David E.

Johnson
John L. Hilbum
Johnny R. Johnson
 V

KVL

v—it

L-u *
 h
c

7-

,v

. ) - lo

Qv,
 n

dJl.

k /if

Digitized by the Internet Archive

in 2013

http://archive.org/details/basicelectriccirOOjoh_oOe
 BASIC
ELECTRIC CIRCUIT
ANALYSIS
 BASIC
ELECTRIC CIRCUIT
ANALYSIS

D.E. JOHNSON, J.L HILBURN, and J.R. JOHNSON

Department of Electrical Engineering

Louisiana State University

Prentice-Hall, Inc. Englewood Cliffs, New Jersey 07632

Library of Congress Cataloging in Publication Data

Johnson, David E
Basic electric circuit analysis.

Includes index.
1. Electric circuits. I. Hilburn, John L., 1938-

joint author. II. Johnson, Johnny Ray, joint

author. III. Title.
TK454.J56 621.319'2 77-24210
ISBN 0-13-060137-3

1978 by Prentice-Hall, Inc., Englewood Cliffs, N.J. 07632

All rights reserved. No part of this book

may be reproduced in any form or
by any means without permission in writing
from the publisher.

Printed in the United States of America

10 9 8 7 6

Prentice-Hall International, Inc., London

Prentice-Hall of Australia Pty. Limited, Sydney
Prentice-Hall of Canada, Ltd., Toronto
Prentice-Hall of India Private Limited, New Delhi
Prentice-Hall of Japan, Inc., Tokyo
Prentice-Hall of Southeast Asia Pte. Ltd., Singapore
Whitehall Books Limited, Wellington, New Zealand
To Frances, Meme, and Betty
 CONTENTS

PREFACE x/77

1 INTRODUCTION 7

1.1 Definitions and Units 7

1.2 Charge and Current 4

1.3 Voltage, Energy and Power 7

1.4 Passive and Active Elements 10

1 .5 Circuit Analysis 12

2 RESISTIVE CIRCUITS 75

2.1 Ohm's Law 15

2.2 Kirchhoff's Laws 19
2.3 Series Resistance and Voltage Division 25
2.4 Parallel Resistance and Current Division 30
2.5 Analysis Examples 35
2.6 Ammeters, Voltmeters and Ohmmeters* 40
2.7 Physical Resistors* 43

3 DEPENDENT SOURCES 49

3.1 Definitions 49
3.2 Circuits with Dependent Sources 50
3.3 Operational Amplifiers 52
3.4 Amplifier Circuits 54

vii
t/iii Contents

4 ANALYSIS METHODS 61

4.1 Nodal Analysis 62

4.2 An Example 65
4.3 Circuits Containing Voltage Sources 67
4.4 Circuits Containing Op Amps 71

4.5 Mesh Analysis 73

4.6 Circuits Containing Current Sources 76
4.7 Duality* 78

5 NETWORK THEOREMS 89

5.1 Linear Circuits 89

5.2 Superposition 93
5.3 Thevenin's and Norton's Theorems 99
5.4 Practical Sources 106
5.5 Maximum Power Transfer 7 77

6 INDEPENDENCE OF EQUATIONS* 117

6.1 Graph of a Network 777

6.2 Trees and Links 7 75
6.3 Independent Voltage Equations 727
6.4 Independent Current Equations 124
6.5 A Circuit Application 727

7 ENERGY-STORAGE ELEMENTS 752

7.1 Capacitors 132

7.2 Energy Storage in Capacitors 136
7.3 Series and Parallel Capacitors 138
7.4 Inductors 142
7.5 Energy Storage in Inductors 145
7.6 Series and Parallel Inductors 747
7.7 Practical Capacitors and Inductors* 750
7.8 Duality and Linearity 757
7.9 Singular Circuits* 753
 1

Contents jx

8 SIMPLE RC AND RL CIRCUITS 160

8.1 Source-free RC Circuit 160

8.2 Time Constants 164
8.3 Source-free RL Circuit 168
8.4 Response to a Constant Forcing Function 772
8.5 The General Case 176
8.6 A Short-cut Procedure 778
8.7 The Unit Step Function 782
8.8 The Step Response 187
8.9 Application of Superposition 193

9 SECOND-ORDER CIRCUITS 201

9.1 Circuits with Two Storage Elements 201

9.2 Second-Order Equations 203
9.3 The Natural Response 205
9.4 Types of Natural Frequencies 208
9.5 The Forced Response 272
9.6 Excitation at a Natural Frequency 215
9.7 The Complete Response 278
9.8 The Parallel RLC Circuit 222
9.9 The Series RLC Circuit 227
9.10 The Complete Response of the RLC Circuit 229
9.1 Alternate Methods for Obtaining
the Describing Equations* 232

10 SINUSOIDAL EXCITATION AND PHASORS 243

10.1 Properties of Sinusoids 244

10.2 An RL Circuit Example 248
10.3 An Alternate Method Using Complex Numbers 250
10.4 Complex Excitations 253
10.5 Phasors 256
10.6 Voltage-Current Relationships for Phasors 259
10.7 Impedance and Admittance 264
10.8 Kirchhoff's Laws and Impedance Combinations 268
10.9 Phasor Circuits 277
X Contents

11 AC STEADY-STATE ANALYSIS 280

11.1 Nodal Analysis 280

1 1 .2 Mesh Analysis 285
1 1 .3 Network Theorems 288
11.4 Phasor Diagrams 254

12 AC STEADY-STATE POWER 506

12.1 Average Power 307

12.2 Superposition and Power 313
12.3 RMS Values 377
12.4 Power Factor 319
12.5 Complex Power 323
12.6 Power Measurement 327

13 THREE-PHASE CIRCUITS 333

13.1 Single- Phase, Three-Wire Systems 334

13.2 Three- Phase Y-Y Systems 338
13.3 The Delta Connection 345
13.4 Y-A Transformations 343
13.5 Power Measurement 353

14 COMPLEX FREQUENCY AND NETWORK FUNCTIONS 360

14.1 The Damped Sinusoid 360

14.2 Complex Frequency and Generalized Phasors 363
14.3 Impedance and Admittance 366
14.4 Network Functions 370
14.5 Poles and Zeros 373
14.6 The Natural Response from the Network Function 375
14.7 Natural Frequencies* 378
14.8 Two- Port Networks 380
 Contents xi

15 FREQUENCY RESPONSE 389

15.1 Amplitude and Phase Responses 389

15.2 Filters 392
15.3 Resonance 395
15.4 Bandpass Functions and Quality Factor 397
15.5 Use of Pole-Zero Plots 401
15.6 Scaling the Network Function 404
15.7 The Decibel 408

16 TRANSFORMERS 416

16.1 Mutual Inductance 477

1 6.2 Energy Storage 425
16.3 Circuits with Linear Transformers 428
16.4 Reflected Impedance 432
16.5 The Ideal Transformer 434
16.6 Equivalent Circuits 440

17 FOURIER METHODS 446

17.1 The Trigonometric Fourier Series 446

17.2 Symmetry Properties 452
17.3 The Exponential Fourier Series 455
17.4 Response to Periodic Excitations 459
17.5 Frequency Spectra 461
17.6 The Fourier Transform 453
17.7 Fourier Transform Operations 466

18 LAPLACE TRANSFORMS 473

18.1 Definition 474

18.2 Some Special Results 476
18.3 The Inverse Transform 450
18.4 The Impulse Function 484
18.5 Applications to Differential Equations 487
18.6 The Transformed Circuit 492
xii Contents

APPENDIX A DETERMINANTS AND CRAMER'S RULE 499

APPENDIX B GAUSSIAN ELIMINATION 504

APPENDIX C COMPLEX NUMBERS 507

APPENDIX D EULER'S FORMULA 513

INDEX 575
 PREFACE

This book was written for a one-year course in linear circuit analysis in the sophomore
year. Such a course is basic in electrical engineering and is usually the first encounter
of the student with his or her chosen field of specialization. It is imperative, therefore,
for the textbook used to cover thoroughly the fundamentals of the subject and at the
same time be as easy to understand as it is possible to make it. These have been our
objectives throughout the writing of the book.
Most students, when they take this subject, will have studied electricity and mag-
netism in a physics course. This background is helpful, of course, but is not a prereq-
uisite for reading the book. The material presented here may be easily understood
by a student who has had a basic course in differential and integral calculus. The differ-
ential equations theory required in circuit analysis is fully developed in the book and
integrated with the appropriate circuit theory topics. Even determinants, Gaussian
elimination, and complex number theory are presented in appendices.
The operational amplifier is introduced immediately after the discussion of the
resistor, and appears, as a matter of course, along with resistors, capacitors, and
inductors, as a basic element throughout the book. Likewise, dependent sources and
their construction using operational amplifiers are discussed early and are encountered
routinely in almost every chapter.
To aid the reader in understanding the textual material, examples are liberally
supplied and numerous exercises, with answers, are given at the end of virtually every
section. Problems, some more difficult and some less difficult than the exercises, are
also given at the end of every chapter. A been made to include a
special effort has
number of problems and exercises with realistic element values. Of course, network
scaling, which is also presented, can be used to make almost all of the remaining
problems practical. In particular, in the chapter on amplitude and phase responses,
problems are given that relate to electric filters, which, of course, are very useful

xiii
xiv Preface

circuits. Active filters, using operational amplifiers, as well as passive filters, are used
as examples. Finally, a select few of the exercises and problems are used to extend the
theory discussed in the chapters. In this way, optional material is included without

adding to the text of the chapter.

