Off

FIG 1-1.

Stepwise

FIG. 1-2.

Continuous

FIG. 1-3.
 INTRODUCTION TO SYSTEMS
4

1.3. BLOCK DIAGRAMS

A block diagram is a shorthand method of showing how one quantity

depends on another. A generalized block diagram is shown in figure 1-4.

Actuating
Signal

Directly Controlled
Variable

FIG. 1-4. Generalized Block Diagram of a Closed Loop System

A simple proportional servo, however, may be better represented by

the block diagram of figure 1-5. Using as an example a mechanical
positioning system, the blocks shown represent the following elements:

1. The input element provides the reference input signal which initiates
and governs the response of the system. This is the reference against

tS
-p

Controlled
Variable

FIG. 1-5. Expanded Block Diagram of a Closed Loop System

1.3 Block diagrams 5

which the actual performance of the system is compared. It is some¬

times called the command, and it may be a reference voltage from a
potentiometer, a linear transducer, a synchro transmitter, or a
tachometer.
2. The output element is that part of the system which actually produces
the output being controlled, and it is acted on by the other com¬
ponents. It is also referred to as directly controlled variable or
controlled system. In a servomechanism this would be a function of
mechanical position or rate.
3. The error-measuring element is a comparator, an error detector or a
summation junction which compares the instantaneous values of the
reference input and the controlled output. This junction can be a
resistance network, an impedance network, etc., and may even be
contained within an amplifier.
4. The control element consists of all the parts of the system which have
any effect on the controlled output. This element contains the power
amplifiers needed in the system, the power inputs, all the control
elements which have an effect on control windings of devices, grid
control of a tube, base control of a transistor, etc.
5. The feedback element is the device that converts the controlled output
into a feedback signal which is an analog of the output. This signal
must be of the same form and scale as the signal from the reference.
The difference between the two signals is the error.

A truer picture of a servomechanism using an electric motor as the

actuator is that represented in figure 1-6. Most of the functional blocks

FIG. 1-6. Electrical Summing Junction Block Diagram

 introduction to systems
6

have been kept, but the error measuring elements block and control
elements block have been replaced with the summation junction symbol,
an amplifier and a motor. The true variable of a servomechanism is a
mechanical shaft angle; thus a dotted line is used with the output desig¬
nated as 60. A mechanical shaft angle must then be the reference input or
command, 6t. The reference input and feedback elements convert these
mechanical signals into electrical signals. Electromechanical transducers
such as precision potentiometers, variable reluctance transducers, re¬
solvers and synchros perform these conversions.
In order to convert the block diagram of figure 1-6 into a repre¬
sentation of a rate servo, all that is needed is to change the input and the
output to the first derivative of position, i.e., speed.

1.4. CLOSED-LOOP APPLICATIONS

The following figures, figures 1-7, 1-8 and 1-9, show three different
closed-loop systems featuring feedback from output to input. The
characteristics of all these systems are (1) fast response, (2) high accuracy
lack of error or deviation, and (3) stability under dynamic responses.
Figure l-7(a) shows the schematic wiring for a dc generator with an
exciter. The voltage across the field of the exciter is the difference between
the reference voltage and the feedback voltage from the generator
armature. If load is suddenly increased on the generator, the generator
voltage will tend to decrease, thereby increasing the differential between
the reference and the load voltage. This means that there is a larger
voltage across the field of the exciter, and that the voltage output of the
exciter is increased. This increased exciter voltage increases the generator
field current, which in turn raises the generator output voltage back to its
present value. Figure l-7(b) shows the block diagram; separate blocks
represent different parts of the circuit.
Figure l-8(a) shows the schematic diagram for a dc motor that
controls the tension of roofing material on a reel drive. In this system, the
roofing paper is a continuous strip and must be rolled on a reel at a
constant tension. If the tension increases, the paper may tear; if the tension
decreases, the layers of paper on the roll become very loose.
As the reel builds up with paper, the tension on the paper would
increase if the speed were kept constant; thus the reel speed must be
controlled to maintain a constant tension. The desired tension is set by
the weight of the rider under which the paper runs. The rider can move
only vertically. The roofing paper supports the rider; therefore, any
change in tension moves the rider either up or down. It moves up for
1 .4. Closed-loop applications 7

Ref Output

FIG. 1-7. Generator Voltage Regulator Schematic and Block Diagram

increased tension and down for decreased tension. Figure l-8(b) presents
the block diagram.
Figure l-9(a) shows the schematic diagram of a gun control servo¬
mechanism where range direction is the controlled parameter. A synchro
pair—a transmitter CX and a control transformer CT—is used to detect
and provide the error signal. An amplidyne is used to provide the ex¬
citation to the armature of the dc driving motor because of its very fast
response and high amplification. This amplification means that the fields
of the amplidyne can be supplied from an electronic amplifier, either tube
or solid state, because the fields do not require very much power. The
input to the amplifier is the “error” signal which indicates how far out of
line the actual gun position is in respect to the desired position.
 INTRODUCTION TO SYSTEMS
8

Roofing

Field
Desired Flux Motor Speed Wind Up Tension Rider
Reference 1 Position
Arm Reel
Position I
I
I

FIG. 1-8. Tension Control Schematic and Block Diagram

The operation of the system is basically as follows:

1. Everything is at rest; the gun is pointing at an angle of 0°; and the

synchros are at null with zero error voltage.
2. Suddenly an airplane appears at 30° from the null.
3. The synchro transmitter is turned 30°; there is now a difference
between the CX and CT stator voltages; thus, an output voltage from
the CT results. This error voltage is amplified and fed to the proper
amplidyne field which generates a different voltage in the amplidyne.
4. The generated voltage in the amplidyne has such a polarity that it
drives the motor in the proper direction and reduces the error voltage
to zero.

Figure l-9(b) is the block diagram.

In all of these systems, the return of the output to the desired point
will probably result in overshooting or hunting. Stabilizing a servo system
1.4. Closed-loop applications 9

I-: — t

A-C

Field
Power Power

FIG. 1-9. Gun Control Schematic and Block Diagram

to prevent or control this overshooting is one of the hardest things to

design into a servomechanism.

QUESTIONS AND PROBLEMS

1. Identify the input and output of the following devices or systems and
draw a block diagram for each:
(a) an automatic refrigerator
(b) Sunbeam “Radiant Toast” toaster
10 INTRODUCTION TO SYSTEMS

(c) a man shooting a deer

(d) a man driving an automobile
(e) an automatic washing machine
(f) a gas furnace controlled by a thermostat
(g) the perspiration system of a human being
(h) an automobile power-steering apparatus
2. A subtractive summing point in a feedback-control loop means that
negative feedback is being employed. Why is this true?
3. Define the following terms:
(a) actuating signal
(b) feedback signal
(c) forward loop
(d) reference input
(e) servomechanism
(f) transducer
(g) summing junction
4. What device must be used between the controlled variable and the
reference if their forms of energy are not the same?
5. Which of the following devices are transducers? Identify the input
and output of each.
(a) a hydraulic jack
(b) a thermocouple
(c) a common distribution transformer
(d) a rack and pinion device
(e) a voltmeter
(f) a vee-belt over a pair of pulleys
(g) a PNP transistor
 TWO

Precision potentiometers

2.1. INTRODUCTION

The wire-wound potentiometer has been utilized as an electro¬

mechanical shaft transducer in electrical systems and instruments for well
over 50 years. With the advancement of servomechanism theory and the
demand by the military for precise electromechanical shaft transducers,
the wire-wound potentiometer has become a very sophisticated device.

2.2. DEFINITION AND SCHEMATIC

What is a precision potentiometer? The Variable Resistive Com¬

ponents Institute defines it as a mechanical-electrical transducer the
operation of which depends on the position of a moving contact (wiper)
relative to a resistance element. It delivers, to a high degree of accuracy,
a voltage output that is some specified function of the applied voltage and
shaft position.
A wire-wound precision potentiometer is characterized by a resistance
element which is made up of turns of wire on which the wiper makes
contact on only a small portion of each turn.
The schematic for a simple wire-wound potentiometer is shown in
figures 2-1(a) and 2-1 (b). However, for the sake of simplicity, we will use
the schematic in figure 2-1(a) throughout this discussion.

11
12 PRECISION POTENTIOMETERS

FIG. 2-1. Potentiometer Schematic Diagrams

2.3. MECHANICAL CONSTRUCTION

The essential details of one company’s design of a wire-wound

potentiometer are shown in figures 2-2(a) and 2-2(b). The mounting plate
is made of aluminum; the shaft is of nonmagnetic stainless steel; the
bearings of stainless steel and are precision made; the housing of brass,
aluminum, or plastic; and the terminals of brass with a gold finish. The
brush arm assembly has contacts of precious metal. The resistance element
consists of a precious-metal alloy or a nickel-chrome alloy wound on a

FIG. 2-2. Exploded Views of Wire-wound Potentiometer Construction (Courtesy of

New England Instrument Co., Natick, Mass.)
2.4. Rotation 13

copper mandrel to provide for a uniform winding of precisely spaced

turns.
The servo mount shown in figure 2-2(b) is characterized by accuracy,
long life, low torque and high reliability, and it is usually motor driven.
The type of bushing mount shown in figure 2-2(a) is usually operated
manually and is used where preciseness is not too important. Thus pre¬
cision sleeve bearings may be used in place of ball bearings.
The overall size of a potentiometer is determined by a number of
factors. Table 2-1 shows how these factors change as the diameter
increases:

TABLE 2-1
Accuracy improves
Resolution improves
Power rating improves
Resistance range improves
Torque increases
Size-weight increases
Nonlinear capability improves

2.4. ROTATION

When the mandrel of a potentiometer is made as shown in figure 2-3,

it is known as a single-turn potentiometer. When a stop mechanism is
built into the potentiometer construction, the rotation of the shaft will
be limited to an angle less than 360°. The mechanical angle with stops
can be any angle desired, e.g., 270°, 300°, 320°, 330°, 335°, 345°.

FIG. 2-3. Single-turn Potentiometer Mandrel

 PRECISION POTENTIOMETERS
14

When an angle of 360° or more is required, the mandrel can be bent

downward at the end, and it can even be continued in the form of a helix.
A winding of three turns, five turns, ten turns or even more can thus be
obtained. A unit built in such a manner is called a multiturn potentiometer.
When only a single turn is used, the unit is called a single-turn continuous-
rotation potentiometer. There is quite a difference between how the brush
arm assembly is built in the single-turn and multiturn units. In a single¬
turn potentiometer, the wiper can be pinned or clamped to the input shaft
so that it turns only with the shaft. However in a multiturn potentiometer,
the slider must move along the axis of the shaft as well as turn on the shaft
so that it will continue to contact the helical winding as the shaft moves.
The design of multiturn potentiometers varies, but there must be trans¬
lator motion as well as rotary motion of the slider.

2.5. MATHEMATICAL RELATIONS

Let us look at figure 2-1(a) for the following analysis. A potenti¬

ometer can be considered “linear” if the resistance per unit length of turn
is constant. Thus:

£ = e_
or r = aR and r = — R (2-1)
a
R ~ 0T dT

where a can only equal or be less than one. If R is assumed to have a

value of one, then r equals a in per unit.
When viewed from the voltage point of view, the potentiometer is a
simple divider with an output voltage measured between the slider or
wiper to one end. Thus:

(2-2)

When viewed as a black box, a potentiometer may be represented as

in figure 2-4. The transfer function relating these is E0/E1N = 0/dT and
should be clearly given for all areas of interest. A few examples are

■5°. = — Linear 0 < 6 < 9 T (2-3)

Ein

= SINE 0 Sine Function 0 < 6 < 360° (2-4)

2.6. Load resistance 15

| Square Function 0 < 0 < 9T (2-5)

E
= 102 (-) 20dB Log Function 0 < 6 <, 0T (2-6)
Ein V /

The potentiometer, then, is used to convert a mechanical change in shaft

position to a corresponding electrical signal or vice versa. Because of its
ability to convert one form of energy to another, the potentiometer is
considered a transducer.

eo = WT. EIn)

FIG. 2-4. Potentiometer Block Diagram

2.6. LOAD RESISTANCE

A precision potentiometer with the load it is driving is depicted in

figure 2-5. This load resistance will introduce an error in the output voltage
e0 which depends on the relative sizes of the load resistance and the
potentiometer resistance. A change in output voltage results when the
current drawn by the load resistance flows through the series portion of
the potentiometer; thus, the voltage drop across this portion changes,
with a resultant change in load voltage e0.
Loading errors for different values of RL are shown in table 2-2 and
can be proven by using calculus. The error difference also depends on

It
1
II
X
Lin Rp
r Rl e0

X V-

FIG. 2-5. Potentiometer Driving a Load

 PRECISION POTENTIOMETERS
16

whether the potentiometer is connected in a single-ended or double-ended

circuit.

TABLE 2-2
Percent Error

Rl to RP Proportion Single-Ended Double-Ended

Rl = Rp 14.8 19.2
Rl = IORp 1.48 1.92
Rl = lOORp 0.148 0.192

2.7. ELECTRICAL SPECIFICATIONS—

WIRE-WOUND POTENTIOMETERS

It is essential to know a number of different electrical characteristics

in order to understand and specify potentiometers. These include:
resolution (travel, voltage, theoretical), linearity/conformity (terminal-
based, zero-based, independent), resistance tolerance, tap location,
power rating, noise, starting and running torque, moment of inertia,
ac characteristics, and electrical overtravel.

Resolution—This parameter is a measure of the sensitivity to which the

output ratio may be set. Figure 2-6 shows the resolution for a wire-wound
potentiometer.
In a wire-wound potentiometer, as the slider moves along the winding,
it may touch each turn of wire at only one point. As this happens, the
slider bounces off the winding so that the voltage varies in steps, rather
than in a smooth linear fashion. The plot for this is shown in figure 2-7;
from this plot, the various definitions of resolution are obtained.

FIG. 2-6. Ideal Resolution FIG. 2-7. Actual Resolution

2.7. Electrical specifications—wire-wound potentiometers 17

1. Travel resolution is the maximum value of shaft travel in one direction

per incremental voltage step in any specified portion of the winding;
2. Voltage resolution is the maximum incremental change in output
ratio with shaft travel in one direction in any specified portion of the
winding;
3. Theoretical resolution is the reciprocal of the number of turns in the
actual electrical angle and is expressed as a percentage.

For the same full range output, a multiturn potentiometer will have
better resolution than a single-turn continuous-rotation unit.

Linearity/Conformity—This parameter is one of the most important in

the selection and application of a precision potentiometer. The terms
linearity and conformity refer to the exactness of the relationship between
the actual and the theoretical function characteristics. Generally, linearity
refers to the straight-line relation of figure 2-6, while conformity refers to
nonlinear relations such as sine, sine/cosine, and square function potenti¬
ometers and how the output “conforms” to the theoretical characteristic.
When a potentiometer is constructed with end taps, there is always
some resistance that the wiper cannot move over. This resistance is
called the end resistance and makes it impossible for the output voltage
to reach either of the terminal voltages applied to the winding ends.
Figure 2-8 shows that the output voltage does not pass through either
zero volts or E volts as is shown in the ideal curve of figure 2-6.
Figure 2-9 shows linearity specified in terms of the maximum deviation
from the ideal curve; it is given as a percentage of the total applied voltage.
The actual characteristic deviates from the ideal characteristic because of
(1) the wire diameter not being constant, (2) unequal spacing of turns,
(3) the turns not being identical in size, and (4) the resolution.

FIG. 2-8. Characteristic of Ideal Po- FIG. 2-9. Terminal-based Linearity

tentiometer with End Resistance
 PRECISION POTENTIOMETERS
18

20] 1.8 V
18.2 V
E = 20 V 18 0 V
Ideal

FIG. 2-10. Output Voltages with and without a High End Trimmer

The effects of end resistance can be nullified by using special potenti¬

ometers called “trimmers” which are connected to either or both ends
of the function potentiometer. Figures 2-10 and 2-11 show two trimmers
connected as high end and low end trimmers. These trimmer potentiom¬
eters can be adjusted to obtain e = 0volts with the wiper at 0° and e =
Evolts with the wiper at 9t. If exact values of the end voltages are not
important, then the trimmers may be adjusted to minimize the maximum
deviation from the ideal characteristic.

FIG. 2-11. Output Voltages with and without a Low End Trimmer

Figure 2-12 shows that the end voltages are exact, but a deviation in
the characteristic does exist. Figure 2-13 shows that the deviation is
minimized by the trimmers. However, this causes some deviation at each
end and is called the best straight line fit to the ideal. When both trimmers
are used, the specification is called independent linearity.
When it is only possible to trim at the high end, the specification is
called zero-based linearity. Figure 2-14 shows this.
Conformity can also be improved by the use of trimmers. Therefore,
it may be desirable to specify conformity in terms of (1) terminal-based,
(2) independent-based, or (3) zero-based. Generally speaking, independent
2.7. Electrical specifications—wire-wound potentiometers 19

FIG. 2-12. Comparison of Actual and FIG. 2-13. Independent-based

Theoretical Function Linearity

FIG. 2-14. Zero-based Linearity

linearity or conformity is the specification most often given by

manufacturers.

Resistance Tolerance—Standard tolerances are usually given as ±5%,

but they may go as low as ± 1 %. The absolute value of the resistance is of
minor interest because the output voltage is some proportion of the
input voltage; the ratio is the important figure.

Tap Location—Taps are made by a single weld to a single turn and are
thus low in resistance and can be located quite accurately. The best
practical tap tolerance can range from ±0.1° up to +1.0°. Tap location
tolerance may have an effect on the accuracy of a potentiometer when it
is used in a servo system, and it should be analyzed carefully.
 PRECISION POTENTIOMETERS
20

Power Rating—The wattage rating of precision potentiometers is not

very large; generally, it is in the order of 1W to 10W, and it is usually
specified at some particular ambient temperature. If the ambient temper¬
ature is greater than 60°C, then the potentiometer must be derated. Under
this condition, the excitation voltage must be reduced. See figure 2-15.

T - °C

FIG. 2-15. Potentiometer Derating Curve

Noise—Noise is defined as any spurious variation in the electrical output

not present in the input. Generally speaking, it can be broken down into
four forms: (1) resolution noise caused by voltage variation as the wiper
moves from turn to turn; (2) shorting noise caused by the different number
of turns short circuited by the wiper; (3) loading noise caused by a change
in wiper-to-winding resistance as the wiper moves; and (4) generated
noise caused by thermocouple effect, etc.
The noise parameter is quite often given in ohms; the value of 100
ohms/potentiometer is about the maximum. Of the four forms of noise,
the effect of loading noise is the most pronounced because of the limitation
it imposes on the maximum permissible rotation speed; the limit is in the
neighborhood of 300 rpm.

Starting and Running Torque—Starting torque of a precision potenti¬

ometer is in the order of one-fifth to one-ounce inch. After a potentiometer
is moving, less torque is required to keep it moving. Thus it can be seen
why precision ball bearings are used in servo mounts whereas bronze
sleeve bearings are used in bushing mounts.
A larger number of turns results in a greater torque, if the same full
range of shaft output is required.

Moment of Inertia—The moment of inertia of precision potentiometers

is usually given in gram-centimeters. On multiturn units, this factor may
have an appreciable effect on system’s performance.
2.7. Electrical specifications—wire-wound potentiometers 21

AC Characteristics—As long as the windings are considered pure

resistance, the mathematical relations are the same for either ac or dc
circuits. However, many turns of fine wire on the mandrel lead to in¬
ductive and capacitive effects between turns and between the turns and the
copper mandrel. On 60 or 400 hertz systems, a reactive load of megohms
may result. Also shielded leads may be required between the potentiometer
and the load. Therefore, when used on ac, potentiometers of the single¬
turn continuous-rotation type with lower winding resistance are preferred.

Electrical Overtravel—Figures 2-16 and 2-17 show the mechanical

and electrical travel of a wire-wound potentiometer as well as the
overtravels.

Continuity-
Angle
Overtravel
Winding t Angles
Angle

FIG. 2-16. Mechanical Overtravels and Heat Loops

The total mechanical travel is the travel of the shaft between specified
stops. The stops are positioned so that the wiper comes to rest on the
overtravels, except in continuous-rotation potentiometers. The electrical

/VWWWVWVWWWWWX
-Total Mech. Travel-|-
Theoretical I
— Elec. Travel -H Mech
Overtravel
-Actual Elec Travel ■(Typ)

FIG. 2-17. Overtravel, Continuity, and Winding Angle

 o
II
CD

0-V\AAAAW^VWSA^WV-0-| |

(O
c
o co
’•*—
o E
c o
D cp
Ll_ g
"O
v_
Q
o
TD 5
C
o o
CO O

22
 x o
c
0)

00
O'

X
c
<r <
o
c
0)
O'
| ctT

FIG. 2-18. Typical Resistance Functions for Non-linear Potentiometers (Courtesy of Bourns Inc., Riverside, Calif.)

23
 PRECISION POTENTIOMETERS
24

winding angle is measured from the first to the last turn of the resistance
element. The electrical continuity angle includes the winding angle and
the overtravel angles.

2.8. NONLINEAR FUNCTIONS

Many applications require functions that are not linear, functions

such as a sine wave output, a cosine output, a log output, a square output.
Most manufacturers can supply just about any function required or
desired. See figure 2-18.
A nonlinear function can be achieved in a number of different ways
(see figure 2-19) such as:

1. variable spacing of turns

2. change of wire size
3. change of wire type
4. contoured card mandrel
5. stepped card mandrel
6. square card mandrel

Some potentiometers have a large number of taps, and, by using

external resistors, a variety of nonlinear functions can be obtained.

2.9. NON-WIRE-WOUND POTENTIOMETERS

The strict definition of a non-wire-wound precision potentiometer is a

potentiometer characterized by the continuous nature of the resistance
element in the direction of wiper travel. To achieve this characteristic,
the resistance element is made of a thick film of metal, carbon, or ceramic/
metal substance which is deposited on a nonconducting base. Such
elements are continuous, noninductive, and offer a smooth path for wiper
action.
These film potentiometers have a number of advantages over the
wire-wound potentiometers, among which are: (1) wear and wiper bounce
are kept very low; (2) higher wiper speeds can be utilized before continuity
loss results, speeds to 2000 rpm; (3) failure of the resistance element is
much lower; and (4) current surges very seldom cause element burnout.
The most important advantage is that resolution is much improved.
Figure 2-20(a, b, c) shows the essential details of three different types
of non-wire-wound potentiometers.
2 9. Non-wire-wound potentiometers 25

dR/d# = W = K

FIG. 2-19. Typical Card Shapes (Courtesy of Bourns Inc.,

Riverside, Calif.)
26 PRECISION POTENTIOMETERS

B - Flat Track

C - Bulk Element

FIG. 2-20. Exploded Views of Non-wire-wound Potentiometers (Courtesy of New

England Instrument Co., Natick, Mass.)

2.10. POTENTIOMETER CONNECTIONS

A potentiometer can be connected into a circuit in either of two ways.
When connected as in figure 2-21, it is called single-ended excitation
because one end of the winding is connected to ground. When connected
as in figure 2-22(a) or (b), it is called double-ended excitation because
either the center tap of the potentiometer is grounded or the center of the
excitation voltage is at ground potential. The load is connected between
the ground and the wiper.

2.11. OPERATION IN A SYSTEM

The potentiometer functions in a servo loop as a reference input
element or a feedback element. In either case, it is used in making the
conversion from mechanical position (shaft angle) to the electrical
representation of this position.
2.11. Operation in a system 27

FIG. 2-21. Single- FIG. 2-22. Double-ended Excitation

ended Excitation

A common connection for two potentiometers is that of a bridge

circuit with the error voltage being measured between the two sliders, as
is shown in figure 2-23(a). The diagram showing how the potentiometers
might be shown in a block diagram is given in figure 2-23(b).

FIG. 2-23. A Potentiometer Error Channel and Block Diagram

Position Feedback. Figure 2-24 shows the connection of a potenti¬

ometer used as a feedback element in a servo loop that controls mechanical
position.
The servo motor, which is a two phase ac type, drives the wiper of the
feedback potentiometer as well as the load. This output 60 is the assumed
angular position of the load. The voltage taken from the wiper is the
feedback voltage ef, while the input voltage ec taken from the reference
potentiometer is the command which directs the loop to position the
output shaft at a particular angle.
When ec is larger than ef, the error voltage ee will have a certain
polarity when it is fed to the ampl fier; conversely, when ef is larger than
28 PRECISION POTENTIOMETERS

Command Servo

FIG. 2-24. Mechanical Position Servo with Potentiometer Transducers

ec, ee will be reversed in polarity. This reversal in polarity will cause the
servo motor to drive in the correct direction in order to reduce the error
voltage to zero. The loop is now at the null position.
This loop is capable only of accepting positive input voltage and
driving toward positive values of dQ. If it is desired to have the loop
respond to negative signals and to drive toward a negative angle, then a
double-ended circuit must be used.
If Ec in figure 2-24 is held at a constant value and Ef, the feedback
source, is made variable, then the computation of a reciprocal results.

Multiplication. Figure 2-25 shows the connection of a potentiometer

used to provide multiplication of two inputs with one input being variable
and one being constant. As the mathematics shows, this multiplication
also uses division by a constant.

2.12. ZEROING A POTENTIOMETER

In testing or using a potentiometer, care must be taken to ensure that

the dial or pointer position at zero degrees gives the correct voltage
(resistance) output that results from the particular potentiometer. This
output voltage will vary from zero to a finite value depending on whether
the potentiometer is a single-turn limited-rotation or continuous-rotation
potentiometer of all types and/or functions.
2.12. Zeroing a potentiometer 29

Command Servo

at Null ex = ef
ey 6y
Therefore a = — and e0 = ■=- Ey
tF

FIG. 2-25. A Multiplying Servo with Potentiometer Transducers

A potentiometer may be checked with a high impedance ohmmeter

in order to measure output or function resistance. A low voltage may be
applied across the two fixed end points (never apply voltage between a
fixed end and a tap) and the output voltage(s) read. If the output voltage
is incorrect, then the pointer or dial may be loosened and rotated and then
clamped in the correct position. In some cases, depending on the
mounting, the body of the potentiometer may be rotated while the pointer
or dial is held; it is then clamped in the new position.

PROBLEMS

1. Two potentiometers of 50K ohms each are connected across 110V dc

as an error detector. Determine the following:
(a) the wiring diagram, (b) the relative position of the wipers to give
an error voltage of 40V if they are single-turn, (c) three-turn,
(d) ten-turn.
2. Two 4W potentiometers of unknown resistance with 360° electrical
travel are connected across 60V ac as an error detector. A voltmeter
30 PRECISION POTENTIOMETERS

between the slider indicates a voltage of 48V. The reference potenti¬

ometer is set at 310°. What is the position of the feedback or follow
potentiometer? What are their resistances? Sketch the ac wave at
each potentiometer slider when an error voltage of 48V exists.
Sketch the wave when the error voltage is driven to zero.
.
3 A potentiometer is designated as 360° sine. What does this mean?
Sketch the output wave that could be obtained.
.
4 A potentiometer is designated as 180° sine/cosine. What does this
mean? Sketch the output wave shapes.
.
5 A tangent function potentiometer is being used in a servo. If 30V dc
is impressed across the potentiometer, determine the slider voltage
for angles of 45°, 60°, 75°, 90°, 180°, 270°, 315°, 360°. Plot the
output. Hint:

Let E0 - la" <75°X)lj ; X = ^

(tan 75°)(2) 9,

6. A wire-wound potentiometer has an electrical rotation of 3590°. It

has 4800 turns. What is the theoretical resolution?
7. A three-turn wire-wound potentiometer has 2100 turns of wire on its
mandrel. What is the theoretical resolution?
8. Potentiometer resistance elements are sometimes made of conductive
films applied to glass plates. This is done in order to: (a) increase
resolution, (b) increase linearity, (c) increase power handling ability,
(d) reduce friction. Which answer is correct and why?
9. If the brush connections to a sine/cosine potentiometer are reversed,
what will happen to the output voltages?
.
10 A 5K potentiometer having 350° of total travel is loaded with a 6K
load resistor. If the applied voltage is 15V, what is the theoretical
voltage supplied across the load at 125° of slider movement?
.
11 (a) In the circuit of problem 10, prove that when the 6K potentiometer
is removed and the resistance between slider and ground is called
Rx, the output voltage will be:

(b) Prove that after replacing the 6K load, the output voltage will
be:

Rx + Rl
V0
Rp - R* +
RxRl
2.12. Zeroing a potentiometer 31

.
12 What is the difference between voltage resolution and angular
resolution?
.
13 What is meant by the term linearity?
14. What is the difference between linearity and conformity?
15. Explain the difference between winding angle and continuity angle.
16. Give one advantage and one disadvantage of a multiturn potenti¬
ometer as compared to a single-turn potentiometer.
.
17 Explain the difference between terminal linearity and independent
linearity.
.
18 What are trimmer potentiometers and why are they used?
19. The following diagram shows a servo loop with a feedback potenti¬
ometer. (a) Prove that the equation is correct, (b) Recopy the
diagram and mark on it (1) where the error voltage would be found,
if and when it exists, and (2) where the feedback voltage is found.

.
20 Two three-turn potentiometers of 10K ohms are connected across
36Y dc as an error detector. Three turns of the potentiometers
correspond to 352° of mechanical movement. Find the wiper position
of the feedback potentiometer to give an error voltage of 18.8V if
the command potentiometer is set at 94°.
.
21 On a small aircraft landing field, the shaft that the wind sock is mounted
on is connected mechanically to the shaft of a sine/cosine potenti¬
ometer. The potentiometer is excited from 50V ac. Determine the
wind direction if the readouts of the x and y coordinates are -23.8V
and 44.4V respectively.
.
22 The wattage rating of single-turn potentiometers that have theoretical
electrical angles of 354° and resistances of 20K ohms is 5W for wire
wound and 2.5W for plastic for temperatures up to 85°C. At 150°,
the wattage rating is zero. Determine the wattage that each of the
potentiometers can handle if they are used at an ambient temperature
of110°C.
 THREE

Variable reluctance transducers

3.1. INTRODUCTION

Another type of transducer is the variable reluctance magnetic type

in which the magnetic reluctance of a transformer is changed by changing
either the position of the secondary coil in respect to the primary or the
position of the iron core in respect to the coils. The three most common
devices that utilize this principle are the E-core transformer, the linear
variable differential transformer (LVDT), and the microsyn. The first
two sense changes in position along a straight line path; the microsyn
senses change in angular position.

3.2. PRINCIPLE OF OPERATION

The basic principle underlying the operation of any of these trans¬

ducers is that the ac magnetic coupling between two or more coils can be
changed simply by changing their relative mechanical position or by
changing the position of a separate moveable core.

E-core Transformer. The E-core transformer has a magnetic circuit

that has an I and E core as shown in figure 3-1.
There are three windings wound around the E-core. The center
leg winding is the primary. The two windings on the outside legs are the
secondaries and they are connected in series so that their voltages oppose
each other. In the exact zero position, the reluctance of the magnetic path

32
3.2. Principle of operation 33

v>-

"A”

- In

FIG. 3-1. Schematic of an E-core Transformer

through both cores is the same; thus, the voltages induced in coils A and
C are equal and opposite, and hence the output voltage, e0, is zero.
When a change in system position causes the I-core to move, the reluctance
changes, and the voltages induced in coils A and C change. When the
voltage in coil A is larger than in coil C, the I-core has moved toward A
and away from C. If the core is moved toward C and away from A, then
the voltage across C will be larger than across A. In either case, the net
voltage across the output terminals is different from zero, depending on the
offset. In the first case, the phase of the output voltage is the same as the
phase of the voltage across A; in the second case, it is the same phase as
the voltage across C. Thus, the amplitude of the output voltage gives a
magnitude indication of the core displacement, and the phase of the
output voltage indicates the direction of the displacement.

Linear Variable Differential Transformer. The LVDT is the most

commonly used variable reluctance element. Figure 3-2 shows its principal
parts.
A primary coil is wound around the center portion of the form, and
two identical secondaries are wound around each end of the form. A
magnetic core rod is located inside this coil assembly. The secondary
coils are connected externally in a series-opposing circuit. The motion
of the magnetic core will vary the mutual inductance of each secondary
to the primary which, in turn, will vary the voltage induced in each
secondary.
If the core is in the exact center between the secondary coils, then the
voltage induced in each coil will be the same and 180° out of phase with
the other so that there will be zero output voltage. This is the null position.
 VARIABLE RELUCTANCE TRANSDUCERS
34

FIG. 3-2. Schematic Diagram of a Linear Variable Differential

Transformer

When the core is moved from this null position, the voltage induced
in the secondary coil toward which the core is moved will increase, an
the voltage induced in the other secondary will decrease. This produces a
differential voltage output that varies in a linear manner with the core
position. If the core is moved the other way, the output voltage would be
shifted in phase by 180°. When voltage is plotted against core position,
within the range of linearity or just beyond, the graph is a straight line
going through the origin and lying in quadrants I and III. When voltage
is plotted as a positive quantity, regardless of phase, the graph is V shape
See figures 3-3 and 3-4.

Core Position

FIG. 3-3. Output Voltage and Its FIG. 3-4. Absolute Magnitude of
Phase Reference Output Voltage

DC Operated LVDT. The LVDT just described works on ac and, to a

certain extent, is the workhorse of the reluctance transducers. However,
the advent of microelectronic circuitry and the simplicity of dc operation
3.2. Principle of operation 35

FIG. 3-5. Block Diagram of a DC-LVDT

has led to the design and construction of the DC-LVDT. The heart
of this device is still an AC-LVDT, but now electronic circuits have been
added; this is shown in the block diagram of figure 3-5.
The diagram shows that the module contains a complete passive
demodulator/dc amplifier signal conditioner.
The full-scale output voltage does not depend on the dc input as it
does in an AC-LVDT. Thus the term “sensitivity” which describes the
output of an AC-LVDT cannot be used to describe the DC-LVDT.
Instead, the output scale factor is specified in volts per inch of core travel
or in millivolts per mil.

Rotary Variable Differential Transformer. The differential trans¬

former used for measuring angular displacement is known as the Rotary
Variable Differential Transformer (RVDT); it is shown in figure 3-6.

Input Shaft

Secondary
No. 1
Wire Leads

FIG. 3-6. Simplified Cross-section of an RVDT {Courtesy of Schaevitz Engineering,

Camden, N.J.)
 VARIABLE RELUCTANCE TRANSDUCERS
36

l/>

o
>
3
CL

FIG. 3-7. Output Characteristics of an RVDT

Note the difference in the core designs of the RVDT and the LVDT.
The RVDT uses a heart-shaped core of magnetic material with the shape
determining the linear output range of the element over its specified range
of rotation. While most RVDT’s allow a full 360° of rotation, only about
+ 40° to ±60° of rotation are used for the useful output (see figure 3-7).
However, the latest designs have extended this linear range to as much as
+ 340°. The output of these later units increases linearly to a maximum
at 340° and then drops nonlinearly to zero volts.
The RVDT is available in both ac and dc types; dc units contain
integral thick film signal conditioning, as is shown in figure 3-8.

FIG. 3-8. Electronic Block Diagram of an RVDT

3.3. USE IN SERVO SYSTEMS

The excellent linear output characteristic of an LVDT or RVDT

makes these devices extremely useful as null-position transducers in
3.3. Use in servo systems 37

"-It

FIG. 3-9. Null Balance System

follow-up servo systems. A follow-up servo system maintains a fixed

positional relationship between two or more moveable members. If the
command member changes position, then the servo system moves the
feedback or “slave” members in a manner that restores their positions
relative to the command. Figure 3-9 is a representative schematic wiring
diagram of such a system.
Electrically, the system works as follows. At null, the two LVDT
cores are at the same position; consequently, their output voltages are
equal and 180° out of phase so that the error voltage, ee, is zero with the
motor at rest. If the core position of the command LVDT is moved in
an upward direction, the magnetic coupling between the secondary coils
will change so that exl increases in a positive direction. This voltage, eT1,
thus becomes larger than ex2, and an error voltage, ee, appears at the
input to the amplifier. This error voltage is amplified and fed to the
control phase of the two-phase servo motor so that the motor starts and
runs in the direction of rotation that will drive the core of the following
LVDT until it matches the core position of the command LVDT. At this
point, the error voltage is again zero, and the motor stops. Conversely,
when the command LVDT is moved in the reverse direction, reverse
motor action will take place.
38 VARIABLE RELUCTANCE TRANSDUCERS

d#,
dt

01

FIG. 3-10. Block Diagram of an LVDT Error Channel

A diagram showing how the two LVDT’s might be shown in a block

diagram is displayed in figure 3-10.

3.5. ZEROING AN LVDT

An LVDT should be mounted on a support that is free from vibrations

and that is a safe distance away from any strong magnetic fields that may
emanate from other power transformers, motors, relays, contactors, etc.
In terms of zero, the LVDT armature should be in the exact center of
its coil. If the electrical zero output does not agree with the mechanical
shaft zero position, then the following adjustments can be made:

1. For fine adjustments, the zero-adjusting screw is rotated until the

servo shaft reaches its zero position.
2. For large adjustments above 2°/0 of the total range, the flexure plate
is adjusted by the zero-adjusting screw until the plate is in a vertical
position. The coil is then physically moved until the servo shaft
nears its zero position, and it is then locked in place.
3. If an LVDT is being used by itself, then the armature core can be
held at exact zero as given by the output voltage, and the body of the
transformer moved until this is reached. The body is then clamped
in place.