The first nine chapters of the book are devoted to terminology and time-domain
analysis and the last nine chapters deal with frequency-domain analysis. Some sections
and chapters, identified by asterisks, may be omitted without any loss in continuity.
Among these is the chapter on network topology, an interesting subject which could
be covered entirely, in part, or not at all. Also those on Fourier methods and Laplace
transforms are often reserved for a succeeding course in linear systems, but their
if there is time to cover them.
essential ingredients are here
There are many people who have provided invaluable assistance and advice con-
cerning this book. We are indebted to our colleagues and our students for the form the
book has taken, and to Mrs. Dana Brown and Mrs. Norma Duffy, who provided the
expert typing and draftsmanship, respectively. A special note of thanks is due Profes-
sors M. Van Valkenburg, A. P. Sage, S. R. Laxpati, and S. K. Mitra, who reviewed
E.
the manuscript and made many helpful comments and suggestions.

Louisiana State University David E. Johnson

Baton Rouge, La. JoHN L H ilburn
Johnny R. Johnson
 BASIC
ELECTRIC CIRCUIT
ANALYSIS
 1

INTRODUCTION

Electric circuit analysis, in nearly every electrical engineering curriculum, is the first

course taken in the major area by an electrical engineering student. Virtually all

branches of electrical engineering, such as electronics, power systems, communication

systems, rotating machinery, and control theory, are based on circuit theory. The only
topic in electrical engineering more basic than circuits is electromagnetic field theory,
and even there many problems are solved by means of equivalent electric circuits. Thus
it is no exaggeration to say that the basic circuit theory course a student first encounters

in electrical engineering is the most important course in his or her curriculum.

To begin our study of electric circuits we need to know what an electric circuit is,
what we mean by its analysis, what quantities are associated with it. in what units
these quantities are measured, and the basic definitions and conventions used in
circuit theory. These are the topics we shall consider in this chapter.

1.1 DEFINITIONS AND UNITS

An electric circuit, or electric network, is a collection of electrical elements intercon-

nected in some specified way. Later we shall define the electrical elements in a formal
manner, but for the present we shall be content to represent a general two-terminal
element as shown in Fig. 1.1. The terminals a and b are accessible for connections with
other elements. Examples with which we are all familiar, and which we shall formally
consider in later sections, are resistors, inductors, capacitors, batteries, generators, etc.

More complicated circuit elements may have more than two terminals. Transistors
and operational amplifiers are common examples. Also a number of simple elements
may be combined by interconnecting their terminals to form a single package having
any number of accessible terminals. We shall consider some multiterminal elements
later, but our main concern will be simple two-terminal devices.
Introduction Chap. 1

An example of an electric circuit with six elements is shown in Fig. 1.2. Some
authors distinguish a circuit from a network by requiring a circuit to contain at least
one closed path such as path abca. We shall use the terms interchangeably, but we may
note that without at least one closed path the circuit is of little or no practical
interest.

FIGURE 1.1 General two-terminal FIGURE 1 .2 Electric circuit

electrical element

To be more specific in defining a circuit element we shall need to consider certain

quantities associated with it, such as voltage and current. These quantities and others,
when they must be carefully defined. This can be done only if we have a standard
arise,

system of units so that when a quantity is described by measuring it we can all agree

on what the measurement means. Fortunately, there is such a standard system of units
that is used today by virtually all the professional engineering societies and the
authors of most modern engineering textbooks. This system, which we shall use
throughout the book, is the International System of Units (abbreviated SI), adopted in
I960 by the General Conference on Weights and Measures.
There are six basic units in the SI, and all other units are derived from them. Four
of the basic units, the meter, kilogram, second, and coulomb, are important to circuit
theorists, and we shall consider them in some detail. The remaining two basic units are
the degree Kelvin and the candela, which are important to such people as the electron
device physicist and the illumination engineer.
The SI units are very precisely defined in terms of permanent and reproducible
quantities. However, the definitions are highly esoteric and in some cases are com-
prehensible only to atomic scientists. 1
Therefore we shall be content to name the basic
units and relate them to the very familiar British System of Units, which includes
inches, feet, pounds, etc.
The basic unit of length in the SI is the meter, abbreviated m, which is related to the
British system by the fact that 1 inch is 0.0254 m. The basic unit of mass is the kilogram

'Complete definitions of the basic units may be found in a number of sources, such as, for example,
"IEEE Recommended Practice for Units in Published Scientific and Technical Work," by C. H. Page
et al. {IEEE Spectrum, vol. 3, no. 3, pp. 169-173, March 1966).
Chap. 1 Introduction 3

(kg),and the basic unit of time is the second (s). In terms of the British units, pound- 1

mass is exactly 0.45359237 kg. and the second is the same in both systems.
The fourth unit in the SI is the coulomb (C). which is the basic unit used to measure
electric charge. We shall defer the definition of this unit until the next section when we
consider charge and current. The name coulomb was chosen to honor the French
scientist, inventor, and army engineer Charles Augustin de Coulomb (1736-1806), who
was an early pioneer in the fields of friction, electricity, and magnetism.
We might note at this point that all SI units named for famous people have
abbreviations that are capitalized. Otherwise, lowercase abbreviations are most often
used. It is also worth mentioning that we could choose units other than the ones we
have selected to form the basic units. For example, instead of the coulomb we could
take the ampere (A), the unit of electric current to be considered later. In this case the
coulomb could then be obtained as a derived unit.
There are three derived units in addition to the ampere that we shall find useful in
circuit theory. They are the units used to measure force, work or energy, and power.
The fundamental unit of force is the newton (N). which is the force required to accel-
erate a 1-kg mass by 1 meter per second per second (1 m/s 2 ). Thus 1 N = 1 kg-m/s 2 .

The newton is named, of course, for the great English scientist, astronomer, and
mathematician Sir Isaac Newton (1642-1727). Newton's accomplishments are too
numerous to be listed in a mere chapter.
The fundamental unit of work or energy is the joule (J), named for the British
physicist James P. Joule (1818-1889). who shared in the discovery of the law of
conservation of energy and helped establish the idea that heat is a form of energy. A
joule is the work done by a constant 1-N force applied through a 1-m distance. Thus
1 J = 1 N-m.

The last derived unit we shall consider is the watt (W). which is the fundamental
unit of power, the rate at which work is done or energy is expended. The watt is
defined to be J/s and is named in honor of James Watt (1736-1819), the Scottish
1

engineer whose engine design first made steam power practicable.

Before we leave the subject of units we should point out that one of the greatest
advantages the SI has over the British system is its incorporation of the decimal
system to relate larger and smaller units to the basic unit. The various powers of 10 are
denoted by standard prefixes, some of which are given, along with their abbreviations,
in Table 1.1.

TABLE 1 .1 Prefixes in the SI

Multiple Prefix Symbol

10 9 Giga G
10 6 Mega M
10 3 Kilo k
10-3 Milli m
irr 6 Micro M
10- 9 Nano n
10-12 Pico P
 Introduction Chap. 1

As an example, at one time a second was thought to be a short time, and fractions
such as 0.1 or 0.01 of a second were unimaginably short. Nowadays in some applica-
tions, such as digital computers, the second is an impracticably large unit. As a
result, times such as 1 nanosecond (1 ns or 10" 9 s) are in common use. Another com-
mon example is 1 gram (g) = 10" 3 kg.

EXERCISES
1.1.1 Find the number of millimeters in 10 km. Am. 10 7

1.1.2 If a mile equals 5280 ft, how many miles in 10 km? Ans. 6.2137

1.1.3 Find the work in millijoules done by a constant force of 25 /zN applied to a
mass of 4 g for a distance of 10 m. Ans. 0.25

1.2 CHARGE AND CURRENT

We are all familiar with gravitational forces of attraction between bodies, which are
responsible for holding us on the earth and which cause an apple dislodged from a tree
to fall ground rather than to soar upward into the sky. There are bodies, how-
to the
ever, that attract each other by forces far out of proportion to their masses. Also, such
forces are observed to be repulsive as well as attractive and are clearly not gravita-
tional forces.
We explain these forces by saying that they are electrical in nature and caused by
the presence of electrical charges. We explain the existence of forces of both attraction
and repulsion by postulating that there are two kinds of charges, positive and negative,
and that unlike charges attract and like charges repel.
As we know, according to modern theory, matter is made up of atoms, which are
composed of a number of fundamental particles. The most important of these
particles are protons (positive charges) and neutrons (neutral, with no charge) found
in the nucleus of the atom and electrons (negative charges) moving in orbit about the

nucleus. Normally the atom is electrically neutral, the negative charge of the electrons
balancing the positive charge of the protons. Particles may become positively charged
by losing electrons to other particles and become negatively charged by gaining
electrons from other particles.
As an example, we may produce a negative charge on a balloon by rubbing it
against our hair. The balloon will then stick to a wall or the ceiling, which are
uncharged. Relative to the negatively charged balloon the neutral wall and ceiling are
oppositely charged.
We now define the coulomb (C), discussed in the previous section, by stating that
10" 19
the charge of an electron is a negative one of 1.6021 x coulombs. Putting it

another way, a coulomb is the charge of about 6.24 x 10 18 electrons. These are, of
 £Ou\o^? * \^( *

Chap. J Introduction

course, mind-boggling numbers, but their sizes enable us to use more manageable
numbers, such as 2 C, in the circuit theory to follow.
The symbol for charge will be taken as Q or q, the capital letter usually denoting
constant charges such as Q=
4 C, and the lowercase letter indicating a time-varying
charge. In the latter case we may emphasize the time dependency by writing q(t). This
practice involving capital and lowercase letters will be carried over to the other
electrical quantities as well.
The primary purpose of an electric circuit is to move or transfer charges along
specified paths. This motion of charges constitutes an electric current, denoted by the
letters i or /, taken from the French word "intensite." Formally, current is the time

rate of change of charge, given by

(L1)
/ My>= ifrnf* '=J
The basic unit of current ampere (A), named for Andre Marie Ampere (1 775—
is the
1836). a French mathematician and physicist who formulated laws of electromagnetics
in the 1820s. An ampere is 1 coulomb per second.

In circuit theory current is generally thought of as the movement of positive

charges. This convention stems from Benjamin Franklin (1706-1790), who guessed
that electricity traveled from positive to negative. We now know that in metal con-
ductors the current movement of electrons that have been pulled loose from the
is the
orbits of the atoms of the metal. Thus we should distinguish conventional current (the
movement of positive charges), which is used in electric network theory, and electron
current. Unless otherwise stated, our concern will be with conventional current.
As an example, suppose the current in the wire of Fig. 1.3(a) is / = 3 A. That is,

3 C/s pass some specific point in the wire. This is symbolized by the arrow labeled 3 A,
whose direction indicates that the motion is from left to right. This situation is

equivalent to that depicted by Fig. 1.3(b), which indicates —3 C/s or —3 A in the

direction from right to left.

Figure 1 .4 i flowing from the left

represents a general circuit element with a current
toward the right terminal. The total charge entering the element between time t and t
is found by integrating (1.1). The i^sult_is__

9t = ?(0 - lOo) (1.2)

T-Jt
We should note at this point tTTaT"W€^axe-<orf5id€ring the network elements to
be electrically neutral. That is. no net positive or negative charge can accumulate in

the element. A positive charge entering must be accompanied by an equal positive

(a) (b)

FIGURE 1.3 Two representations of the same FIG U R E 1 .4 Current flowing in a gen-
current eral element
Introduction Chap. 1

charge leaving (or, equivalently, an equal negative charge entering). Thus the current
shown entering the left terminal in Fig. 1.4 must leave the right terminal.
There are several types of current in common use, some of which are shown in
Fig. 1.5. A constant current, as shown in Fig. 1.5(a), will be termed a direct current, or
dc. An alternating current, or ac, is a sinusoidal current, such as that of Fig. 1.5(b).
Figures 1.5(c) and (d) illustrate, respectively, an exponential current and a sawtooth
current.

DC

(a)

(c)

FIGURE 1.5 (a) dc; (b) ac; (c) Exponential current; (d) Sawtooth
current

There are many commercial uses for dc, such as in flashlights and in power sup-
plies for electronic circuits, and, of course, ac is the common household current found
all over the world. Exponential currents appear quite often (whether we want them or
not!) when a switch is actuated to close a path in an energized circuit. Sawtooth waves
are useful in equipment, such as oscilloscopes, used for displaying electrical charac-
teristics on a screen.

EXERCISES
1.2.1 Find the charge in picocoulombs represented by 10,000 electrons.
Ans. 0.0016021

1.2.2 The total charge entering a terminal of an element is given by

q(t) = 10/ 2 - 2tC

Find /(/) at t = and at t = 1 s. Ans. -2 A, 18 A

Chap. 1 Introduction 7

1.2.3 The current entering a terminal is given by

f(0 = 30/ :
-4/ A

Find the total charge entering the terminal between / = 1 s and / = 3 s.

Ans. 244 C

1.3 VOLTAGE, ENERGY, AND POWER

Charges in a conductor, exemplified by free electrons, may move in a random manner.

However, we want some concerted motion on their part, such as is the case with an
if

electric current,we must apply an external or so-called electromotive forc e (EMF).

Thus work is done on the charges. We shall define voltage "across" an element as the
work done in moving a unit charge (+1 C) through the element from one terminal to
the other. The unit of voltage, or potential difference, as it is sometimes called, is the
volt (V), named in honor of the Italian physicist Alessandro Guiseppe Antonio
Anastasio Volta (1745-1827).
Since voltageisjhe number of joules of work performed on coulomb, we may 1

say thatfl V is 1 J/C. Jhus the volt is a derived SI unit, expressible in terms of other
units. J—
We shall represent a voltage by v or Kand use the +, — polarity convention shown
in Fig. 1.6. That is. terminal A is v volts positive with respect to terminal B. Putting it

another way in terms of potential difference, terminal A is at a potential of v volts

higher than terminal B. In terms of work, it is clear that moving a unit charge from B
to A requires v joules of work.
Some authors prefer to describe the voltage across an element in terms of voltage
drops and rises. Referring to Fig. 1.6, moving from
a voltage drop of v volts occurs in
A to B. In contrast, a voltage rise off moving from B4e A.
volts occurs in
As examples. Figs. 1.7(a) and (b) are two versions of exactly the same voltage. In
(a), terminal A is +5 V above terminal B, and in (b), terminal B is —5 V above A
(or -f-5 V below A).
In transferring charge through an element work is being done, as we have said. Or,

putting it another way, energy is being supplied. To know whether energy is being

5V -5V

A O- -O B

+ v - (a) (b)

FIGURE 1 .6 Voltage polarity convention FIGURE 1 .7 Two equivalent voltage representa-

tions
 Introduction Chap. 1

supplied to the element or by the element to the rest of the circuit, we must know not
only the polarity of the voltage across the element but also the direction of the current
through the element.If a positive current enters the positive terminal, then an external
force must be driving the current and is thus supplying or delivering energy to the
element. The element is absorbing energy in this case. If, on the other hand, a positive
current leaves the positive terminal (enters the negative terminal), then the element is

delivering energy to the external circuit.

As examples, in Fig. 1.8(a) the element is absorbing energy. A positive current
enters the positive terminal. This is exactly the case also in Fig. 1.8(b). In Figs. 1.8(c)
and (d) a positive current enters the negative terminal, and therefore the element is

delivering energy in both cases.