3.6. THE MICROSYN

The microsyn is a reluctance type transducer that is used when the

angular movement being controlled is only a few degrees. It is an in¬
ductive device that has primary and secondary windings wound on poles.
3.6. The microsyn 39

Two poles, 180° apart in mechanical space, comprise the primary, and
two poles, also 180° apart, comprise the secondary. The rotor is a bar of
low reluctance iron with no windings. Figure 3-11 shows the rotor in the
zero position; the flux paths are indicated by the dotted lines. In this
position, the turns of the output windings are not cut by flux created by
the exciting current; therefore, there is no voltage induced in the secondary
windings. When the rotor is moved away from the zero position in the
clockwise direction, an emf of a certain magnitude (depending on dis¬
placement) and phase will be induced in the secondary windings, and an
output voltage will result. When the rotor is turned in the opposite
direction, the magnitude of the voltage will again be determined by the
displacement, but it will now be of opposite phase. The microsyn can be
constructed so that it has excellent resolution and is extremely linear. It
may also be wound to give a number of functional forms.
In addition to its application as a generator of an electrical signal, the
microsyn may also be used as a torque transmitter when the torque is
proportional to the product of two currents. The windings are now called

Stator

FIG. 3-11. Elementary Microsyn

40 VARIABLE RELUCTANCE TRANSDUCERS

the reference and control winding rather than primary and secondary.
The amount of torque is determined by the magnitude of the currents;
the direction in which the torque acts is determined by their phase relation,
or polarity. Since torque is the principal thing here, transformer action
does not enter in; therefore, the exciting currents can be either ac or dc.
 FOUR

Resolvers

4.1. GENERAL

An induction resolver is essentially a transformer with a rotary

electromagnetic coupling between the primary and secondary windings.
The windings are located on the stator and rotor in such a manner that the
output voltage has an amplitude which is proportional to the sine or cosine
of the input shaft position. In normal operation, the resolver stator, which
is similar to a transformer primary, is excited with alternating voltage.
The coupling between the rotor and the stator is magnetic, the same as
between the primary and secondary of a transformer. In a resolver,
however, the position of the rotor can be changed with respect to the
stator; this allows the magnetic coupling to vary with shaft rotation. This
variable coupling produces an output voltage which has an amplitude
that is equal to the stator voltage multiplied by the sine or cosine of the
rotor shaft’s angular position with respect to the stator. See figures 4-1
and 4-2.

SI R 1 SI R1

S3 R3

FIG. 4-1. Two-winding Resolver FIG. 4-2. Three-winding Resolver

41
42 RESOLVERS

4.2. FOUR-WINDING RESOLVER

To make induction resolvers even more useful, two windings are

provided on both the stator and the rotor. The pairs of windings are
located at right angles to each other. Complex trigonometric functions
involving both sines and cosines can be solved when both stators are
excited. The voltage transformation ratio from primary to secondary is
usually 1:1, although other values are used. Other windings may be
added for feedback or compensation in order to achieve greater accuracy
and to make the resolver less sensitive to environmental variations and
load and source impedance variations. See figure 4-3.

FIG. 4-3. Four-winding Resolver

The flux created by ES1 induces a voltage in R1-R3 that is determined

by the cosine of the angle whereas the flux created by ES2 induces a voltage
in this same coil that is determined by the sine of the angle.
Figure 4-4(a) depicts two stator windings at right angles to each
other. Assume that each winding is excited with 10V ac. Each winding
will have a magnetic field of its own which will be at a right angle to the
other. The resultant of these fluxes will be at 45° and equivalent to 14.14V.
However, this flux vector may assume any position depending on the
amplitude and polarities of the exciting currents; its angular position is
equal to the arc tangent of the ratio of the two stator currents.
Figure 4-4(b) shows the same stator windings with one rotor winding
at a right angle to one stator and parallel to the other stator. The output
voltage of this winding is then 10V.
When two rotor windings are used as in figure 4-4(c), the voltages are
proportional to the magnitude of the flux vector and to the sine or cosine
of the angle of rotation of the rotor winding with respect to the vector.
 in
i 05
sT c
cr *°c
Ll in in
L-l Q
o
CO X
o
1 ~ o
0) 1 —
E 05 0 1
A 01 = ZS3

o co O
if) O to p
■—- ° o —

> CM o X

o LU y O

n II II II
CM
(T a:
LU LU
O
>
O
CO
CD
tr
CO
o o
CO CO
x
LU
O O
+ O)
A 01 = zs3

+ +
CD
■*->
o
cn O
O
o >
yo coo
o x > o
w 0 o o O
Lj y T)
d>
o
II II II II
o
■O
cc c
LU

O
A 01

II
to
LU

43
44 RESOLVERS

-S2
S4 S2
SI R1

sin 6 (4-1)
sin 9 (4-2)

FIG. 4-5. Basic Schematic of a Four-coil Resolver at

Electrical Zero

4.3. ELECTRICAL ZERO

The most important application factor is the electrical zero and the
corresponding electrical relations as referred to the resolver terminals.
Figure 4-5 shows a four-coil resolver with fixed primaries (stators) and
rotating secondaries (rotor). The relative phase of the windings is shown
by the polarity dots; the polarity of the voltage is analogous to the direc¬
tion of the voltage vector. For the zero position, stator coil 1, SI-S3,
induces maximum voltage in rotor coil 1, R1-R3, but zero voltage in rotor
coil 2, R2-R4. Similarly, S2-S4 induces maximum voltage in R2-R4 but
zero in R1-R3. Equations 4-1 and 4-2 give the voltages.
Standard rotation for positive values of shaft angle 0 is counter¬
clockwise (CCW). As the shaft is turned CCW from zero, each stator
winding induces voltages in both rotor windings. These voltages vary in
magnitude and phase (direction) in accord with the sine-cosine relations
in each quadrant. Figure 4-6 shows a rotor angle of 45°. The phase of the

FIG. 4-6. Determination of In-phase and Out-of-phase Voltages for a Rotor

Position of 45°
4.5. Composition 45

voltages induced in the rotor windings, as a result of the current flowing

in the S2-S4 winding, is indicated by black dots. Both the voltages at R1
and R2 are in phase with the voltage at S2 because both the sine and
cosine of 45° are positive. The phase of the voltage induced in the rotor
windings, as a result of the current flowing in the SI-S3 winding, is
indicated by open circles. The voltage at R2 is in phase with the voltage
at SI whereas the voltage at R4 is out of phase with the voltage at SI
because the cosine of 45° is positive while the sine of —45° is negative.

4.4. RESOLUTION

Of the basic resolver computations—resolution, composition, and

transformation of axes—resolution is the simplest and is used to convert
the polar coordinates of a vector to rectangular coordinates (see figure
4-7). A vector R L9 can be “resolved” into two components, x and
y, because x = R cos 9 and y = R sin 9.
The electrical inputs to the servo loop of figure 4-7 are R and 9; the
loop operates as a potentiometer follow-up system which converts
electrical input, 9, to the corresponding angular shaft position, 90, of the
resolver. The exciting voltage applied to the stator winding is converted
to the required rectangular coordinates.

4.5. COMPOSITION

Composition converts two rectangular coordinates to their corre¬

sponding polar coordinates. It is the inverse of resolution and always
requires a servo loop with a servo motor and amplifier (see figure 4-8).

90

(4-3)

(4-4)

FIG. 4-7. Resolution—Converting a Polar Coordinate to its Rectangular Component

46
RESOLVERS

FIG. 4-8. Composition—Converting Two Rectangular Coordinates to a Polar

Coordinate

The inputs ES1 = y and ES2 = x are applied to the two stator windings,
and the outputs are taken from the two rotor windings. The voltages
ES1 and ES2 form a resultant magnetic flux vector which lies at the angle 9
to the x-axis.
This magnetic flux vector induces voltages in R1-R3 and R2-R4.
The R2-R4 output is applied to the amplifier in the proper phase in order
to drive the motor and the resolver shaft CCW for positive angles of 9.
The servo motor turns the shaft through the angle 9 until the null is
reached; this occurs when the Rl-R3-axis is at right angles to the flux
vector; the voltage applied to the amplifier then becomes zero. The
R2-R4-axis is then parallel to the flux vector, and maximum voltage ap¬
pears across the R2-R4 terminals. The voltage at R1-R3 is the error
voltage in the servo loop, the voltage at R2-R4 is the resultant R, and the
dial indication is the desired angle 9.

4.6. ROTATION OF COORDINATE AXES

Many ships, aircraft, and space vehicles require information generated

in one coordinate system to be converted into another coordinate system.
For example, components of velocity measured with an inertial navigator
must be converted into aircraft pitch-and-roll velocities. Thus, rectangular
coordinates from one set of axes must be rotated to a second set of axes
that are at some angle relative to the first. Another example would be the
problem of converting data in terms of a ship’s heading into data in terms
of true North in order to furnish target latitude and longitude data to
4.6. Rotation of coordinate axes 47

FIG. 4-9. Latitude and Longitude of a Target

other ships in the fleet. The course the ship is on is called the relative¬
bearing coordinates with North being called the true-bearing coordinates.
See figure 4-9.
Voltages proportional to relative-bearing coordinates x and y are
applied to the respective stator windings of a resolver; thereby producing a
resultant magnetic flux vector. The rotor windings are then rotated from
their zero position through an angle equal to the course angle of the ship
from North. The voltages induced in each rotor winding are proportional
to the sine or cosine of its angular position with respect to the flux vector,
not with respect to zero. If the reference is made with respect to zero,
then the outputs are found from equations 4-1 and 4-2. The following
two examples are used to illustrate the steps necessary in finding the
true-bearing coordinates.

Example 4-1

See figure 4-10. Six volts (representing an x-coordinate of +6) and 8V

(representing a y-coordinate of - 8) are applied to primaries S2-S4 and
SI-S3 respectively; the course angle is 37° NNE. The coordinate y =
— 8 is represented by a phase reversal of S1-S3; thus, the resultant
magnetic field lies in the fourth quadrant. The rotor windings are now
rotated through an angle of 37°; therefore, the voltage induced in each is
proportional to the component of the resultant field which lies parallel
to the axis of the rotor winding R1-R3. The voltage induced in R2-R4 is
zero because its axis is at right angles to the resultant field. The true
bearing coordinates are x = 0 and y = —10, and the target is located
10 units due South of the ship.
48
Axis

It

11
o
>

+
10
rO

'd'
id
LlI
I

II
<1>
c

LlI

D I

II
w (X)
tnO

CO
x

in „
cj LlI CO
o O
g®

II
N X CO
FIG. 4-10. Conversion of Relative Bearing Coordinates to True Bearing.Coordinates
4.6. Rotation of coordinate axes 49

Example 4-2

The radar on an aircraft carrier picks up the blip of an enemy ship and
supplies the distance and angle that the carrier is from the target. Other
ships in the fleet must be informed of the true-bearing coordinates so
that they can converge on the target. The carrier is steaming on course
on a relative-bearing axis (x-axis) which lies 20° NNE of true North. The
enemy ship is 36 miles from the carrier and at an angle of 50° NNW from
it. Determine:

1. the wiring diagram

2. the relative-bearing coordinates
3. the true-bearing coordinates

Solutions

(1) IRS 2RS

(2) Using the resolver equations 4-1 and 4-2, we will solve for ER1 and
Er2 of IRS which will be the x- and y-(relative-bearing) coordinates.

Erj = ES1 cos 9 + ES2 sin 9

= 36 cos 50°
= 36 x 0.643
= 23.3V or 23.3 miles

Er2 = Es2 cos 9 — ES1 sin 9

= 0 — 36 sin 50°
= -36 x 0.765
= —27.6V or 27.6 miles

Note: The negative sign indicates an out-of-phase ac relationship

only; it has no bearing on the solution.
50
RESOLVERS

(3) Er1 and ER2 found in step 2 are the excitation voltages E^ and Es2
respectively for IRS. Again, using equations 4-1 and 4-2 except
using angle alpha, d, rather than theta, 9, we will solve for ER1 and
Er2 of 2RS which will be the north and west (true-bearing)
coordinates.

N: ER1 = ES1 cos d + ES2 sin d

= 23.3 cos 20° + 27.6 sin 20°
= 23.3 x 0.940 + 27.6 x 0.343
= 21.9 + 9.47
= 31.36V or 31.36 miles north

W: ER2 = ES2 cos d — Esj sin d

= 27.6 cos 20° - 23.3 sin 20°
= 27.6 x 0.940 - 23.3 x -0.343
= 25.94 - 7.99
= 17.95V or 17.95 miles west

The north and west coordinates may also be found from

N = R cos 9 cos d + R sin 9 sin d

= 36 x 0.643 x 0.94 + 36 x 0.765 x 0.343
= 21.76 + 9.45
= 31.21V or 31.21 miles Ck.

W = R sin 9 cos d — R cos 9 sin d

= 36 x 0.765 x 0.94 - 36 x 0.653 x 0.343
= 25.89 - 7.94
= 17.95V or 17.95 miles Ck.

R cos 9 cos d + R sin 9 sin d may be reduced as a trigonometric

identity to R(cos 9 cos d + sin 9 sin d) and then to R cos (9 — d).
R sin 9 cos d — R cos 9 sin d may be reduced to R(sin 9 cos d —
cos 9 sin d) and then to R sin (9 — d); therefore

N = 36 cos (50° - 20°)

= 36 cos 30°
= 36 x 0.866
= 31.2V or 31.2 miles Ck.

W = 36 sin (50° - 20°)

= 36 x 0.5
= 18V or 18 miles Ck.
4.7. Data transmission—angular addition/subtraction 51

Vector Diagram IM

4.7. DATA TRANSMISSION-

ANGULAR ADDITION/SUBTRACTION

Figure 4-11 illustrates the use of two resolvers in cascade to give

angular addition or subtraction. Data, in the form of angular shaft
positions, are transmitted from one place to another without any mechan¬
ical connections. Only one of the stator windings is used with the other
winding being short-circuited. Figure 4-11 shows that the two resolver
rotors are displaced 30° from their zero position. Ten volts is applied to
S2-S4 of resolver No. 1. The voltage induced in R2-R4 is 8.66V, and this
in turn is applied to S2-S4 of resolver No. 2. The 5V induced in R1-R3
of No. 1 is fed to the S1-S3 winding of No. 2. The voltages induced in
windings R1-R3 and R2-R4 of No. 2 add, giving 8.66V and 5V respec¬
tively. These values represent the sine and cosine of 60° which is the sum,
30° + 30°, of the two resolver rotor angles.
Figure 4-12 shows angular subtraction where the same angular and
voltage values are assumed. The difference between figures 4-11 and 4-12
is that the connections to S2-S4 of No. 2 have been reversed. The resultant
voltages are therefore 0V and 10V, respectively, which represent the sine
and cosine of 0°, (30° — 30°).
When the output ER1 of resolver No. 2 in figure 4-12 is applied to an
amplifier, and the amplifier output is in turn fed to a servo motor, a servo
loop, where one resolver follows another, will be obtained. As an example,
let ES2 = 10V, 0 = 40° (resolver No. 1), and a = 30° (angular position
of resolver No. 2). Before the servo loop starts to operate, the voltage
output of R2-R4 of 2RS will be 9.85V, and the voltage output of R1-R3
 m
in o
>

o a X
10
a > <x>
X _c 10
c
to in
o CO 10 co
CM
1 m
CO CO CO II
Ld to it Ld
II m
10 +
1 CO + m
in
10
a 6 a 10
OJ
CO CO 00
o X +
i o
i
CJ
CM
10 o o ro
CO 10 in CO X m
Ld 0D N- Ld m

II II II II II ll
OJ
or

FIG. 4-11. Resolver Schematic Showing Electrical Data Addition

Ld
o
O
rO
II
a

co
cr
CvJ

a
+
S
CO
o
o
CM
to
Ld

II
CM
a:
eg Ld
(/)

52
 in
in a O
O c X
a CO
X > ID
c <\J
CO
in o CO
ID >
('~S
w
1 1 LU 00
to
LU 1 1 II
ID II
ID ■— ' rO
1 CO in + rO
+
a C\J ID
O a ID
CO X
o 1 CO 00 1
o
o ID m o rO
6
C\J <D N ro
CO CO X
UJ CO 1 LU in
II II ii II ll II c
o
£K a: V-*
o
LU LU o CO
O <-»
ro -Q
D
CO
CO
♦-»
CO
Q
5.0 V

CO "co
cr o
c\j
o
CD

CD
c
(\J
U1 > 5
<D o
JZ
LO CO
00
I CO
E
CD
x:
o
CO
»_
CD
>
O
CO
CD
0C
CN
i

CD

53
54
RESOLVERS

to the amplifier will be —1.74V. The minus sign indicates an out-of-

phase voltage. The servo motor will then rotate in the correct direction,
determined by the phase of Eri, in order to drive the rotor of resolver
No. 2 to the 30° position. At this point, the voltage to the amplifier will
be zero, the motor will stop, and the data transmission will be 0 - a = 0.
In a circuit such as this, resolver No. 1 is the reference element, and re¬
solver No. 2 is the feedback element and error detector.

4.8. PERFORMANCE CHARACTERISTICS

The equations and examples shown are for an ideal resolver. The
practical resolver differs from the ideal in the following respects:

1. Functional error—The output voltages that are induced are not

exactly sine or cosine functions of the rotor angle.
2. Interaxis error—The voltages at the null positions are not zero but
some small value.
3. Null output—Null voltage or quadrature voltage exists due to the
quadrature component.
4. Transformation unbalance—The transformation ratio between the
pairs of windings is not exactly 1:1.
5. Proportional voltage error—A deviation may exist between the actual
mechanical shaft angle and the ideal position required to generate a
perfect sine or cosine function.
6. Phase shift—A phase shift of the secondary voltage with respect to
the primary voltage may occur.

4.9. ZEROING A RESOLVER

In testing or using a resolver, great care must be used in the mechan¬

ical means used to position the rotor. On many units, a calibrated dividing
head has proved to be the most satisfactory. The error in the head should
be less than 30 seconds of a degree of angle. There may also be a line
on the housing and a radial index line on the end of the shaft for the zero
reference.
The procedure for zeroing is simply to connect one of the stator
windings (Type A resolver) across rated voltage and then to measure the
output voltages with a precision voltmeter. One rotor winding should
show maximum voltage corresponding to the cosine of 0°, while the other
should give zero voltage corresponding to the sine of 0°. If the voltages
4.9. Zeroing a resolver 55

are not maximum or zero, then the housing of the resolver should be
rotated, while the pointer is held at zero, until these conditions are
achieved. Then the housing should be clamped in this position. Figure
4-13 shows the exploded view of a four-coil resolver.

PROBLEMS

1. An AC LVDT has the following data: input 6.3V, output 5.2V,

range +0.50 inches. Determine:
(a) the wiring diagram with polarities
(b) the plot of output voltage vs. core position for a core movement
going from +0.45 inches to —0.30 inches
(c) the output voltage when the core is —0.25 inches from center
(d) whether this output voltage is in phase or out of phase
2. A four-coil resolver is used to convert slant range and elevation angle
to horizontal range and height. Assume that one volt is proportional
to 500 feet.
(a) With 5V applied to stator winding S2-S4 and the rotor rotated at
20°, determine the horizontal range and height in volts and feet.
(b) Repeat problem 2a for 18V excitation and 210° rotation.
(c) Repeat problem 2a for 12V excitation and 320° rotation.
3. The resolver of problem 2 is used to convert horizontal range and
height to slant range and elevation.
(a) With 12V applied to S1-S3 and 18V applied to S2-S4, determine
the slant range in volts and feet and the elevation angle.
(b) Repeat problem 3a for 10V and 15V excitation.
(c) Repeat problem 3a for 24V and 8V excitation.
4. For a resolver rectangular-to-polar transformation loop, assume
x = 17V and y = 17V and that the output shaft is initially at 30°.
What is the voltage applied to the amplifier? What is the null angle?
5. The radar on an aircraft carrier picks up the blip of an enemy ship
and supplies the distance and angle that the carrier is from the target.
Other ships in the fleet must be informed of the true-bearing co¬
ordinates so they can converge on the target. The carrier is steaming
on course on the x-axis, which lies 40° NNE from true North. The
radar indicated that the enemy ship is 42 miles away at an angle of
70° NNW. Determine:
(a) the relative-bearing coordinates in volts and feet
(b) the true-bearing coordinates in volts and feet
(c) the vector diagram
56
REAR END BELL
WITH BEARING

FIG. 4.13. Exploded View of a 4-coil Resolver (Reeves Instrument Corp.)

4.9. Zeroing a resolver 57

6. For a resolver data transmission loop, assuming ES1 = 20V, what is

the voltage applied to the amplifier for the following conditions:
(a) 6 = 30°, oc = 26°
(b) 0 = 30°, a = 34°
7. Two four-coil resolvers are used in a data transmission loop where the
output voltages of the second resolver represent angular addition of
the rotor positions. With 10V applied to stator windings S2-S4 of
resolver 1, and the two rotors set at 30° and 40° respectively,
determine:
(a) the voltage applied to both stator windings of resolver 2
(b) the output voltages of resolver 2
(c) if angular addition is being accomplished; check by making a
vector diagram
8. The connections from the rotor winding R2-R4 of resolver 1 to the
stator winding of resolver 2 are reversed. With the same excitation
voltage and rotor positions as in problem 1, determine:
(a) the voltages applied to the stator winding of resolver 2
(b) the output voltages of resolver 2
(c) if angular subtraction is being accomplished; check by making a
vector diagram
 FIVE

Synchros

5.1. GENERAL

Synchros are electric machines that are widely used in modern control
and signaling applications where it is often necessary to control the
angular position of one shaft by positioning another shaft. They are used
as an electrical means of transmitting shaft position information, even
though a considerable distance may lie between the two shafts. The two
shafts are tied together electrically in such a manner that when one is
turned, the other will turn in exactly the same way or in a predicted way.
There is no mechanical connection between the two shafts, and this
represents the great advantage of the use of synchro devices.
The name synchro originated in the United States Navy, and it means
a device that is “self-synchronous”; that is, they are “synchronous” with
themselves. Other common names for the same machine are selsyn,
autosyn, teletorque. The names magnesyn and telegon, however, are not
to be confused with a synchro because their principle of operation is
different. The different kinds of synchros can be classified approximately
by size and accuracy as:

1. power synchros
2. instrument synchros
3. control and indicator synchros

Power synchros are really wound-rotor induction motors whose

rotor voltages must be of the same value and whose rotors must operate
at the same per unit speed. They may or may not have the same number

58
5.2. Basic synchro systems 59

TABLE 5-1
Type Nomenclature Input(s) Outputs)

Transmitter Control—CX Mech. Shaft Angle Elec. Synchro Data

Torque—TX

Receiver CR Elec. Synchro Data Mech. Shaft Angle

TR

Control CT 1. Elec. Synchro Data Error Voltage

Transformer 2. Mech. Shaft Angle

Differential CDX 1. Elec. Synchro Data Elec. Synchro Data

Transmitter TDX 2. Mech. Shaft Angle (Sum or Difference)

Differential TDR 1. Elec. Synchro Data (A) Mech. Shaft Angle

2. Elec. Synchro Data (B) (Sum or Difference)

of poles. However, their open-circuit rotor voltages must be the same so

that their rotor voltages will match at any speed. Some examples of their
use are:

1. operation of large valves

2. printing presses
3. kiln feeders
4. exact synchronism of the ends of a vertical lift bridge

Instrument synchros are usually extremely small and light. They are
used only to carry a light-weight pointer to indicate position because
they develop a very small torque. They are constructed like a miniature
rotor-wound induction motor except that the rotor is generally wound for
single-phase excitation and has two definite field poles.
Control and indicator synchros are constructed in the same manner
as instrument synchros, although they may be smaller in size. Power
and instrument synchros are usually not too accurate whereas control and
indicator synchros have errors that are measured in tenths of degrees or
less.
Table 5-1 presents the different synchros and their data in terms of
input and output.

5.2. BASIC SYNCHRO SYSTEMS

Synchro systems can be divided roughly into two groups depending

on whether they are used to give:

1. torque transmission, or
2. voltage indication
60 SYNCHROS

FIG. 5-1. Connections of a Torque Transmitter and Receiver

The first type of system is made up of two synchros between which

torque is transmitted electrically in order to rotate the shaft of the motor
or receiver synchro in synchronous position or correspondence with the
shaft of the generator or transmitter synchro. The receiver shaft may
carry a dial for indication purposes, or it may drive some other small
mechanical device. A typical circuit is shown in figure 5-1.
The second type of system is also made up of two synchros connected
together electrically in order to induce voltage in the synchro control
transformer. This voltage has a polarity and a magnitude which are
functions of the relative direction and amount of displacement between
the shafts of the two synchros. This voltage can be used for indication
purposes, but it is usually used as a control voltage for a much larger
servomechanism system. A typical circuit is shown in figure 5-2.

0°

FIG. 5-2. Connections of a Torque Transmitter and a Control Transformer

5.4. Electrical zero 61

5.3. BASIC SYNCHRO THEORY

The rotor of a synchro is wound with a single-phase concentrated

winding, and the stator is wound with three distributed windings that may
be connected in either wye or delta to give poles 120° apart. Electrically,
in normal operation, the synchro acts as a transformer, and voltages and
currents in the unit are all single-phase. The rotor winding is excited by
an ac source and by transformer action between the rotor and stator
windings, voltages are induced in the three stator windings, the magnitude
of these voltages depends on the angular position of the rotor with respect
to the stator. These voltages are of line frequency, and they are in phase
with the rotor voltage. The voltage in each coil of the stator varies
sinusoidally with the angle of the rotor with respect to that coil.
The most common transformation ratio between the rotor and a
stator coil is 2.2:1 step-down. Thus, with 115V across the rotor, there is a
maximum of 52V across any one stator coil; this maximum occurs when
the stator coil is aligned with the rotor. The stator voltage value for any
rotor angle may be found by applying the formula:

E = 52 cos <fi (5-1)

where E is the maximum voltage in any one stator coil, and <p is the angle
between rotor and stator coil.

5.4. ELECTRICAL ZERO

In measuring the angular position of a synchro shaft, some point

must be chosen as the reference or starting point. This reference is usually
taken as being the position in which the axis of the rotor is lined up with
the axis of the stator coil No. 2. Thus, the rotor position is described in
terms of the number of degrees that the rotor is rotated away from this
reference position. This reference point is called electrical zero. Under
these conditions, the voltage between SI and S3 must be zero, and the
phase of the voltage at S2 must be the same as the phase of the voltage at
R1 (see figure 5-3).
A synchro is, in effect, a transformer that has one coil which may be
rotated through 360°. The magnetic field within the synchro may also be
rotated through 360°. If an electromagnet is inserted in this field, and it is
pivoted in such a manner that it is free to turn, then it will always tend to
line up with the magnetic field. This is the basic principle underlying all
synchro operation. In figure 5-3, the three stator coils are shown 120°
apart. If SI and S3 are tied together, and ac is impressed across S2 and
62 SYNCHROS

0°
0° 0°

26 V
52 V
26 V

FIG. 5-3. Schematic and Resultant Fields for Electrical Zero

S3, then the field will be as is shown by the arrows. The rotor coil, if
connected across ac, will now line up with the field of the stator and will
stay in this position even though the alternating wave goes through its
complete cycle. It is possible to get the rotor to change its position by
180°; this will be discussed later.
For the conditions shown in figure 5-3 with 115V across the rotor
coil, there must be 78V measured between S2 and S3, also between S2 and
SI, while zero volts is measured between SI and S3. Thus, the resultant
field is proportional to 78V and points at the zero degree position. The
78V figure comes about because the coil S2 is in phase with the rotor and
has 52V induced in it; coils SI and S3 meanwhile make an angle of 60°
with the rotor and therefore have 26V induced in them. The vector sum
is 78V.
In any particular synchro, these voltages in the electrical zero position
are as indicated in figure 5-4, all are in phase with the rotor voltage.
Figure 5-4(a) shows the voltage measured from SI to S2 when the rotor is
turned to any given angular position; figure 5-4(b) shows the voltage
measured from S2 to S3; and figure 5-4(c) shows the voltage measured
from S3 to SI. For the zero degree position, the voltages may be read in
the following manner:

• The voltage from SI to S2 equals 78 volts and is 180 degrees out of

phase with the voltage from R1 to R2.
• S2 to S3 equals 78 volts; in phase with voltage from R1 to R2.
• S3 to SI equals zero volts.

Voltages above the zero-axis are in the same phase as the rotor
voltage R1-R2 while voltages below the zero axis are out of phase with
5.5. Summary 63

(a)

Esi — S2 ~ K cos[6 — 240)

- K cos8 (5-2)

(b)

ES2-S3 = Kcos©
- K cos (S — 120) (5-3)

(c)

ES3_ s, = K cos (Q — 120)

- K cos [6 — 240) (5-4)

Rotor Shaft Position

CCW Rotation

FIG. 5-4. Phase Relationship of Voltages between Stator Terminals of a Synchro

rotor voltage R1-R2. The stator voltages are either in phase with the rotor
voltage or 180° out of phase; no position in between exists. Thus, for a
90° CCW rotational position of the rotor, the voltage from SI to S2 and
from S2 to S3 must be half as large as the voltage from S3 to SI, and it
must be 180° out of phase with voltage from R1 to R2 (counteracting the
in-phase voltage of S3 to SI).

5.5. SUMMARY

The most important factor in analyzing synchro operation is the

resultant field that is produced. Each stator coil produces a field of its
own with a magnitude proportional to the voltage induced in it. The
three stator fields then combine to form a resultant field with a fixed
64 SYNCHROS

magnitude. To determine the direction of the resultant field, the following

rules might be helpful.

1. In a transmitter, the field of any stator coil always opposes the field
of a rotor coil; therefore, the resultant field in the stator always
opposes the rotor field.
2. The resultant field in a receiver is in opposition to the resultant field
of a transmitter stator (or in the same direction as the transmitter
rotor field).
3. The field of a receiver rotor tends to line up in the same direction as
the resultant field of the receiver stator.
4. When the rotor coils of a transmitter and receiver are lined up, the
rotor coils induce equal and opposite voltages in their respective
stator coils, thus causing the resultant fields to cancel. Therefore, no
current flows between the two stators.

The fields of the coils in a stator, therefore, combine vectorially to

produce a single field of a definite strength and in a definite direction.
The movement of a receiver rotor occurs because the rotor’s magnetic field
aligns itself with the resultant stator field that is created. Thus, in any
transmitter, the rotor and stator fields always oppose one another;
however, in a receiver, they always aid one another.

5.6. HOW A RECEIVER FOLLOWS A

TRANSMITTER

A fundamental circuit showing the electrical connections between a

synchro transmitter and a receiver is illustrated in figure 5-1. Even though
the distance between the units may be considerable, the rotor coils must
be connected to the same single-phase supply source. If the transmitter
rotor is set on the zero position, the stator voltage pattern will be that
shown in figure 5-4. If the receiver is not on the zero position, then current
will flow in the stator coils of the receiver, and its rotor, if free to turn, will
line up with the zero position. Thus, the voltages induced in the two stator
windings are equal and opposite to each other, and, therefore, no current
will circulate in the windings.
If the transmitter shaft is turned to the 30° position, and the rotor of
the receiver is momentarily held in its original zero position, then the
voltage balance between the two stator windings is changed, and circu¬
lating currents will flow between the two windings. Figure 5-5 shows this
voltage imbalance, and the current direction is indicated by the arrows.
5.7. Synchro torque 65

O o
30 0

Y X

TX TR

FIG. 5-5. Electrical Relations between a Transmitter and Receiver when

They Are at Different Mechanical Angles

This current flow produces a torque which tends to turn the rotor of the
receiver to a position that is in alignment with the stator field. When
alignment occurs, the voltages induced in the stator windings of the
receiver are equal and opposite to the voltages induced in the stator
windings of the transmitter.
It is important to realize that there is also a torque produced in the
transmitter that is equal to that produced in the receiver if both synchros
are not at the same shaft angle. If the transmitter is turned to 30° and
then released before the receiver has “followed-up,” the transmitter would
be forced to turn backwards towards 0°, and the system would synchronize
at an angle between 0° and 30°. This is the reason why the transmitter is
generally not allowed to turn by itself but is held by the position of the
input shaft.

5.7. SYNCHRO TORQUE

When a receiver is in correspondence with a transmitter, there is no

torque being produced and no current circulating between the units.
However, it is quite difficult to get the two to line up exactly without some
lag of the receiver shaft behind the transmitter shaft. The smaller this lag
is, the greater the accuracy of the system is. The rotors of two units in a
typical system may line up with a misalignment of less than 0.5°. This
misalignment is called the angular error. If the receiver is prevented from
turning so that it cannot line up properly, a torque is produced which
tends to turn the receiver rotor to the zero position. The currents creating
this torque are large and will cause overheating if the misalignment is
maintained.
66 SYNCHROS

The relationship between torque and angular error for any particular
synchro unit is such that the torque increases to a maximum at an angular
error of 90° and then decreases to zero at a shaft displacement of 180°.
This 180° position is unstable because, if the displacement changes
slightly in either direction, a torque will be produced which will quickly
snap the rotor back to the zero position. This means that when the re¬
ceiver is suddenly called upon to shift to a new position, it will have a
tendency to hunt or oscillate. For this reason, a small, friction-driven
flywheel is mounted on the receiver shaft to act as a damper. This also
means that synchro units will always synchronize when their rotors are
excited, and, from this, we can see why the system is called self-
synchronous.
The torque of a synchro is given in terms of torque gradient which is
a unit based on a 1° angular error. If the torque gradient is listed as
0.6 ounce-inch per degree, it would mean that a torque of 0.6 ounce-inch
would pull the receiver rotor one degree away from the transmitter rotor.
For small angles, the torque increases almost linearly. To keep the
angular error as small as possible, the torque gradient is made as large as
possible.

5.8. WIRING ERRORS

Since there are five interconnecting leads between a transmitter and a

receiver, any reversal of connections or change in connections will have a
material effect on the operation of the system as a whole. This may be
desirable when it is necessary to have a fixed, angular displacement
between the positions of the two shafts or when the direction of rotation
should be opposite between the two shafts.
Reversal of the rotor leads will produce a 180° displacement of
shafts, but the direction of rotation will remain the same. Reversing the
rotor leads and reversing S1-S2 leads will produce 60° shaft displacement
as well as a reversed direction of rotation. If the rotor leads are reversed
back to their original connection, and S1-S2 leads remain reversed, then
240° shaft displacement and a reversed rotation will result. A cyclic
interchange of stator connections (transmitter S1-S2, S2-S3, S3-S1) will
produce 120° shaft displacement with direction of rotation remaining the
same. Therefore:

1. Rotor Wiring Errors always introduce 180° mechanical shaft errors

but with correct direction of rotation.
2. Stator Wiring Errors involving any pair of leads always give reversed
direction of rotation; those involving S1-S2 or S2-S3 give 120° shaft
5.9. Differential Synchro 67

displacement as well; and those involving S1-S3 give no shaft dis¬

placement.

5.9. DIFFERENTIAL SYNCHRO

A differential synchro is a unit that is used to modify a basic synchro

system. It may indicate the difference between the relative positions of
two transmitter shafts; it may sum up the angular displacements of several
shafts and represent the total as the displacement of one shaft; it may
meter the rate of fuel flow in a multiengine plane; it may modify the angle
between a transmitter and receiver shaft; or it may have any of a multitude
of other uses.
A differential unit is constructed very much like a synchro transmitter
except that the rotor has three sets of coils that are equally spaced and
connected to produce poles 120° apart. Both stator and rotor windings
have the same number of turns; therefore, the transformation between
the corresponding stator and rotor coils is 1:1. This means that the
differential, when the rotor is held in position, will have the same voltage
at the rotor terminals as at the stator terminals. A differential thus can be
constructed and connected to work as a differential transmitter or differ¬
ential receiver.
If a differential transmitter is inserted between a transmitter and a
receiver, and the rotor of the differential is not free to move, then the
receiver rotor follows the transmitter rotor just as if the differential were
not in the circuit.
Figure 5-6 illustrates the connections for a differential transmitter
that is inserted between a transmitter and a receiver. Here, the transmitter
rotor is held at zero position, and the differential rotor is turned to the 120°
position. The maximum voltage is now induced in coil R1 of the differen¬
tial rotor, and, since R1 is connected to SI of the receiver stator, the
resultant field is now in the direction of SI, and the receiver rotor turns
clockwise to the 240° position. Therefore, the position of the receiver
rotor is equal to the position of the transmitter minus that of the differen¬
tial transmitter. Thus, the following rule holds true: TX - TDX = TR.
If one desires to add the position of the differential transmitter to
that of the transmitter, then it is necessary to reverse the connections to
S1-S3 and R1-R3 of the differential. If the circuit is analyzed, then the
following rule can be proven: TX + TDX = TR.
If only the transmitter is reversed, that is S1-S3, the rule becomes:
XX + TDX = -TR. If only the receiver connections S1-S3 are re¬
versed, then the rule becomes: TX - TDX = -TR.
68
FIG. 5-7. Connection of a Torque Differential Receiver that Will Give Subtraction
5.10. Synchro control transformer 69

A differential receiver is inserted between two transmitters in order

to indicate the angle between their shafts. If there is no angular difference
between the shafts of the two transmitters, then the shaft of the differential
will be at the zero position. However, if the transmitters are set to two
different angles, then the differential will indicate the difference between
the two. This is illustrated by figure 5-7, which shows that 1TX —
2TX = TDR.
If one desires to get a minus TDR reading, then S1-S3 should be
reversed for both transmitters so that 1TX — 2TX = —TDR. If one
desires to add the two transmitter readings, then only R1-R3 of the
differential need be reversed; under this condition, 1TX + 2TX = TDR.
If only S1-S3 of the differential is reversed, then 1TX = 2TX = —TDR.
It can be seen that in either of the differential connections, a differen¬
tial is not connected to the source of excitation; therefore, its exciting
current must come from either of the standard synchros to which it is
connected. It is normal practice to supply this excitation to the stator of
the differential only, and this means that the synchro transmitter must be
designed to carry this extra current. A specially designed unit, called a
synchro exciter, is quite often used, and it may operate in the system
either as a transmitter or as a receiver. In still other cases, balanced static
capacitors are connected across the stator terminals of the differential
unit. These capacitors have such a rating that the leading current they
take is exactly equal to the lagging current of the differential stator
winding. Thus, the synchro transmitter is required to furnish only the
energy component of current, which is perhaps 10% of the total excitation
current.