2A 2A 2A 2A
+ o- + 0-

5V 5V 5V 5V

7 7
(b) (c) (d)

Figure 1 .8 Various v-i relationships

Let us consider now the rate at which energy is being delivered to or by a circuit
element. If the voltage across the element is v and a small charge Aq is moved through
?, . ( <

the element from the positive to the negative terminal, then the energy absorbed
by the element, say Aw, is given by
Aw = v Aq

If the time involved is At, then the rate at which the work is being done, or the energy
w is being expended, is given by

lim -r—
Aw = ,.„
lim v
.. Aq
-j^-
Af-0A? Ar^O A/

or

dw dq
(1.3)
dt dt v

Since by definition the rate at which energy is expended is power, denoted by p, we

have
dw
y (1.4)
dt

We might observe that (1.4) is dimensionally correct since the units of vi are (J/C)(C/s)
or J/s, which is watts (W), defined earlier.
i are generally functions of time, p given by (1.4) is a time-varying
Since v and
quantity. sometimes called the instantaneous power because its value is the power
It is

at the instant of time at which v and i are measured.

 VJori
C/?3D. / i
,1 Introduction Q

Summarizing, the typical element of Fig. 1.9 is absorbing power, given by p = vi.

If either the polarity of v or i (but not both) is reversed, then the element is delivering
power, p vi, to the external circuit. Of course, to say that an element delivers a
negative power, say — 10 W, is equivalent to saying that it absorbs a positive power, in
this case +10 W.

FIGURE 1.9 Typical element with

voltage and current

As examples, in Figs. 1.8(a) and (b) the element is absorbing power of

p = (5)(2) = 10W. [In Fig. 1.8(b) the 2 A leave the negative terminal, and thus
2 A enter the positive terminal.] In Figs. 1.8(c) and (d) it is delivering 10 W to the
external circuit, since the 2 A leave the positive terminal, or, equivalently, —2 A
enter the positive terminal.
Before ending our discussion of power and energy, let us solve (1.4) for the energy
w delivered to an element between time t and /. We have, upon integrating both sides
between t and /.

w(t) — w(t ) = f vi dt (1.5)

•Ito

For example, the energy delivered to the element of Fig. 1.8(a) between / = and
/= 2 s is given by
2
w(2) - vr(0) = f (5)(2) dt = 20 J
Jo

Since the left member of (1.5) represents the energy delivered to the element
between t and t, we may interpret w(t) as the energy delivered to the element between
the beginning of time and t and w(t ) as the energy between the beginning of time and
t . In the beginning of time, which let us say is / = -co, the energy delivered to the
element was zero; that is,

U'(-oo) =
If t = — <x> in (1.5), then we shall have the energy delivered to the element from the
beginning up to t, given by
t

w(t)=\ vidt (1.6)

This is consistent with (1.5) since

w(t) = j
vi dt

= vi dt + vi dt
J-OO J to
1 Introduction Chap. 1

By (1.6) this may be written

w(t) = w(t ) + I" vi dt

>to

which is (1.5).

EXERCISES
1.3.1 The element shown is
is absorbing
a
^
power of p = 20 W. Find the current entering
terminal b. Ans. —4 A

5V

bo-

EXERCISE 1.3.1

1.3.2 Find the energy delivered to the element of Ex. 1.3.1 between and 2 s.

Ans. 40 J

1.4 PASSIVE AND ACTIVE ELEMENTS

We may classify circuit elements intotwo broad categories, passive elements and
active elements, by considering the energy delivered to or by them.
A circuit element is said to be passive if the total energy delivered to it from the
rest of the circuit is always nonnegative. That is, referring to (1.6), for all t we have

wit) = j" p(t) dt = vi dt>0 (1.7)

J"

The polarities of v and i are as shown in Fig. 1.9. As we shall see later, examples of
passive elements are resistors, capacitors, and inductors.
An active element is one that is not passive, of course. That is, (1.7) does not hold
for all time. Examples of active elements are generators, batteries, and electronic
devices that require power supplies.
We are not ready at this stage to begin a formal discussion of the various passive
elements. This will be done in later chapters. In this section we shall give a brief

discussion of two very important active elements, the independent voltage source and
the independent current source.
 .

Chap. 7 Introduction 11

An independent voltage source is a two-terminal element, such as a battery or a

generator, that maintains a specified voltage between its terminals. The voltage is

completely independent of the current through the element. The symbol for a voltage
source having v volts across its terminals is shown in Fig. 1.10. The polarity is as
shown, indicating that terminal a is v volts above terminal b. Thus if v > 0, then
terminal a is at a higher potential than terminal b. The opposite is true, of course, if

v< 0.

In Fig. l.lO, the voltage v may be time varying, or it may be constant, in which
case we would probably label it V. Another symbol that is often used for a constant
voltage source, such as a battery with V volts across
terminals, is shown in Fig. its 1 . 1 1

In the case of constant sources we and 1.11 interchangeably.

shall use Figs. 1.10
We might observe at this point that the polarity marks on Fig. 1.11 are redundant
since the polarity could be defined by the positions of the longer and shorter lines. We
shall leave the polarity marks off in most cases in the future. There are times, however,
in analyzing circuits when it is convenient to use the polarity marks.
oa p a

6
6b 6b

FIGURE 1.10 Independent voltage FIGURE 1.11 Constant voltage source

source

An independent current source is a two-terminal element through which a specified

current flows. The current is completely independent of the voltage across the element.
The symbol for an independent current source is shown in Fig. 1.12, where i is the
specified current. The direction of the current is indicated by the arrow.
Independent sources are usually meant to deliver power to the external circuit and
not to absorb it. Thus if v is the voltage across the source and its current i is directed
out of the positive terminal, then the source is delivering power, given by p = vi, to
the external circuit. Otherwise it is absorbing power. For example, in Fig. 1.13(a) the

2A 2A

+ +
12 V 12 V
~

o o
(0) (b)

FIGURE 1.1 2 Independent current FIG U R E 1 .1 3 (a) A Source delivering and (b) Absorb-
source ing power
 —

12 Introduction Chap. 1

battery is delivering 24 W to the external circuit. In Fig. 1 . 1 3(b) the battery is absorbing
24 W, would be the case when it is being charged.
as
The sources that we have discussed here, as well as the circuit elements to be
considered later, are ideal elements. That is, they are mathematical models that approxi-
mate the actual or physical elements only under certain conditions. For example, an
ideal automobile battery supplies a constant 12 V, no matter what external circuit is
connected to it. Since its current is completely arbitrary, it could theoretically deliver
an infinite amount of power. This, of course, is not possible in the case of an actual
device. A real 12-V automobile battery supplies approximately constant voltage only
as long as the current it delivers is low. When the current exceeds a few hundred
amperes, the voltage drops appreciably from 12 V.
We shall consider practical sources in a later chapter and see under what conditions
they may be approximated by ideal sources. Also later, we shall consider dependent
sources whose voltage (or current) is controlled by another voltage or current some-
where else in the circuit.

EXERCISE
1.4.1 Find the power being supplied by the sources shown.
Ans. (a) 36 W, (b) -40 W, (c) -2 W
IA
t
'
o

-2V

(a) (b) (c)

EXERCISE 1.4.1

1.5 CIRCUIT ANALYSIS

Let us now look at the words circuit analysis, which are contained in the title of the
book, and see what they mean. Generally, if an electric circuit is subjected to an input
or excitation in the form of, say, a voltage or a current provided by an independent
source, then an output or response is produced. The output or response may also be a
some element in the circuit. There may be, of
voltage or a current associated with
course, more than one input and more than one output.
There are two main branches of circuit theory, and they are derived from the
following three key words: input, output, and circuit. The first branch is circuit
Chap. 1 Introduction 13

analysis, which, given the circuit and the input, is concerned with finding the output.
The other branch is circuit synthesis, which, given the input and output, is concerned
with finding the circuit itself.

Network synthesis is much more complex in general than analysis and will proba-
bly be encountered by the student in a later course. Circuit analysis will be our concern
in this book. We may be interested only in finding one or more outputs, such as a
voltage or current existing somewhere in the circuit, or in determining the energy or
power delivered one element or another. Or we may wish to perform a complete
to
analysis, finding every unknown current and voltage in the circuit. In any case, in the
chapters to come we shall develop systematic methods of analysis that can be applied
generally to any circuit of the type we consider.

PROBLEMS
1.1 Let f(t) in the graph shown be the charge q{t) in millicoulombs that has entered
the positive terminal of an element as a function of time, (a) Find the total charge
that has entered the element between 3 and 7 s, (b) the current entering the element
at 5 s, and (c) the energy absorbed by the element between 1 and 7 s if the voltage
across the element is 10 V.

f(t)

4 -

T (seconds)

PROBLEM 1.1

1.2 Let/(/) in Prob. 1.1 be the current i(t) in milliamperes entering an element ter-

minal as a function of time, (a)Find the charge that enters the terminal between
and 7 s and (b) the rate at which the charge is entering at t = 5 s.

1.3 The terminal voltage of a voltage source is given by v = 10 cos 100/ V. If the
charge leaving the positive terminal is q = 6 sin 100/ mC, find the power being
supplied by the source at t = 71/100 s.

1.4 For the element shown, the charge entering the positive terminal is

q = —2e' 05 '
C. Find the power delivered to the element as a function of time
and the total energy delivered to the element between and 2 s if (a) v = 6/',

(b) v = 3 dijdt, and (c) v = 2 i dt + 4. (Note v : is in volts if / is in amperes.)

14 Introduction Chap. 1

o-
+

PROBLEM 1.4

1.5 If the element in Prob. 1.4 is a battery with a constant terminal voltage of v = 12 V
and the current i is 2 A, then the battery is in the process of being charged. (It is

absorbing rather than delivering power.) (a) Find the energy supplied to the bat-
tery in 2 h (hours), (b) Find the charge delivered to the battery in this time. Note
the consistency of the units, 1 V = 1 J/C.

1.6 Suppose the voltage v in Prob. 1.5 varies linearly from 4 to 12 V during the 2 h
that the battery is being charged. If / = 2 A is constant during this period, find
(a) the total energy supplied and (b) the total charge delivered to the battery.

1.7 In the element of Prob. 1 .4, let i(t) = for t < and i(t) = 6 sin It A for t > 0.

If v = 4 dijdt V, show that the energy delivered to the element is nonnegative for
all t > 0.
1.8 In Prob. 1 .7, find the total charge delivered to the element and the power absorbed
at / = 7r/4 s.
 1-*'"7
2.
2
RESISTIVE CIRCUITS

The simplestand most commonly used circuit element is the resistor. All electrical
conductors exhibit properties which are characteristic of a resistor. When currents
flow in conductors, electrons which make up the current collide with the lattice of
atoms in the conductor. This, of course, on the average, impedes the motion of the
electrons. The larger the number of collisions, the greater the resistance of the con-
ductor. We shall consider a resistor to be any device which exhibits solely a resistance.
Materials which are commonly used in fabricating resistors include metallic alloys and
carbon compounds.
In this chapter, we shall first introduce the terminal relations for a resistor based on
Ohm's law. Two laws necessary for systematic solutions of networks, known as
Kirchhoff's laws, are then examined. With these laws, we shall begin our study of
by finding solutions for "single-loop" and "single-node-pair" resistive
circuit analysis
networks having independent sources as inputs. We shall conclude the chapter with a
discussion of simple measuring instruments followed by a discussion of practical
resistors.

2.1 OHM'S LAW

Georg Simon Ohm (1787-1854), a German physicist, is credited with formulating the
current-voltage relationship for a resistor based on experiments performed in 1826. In
1827 he published the results in a paper titled "The Galvanic Chain, Mathematically
Treated." As a result of this work, the unit of resistance is called the ohm. It is ironic,

however, that Henry Cavendish (1731-1810), a British chemist, discovered the same
results 46 years earlier. Had he not failed to publish his findings, the unit of resistance
might well be known as the caven.

15
16 Resistive Circuits Chap. 2

Ohm's law states that the voltage across a resistor is directly proportional to the
current flowing through the resistor. The constant of proportionality is the resistance
value of the resistor in ohms. The
symbol for the resistor is shown in Fig. 2.1.
circuit
For the current and voltage shown, Ohm's law is

m = m) (2.i)

where R >
is the resistance in ohms.

Rearranging (2.1) into the form R = v(t)/i(t), we see that

1 ohm = 1 V/A

The symbol used to represent the ohm is the capital Greek omega (Q).
Since R is constant, (2.1) is the equation of a straight line. For this reason, the

resistor is called a linear resistor. A graph of v(t) versus i(t) is shown in Fig. 2.2, which
is a line passing through the origin with a slope of R. Obviously, a straight line is the
only graph possible for which the ratio of v{t) to /(/) is constant for all /'(/).

Resistors whose do not remain constant for different terminal currents

resistances
are known For such a resistor, the resistance is a function of the
as nonlinear resistors.
current flowing in the device. A simple example of a nonlinear resistor is an incandes-
cent lamp. A typical voltage-current characteristic for this device is shown in Fig. 2.3,
where we see that the graph is no longer a straight line. Since R is not a constant, the
analysis of a circuit containing nonlinear resistors is more difficult.

v(t)

i(t)

+
Ht)

v(t)

FIGURE 2.1 Circuit FIGURE 2.2 Voltage-current FIGURE 2.3 Typical voltage -
symbol for the resistor characteristic for a linear resistor current characteristic for a non-
linear resistor

In reality, all practical resistors are nonlinear because the electrical characteristics
of all conductors are affected by environmental factors such as temperature. Many
materials, however, closely approximate an ideal linear resistor over a desired operating
region. We shall concentrate on these types of elements and simply refer to them as
resistors.
An examination of (2.1), in conjunction with Fig. 2.1, shows if /'(/) > (current
entering the upper terminal), then v(t) > 0. Thus the current enters the terminal of
higher potential and exits from that of the lower potential. Next, suppose that/(r) <
 7

Chap. 2 Resistive Circuits 1

(current entering the lower terminal). Then v(t) < 0, and the lower terminal is higher in
potential than the upper one. Once again, the current enters the terminal of higher
potential. Since charges are transported from a higher to a lower potential in passing
through the resistor, the energy lost by a charge </ [energy = qv(t)] is absorbed by the
resistor in the form of heat. The rate at which energy is dissipated is, by definition, the
instantaneous power

(2.2)

A graph of (2.2), shown in Fig. 2.4, reveals that p{t) is a parabolic (and thus
nonlinear) function of i{t) or v{t) which is always positive. (The horizontal scales are,
of course, different in the two cases.) Thus, for a linear resistor, the instantaneous
power is nonlinear even though the voltage-current relationship is linear.

The condition for passivity, given in (1.7), is

h<0 = p(t) dt>0

J"

Therefore, since p(t) is always positive, we see that the above integral is positive and
that the resistor is, indeed, a passive element.
The beginning student often encounters difficulty in determining the proper
algebraic sign in applying (2.1) when the voltage assignment differs from that of Fig.
2.1. Consider, for example, the assignment of Fig. 2.5. Comparing these figures, we see

that t>,(?) = — v(t) = —Ri(t). Therefore, when the voltage assignment is such that the
assumed current enters the terminal of lower potential (— ), a minus sign must be
employed in using (2.1). It should be noted that

„(a
py)
_ 1& - r-*>i(')]
2
_ vM
R R R

Hence the voltage assignment has no effect on the sign of p{t).

Another important quantity which is very useful in circuit analysis is known as
conductance, defined by

G =1 (2.3)

The unit of conductance is the mho (A/V), which is ohm spelled backwards. The
symbol representing the mho is an inverted omega (13). Combining (2.1)— (2.3), we see
that alternative expressions for Ohm's law and instantaneous power are

i(t) = Gv(t) TrGv ( 2 4)

-

and

pit) = ^= Gv\t) (2.5)

18 Resistive Circuits Chap. 2

As an example of the application of Ohm's law, consider finding the current for the
circuit of Fig. 2.6 in which a 12-V storage battery is connected to a 1-kQ resistor.
From (2.3) and (2.4), G = j^ = 10" 3 yand/ = 10" 3 x 12 A= 12 mA. Also, (2.5)
yields p(t) = 10" 3 x 12 2 = 144 mW, which is the minimum power rating or wattage
required for the resistor to ensure that it will not be damaged as a result of over-
1
heating.

p(t)

i(t)

(t) (or v(t))

/.(t) 12 V z=Z < i kn,

-f

FIGURE 2.4 Graph of the instan- FIGURE 2.5 Resistor F I GURE 2 . 6 Simple resistive
taneous power for a resistor with a reversed voltage circuit

assignment

The current in this example is a direct current since its value is invariant in time.
Suppose we now replace the 12-V battery by the time-varying voltage v{t) = 10 cos t V
and repeat the above procedure. The current is

l0
i(t) =
^ 1
A= 10 cos t mA
and the instantaneous power is

pit) = 0.1 cos 2 /W

which is always positive. The current, in this case, is an alternating current.