5.10. SYNCHRO CONTROL TRANSFORMER

A synchro control transformer is constructed very similarly to a

differential unit except that it has a single rotor winding like a transmitter.
It is used with a synchro transmitter as is shown in figure 5-2. The rotor
winding is not connected to the supply source, and, under normal con¬
ditions, it carries no current. It is used to produce an output voltage that
is proportional to the angular difference between the two shafts. This
output voltage is called an error voltage because it exists only when the
shafts are in “error” or misaligned.
If the shaft of the transmitter is set at zero position, then the magnetic
field in the control transformer is in line with phase S2, and, if the shaft
of the control transformer is turned to the zero position, then its winding
axis will be at 90°, while the magnetic field and the output voltage will be
at zero position. For any other position of the control transformer shaft,
70 SYNCHROS

e,

FIG. 5-8. Connection of a Transmitter and Control Transformer with Amplifier and
Servomotor

a voltage will be induced in the rotor coil that varies with the sine of the
angle between the shafts. In the electrical zero position, coils SI and S3
induce voltages in the rotor, but these voltages are equal and opposite in
phase; therefore, the net result is zero.
When the rotor is turned so that it lines up with S2 (90° from zero
position), the voltage induced in it will be a maximum of 57.5V. Rotating
the shaft another 180° will again induce a maximum voltage but of opposite
phase. Thus, the error voltage from a control transformer varies both in
magnitude and in phase. The output voltage is thus given by e0 =
57.5 sin (0X — 60) with 0X being the transmitter position and 0o being
the control transformer position. This makes the control transformer very
useful in any system where position is being controlled. In these position
control systems, the output voltage is normally fed into an electronic
amplifier as an error signal. For this reason, the control transformer is
quite often called an error detector. Figure 5-8 shows a circuit that con¬
trols the angular position of an output shaft. In this circuit, any change in
input shaft position must be followed by a corresponding change in
output shaft position.
The rotor of the synchro transmitter is mechanically connected to the
input shaft. The rotor of the control transformer is mechanically con¬
nected to the output shaft. The electrical output of the control trans¬
former rotor is fed into the input stage of a power amplifier. A two-phase
ac control motor is used to furnish the mechanical power to rotate the
5.11. Geared synchro systems 71

output shaft and its associated load. The input to the control winding of
the control motor is supplied by the amplifier. When the output shaft is
in the correct position, the error voltage input to the amplifier is zero, and,
therefore, there is no voltage impressed across the control winding of the
control motor. Thus, the motor does not turn. If any change occurs in the
input shaft position, then the control transformer generates an error
voltage, and correction takes place to bring the output shaft into alignment
with the input shaft. To operate satisfactorily, the amplifier must be
phase-sensitive so that a change in either direction of rotation may be
corrected for.

d_
dt

-8o

FIG. 5-9. Block Diagram of a Synchro Error

Channel

If one desires to have a dc control motor correct the angular dis¬

crepancy, then the output of the control transformer must be rectified in
order to produce a proportional dc voltage that is also polarity sensitive.
A diagram showing how the CX and CT synchro pair might be
shown on a block diagram is presented in figure 5-9.

5.11. GEARED SYNCHRO SYSTEMS

When the rotors of synchro systems turn together at the same speed,
the system is known as a one-speed system. However, due to internal
friction and other errors in the system, an angular displacement, even at
no-load, exists between the input and output shafts. This is called no-load
error. If any external load is applied, then a load error of a greater
magnitude arises. To achieve greater accuracy, geared systems called
double- or two-speed systems are used. To understand why this is true,
consider a synchro transmitter-receiver system used for data transmission.
Suppose that 360° of rotation represent 1200 miles, and that the accuracy
of the one-speed system is 15 minutes of a degree or 1 part in 1440. If
an accuracy of 0.0057o is required, then the error must be held to 0.06
72 SYNCHROS

mile. In order to achieve this accuracy, a 15X (15-speed) synchro pair

could be used, thus

15 revolutions = 1200 miles

1 revolution = 80 miles
15 x 80 miles = 0.055 mile
60' x 360°

Input Output

B1
B2
B3

Errors No Lood Load

loz. in loz. in
Input Output Input Output
i I

1 Speed System

No Load Lood
1 oz in. 1 oz in
Input Output Input Output
9.9307
I
I
—1

59.584°] 9.9,307c

(a) (b)

6 Speed System

FIG. 5-10. Comparison of Errors, No Load and Full Load, for a IX versus a 6X
System
5.12. Synchronism 73

The advantages of a double-speed system over a single-speed system

can be easily seen in figure 5-10. Here, a comparison is made between
the errors, no-load and full load, for a one-speed and a six-speed system.
Internal frictions are such that an error of 1 ° can be considered as common
to both systems under no-load conditions of figure 5-10(a). For the one-
speed system, the output error will be 1° and for the six-speed system it
will be 0.17°. This is a ratio of about 6:1 which shows that no-load errors
are inversely proportional to the gear ratio.
Under load conditions of one ounce-inch of torque, the load error
will be 2.5° for the one-speed system, and 0.0693° for the six-speed system.
These load errors are found by assuming that the load torque is such that
no-load errors can be considered zero and also that the torque gradient
for the synchros is 0.4 ounce-inch per degree. With a 6:1 reduction twice
between input and output, the load error in a double-speed system is
inversely proportional to the square of the gear ratio.

5.12. SYNCHRONISM

In a single-speed system, self-synchronization is always achieved

when power is turned on. However, in any geared system, there are the
same number of positions of the output shaft where the receiver will
align as there is in the number of the gear ratio. Thus, if the system for
any reason is shut down, the receiver shaft may align in a different position.
If a control transformer is used in place of a receiver, then there would be
six different null output voltages; they would occur at 60°, 120°, 180°,
240° and 300° (multiples of 360°/6). The way to overcome these multiple
nulls is to use two sets of transmitter-receivers or transmitter-control
transformers. One pair is connected to controlled shafts, and the other
pair is geared to run at the higher speed. Coarse error voltage is obtained
from the former, fine information from the latter.
The principle of the double-speed system is demonstrated con¬
tinually in one’s wrist watch or the clock on the wall. The hour hand and
minute hand give coarse and fine data respectively. In a similar manner,
the coarse CX-CT pair prevents misalignment because it does not have
multiple nulls, while the fine CX-CT pair increases accuracy.
The double-speed system shown in figure 5-10 has a one-speed (coarse)
system in conjunction with a six-speed (fine) geared system. However,
some device is required at the input to the amplifier which selects the data
to be used in operating the servo motor, a device which measures the level
of the coarse error. When this level is above a certain value, the coarse
error signal is amplified and fed to the servo motor. When the coarse
error signal drops below this value, the fine error signal controls the servo
74 SYNCHROS

FIG. 5-11. Determination of Proper Switching Between IX and 6X Synchros

motor. In selecting this error level, two criteria must be satisfied; they are:

1. The output shaft error must be less than one-half revolution of the
fine synchro; otherwise, the fine error signal will cause the motor to
be driven to an incorrect null position.
2. The level of the voltage applied to the motor should not change
suddenly when switching occurs between signals; otherwise, an
undesired transient effect will result. Both signals should be large
enough to saturate the amplifier and to make it insensitive to signal
changes.

Figure 5-11 shows the graph of the 6X and IX synchro CT output

voltages vs. output shaft positions. If the output shaft position is between
0° and 30°, then both voltages will cause positive rotation toward the
commanded null position at zero. At 30°, the fine CT has rotated 30° x 6
or 180°. Beyond 30°, the fine CT has an output of opposite polarity, and,
if it is applied to the servo motor, it would drive it away from zero.
Therefore, control must be shifted to the coarse CT before the fine CT
turns 180°. The shaded area shows the error angle over which the fine CT
may be used. Outside this area, the coarse CT must take command.
Thus, switching anywhere in this region will theoretically satisfy criteria 1.
However, if switching takes place too close to 180°, the level of the fine
signal might be too low to saturate the amplifier. In actuality, switching
should be done at point A or thereabouts. The devices that are used to
accomplish this switching are called cross-over systems or synchronizers.
Relays or all electronic circuitry can be used to recognize the level of
voltage at point A.
As the system approaches the 180° position of the coarse CT, its
output voltage drops to a very low level, point B. An error angle this large
5.12. Synchronism 75

Coarse CT +

would permit the crossover system to deenergize and to shift control back
to the fine CT. However, at this point, the fine CT voltage is out of phase,
and thus the motor is driven to a false null at 180°. This always happens
when even gear ratios are used, unless something is done to prevent it.
Even so, even gear ratios are preferred over odd gear ratios. In order to
avoid this selection of an incorrect null, a voltage called a stick-off voltage
is added in series with the output of the coarse CT, as is shown in figure
5-12. The effect of this voltage, when added to the coarse voltage, is to
shift the two null positions of the coarse unit in opposite directions. Its
amplitude is selected so that the new coarse CT null occurs when the fine
CT completes an additional rotation of 90° past 180°.
To get the proper nulls at zero error angle, it is now necessary to re¬
zero the coarse CT so that it produces a null once again at zero degree
shaft position. The new alignment of the coarse and fine nulls is shown in
figure 5-13; both CT’s now have the same phase polarity at each null.

Coarse and Fine Nulls

FIG. 5-13. Correlating Coarse and Fine Nulls Using a Stick-off Voltage
76 SYNCHROS

PTl In IT'PV-Antenna

Servo
Motor

Syn. Servo
Network Ampli.

h 'Coarse

FIG. 5-14. Double Speed Transmission System

Figure 5-14 shows a 36:1 speed system that is used by the United
States Navy to control antenna positions.
To achieve even greater accuracy, the gear trains could be eliminated
so that gear inaccuracies will not impose a limit on total system accuracy.
For example, a modern gear train has an error of around 2 minutes of
angle, while CX and CT synchros are now available with less than 2
minutes of error; thus, the gear train governs the total error. If the mechan¬
ical gear train could be replaced with electrical gearing, then the total error
could be reduced. The use of multipole synchros does exactly this.
When multipole synchros are used in two-speed systems, the gear
inaccuracies are replaced with the synchro inaccuracies which include any
errors caused by the electrical gearing, and they are included in the
manufacturer’s specified synchro error. Inaccuracies in the range of 5 to
10 arc seconds or lower are now available; therefore, the overall system
accuracy is also this amount. Other advantages of the multipole synchro
are its driving torque, size, and shape.
5.15. Zeroing a transmitter—TX and CX with a voltmeter 77

5.13. THE NEED FOR ADJUSTMENT

In order for synchros or any other transducers to work together

properly, it is necessary that they be correctly connected and aligned to
each other and to the other devices with which they are used.
Electrical zero is the reference point for aligning all synchro units.
The mechanical reference point for any devices connected to the synchros
depends on the particular application. For example, on board ship where
course data are being checked, the reference would be true North. On a
gun director servo, the reference could be a specific distance or an angle.
No matter what the system, the mechanical and electrical zeroes must
coincide.
The best way to accomplish this alignment is to align all synchros to
electrical zero. The major advantage here is that if wiring errors occur,
the trouble in the system will show up in the same way. Sometimes, a
synchro can be zeroed by adjusting it mechanically by physically turning
the synchro rotor or stator.
A synchro transmitter is on electrical zero when rated voltage on the
rotor winding, R1 and R2, will induce no voltage between SI and S3,
and when the rotor and stator voltages subtract if R1 is connected to S2.
A synchro control transformer is on electrical zero when rated voltage
connected between SI-S3 and S2 induces no voltage in the rotor, and when
the rotor and stator voltages subtract if R1 is connected to SI.
A differential transmitter is on electrical zero when rated voltage
connected between SI-S3 and S2 induces no voltage between R1 and R3
of the rotor, and when the rotor and stator voltages subtract if R2 is
connected to S2.

5.14. ZEROING PROCEDURES

There are a number of methods that may be used to find electrical

zero, but two major steps should be used. The two steps are a coarse or
approximate setting and a fine setting. The coarse setting may be ac¬
complished by a physical adjustment because the frame may have an
arrow stamped on it, and a line may be engraved on the shaft end. When
these line up, coarse adjustment is made. If these marks are not on the
synchro, then an electrical test must be made.

5.15. ZEROING A TRANSMITTER—

TX AND CX WITH A VOLTMETER

When rated voltage is applied to R1 and R2 as shown in figure

5-15(a) and a VTVM is connected between SI and S3, the voltmeter should
78 SYNCHROS

0-250 V

(b) Fine Setting

FIG. 5-15. Electrically Zeroing a Transmitter

read zero or minimum when the synchro is at the zero position. However,
a zero or minimum voltage will also result at 180°, unless the phasing
between the rotor and stator voltages is carefully checked. Therefore,
proceed in the following manner:

1. Connect the synchro and voltmeter as shown in figure 5-15.

2. Turn the rotor or stator until the meter reads a minimum voltage,
about 40V. This is the coarse setting. See figure 5-15(a).
3. Reconnect the synchro and meter as shown in figure 5-15(b). Adjust
the rotor or stator for a null or minimum reading of less than 1.0V.
This is the fine setting, and the synchro can be clamped at this
position.

5.16. ZEROING A RECEIVER—

TR AND CR BY ELECTRICAL LOCK

Remove all connections and reconnect as illustrated in figure 5-16.

The shaft and pointer will turn to 0°. Set the dial or pointer at this exact
zero position. The voltage applied to R1-R2 should be 0.866 of the source
voltage because this is the normal voltage induced in S2-S3 at 0°. If 78V
is not available, then 115V may be used for only a few seconds because
higher than rated voltage causes the synchro to overheat.
If the rotor is not free to turn, then a receiver can be zeroed in a
manner similar to that used for zeroing a transmitter.
5.18. Zeroing a differential transmitter—TDX using a voltmeter 79

FIG. 5-16. Zeroing a Receiver by Electrical

Lock

5.17. ZEROING A CONTROL TRANSFORMER—

CT USING A VOLTMETER

1. Connect the control transformer and voltmeter as shown in figure

5-17(a).
2. Turn the rotor or stator to obtain a minimum voltage reading; this
will be the coarse setting.
3. Reconnect the voltmeter as shown in figure 5-17(b) and adjust the
rotor or stator for the null or minimum voltage. Clamp the CT in
this position.

5.18. ZEROING A DIFFERENTIAL TRANSMITTER—

TDX USING A VOLTMETER

1. Connect the differential transmitter and voltmeter as shown in figure

5-18(a).

FIG. 5-17. Electrically Zeroing a Control Transformer

80 SYNCHROS

(b) Fine Setting

FIG. 5-18. Electrically Zeroing a Differential Transmitter

2. Unclamp the differential and turn it until the meter gives minimum
reading. This is the approximate zero.
3. Reconnect the unit as shown in figure 5-18(b) and turn the differential
until a null reading is obtained. This is the fine zero setting.

PROBLEMS

1. A transmitter is at zero degrees; the voltage from R1 to R2 must be

in phase with the voltage from: (a) S2 to S3, (b) SI to S2, (c) SI to
S3, (d) the common point to S2. Why?
2. A receiver has its shaft in the 0° or 180° position when the axis of the
rotor coil lines up with the: (a) axis of stator coil S2, (b) axis of stator
coil SI, (c) axis of stator coil S3, (d) midway between the axis of
coils S2 and S3. Why?
3. A synchro system is a: (a) single-phase system, (b) three-phase
system, (c) two-phase system. Why?
4. Maximum current flows in the stator coils of a receiver when its rotor
position: (a) is the same as that of the transmitter, (b) differs from the
transmitter by 180°, (c) differs from the transmitter by 90°, (d) is
displaced by 120° from the transmitter. Why?
5. When the shaft of a transmitter is set at the 120° position, the voltage
measured between SI to S3 will be: (a) 78V, (b) 52V, (c) 26V. Why?
6. Determine the voltage measured between each pair of stator leads of
a transmitter for the following shaft positions: (a) 25°, (b) 150°
(c) 240°, (d) 330°.
5.18. Zeroing a differential transmitter—TDX using a voltmeter 81

.
7 A transmitter and receiver are zeroed properly and are then con¬
nected together in straight cyclic connection. Rotate the transmitter
shaft to 180° and, at the same time, hold the receiver shaft at the 90°
position. Make up a schematic diagram of the connections showing
the stator coil voltages for both units.
8. Allow the shaft of the receiver in problem 7 to come into alignment
with the shaft of the transmitter and then make up a schematic
diagram of the connections showing the stator voltages for both units.

For Problems 9 through 16 use Magnetic Field Analysis as proof.

9. A transmitter and a receiver are connected together with SI and S3

reversed at the receiver. If the transmitter shaft is turned to 335°, the
receiver shaft will indicate: (a) 25°, (b) 155°, (c) 335°, (d) 215°. Prove
your answer by a schematic diagram, also show the original starting
position.
.
10 A transmitter is connected to a receiver in the following manner:
S1-S2, S2-S1, S3-S3, Rl-Rl, and R2-R2. If the transmitter is
turned to 240°, find (a) the position of the receiver, and (b) the
starting position.
.
11 A transmitter is connected to a receiver in the following manner:
Sl-Sl, S2-S3, S3-S2, R1-R2, and R2-R1. The transmitter is set at
60°, find the position of the receiver.
.
12 Repeat problem 10 except reverse the rotor connections to the
receiver.
.
13 A differential transmitter is inserted between a transmitter and a
receiver. The transmitter stator is connected to the differential stator
in the following manner: S1-S3, S2-S2, and S3-S1. The differential
rotor is connected to the receiver stator in the following manner:
R1-S3, R2-R2, and R3-S1. The rotors of the transmitter and re¬
ceiver are connected Rl-Rl and R2-R2. Determine the position of
the receiver if the transmitter is set at 90°, and the differential is set
at 150°.
.
14 A transmitter is connected to a differential transmitter in the following
manner: S1-S3, S2-S2, and S3-S1. The differential in turn is con¬
nected to a receiver by Rl-Sl, R2-S2, and R3-S3. The rotors of the
transmitter and receiver are connected Rl-Rl and R2-R2. Determine
the position of the receiver if the transmitter is set at 50°, and the
differential is set at 80°.
.
15 A transmitter is connected to a differential transmitter in the following
' manner: Sl-Sl, S2-S2, and S3-S3. The differential in turn is con¬
nected to a receiver by Rl-Sl, R2-S2, and R3-S3. The rotors of the
82 SYNCHROS

transmitter and receiver are connected Rl-Rl and R2-R2. Determine

the position of the receiver if the transmitter is set at 130°, and the
differential is set at 30°.
.
16 Two transmitters are connected to a differential receiver in cyclic
fashion. Determine the reading of the receiver when 1TX reads 135°
and 2TX reads 75°.
.
17 The axis of the rotor in a synchro control transformer when the zero
position is: (a) at right angles to the axis of S2, (b) lined up with the
axis of SI, (c) lined up with the axis of S2, (d) lined up with the axis
of S3. Why?
.
18 A transmitter, control transformer, servo amplifier and servo motor
are connected as an instrument servomechanism. Assuming that the
transmitter is set at zero degrees, determine the input voltage to the
amplifier for the following cases and indicate which voltages are
out of phase:
(a) Transmitter input to 50°
(b) Transmitter input to — 135°
(c) Transmitter input to 210°
(d) Transmitter input to —210°
(e) Transmitter input to —60°
.
19 A 24-speed geared synchro system has a no-load error of 1.75°.
What is the error at the output?
.
20 What is the relationship between the gear ratio and the load error in
a geared 24-speed system? Assume that the load error is 1.6°.
 SIX

Mathematics of block diagrams

6.1. GENERAL

The purpose of a block diagram in any system, whether it be

electrical, mechanical, hydraulic, pneumatic, chemical, etc., is to simplify
the knowledge of the operation of the total system by conveying a better
understanding of the relationship of the various parts. Only essential
details are shown in order to give functional operation of all elements
within the total system. The block diagram is combined with the transfer
function of each element in the system to give a clear picture of the input
and output signals. The transfer function which is defined as the ratio of
the output signal to the input signal is placed within the block representing
each element.
The techniques used in transforming and/or reducing block diagrams
from very complicated structures to a single block are based on the use of
a few identities and algebraic rules. To assist in performing the required
algebra, other definitions of symbols concerning block diagrams need to
be made. These symbols are:

1. The block represents the transfer function of the system element.

83
84 MATHEMATICS OF BLOCK DIAGRAMS

2. The summing junction represents the place where addition or sub¬

traction of different signals is done.

3. The take-off points represent the places where more than one input
or output signal is used.

4. The directional arrows represent the flow of the signal.

The identities and other short cuts used in simplifying complex

block diagrams are shown in the following pages and examples.

6.2. DIAGRAM MATHEMATICS

An open-loop control system that has a gain of G is shown by the

block diagram depicted in figure 6-1.
A closed-loop system that has a gain of G is shown by the block
diagram depicted in figure 6-2. Figure 6-2(a) shows the feedback to be
unity or 1.0, while figure 6-2(b) shows a feedback factor of H which can
be varied.
In equations 6-2 and 6-3 when the denominator has a plus sign, we
know that the feedback voltage, ef, subtracts from the reference or input
voltage, et and gives negative feedback. Negative feedback always gives
stable system operation with the following four characteristics: (1) gain
is reduced, (2) stability is increased, (3) response time is slowed down, and
(4) bandwidth is increased. Negative feedback is the foundation upon
which all automatic control is built; it is illustrated graphically by figures
9-2, 9-3(b, c, d).
On the other hand, when the denominator has a minus sign, we know

e1 eo
eo eo ~ Ge-i = G (Open Loop Gain) (6-1)

FIG. 6-1. Open-loop Diagram and Gain

6.2. Diagram mathematics 85

(b)
CD
II

<D

d)
II
<D
1

o
ee ~ er ~ ef ef
= Gten - e0) e0 = G(en - He0)
— Ge1 — Ge0 = Ge! — HGe0
= e0 + Ge0 Ge, = e0 + HGe0
= e0(1 + G) = e0(1 + HG)
eo . G (6-3)
- G (6-2)
1 + G ei 1 + HG

FIG. 6-2. Derivation of System Gain for Closed-loop Systems

that the feedback voltage, ef, adds to the reference or input voltage, ej
and gives positive feedback. Positive feedback leads to unstable system
operation because, as the resultant denominator approaches zero, the
oscillations become larger and larger. When the open-loop gain, G or HG,
equals unity, the system becomes a pure oscillator because overall gain is
infinite. This is shown by curve A in figures 9-3 and 9-4. What is actually
happening is that the feedback signal strengthens the error signal so that
the actual error is compounded to the point where the system saturates,
and considerable damage may result to any rotating components and/or
other devices. Positive feedback has the following characteristics: (1)
very high gain (up to infinity), (2) very unstable, (3) extremely fast response
time (digital in nature), and (4) a very narrow bandwidth. Positive feed¬
back is alright in some electronic devices, but it is not used in automatic
control systems.
The systems of figures 6-2(a) and 6-2(b) can now be represented by
single blocks. See figures 6-3(a) and 6-3(b).
The open-loop gain in the system that is shown in figure 6-3 is equal
to G or HG. This gain may be used to determine the steady-state fre¬
quency response of a closed-loop system using the open-loop transfer
function. A transfer function allows the various elements of a servo
system to be changed in block form for the purposes of simplifying or

G G
1 + G eo ei 1 + HG

(a) (b)

FIG. 6-3. Simplified Block Diagrams of the Closed-loop

Systems of Fig. 6-2
86 MATHEMATICS OF BLOCK DIAGRAMS

reducing. The transfer function may be considered an operator because

it operates on the input to give an output. With the use of a number of
different rules, given in figure 6-4, blocks may be combined into a single
block, a pick-off point or a summation junction may be shifted ahead of
or behind blocks, all for the purpose of diagram reduction.

|Fundomentolsl

Point

| Identities 1
Rules Block Diagram Equiv. Block Dioq Equotion

C = RG,G2 (6_4)

C = RG1 + RG2
(6-5)

3. Moving o R- RG (6-6)
Pick Off
Point Beyond
a Block R

4 Moving o R- C R C C = RG
Pick Off
Point a
Block Ahead C
c

5. Moving a C = R,G ±
Sum. Point R2G (6-7)
Beyond a
Block

FIG. 6-4. Basic Block Diagram Manipulations

6.2. Diagram mathematics 87

Rules Block Diagram Equiv. Block Dioq Equation

6. Moving a C = R,G ± R2
Sum. Point
(6-8)
Ahead of
a Block

G G
b * b L — K
7. Unity Feedback 1 + G . 1 + G .
(6-9)

1 + HG
(6-10)

[Short Cuts]

3. The Input/Output Ratio Consists of Terms Equal to the Number of Possible

Paths from Output to Input, Going Against the Arrows Along the Forward Path
and With the Arrows Along the Feedback Path. Each Term is Equal to the
Product of All the Feedback Functions Along One Such Path Divided by the
Product of All the Forward Functions.

FIG. 6-4. (Continued)

88 MATHEMATICS OF BLOCK DIAGRAMS

Note: Number of Paths = Number of Terms

R. _ C , G3 , _1_ _ Gi G2G4 + G2G3 + 1 o _C_ _ gig2

C 4 ^ GiG2 Gi G2 R G1G2G4+ G2G3-M

Block Manipulations are Permissible Which do Not Alter

(1) The Product of The G's Between Input and Output or (2) The
Product Around a Feedback Loop

Block Diagrams For Networks

Impedance Admittance

Resistance

Inductance

Capacitance

Examples

1.

FIG. 6-4. (Continued)

6.2. Diagram mathematics 89

—H□— - C—J—1
'2
6 “ ; e
r il - \z. °21

Move X- to Left and Move x-to Right

Ri L2S
Change Order of of Takeoff Point
Summing Points

Reduce Inner Loops to Equivalent Blocks

1 /<C?\ R C2S 1
R^s 4- 1 R2 C2S + 1 C2 s eo
R,

Combine Blocks in Loop, Reduce to Equivalent Block

and Combine.

1
e RiR2C1C2s2 4- (R, Cn + R2C2+ RiC2)s + 1

FIG. 6-4. (Continued)

90 MATHEMATICS OF BLOCK DIAGRAMS

Some example problems will now illustrate the theory presented.

Figure 6-5 shows the reduction of a block diagram.

Example 6-1

/
Gi G2 G3
(f) Ultimate
ei
1 - G2(H - G,) Simplification

FIG. 6-5. Reduction of a Block Diagram

6.2. Diagram mathematics 91

Example 6-2

Given the network shown in figure 6-6. Find the ratio C/R by two
methods.

C = 20

“y”

Method 1 Method 2

Signal Values at

d- ?p-s
E = 20 x 1.2 = 24
F = 20 x 0.8 = 16
B = 24 x 5 = 29
29
A = = 4 83
R = 483 + 16 = 20.83
C.= 20
0.96 Ans
R 20.83

6 x 0.69
I + 0.8(6 x 0.69)
4.14
= 0.96
1 + 3.31
Ans
CK

FIG. 6-6. Determining Output/Input Ratio

Example 6-3

Given the network shown in figure 6-7, find the open-loop gain.

PROBLEMS

1. The following diagram illustrates a voltage divider network:

 00
ro
+
CO CO
(\J CO
+
<VJ
CO +
CVJ
CO
OO

00
fO
+
+■

FIG. 6-7. Determining Open-loop Gain

CO
V 00 CO
00
+
OJ
CO +
CO
00

00
-h*

92
6.2. Diagram mathematics 93

(a) Write an equation for E0 that will give an open-loop system, i.e.,
as a function of Ej, R, and R2.
(b) Draw the block diagram for problem la.
2. Use the voltage divider network of problem 1.
(a) Write the equation for E0 in closed-loop form, i.e., as a function
of Ei, E0, R] and R2.
(b) Draw the block diagram for problem 2a.
(c) Write the closed-loop transfer function equation.
3. Simplify and determine the open-loop transfer function for the
following system.

4. Determine the values of R and C of the following system by two

methods.

5. Find the transfer function for the following system.

6. Find the transfer function for the following system.

94 MATHEMATICS OF BLOCK DIAGRAMS

7. Reduce the following network to one overall block diagram.

8. Determine the ratio C/R for each of the following systems.

A R ► C

B R

9. Determine the ratio C/R of the following system by two methods.

10. Determine the ratio of C/R of the following system.

6.2. Diagram mathematics 95

11. Determine the ratio of C/R of the following system.

 SEVEN

Mechanical electrical
characteristics

7.1. GENERAL

In most servo systems, a drive motor is used with other devices such
as a gear train, a damper, or a tachometer-generator to position or to
rotate a load. The drive motor can be an air motor, a hydraulic motor, or
an electric motor. This discussion will concentrate on the electric motor
only.
In order to analyze and to determine overall system performance,
some mechanical characteristics will be specified, and then the electrical
quantities which are analogous will be identified.

7.2. ROTARY MECHANICAL MOTION

Forces on a Body. A body which is set in motion by a force has a

number of forces opposing it, i.e., the force is relative to the mass and
acceleration and the constants of friction and velocity. This force is
equal to
f = Ma + Fv (7-1)

Moment of Inertia. Inertia is the property of a body which resists a

change in its velocity in proportion to its mass. Moment of inertia refers
to the mass of a body in relation to its center of rotation when it is under¬
going a change in velocity. It is equal to

J = Mr2 (7-2)

96
7.1. General 97

It may be expressed in gram-centimeter squared, ounce-inch second

squared or dyne-centimeter squared.

Friction. Friction is generally defined as that resistance to motion which

occurs when one body is moved in respect to another. In servo systems,
the general term friction can be broken down into the following:

1. Static friction is the friction that exists when the relative motion is
zero. It is often called stiction. In a conventional electric-motor
drive system, it would be referred to as the starting friction that the
motor must overcome in order to set the load in motion.
2. Coulomb friction is sometimes called kinetic friction, and it is constant
and completely independent of relative motion. In terms of surfaces,
it would be the friction existing between perfectly dry and smooth
surfaces.
3. Viscous friction is friction which is proportional to relative motion or
velocity and can be likened to the movement of a body through a
medium such as gas or liquid. This is the friction specified in the
friction constant in equation 7-1. It may be expressed in dyne-
centimeter per radian per second, or ounce-inch per radian per
second.

Angular Position. Angular position or angular displacement is generally

defined as the angle through which a shaft is turned as it moves to a new
position, and it is measured in degrees or radians.

Angular Velocity. Angular velocity is generally defined as a change in

angular position per unit of time, i.e., the first derivative of position in
respect to time. It may be expressed in revolutions per minute, degrees
per second or radians per second, and it is given by

do = e or 9 = - (7-3)
dt t t

Angular Acceleration. Angular acceleration is generally defined as a

change in angular velocity per unit of time, i.e., the second derivative of
position in respect to time. It may be expressed in radians per second
squared, and it is given by

d^O = e_
(7-4)
d7 ~ t2
98 MECHANICAL ELECTRICAL CHARACTERISTICS

Torque. Torque is generally defined as the application of a force through

a radius, i.e., a distance measured from the center of rotation. It may be
expressed in Newton-meter, pound-foot, ounce-inch, dyne-centimeter,
etc., and it is given by

T - fr (7-5)

When defined in terms of moment of inertia and angular acceleration,

inertial torque is given by

Tj = W = J ^ (7-6)
dt2

When defined in terms of viscous friction and angular velocity, friction

torque is given by

df)
Tf = ¥9 = F — (7-7)

Power. Power is generally defined as the time rate of doing work or a

force acting through a distance per unit of time. It may be expressed in
horsepower or in watts, depending on size of the drive motor. It is given by

/VI ryi

— but f = - and d = rd
t r
Therefore
T rd
P
r t
However,
dO
rd so
dt

In the “English System,”

„ 2n x rpm x 746
r — - X (7-8)
33,000 16 x 12

which reduces to P = TS/1352 with P in watts, T in ounce-inch, and S

in rpm.
In the “Metric System,”

T x 141.6 2n
■- x S — (7-9)
1352 60
7.3. Gear trains 99

which reduces to P = Tv with P in watts, T in Newton-meters, and v in

radians per second.
If the servo uses less than 100W, then it is generally classified as an
instrument servo, while if it uses over 100W, the classification falls in
the power range.

7.3. GEAR TRAINS

In most servo systems, a gear train is used to connect the servo motor
to its load and to transducers used for feedback. A gear train usually
accomplishes the following things:

1. It reduces the high speed of a servo motor to the required load speed.
2. It changes the torque of the motor to the required load torque.
3. It may change the direction of rotation of the output shaft in respect
to the motor shaft.

A well designed gear train should introduce a minimum of inertia and

friction into the system, and, in this respect, it acts the same as an electrical
transformer of very high efficiency. The preciseness with which the gears
are made and the ratio selected to satisfy the system requirements should
provide for optimum performance. The symbol for a gear train in a servo
block diagram appears in figure 7-1.

FIG. 7-1. Block Diagram Symbol for a

Gear Train

N is the overall gear ratio of all meshes (Nx x N2 x N3, etc.) where
a mesh consists of two gears such that Nx = NL/NM, where NM equals
the number of teeth on the motor (drive) gear and NL equals the number
of teeth on the load (driven) gear.
Figure 7-2 shows a motor driving a load through a gear train with
associated parameters. These parameters are related in the following
manner:

Tm — torque on motor shaft

Tl — torque on load shaft
a>M — angular velocity of motor
coL — angular velocity of load
NM — number of teeth on motor gear
100 MECHANICAL ELECTRICAL CHARACTERISTICS

Nl — number of teeth on load gear

N — gear or speed ratio, motor to load
Fm — viscous friction of motor
Fl — viscous friction of load
JM — actual motor inertia
JL — actual load inertia
rM — radius of gear on motor
rL — radius of gear on load
F — force transmitted between gears
v — velocity of gears

The linear velocity of both gears will be the same, and the forces at
the point of contact will be equal and opposite. Therefore

v = rMcoM = rLcoL and °h. = —

— = —=-
®M N

rM _ *M
CO. - -rr
rL N

F — — = — (7-10)
'M

TL = — T,M NT,M (7-11)

Tm®m — ~ (Ncol) — Tlcol (7-12)

7.3. Gear trains 101

P, the horsepower transmitted by the gear train, is equal to

Tl^l _ Tm(Qm
(7-13)
550 ~ 550

while the torque required to accelerate the load inertia alone is

dcoL
Tl — Jl ~ — JLaL (7-14)
dt

where dco/dt = a = angular acceleration in rad/s2. Therefore, the power

required by the load at the motor shaft is equal to

1 (r dC0L'
P =
550 V L dt ,

1 d(coM/N) coM
CO. = - Ji-
L 550 dt N

1 , /iV dco M
550 L(n) dt

1 dcoM
(7-15)
550 JlM M dt

where JLM = JL(1 /N)2 and it is the equivalent inertia load referred to the
motor shaft.
This illustrates that the torque required at the motor to accelerate
the load can be determined by considering the load to be of an equivalent
inertia, JLM, and to be mounted directly on the motor shaft. Thus, total
inertia referred to the motor, JTM, that the motor is required to accelerate
is equal to

Jtm — Jm + JLm — Jm + ^ (7-16)

Conversely, if motor inertia is referred to the load shaft, the total inertia
of the load and motor referred to the load JTL, is equal to

Jtl — Jl + Jml — Jl + Is!

(7-17)
102 MECHANICAL ELECTRICAL CHARACTERISTICS

Thus, load, motor torque, speed, and inertia may be referred from one
shaft to the other of a gear train by some function of the gear ratio as
follows:

Tl = NTm; = ^7 ; Jlm = i ; and W = N!JM (7-18)

N N2

Any other torques that may be proportional to velocity or position of the

motor and load must also have their system constants referred across the
gear ratio by the N2 or 1/N2 factor, thus

FMl = N2Fm and FLM - ^ (7-19)

N2

where FML is the viscous friction of the motor referred to the load and
Flm is the viscous friction of the load referred to the motor. Furthermore,

Ftm = Fm + Flm = Fm + —(7-20)

N2

Gearing inefficiency results in the reduction of the torque transmitted

through the gear train from the drive gear to the driven gear below that
value that would normally appear. Thus, when acceleration requirements
of the load are calculated, the load torque and, therefore, the equivalent
load inertia must be greater by a factor of 1/u, where u is the decimal gear
efficiency.

TABLE 7-1
Symbol Parameter SI English and/or CGS
J moment of inertia kilogram-meter2 oz-in-s2
gm-cm2
dyne-cm2
F viscous friction oz-in/rad/s
dyne-cm/rad/s
T torque Newton-meter lb-ft
oz-in
dyne-cm
P power watt watts
ergs/s
a angular acceleration radian/second2 rad/s2
deg/s2
V angular velocity radian/second rev/min
rad/s
deg/s
e position (displacement) radian radian
degree
7.4. Units and conversion factors 103

7.4. UNITS AND CONVERSION FACTORS

In many small instrument servos, torque is often expressed in units of

ounce-inch, while moment of inertia is given in units of gram-centimeter
squared. Thus, it is necessary to be able to go from one set of dimensions
to another. This is easily done if conversion factors between systems are
understood and used properly.
Table 7-1 shows a number of the most useful parameters in both the
SI and English and/or CGS units, while table 7-2 shows some of the most
useful conversion factors. For a more complete table of metric and other
units, see Appendix A.