EXERCISES
2.1.1 The terminal current of a 20-kQ resistor is 10 mA. Find (a) the conductance,
(b) the terminal voltage, and (c) the minimum wattage of the resistor.
Arts, (a) 5 x 10" 5 15, (b) 200 V, (c) 2 W
2.1.2 The instantaneous power absorbed by a 1-kQ resistor is 10 sin 2 377/ W. Find
v(t) and i(t). Am. 100 sin 377/ V, 0.1 sin 377/ A
J
The wattage of a resistor is based on the average power (to be discussed in Chapter 12). For
direct currents, the instantaneous power equals the average power.
 9

Chap. 2 Resistive Circuits 1

2.1.3 Find v(t) and p(t). Ans. -50 V, 250 W

v(t) > ion

5A

EXERCISE 2.1.3

2.2 KIRCHHOFF'S LAWS

The solution of single-resistor circuits, such as those in the previous examples, may be
found using Ohm's law; however, more complicated circuits require the use of two
laws first stated by the German physicist Gustav KirchhoiT(l 824-1887) in 1847. 2 The
two laws are formally known as Kirchhoff 's current law and Kirchhoff 's voltage law.
These laws, together with the terminal characteristics for the various circuit elements,
permit systematic methods of solutions for any electrical network. We shall not attempt
to prove Kirchhoff *s laws here since the concepts necessary for the proof are developed
in a later, interesting study of electromagnetic field theory.
A circuit consists of two or more circuit elements connected by means of perfect
conductors. Perfect conductors are zero-resistance wires which allow current to flow
freely but accumulate no charge and no energy. In this case, the energy can be con-
sidered to reside, or be lumped, entirely within each circuit element, and thus the
network is called a lumped-parameter circuit.
A point of connection of two or more circuit elements is called a node. An example
of a circuit with three nodes is shown in Fig. 2.7(a). Node 1 consists of the entire con-

Node i

Node I

Node 3^
© Node 3

Node 2
Node 2
(o) (b)

FIGURE 2.7 (a) Three-node circuit; (b) Three-node circuit redrawn

2 ActualIy, even single-resistor circuits tacitly require KirchhofTs laws.

 /

20 Resistive Circuits Chap. 2

nection at the top of the circuit. The beginner quite often mistakes points a and b for
nodes. It should be noticed, however, that a and b are connected by a perfect con-
ductor and can be considered electrically as being identical points. This is readily
demonstrated by redrawing the circuit in thewhere at node 1 all form of Fig. 2.7(b),
connections are shown at a single point. Similar comments apply for node 2. Node 3 is
required for the interconnection of the independent voltage source and the resistor.
With these concepts, we are now ready to discuss the all-important laws of Kirchhoff.
Kirchhoff's current law (KCL) states that

The algebraic sum of the currents entering any node is zero.

To demonstrate the use of this law, consider currents flowing into a node, as shown in
Fig. 2.8. KCL states that

jl + 2 + (-/s) + i4
= <L
where we recall that i3 flowing out of the node is equivalent to — i3 entering the node.
For the sake of argument, let us suppose that the sum is not zero. In such a case, we
would have
U +h- /s + /« - * 9* o
where and hence must be the rate at which charges are accumulated
*P has units of C/s
in the node. However, a node consists of perfect conductors and cannot accumulate
charges. In addition, a basic principle of physics states that charges can neither be
created nor destroyed (conservation of charge). Therefore, our assumption is not
valid, and ¥ must be zero, demonstrating the plausibility of KCL.
Suppose in our example we now consider the sum of the currents leaving the node.
From Fig. 2.8, the sum is

—/, - i2 + i3 —U =
or, multiplying both sides by —1, we see that

i, + i2 — i3 + i4 =
which is identical to our previous result. This demonstrates an equivalent statement for
KCL, which states that

The algebraic sum of the currents leaving any node is zero.

Let us now rearrange the above equation in the form

^^K^r + i2 *
4 =h
where /,, i 2 , and /4 are entering the node and i3 is leaving. This form of the equation
illustrates another statement for KCL, stated as

The sum of the currents entering any node equals the sum of the currents leaving the
node.
Chap. 2 Resistive Circuits 21

In general, a mathematical expression of KCL is

N
E =o /. (2.6)

where i„ is the nth current entering (or leaving) the node and N is the number of node
currents.
As an example of KCL. let us find the current i in Fig. 2.9. Summing the currents
entering the node, we have

5 + -
/ (-3) -2 =
or
i = -6 A

FIGURE 2.8 Currents flowing into a node FIGURE 2.9 Example of KCL
AWe = —6 A
note that / entering the node is equivalent to 6 A leaving the node.
Therefore it is not necessary to guess the correct current direction prior to solving the
problem. We still arrive at the correct answer in the end.
If we sum the currents leaving the node, we have

-5-/+ (-3) + 2 =
or
/= -6 A

which completes a demonstration of two of the three forms of KCL.

We now move on to KirchhofT's voltage law (KVL). which states that

As an
The algebraic sum of the voltages around any closed path is zero.
(&Q
illustration, application of this statement to the closed path abcda offFii
Fig. 2.10
gives

v, + t' 2 V-, (2.7)

22 Resistive Circuits Chap. 2

where the algebraic sign for each voltage has been taken positive when going from +
to — (higher to lower potential) and negative when going from — to + (lower to
higher potential) in traversing the element.
As in the case of KCL, we shall not attempt a proof of KVL. However, to illustrate
the plausibility of (2.7), let us assume that its right member is not zero. That is,

—v + x
v2 — v3 = <S> ^
The left member of this by definition the work required to move a unit
equation is

charge around the path adcba. A lumped-parameter circuit is a conservative system,

which means that the work required to move a charge around any closed path is zero.
(Proof of this will come in a later course involving an interesting study of electro-
magnetic theory.) Thus, our assumption is not valid, and O is indeed zero.
We should point out, however, that all electrical systems are not conservative. In
fact, electrical power generation, radio waves, and sunlight, to mention only a few, are
consequences of nonconservative systems.
The application of KVL is independent of the direction in which the path is

traversed. Consider, for example, the path adcba in Fig. 2.10. Summing the voltages, we
find

V3 — V2 + *>1 = °

which is equivalent to (2.7).

In general, a mathematical representation for KVL is
N
(2.8)

where v„ is the nth voltage in a loop of N voltages. The sign of each voltage is chosen
as described earlier for (2.7).
As an example of the use of KVL, consider finding v in Fig. 2.1 1. Traversing the
circuit in a clockwise direction, we find

1 - [r^
+ v -
—V/r-
T
+W ^
^AA/
a a + 2V -
FIGURE 2.10 Voltages around a closed path FIGURE 2.11 Circuit to illustrate KVL

(r
Chap. 2 Resistive Circuits 23

or v = —3 V. Suppose we now perform a counterclockwise traversal. In such a case,

5 + 2-10-v =
or v = —3 V. which, of course, is the same result obtained for the clockwise traversal.
Still another version of KVL for Fig. 2.1 1 yields

10 +v=2+3
where the sum of the voltages with one polarity is equated to the sum of the voltages
with the opposite polarity. (Stated another way, the voltage rises equal the voltage
drops.)
In each of the previous examples, KVL has been applied around conducting paths,
such as abcda above. The law, however, is valid for any closed path. Consider, for
instance, the path acda of Fig. 2.11. We note that movement directly from a to c is not
along a conducting path. Applying KVL to this closed path yields v ac + 10 — 2 = 0,
where vac is the potential of point a with respect to c. Thus, v ae = — 8 V. We could
also have chosen the path abca, for which

-5 + v — vac = —5 - 3 - v ac =
Therefore. vac = — 8 V. which demonstrates the use of different closed paths to
obtain the same result.
As another example of the application of KCL
and KVL, consider finding i x and
vx in the network of Fig. 2.12. Summing
node a gives
the currents entering
—4 — 1 + /, = 0, or i, = 3 A. At node b, — /, + 2 — i 2 = 0, or i 2 = — 1 A. At
node c, i 2 + i 3 — 3 = 0, or i 3 = 4 A. Therefore, at node d, —i x — 1 — i 3 = 0, or
i x = —5 A. Next, KVL about the path abcda gives — 10 + v 2 — v x = 0. From Ohm's

law, t\ = 5/ 2 = -5 V. Therefore, v x = — 15 V.
Before concluding our discussion of Kirchhoff's laws, consider the network of
Fig. 2.13. in which several elements are shown within a closed surface S. We recall

that the current entering each element equals that leaving the device so that each
element stores zero net charge. Therefore, the total net charge stored within the
surface is zero, requiring that

h +h+ ii +U=
This result illustrates a generalization of KCL, which states

The algebraic sum of the currents entering any closed surface is zero. 3

To illustrate the plausibility of the generalized KCL, let us write KCL equations at
nodes a, b. c, and d of Fig. 2.13. The results are

3
The surface cannot pass through an element which is considered to be concentrated at a point
in lumped-parameter circuits.
24 Resistive Circuits Chap. 2

SS\\>5 \ .CLOSED
§f/ \V \^ SURFACE S

FIGURE 2.12 Network for example of KCL FIGURE 2.13 Network for generalized KCL
and KVL
'
= — l\ if.

—h -I*

= —h + it

m = h+h
Adding these equations yields

'l +h+ *3 + '

4 =
as previously noted.
From the generalized KCL, we see immediately in Fig. 2.12 for a surface enclosing
points a, b, c, and d that —i x — 4 + 2 — 3 = 0, or ix = —5 A.

EXERCISES
2.2.1 Identify the nodes in the figure.

ion

£> -AAA/—

i
Vw-
2fl

3.5 A 10 a (t) 2A
•"O
3A
-AAA^
IA

EXERCISE 2.2.1
Chap. 2 Resistive Circuits 25

2.2.2 Determine (a) /,. (b) /, . and (c) the potential of a with respect to h.

Am. (a) -3 A, (b) 4 A, (c) -8 V

2A

EXERCISE 2.2.2

2.2.3 In Ex. 2.2.1, find v x . Am. 22 V

2.3 SERIES RESISTANCE AND VOLTAGE

DIVISION

Now that the laws of Ohm

and Kirchhoff have been introduced, we are prepared to
We shall consider a simple circuit to be one
begin analyzing simple resistive circuits.
that can be completely described by a single equation. One type, which we shall
consider in this section, is a circuit consisting of a single closed path, or loop, of
elements. By KCL each element has a common current, say i. Then Ohm's law and
KVL applied around the loop yield a single equation in i that completely describes the
circuit.

Elements ar e said to be connected in ser ie s when they al l ca rry the same current.
Clearly, the networks of this section consist entirely of elements connected in series.
An important circuit of this type, consisting of two resistors and an independent
voltage source, provides an excellent starting point. We shall first analyze this special
case and then develop the more general case.
A single-loop circuit having two resistors and an independent voltage source is
shown in Fig. 2.14(a). The first step in the analysis procedure is the assignment of
currents and voltages to all elements in the network. In this circuit, it is obvious from
KCL that all elements carry the same current. We may arbitrarily call this current / in
the direction shown (clockwise). The novice often attempts to guess the true current
direction in making assignments. Such an assignment is not necessary, and is not
usually possible, even for the expert. We next make the voltage assignments for i?j and
R 2 as i\ and r 2 respectively. These assignments are also arbitrary but in the figure have
,

been chosen to satisfy Ohm's law for a positive algebraic sign.

26 Resistive Circuits Chap. 2

(a) (b)

FIGURE 2.14 (a) Single-loop circuit; (b) Equivalent circuit

The second step in the analysis is the application of KVL, which yields

v = Wj + v2
where, from Ohm's law,
t>, = A\/
(2.9)
R 2i

Combining these equations, we find

v = R^i +R 2
i

Solving for i yields

(2.10)
R, + R2
Let us now consider a simple circuit consisting of the voltage source v connected to
a resistance R s , as shown in Fig. 2.14(b). If R s
is selected such that

v
1
> = R,
(2.11)

then the network is called an equivalent circuit of Fig. 2.14(a) because an identical
current (response) is produced for a voltage v (excitation). In general, two circuits are
said to be equivalent when they exhibit identical voltage-current relationships at their
terminals.
Comparing (2.10) and (2.11), it is obvious that

R = s R, +R 2 (2.12)

Interpreting (2.12), we see that if the series combination of R {

and R 2 is replaced by a
resistor R s , the same current flows from the source v. Therefore R s
is the equivalent
resistance of the series connection.
Combining (2.9) and (2.10), we see that

R,
v, =
R + R2
l

(2.13)
R*
R, A'-
Chap. 2 Resistive Circuits 27

The potential of source r divides between resistances /?, and /?, in direct proportion to
their resistances, demonstrating the principle of voltage division for two series resistors.
We s ee in (2. 1 3) that the greater voltage appears across the larger resistor,
I he instantaneous powers absorbed by /?, and R 2
are

Pl
Pi = i^
R x
(R.
*<
+ R.Y
*>
1 V=*R
and
_v\
~ _ R2 V 2
Pl
R -(R + R y2 i 2

respectively. The total power absorbed is

p ^ p> = RT^r
v2
l
I

iR7+Rj
v \
=
== VI

The p ower delivered by the source also equals vi. indicating that the powe r delivered by
R and R 2 This result is
tne source eq uals tha t absorbed by t
. known as conservation oj
power (sometimes also referred to as Tellegen s theorem), a property which is often
useful in circuit analysis.
Let us now digress briefly and repeat the analysis for the counterclockwise current
assignment of Fig. 2.15. Application of KVL yields
V = vl + vz

in which

»i = ~ R xh
v2 = -R 2i x

Combining these equations, we have

v = -RJ, - R 7ix

from which

i, =
R ^-R
t 2

Comparing our result with (2.10). we see that i, = — /'.

which is also evident from
an inspection of Figs. 2.14(a) and 2.15. Hence, if the proper direction for positive
current happens not to be chosen, a negative sign will result. This, of course, by
definition means the positive current flows in the opposite direction. Similar state-
ments apply to voltage assignments. If the polarity for positive voltage happens not to
be chosen, a negative sign will occur in its solution, leading, of course, to the correct
answer.
As an example of the utility of the preceding analysis, suppose v = 120 sin t V,
/?, = 90 Q. and i\ = 72 sin t V in Fig. 2.14(a). Let us now determine R 2 ,
/',
and the
instantaneous power associated with each element. From (2.13)
28 Resistive Circuits Chap. 2

90
72 sin t 120 sin t
90 + R,

which yields R =2 60 Q. Hence, R = s

and, from (2.10), i = 120
i?, + i? 2 = 150 fi,

sin //150 = 0.8 sin / A. The instantaneous power to and R 2 is p = RJ 2 = 57.6 R t x

sin 2 1 W and p 2 = R 2 i 2 = 38.4 sin 2 t W. Thus, the power delivered by the source is
96 sin 2 / W. We also observe that the power absorbed by R and R 2 should equal that x

which is absorbed by R s in Fig. 2.14(b). The power to R s is RJ 2 = 96 sin 2 t W.

v l|R|

6 V
2|R 2

+ '

n|R n
v

FIGURE 2.15 Single-loop circuit with FIGURE 2.16 Single-loop circuit with

counterclockwise current N series resistors

Let us now extend our analysis to include the series connection of A^ resistors and
an independent voltage source, as shown in Fig. 2. 16. KVL gives

v = v. + vN
in which

v2 =R 2i

(2.14)

vN = RN i

Therefore
v = R,i +R + 2i . . . + RN i

Solving this equation for / yields

v
i = (2.15)
/?, +R 2 R*

Let us now select R s

in the circuit of Fig. 2.14(b) so that (2.1 1) is satisfied. Equiva-
lence of (2.11) and (2.15) requires that
N
R, = R, + R + 2 . . . + RN = -£R„ (2.16)

Si s
Chap. 2 Resistive Circuits 29

Therefore the equivalent resistance of N series resistors is simply the sum of the
individual resistances.
Substituting (2.15) and (2.16) into (2.14). we find

— R,
tllv
R,

(2.17)

Rv

which are the equations describing the voltage division property for TV series resistors.
Again, we see that the voltage divides in direct proportion to the resistance.
The instantaneous power delivered to the series combination, from (2.2) and
(2.17). is

vl
' = # + #+•••+*
R R, 2
~ R:
t
9
v + . + ^v 2

R] Ri

= VI
R.

This power is equal to that delivered by the source, verifying conservation of power for
the series connection of TV resistors.