TABLE 7-2
Conversion Factors for Servo Calculations
Multiply By To Obtain Multiply By To Obtain

Length Power
meters 39.37 in (oz-in)(rpm) 7.395 x io-4 watts
meters 3.2808 ft (oz-in)(deg/sec) 1.232 x 10“4 watts
inches 2.54 cm (oz-in)(rpm) 9.917 x io-7 hp (English)
feet 30.48 cm (oz-in)(deg/sec) 1.653 x io-7 hp (English)
hp (metric) 1.0138 hp (English)

Mass Angular Velocity

oz 28.35 gm rpm 0.1047 rad/s
lb 453.6 gm rpm 6 deg/s
slug 1.459 x io-4 gm rps 6.2832 rad/s
oz 514.7 slug deg/s 1.745 x io-2 rad/s
lb 3.108 x io-2 slug

Moment of Inertia Torque Gradient

oz-in2 182.9 gm-cm2 oz-in/deg 0.2984 lb-ft/rad
lb-ft2 4.214 x 10s gm-cm2 oz-in/deg 4.046 x 106 dyne-cm/rad
slug-ft2 1.356 x 107 gm-cm2
oz-in2 1.349 x 10~5 slug-ft2
lb-ft2 3.108 x 10“2 slug-ft2 Torque-Inertia Ratio
kg-m2 141.612 oz-in-sec2 oz-in/gm-cm2 7.062 x 104 rad/s2
oz-in/oz-in2 386.1 rad/s2

Force
N (newton) 3.5969 OZ

N(newton) 0.2248 lb
gm 980.7 dyne
00
to
’•o

I04 dyne
X

oz
lb 4.448 x 105 dyne

Torque
Nm 0.7375 lb-ft (Unless otherwise specified, the
Nm 141.612 oz-in units “oz,” “lb,” and “gm” are
Nm 10.197 x 10* gm-cm in units of force.)
gm-cm 980.7 dyne-cm
oz-in 7.062 x 104 dyne-cm
lb-ft 1.356 x 107 dyne-cm
oz-in 72.01 gm-cm
lb-ft 7.233 x 10-5 gm-cm
104 MECHANICAL ELECTRICAL CHARACTERISTICS

7.5. SERVO CHARACTERISTICS

When discussing the characteristics of a servo system, the accuracy

and sensitivity of the elements used to represent the command and the
controlled output must be considered, both for the static condition and
for the dynamic condition. The accuracy with which the elements that
are used as error detectors or error correctors are built determines the
amount of error that will result when an instrument servo is assembled.
Linearity and/or conformity express the accuracy of a precision potenti¬
ometer or an LVDT. Resolver or synchro accuracy is given in terms of
seconds or minutes of angles. A tachometer-generator, either ac or dc,
would have its accuracy given in terms of linearity in respect to volts
per revolution.
Sensitivity refers to how well a servo can satisfy the fundamental
equation that et = ef at null. Regardless of whether the servo is an on-off
or a bang-bang type, a proportional type or a rate type, the difference
between the command, 9„ and the output, 0o, must increase to some value
before any change is initiated. This value gives the sensitivity of the servo
and can be thought of as a dead band, dead space or dead time. Resolu¬
tion of the transducers used as error detectors has to be considered when
determining the dead band. Electrical noise, lost motion in gear trains or
drives, and velocity and acceleration lag, all must be considered. The
minimum voltage, either positive or negative, over which a servo motor
will respond can also be thought of as dead band.
The sensitivity depends, basically, on the torque gradient or torque
constant, Kx, of the servo, and the magnitudes of the torques that the
motor must overcome before it starts to rotate. The torque constant is
the output torque per unit error angle, and it is equal to

T = Kx0 = KXV or Kx = - = — (7-21)

e vc

Let us consider a simple position servo of the proportional type. In

this servo, the driving torque is proportional to the difference between the
command and the controlled output. This output is normally made up
of two torques, one required to overcome the viscous friction of the load
and the other required to overcome the inertia of the load. The friction
load torque is proportional to the control shaft velocity, while the inertia
load torque is proportional to the control shaft acceleration. Equation
7-22 gives the relation between these torques and the error.

Kx@i = J@q + F0O + KX0O

Kx0i KX0O ~ J0O T F0O (7-22)
Kx(0j - d0) = W0 + F60
7.6. Electrical circuit analogy 105

If the load were made up of viscous friction only, then the error would
be converted directly into output velocity or speed. At zero error, all
velocity or motion stops, and the system is at rest. There is no stability
problem, but the response time is slow. If the load were made up of
inertia only, then continuous oscillation would result. This comes about
because, as the error is reduced to zero, acceleration becomes zero but
torque still moves the load at a constant speed. As the load overshoots,
torque is developed in the opposite direction which decelerates the motor
through the zero velocity point and, then reversing the movement, it tries
to find the zero error point. This process continues with sustained
oscillation.

7.6. ELECTRICAL CIRCUIT ANALOGY

A mechanical servo system which is composed of inertia and friction

has an electrical circuit which is analogous to it, i.e., the same mathematical
equations will describe their operation. The electric circuit shown in
figure 7-3 is a simple R-L-C series connection.

R L

T
E
T
Eo
i i
FIG. 7-3. Electrical Analogy of a Mechanical Servo

In this circuit, voltage is the command which determines the charge

of the capacitor, and it is thus analogous to torque, which determines the
movement of a shaft. In the electric circuit, the current flow and the rate
of charge, Q, is determined by the resistance, R. In a mechanical servo,
the speed at which the shaft rotates is opposed by the friction constant, F.
In the electric circuit, the increase or decrease of current is opposed by
inductance, L, while, in the servo, the acceleration of the shaft is opposed
by the moment of inertia, J. The error in the mechanical servo is 6 which
is the difference between the input angle and output angle, while, in the
circuit, the error is the difference in charge between that which the input
voltage and the output voltage will move through the circuit. In the
mechanical servo, the torque constant is KT, while, in the electric circuit,
the corresponding quantity is 1/C.
106 MECHANICAL ELECTRICAL CHARACTERISTICS

Analogous quantities can be tabulated, as in Table 7-3.

TABLE 7-3

Servo Quantity Electric Quantity

e, CE, = Q,
Go CEo = Qo
6 = 6i — 0q Q = Qi — Qo
T = Kt0
*-§
F R
J L

The reason for this determination of analogous quantities is that the

behavior of the electric circuit can now be analyzed through testing, and
the performance of a servo can thus be predicted. For example, damping
of a servo may be achieved by mechanical means such as a flywheel or by
a resistor-capacitor network with the same end results.

7.7. TEST COMMANDS AND SERVO

RESPONSES

General. When testing a servo, both static and dynamic characteristics

must be considered in regard to accuracy, sensitivity, stability, response,
etc. Ideally, perfect accuracy would always be achieved with a resulting
infinite sensitivity, but, in practice, there is always some static error, i.e.,
117V +0.5V or 360° +0.1°. Accuracy would be

— x 100 = 0.428% or — x 100 = 0.028%

117 360

Dynamic characteristics are more important in the study and control

of servo systems than are static characteristics. The term dynamic means
a type of command that is applied to a servo in an everchanging form.
Therefore, in the testing and evaluating of a servo, some basic commands
are utilized. These are the step-position, the step-velocity and the
sinusoidal command.

Step-Position Command. This is a situation where the command changes

very abruptly from one fixed value to another (either higher or lower).
Figure 7-4 shows that the initial value of the command, 9, is zero and,
after the unit step change, it rises to a new fixed value.
7.8. Response of a servo 107

25°(3) t = 0.2 Sec.

T ime

FIG. 7-4. Step Position Command

Step-Velocity Command. This is a situation where the rate of the com¬

mand changes quite rapidly from one value to another. Figure 7-5 shows
that the initial value of the command, 6, is zero, and, after the command,
it has a different value. This is often called a ramp function.

FIG. 7-5. Step Velocity Command

Sinusoidal Command. The magnitude of the command varies as a

sinusoidal function of time. Figure 7-6 shows such a command with the
value of the sinusoid being in the range of 0.1 to 20 hertz.

FIG. 7-6. Sinusoidal Command

7.8. RESPONSE OF A SERVO

The study of the response of a servo to these different commands can

be quite complicated because the form of the command may not be a
separate one but rather a combination. To simplify the evaluation process,
108 MECHANICAL ELECTRICAL CHARACTERISTICS

the response can be divided into two different cases, and then each of these
cases can be considered separately. These cases are:

1. The steady state case where the form of the command is constant so
that any effects from any previous change in form have died out. A
sinusoidal waveform that is fixed in amplitude and frequency would
be considered constant.
2. The transient case where the form of the command is changing so that
the response of the system is still feeling the effects of the change.

Evaluation of a servo is further broken down into the time domain

region or the frequency domain region.

1. The time domain region where the system response is plotted on a

graph as the dependent variable with time as the independent variable
or as a trace on an oscilloscope.
2. The frequency domain region where the system response is plotted on
a graph (semilog paper) with its amplitude (20db log) and phase lag
or lead as the dependent variable with frequency in hertz or radians/
sec as the independent variable.

Figure 7-7 shows how the various commands are used to study the
responses for these cases and regions.

Transient Steady State

Time Domain Frequency Domain

_d oj = Rad/ Sec (f in Hertz)

dt
or
S

Root Locus
1
Bode Plot
or
Nyquist
or
Nichols Chart

FIG. 7-7. Correlation of Test Commands for Transient and Steady State Test Signals
7.8. Response of a servo 109

SOME EXAMPLE PROBLEMS

Example 7-1
A servo motor has the following data: 20; 115V; 20W; Ts = 8 oz-in;
S0 = 6000 rpm; SFL = 4500 rpm; Jm = 0.25 oz-in-s2; Jt = 24 gm-cm2.
Determine the following: (1) angular velocity in rad/s at full load, (2)
torque developed at full load, (3) acceleration of load in rad/s2 and
(4) speed in rpm for part 3.

SFL x 2n 4500 x 6.28

(1) VelAng = = 471 rad/s ans
60 60

P x 1000 20 x 1000
(2) T = 6.0 oz-in ans
0.746 x SFL ~ 0.746 x 4500

1 oz-in-s2
(3) JL in oz-in-s2 = 24 gm-cm2 x
7.06 x 104 gm-cm2

= 3.39 x 10“4 oz-in-s2

T 6.0 oz-in
0 = - 17,700 rad/s2 ans
J 3.39 x 10 4 oz-in-s2

(4) 5 II x 60 =
17,700 rad/s2
x 60 = 3185 rpm ans
2n 6.28

Example 7-2
An inertia load of J = 10.6 gm-cm2 is subjected to a torque of 3.5 oz-in.
What acceleration in rad/s2 is imparted to the load?

3.5 oz-in
1 oz-in-s
10.6 gm-cm"
7.06 x 10 gm-cm

= 23,300 rad/s2 ans

Example 7-3
A servo motor is rated at 4W. If it is delivering 2.4 oz-in of torque to its
load, at what speed is it running?

P x 1352
P = and S
T

13 5 2
S = 2250 rpm ans
2.4
110 MECHANICAL ELECTRICAL CHARACTERISTICS

Example 7-4
A step position command of 12° and a step velocity command of 16 deg/s
are both introduced at time t = 0. At what time will the step velocity
command be twice the step-position command?

12° 16 deg/s S.V. = 2 S.P.

t = — x 2 = 1.5 s
16

S.P. S.V.

PROBLEMS

1. A shaft is rotating at 3600 rpm. What is its angular velocity in rad/s?

.
2 A shaft is rotating at an angular velocity of 293 rad/s. What is its
speed in rpm?
.
3 What is the angular velocity in deg/s for a motor that is rotating at
2400 rpm?
.
4 A motor is rotating at an angular velocity of 12 deg/s. What is its
speed in rpm?
5. A motor is rotating at 32 rad/s, and the power input is 24W. What is
the driving torque at the shaft?
6. A step-position command of 15° and a step-velocity command of
18 deg/s are both introduced at time t = 0. At what time will they
be equal to each other?
7. A step-position command and a step-velocity command are both
equal at time t = 2.4 s. If the step-position command is 9.8°, what
is the step-velocity command in deg/s?
8. A servo motor develops a torque of 0.12 oz-in at 3200 rpm. What is
the power developed?
9. A load has an inertia, J, of 21.2 gm-cm2. The torque delivered to this
load by the motor is 7.2 oz-in. Find the acceleration of the load in
rad/s2.
.
10 For the motor and load of problem 9, find the speed in rpm.
11. A servo system has the ability to position a pointer with an accuracy
of 24 minutes out of 360°. The 360° represent one mile in distance.
What is the accuracy in percent?
7.8. Response of a servo 111

.
12 A proportional zero has a moment of inertia of 2.36 gm-cm2. What
is the moment of inertia in oz-in-s2?
.
13 A 4W servo motor has a load speed of 4800 rpm. What is the torque
in oz-in?
.
14 A potentiometer of 10K ohms is connected across 48V. It can rotate
through an angle of 353°. What is the voltage gradient? What per¬
cent accuracy will allow a deviation of 0.065 volts/degree?
 EIGHT

Analysis of servomotors and

generators

8.1. GENERAL

In early servomechanisms, both ac and dc servomotors had been used;

with the ac predominating, mainly because an ac squirrel-cage induction
motor is rugged, easy to use, and performs well. Even though the dc motor
was more efficient, it was less reliable because of brush problems and radio
frequency interference. DC motors were used mainly for high power
servomechanisms, while the ac motor dominated the instrument servo
field.
However, the introduction of new magnetic material has brought
forth radically new and superior motor configurations which has led to
increased performance advantages of the dc motor. The brush problem
has been resolved so that increased reliability and reduced RF1 have been
achieved. AC amplifiers are inherently more stable than dc amplifiers,
but they are also more expensive. AC line power is more readily available
than dc power, but most amplifiers require a dc supply. DC signals and
dc feedback are more practical because phasing problems are eliminated
and compensation is simplified.

8.2. APPLICATION FACTORS

Servomechanisms which generate less than 100W are considered

instrument servos, and, in this power range, the two-phase ac servomotor
has been the major selection especially when signals, power input and

112
8.3. Fundamental equation 113

excitation levels come from a carrier frequency of either 60Hz or 400Hz.

Low level dc signals, coming from thermocouples, strain gages, etc., are
fed through a modulator, which converts the dc to ac, are amplified, and
are then used to drive an ac motor.
DC servomotors should be used where the required output demands
fast response and high efficiency; also, a dc system is simpler and lower in
cost. The systems are simpler and smaller because present-day, operational
amplifiers practically eliminate the drift that was characteristic of older dc
amplifiers. Heat sinking and cooling method advances as well as pulse
width modulation techniques are also factors. DC motors can deliver peak
power well above the steady power rating; for these reasons, a smaller
motor is used. As an example, chart recorders which require a wide
bandwidth almost invariably use dc motors. Large inertial loads are
another example. In general, where applications are for moderately
accurate, short life (less than 1000 full-speed operating hours), non-
militarized systems, dc systems should be used. If the requirements are
for high accuracy, long life, rugged and compact systems, then ac systems
should be used. An exception to this is the use of dc torque motors which
eliminate the need for gear trains.
Electrically, torque is produced in a motor by interaction between
armature current and field flux. Maximum torque is obtained when the
current and flux are in time phase together. Therefore, in ac motors, space
and time-phase deviations result in a lower torque output. In an ac motor,
transformer action through the stator magnetic field generates the rotor
current and flux. As very small transformers are inefficient, so are small
induction motors. In addition the peak flux density is zr/2 times the
average flux density; therefore, the iron in the induction motor is increased,
by 57% to be exact, which results in inefficiency. DC motors receive their
rotor (armature) energy directly by conduction through the brushes. The
iron losses are considerably less, and, even though armature reaction is
present, it lies along the interpolar axis so that less iron is required in the
dc motor. All of these factors lead to higher dc efficiency.

8.3. FUNDAMENTAL EQUATION

The fundamental equation (applicable to both two-phase induction

motors and dc servomotors) relates motor torque, speed, and power.

Ts = ^ PR (8-1)
S„

where Ts is the stall torque in oz-in, S0 is the no load (free) speed in rpm,
and PR is the stall power to rotor in watts.
114 ANALYSIS OF SERVOMOTORS AND GENERATORS

Speed

FIG. 8-1. A Comparison of Motor Power Output and

Motor Torque

The motor must be capable of producing more power than the load
requires. The speed of the motor when it is driving the load at rated
voltage should be above one-half the free speed. This allows the motor to
operate on the negative slope of the power curve. Thus, as the load
increases, the operating point moves toward the point of maximum power.
As the maximum speed decreases, the theoretical torque per watt increases.
The inherent damping—which equals Ts/S0, slope of the lines—increases
rapidly as the free speed decreases (see figure 8-1). As the motor speed
is reduced in respect to the load, less gearing is required which is why dc
torque motors usually serve as direct drive in closed-loop servos.
AC servomotors have numerous poles with a lower synchronous
speed which increases internal damping. However, the efficiency of power
transfer to the rotor decreases rapidly as the number of poles increases.
Thus, ac motors have a limit to the number of poles; dc motors do not
have this limit because dc free speed decreases as magnetic field strength
increases. The newer magnetic materials allow free speed to be reduced
to a level well below the value for comparable ac motors.
Torque efficiency in dc motors is almost 1007o; this means that nearly
all the input stall power is converted to torque. In an ac motor, it is a
very complex function. Brush friction and voltage drop are the principal
loss factors in small dc motors. The basic torque equation shows that
developed torque per watt in a dc motor depends only on the free speed
which is inversely proportional to magnetic field strength. In ac motors,
the magnetic field strength comes from the transformer action of mag¬
netizing current. For efficient operation, this means very small air gaps
are needed.
8.4. AC servomotors 115

8.4. AC SERVOMOTORS

The conventional ac servo motor is a two-phase induction motor

which usually contains a squirrel-cage rotor and operates on the same
principles as a transformer with the stator acting as the primary winding
and the rotor acting as the secondary winding. Usually, a fixed voltage is
applied to the reference phase, and a variable voltage is applied to the
control phase. These voltages create a rotating magnetic field in the stator
which, by transformer action, induces currents in the rotor; the rotor will
then revolve in the direction of the rotating magnetic field. The torque
required by the load determines the rotor speed.
If a two-phase power supply is not available, then the motor can be
connected across a single-phase supply. However, in order to get the 90°
effect of a two-phase supply, it is necessary to connect a capacitor either in
series or in parallel with the reference winding. The control winding is then
connected across the output from a servo amplifier which means that there
is a variable voltage across it, and thus there is a variable torque output
of the motor.
Figure 8-2 shows the approximate normal characteristics of a two-
phase motor having essentially straight speed torque curves. Most of the
performance characteristics of the motor can be predicted from these
curves. The shape of these curves is governed almost entirely by rotor
resistance. With high resistance, a more linear curve results, but the
maximum torque is less. The more sloping the curves, the more that
damping becomes inherent in the motor. The speed at which the motor
operates with rated voltage on both windings and no external load is
called the free speed. The torque at zero speed is the stall torque.

FIG. 8-2. Speed Torque Characteristics for Different

Values of Control Voltages
116 ANALYSIS OF SERVOMOTORS AND GENERATORS

The available torque is given by equation 8-2:

T = Ts — — x — (8-2)
S0 dt

However, Ts = KTVC where KT is the proportionality constant between

stall torque and control field voltage. The torque, T, that is generated to
give acceleration at any speed or voltage is

Ts dO (8-3)
T - KXVC
S0 dt

and as this torque accelerates the rotor inertia, J:

JsdO
KtVc (8-4)
S0 dt

This torque consists of two components. One is utilized to accelerate the

rotor inertia while the other accelerates the viscous friction constant,
Ts/S0. This constant is the slope of the speed-torque curve, is represented
by F, and represents the viscous damping of the motor and is usually
expressed in dyne-centimeters per rad per second, or in ounce-inch/radian
per second. See figure 8-3.

Torque

FIG. 8-3. Speed Torque Relationships for a Servomotor

8.5. DC servomotors 117

^ = KtVc(1-£ T/Tm) (8-5)

Time

FIG. 8-4. Motor Speed as a Function of Time

As the control field voltage is reduced, the slope of the speed-torque

curves (or the damping) decreases. Damping at low control voltages is
about one-quarter to one-half the damping at rated voltage. This is an
important consideration, especially in a position servomechanism, because
the operating point is usually at or near zero control voltage and zero speed.
It is at this point that stability is the hardest to achieve because of gearing
backlash, transducer resolution and other effects.
If rated voltage is applied to the motor at rest, mosi of the torque that
is developed will be used to accelerate the rotor inertia because the friction
constant at the instant of starting is zero. However, as the speed increases,
the viscous friction or damping will come into effect more and more,
and the acceleration will decrease. The relationship between instantaneous
speed, d0/dt, and time at any instant is an exponential function, as is
shown in figure 8-4. J/F is the time constant of the motor.

- = KtVc(1 - e~T/™) (8-5)

dt

8.5. DC SERVOMOTORS

It was mentioned before that the need for high power output and fast
response has brought the dc motor back into the servo picture. The
different types of dc motors are the wound-field type (shunt, series, and
compound wound), the printed-circuit motor, the moving-coil motor, the
direct-drive dc torquer, the permanent magnet motor, and the stepper
motor. The stepper motor will not be covered in this chapter, but it will
be discussed in Chapter 12 mainly because it is a digital device, rather than
one of continuous rotation.
118 ANALYSIS OF SERVOMOTORS AND GENERATORS

+ +

V,+

COM

Armature Control Field Control

(a) (b)

FIG. 8-5. Basic dc Wound-field Motor Connections Using Transistors

Wound-Field Type

DC motors can be controlled by varying the current to either the armature

or the field. In many motors, the field winding is split into two sections.
If both windings have the same current flow, then the two magnetic fields
oppose, and no motor action takes place. However, if one of the field
voltages is higher than the other, then the motor will run with the direction
of rotation being governed by the winding that has the largest voltage and
current. The motor is called a split-series field motor; its connection is
shown in figure 8-5.
In figure 8-5(a), the motor is connected for armature control with the
fields acting as the load for the NPN transistors. The polarity and value
of the voltage at VI or V2 will determine the direction of rotation because
this governs which field is excited. The counter emf of the motor will
appear between the emitter and ground so that VI or V2 must be in¬
creased in order to increase the speed. When VI is increased, the transistor
is momentarily forward biased to a greater degree which increases the
collector current and field current. The increase in field current increases
the motor current which causes an increase in motor speed. The counter
emf will thus increase; this tends to reduce the emitter and collector
current, and the speed stabilizes. If VI is held constant, then as load is
increased, the speed will decrease, the armature and field current will rise
8.5. DC servomotors 119

a little, and so will the torque. This results in a sloping speed-torque

characteristic which is desired. In either case, VI or V2 being used, a
certain minimum voltage is required to start the motor rotating in either
direction.
In figure 8-5(b), the motor is connected for field control with the
armature acting as the load for the NPN transistors. The transistor current
is determined by the value of the voltage at VI or V2. In this connection,
if the motor is unloaded, the speed will increase to a very high value for a
small change in input voltage. The reason for this is that for this transistor
connection, the motor has the no-load high speed characteristic of the
series motor; thus, as biasing is increased, the field current rises with a
corresponding rise in speed. As load is applied, the motor slows down very
rapidly in order to allow the collector current to increase somewhat so
that the required torque is produced.
Of the two connections, armature control requires a larger driving
signal, but it is much easier to control; field control is quite a bit more
sensitive, but it is harder to control in a system in terms of stability.

Permanent Magnet (PM) Type

A permanent magnet (PM) motor is one in which a permanent magnet

replaces the field windings, thus saving on copper, armature size, and
power to energize the field. The PM motor is usually well compensated
by commutating windings so that the field magnets are not demagnetized
whenever the armature voltage is suddenly reversed or sudden overloads
or stalling result. Operating in high temperatures may also cause a
decrease in PM flux. The main advantages of a PM motor are that they
(1) are smaller and lighter than wound-field motors of the same rating,
(2) have a linear speed-torque relationship, (3) have a high starting torque,
and (4) only two wires are needed, rather than four.
The magnets for almost all PM motors are made from a form of
Alnico VIII alloy cast in a circular ring and completely surrounding the
armature. However, a new magnet material composed of samarium (a
rare earth element), cobalt, and some additives imparts flux properties
that are quite a bit above those of the Alnico series. Unfortunately, this
material is extremely expensive, so motors that are made with the
samarium magnet are restricted to special applications, mostly military,
medical pacemakers, and some computer areas. Their biggest advantages
are their low moment of inertia, fast response and high torque.
Another type of PM motor that may eventually come into use as a
servo motor is the brushless dc motor that uses the samarium magnets and
electronic, rather than mechanical, commutation. The electronic com¬
mutation is achieved by using several different techniques such as optical,
rf, resolver, and Hall-effect semiconductor devices.
120 ANALYSIS OF SERVOMOTORS AND GENERATORS

Direct-Drive Torque Motor Types

The direct-drive torque motor is a permanent magnet type of motor that

can rotate continuously after the introduction of an electric signal so as to
maintain zero error in a position or speed control system. They are
constructed with a frameless type of housing, of large diameter but rather
narrow, and they have fairly large axial holes through the rotor for con¬
nections to various shafts and bosses. They operate well at low speeds
with high torque to inertia ratio and with zero-coupling loss or backlash
because of the direct load connection. Commutation at the higher speeds
causes sparking and also contributes to sensitivity ripple; ripple is a
variation in torque or generated voltage as the brushes switch the con¬
ductors. Cogging, which is the change in air gap reluctance as the rotor
revolves, is a residual effect that sometimes limits the motor torque
performance. Building the rotor with skewed laminations and bridging
the poles may help to eliminate cogging. Another way is to use a slotless
rotor.
The two methods of controlling power to dc torque motors are pulse
width modulation and proportional control. Pulse width modulation uses
either SCR’s or transistors with high efficiency, but it gives sluggish
responses with fairly large ripple torque. Proportional control is usually
accomplished via a bridge circuit of transistors with two operating as
linear amplifiers and two operating to switch the power supply connections
to the motor.
The direct-drive dc torque motor eliminates the gear train used with
most servomotors, and, thus, the system must be analyzed somewhat
differently because of its large bandwidth and high performance.

Moving-Coil DC Motor Types

The moving-coil motor is a permanent magnet type of motor that has

either a flat disc armature or a shell armature that contains no iron. The
copper armature conductors in the shell style are held in place by non¬
magnetic materials such as epoxy, polymer resins and fiber glass. The
armature of the disc style may be made by photoetching a disc or from
preformed segments. The disc type motor is often called the printed-circuit
motor. The active armature conductors of either style move through an
air gap of high flux density but with no reluctance torque effect or cogging.
The moving-coil motor has the highest acceleration capability of any
dc motor, /.<?., several thousand start-stop duty cycles per second, so it is
generally used for incremental motion control rather than for continuous
high speed performance. The torque to inertia ratio of a shell type
armature may reach 1,000,000 rad/s2. Because of these desirable character¬
istics, these motors are used extensively in the data-processing industry,
8.5. DC servomotors 121

numerical control of machine tools, textile drives and any application

requiring high efficiency, high performance, and high acceleration rates.

Printed-Circuit Motor Type

The printed-circuit motor is in reality the disc moving-coil type with a

printed-circuit armature/commutator that gives a very compact, low
inertia unit. The armature is made of copper with a large number of
commutator segments which ensure a smooth, noncogging output mainly
because the inductance of the printed armature is so low that it is con¬
sidered zero for all practical purposes. This is true because the armature
has very few turns with poor inductive coupling to the permanent magnet
field structure.
The principal advantages of the printed-circuit motor are (1) low
time constant, (2) high pulse torque, (3) minimal cogging, (4) wide speed
range, and (5) efficient heat transfer. These advantages have allowed the
motor to be used successfully in both velocity and positioning servo
systems; in incremental motion drives such as punched-card readers/
punches, programmable stepping servos, paper feed mechanisms for
printers, line-by-line reading and punching of paper tape, disc file magnetic
head actuator; and in many other systems. In most of these applications,
a low inertia load must be started and stopped rapidly and repeatedly
either in fixed increments or in a random but programmable manner.
Such a motor is necessary to drive successfully the load at the high
repetition rates involved. In some cases, the stepping motion of the servo
may have to be inhibited in order to slew through large shaft displacements.
This is true because any time given up to motion is really a lost time; thus,
the slew operation must be completed at the highest speed compatible
with the system’s capability.
Figure 8-6 shows a block diagram of an incremental servo of Type 0
(see figure 10-7). The motor has a standard dc tachometer, and it is driven
from a servo amplifier. In this type system, under normal operation as a
regulator, any difference between the input voltage and the tachometer
voltage will just be sufficient to drive the servo motor and to maintain the
desired load speed.
Incremental motion control is obtained by operating the servo in a
nonlinear mode. Explanation for this is given by reference to the wave¬
forms of figure 8-7 and is outlined as follows.

A step command voltage, 1-2, is applied to the input at time t0.

Since the motor cannot respond instantly, a large error signal, 3-4,
which exceeds the saturation level of the amplifier is produced; thus,
the full amplifier output voltage, 5-6, is applied to the motor. The
122 ANALYSIS OF SERVOMOTORS AND GENERATORS

FIG. 8-6. Block Diagram of an Incremental Servo

Step Command
Voltage
Voltage
Error
Voltage
Tach
Motor Voltage
and Current

FIG. 8-7. Waveforms of Incremental Servo Response to a Step Command

8.6. Evaluation factors 123

motor then accelerates, 7-8, and, as this happens, the tachometer

voltage increases and reduces the error until the error signal falls
within the tolerance band, 9, and the system stabilizes.

If the step voltage drops to zero, at 10, the reverse of the above
operation will take place and can be easily traced. The level of
voltage and current required to supply the load torque is shown
at 11.

8.6. EVALUATION FACTORS

Other factors to be considered when analyzing motor operation and

characteristics are (1) dead band, (2) torque constant, (3) velocity constant,
(4) moment of inertia, and (5) the motor transfer function. Let us examine
each of these in order.

Dead Band

In the strict sense, dead band is defined as the specified range of values
over which the incoming signal can be varied without causing the output
to respond. In a servo motor, it is the range of applied voltage just before
the motor starts to rotate in either direction for a given amplifier gain.
For an ac servo motor, this means that there is full voltage on the reference
phase and zero volts on the control phase. As the control-phase voltage
is slowly increased, a point is reached at which the motor just starts to
turn. If the polarity of the voltage is reversed, the same thing will happen,
but the motor will start in reverse. For a dc servo motor, it is the applied
voltage, either to the field or armature depending on how they are con¬
trolled, which will just cause the motor to start. This starting point is
also called the threshold, and it is clearly shown in figure 8-8. The dead

FIG. 8-8. Deadband of a Servomotor

124 ANALYSIS OF SERVOMOTORS AND GENERATORS

band is due to a number of different things including friction of the gear

train and load, inertia of the load and motor rotor, and certain magnetic
effects.
If a potentiometer or synchro transducer with a sensitivity in volts
per degree is supplying the error signal to the amplifier, then the dead band
can be determined by applying equation 8-6 with twice the angle being
the dead band.

2 x Vc = D.B.° x Ks x Ka
?V
D.B.° = (8-6)
kska

with Ks being CT sensitivity in volts/degrees, KA being amplifier gain an

volts/volt, and Vc being control voltage in volts.
It is clear from equation 8-6 that the sensitivity of a system can be
improved by reducing the dead band, and the easiest way to do this is to
increase the gain of the amplifier. However, increasing the gain may lead
to instability; therefore, a compromise must be reached. It can also be
seen that if an application requires a certain dead band and the transducers
and motor have been selected, then the required gain can be calculated so
that an amplifier can be specified.

Torque Constant—Kx

The torque constant is the torque per armature ampere or the torque per
armature volt that the motor can develop with constant field flux. In
servo treatment, it is generally considered to be the ratio between the stall
torque and control voltage. For an ac servo motor, it means rated voltage
on the reference phase with the starting or stall torque being directly
proportional to the voltage on the control phase. For a dc servo motor,
it is the stall torque developed for the impressed voltage on either the field
or armature. For either motor, it is given by

with Ts being stall torque and Vc being control voltage.

Velocity Constant—Kv

The velocity constant is a proportionality constant between motor speed

and control voltage and represents the speed in radians per second that
the motor can develop for every one volt that is impressed on it. It is
expressed as

(8-8)
8.6. Evaluation factors 125

with S being speed or velocity and Vc being control voltage. Kv may also
be found from the torque constant, KT, and the viscous damping factor,
F, of the motor. The damping factor is equal to Ts/S0, so:

(8-9)

but Kt = Ts/Vc and F = Ts/S0; therefore:

Ts/Vc _ S0
Ts/S0 Vc

with S0 being in rad/s and Vc being in volts. This constant is important

when the steady-state response of a servo is being considered after it is
subjected to a step-velocity command because it represents how fast a
motor will respond to a one volt error change.

Moment of Inertia

In many instrument type servos, the load consists only of a dial pointer or
data device such as a recorder pen; therefore, the moment of inertia of
the motor rotor represents the main retarding force to angular motion.
This moment of inertia, as is outlined in section 7-2, is usually expressed
in gram-centimeter squared or ounce-inch-second squared.

Motor Transfer Function

This is the ratio between the motor output, in terms of angular movement,
and the input control voltage; it combines most of the motor terms
discussed in sections 7-2 and 8-4. It may be inserted into a system block
diagram in the following way:

Kt0 = J9 + F0O but Kt0 £ KtVc (8-10)

Now convert equation 8-10 to the LaPlace operator

KtVc = Js20o + Fs0o

Factor 0o out so

KXVC = 0o(Js2 + Fs)

Rearrange

90 _ Kx
Vc Js2 + Fs
126 ANALYSIS OF SERVOMOTORS AND GENERATORS

Divide by F:

0o kt/f
Vc (J/F)s2 + S

Kv and J/F = t (time constant)

. o0 Kv Kv
(8-11)
" Vc s + TS2 S(1 + TS)

Kv
Vc - d0
S(1 + TS)

8.7. SERVO GENERATORS

Many applications in servo systems, especially in the instrument

servo range, require a voltage which is proportional to the rate of change
of some function such as position or speed. This voltage may be used as
a shaft speed readout or as a stabilization voltage in a closed-loop system.
Such a device is called a rate generator or a tachometer-generator.
The ac rate generator usually consists of a two-phase stator, a drag
cup, and a rotor of nonmagnetic material. An ac reference voltage is
applied to one phase, thus setting up a magnetic field. The second phase
is located 90° mechanically from the first. When the drag cup is rotated
between the two windings, a second field is produced at approximately
right angles to the first. This field produces an output voltage which is a
direct function of the rotor shaft velocity. A change in direction of rotation
causes a 180° phase shift in the polarity of the output voltage. The positive
direction of rotation gives an output voltage that is in-phase with the
reference input. Conversely, an out-of-phase voltage means reverse
rotation. Figure 8-9 shows the schematic diagram.
The rate generator is used for three basic applications, i.e., (1) as
speed or rate control for a servo, (2) as integrating or angular position
servos, and (3) as a stabilizing device for a position servo. When it is used
in a speed control system, the output of the rate generator is compared to
the command or reference input signal. If the desired output of a speed
control system is angular position rather than angular speed, then
integration is performed by the rate generator. The third application,
which is velocity damping, uses the rate generator to simulate a higher
viscous friction. Increasing the viscous friction in respect to the moment
of inertia decreases the system time constant, and thus permits higher
8.7. Servo generators 127

FIG. 8-9. A Two-phase Rate Generator

amplifier gain. Higher gain with the same stability usually means better
system operation.
The dc tachometer-generator usually consists of a permanent magnet
field with a wound armature. The output voltage of the armature is pro¬
portional to the shaft velocity and is quite linear; therefore, over its range
of operation, it is free from random input voltage changes in the operating
frequency ranges. Also the voltage gradient, i.e., volts per rpm, should
be stable with any changes in ambient or device temperature. The polarity
of the voltage is determined by the direction of rotation.
The rate generator or tachometer-generator is quite often constructed
integrally with the servo motor on the same shaft. This makes mounting
easier and, at the same time, still gives high performance characteristics.
Some example problems will now illustrate the theory presented.

Example 8-1

Given the following servo motor data: 20 115V 20W Ts — 8 oz-in;

S0 = 6000 rpm; SFL = 4500 rpm; and JM = 6.5 x 10“4 oz-in-s2.
Determine the following: (1) slope of characteristic curve, (2) torque
constant, KT, in oz-in/V, (3) velocity constant, Kv, in rad/s/V, (4) time
constant, T, and (5) transfer function in frequency response form.

Solution

(1) Slope F = _ Zs =-8 oz-in_ x _ _o.Ol3 oz-in/rad/s ans

S0 6000 rpm 2n

17 _ S
T _ 8 oz in _ q Q7 oz_jn/y ANS
(2) JVt — -
vc 115V

(3) = 0 07 oz-in^V = 5.36 rad/s/V ans

I
<

0.013 oz-in/rad/s

6.5 oz-in-s2 x 10- = 0 05 s ANS

(4)
I
Til
I

0.013 oz-in/rad/s
128 ANALYSIS OF SERVOMOTORS AND GENERATORS

(5) k = Kv = 5-36 = 5.36 x 20 = 107 an$

' ' Vc s(l + is) s(l + 0.05s) s(20 + s) s2 + 20s

Example 8-2

A two-phase servo motor will start when the applied voltage to its control
phase equals 6% of rated voltage. The CT of the synchro pair used as
transducers has a sensitivity of 0.64 volts/degree. The rated control
voltage is 30V. The amplifier gain equals 36dB. Determine the following:
(1) the dead band in degrees, and (2) the gain if the dead band is tripled.

Solution

(1) 2VS = D.B.° KsKa

Ka = 36dB or 63

j)g°_2 x 30 x 0.06V 3.6°

= 0.089° ans
~~ 63 x 0.64V/deg 40.32

2 x 30
(2) Ka
_ 3 x 0.089 x 0.64

= 351 Ka in dB (m = 20 Log M) M = 351 m = 20 x 2.55

Ka = 50.9dB ans

Example 8-3

A dc permanent magnet servo motor has the following data: 50V

1800 rpm Kx = 5.25 oz-in/V; and Ria = 8 ohm. Determine the free
speed and stalled torque for applied voltages of 50V, 40V, 30V and 20V.
Also sketch the curves for the different voltages.

Solution

50 . V So T
kg - v - 0.928V/rpm
S0 1800 50 1800 262.5
Ts = KXVC = 5.25 oz-in/V x 50V 40 1430 210
= 262.5 oz-in 30 1070 158

c _ 40 20 715 105
4- O x 1800 = 1430 rpm
0.028

T 40
TS40 - - X 262.5 = 210 oz-in
8.7. Servo generators 129

PROBLEMS

1. A dc servo motor has the following data: Permanent magnet 120V

1500 rpm KT = 105.6 oz-in/armature ampere; and Ria = 6 ohms.
Determine the free speed and stalled torque for different applied
voltages of 120V, 90V, 70V, 50V and 30V. Also plot the family of
curves.
2. A two-phase servo motor will start when the voltage applied to its
* control phase equals 5% of rated. The sensitivity of the synchro CT
is 0.393 volts/degree. The rated control phase voltage is 35V. Cal¬
culate the amplifier gain if the desired dead band equals 0.1°.
3. A two-phase servo motor will start when the applied voltage to its
* control phase equals 5% of rated. The sensitivity of the synchro CT
is 1.0 volts/degree. The rated control phase voltage is 26V. The
amplifier gain equals 100 or 40dB. Calculate the dead band in
degrees.
4. A permanent magnet dc motor has the following parameters:
armature inertia 2.0 oz-in-s
Ts at 30V 5 oz-in
S0 at 30V 4800 rpm
Determine (a) friction constant, F, in oz-in/rad/s, and (b) the time
in seconds it takes the motor to reach a speed of 3000 rpm.
5. A servo motor, Kearfott R110-2, has a mechanical time constant of
0 038 seconds and an electrical time constant of 0.0007 seconds. It
has a velocity constant of 12 rad/s/volt. Determine the motor transfer
function.
130 ANALYSIS OF SERVOMOTORS AND GENERATORS

6. A two-phase servo motor has the following data:

Phase voltages = 115V; Ts = 5 oz-in; S0 = 4500 rpm;

and

J = 0.2 oz-in-s2

Determine the motor transfer function.

7. A constant speed application takes a regulating signal from a
tachometer-generator sensing device. The regulator requires a signal
change of 0.1 V to produce the necessary corrections.
(a) If the allowable error at maximum speed is 1%, what voltage
must the tachometer-generator produce at maximum speed?
(b) What tachometer-generator voltage is required at the maximum
speed to reduce the allowable error to 0.5%? 0.25%?
(c) What conclusion can be made relative to the magnitude of the
signal level where a choice exists?
8. A speed-indicating device consists of a two-pole permanent magnet
dc tachometer-generator with a dc voltmeter connected to its ter¬
minals. The generator has an armature with 1,000 conductors having
two parallel paths and a resistance of 1,000 ohms between brushes.
The air gap flux is 36,000 lines per pole. The dc voltmeter which is
connected directly across the brushes has a resistance of 2,000 ohms.
What speed in RPM will be indicated by a reading of 5V on the
voltmeter?
9. What is the transfer function of the tachometer-generator of problem
8?
10. A servo motor has a time constant of 0.026 seconds, a free speed of
6500 rpm and a rated control voltage of 26V. Determine the motor
transfer function.
 NINE

Stability in linear servo systems

9.1. GENERAL

A system is defined as being stable if it will return to a state of

equilibrium in some definite amount of time after it has been subjected to
a disturbance. This time length usually has certain limits which are deter¬
mined by the system’s application.

9.2. SYSTEM RESPONSES

A simple closed-loop servo system using a synchro transmitter and

control transformer as transducers is shown in figure 9-1. The error
voltage from the CT is fed into an amplifier which controls the servo motor
which is driving a load and the rotor of the CT. The system transfer
function is also given.
Figure 9-2 shows the response curves for such a servo that has been
subjected to a step-position command. The original controlled output had
been on the x-line, and the step-position command now drives it to the
y-line; the system responds along one of the three curves, depending on the
damping.
Curve 1 shows the response of an underdamped system with its peak
overshoot, oscillation period and settling time. The settling time is the
time it takes for the system to come within the tolerance band by the fourth
time constant, or after 98 % change has resulted. In addition, the system
could hunt continuously. For proper operation of this servo, the feedback

131
132 STABILITY IN LINEAR SERVO SYSTEMS

0o

8 , S(1 4- STm)

K = KsKaKvN

FIG. 9-1. A Simple Closed-loop System

A - Peok Overshoot
B - Oscillation Period
C - Tolerance Band
D - Response Time
E - Settling Time
F — Rise Time

FIG. 9-2. System Response Curves for Three Conditions of Damping

voltage from the output opposes or subtracts from that of the reference
input, i.e., negative feedback; therefore, as the error increases, the restoring
torque, which tends to reduce the error, also rises.
As the time lag between the system input and output becomes greater
9.2. System responses 133

Steady State Error

A — Undamped
B — Underdamped
C - Critically Damped
D - Overdamped

FIG. 9-3. Comparisons of Steady State Errors and Restoring Torques for Different
Conditions of Damping

and greater—over one-half cycle—the feedback voltage from the output

transducer reverses polarity, and, instead of subtracting from the input, it
now adds to it, i.e., positive feedback. Thus, the input to the amplifier and
thence to the motor drives the motor in the same direction as the input is
displaced. This means that the error is increased rather than decreased,
and, if it is allowed to build up, it will eventually saturate the system so
that it will oscillate continuously. Figures 9-3 and 9-4 show the steady-
state error and restoring torques that can exist for the different damping
ratios.
134 STABILITY IN LINEAR SERVO SYSTEMS

Torque Proportional to
Angular Displacement

Torque in Wrong Direction to Cancel Error,

Actually Proportional to Past Error. Damping
Torque 180° Out of Phase, Maximum Oscillation
Occurs
Angular Displacement

FIG. 9-4. Analysis of Damping Torques and an Error Signal

In analyzing system performance, two questions can be asked:

1. How close can the output shaft be positioned in respect to the input
shaft? The answer here depends on static and running friction and its
damping effect as discussed in Chapter 8, p. 117. When the torque at
9.2. System responses 135

FIG. 9-5. Position Error for a Step Position Command

the output drops below that required to overcome the running-friction

torque, a minimum positional error will exist. However, static
friction is usually much larger than running friction so that when the
input shaft is displaced again, a large difference in shaft position with
a resultant large error voltage will be required to overcome the
breakaway or stiction torque. Since this is the maximum positional
error, this is the one usually considered in design calculations as the
minimum static error of the system. Figure 9-5 shows this positional
error. The term position error constant, KPE, reflects this misalignment
between input and output and can be found from

where Ee is the error voltage and d is the shaft displacement. Figure

9-6 shows the transient error curves for different vicsous-damped
servos with different friction constants and damping ratios. It can be
seen that the relative position error is about the same, but that the
oscillations die out faster depending on the damping.
2. How close can the output shaft follow the input shaft if the input
shaft rotates at a constant velocity? When the input shaft is started
running at a constant speed, the output shaft will try to follow it.
Again, because of system friction, there will be an opposing force
which'will cause the output to oscillate about the input shaft speed
until it settles down. However, a difference in shaft positions will now
exist, even though both are running at the same speed. This is called
steady-state velocity error or velocity lag error, the value of this error
is determined by the amplifier gain and the magnitude of the viscous
friction (see equation 8-8). Figure 9-7 shows different viscous-damped
servos with different damping ratios and friction constants for the
136 STABILITY IN LINEAR SERVO SYSTEMS

Cl>£2>£3

FIG. 9-6. Transient Error Responses for Different Damping Ratios—Step Position
Command

FIG. 9-7. Transient Error Responses for Different Damping Ratios—Step Velocity
Command

same step-velocity command. A comparison of these curves shows

that the servo with the largest steady-state error will have fewer
oscillations.

The actual damping of this system is related to the critical damping

in the following manner:

Damping ratio (zeta) = -aCtUal dampin8 (9-2)

critical damping
9.3. Servo damping 137

The natural frequency at which the system will break into oscillations
is

Natural Frequency = coN = — rad/s (9-3)

These equations show that the system gain, K, must be adjusted quite
carefully because, when it is changed, the damping ratio varies inversely
while the natural frequency varies directly. They also show that as the
inertia, J, is decreased, it has the desirable effect of increasing ( and toN.
To improve damping, various methods of decreasing J and/or increasing F
are used.

9.3. SERVO DAMPING

Most textbooks and manuals on servomechanism design derive their

analysis on the assumption that the systems are linear and that the servo
motor is linear. Idealized motor characteristics are shown in figure 9-8,
and, from these, the designer makes his choice of components, defines any
significant parameters, and estimates system performance with whatever
experimental adjustment is required.
In most single-loop, second-order systems, the servo motor presents
the most nonlinearity and the slope of the motor torque vs. the speed
characteristic represents the motor damping as given by:

(9-4)

XJ
<D
<V
Cl
CO

Torque

FIG. 9-8. Idealized Motor

Characteristics
138 STABILITY IN LINEAR SERVO SYSTEMS

~D
CD
CD
CL
CO

Torque

FIG. 9-9. Actual Motor Output with

Linear Characteristics

The damping is directly related to the velocity constant, K, and the

motor time constant, tm, with both of these factors decreasing as the
damping decreases. Damping also varies directly with stall torque and
inversely with free speed.
Figure 9-9 shows the actual speed torque characteristics of a motor
with a linear relationship. At low control voltage and low speeds, which
represent static operation of a positioning servomechanism, the damping
is about one-half the damping derived from the straight line, idealized
curve. The actual slope is almost double what is expected when linearity
is assumed. If the speed torque characteristic has a bulge, then damping
varies with speed in a more complex manner. The larger the servo motor,
the greater the curvature (see figure 9-10).

X>
a>
a)
cl
CO

Torque

FIG. 9-10. Actual Motor Output with

Non-linear Characteristics
9.4. Damping methods 139

Using the actual curves, figure 9-10 leads to much better correlation
between theory and experiment. The most important special condition
occurs at low speed and low control voltage where damping is at a minimum
and the tendency to oscillate is greatest. For slewing or steady operation
at a specific speed, the damping associated with stability is taken as the
slope at the actual operating point.
To obtain the parameters for equations 9-2 and 9-3 which are most
suitable for a particular application, the servo designer has a number of
damping techniques available. Some of these techniques are described
below.

9.4. DAMPING METHODS

Sections 9.4(A) through 9.4(C) illustrate and describe different types

of mechanical devices or networks that are used for stabilization purposes,
while sections 9.5(A) through 9.5(E) show the changes in the block
diagram and transfer function that occur when these methods are used.

Inherent Damping. The following methods increase the natural damp¬

ing of either ac or dc servo motors without adding mechanical com¬
plications:

1. reducing the skew of the rotor

2. reducing the inertia of the rotor
3. increasing the resistance of the rotor
4. using dc on the motor windings

Use of these techniques will allow the free speed of a certain six-pole,
400 Hz motor to be reduced from 6000 rpm to as low as 4500 rpm.
However, reducing the rotor skew or using dc has disadvantages which are:

1. an increase in the starting voltage which decreases system sensitivity

2. a reduction in the slewing speed
3. an increase in power losses
4. a tendency to cause motor dither at system null

Reducing the inertia of the rotor increases the damping ratio and also
increases the speed of response of the system.

Viscous or Velocity Damping. Viscous and velocity or rate damping

techniques are those in which the amount of damping is directly propor¬
tional to the speed of rotation of the motor. The damping signal is pro¬
duced electrically from a transducer at the output. The signal is fed back
140 STABILITY IN LINEAR SERVO SYSTEMS

into the closed-loop with a polarity that (1) makes an increasing error
appear larger, and (2) makes a decreasing error appear smaller. The most
commonly used methods are listed below with all the methods introducing
velocity lag into the system, i.e., a steady-state error exists between input
and output.

1. Viscous damping is achieved by different devices.

(a) Air or oil dashpots, friction discs, brakes, gear type devices,
magnetic brakes, eddy current brakes, fluids containing magnetic
particles, etc.

FIG. 9-11. Magnetic Viscous Damping Added to a Servo Motor (Cedar Engineering
Div., Control Data Corp.)
9.4. Damping methods 141

(b) An aluminum drag cup rigidly connected to the motor shaft

rotating in a magnetic field produced by two permanent magnets
(see figure 9-11). One of the magnets is usually adjustable so the
damping is adjustable.
These methods have little effect on starting voltage or torque, but they
do increase the motor inertia and reduce the system slewing speed.
2. Velocity damping is obtained by means of feedback which comes from
a tachometer-generator mounted integrally with the motor. If it is
an ac tachometer, then it consists of a two-phase stator, a drag cup,
and an adjustable core. One winding is excited from a reference
source. The second winding has voltage induced in it because of the
interaction of the drag cup and the reference winding. This voltage
signal, which is proportional to the rate of change of the output, is fed
to the amplifier in opposition to the error signal between the output
and the input. The amplifier thus responds to the two signals, (1) due
to the position error, and (2) due to a change in velocity. If it is a dc
tachometer, then it may have a permanent magnet field with a wound
rotor with the signal being dc, instead of ac (see figure 9-12). This
method is very flexible because it allows a wide range of damping
ratios over a wide range of speeds. The major disadvantages are:
(a) high cost
(b) introduction of noise
(c) an increase in motor inertia

Adjustable Core

Drag Cup Tachometer Stator

FIG. 9-12. A Motor-tachometer

142 STABILITY IN LINEAR SERVO SYSTEMS

Flywheel Bearings Drag Cup

(d) velocity lag

(e) introduction of temperature and frequency sensitive variable
3. Counter emf damping is produced when the counter emf of a dc
servometer is used as the measure of speed. This generated voltage
is fed directly into the amplifier but in opposition to the position
error signal.

Inertial Damping. Inertial damping is a form of viscous damping which

does not introduce velocity lag. It consists of an aluminum drag cup
rigidly connected to the motor shaft with a permanent magnet flywheel
which is also on the shaft on ball bearings (see figure 9-13). The only time
this drag cup has any effect is during acceleration. When the motor is
9.5. Mathematical analysis 143

running at a constant speed, the viscous coupling between the drag cup
and the flywheel drives the flywheel at an essentially constant speed. This
method does not affect motor performance very much, but it does intro¬
duce some errors during acceleration.

9.5. MATHEMATICAL ANALYSIS

Inherent and Viscous Damping. The transfer function of a linear ac

servo motor can be written as in equation 9-5; J includes the inertia of the

_ Kt/F (9-5)
Vc s[(J/F)s + 1]

drag cup and the rotor, and F includes the viscous damping as well as the
inherent damping. F is determined from the slope of the characteristic
curve and is given by equation 9-4.

Tachometer Feedback. When a tachometer is used for damping pur¬

poses, the block diagram in figure 9-1 is changed to that in figure 9-14.
The transfer function for the inner loop is given in equation 9-6, with F in
equation 9-5 being replaced by F + KAKTKG. J includes the inertia of
the rotor and the drag cup.

kakt
F + Kj Kq (9-6)
B*
VA
S( F + KAKTK6S+1)

FIG. 9-14. Addition of a Tach Generator for Damping

144 STABILITY IN LINEAR SERVO SYSTEMS

FIG. 9-15. Addition of a Corrective Network for Damping Purposes

Inertial Damping. The transfer function of an inertial-damped motor

can be written in the form given in equation 9-7; the values of J, F, Jd
and Fd are determined by using the desired corner frequencies from a
Bode plot. This type of motor responds to a step input with a large initial
overshoot which is quickly attenuated. Critical damping is possible. Its
most common application is in systems which require high velocity
constants and low accelerations, but which cannot tolerate a velocity lag.

__ kt/f
Vc " (9-7)
+ s + 1
F F[(Jd/Fd)s + 1],

Error Rate Damping. Error rate damping is also a form of damping

which does not introduce velocity lag. It involves the use of components
whose voltage output varies as the rate of change of error, that is during
acceleration or deceleration periods. These are three or four terminal
networks composed of different configurations of resistance and capac¬
itance; they are called (1) phase lag, (2) limited-phase lag, (3) limited-phase
lead, and (4) lag-lead networks. For a more complete description of each
and its transfer function, see section 10.2. Figure 9-15 shows the addition
of a limited-phase lag network in a closed-loop system.

The Characteristic Equation. Let us look again at the equation 7-22

on page 104.

Kjd, = J 90 + F 60 + K T0O or K Tdl = J 6>0 + F -0O + Kx6>0

dt dt
(9-8)

which we now recognize as being in the quadratic form. Now let us equate
the transient part to zero.

+ F— + K = 0 (9-9)
dt2 dt
9.5. Mathematical analysis 145

The roots of this equation, and r2, will now be as follows with three
possible solutions:

-F ± VF2 - 4KJ /F2 K

ri, r? — F + (9-10)
1 2 2J 2J Asi 4J2 J

1. Negative, real and unequal roots. This solution leads to an over¬

damped system.

F2 K
or ?J->i (9-11)
4J2 > 7 2 VKJ

2. Negative, real and equal roots. This solution leads to a critically

damped system.

_ K or F 1_
(9-12)
4J2 J 2 7kj

3. Conjugate and complex with negative real parts. This solution leads
to an underdamped system.

F2 K (9-13)
or c = l < 1
4J2 J 2 VKJ

Thus, for critical damping, the friction constant, Fc, should equal
2\/KJ- The ratio of actual damping to critical damping is denoted by
zeta (C) and £ = F/Fc = F/2\/KJ. The frequency at which the system
will break into oscillation is called the natural frequency and is given by
WN = Vk/J in radians/second or fN = (1 /(27r))(VKL/J) in hertz. The
bandwidth of the system will now be the range of frequencies the system
can follow over which the amplitude of the output stays above the half
power point or above — 3dB.
Some example problems will now illustrate the theory presented.

Example 9-1
A viscous damped servo has a natural angular frequency of 15 rad/s. A
step velocity command of 50 deg/s is introduced to the servo. Under the
assumption that the servo is critically damped determine: (1) the steady-
state error, (2) the damping time constant, and (3) the maximum step-
velocity command that will limit the steady state error to 0.5 .
146 STABILITY IN LINEAR SERVO SYSTEMS

_ 2£ x SVC = 2 x 10 x 50 deg/s 6.66° ans

coN 15 rad/sec

2. t = —= --- = 0.066 s ans

£cuN 1.0 x 15 rad/s

3. S.V.C. = 5^ = ft5° x 15 rad/s - 3.75°/s ans

2C 2 x 1.0

Example 9-2

A mechanical position servo is shown in the following diagram. The load

is an instrument pen recorder with the following specifications:

CT Amplifier Motor Gear Train

fli 90

Specs: (1) Rate of pen movement—50 deg/s.

(2) Threshold—<0.12°
(3) Scale factor—415 mV/deg approximately
(4) Damping ratio—>0.4
(5) Input transducer signal—11.8V line-to-line
(6) Power supply—260V 400 Hz.

The following synchro and motor will be used: Size 8 CT 41 lmV/deg

sensitivity 11.8V and a size 8 motor with the following specs: Ts = 0.15
oz-in; S0 = 6000 rpm; Vc = 26V; J = 0.48 gm-cm2; and starting
voltage 10% of Vc.

We will now determine the following things:

(a) Gear Ratio PenRate x 60 s

rpm
360° H

50 deg/s x 60 s
rpm = = 8.3 rpm
360°

S 6000 „„„ I
Gear Ratio = - = 723 or N = —
rpmPEN 8.3 725
9.5. Mathematical analysis 147

(b) Vband = 0.12° x 411 mV/deg = 0.049V

(c) Amplifier Gain _ Vc x %Vs.t. _ 26 x 0,1 _

Ka a " VD B " 0.049

(d) CT Gain Ks = 57.3°VSEN. = 57.3° x 0.411 = 23.6 V/rad

Ks

(e) Torque Constant K j[s _ °-15 oz~in = 0.0058 oz-in/V

Kx T Vs 26V

(f) Overall Gain K = KSKAKTN

K 1
= 23.6V/rad x 53 x 0.0058 oz-in/V x —

= 0.01 oz-in/V

(g) Damping Characteristic p _ Tg _ 0,15 oz-in ^ 60

F S0 6000 rpm 2n

= 2.39 x 10-4 oz-in/rad/s

(h) Velocity Constant = Kj = 5.8 x 10~3 oz-in/V

Kv v F 2.39 x 10“4 oz-in/rad/s

= 24.3 rad/s/V

2.39 x 10“4 oz-in/rad/s

(i) ^ = F_ =
2\/KJ 2yJ 1.0 x 10"2 oz-in/V x 6.8 x 10 6 oz-in-s2

= 0.48 gm-cm2

= Q'48- = 0.068 x 10"4 oz-in-s2

7.062

2.39 x 10"4 oz-in/rad/s

2 x 2.61 x 10"4 oz-in-s

= 0.46

(j) Transfer Function _ _J _ 6.8 x 10 6 oz-in-s2 _ q q285 s

T " F 2.39 x 10"4 oz-in/rad/s

T.F. =
K, 24.3
S(1 + ts) S(1 + 0.9285s)

24.3 x 35.1 _ 852_

s(35.1 + s) _ s(s + 35.1)
148 STABILITY IN LINEAR SERVO SYSTEMS

If this damping ratio, ((), is not large enough, then tachometer feed¬
back can be adjusted to bring it above some minimum value. Reduc¬
ing the inertia of the motor will also do this, i.e., using a motor with
a higher free speed and a lower moment of inertia.

QUESTIONS AND PROBLEMS

1. A feedback potentiometer has a total travel of 350°. Determine the

transfer function if the supply voltage is 45V.
.
2 Give a clear and concise definition of the following terms:
(a) natural frequency (f) phase lag
(b) deadband (g) velocity-lag error
(c) zeta (h) position error
(d) settling time (i) error rate damping
(e) velocity damping
3. A servo has a natural frequency of 7.5 Hz and a moment of inertia
of 0.308 oz-in-s2. What is the torque gradient in oz-in/deg? If the
damping ratio is equal to 0.65, what is Fc in oz-in-s/rad?
4. A servo has a torque gradient of Kx = 10 oz-in/deg. The moment of
inertia is 0.624 oz-in-s2. What is the natural frequency?
5. The friction constant, F, of the servo problem 4 is 18 oz-in/rad/s.
What is the damping ratio, zeta?
6. A viscous-damped servo driving a position follow up system has a
natural angular frequency of 12 rad/s with an input velocity of 40
deg/s. Determine:
(a) the steady-state error for critical damping when zeta equals 1.0
(b) the steady-state error when zeta equals 0.8
(c) the damping time constant for problem 6a
(d) the damping time constant for problem 6b
7. A viscous-damped servo has a natural frequency of 10 rad/s and a
zeta of 0.8. A step-velocity command of 80 deg/s is introduced.
Determine:
(a) the steady-state error
(b) the maximum displacement of the output during the transient
period
8. Repeat problem 7 for a zeta of 0.9.
9. For the servo in problem 6, calculate the proportional change in zeta
that will reduce the steady-state error for critical damping by 20%,
i.e., to 0.8 of the value found in problem 6a.
9.5. Mathematical analysis 149

10. The torque gradient of a servo motor is 12 oz-in/deg. It has a natural

frequency of 6 Hz. What is the moment of inertia in oz-in-sec2?
11. A positioning servo system uses synchros as error detectors. The
system has the following load requirements: (1) rate of moment =
50 deg/s, and (2) a deadband of 1.0°. Using the motor of Example 8-1,
determine the following, assuming that the motor will start the load
with 10% voltage. Also assume that the CT has a sensitivity of
415mV/deg.
(a) the gear ratio
(b) the amplifier gain
(c) the overall gain, K
(d) the damping factor
 TEN

Frequency response in servo

systems

10.1. ANALYSIS USING NYQUIST PLOTS

To test a servo such as that mentioned in Chapter 7, many engineers

apply a step-position command to the input and then measure or calculate
the ensuing overshoot, errors, rise time, settling time, etc. However, many
engineers prefer to apply a sinusoidal command and then to measure the
loop gain and phase lag for a range of input frequencies.
To do this, the loop is opened at the error detector, R is reduced to
zero, and a sinusoidal command of low frequency and low constant
voltage is applied in place of the error singal (see figure 10-1).
The measurements now consist of the voltage at F and the phase
angle by which F lags E. These readings are taken for a range of fre¬
quencies so that a polar plot of the feedback voltage can be plotted vs.
phase angle. To find the gain, the ratio of F to R must be used under
closed-loop conditions. This is done by reconnecting F to the error
detector so that E = R — ForR = F + E. Since E is a constant value,
the length of vector R on the plot is the vector sum of a — E and a varying

FIG. 10-1. Opening of a Feedback loop

150
10.1. Analysis using Nyquist plots 151

Example 10-1
Nyquist Plot

C _ 10
R (1 + jcurOll + j<jL>r2)

_10_
R (1+ j0.05ud(1 +jOlud

F = 10 A, = (12 + (0.05oi)2)2
A,A2
A2 = (12 + (0 leu)2)?

8j = + 02 9j = -Tarf1 0.05cu - Tan-1 Olcu

Volts

UJ f F 9, e2
Rad/Sec Hz Volts Deg Deg Deg
2 0 32 9.7 -6 -II -17
6 096 8.2 -17 -31 -48
8 1.28 7.2 -22 -39 -61
10 1.6 6.3 -27 -45 -72
20 3.2 3.2 -45 -63 -108
40 6.4 l.l -63 -76 -139
80 12 8 0.3 -76 -83 -159
200 32 0 1 -84 -87 -171

F. Vector E is now drawn so that its tail is at a — 1. Now any line drawn
from the -1 point to any point on the polar plot equals R. Example 10-1
clearly illustrates how this is done.
The ratio of F/R is now considered for three cases:

1. When F is smaller than R, the feedback lags the input by an angle

up to 90°.
152 FREQUENCY RESPONSE IN SERVO SYSTEMS

01 02 05 10 2 5 10

<u -Rad/sec.

FIG. 10-2. Phase Lag Network

2. When F is 1.414 times larger than R, the angle of lag is 135°.

3. When F is about four times larger than R, the angle of lag is about
180°.

The criterion that determines whether a system will be stable or

unstable is called the Nyquist criterion. It is as follows:

1. If the curve passes to the right of the — 1 point, then the system is
absolutely stable.

e0 w = ct ej e0 w = 0

w - Rad/sec.
At A,^ = 4^ =1 Rad/sec

At B,w2= ^ = 5 Rad/sec

FIG. -10-3. Limited Phase Lag Network

10.3. Analysis using Bode plots 153

At D> w2= ok = 20 Rad/sec

FIG. 10-4. Limited Phase Lead Network

2. If the curve passes through the — 1 point, then continuous oscillation

will result, and the system is absolutely unstable.
3. If the curve passes to the left of the — 1 point, then the system will be
unstable with the gain determining the instability.

10.2. NYQU1ST PLOT CORRECTION

The upper limits for useful stability is in the range of 1.5 to 2.0 for the
F/R ratio. Corrective networks such as those shown in figures 10-2
through 10-5 change the gain of the system, and thus the shape of the
polar plot so that the -1 point is no longer encircled. Figure 10-6 shows
a system that has been stabilized in a conditional sense by a Notch network,
figure 10-7 shows the different types of servos with their transfer functions
and Nyquist plot corrections.

10.3. ANALYSIS USING BODE PLOTS

A control system consists of devices connected by signal paths. Each

device may change the magnitude of its input signal and the time of this
signal depending on the frequency of the input signal. A dc signal can e
attenuated in size but not in time, while an ac signal can be attenuated in
both size and time (phase).
 e£ _ (1 4- jtur2) (1 + jgjr3)
(10-4)
ei (1 + joiTi) (TTjurrTT

Ri + R2 _ 1.0+0 25 _c
R2 025 3

r, = aR2C2 = 5 x 0 25 x 3 2 = 4 sec
Ci r2 = R2C2 = 0 25 x 3.2 = 0 8 sec
R,
0 25
1M
X r3 = R, C! = 1 0 x 0 25 = 0.25 sec
e1 0.25 M
eo u = ^r = nr = 0 05 560
3.2
At 4,00,= 4- = 0 25 Rad/sec.

At B,oj2= = 1 25 Rad^ec

At 0,(^3=,^ - 4 Rad/sec

01 02 0 5 12 5 10 20 50

FIG. 10-5. Lag Lead (Notch) Network

FIG. 10-6. Notch Network Used to Stabilize a System (Conditionally)

154
 N
3 cn
d
CM <D 0) 3
O CL ■O
<D CM o O o c
o
CL PO u
> CO $
o Q o
g o 'c OD
N k.
0)
.2 a> 3 3T 0)
~ > o
o o C\J O o
Q_ UJ 1 <

CM

FIG. 10-7. Types of Servos and Nyquist Correction

.a
+ <D <i> a>
c ~ o k. CL T3 cn
o ■=.
o o o o
o O co & 3
O to rO
X) cn a >*
<D
k. c ^
I § a
cr o
Iw s
O O w- <D
cr
7 O
1 CM
0)
>
CL Z 3

O .a >v
<D C O o
CL o rO + TJ
O k_
° O
“5 a> 2
o
3 CO 3
S o
a __ — <D
Q o
E O
•D <D
cn >
a> <d
O CO CO
cr 3 z

cn
c c
CL
o o o
o Q_
tn O
<
CO M
. *2 ” at k_
~D r £ 3 3 a. -g o
<D
o o is cr O
CO <D
cn ° o
3
t
COS CO CO CO
3 H* 3

155
156 FREQUENCY RESPONSE IN SERVO SYSTEMS

The change in size between the input and output is called the gain or
amplification of the device or system, and it may be designated by the
letter K with the proper subscript or by the letter G. The change in the
time of the signal is simply the number of degrees by which the output
differs from the input or by how much the output lags the input. This time
change is usually called the phase angle or phase difference (lag or lead),
and it is given in degrees. It is considered a positive phase difference when
the output leads the input and negative when the output lags the input.
The frequency response of any device or system is the table of values
of K or G and the phase angle, 0, which results when a sinusoidal test
signal is varied over a range of frequencies, usually from 0.1 to 10 or 20
or more hertz. These values may be plotted as two graphs on the same
paper and are known as Bode plots. It is normal practice in the Bode plot
or diagram to use a vertical linear scale for the gain and phase angle and
a horizontal log decade scale for the frequency (semilog paper). The
gain of the device or system is converted to decibels as A or M (dB) =
20 log10K or G, thus allowing one to use the linear scale. The horizontal
log decade scale is the frequency, co, in radians per second.
In the analysis of device or servo operation, two methods are used:
one is by experiment and the other is an analytical procedure. In the
experimental method, a sinusoidal signal from a servo analyzer or function
generator is applied to the input of a system, and the output is measured.
The ratio of the output to the input is determined, and the gain in decibels
is computed. A dual-beam oscilloscope can show the relationship as well
as the phase angle. However, this method may be difficult; therefore, the
analytical method is used because it is simpler and easier.
In the analytical method, the mathematical equation, i.e., transfer
function, that represents the system is used. The transfer function of a
device or system is usually given by an equation ranging from a constant,
to an algebraic lelation, to a differential equation. If the system’s transfer
function is a differential equation, then the use of LaPlace transforms will
change this into an algebraic equation. LaPlace transforms allow one to
go from the time domain to the frequency domain. This operational
method transforms a differential equation from a function of time f(t) to
an algebraic equation F(s) which is a function containing the LaPlace
operator s = a + jco, consisting of a real portion a and an imaginary
portion jco. However, to make the Bode plot it is necessary to change the
operator s to frequency in radians per second or hertz, and this is done by
simply using the real portion a and the imaginary portion jco.
Logarithmic values are used because of the ease with which multi¬
plication and division of numbers are handled. For example, if two
devices are cascaded, then the overall gain is the product of the individual
gains AjA2, but, in decibels, the gain is the sum of dBAj + dBA2. If the
10.4. Bode plot construction 157

transfer function has a reciprocal such as 1/(1 + jcot), then the gain in
decibels is the negative value of dbcox. The logarithm of the frequency, to,
is not quite exact since the dimensions of co are radians per second. Some
texts use co/co0 or cot as the semilog scale for frequency, but, in this text,
we will state that co represents a dimensionless number that has the
numerical value of frequency in radians per second.

10.4. BODE PLOT CONSTRUCTION

A Bode diagram can be plotted for any transfer function with this
function representing any of the following factors:

1. K—a constant.
2. jco or any power of jco, such as (jco)2.
3. 1 + jcox or any power of 1 + j cot, such as (1 + jcox)2.
4. 1 + jcoKj + (jco)2K2 (This is a quadratic factor with complex roots
and it will not be covered here.)
5. Factors 2 and 3 give a leading function. The reciprocal of either of
these, which is the usual thing, gives a lagging function.

When the diagram is actually constructed, the following things can

be noted:

1. Straight lines are used to approximate all magnitude curves.

2. Semilog paper is used.
3. Values of co are plotted on the abscissa, but the graph is log co because
of the log divisions.
4 The ordinate is also a log value of gain, dBA. Therefore, the fre¬
quency curve is a log-log plot.
5. The phase angle or phase shift curve has the same horizontal axis
divisions as the gain curve; thus, its values always correspond to the
gain values at any frequency.
6. When functions are multiplied, their phase angles are added.

When considering any constant, K, we find that it is plotted as a

straight horizontal line with 20 log K or KdB being a constant. If K
equals 1.0, then 20 log 1 equals 0, and the line falls on the OdB-line. K
causes no phase shift, so the phase angle is 0 .
The second factor, jco, has magnitude and phase angles. The value
of j is 90°, j2 is 180°, j3 is 270°, etc.; co is the distance up or down the
158 FREQUENCY RESPONSE IN SERVO SYSTEMS

j-axis. The log gain is in decibels and equals 20 log jco, or 20 log (jco)2
which equals 40 log |jco|. Each jco turn has an upward sloping line at
20dB/decade, thus (jco)2 slopes at 40dB/decade, etc. If the term is 1 /jco,
then 20 log |l/jco| equals 20 log (jco)-1 which gives a negative slope of
— 20dB/decade.
The third factor, 1 + jcox, can be analyzed in the following manner:
t is the time constant and co is the frequency variable. If co is much less
than 1/t, then jcox becomes quite small (negligible), and 1 + jcox becomes
almost 1. Thus, at low frequencies, the function on the diagram is K
equals 1 with a phase angle of 0°. If co is much greater than 1/t, jcor is
very much larger than 1, and 1 + jcor becomes almost equal to j cor. On
a Bode diagram, jco and jcor plot the same except for a parallel shift of
the log gain line. Thus, jcox has a slope of 20dB/decade and a phase
angle of 90°. When this line is drawn, we will find that it intersects with
the K-line at the point called the corner frequency or breakpoint where
1/t becomes equal to 1 (see example 10-2).

Example 10-2
Let us analyze the following function:

C = K _ 20 _ 20
R jco(l + jcox) jco(l + j0.25co) jco[l + j(co/4)]

which has three elements. Each of these elements can be plotted separately
on a Bode diagram, and then their log magnitude values must be added.
For small values of co, the element 1/(1 + j0.25co) becomes 1/1 or 1, and
the total function is 20/jco which has a slope of — 20dB/decade.
In order to start the line, one point is needed so 20 log K/jco =
20 log K/co, when K = 20 we have 20 log K/K = 20 log 1 = 0. There¬
fore, the initial slope either goes through the point co = K on the OdB line
or would go through this point if the line were extended. This is the anchor
point in all asymptotic or straight line Bode diagram plots with jco in the
denominator.
The second line comes from the function 1 + j0.25W because of very
large values of jco which effectively “swamp” the 1 in the function. If
co = 100, then 1 + j0.25co becomes 1 + j25, the absolute value of which
is 112 + 252|. At these high values of co, 1/(1 + j0.25co) approaches 1 /jco
and the total function is (20/jco)(l/jco) = 20/(jco)2 which has a slope of
— 40dB/decade and starts at the breakpoint. In this case, the breakpoint
is where 1 + jco/4 becomes 1 + jl which is when co = 4. Therefore,
from the intersection of the -20dB/decade line and the frequency co = 4,
the final line is drawn at a slope of -40dB/decade (see figure 10-8).
The phase shift curve can also be plotted on the Bode diagram. For
this function, the gain K = 20 is a constant at 0°, so it does not affect the
phase plot. The phase shift actually starts at 90° because of the single jco
 s9aj6aQ-6Di asDiid

FIG. 10-8. A Bode Diagram of the Function

159
 o
o

FIG. 10-9. A Typical Bode Diagram

CJ
lO <D
<0 (/)
c\i \
T3
O
cr
p
cvi i
*

o
00
o
o
lO
o
o
rO
o
OJ
o

S|9qp9Q Uj UID0

160
10.5. Summary 161

term; thus, when to = 0, we have 20/jco, but 20 does not affect the phase,
so 1/jco = 1/90° = -90°. The term 1 + jco/4 combines with the single
term 1/jco to give angles between 90° and 180°. At high frequencies, we
have (l/jco)[l/(l + jco/4)] = l/(jco)2 which is 1/180° = -180°. See
figure 10-8.

Example 10-3
Let co = 20 then

C 20 _ 20
R ~ j20(l + J20/4) ~ j20(l + J 5)
dBA = 20 log 20 - 20 log 20 - 20 log 5
= 26.02 - 26.02 - 13.98 = -13.98 (see Pt. 1)
0 = 0° - 90° - tan-15 = 0° - 90° - 78.5° = -168.5°

10.5. SUMMARY

The following statements can now be made about the Bode plot in
terms of stability.
1. If the phase shift angle is 180° or higher at unity gain (OdB), then the
system may be completely unstable when the loop is closed.
2. The closer the slope of the gain curve is to a -20dB/decade slope
when the plot crosses the unity gain (OdB) line, the more stable the
system will be.
3. The gain crossover frequency is the frequency at which the gain curve
crosses the unity gain (OdB) line. The phase angle is noted at this
point, and, if it is a negative 180° or higher, then the system will be
inherently unstable.
4. To insure stability, the phase angle at the gain crossover frequency
is kept well under — 180°, the amount under is called tho. phase margin.
A phase margin of 40° (phase angle: - 180° + 40° = - 140°) to 60°
(phase angle: - 180° + 60° = - 120°) is desired for most servos.
5. Finally, if it is impossible to get 180° phase shift or if such a phase
shift will occur with extremely low gain, then the system will be
inherently stable.
Figure 10-9 shows a typical Bode diagram with the above information
clearly labeled.

Example 10-4
The following is the open-loop diagram of a servo system.
 6aQ - 3|6u\/ asDiid

- Rad/sec.

162
10.5. Summary 163

1. Write the open-loop transfer function in frequency response form.

2. Calculate the phase angle for <x> = 2, 4, 8, 20, 40, 80 rad/s.
3. Draw the Bode diagram-gain and phase plot on 3 cy. semilog paper
(1) start with co = 0.1 (2) dB scale from -30 to +40 for major
divisions. Show the anchor point.
4. What is the phase margin?
5. What is the gain crossover frequency?
6. Determine from the Bode plot the gain this system will have for a
signal frequency of 30 rad/s. Check by calculation.
7. As the gain of this system is reduced it becomes less and less unstable.
At what gain in dB and signal frequency will it become or start to
become more stable? Determine the value in K that corresponds to
this point.
8. If you did not want to reduce gain then what could be done to
stabilize the system? Explain how and why.

Solution:
1. In order to write the open-loop function, it will be necessary to change
the LaPlace operator “S” to the frequency form of “jco” and also to
calculate the time constant tau (x). Thus

T = 1 _ 0 3 Qz~'n~sec2 _ 0.125 sec

F 2.4 oz-in/rad/sec

90 _ KaKv _ KaKv
e s2x + s s(xs + 1)

and s = jco so

e_o __ KAKy = 5 X 12 = 60
e jco(l + jcox) jco(l + jcoO. 125) jcu(l + j0.125co)

Normalizing the equation gives

e jco[l + j(<w/8)]

j<u[l + j(<u/8)]
164 FREQUENCY RESPONSE IN SERVO SYSTEMS

tan 1 jco = —90°, tan 1 j(co/8) at co = 2,0= — 14°, 0T = —90° +

_14° = -104°

CO 0t
2 -104°
4 -117°
8 -135°
20 -158°
40 -169°
80 -174°

3. Anchor point = -- = — = 60 rad/s

jco jco

Corner frequencies:

1 • CO
<D Cl = JQJ = 1 «C2 = J - = 8
o

4. Phase margin 0PM (solid curves) at a gain of OdB

180° + (-160°) = 20° ans

5. Gain crossover frequency coG c = 22 rad/s.

6. W = 30 rad/s from curve gain in dB = —5.5
dBcal. = 20 log 60 - 20 log 30 - 20 log (30/8) = 35.6 - 29.5 -
11.5 = -5.4 CK.
7. System stable at a 0PM = 40° + from graph co = 9.6 rad/s with
K = 11.5 or 21dB (see construction).
8. A tach. generator could be added (see example 9-2) and then have
the output attenuated so that the viscous damping will be increased
by a factor of as much as 5 or 6. Thus, let FTG = 12 oz-in/rad/s
and JTG — 0.03 oz-in-s2. The tach. generator inertia and viscous
friction will add to those of the servo motor (see equation 9-6) where
KaKtKg = FTG and J = JM + JTG. Therefore a new time constant
must be calculated and a new transfer function written.

, _ 0.3 + 0.03 0.33

0.023 s
_ 2.4 + 12.0 14.4

and

0o = 60 = 60
so coC3 = 43.5
e jco(l + j0.023co) jco[l + j(co/43.5)]
10.5. Summary 165

The new phase lag points are now —92°, —95°, —100°, —115°,
— 133°, — 151°, which shifts the phase curve up and to the right. The
new gain curve is shown by the dotted line extension with a new
corner frequency at a>C3. The new crossover frequency is 52 rad/s.
The new phase margin 0pM = 180° + (—140°) = 40°, which is
acceptable.

PROBLEMS

1. An amplidyne voltage regulator is wired according to the following

diagram.

The amplidyne has the following parameters: 3KW, 125V, 1800 rpm.
Quadrature field = Rq = 1.4 ohms; Lq = 9.138 henries. Control
field Rf = 1000 ohms, Lf = 36 henries. Gain between E and F is 10 at
a>z=0. The generator field Rf - 75 ohm, Lf = 38 henries. The
amplifier resistance can be taken as 1200 ohm. Consider the system
as having three time constants, Tx = generator field, T2 = amplifier
quadrature field, T3 = amplifier and amplidyne control field.
Assuming that there is a 1.0V input at E, calculate the voltage at F
and the phase angle for input frequencies of co = 0, 1, 2, 3, 4, 5,
7.5, 10, 15, 20, 30 and 40.

Open Loop—Switch S at position A

(a) Assuming that there is a 1.0V input at E, calculate the voltage at F
and the phase lag in degrees for the following values of co:
co = 0, 1, 2, 3, 4, 5, 7.5, 10, 15, 20, 30 and 40.
(b) Plot the results as a Nyquist plot on graph paper.
(c) Determine if the regulator will be stable or unstable when the
loop is closed.

Closed Loop—Switch S at position B

(a) If the gain is increased so that F becomes equal to or greater
than E (at the 180° position), what will happen to system
operation?
166 FREQUENCY RESPONSE IN SERVO SYSTEMS

(b) If the plotted curve passed exactly through the — 1 point, what
would be true about the ratio of F to R? What is true about the
stability?
2. The servo system that is shown in the following figure has a magnetic
amplifier that has a time delay of t = 0.05 seconds. Using a Bode
diagram, find the frequency where the open loop has a phase lag of
180°, and also determine the amplifier gain, KA, so that the overall
gain is 1.0 at this frequency (stability limit).

3. Plot the Bode diagram for K = 2.5, K = 15, and K = 25.

4. Plot the Bode diagram for (a) (jco)2, (b) (jco)4, and (c) (jco)3/(jco)2.
5. Plot the Bode diagram for (a) l/(jco)3, (b) l/(jco)2, and (c) (jco)/(jco)3.
6. Plot the Bode diagram for the function 1/(1 + 0.1 jco) by calculating
gain and phase points for co = 0.5, 2, 5, 8, 10, 12, 15, 20, 25, 40
and 100.
7. Plot the Bode diagram for the function 20/[(l + 0.1jco)jco].
(a) Find gain and phase shift at a signal of 100 rad/s.
(b) If C equals 1/c/cf, find R to give the above function with an
amplifier gain of 20.
8. Plot the Bode diagram for the function 15/[jco(l + 0.5jco)].
9. Plot the Bode diagram for the function 15/[s2(l + 0.125s)].
(a) Find gain and phase shift at a signal of co = 2.0, 15 and 20.
(b) Find the gain crossover frequency, and the phase crossover
frequency.
(c) What is the phase margin? the gain margin?
(d) Can this system be stabilized?
10. Plot the Bode diagram for the following system.

km

v
24.9
Vy ''A
s (1 + 0.023s)

(a) Plot the Bode diagram of KM only, KA only, and KA and KM.
(b) What is bandwidth of KM? KA and KM?
10.5. Summary 167

(c) What is the phase margin? gain margin?

(d) What is the phase crossover frequency? gain crossover frequency?
(e) Determine Kv.
(f) Identify KA and KM.
(g) If F is equal to 21.75 oz-in/rad/s, find J.
11. The following is the open-loop diagram of an amplifier and servo
motor and load.

K- Ka = 25

O.I25s2 + s Kt = 1.8

(a) Write the transfer function C/R in the frequency response form.
(b) Plot the Bode diagram.
(c) Calculate and plot the phase angle for co = 2, 6, 8, 12, 20 and 40.
(d) What are the breakpoints?
(e) What is the gain crossover frequency? the phase crossover
frequency?
(f) What is the phase margin?
(g) What gain will this system have for a co of 20? Check by cal¬
culation.
(h) As the gain of this system is reduced, it becomes less and less
unstable. At what value of gain will it become or start to become
stable?
12. A speed regulator with unity feedback has the following open-loop
transfer function

S =_K_
E (1 + 0.04s)(l + 0.0125s)

With the use of a Bode diagram find the value of gain K that will give
a phase margin of 40°.
 ELEVEN

Servo trainers

11.1. GENERAL

Most servo training kits are built in what is called modular or indivi¬
dual pieces or parts that may easily be connected together, both mechan¬
ically and electrically. They are primarily intended for experimental work
in a laboratory giving a general qualitative grasp of closed-loop techniques
as well as some more precise theoretical and analytical techniques.
This chapter examines one such trainer; it is built by Feedback, Inc.,
Berkeley Heights, N.J., especially for those students or readers who are
just starting their study of the servo control field. Figures 11-21 through
11-24 show the different units of their modular dc and hybrid ac servo.
The descriptions that follow, i.e., the theory of unit operation of the
Feedback equipment could, without too much difficulty, be adapted to
cover the trainers built by other manufacturers. Figures 11-25 and 11-26
show the parts that make up the trainer built by Digiac Corp., Smithtown,
N.Y. Figures 11-27 through 11-29 show the parts that comprise the
trainer built by Electro-craft Corp., Hopkins, Minn.

11.2. THEORY COVERAGE

This section is mostly qualitative and covers the following:

1. Measurement of motor characteristics—field and armature control;

transient response; Tach generator calibration

168
71 3. Motor characteristics and operation 169

2. Properties of an operational amplifier; Speed control—forward and

reverse; Effect of forward gain
3. Motor speed control behavior; Determining deadband; The error
channel and closing the loop
4. Reduction of overshoot and settling time—velocity feedback; Effect
of inertia; Stabilization; Field control
5. Synchro link; Demodulator action; Closed loop—synchro error link
6. Simulated relay unit; Relay system—open loop; Relay system—
closed loop

11.3. MOTOR CHARACTERISTICS AND

OPERATION

In any electrical-position or speed-control system, an electric motor

is used along with its power supply and amplifier stage to control the power
input in response to a lower level control signal. The motor should be
reversible.
Figure 11-1 (a) shows the armature connections, and figure 11-1 (b)
shows the field connections. In either circuit, a positive voltage at Vj will
forward bias the NPN transistor and will cause rotation in one direction.
If this positive voltage is applied to terminal No. 2 on the servo amplifier
instead of to terminal No. 1, the motor will reverse direction of rotation.
Each connection has different characteristics, and each has certain
advantages and disadvantages.

FIG. 11-1. Basic Servo Motor Circuitry

170 SERVO TRAINERS

In the connection in figure 11-1 (a), the armature back emf appears
between the emitter and ground, and thus the control voltage Vj or V2
must be increased in order to increase the motor speed. If the motor is
loaded, the speed decreases and the current rises to provide the torque that
is needed to keep the load moving with Vl or V2 being kept constant.
A certain minimum motor voltage is required to start rotation.
In the connection in figure 1 l-l(b), the transistor current is determined
largely by the input signal. Therefore, when the minimum voltage is
reached and there is no load, the motor speed increases greatly with a
small increase in input. Conversely, if the motor is loaded, then the speed
will drop very rapidly. This makes the motor more difficult to control.

11.4. TRANSIENT RESPONSE

The first part of the first experiment, (see figure 11-2 for connections),
will be concerned with the steady-state performance of the motor after a
command signal has caused the motor to respond and then to settle down.
Under ideal conditions, a motor, regardless of the type, should respond
instantly to a command signal, and bring the system into correspondence.
However, in any practical motor, there is always a time lag between the
application of a command signal and the motor response. This lag exists
because a torque must be generated to accelerate the armature and con¬
nected load, and this takes time; thus, the speed rises along an exponential
curve (see figure 8-4, page 117). This exponential rise is exactly the same as
the voltage rise across a capacitor in an RC circuit when the circuit is
subjected to a step-voltage command. The time constant, tm, for the
motor curve is determined by the parameters of mechanical friction,
inertia, armature resistance, and back emf. Figure 11-3 illustrates this,

Power
Servo Supply
—jl Amplifier “|+-
Attenuator
■—|Com.
Unit
i 1“
I i_ .J
3
4 ■ r~
I II

^LL_
i Motor 8
I Tach I
i Com.
T-

FIG. 11-2. Connections for Testing a Servo Motor

11.5. Speed control 171

8q = ktvc (1-6 _T,TM )

80 = Velocity

Vc = Volts

Kt = Speed/Volts Applied

= Motor Time Constant

FIG. 11 -3. Motor Characteristics for the Basic Methods of Control

and it also shows the difference in response when additional inertia is

added to the motor shaft. Adding load will give the same response.

11.5. SPEED CONTROL

Some of the general characteristics of a closed-loop system may be

investigated by or through the use of a simple speed-control system. In
order to control speed, a signal that is proportional to that speed must be
generated, and this may be done through the use of a tachometer-generator.
The tachometer-generator is driven by the motor, directly or through
gearing, and thus generates a voltage which is fed back to the comparator
or summing junction. Here the voltage is compared with the reference
voltage, and the difference between the two is the error signal which
drives the motor to its preset speed level. If the gain of the forward path
(open-loop gain) is large, then only a small error signal is required to
operate the motor. As the forward gain is reduced, the error signal must
increase in a still greater proportion to drive the motor. With gain very
172 SERVO TRAINERS

high, the motor is essentially controlled by the reference voltage. This

may be proved in the following manner:

90 = Ke f)0 = velocity K = gain e = error (11-1)

e = V,-U (11-2)

KG0O — tachometer voltage combining equations 11-1 and 11-2

60 = K(VC — Kg$0) = KVC — KKg0o and dividing by 90

KV KV
1 = ^ - KKg and = 1 + KKg
^0

KV„
9n = (11-3) If K is large, then 60 =
1 + kkg KG

If Kg = constant then 0o = Vc
Also it can be shown that the error falls as the forward gain is
increased.

e = Vc — KKGe and dividing by e

1 = ^ - KKg and ^ = 1 + KKg

e e

Vc
(11-4)
1 + kkg

11.6. THE OPERATIONAL AMPLIFIER

To investigate the response of the motor to a step-position command

during the transient state, an operational amplifier will be used. This
operational amplifier will compare signals and will also provide a high
gain that can be varied. Figure 11-4 shows the three different operational
amplifier circuits that can be used and their resultant voltage outputs.
Figure ll-4(a) is a summing circuit, figure 11 -4(b) introduces a scaling
factor, and figure ll-4(c) gives an output that is comparable to that of an
ideal motor. When the circuit of 11 -4(b) is used with a feedback signal
plugged into the external feedback sockets, full amplifier gain is not
available. Under any condition, the maximum amplifier output is about
12V. When Rx = R2 (control switch to the left), the output is the sum
of the input voltages but with a minus sign. If Rx # R2, then a scaling
factor is introduced, and, when Rx = R2 and a capacitor is in parallel
173
174 SERVO TRAINERS

FIG. 11 -5. Connections for a Speed Control System

with the feedback resistor (control switch in the middle), the output
voltage has the form shown in figure 11-3. When the control switch is at
the right, provision is made to connect an external feedback signal around
the amplifier across terminals 4 and 5. If it is necessary to zero the oper¬
ational amplifier, it may be done by measuring the output voltage at
terminal 6 with no inputs at terminals 1, 2, or 3 and with the control
switch at the left. The output should be zero, thus showing that the + 15V
supplies balance internally. If the output is not zero, then adjust the zero
adjustment knob until the output is zero.

11.7. REVERSIBLE SPEED CONTROL

When a servo system requires control in either direction of rotation,

it is necessary to reverse the reference voltage; this means that the final
servo amplifier must have two input sockets. The operational amplifier
does not have such reversible ability, so a preamplifier unit is introduced
between the attenuator unit (reference potentiometer) and the servo
amplifier.
The preamplifier has two output sockets and two input sockets as
shown in figure 11-6. A positive voltage applied to either input gives a
positive output at one output socket; a negative voltage applied to either
input gives a positive output at the other output socket. These two output
sockets then provide the drive for both inputs on the servo amplifier, and
the motor will respond in either direction. The overall gain of the pre-
11.7. Reversible speed control 175

+ 15 V

Ref i-;
1 Pre | i Servo i
i i
-i Amp l Amp i
I_I i__ _j

-15 V

FIG. 11-6. Connections for Reversible Speed Control

amplifier is about 25, and maximum output is about 12V. Figure 11-7
shows the better speed control and gain associated with the preamplifier.
If it is necessary to zero the preamplifier, it may be done by measuring
the output at terminals No. 3 and No. 4 when there is no input to terminals
No. 1 and No. 2. These output voltages should be approximately 1.0 to
1.2V, thus showing that the + 15V supplies balance internally. If not, then
adjust the zero adjustment knob until the voltages balance. If they still
will not balance, then the “preset” zero can be adjusted.
In the simple speed-control system that is shown in figure 11-5,
whenever a step-position change of voltage is made, the response is such
that the motor speed comes up to the final value without overshooting, as

FIG. 11-7. Speed Control Characteristics for Different Values of Gain

176 SERVO TRAINERS

FIG. 11-8. Motor Responses for the Speed Control System

is illustrated in figure 1 l-8(a). As long as the signals in the system do not

exceed the operating range of each unit, /.<?., saturation of an amplifier,
the response follows an exponential curve. If the maximum signal ranges
are exceeded, then response will be slower.
If an additional time constant should take place in the forward path,
then the response may be such that overshooting takes place. The reason
for this is that the extra time delay prevents the increased signal to the
servo amplifier from being reduced as quickly as the error is reduced; thus,
the motor may oscillate before settling down. This is illustrated in figure
ll-8(b). To observe the effect of an additional time delay on motor
performance, the RC network in the operational amplifier should be
inserted in the forward path.
In the speed-control system that is shown in figure 11-9, the output
of the tachometer-generator is fed into the operational amplifier, rather
than the preamplifier, so that the error signal can be displayed. The error
is the difference between the command voltage and the tachometer
voltage. When the error is zero, the system is running at some definite
speed and is termed a velocity control system.
Figure 11-10 shows the display of the velocity signal or tachometer
output and the error signal for different values of gain. It clearly shows
that when gain is high, deadband is small, and, when gain is low, deadband
is large. It also shows that high gain gives very good speed response with
an error of peak value only at the instant of motor reversal. However,
when gain is reduced, the response is still good, but the final speed of the
motor is reduced and there is a larger error.
When an additional time delay or a larger load is added, the system
becomes more and more oscillatory. If both occur, then the response is
slowed down greatly and the error change is also slowed down. This may
be caused by the fact that the power supply is overloaded, and it just
cannot supply the necessary power.
1

£
o
4-*
C/)
>*
cn
~o
4->
c
o
o
>
‘o

(D
>
<D
4->
_<d
Q.
E
o
o
<

CD

177
178 SERVO TRAINERS

High
Gain L Error

Med.
Gain V

Low
Gain

Added
Time
Delay

FIG. 11-10. A Comparison of Velocity and Error Signals for Different Values of Gain
in a Velocity Control System

11.8. THE ERROR CHANNEL

Another scheme for investigating the characteristics of a closed-loop

system is that of mechanical position control (a true servomechanism)
where a motor drives an output shaft to the same angle as that of an input
shaft. Potentiometers, synchros, and rotary LVDT’s are excellent trans¬
ducers to use in such a system. In dc servos, potentiometers are used
most often.
A reference or input potentiometer and an output or feedback
potentiometer form an error channel when properly connected with their
slider voltages summed via the operational amplifier. To properly connect
the potentiometers, their corresponding winding ends, (1, 2) and (2, 1),
must be connected to equal and opposite dc voltage supplies so that, if
the two shaft angles are equal, the slider voltages will be equal and will
cancel (see figure 11-11). Negative feedback is thus obtained. Any sudden
change in angular position between the two shafts will give an error signal
which, in turn, will drive the forward path. When the loop is closed by
connecting the slider of the output potentiometer to the operational
amplifier, the polarities of the signals must be carefully observed because
the error signal should cause the servo motor to rotate in a direction to
reduce the error, not to increase it.
11.9. Genera! performance 179

Error

KE (0,-00)

Input-1 1---
Output From Motor

FIG. 11-11. Potentiometer Error Channel for a Position Control System

11.9. GENERAL PERFORMANCE

Two characteristics of a simple position servo that are of interest are

the deadband and the transient response. The deadband is the minimum
signal that is required by the system in order for it to respond. In the
first experiment, the motor did not start until a minimum voltage was
applied, this is the deadband, and it is inversely proportional to gain.
The transient response is studied in terms of overshoot and settling
time. Overshoot takes place because of system response to an input
change. If a step-position command occurs, an error signal is set up which
will drive the motor until it passes through the alignment point. At this
point, the motor has velocity and will drive on through the point; thus,
the error signal reverses, and the motor slows down, stops, reverses, and
tries to drive to realignment in the reverse direction. But again, overshoot¬
ing may result until the system finally settles down. For low values of gain,
little overshooting takes place, but, as gain is increased, the overshooting
and the settling time both increase as is shown in figure 11-12.

FIG. 11-12. Transient Motor Conditions

180 SERVO TRAINERS

It can now be seen that, if gain is low, the deadband is large and the
system responds slowly; as gain is increased, deadband is reduced and
oscillations now occur, but system response is quicker. Ideally, gain
should be high, deadband low, and response time low with little or no
overshoot. One way to achieve this is to add additional load via the
magnetic brake, but this means that additional power is required from the
motor, which may not be available.
A second way to achieve this is to use a signal that is proportional to
motor speed or velocity, i.e., a signal from the tachometer-generator. If
the drive power could be reversed before the system reaches the alignment
point, then the motor would already be slowing up as this point is reached,
and the overshoot would be greatly reduced. This is shown in figure
11-13(a). The signal from the tachometer-generator is a direct measure of
the output shaft speed which may be increasing as the system comes into
alignment. If this signal is fed back into the summation point and is
negative when it is added to the error, then the sum of the signals becomes
zero and may even reverse before the system reaches alignment. Therefore,
the motor drive reverses before alignment, and the system slows down
without the excessive overshoot. The major disadvantage is that the speed
of response decreases as velocity feedback is increased. See figures 11-13(b)
and 11-13(c). Figure 11-13(b) is the general block diagram and shows that
the tachometer-generator is connected in a second or inner loop. Polarities
must be carefully observed around both the outer and inner loops in order
to ensure that the system responds correctly.
A third way is to increase the effective inertia of the motor by adding
more inertia to the motor shaft by using an inertia disc. If this inertia disc
is placed on a shaft that is geared to the motor, then its effect will be quite
a bit different. This is because of the gear reduction ratio, N. Inertia on
a low speed shaft would be reduced by 1/N2, or by about 1/900 at the
armature for this servo.

11.10. INSTABILITY AND STABILIZATION

In the experimental work done so far, any transients that have

occurred have died away, thus giving a stable system. However, in many
practical systems, the oscillations may continue, rather than decrease, and
the system becomes unstable. This is particularly true when there is a lot
of inertia but little or no friction.
Any additional time delays in the forward path that have been
introduced for whatever reason will also lead to instability. Filters that
are used to eliminate noise or other spurious signals contribute to the
problem of stabilizing a system.
 No Velocity Feedback

FIG. 11-13. Effect of Tach Generator (Velocity) Feedback

181
182
Z)

o
dCL
. •
Q. O
Disturbance

FIG. 11-14. A Complete Mechanical Position Control System with Signal Traces
11.12. Relay-controlled systems 183

Increasing the sensitivity of a system (reducing deadband) by using

field control, rather than armature control, may also lead to instability,
unless other corrections are taken. The speed of an unloaded motor with
field control increases to a very high value with low input (see figure 11-3).
Velocity feedback shifts the characteristic to a level that is closer to that
of an armature-controlled motor.
Figure 11-14 shows the complete mechanical position control system.
The oscilloscope traces show the effect an outside disturbance has on the
desired position with resulting error and finally the velocity stabilization.

11.11. AC CONTROL SYSTEMS

All the control systems studied on the previous pages have used a dc
error channel, a dc servo motor with potentiometers being used as the
error detectors. If one desires that synchros be used as error detectors,
then the ac error signal must be demodulated to provide a dc signal to
drive the servo motor. There is no requirement for an operational
amplifier because the difference between input and output signals is
actually produced in the synchro CT. A hybrid system using synchros
and a demodulator is shown in figures 11-15 and 11-16.

11.12. RELAY-CONTROLLED SYSTEMS

There are servo systems in which the motor is controlled by a relay

rather than by an amplifier. Figure 11-17 shows the general schematic
diagram of such a system. The error signal will energize the relay coil
which in turn will switch power to the motor. The relay may be three
position, ione position drives the motor in a forward direction, a second

,_Command
/

rv / CT

FIG. 11-15. Synchro Transducers

184
Command

<
o
<
CM
O
i
i
i
+
in
>

—i
_J n
—

CL
4-

3
Ti

o
c
O
"
a.<7)

_> c£
C7>
'

i—I1,
>
<D
■S
5

8
S

FIG. 11-16. A Complete Mechanical Position Control System with Synchro Transducers
11.12. Relay-controlled systems 185

Output

FIG. 11-17. Basic Relay Connections

position drives the motor in the reverse direction, and a third position
results when the relay is deenergized and the motor is stopped. The relay
might be two position which means that power is always applied to the
motor, either in one direction or the other.
In reality, the relay is an amplifier in which a small amount of coil
power controls contacts, which in turn control a larger amount of power
that is applied to the motor. However, the power that is applied to the
motor is not of a continuous adjustment as it would be with an amplifier,
therefore, the performance of the relay system is not as precise as that of a
power amplifier.
The characteristics of relays may be represented by plotting output
against input. A two-position relay that controls power in either direction
and changes over when the coil signal (error) changes sign has the
characteristics shown in figure 11-18(a); it is an “ideal” one. Hysteresis,
which most two-position relays have, is shown in figure 11-18(b), the relay
cannot operate there until a positive or negative signal voltage of a

Output

Ir
For. —

Rev.
Input -e2 + e2 -e,
+ e' J.1
Ideal Hysteresis Deadband or Deadband
Off Position and Hysteresis
Two Position Three Position

(a) (b) (c) (d)

FIG. 11-18. Characteristics of a Two-position and a Three-position Relay

 186 SERVO TRAINERS

FIG. 11-19. Effect of Adjustable Slope and Deadband

minimum value is present. A three-position relay with an “off” position

gives the deadband characteristic that is shown in figure 1 l-18(c); here,
an input of plus or minus e is required to operate the relay. Figure 11-18(d)
shows the characteristic of a three-position relay that has both deadband
and hysteresis on the switching level.

Input

+ 15 V

i Op. Output
Amp. Relay

T
- 15 V
i_

+ 15 V

1 Gain

Output

15 V

FIG. 11-20. A Relay Controlling a Mechanical Position Servo

The unit that is used to simulate these characteristics does so elec¬

tronically, and it also provides some additional characteristics which relays
do not have, namely (1) limiting with adjustable slope and deadband, and
(2) adjustable backlash. Because it is a small unit, it does not have enough
power to operate the motor directly; therefore, the output feeds the
normal preamplifier and power amplifier. Figures 11-19 and 11-20 show
the experimental connections.
FIG. 11-21. DC Servo Trainer (Courtesy of Feedback, Inc., Berkeley Heights, N.J.)

FIG. 11-22. Miscellaneous Servo Devices, Synchros,

etc. (Courtesy of Feedback, Inc., Berkeley Heights, N.J.)

187
188 SERVO TRAINERS

FIG. 11-23. Miscellaneous Servo Devices, Modulator/Demodulator, etc.

(Courtesy of Feedback, Inc., Berkeley Heights, N.J.)

FIG. 11-24. AC Servo Trainer with Logic Bits and Shaft Encoder (Courtesy
of Feedback, Inc., Berkeley Heights, N.J.)
11.13. Digiac corporation servo trainer 189

11.13. DIGIAC CORPORATION SERVO

TRAINER

The philosophy of this company in the design and production of its

Servomechanisms trainer is that the experience that is gained in the
laboratory is one of the most important aspects in preparing a student
for work in the field of automatic control of machinery.
The laboratory equipment shown in figures 11-25 and 11-26 uses dc,

FIG 11 -25. Servomechanisms Trainer (Courtesy of Digiac Corp., Smithtown,

N.Y.)

rather than ac, because the dc servomotor and its associated devices dem¬
onstrate in a much clearer fashion the basic principles of feedback control.
The approach is essentially practical with very little dependence on higher
mathematics, i.e., the calculus. The experimental procedures that can be
followed with this equipment lead from the very simple to the more
complex. However, the equipment has a flexibility that allows a student
to do more advanced experiments if he so desires.
190 SERVO TRAINERS

FIG. 11-26. Servo Motor and Transducer (Courtesy of Digiac Corp.,

Smithtown, N.Y.)

11.14. ELECTRO-CRAFT CORPORATION

SERVO TRAINER

The philosophy of this company in the design and production of its

Servo Trainer or Control Systems Laboratory equipment is that the
experience that is gained in the laboratory is one of the most important
aspects in preparing a student for work in the servo or automatic control
field.
The laboratory equipment (Motomatic Control System) is dc in
nature because the dc servomotor and controller principle has proved to
be far superior to the traditional ac motor approach for many industrial
applications. All the principles of feedback control can be demonstrated
in a much more lucid fashion by using dc rather than ac servo devices. The
equipment has been designed in such a manner, i.e., flexibility, that many
different experiments may be performed other than those that are con¬
sidered standard. The following figures 11-27 through 11-29 illustrate
very graphically what this equipment is.
7 7.14. Electro-Craft corporation servo trainer 191

FIG. 11-27. Motomatic Servo Trainer (Courtesy of Electro Craft Corp.,

Hopkins, Minn.)

FIG. 11-28. Servo Motor and Speed Control Unit (Courtesy of Electro
Craft Corp., Hopkins, Minn.)
192 SERVO TRAINERS

FIG. 11-29. Motomatic Servo Experimenter (Courtesy of Electro

Craft Corp., Hopkins, Minn.)
 TWELVE

Other electro-mechanical devices

SHAFT ENCODERS

12.1. INTRODUCTION

When one desires to change an analog signal such as speed, voltage

or position to a digital signal to obtain greater accuracy, or when it is
desired to operate with very small error signals, it is best to use a digital
servomechanism. In order to deal with both signals, i.e., to go from
analog to digital, an A to D converter is needed. This conversion may be
accomplished by a shaft encoder in which the shaft position is an analog
measurement, and it may represent the direction of a radar screen, the
position of the rudder of an aircraft or any number of different automatic
control applications.
An encoder is an electromechanical device that can be fastened to a
rotating shaft to produce a series of pulses to indicate shaft position. As
it has a shaft, housing, and electrical outputs, it looks somewhat like a
motor. In reality, it is simply a rotating switch with a number of poles
and positions that vary depending on the type of encoder.
Encoding, which means assigning digital values to mechanical
motion takes place when the shaft of the encoder is coupled, either through
gears or a flexible coupling, to the output shaft of the system under test
or study. It thus produces coded combinations of outputs from analog
inputs, namely, analog to digital conversion. Two basic types of encoders
are (1) the incremental or pulse type, and (2) the absolute type. These two
types look somewhat alike, but they differ quite a bit in their method of
operation.
193
194 OTHER ELECTRO-MECHANICAL DEVICES

FIG. 12-1. A Simple Gear-type Mechanical Encoder

12.2. INCREMENTAL ENCODER

The basic operation of an incremental encoder is that, as its shaft is

turned, a stream of pulses is generated on a pair of output wires. Each
pulse represents some fraction of a revolution of the shaft. A mechanical
analogy would be a gear with a small number of teeth rotating so that the
teeth close a snap-action, spring-return microswitch. Each time a tooth
touches the switch and closes the circuit, an output voltage can be
measured (see figure 12-1).
One complete revolution of the gear would produce eight pulses, each
pulse representing 45°. At each tooth, the pulse is the same as the one
before it; therefore, the angle is determined by storing and counting the
pulses. Thus, the count at any time is the number of angular increments
through which the shaft has been rotated.
The most common way to generate a pulse is through the use of optics.
Here, a disc that has been constructed with alternating transparent- and
opaque-rimmed sections is rotated between a light source and a photo¬
detector. The opaque or dark segments of the disc serve to interrupt the
light source, and thus a pulse is created. Figure 12-2 shows such a disc.
In this example, however, the electronics cannot “know” the direction of
shaft rotation because either direction gives the same pulse trains. An
encoder built in this manner is called a unidirectional incremental encoder.
To become bidirectional, more complex circuitry is required. Also, if
power fails or is turned off, the count is lost. Another disadvantage is that
extraneous electrical noise or pickup might very well be counted as a
pulse and thus produce a wrong count.
Advantages really depend on the application. For example, (1) if rate
or velocity is being measured, (2) if counting of pieces or parts is required,
(3) if a zero reset regardless of shaft position is required, (4) if the shaft
always rotates in one direction, or (5) if loss of power or noise is not a
problem, then an incremental encoder will be the best to use.

12.3. ABSOLUTE ENCODER

The basic operation of an absolute encoder is that it consists of a

coded disc and brushes that sense the code pattern. There are usually
12.4. Basic techniques 195

Light
o [].. ... Output

<4=
Disk

FIG. 12-2. A Simple Optical Disc Encoder

several concentric tracks on a disc with each track being composed of

alternately conducting and nonconducting segments. The innermost track
will have fairly long segments but only a few of them, while the track at
the outside of the disc will have a large number of smaller segments. Each
track will thus differ in segment length and number as the track diameter
changes. Depending on the code being used, each of the different tracks
may represent distinct digits or the output wires which make up a digit.
Figure 12-3 shows an elementary type. These tracks normally mean that
the absolute encoder has quite a few more leads than the incremental type.
If slip rings or long leads are required to bring the outputs out, then there
will be an increased cost. Other disadvantages are that rate or velocity
data cannot be measured easily, and, when zeroing is required, it cannot
be done easily, unless it operates into a computer.
Advantages are that they are insensitive to noise or pickup, a count
cannot be lost because of power failure, and digital outputs wdl follow
changes in shaft rotation without any problem.

FIG. 12-3. A Simple Coded Disc Encoder

12.4. BASIC TECHNIQUES

As was stated in the beginning, an encoder is like a rotating switch

with many poles and positions. Now there are basically two ways of doing
this switching; they are referred to as (1) contacting, and (2) noncontacting.
In the contacting style, a wire brush moves along a coded disc or com¬
mutator, thus the name brush type encoder. In the noncontacting style, a
circuit parameter such as voltage, current, or resistance changes quite
abruptly in response to the switching of a light beam, a magnetic field or
196 OTHER ELECTRO-MECHANICAL DEVICES

capacitive coupling. The most common noncontacting method is through

the use of optics (light beams).
As to which one is best to use, it is again a toss up, and it depends on
the application. Thus, pick an optical or incremental encoder when (1)
more than 10,000 counts/turn accuracy is required, (2) very, very low
starting torques of less than 0.1 oz-in are required, and (3) high speeds
especially above 5000 rpm and of continuous nature are required. A
brush type or absolute encoder would be picked for applications when (1)
environmental problems such as dripping liquid, acidic atmospheres, and
heavy shocks are prevalent, (2) high wattage of 50W or more is needed to
the load, and (3) other applications where long life and reliability are
demanded.

12.5. LOGIC AND CODES

In any shaft encoder, it is quite difficult to synchronize positionally

each brush/track or optic circuit combination with all the others. To
prevent signals that are out of sequence from reaching the output, logic
circuits are used to ensure that all brushes and tracks or optics are pre¬
cisely in alignment. Logic circuits help to eliminate electrical noise
generated by the brushes or optics, and they also help to provide longer
life for the encoder.
When encoders were first used, industry was using counting systems
based on the powers of ten; however, as computer technology became
more advanced, a different counting system was used. A relay has an
on-off characteristic the same as a transistor so that a counting system
could be used. Thus, the number 3150 would be represented as
1 x 211 + 1 x 210 + 1 x 26 + 1 x 23 + 1 x 22 + 1 x 21; all other
coefficients of two are zero. The number is written as 110001001110 and
would be instrumented as a 12-wire line. The wires connecting or repre¬
senting the “Ts” would have voltage on them relative to a common point,
while the “0’s” are at zero potential (see figure 12-4).

1
1 All Wires With “l,s”
0 Have Voltage With
0 Respect to Ground,
0
1 The Others are at
0 Zero Potential.
0
1
1
1
0 12 Wire Line

FIG. 12-4. A 12-wire line in a Simple Counting

System
12.5. Logic and codes 197

TABLE 12-1
A Comparison of Three Counting Codes
Decimal Binary Gray

0 0 000
1 1 001
2 10 Oil
3 11 010
4 100 110
5 101 111
6 110 101
7 111 100

If one desires to combine both the decimal and the binary system,
then the Binary-Coded Decimal (BCD) system in which four binary rows
are used for each digit used. The number 3150 is now represented as:

23 22 21 2°
\\ \ I
0011 0001 0101 0 0 0 0

3 1 5 0

Shaft encoders that use brushes create problems in that, if one of the
brushes accidently touches an adjacent row because of mechanical
problems, then large errors in the count could result. To prevent this, the
Gray code is used so that any chance of error is eliminated; however,
additional brushes and more electronic circuits are required. To go from
binary to the Gray code is a simple matter. The number written in binary
is converted by changing any digit immediately preceded by a “1” to its
opposite state. The number 3150, for example, would change from
110001001110 to 101001101001. Table 12-1 shows the relationship
between decimal, binary and Gray code numbers.
If we look at the binary code, we can see why the Gray code is more
advantageous to use. When the binary code advances from 3 to 4 (011 to
100), all three bits must change state from “0” to “1” or from “1” to “0.”
However, it is almost impossible to design mechanical switches that will
all change state at exactly the same time. Therefore, if some switches open
or close sooner or later than others, the transition from Oil to 100 may
appear to go through several states such as Oil-Gil-GO 1-GOO. The
false states may only last for a very short period of time, but a digital
computer might respond to them, and erroneous outputs could result.
Note that as the Gray code is advanced from number to number, only
one bit of the code changes state.
Figure 12-5 shows a binary-coded disc, and figure 12-6 shows a
Gray-coded disc. In these discs, the black sections may represent con-
 OTHER ELECTRO-MECHANICAL DEVICES
198

FIG. 12-5. Binary Code Disc

ducting surfaces so that brushes riding on the disc at each of the con¬
centric rings can give an output whenever they make contact with a
conducting surface, or they may represent opaque sections on a trans¬
parent disc so that light is blocked for the proper digits.
The application in which the encoder is used really determines which
code is the best to use. Industrial applications have used the decimal code
for years, especially for outputs that are displayed, so they probably will

FIG. 12-6. Gray Code Disc

12.7. Types 199

continue to use this code. If one desires to have the readout used with a
computer, then the BCD might be best. Binary code is the heart of
computer language, yet the Gray code is capable of a greater number of
bits in relationship to the pulses generated.

STEPPER MOTORS

12.6. GENERAL

A stepper motor is an electromagnetic device that converts digital

pulse inputs into analog output shaft movements with these steps being
equal to the number of input pulses. They can thus be used for remote¬
positioning control systems in a wide variety of industrial drives such as
printers, tape drives, capstan drives, machine tool and numerical control
drives, and process control drives. They also can be used for variable
speed drives such as curve tracers, and feed drives in chart recorders.
Advantages of the stepper motor are (1) that being a digital device, it
interfaces readily with a computer and other digital components, (2) it has
long life and fast response, and (3) it has good open-loop stability when
properly matched to the system’s requirements. This means that other
transducers such as a tachometer-generator, synchros, and potentiometers
can be eliminated.

12.7. TYPES

Five basic types of stepper motors are in general use; they are (1)
variable-reluctance type, (2) solenoid-ratchet type, (3) permanent magnet
type, (4) synchronous type—phase controlled, and (5) harmonic-drive type.
The two most widely used types are the variable reluctance and the
permanent magnet. Variable-reluctance types usually provide higher
torques, larger stepping rates, and smaller stepping angles than do the
permanent magnet types, although the newer permanent magnets are
about the same in performance. The permanent magnet type, however,
does have two other important advantages, and these advantages are
(1) better damping, and (2) small overshoot.
The permanent magnet type usually consists of a cylindrical perma¬
nent magnet rotor rotating within a laminated, slotted stator which
contains the windings. The rotor is magnetized along a diameter and may
have one or more pairs of poles. The stator winding may be two, three,
four, six or eight phases. In operation, the rotor lines up with the stator
200 OTHER ELECTRO-MECHANICAL DEVICES

FIG. 12-7. Simplified Diagram of a 2-pole Stator

magnetic field produced by applying a dc voltage to the stator coils. When

the polarity of the dc voltage is switched in a definite order, the stator field
will rotate in either clockwise or counterclockwise steps.

12.8. SWITCHING MODES—

PERMANENT MAGNET TYPE

For simplicity in the following discussion, the windings will be

considered to be wound on salient poles. Figure 12-7 shows such an
arrangement for a two-pole stepper motor (two poles in each of the two
phases).
With the positive of the supply connected to SI and the negative
connected to S3, a magnetic field is produced by the stator. The rotor lines
up with this stator field as is shown in figure 12-8(a) (solid shading indicates
a north pole). If the positive is now switched from SI to S2 and the
negative from S3 to S4, then a magnetic field is produced as is shown in
figure 12-8(b). The field has now rotated 90°. Subsequent polarity changes
as are shown in figures 12-8(c) and 12-8(d) rotate the stator field, and thus
the rotor, through 360° in 90° increments. It should be noted that each
position is unique, and that no positional ambiguity exists.
The sequence of polarities to obtain 90° stepping as described above
is given in table 12-2.

TABLE 12-2
Switching Mode "A" (90° Step Angle)
Step Position SI S2 S3 S4

1 Fig. 12-8(A) + 0 —
0
2 Fig. 12-8(B) 0 + 0 —

3 Fig. 12-8(C) — 0 + 0
4 Fig. 12-8(D) 0 — 0 +
12.8. Switching modes—permanent magnet type 201

Another method of obtaining a 90° stepping mode is to use the resultant

of two fields as is shown in figure 12-9.
The sequence of polarities to obtain 90° stepping as shown in
figure 12-9 is given in table 12-3.
Since with this method two windings are energized at the same time,
the current and power input are 100% higher than with the previous
method, but the torque is only 41 % higher.
It is obvious from figures 12-8 and 12-9 that a proper combination
of the two methods will result in a 45° stepping mode. The polarity
sequence is given in table 12-4.
Switching the polarities as is shown in tables 12-2, 12-3 and 12-4 is
readily accomplished with mechanical switching devices. To do this

TABLE 12-3
Switching Mode "B" (90° Step Angle)
Step Position SI S2 S3 S4

1 Fig. 12-9(A) + + - —
2 Fig. 12-9(B) — + + —
3 Fig. 12-9(C) — — + +
4 Fig. 12-9(D) + -1-
 202 OTHER ELECTRO-MECHANICAL DEVICES

FIG. 12-9. Step Positions—Switching Mode "B"

electronically with solid-state components is rather complex because two

polarities have to be switched.
In order to simplify the electronic driving circuit, stepper motors are
usually made with a center-tap in each winding. They are then called
four-phase stepper motors. A simplified diagram is shown in figure 12-10.
By connecting both center-taps to one terminal, e.g., the positive, of
the supply and switching the other terminal as is shown in tables 12-5 and
12-6, rotation of the stepper motor is the same as in figures 12-9 and 12-10.
By combining the switching modes of tables 12-5 and 12-6 in the
proper sequence, 45° stepping is also obtained as is shown in table 12-7.

TABLE 12-4
Switching Mode "C" (45° Step Angle)
Step Position SI S2 S3 S4
1 Fig. 12-8(A) + 0 _ 0
2 Fig. 12-9(A) + + — _
3 Fig. 12-8(B) 0 + 0 _
4 Fig. 12-9(B) — + 4- —

5 Fig. 12-8(C) — 0 + 0
6 Fig. 12-9(C) — —
+ +
7 Fig. 12-8(D) 0 —
0 +
8 Fig. 12-9(D) + — — +
12.8. Switching modes—permanent magnet type 203

FIG. 12-10. Simplified Diagram of a 4-phase Stator

TABLE 12-5
Switching Mode "D" (90° Step Angle)
Step Position SI S2 S3 S4

1 0 0 — 0
2 0 0 0 —
3 — 0 0 0
4 0 — 0 0

TABLE 12-6
Switching Mode "E" (90° Step Angle)
Step Position SI S2 S3 S4

1 0 0 — —

2 — 0 0 —
3 — — 0 0
4 0 — — 0

TABLE 12-7
eno

Switching Mode "F" Step Angle)

Step Position SI S2 S3 S4

1 0 0 — 0
2 0 0 — —
3 0 0 0 —
4 — 0 0 —
5 _ 0 0 0
6 _ — 0 0
7 0 — 0 0
8 0 — — 0
204 OTHER ELECTRO-MECHANICAL DEVICES

1 2 345678 9 10 II 12

FIG. 12-11. Switching Waveforms for 30° Steps

12.9. A COMPARISON OF SWITCHING

MODES

A comparison of the various switching modes is given in table 12-8.

Mode “E” is used as the basis for comparison since it gives the highest
torque for the simplest electronic driving circuit.
Four-phase stepper motors can of course be used as two-phase
motors by leaving the center-taps disconnected.
Three-phase stepper motors operate in a similar manner to those
types that have already been covered. The switching wave forms obtained
from 30° stepping are shown in figure 12-11.
Two ways in which the switching of the dc driving voltage can be
accomplished are through stepping transmitters (mechanical drives) and

TABLE 12-8
Comparison of Switching Modes
Switching Total Current Torque Per Unit
Mode Torque and Power Input of Power Input
A 70% 25% 280%
B 100% 50% 200%
C 85% Avg. 37% Avg. 230%
D 70% 50% 140%
E 100% 100% 100%
F 85% Avg. 85% Avg. 100%
12.10. Developed torque of stepper motors 205

U) + V
a v
o> j/>
(7i 3
CL 0
+ V

to
0
+v
CM
if)
0
+V
ro
If)
0- - -
+ V- - i—
V
If)
0 - -

4 1 2 3 4 1 2

Rotation — CCW

FIG. 12-12. Typical Logic Waveforms from Electronic Driver

all electronic drivers. The mechanical transmitter can be fitted with

microswitches operated by an eccentric shaft and the correct gearing so
that high resolution is obtained. If diodes are used for suppression, then
correct power line polarity must be observed. Figure 12-11 shows the
voltage waveforms developed across a three-phase stepper motor.
The electronic driver is an all-transistorized logic driver that accepts
pulses on one or two lines (clockwise or counterclockwise) and processes
these pulses via two flip-flops operated in a forced count sequence. The
output of the flip-flops is then fed through buffer amplifiers to drive the
stepper motor. Figure 12-12 shows some typical logic waveforms.

12.10. DEVELOPED TORQUE OF

STEPPER MOTORS

Static Conditions. At rest, in a particular step position, the rotor is

lined up with the stator field, and no torque exists to move it to a different
position. If the rotor is moved away from this lined-up position by
external mechanical means, then a restoring torque will be developed.
It should be noted that at 90° displacement maximum torque is
developed. This is called holding torque and is used as a figure of merit.
The stepper motor cannot drive a load requiring a torque equal to the
206 OTHER ELECTRO-MECHANICAL DEVICES

holding torque since this occurs only at 90° displacement. At other

positions, the torque is lower. Pull-out torque (sometimes called running
torque) is the proper indication of the loading capacity of the motor. The
pull-out torque varies with the switching rate, and it is limited by the
inherent damping in the motor. Load inertia has little or no effect on
pull-out torque.

Dynamic Conditions. If instead of moving the rotor as described under

“Static Conditions,” the polarities are changed so that the stator field
moves to the next step position, then the same relative condition occurs.
In the case of a 90° stepper motor, maximum torque is developed at the
moment of switching. Because of this torque, the rotor moves, and,
without any load connected to it, it will line up with the stator field at the
new step position.
However, if a load is connected to the rotor, then the load torque
tends to hold the rotor at a certain displacement from the stator field
position. The combined inertia of the load and rotor has an oscillatory
effect. Damping is present in the form of friction, eddy currents and
hysteresis.
In between the regions of operation mentioned above, there is a range
of switching rates where the stepper motor behaves erratically. Stepping
still occurs but with irregular step angles. The pull-out torque varies
considerably, and it is at certain switching rates lower than at higher rates.
By careful selection of load inertia and additional damping, this effect can
be minimized.
If the switching rate is increased and operation becomes stable as
mentioned above, there comes a point, with further increase in switching
rate, where the motor cannot accelerate the load plus motor inertia from
standstill to synchronous speed. Here, the pull-in rate, which is the top of
the response range, is reached. This means that between the zero switching
rate and the pull-in rate the motor can start, stop and reverse on command.
A small nebulous range around the pull-in rate exists wherein the stepper
motor can bring the load to synchronism, but not without losing some
steps. The pull-in rate varies with the total inertia. In the response range,
the stepper motor can be successfully used as an open-loop positioning
device with no-load step position errors of about 1°. A load that is applied
to the stepper motor will increase the line-up inaccuracy by an amount
determined by the torque gradient.
If the switching rate is increased still further while the motor is
running, the slew range is reached. In this range, the motor cannot start,
stop or reverse on command, but it develops enough torque to overcome
the load torque. This is a useful range, provided means can be devised to
bring the motor to these speeds out of the response range.
12.11. Use in control systems 207

FIG. 12-13. Stepper Motor Characteristics

The switching rate can be increased until the pull-out range is reached,
at which point the inherent damping decreases the available torque to
below the level of the load torque. This is at the top of the slew range.
A summary of the performance characteristics explained above is
given in figure 12-13.

12.11. USE IN CONTROL SYSTEMS

There are two distinctly different ways of using stepper motors in

control systems. One is the open-loop system, and the other is the
positional feedback or closed-loop system. Both have advantages over
conventional servo systems. Figures 12-14 through 12-16 show some block
diagram examples of stepper motor uses in these two modes.
The open-loop stepper motor system is quite unique, and it has no
exact equivalent in the conventional servo system. The stepper motor is a
digital device whose output in shaft angular displacement is completely
determined by the number of input pulses. The motor can thus position
an object accurately on an input command in the form of a number of
pulses. However, the motor must be used in its response range, i.e., where
it can start, stop and reverse on command. An important advantage of the
open-loop system is that any error is noncumulative. The elimination of
208 OTHER ELECTRO-MECHANICAL DEVICES

Output

Computer Logic Stepper Geor Resolver

Driver Motor Troin

FIG. 12-14. Digital-to-analog-Conversion

Computer Logic Stepper Lead

Driver Motor Screw

FIG. 12-15. Machine Tool Positioner

CT

FIG. 12-16. Analog-to-digital Shaft Angle Converter

the feedback loop and any special circuitry used for damping results in
quite a savings in design time and installation.
In the closed-loop or positional feedback system, the motor is used
like a conventional servo motor. In such a feedback system, a signal from
the output is fed back and is used to operate a gate controlling the pulses
from a pulse generator. For most applications, the motor must be
operated within its response range, the same as with the open-loop system.
However, when velocity feedback or a signal similar to velocity feedback
is used, the motor can be started in its response range and from there
accelerated into its slew range, and amazingly high performance can then
be obtained.

Note: Much of the material in Sections 12.6 through 12.11 was excerpted by
permission from a brochure of Muirhead Instruments Ltd., Stratford, Ontario
Canada.
 THIRTEEN

Microprocessors, computers, and

servomechanisms

13.1. GENERAL

One of the first servomechanisms to be used commercially was the

flyball governor invented by James Watt in 1788 to control the speed of a
steam engine automatically. One of the first computers to be used was the
abacus invented by the Chinese and, to some extent, still used by them.
From the advent of the abacus and the flyball governor to the modern
servo and electronic computer, much has happened. The advancement
of the electronic art from tubes to transistors and integrated circuits has
provided the control industry with equipment and measuring instrumen¬
tation giving increased precision, reliability and economy. Computers,
both analog and especially digital, are used more and more to achieve the
ultimate in performance in systems where many variables are measured
and controlled simultaneously.
Even though we think of computers as being all electronic, and we
know that their central processors are, modern computer systems also
incorporate a large amount of electromechanical instrumentation. The
interfaces for instrumentation such as printers, tape drives, computer
peripherals, and control transducers require very precise control devices
which can control with high precision the incremental motion required.
Some of these transducers may be synchros, resolvers, LVDT’s, and
different types of servo motors, especially the stepper motor. Figure 13-1
is a typical block diagram of a dc velocity servo system being commanded
by binary words on two parallel lines with each of the words denoting
increment size and direction respectively. This system has the usual servo

209
210 MICROPROCESSORS, COMPUTERS, AND SERVOMECHANISMS

Digital Control Servomechanism

I___J i_I

FIG. 13-1. Digital Control of a dc Servo System

components plus a digital and velocity transducer and a digital up/down

counter with its associated circuitry.
The typical sequence of events that takes place after a single increment
input is as follows: The single word input is entered into an up/down
counter whose output is then decoded and fed into a nonlinear digital to
analog converter. From the converter, the signal is fed to a dc velocity or
rate servo whose permanent magnet motor then accelerates, driving its
load and the position encoder. The position encoder then feeds into the
gate and the up/down counter until the number counts down to its pre¬
determined value. When this value is reached, the signal to the amplifier
decreases, and the motor decelerates. The velocity tachometer (tachometer-
generator) is also driven by the dc motor, and it feeds a signal proportional
to speed to the servo summing junction for stabilization purposes.

13.2. THE MINICOMPUTER

The use of digital electronics in control has probably grown the

fastest in the process and machine tool industry chiefly because of the
advent of the minicomputer. The development of the minicomputer really
came about because one company, Digital Equipment Corporation, did
not believe that, for a computer, bigger and faster was necessarily better.
Instead, they came out with a computer that could be sold to those in the
industrial field who could not afford the high cost of the larger computer.
Minicomputers now give a large number of users enormous computing
power at relatively low cost. As the prices drop, it becomes more and
more economically feasible to use them in greater varieties of applications.
13.3. The microprocessor 211

Some people believe that there will be a computer in every home within
a decade.
The typical minicomputer is a small (about the size of a portable
television set), general purpose, digital computer that is capable of pro¬
viding a minimum of 4k (4096 words) bits of programmable memory with
associated power supply, teletypewriter, programmer console and related
hardware; it sells for about $20,000. The initial application of a mini¬
computer was in the engineering and scientific field where it supplemented
the technician in the control of experiments and tests and in instantaneous
data reduction and interpretation. Analytical instrumentation and research
in the fields of gas chromatography, crystallography, geophysics, wind
tunnel testing, cardiology, spectroscopy, etc., are ideal places for the
minicomputer. However, since research and development is somewhat
stable, this field became somewhat saturated; therefore, the minicomputer
moved into the industrial control field.
In the industrial control field, minicomputers have had wide accept¬
ance in industries where (1) labor costs are a large factor in the total cost
picture, (2) close tolerances of products are required, and (3) many
complex variables are involved. In these areas, the minicomputer out¬
shines the large process computers mainly because it is being used as a
local-loop controller feeding data into a larger data-processing computer.
Minicomputers are being used extensively in the small, discrete parts
manufacturing industries where production control, storage, and expedit¬
ing and shipping problems are becoming more acute. Other areas of use in
industrial control are automotive testing and checking, automatic inven¬
tory control, electrical component testing, production-line control and
scheduling, numerically controlled machine tools, and data acquisition
and reduction. One of the largest areas of use will be in the data com¬
munications field where, for example, messages are switched between
terminals or points, or where an exchange of information will take place
between remote terminals and a central process computer. The one thing
that has probably made the minicomputer so much in demand is the
decreasing cost of its heart, i.e., central processor unit (CPU), even though
the peripheral equipment used to make up a complete system has not
declined in price very much.

13.3. THE MICROPROCESSOR

The microprocessor, which moved out of the laboratories in 1973,

will begin to take its place in the digital electronics field in the future and
may well be the most important advance since the transistor was developed.
The microprocessor may be likened to a minicomputer with a different
212 MICROPROCESSORS, COMPUTERS, AND SERVOMECHANISMS

CPU unit fabricated on one tiny piece of silicon that gives all the advan¬
tages of digital computing techniques at an order of magnitude lower in
price. The heart of the microprocessor is its central processing unit that is
made up of large scale integration (LSI) circuitry and that has the ability to
process both arithmetic and logic data in a program control mode. It has
filled a gap in the industrial control field for those low-cost applications
where the relatively large and costly software requirements of the mini¬
computer are not necessary. Thus, all the advantages of a computer are
realized through microprocessors without the attendant disadvantage of
using or developing such software. Programs must be developed and
debugged beforehand, although they are easier to do because of the
microprocessor’s four-bit design and ease with which it can handle BCD
arithmetic. Thus, the general application area for a microprocessor lies
between the minicomputer and those devices which use LSI circuitry such
as the four-function, hand-held calculator. Because of its small size, low

FIG. 13-2. Traffic Control through a Minicomputer (Courtesy of Chilton Tech¬

nological Seminars, Radnor, Pa.)
13.3. The microprocessor 213

FIG. 13-3. A Microprocessor (Courtesy of Pro-Log Corporation, Monterey, Calif.)

power consumption, and reliability, the microprocessor will be ideally

suited for use in appliances such as electric stoves, automatic washers and
dryers, and refrigerators as well as in automobiles, cash registers, etc.
One very large field, that is now 75% controlled through electro¬
mechanical relays and controllers, is traffic control. In this field, the
microprocessor will probably act as the CPU for a more complex com¬
puter. Figure 13-2 shows a block diagram of such a computer controller.
This system can be easily changed to fit special traffic conditions or laws
by simply using different software because all that is really needed here is
counting and simple bit manipulation. As traffic moves more efficiently
through intersections, less energy is wasted with attendant reduction in
pollution.
In reality, the microprocessor consists of control and arithmetic, and
it is the heart of a digital subsystem known as a microprogrammable
processor or microcomputer. A microprocessor is pictured in figure 13-3.
Many semiconductor manufacturers now offer the microprocessor as a
standard digital logic building block which will allow the system designer
to design the required control system. In many of these control appli¬
cations, a piece of hardware performs a certain job over and over on
demand with little data acquisition. Every job is usually unique which
means no canned program. It may be for a high volume market or for a
214 MICROPROCESSORS, COMPUTERS, AND SERVOMECHANISMS

FIG. 13-4. Control System Showing a Microprocessor as a Logic Element with

Different Loads (Courtesy of Pro-Log Corporation, Monterey, Calif.)

one-of-a-kind nature. Two characteristics are inherent in control of this

nature, i.e., (1) when the power is turned on, the system does the job, and
(2) all hardware of the system is necessary to do that job.
Figure 13-4 shows a microprocessor system used as a logic element to
supply different types of loads. Here, it is treated as a black box with a
number of input gates and output latching drivers. These 1/0’s are wired
to loads which might be switch contacts (relay contacts, keys, thumbwheel
switches, push buttons, relay coils, lamps, motors, etc.), and which are
then wired in a matrix configuration where one axis is driven by the
outputs, and the other axis is sensed by the inputs. Functions and infor¬
mation of these loads are all given in the program. The microprocessor
can thus perform timing sequences, make decisions, convert codes,
analyze and linearize curves, do arithmetic, etc. When used in this manner,
the microprocessor eliminates component wiring (which other systems
require), and it instead has specific coding put into the Program Memory
 <
Machine Motions, Linear or Rotary,
Which are to be Controlled, Measured,

m
u
o
w
O
4-
C

o E
<D
j
/

J
1
FIG. 13-5. The A-B-C-D's of Industrial/Computer Control (Courtesy of Astrosystems, Inc., Lake Success, N. Y.)

215
216 MICROPROCESSORS, COMPUTERS, AND SERVOMECHANISMS

so that the first characteristic above is satisfied whenever power is turned

on. Programmed read only memory (PROM) and read only memory
(ROM) chips are coded and used as the Program Memory.

13.4. CONCLUSION

Microprocessors and microcomputers will establish themselves more

and more in the electronic and industrial control fields, and they may play
a larger part in the control of a servo through digital means. There is no
question that whole new, specialized application areas will be opened up
in the consumer, communication and industrial sectors provided that the
support and software for these applications can be married. Figure 13-5
illustrates the A-B-C-D’s of such a marriage.
 FOURTEEN

Laboratory experiments

14.1. GENERAL

This chapter is made up of a number of laboratory experiments

designed to give the student or reader a thorough understanding of
elementary servo principles using one of the modular servo trainers that
are illustrated and discussed in Chapter 11. Each experiment will take
about two hours to perform and another two to three hours to write up,
should that be required. The experiments should be done in the order
listed as each succeeding experiment is more complex and depends on
information learned from the previous one.

14-2. Title: DC Servo Motor Characteristics

Object: To investigate and to take measurements of dc servo motor
characteristics, to calibrate a dc tachometer-generator, and to record
the steady-state and transient response of the motor
Apparatus: Modular servo system trainer, VTVM, dc oscilloscope or
recorder
Procedure:
I. Motor operation
1. Mount the servo amplifier, the power supply, and the motor on
the steel baseboard. Plug the motor into the servo amplifier and
the servo amplifier into the power supply. See figure 14-2.
2. Connect one of the potentiometers in the attenuator unit across
the dc supply with the slider being connected to input terminal

217
218 LABORATORY EXPERIMENTS

Com.

(b) Field Control

FIG. 14-1. Motor Connections

No. 1 of the servo amplifier. This will provide a variable input

to the motor through the amplifier.
3. Connect the motor so that armature control of the motor will be
obtained. See figure 14-l(a).
4. Set the potentiometer to give zero volts. Turn the power supply
on and the meter should read about 0.5 ampere. This is due to
basic circuits in the servo amplifier, and it does not affect the
motor connections. Turn the potentiometer up, and, when the
current has increased to about one ampere, the motor should
start with the speed being controlled by the voltage from the
potentiometer. When voltage is applied to the other servo
amplifier input jack, the motor will reverse its direction of
rotation.

r ~!
flower
r Servo + Supply
Attenuator 1 Amplifier
Com. 1
Unit -■2
- |
I -1
2 3
4 “i~ r II
±1
LL
Motor &
Tach.
1 Com. _I
T

FIG. 14-2. Motor Experiment

LABORATORY EXPERIMENTS 219

60 = K-rVc (1 -e-T,TM )

60 = Velocity
Vc = Volts

Kt = Speed/Volts Applied

rM = Motor Time Constant

FIG. 14-3. Motor Characteristics

5. Record the voltage at which the motor just starts to rotate, for
both directions of rotation, by measuring the voltage at the input
to the servo amplifier.
II. Tachometer-generator calibration—Kc factor

1. The motor speed may be computed from the tachometer-

generator voltage once the factor KG regulating speed and emf
has been found. To determine this factor (which is measured in
volts/1000 rpm), the motor must be driven at a known speed and
the tachometer-generator output must be measured. Determine
Kg by using a strobotach.

III. Measurement of motor characteristics

1. (a) Armature Connection-Speed vs. Voltage—Increase the

potentiometer voltage until the motor just starts, record this,
and then record output of the tachometer as indicated on the
voltmeter for six settings of the potentiometer between the
starting point and 10. Measure input voltage at the servo
amplifier for each of these points. See figure 14-3(a).
220 LABORATORY EXPERIMENTS

To obtain the speed-torque characteristic, locate the magnet

brake assembly correctly and then set the brake to maximum
(10), and increase the input voltage to maximum. Leave the
input voltage constant and swing the brake back taking
readings of tachometer voltage for five settings of the brake
between 2 and 10. See figure 14-3(b).

2. Field Connection—Speed vs. Voltage—Arrange the field con¬

nection (see figure 14-1,b); then swing the brake magnet clear.
Increase input voltage slowly until the motor just starts and
record potentiometer setting. Increase the potentiometer setting
until the tachometer output is 5V and note this setting. If
possible, repeat for a tachometer output of 9V. Again see
figure 14-3(a). Set the brake to maximum and adjust the input to
give maximum current. Keep the input constant and swing the
brake back taking readings of tachometer voltage for five settings
of the brake between 2 and 10. Again see figure 14-3(b).

IV. Transient response of motor

1. Connect the motor for armature control with no disc on the

output shaft. Using a servo analyzer or function generator,
apply a square wave of about 6-8V amplitude at a frequency of
0.15 Hz to one servo amplifier input jack (the potentiometer
voltage input may be disconnected or reduced to its lowest
value). The motor will now respond to the positive half wave.

2. The change in the speed increase of the motor may be examined

by connecting the tachometer-generator output to a dc oscillo¬
scope or recorder. An exponential rise will be observed. Note
the difference in response time for the following cases:
1. no brake disc
2. brake disc only
3. brake disc under load
4. inertia disc

See figure 14-3(c).

3. Connect the tachometer-generator output to one input of the

two channel recorder and make a trace of the speed charac¬
teristic with the brake disc mounted.

4. If time permits, mount the inertia disc on the low speed output
shaft and note effect of the gear train between the motor and
the load.
LABORATORY EXPERIMENTS 221

RESULTS

1. Make a simple block diagram for the motor connection shown in the
Procedure, part Ill, step 1.
.
2 Determine the factor KG and then plot speed vs. input volts for both
connections of part III. Also plot speed vs. brake scale for both sets
of readings.
3. From the recorded trace of the transient response, determine the
motor time constant (it should be about 0.25 s).
.
4 It may be said that the armature connection requires more driving
power. However, it is easier to control whereas the field connection is
more sensitive but less desirable in a servo system in relation to the
system stability. Which set of curves shows this to be true? Why?
5. What effect does the inertia load have on motor response? Why?
6. Is the system shown in Procedure, part III (Results 1) an open-loop
or closed-loop system? Why?
.
7 Which method of control, armature or field, will have the higher gain?
Why?

14-3. Title: DC Servo Motor Speed Control System

Object: To investigate a simple speed control system with its associated

operational amplifier and preamplifier, to determine tachometer-
generator polarity, and to ascertain the effect of forward path gain on

performance

Apparatus: Modular servo system trainer, VTVM, dc oscilloscope,

recorder
Procedure:

I. Properties of an operational amplifier

1. Connect the ± dc supplies and set the control switch to the left
position. Measure the output which ideally should be zero, and
adjust the zero switch, if necessary. Connect the two potenti¬
ometers in the attenuator unit across the dc supplies so that two
independent variable voltages, VI and V2, are available at the
sliders. Connect the sliders to two of the inputs as is shown in
figure 14-4(a) and set each slider to 2.5V so that 5V output is
obtained. The output of the amplifier can now be checked. It
should sum the inputs, but with a reversed sign. Do this in three
steps for the polarities as shown.
1. Vx POS. V2 POS.
222
FIG. 14-4. Operational Amplifier Circuits
LABORATORY EXPERIMENTS 223

2. Yj POS. V2 NEG.
3. Vj NEG. V2 NEG.
2. Set the control switch to “EXT. FDBK.” Connect the gain
control potentiometer in the attenuator unit to the amplifier
output with the slider being connected back to one of the input
jacks as is shown in figure 14-4(b). Apply 2V to one input jack
and measure the output to check that the gain is inversely pro¬
portional to the output potentiometer setting, i.e., for 50%
setting, the output should be 4V for a gain of 2. Do this in four
steps for the following settings.
1. 25% setting
2. 50% setting
3. 75% setting
4. 90% setting
II. Simple speed control system

1. Connect the attenuator unit, operational amplifier, servo

amplifier, power supply, and motor/tachometer-generator units
as is shown in figure 14-5. Arrange the motor for armature
control and swing the brake clear. Set the gain of the oper¬
ational amplifier to unity by setting the gain potentiometer to
100%. DO NOT connect in the tachometer-generator. Check
to see that if the reference potentiometer is turned up, the motor
will rotate. Set the operational amplifier switch to normal
(CCW).
224 LABORATORY EXPERIMENTS

2. In order for the system to operate properly, the tachometer-

generator polarity must be checked. The output of the tach¬
ometer must oppose the reference voltage at the input. Ground
one side of the tachometer and then connect the other side
to an operational amplifier input. If the speed decreases,
then negative feedback is obtained; if the speed increases to a
very high value, then positive feedback is obtained. Finally,
connect the tachometer for negative feedback.

III. Effect of forward gain on system performance

1. As the forward gain is increased, the speed drop for a given load
should decrease. This is so because the power to the amplifier
increases with gain.

2. Be sure the tachometer-generator is connected for negative

feedback. Adjust the reference voltage to give a speed of 1000
rpm with the operational amplifier gain set to 1.0 and the brake
swung clear.

3. Swing the brake forward and then measure (1) motor input
voltage at the servo amplifier, and (2) motor speed (via tachom¬
eter voltage). Do not adjust the reference potentiometer while
taking these readings.

4. Repeat part III, step 3 for operational amplifier gains of 5 and 10,
each time readjusting the reference voltage to give a no-load
speed of 1000 rpm. When motor current approaches a value of
2A, the speed may fall quite rapidly as load is increased. This
shows that the maximum power that can be supplied depends on
system capacity and not on feedback.

IV. Reversible speed control

1. Insert the preamplifier unit instead of the operational amplifier

in the circuit of figure 14-6 and connect the reference potenti¬
ometer across the ± dc supplies as is shown. Arrange the motor
for armature connection and swing the brake clear. Connect
the reference voltage and tachometer-generator output to the
preamplifier. Be sure the tachometer is connected to give
negative feedback.

2. Adjust the speed to 1000 rpm at no load. Be sure that the motor
is running in the same direction as it previously was.

3. Swing the brake forward and perform the same test as in part III,
step 3. Check that the drop in speed is much less because the
amplifier gain is about 25 rather than 10 (see figure 14-7).
LABORATORY EXPERIMENTS 225

+ 15 V

1
Ref
-1 Pre Servo
i
2
Amp Amp

-15 V

FIG. 14-6. Reversible Speed Control

Speed

FIG. 14-7. Speed Control Characteristics

RESULTS

1. Make a simple block diagram for the motor connection of the

Procedure, part I.
2. Plot speed versus brake setting for the load tests for gains of 1, 10 and
25. On a separate graph sheet, plot motor input volts versus brake
setting for the same load tests.
3. Why was it so important to determine the correct polarity of the
tachometer-generator for system operation?
226 LABORATORY EXPERIMENTS

4. The curves show that as the forward gain of the operational amplifier
is increased the amount of error required to drive the system decreases.
Why? Also prove by formula.

5. What is the main purpose of the operational amplifier?

6. What two benefits resulted in the system operation when the pre¬
amplifier was used?
7. Show by formula and calculation that the results of Procedure, part I,
steps 1 and 2 came out as expected.

14-4. Title: DC Servo Speed and Position Control System

Object: To investigate the behavior of a speed control system under

various operating conditions, and to set up the error channel that is
associated with a mechanical position control system, closing the
loop, and general performance characteristics such as deadband

Apparatus: Modular servo system trainer, dc oscilloscope, VTVM,

recorder (if needed)

Procedure :

I. Investigation—motor speed control behavior

1. Connect the attenuator unit, operational amplifier, preamplifier,

servo amplifier, the input potentiometer, power supply, and
motor/tachometer-generator unit as is shown in figure 14-8.
Arrange for armature control. Set the gain potentiometer in the
attenuator unit to zero, the input potentiometer to zero, and the
velocity potentiometer to 10. Switch the power “on” and adjust
the preamplifier zero to get a minimum reading on the ammeter.

2. Set the operational amplifier switch to normal (CCW) and then

set the gain potentiometer to 10. Remove the lead from the
velocity potentiometer to socket 3 on the operational amplifier
and adjust the operational amplifier zero to get a minimum
reading on the ammeter. Reconnect the lead to socket 3 and
adjust the input potentiometer until the motor stops. The input
potentiometer should be at zero.

3. Set the gain and tachometer potentiometers to 10. Using a servo

analyzer, apply a square wave of ±3V at 0.2 Hz at socket 2 of
the operational amplifier. Display the velocity signal (output
of the tachometer-generator) and the error signal (output of the
operational amplifier) on a dual-channel scope. The display
should be as shown in figure 14-9(a); also note the ammeter
reading.
I

E
(D
•*->
(/)
>»
(/)
o
+-»
k_

c
o
o
>
‘o
_o
<X>
>
00

<3

li-

227
228 LABORATORY EXPERIMENTS

High
i
Gain

r _ Error

Med.
Gain L \

Low
Gain V

Added
Time
Delay

FIG. 14-9. Velocity Control Signals

4. Reduce the gain to 5 and repeat step 3. Reduce gain to 1 and

repeat step 3.
5. Introduce an additional time delay in the forward path by setting
the switch on the operational amplifier to the center position.
Repeat the display of the tachometer and error with the gain
potentiometer at 10. See figure 14-9(b).
6. Set the selector switch on the operational amplifier to the normal
position, tachometer potentiometer at 0 and gain potentiometer
at 10. Fit the brake assembly and adjust it until the tachometer
display is the same size as was seen in step 3. Note the ammeter
readings.
II. Determining deadband

1. Disconnect the square wave from the operational amplifier so

that the motor is controlled from the input potentiometer only.
2. Measure deadband by finding the total angle through which
the motor just starts to rotate in either direction. Start with the
gain at 1.0.

note: The deadband width can be determined ahead of time.

In section 14-2, Procedure, part III, the minimum voltage to
start the motor rotating was found (approximately 5V). The
LABORATORY EXPERIMENTS 229

l
Error

KE(a,-e0)

Input-1 1-
Output From Motor

FIG. 14-10. Error Channel Connections

preamplifier gain is about 25; therefore, the minimum voltage

at the input to the preamplifier should be about +0.2V. With
gain at 1.0, then about +2V will be required at the output of the
operational amplifier. If the position error constant is about
0.1 volt/deg, then ±20° are required to give +2V for a total
deadband width of about 40°. The deadband should be equal
around the zero position; if it is not, adjust the preamplifier zero
until it is.
3. Repeat this deadband test for gains of 2, 5, 8 and 10.
III. The error channel and closing the loop

1. Disconnect the lead from the tachometer potentiometer to the

operational amplifier. Also disconnect the input to the pre¬
amplifier. Set the gain potentiometer at 0.
2. Mount the output potentiometer on the motor shaft extension
and then connect to the ± dc supply as is shown in figure 14-10.
Connect the slider to an input of the operational amplifier as is
shown.
3. Set the control switch of the operational amplifier to give noimal
addition. Measure the voltages between each potentiometer
slider and common, and turn the shafts until zero voltage is
obtained. Both dials should be at or about zero. If they are not,
then zero them.
4. The output of the operational amplifier should now be zero.
Rotate the input shaft and check that this output depends on
shaft displacement. Determine the position error constant,
Kpe; it should be about 0.1 volt/deg, i.e., a shaft displacement
of 10° should give IV error.
5. Reconnect the lead from the operational amplifier to the pre¬
amplifier. Introduce a small displacement between the input
230 LABORATORY EXPERIMENTS

shaft and the output shaft, say 30°. Now increase gain slowly.
If the output shaft moves toward the input, the connections are
correct. If the output shaft moves in the reverse direction or
settles (oscillates) at 180°, then reverse either the leads to the
servo amplifier and/or the + connections to the output poten¬
tiometer. The system should now align.

RESULTS

1. Make a simple block diagram for the motor connection of the speed
control system.
2. Explain why the deadband decreased as the gain was increased.
3. In setting up the error channel, Procedure, part III, if the input and
output potentiometers are connected in reverse to the dc supplies, the
output potentiometer will lock 180° out of phase to the input
potentiometer. Why is this true?
4. How did the measured position error constant, KPE, compare to the
predicted value? Show how this is found?
5. In the speed control system, the displays indicated that the transient
error was the same for all values of gain, but that the steady-state
error increased as the gain was decreased. Why is this true?
6. When the additional time delay was introduced, the speed response
showed that the motor now oscillated as well as the error. Why is
this true?
7. Explain how the error is found in the speed control system and where
it is displayed.

14-5. Title: Testing of a DC Servo Position-Control System

Object: To investigate the reduction of overshoot and settling time

through velocity feedback of a mechanical position-control system,
the effect of added inertia, and the stabilization of the control system

Apparatus: Modular servo system trainer, dc oscilloscope with four-trace

plug-in, and recorder
Procedure :

I. Reduction of overshoot and settling time—velocity feedback

1. The complete mechanical position control system with the

command and output potentiometers mounted and connected
to the dc supplies and the operational amplifier is shown in
figure 14-11. The output of the operational amplifier which is
 Position Feedbock

I
I

b
FIG. 14-11. Position Control System

231
232 LABORATORY EXPERIMENTS

the difference between these two, i.e., the error, is connected to

a potentiometer in the attenuator unit and to the preamplifier.
The output from the tachometer-generator is also fed to the
preamplifier through a potentiometer.
2. As can be seen, this system contains two feedback loops—the
outer loop through the error potentiometer and the inner loop
through the tachometer-generator. In order for the system to
operate properly, the polarities of both loops must be considered
separately. Set both the error and velocity potentiometers to
zero and then proceed in the following manner:
Error Potentiometer—The correct polarity was determined in
section 14-4, part III, under the section on Closing the Loop;
therefore, refer to this section.
Velocity Potentiometer—With this potentiometer at zero, start
the motor rotating by turning the preamplifier zero adjustment.
Increase the velocity potentiometer setting, and, if the polarity
of the tachometer-generator is correct, the speed should decrease.
If it does not, then reverse the connections to the tachometer-
generator. Rezero the preamplifier.
3. To determine the effect of velocity feedback, set the gain of the
error potentiometer to 10 and then turn the command shaft
sharply through about 45° for several different settings of the
velocity potentiometer. Observe the results.
4. Using a servo analyzer, apply a square wave of about ±3V at
0.2 Hz to the input of the operational amplifier. Display the
signal of the output potentiometer and the error on a storage
scope, channels 1 and 2. Examine and record the position
response and error signal for various velocity potentiometer
settings.
II. Effect of added inertia and/or time delay

1. Mount the inertia disc on the motor shaft and repeat the test of
part I, step 4.
2. Mount the inertia disc on the shaft opposite the output poten¬
tiometer and repeat the test of part II, section 1. Note carefully
any differences in the form of the responses.
3. Remove the inertia disc and switch in the additional time delay
in the operational amplifier. Investigate the response for zero
damping, underdamping, critical damping, and overdamping.
III. Stabilization

1. Set the switch on the operational amplifier to the normal

position.
LABORATORY EXPERIMENTS 233

2. Connect the four-trace plug-in of the storage scope to the four

“Y” signal points. Be sure the trigger switch is “out” so that the
time base will internally trigger on channel 1. Set the CHOP-
ALT switch to CHOP.
3. Set the gain potentiometer at 10 and then investigate the response
for zero damping, underdamping, critical damping, and over¬
damping. Record these displays. Note that the velocity feedback
signal is in opposition to the error signal. Also note that these
signals will cancel shortly before the error goes through zero.
4. Change the disturbance to a sine wave at 0.1 Hz. Record the
variation in amplitude and phase shift between the signal and
the position signal as the frequency is slowly increased to 5 Hz.

IV. Field control

1. Disconnect the disturbance signal to the operational amplifier.
Change the motor wiring to give field control.
2. Investigate the behavior of the system for small angular inputs to
the input potentiometer for various levels of gain and velocity
feedback.

RESULTS

1. Make a block diagram for this complete system showing all polarities.
2. Most servo motors are built with a relatively long armature and small
diameter. Did any of the results of this experiment show why
this is true?
3. What is meant by the term “velocity feedback,” and why does it
stabilize a system so effectively.
4. Was there any difference in the operation of the system when the
* additional time delay element of the operational amplifier was
switched “in” or “out”? What is the difference due to?
5 What was the difference in the position signal when a sinusoidal
* command was used rather than a step-position command, especially
as the command frequency was increased to a value of 3.0 to 5.0 Hz?
6. Is the following error, /.<?., angular lag between input and output,
affected by changes in level of tachometer feedback?
7. What are the advantages and disadvantages of using field control for
the motor?
8 In Procedure, part III, step 3, a statement was made that velocity and
error canceled shortly before the error went through zero. Was this
true and why?
234 LABORATORY EXPERIMENTS

/-Commond
/

CX / CT

FIG. 14-12. Synchro Link

14-6. Title: A DC Position Control Servo Using

AC Error Detectors
Object: To study and to investigate the operation of continuous rotation
transducers in the error channel and the conversion of an amplitude-
modulated ac error signal to a dc error signal through a demodulator
Apparatus: Modular servo system trainer, synchro units, demodulator,
dc oscilloscope or recorder
Procedure:
I. Synchro link
1. Mount the synchros and the power supply. Connect the
synchros in cyclic connection. Connect R1 and R2 of the
synchro transmitter to the ac output of the power supply.
Connect the —15V supply to the —15V terminal of the small
dc drive motor of the transmitter (see figure 14-12).
2. Set the dial on the transmitter to 0° and monitor the output at
R1 of the synchro transformer on an oscilloscope. Ground R2
of the CT. Note dial settings of the CT for both minimum and
maximum outputs.
3. Drive the transmitter rotor at various speeds by means of the dc
drive control and monitor the output of the CT. In order to
observe the change in phase each time the amplitude of the
signal goes through zero, it will be necessary to use a very low
sensitivity setting on the oscilloscope.
II. Investigation of demodulator action
1. Mount the demodulator unit on the baseboard. Connect the
error output of the synchro CT to input terminals No. 2 and
No. 4 of the demodulator.
LABORATORY EXPERIMENTS 235

2. Set the drive motor on the transmitter to the “off” position and
also the dial to 0°. Monitor the waveform at terminal 9 of the
demodulator and adjust the CT dial until the error voltage at R1
and R2 is zero. It may be necessary to use the zero control on
the demodulator to get this zero output at terminal 9.
3. Measure output voltage at R1 for different dial settings, i.e.,
0°, 30°, 45°, 60° and 90°.
4. Change the dial of the CT until maximum output is obtained at
Rl. Adjust the phase control on the demodulator until the
output at terminal No. 9 is one of the signals that are shown in
figure 14-12. Be sure that the filter switch is “off.”
5. Note the variation in the demodulator output as the motor
driving the transmitter is varied in speed. Note the variations
in output as the filter in the demodulator is switched in, i.e.,
filter switch to demodulator.

III. Closed-loop synchro error link

1. Connect the power supply, the servo amplifier, the preamplifier,

the attenuator unit, the modulator/demodulator, the motor, and
the synchros as is shown in figure 14-13.
2. Set the gain and tachometer potentiometers to 0. Energize the
system and adjust the zero control on the preamplifier so the
motor does not run.
Set the dial on the transmitter so that it is about 30° different
from the CT. Monitor the output at terminal No. 9 of the
demodulator. Set the filter switch to “off” and adjust the phase
control until the output shows a good full wave rectified ac.
Now reset the filter switch to demod.
3. Set the gain to 5 and note system response.
4. Switch the drive motor of the synchro transmitter to “on” and
adjust it until the transmitter rotates at about 1 rps. (This may
be determined by visually watching the 60 Hz track on the dial.)
Note system response for the following settings:
(a) Gain—5 Velocity—1 5 10
(b) Gain—1 5 10 Velocity—5
(c) Synchro transmitter—0 rps
Gain—5 Velocity—1 5 10
5. Mount the output potentiometer to the low speed motor shaft
and connect it to the + dc supply. Apply the disturbance signal
(square wave) between the external input and common on the
demodulator. Monitor the output from the wiper of the output
236
Command
>
>

FIG. 14-13. Position Control—Synchro Link

LABORATORY EXPERIMENTS 237

potentiometer. Investigate the response of the system for the

same settings of the gain and velocity potentiometers as in
part III, step 4.

RESULTS

1. Make a block diagram for this system.

2. In the setup used, what changes the ac signal to the dc signal that is
is needed to drive the servo motor?
3. From observations made while conducting the experiment, which
feedback stabilizes the system the best? Why? How does this
answer compare with the theoretical answer?
4. Does the combination of ac and dc have any adverse effect on the
system as compared with a system that is only dc?
5. What may have been a cause for the instability at the low end of the
frequency range (0.1 to 1.0 Hz)?
6. In the performance of this system in part III, was there any time where
the false null of the synchro would have an effect on the output?
Explain!

14-7. Title: A Relay-Controlled Position-Control System

Object: To study and to analyze the characteristics of a servo relay unit

and the effect that the relay output has on a mechanical position-

control system
Apparatus: Modular servo system trainer, simulated relay unit, and dc
oscilloscope
Procedure:

I. Simulated relay unit

1. Connect the simulated relay unit as is shown in figure 14-14 and

apply a triangle (ramp) wave of 1 or 2 Hz; observe and record
the characteristics for the following relays.
Ideal Set the width and hysteresis potentiometers
to zero and switches Sx to deadband and S2 to
hysteresis.
Deadband Adjust width potentiometer so that deadband
appears.
Hysteresis Set width potentiometer to zero and adjust
hysteresis potentiometer so that hysteresis
appears.
238 LABORATORY EXPERIMENTS

Output

For.

Ramp i Relay i
— —1 *■ Y
Input i-1 Input -e2 + e2
CRO
X
Rev.

Ideal Hysteresis

-e, + e1

Deadband or Deadband
Off Position and Hysteresis

FIG. 14-14. Characteristics of a Relay System

Hysteresis and Adjust both controls.

deadband
Limiting with Set S2 to slope. Adjust width and slope
adjustable slope control for both directions for normal dead¬
and deadband band and maximum slope.
Backlash Set Si to backlash. Adjust width control.
II. Relay-operated system—open loop

1. Connect the following modular units as is shown in figure 14-15:

attenuator, operational amplifier, preamplifier, servo amplifier,
power supply, relay, input and output potentiometers, and motor
with the motor arranged for armature control. Set gain, width
and hysteresis potentiometers to zero and disconnect the
feedback line from the output potentiometer. Adjust the input
potentiometer to mid scale.
2. Turn up the gain control until the motor starts to rotate. The
direction of rotation can now be controlled by CW or CCW
rotation of the input potentiometer with this reversal being very
sharp, indicating an ideal relay characteristic.
 LABORATORY EXPERIMENTS 239

Input

+ 15 V

1 Output
Op Relay
Amp.

T
- 15 V
+ 15 V
Gain
1
Output

- 15 V
FIG. 14-15. Connections of a Relay-controlled System

3. Set St and S2 to deadband and hysteresis respectively. Then for

various settings of the width and hysteresis potentiometers,
observe and record the effect on the output potentiometer as the
input potentiometer is rotated.
III. Relay-operated system—closed loop

1. Reconnect the feedback loop and adjust the controls until the
motor rotates. If the output potentiometer follows the input
potentiometer then the supply connections are correct. If the
output potentiometer does not follow the input potentiometer,
determine what must be done to get it to do so.
2. Note how the motor oscillates when the width and hysteresis
potentiometers are set at zero. Determine how this oscillation
may be eliminated and yet still have the system follow.
3. Apply a square wave of the same amplitude and frequency as the
ramp input of Procedure, part I and observe the effects of this
disturbance.
4. Reduce the amplitude of the square wave and observe the
changes, if any, that this makes.
5. To investigate deadband, remove the square wave and set
hysteresis to zero and width to about 25 %. Rotate the input
potentiometer and note response. Determine how the system
accuracy can be improved.
6. To investigate hysteresis, set deadband and hysteresis to zero.
Turn up the gain until the system oscillates. Determine how the
oscillation frequency can be reduced.
240 LABORATORY EXPERIMENTS

RESULTS

1. Make a block diagram for the closed-loop system indicating where

the feedback loop may be opened.
2. Why does the motor reverse direction so abruptly under open-loop
conditions when the input potentiometer is adjusted?
3. What characteristic was being exhibited by the system under open-
loop conditions for Procedure, part II, step 3?
.
4 One of the general problems of a relay system is that if oscillations
are eliminated, then deadband is increased and accuracy is reduced.
Did any of your results prove this? Why?
5. Was there any advantage to introducing hysteresis in the control of
this system? Why?
6. What are the differences between deadband and hysteresis? Do they
affect each other?
7. What effect does deadband have on system response and gain?
 APPENDIX A

SI (metric) units with English

and CGS equivalents

ENGLISH CGS
si
Unit Symbol Unit Symbol
Quan Unit Symbol
3.28 feet ft 100 centimeters cm
Length 1 meter m
39.37 inches in
2.2 pounds lb 103 grams gm
Mass 1 kilogram kg
second s second s
Time second s
10.76 square feet ft2 104 square cm2
Area 1 square meter m2
centimeters
3.28 feet per ft/s 100 centimeters cm/s
Linear 1 meter per m/s
second per second
velocity second dyne
N 0.225 pounds lb 105 dynes
Force 1 Newton dyne-cm
J 0.738 foot-pound ft-lb f 107 dyne-
Energy 1 joule
force centimeters
or work watt-second erg/s
W 107 erg per
12

Power 1 watt (joule W

second
O
1
X

per second)
horsepower HP
m/s2 3.28 feet per ft/s2 100 centimeters per cm/s2
Linear 1 meter per
second squared second squared
acceleration second squared 107 dyne dyne-cm
1 Newton Nm 0.738 pound-foot lb-ft
Torque
meter 141.9 ounce-inch oz-in centimeter
23.7 pound- lb-ft2 107 gram-square gm-cnr
Moment of 1 kilogram- kg-m2
square feet centimeter
inertia square meter C
coulomb C coulomb
Electric coulomb C
charge ampere A
A ampere A
Electric ampere
current volt V
V volt V
Electric volts
potential farad F
farad F
Capacitances farad henry H
H henry H
Inductance henry ohm n
a ohm fi
Resistance ohm

241
242 APPENDIX A

Other Useful Units

Angular
position rad rad angular deg
Angular
velocity rad/s rad/s 1 “/second
Angular
acceleration rad/s2 rad/s2 1 “/second2
Torque
constant Nm/A oz-in/A
Voltage
constant V/rpm volts/rpm
Damping
constant Nm/rpm oz-in/rpm
Viscous
damping Nm/rad s-1 oz-in/rad s 1
 APPENDIX B

Glossary of terms

Actuating Signal or Error (E)—Reference input (R) minus the primary

feedback (C).
Bandwidth—The range of frequencies of a device within which its
performance, with respect to some characteristic, conforms to a
specified standard.
Block_A part of a control loop that receives an input signal, processes
it, and transmits an output signal; a diagrammatic representation of
an element of a servomechanism or feedback control system.
Block Diagram—Simplified system reducing each element to its transfer
function and schematically representing it as a block.
Bode Diagram—Open-loop plot of phase angle 9 and amplitude ratio x dB
as a function of co; usually plotted on semilog paper with co in
radians per second, and phase angle in degrees.
Carrier Frequency—Constant frequency upon which useable data is
superimposed.
Closed Loop—System where the output is fed back to the input for
constant comparison; the complete signal path in a control system
that is represented as a group of units that are connected in sue a
manner that a signal started at any point follows a closed path and
can be traced back to that point.
Command (R>—The input that is established or varied by some means
external to and independent of the feedback control system under
consideration.
243
 244 APPENDIX B

Control Accuracy—Degree of correspondence between the controlled

variable and the ideal value.
Control Ratio—The frequency response of the controlled variable to the
reference input; under linear conditions, this ratio is expressed
mathematically as C/R = KG(1 + KGH).
Controlled Variable (C)—The quantity, or conditions, of the controlled
system that is directly measured and controlled.
Corner Frequency—Frequency where the open-loop plot of gain versus
frequency changes slope; the product of this frequency in radians per
second and time constant, t, in seconds equals units.
Critical Damping—The point at which system damping is great enough
to overcome all tendencies to oscillate or where the damping ratio
equals one.
Damping—A measure of the restraining force that prevents return to an
equilibrium position; the damping ratio is a measure of the amount
of damping. Critical damping permits the system to return to
equilibrium with no overshoots if there is a damping ratio of 1.0.
Overdamping permits a slow return to equilibrium with no overshoots
if the damping ratio is greater than 1.0. Underdamping permits a
rapid return to equilibrium with overshoots; this results in a settling
out period required to reach steady state equilibrium, and is present
with a damping ratio less than 1.0.
Damping Ratio (0—Ratio of actual damping to critical damping.
Dead Space or Deadband—Maximum deviation on either side of the point
of agreement between input and output for which no corrective
action will take place.

Dead Time or Time Delay A period of delay between two related actions
such as the beginning of a change in an input signal and the beginning
of a related change in the output.
Dynamic Analysis The study of control system performance by analyzing
the effect of disturbance inputs on the controlled variable or in
conditions that affect this variable.
Error (E)—Difference between output and input, or the difference between
an actual value and a desired or reference value; in a control loop,
this error is driven towards a desired minimum.
Error Detector The element or group of elements that convert the
difference between output and input of a system into usable form.
Error Rate Damping—Method of damping in which an additional signal
proportional to the rate of change of error is introduced and added
to the error signal for anticipatory purposes.
Glossary of terms 245

Forward Loop (Path)—In a feedback control loop, the transmission path

from the actuating signal to the output signal.
Frequency Response—Practical or mathematical observation of the ability
of the output to follow the input when the input is varied sinusoidally
over a given frequency range.
Gain Crossover—Point in the plot of loop ratio at which the magnitude of
the loop ratio is unity (Bode diagram).
Gain Crossover Frequency—Frequency at which the open-loop system
gain is unity or zero dB.
Gain Margin—The amount by which the magnitude of the loop ratio of
stable system is different from unity at phase crossover; it is frequently
expressed in decibels (Bode diagram).
Lead or Lag—An advance or delay of the output signal with respect to
the input.
Natural Frequency (coN)—The frequency at which a disturbed second-order
system or component would oscillate with zero or no damping.
Negative Feedback—Subtracting some or all of an output from the input
to achieve a desired effect; in amplifiers, used for gain stabilization.
Open Loop—Condition where efforts are made to make an output agree
with an input without a direct comparison.
Open-loop System—An open-loop system has no feedback, or it has the
feedback loop disconnected at the error summing point.
Phase Angle—A measure of the time by which an output lags or leads an
input; measured in degrees for sinusoidal signals.
Phase Crossover Frequency—Frequency at which the open-loop phase
shift is a full 180° (Bode diagram).
Phase Margin (0M)—The angle of difference between the phase of the loop
ratio of a stable system and 180° at gain crossover (Bode diagram).
Rate Generator—A signal which changes linearly with time.
Reference Input (R)—The reference used to compare the measured
variable to define a deviation of error signal; also referred to as set
point or desired value.
Response Time—Time required for the output of a system or element to
reach a specified value after the application of a step input or
disturbance.
Rise Time—The time required for the output of a system or element to
increase from one specified percentage of the final value to another
after the application of a step input; usually, the specified percentages
are 10%, and 90%.
246 APPENDIX B

Sensitivity—Maximum possible difference between output and input for

which no corrective action will take place.
Servoamplifier—Linear amplifier specifically designed to link error
detector voltage to motor control phase at a higher voltage and
power level than that supplied by the error detector.
Servomotor—1. Any motor used in a servomechanism to correct physically
differences between input and output. 2. Two-phase motor specifically
designed for servo applications and having low torque-to-inertia
ratio, sloping speed-to-torque curve, and torque proportional to
product of the voltage on each phase.
Servomechanism—A system in which the output is mechanically driven by
the difference between the input and the output for the purpose of
making the output agree with the input.
Settling Time—The time required for the absolute value of the difference
between the output of a system or element and its final value to
decline below and remain less than a specified amount after the
application of a step input or disturbance; the specified amount is
often expressed in terms of percent of the final value.
Stability—The property of a system or element that makes its response to
a stimulus die down if the stimulus is removed; a statement that a
system is stable means that the system is stable under all normal
operating conditions and for all types of stimuli normally encountered.
A system may be referred to as being stable in one region of operation
and not in another. If this is the case, the region of stability should
be specified.
Steady State—A stabilized condition in which the output has leveled out,
or has reached a constant rate of change, for a constant input; the
terminology is also applied to a condition in which the input is a
periodic constant amplitude signal.
Steady-state Error—Error that remains after the transient has expired.
Summing Point—A descriptive symbol used in block diagrams to denote
the algebraic summation of two or more signals; the direction of
information flow is indicated by arrows, and the algebraic process of
the summation by plus and minus signs.
Torque Constant—Constant of proportionality between motor stall torque
and control voltage.
Transfer Function—A mathematical relationship between an input signal
and a corresponding output signal; may be given in terms of LaPlace
transforms and as the ratio of output to input.
Transient Response—Output versus time in response to a step input.
Glossary of terms 247

Undamped Natural Frequency (coD)—System oscillatory frequency when

all damping is removed.
Velocity Lag Error—Lag between input and output that is proportional
to the rate at which the input is varying.
Viscous Damping—Utilization of reactive torque, or force, proportional
to speed for braking action.
Viscous Friction—Friction proportional to angular or linear speed.
 APPENDIX C

Transfer functions

Device Transfer Function Dimensions

Potentiometer KP = — V/deg or V/rad

0[N
Synchro Pair
K = — V/deg or V/rad
CX-CT
0.N
Linear Variable c
Differential V
JVj VHT
— Eo V/increment
DISPLACE
Transformer

Amplifier ka = —
EIN

Modulator AC volts/DC volts

Ein

Demodulator K — 1^°. DC volts/AC volts

° E
11 IN

I _ e0
Gear Train
N 0IN

V/1000 rpm or
Rate Generator Kcs = —
e,N V/rad/s

248
Transfer functions 249

Device Transfer Function Dimensions

do kt
AC Servomotor deg/V or rad/V
-IN s(l + STM)

DC Servomotor do _Kt/f_ deg/V or rad/V

Armature Control s[l + (J/F)s](l + sta)

DC Servomotor _Kt_
deg/V or rad/V
Field Control RfJs2(1 + stf)(1 + stm)

E0 l + T jS
Lag Network
E1N 1 + t2s
E0 _ 1 + TjS
Lead Network
E,n 1 + T2S
AC Servomotor d0 _ Kv(l + t2s)
deg/V or rad/V
Inertial Damped E,n (1 + TiSXl + T3s)
AC Servomotor dp = Ky
deg/V or rad/V
Viscous Damped E,n s(l + tms)

Eo_ = K
Magnetic Amplifier Volts/Volt
Ein (1 + TS)
 APPENDIX D

Universal transient response

curves

Universal Transient Response Curves

250
 APPENDIX E

References

Ahrendt, W. R., Servomechanism Practice. New York: McGraw-Hill Book

Company, 1954.
Baeck, Henry, Practical Servomechanisms Design. New York: McGraw-Hill
Book Company, 1968.
Bateson, Robert, Introduction to Control System Technology. Columbus, Ohio:
Charles E. Merrill Publishing Company, 1973.
Brite, Robert, and Carlo Fioranelli, Synchros and Servos. Indianapolis, Ind.:
Howard Sams, 1967.
Bukstein, E., Basic Servomechanisms. New York: Holt, Rinehart & Winston,
1963.
Bulliet, L. J., Servomechanisms. Reading, Mass.: Addison Wesley Publishing
Company, 1967.
Chestnut, Harold, and Robert Mayer, Servomechanisms and Regulating System
Design. New York: John Wiley & Sons, 1951.
DC Motors, Speed Controls, Servo Systems. Hopkins, Minn.: Electrocraft
Corporation, 1973.
DeRoy, Benjamin E., Automatic Control Theory. New York: John Wiley & Sons,
1966.
Hagen, Frank, Servo Engineers Handbook. Worcester, Pa.: Transicoil Inc.,
1968.
Herceg, E., Handbook of Measurement and Control. Pennsauken, N. J.: Schaevitz
Engineering, 1972.
Humphrey, William, Introduction to Servo System Design. Englewood Cliffs,
N.J.: Prentice-Hall, Inc., 1973.

251
252 APPENDIX E

Johnson, Eric R., Servomechanisms. Englewood Cliffs, N.J.: Prentice-Hall, Inc.,

1963.

Kearfott, Technical Information for the Engineer—Number 1. Little Falls, N.J.:

Singer-General Precision, Inc., 1969.

Philco Technological Center, Servomechanisms, Fundamentals and Experiments.

Englewood Cliffs, N.J.: Prentice-Hall, Inc., 1964.

The Singer Company, AC Instrumentation Application Notes. Los Angeles,

Calif.: Singer Instrumentation, 1973.

Weyrick, Robert, Fundamentals of Automatic Control. New York: McGraw-

Hill Book Company, 1975.

Zeines, B., Automatic Control Systems. Englewood Cliffs, N.J.: Prentice-Hall,

Inc., 1972.

-, Servomechanism Fundamentals. New York, McGraw-Hill Book

Company, 1959.
 APPENDIX F

Answers to selected problems

CHAPTER 2

1. (b) 130.9° (d) 1309°

3. 0° 360°
6. 0.0208%
7. 0.0476%
10. 5.357Y
11. (b) 4.5Y
20. 278°
22. Plastic— 1.54W

CHAPTER 3

(see Chapter 4 problem 1)

CHAPTER 4

1. (c) —2.6V (d) out-of-phase

2. (b) Altitude—4500 ft. Range—7790 ft.
3. (c) Slant range—12,650 ft. Elevation angle—71.6°
4. 6.2V 45°

253
254 APPENDIX F

5. (b) n — 36.4 mi. w — 20.96 mi.

7. (b) ER2—3.41V ER1 = 9.4V
8. (a) ES1—5.0V Es2 = -8.66V (c) Vector subtraction

CHAPTER 5

1. d
3. a
4. b
6. (c) St - S2 = 78V, S2 - S = 0, S3 - Sx = 78V
9. a
10. (a) 120°
11. -120° (240°)
13. 300°
15. 100°
17. a

CHAPTER 6

1. (a) E0 = -- Et
R, + R2

2. (a) E0 = E, ^2 — E0 §2
Ki Kx

C _ A + B
' R 1 + HA

4. R = 6 C = 450

7 c __G!G2(G3 + G4)_
’ R G!G2(G3 + G4) + H(G3 + G4) + 1

9. - = 56.3
R

C = GtG2
11
’ R 1 + HlGl + H2GjG2
Answers to selected problems 255

CHAPTER 7

1. 377 rad/s
2. 2800 rpm
4. 2 rpm
6. 0.833 sec.
9. 24,000 rad/s2
.
10 1500 rpm
14. 52.4%

CHAPTER 8

1. V = 70V S — 875 rpm T = 1225 oz-in

V = 30V S = 375 rpm T = 525 oz-in
2. 89
4. (a) 0.01 oz.in/rad/s (b) 200 sec

0o = 4.06
' E _ s(l + 18.8s)
7. (a) 10V
8. 1250 rpm

10. *2 = 1020
E s(38 + s)

CHAPTER 9

3. (a) 11.92oz.in/° (b) 18.92 oz.in-sec/rad.

4. 4.86 Hz
6. (a) 6.66° (b) 5.3° (c) 0.0833 sec (d) 0.108 sec
8. (a) 14.2°

CHAPTER 10

7. (a) — 34.5dB, -175°

9. (a) 11.5dB and -194°, -29dB and -242°, -37dBand -249°
(b) coGC = 3.87 cope — 0
(c) 0PM = -25.5° dGM = 50dB +
 APPENDIX F
256

U' (a) R Jco[l + J(co/8)]

(d) coBP = 8
(e) coGC = 19 coPC = oo
(f) 0PM = 22°
(g) -ldB
(h) 21dB or K = 11.22
Index

ac control system, 183 corner frequency, 158, 160

acceleration lag, 104 critically damped, 132
accuracy, 104, 106
actual damping, 138
actuating signal, 4
damping, 106, 116, 137
amplidyne, 7
damping ratio, 1 36
angular acceleration, 97
damping torque, 134
angular position, 97
damping waveforms, 132
angular velocity, 97
dead band, 104, 123, 179
automatic control, 1
decade,158
decible, 156
backlash, 186 digital control, 199
bandwidth, 145 directly controlled variable, 4
block diagrams, 4, 83 double speed synchro systems, 71
directional arrows, 84 dynamic characteristic, 106
expanded,4
identities and manipulation, 86
reduction, 90
E-core transformer, 32
take off point, 84
electrical summing junction, 5
Bode Plot (see Frequency Response)
error, 4, 104, 1 33
breakpoint, 159, 160
channel, 178
detector, 5
C.E.M.F. damping, 142 measuring element, 5
characteristic equation, 144 signal, 5,178
closed loop control, 2, 6 voltage cancellation, 134
closed loop gain, 84 error rate damping, 144
codes (see Shaft Encoders)
commands (see Test Commands)
computers (see Microprocessors) feedback, 2
conformity, 17 control systems, 1
control element, 5 element, 5
control field winding, 115 factor, 84
controlled variable, 4 loop, 4
conversion factors, 103 signal, 4
257
 INDEX
258

forces, 96 inertial damping, 142

free speed, 115 inertial torque, 98, 104
frequency domain, 108 inherent damping, 139, 143
frequency response—Bode Plot, 153 instability, 180
156 instantaneous speed, 117
analysis, 153-62 instrument servo, 98
construction, 157
experimental method, 156
lag-lead (notch) network, 154
negative slope, 158
LaPlace transform, 156
typical, 160
limited phase lag network, 152
frequency response—Nyquist plot,
limited phase lead network, 153
150
linear variable differential trans¬
analysis, 15 1
former, 33
correction, 152-55
ac operation, 34
criterion, 152
block diagram, 35
measurements, 150
dc operation, 34
opening of a feedback loop, 150
V curve, 34
friction, 97
zeroing, 38
coulomb, 97
logic (see Shaft Encoders)
static, 97
log magnitude, 157
viscous, 97
friction torque, 98, 104
mathematical units, 103
mechanical position control, 6,
gain, 85
180-82
asymptote, 158
microprocessors, 211
crossover, 161
microsyn, 39
crossover frequency, 161
minicomputers, 210
curve, 159
modular, 168
margin, 161
moment of inertia, 96, 125
gear trains, 99
motors
block diagrams, 99
alternating current, 115
gear parameters, 99-102
application factors, 112
generator, rate of tachometer, 126,
direct current, 117
141,143,176
direct drive torque type, 120
application, 126
moving coil type, 120
construction, 126
permanent magnet type, 119
generator voltage regulator, 6
printed circuit type, 121
gun control servo, 7
transistor control, 118
wound field type, 118
free speed, 115
half power point, 145
fundamental equation, 113
hysteresis, 185
power output, 98, 114
speed torque curves, 114-16, 137,
138
idealized motor characteristics, 137 torque production, 113, 116
incremental motion control, 121 transfer function, 125
induction resolvers (see Resolvers) motor characteristics, 169
INDEX 259

natural frequency, 137 rotation, single/multi turn, 13

negative feedback, 84, 132 single ended connection, 26
null position transducer, 36 starting/running torques, 20
Nyquist plots (see Frequency tap location, 19
Response) trimmers, 18
wirewound construction, 12
zeroing, 28
on-off control, 2
open loop,2
open loop gain, 84 rate damping, 144
operational amplifier, 172 reference field winding, 1 15
operator, 86 reference input signal, 4
oscillation, 132 relay controlled system, 183
output element, 5 resolvers, 27
overall gain, 85, 156 composition, rectangular to polar
overdamped, 132 coordinates, 45
overshoot, 132, 179 data transmission-angular
addition/subtraction, 5 1
peak overshoot, 132 determination of phase voltage, 44
phase electrical zero, 44
angle, 156 generation of flux, 44
crossover frequency, 160 induced voltage, 42
curve, 159, 160 performance characteristics, 54
lag network, 152 resolution, polar to rectangular
margin, 161 coordinates, 45
shift, 158 rotation of coordinates axis, 46
position error, 135 zeroing, 54
positive feedback, 85, 133 response time, 132
power, 98 reversible speed control, 174
power servo, 98 rotary variable differential trans¬
precision potentiometers formers, 35
ac characteristics, 21 rotor inertia J, 116
block diagram, 15
conformity, 17
double ended connection, 26 servo
linearity, 17 characteristics, 104
load resistance/loading error, 15 dead band, 104
mathematical relations, 14 electrical circuit analogy, 105
multiplication circuit, 28 sensitivity, 104, 182
noise, 20 steady state response, 108
non linear functions, 23 torque gradient, 104
non wirewound construction, 26 transient response, 108
overtravel, electrical/mechanical, types, 155
21 servomechanism, 1
position feedback circuit, 27 settling time, 131
power rating, 20 shaft encoders
resistance tolerance, 19 absolute type, 194
resolution, 17 Binary Coded Decimal System, 197
 260 INDEX

Shaft encoders (contd.) Synchros (contd.)

brush type, 195 follow up, 64
contacting type, 195 torque, 65
Gray code, 197 voltage phase relations, 63
incremental, 194 wiring errors, 66
non-contacting type, 195 zeroing, 77
switching, brush/optical, 195 synchro geared systems (see Double
slope, 186 speed synchro systems)
slew speed, 206 system gain, 137
speed control, 171 system responses, 131
stability, 1 31
stabilization, 180
tachometer feedback, 143, 180
stall torque, 1 1 5
teachometer voltage, 172
static error, 135
take off point (see Block Diagram's)
static friction, 97
tension control servo, 6
steady state errors, 133-35
test commands
step wise control, 2
sinusoidal, 107
stepper motor
side position, 106
characteristics, 207
step velocity, 107
electronic driver, 205
three position, 1 83
holding torque, 205
time constant, 1 17
mechanical driver, 205
time domain, 108
pull in rate, 206
tolerance band, 132
response range, 206
torque constant, 124
running torque, 206
transducers, 2, 6, 32
slew range, 206
transfer function, 85, 125, 156
switching modes—permanent
transient error, 136
magnet type, 200
comparisons, 200-4 transient response, 170
two position, 185
sequence of polarities, 201
step positions, 201,202
switching voltage waveforms, 204
undamped, 133
types, 199
underdamped, 132
use in control systems, 207
summing junction, 5, 84
synchros variable reluctance transducers
basic systems, 59 E core transformer, 32
basic theory, 61 principle of operation, 33
block diagram, 71 winding configuration, 33
control transformer, 69 velocity constant, 124, 138
data, 59 velocity control system, 176
designation, 59 velocity damping, 141
differential, transmitter/receiver, velocity lag error, 1 35
67 viscous damping, 116, 139
electrical zero, 61 viscous friction constant F, 1 16
flux summary, 63 voltage gradient, 127
 Date Due
t
fl f* V
UUI i 91980
n! -AjO
^2jaar
oct ; *> 2G05
OC: ' > 2005
 TJ 214 M48
Miller, Richard W
ben/omechan s7's devices and 010101 000

0 1 63 01719
TRENT UNIVERS

TJ214 .M48
Miller, Richard W
Servomechanisms

DATE
ISSUED in' ^ *7/

&?6Q8<6