EXERCISES
2.3.1 For the network shown, find (a) an equivalent circuit, (b) the current (c) the ;',

power delivered by the source, (d) v u and (e) the minimum wattage required
for /?!. Am. (b) 0.5 A, (c) 5 W, (d) -4 V, (e) 2 W
ion

EXERCISE 2.3.1
30 Resistive Circuits Chap. 2

2.3.2 In Fig. 2.14 (a), v = lOe"', v 2 = 2e~', and /?, = 16 0. Find (a) R2 ,
(b) the
instantaneous power delivered to R 2 and (c) , the current i.

Am. (a) 4 Q, (b) e~ 2 < W, (c) 0.5<r' A

2.3.3 A resistive load 4
requires 8 V and dissipates 2 W. A 12-V storage battery is

available to operate the load. Referring to Fig. 2.14 (a), if R2 represents the load
and v the 12-V battery, find (a) the current i, (b) the necessary resistance R x ,

and (c) the minimum wattage of R x

. Ans. (a) 0.25 A, (b) 16 Q, (c) 1 W

2.4 PARALLEL RESISTANCE AND CURRENT

DIVISION

Another important simple circuit is the single-node-pair resistive circuit. In analyzing

these networks, we shall first examine a special case and then develop the more
general case, as was performed for the single-loop network.
Elements are connected in parallel when the same voltage is common to each of
them. A single-hode-pair circuit consisting of the parallel connection of two resistors
and an independent current source is shown in Fig. 2.17(a). To begin the analysis, we

first assign voltages and currents to each circuit element. As in the case of the single-
loop circuit, the assignments are completely arbitrary. We have chosen the assign-
ments given by /,, i 2 , and v.

(a) (b)

FIGURE 2.17 (a) Single node-pair circuit; (b) Equivalent circuit

Next, we need to apply either KCL or KVL. Inspection of the circuit shows that a
single node (either the upper or the lower) is common to all elements. This suggests
that for single-node-pair circuits KCL is the most efficient choice. Applying KCL at
the upper node yields

where, from Ohm's law,

/, = = G<o
(2.18)
h = = G2 v
Combining these equations gives

i =Gv+G 2v

4 A load is an element or collection of elements connected between the output terminals. In this

case, the load is a resistor.

Chap. 2 Resistive Circuits 31

and solving for v. we find

v = (2I9 >
Grnr t

In the circuit of Fig. 2.17(b), if Gp is selected such that

vp ~ = ±- = v (2.20)
Gp

then the network is an equivalent circuit to that of Fig. 2.17(a). Comparing (2.19) and
(2.20). we see that

GP = G + G ] Z (2.21)

Clearly. Gp is the equivalent conductance of the two parallel conductances. In terms

of resistances, (2.21) becomes

or

(2.22)

Therefore, the equivalent resistance of two resistors connected in parallel is equal to

the product of their resistances divided by their sum. It is interesting to note that Gp is

greater than either G t

or G2 : therefore. Rp is less than either R x
or R 2 . From this
result, we see that connecting resistors in parallel reduces the overall resistance.
In the special case R = R we
2 {
, see from (2.22) that Rp = RJ2.
Substituting (2.19) into (2.18) gives

G, .

(2.23)
G, .

'i-g—tg;'

The current of the source /'

divides between conductances (7, and G 2
in direct propor-
tion to their conductances, demonstrating the principle of current division. It is common
practice to give the resistor values in circuit diagrams in ohms (resistance) and not
mhos (conductance). In terms of resistance values, (2.23) becomes

R2 .

(2 24)
'

A,

Therefore the current divides in inverse proportion to the resistances. We see that the
larger current flows through the smaller resistor. The power absorbed by the parallel
32 Res/stive Circuits Chap. 2

combination is

Pi — R-lh + *2*2

R\i
2^1 + +R 2^2
(*1 + *2)
2
(#. 2)

*1*2 :2
VI
R, +R 2

which equals that delivered to the network by the current source.

Suppose in Fig. 2.17(a) that R — 3 Q, R 2 = 6 Q, and i 3 A. Then, from
{
= (2.22),
i?, = (3)(6)/(3 + 6) = 2 fi. From (2.24), i, = f(3)
= 2 A, and /2 = |(3) = 1 A. The
voltage v = RJi =R zi 2
= (3)(2) = 6 V. The current of the source flows through an
equivalent resistance of R p Hence the voltage. is also given by v = Rp = i (2)(3) = 6 V.
Let us now consider the more general case of N parallel conductances and an
independent current source, as shown in Fig. 2.18. KCL gives

FIGURE 2.18 Single-node-pair circuit with N parallel conductances

i = /, + i2

for which
/, = G,v

(2.25)

i N = GN v
Therefore we have
i = G{o +G 2
v + . . . + GN v
from which

v — (2.26)
G,+G + 2 ... +G t

If we now select G in Fig. 2.17(b) such that (2.20) is satisfied, then (2.26) requires
that

Gp = G, +G + 2 . . . + GN = £ i = i
G, (2.27)

In terms of resistances, this equation becomes

R
- -+-
R +
/?,
... +- =£
R N i=i
-
Rj
(2.28)
2
 -

Chap. 2 Resistive Circuits 33

Hence the reciprocal of the equivalent resistance is simply the sum of the reciprocals
of the resistances.
Combining (2.25)-(2.28), we find

G,. Rp .

R,,
R,
(2.29)

. _ Gjv .
_ /?„ .

Again the currents divide in inverse proportion to the resistances.

We observe in (2.28) that for N > 2 an expression for R p is more complicated than
(2.22). Formulas could be obtained, of course, for N = 3, 4, etc., but it is usually
easier to apply (2.28) directly. As an example, suppose for N = 3 that i?, = 4 Q,
Rz = 12 Q, and R 3 = 6 Q. Then

R, 12
= TU
and Rp = 2 Q.
Let us now consider finding the equivalent resistance R eq of the network of Fig.
2.19(a). as viewed from terminals x-y. Such reductions are very helpful in analyzing
many types of circuits, as we shall see in the next section. The process is carried out by

7a i n 7 A

*eq^ 12 A 5 n 6 A

(a) (b)

7 A
*o- ^AA/—

:4 a R eq — II A

(c) (d)

FIGURE 2.19 Steps in determining the equivalent resistance of a

network
34 Resistive Circuits Chap. 2

successive combinations of parallel and series connected resistors. In Fig. 2.19(a), the

student often errs in taking combinations such as the 7- and 12-fi resistors to be in
series. We see, at node a, a current in the 7-Q resistor would divide
however, that
between the and 12-Q resistor; hence they cannot be in series. The 1- and 5-fi
1-

resistors, however, would carry the same current. Therefore they are in series, having a

series resistance of 6 Q as shown in Fig. 2.19(b). We now observe that the same voltage
would occur across the 6- and 12-Q resistors, indicating a parallel connection having
an equivalent resistance of (6)(12)/(6 + 12) = 4Q, as shown in Fig. 2.19(c). It is
apparent that the 7- and 4-fi resistors of this network are in series, yielding an equiva-
lent resistance for the entire network of 1 Q [Fig. 2.19(d)]. Therefore, from terminals
1

x-y, the network could be replaced by a single resistor of 1 1 Q. This is useful in deter-
mining, for instance, the power delivered by a source connected to terminals x-y.
Suppose a 22-V source is applied. Then the current flowing from the source is / = ^
= 2A, which gives an instantaneous power p(t) = (22)(2) = 44 delivered to the W
resistor network.

EXERCISES

2.4.1 In the circuit shown, find (a) the equivalent resistance seen from the source
terminals, (b) v u (c) i2 , and (d) i.

Ans. (a) 20 fi, (b) -120 V, (c) -2 A, (d) -6 A

EXERCISE 2.4.1

2.4.2 In Fig. 2.18, if N= 3, R^ = 120 Q, R2 = 60 Q, and v = 10 sin t V and the

instantaneous power delivered by the current source is 5 sin 2 t W, find (a) -/? 3 ,

(b) i, and (c) i} . Ans. (a) 40 Q, (b) 0.5 sin t A, (c) 0.25 sin / A
2.4.3 A load requires 4 A and absorbs 16 W. A current source of 10 A is available.

For the circuit of Fig. 2.17 (a), if G 2 and i the current source,
represents the load
find (a) the voltage v and (b) the required conductance C,.
Ans. (a) 4 V, (b) 1.5 U
2.4.4 Find the equivalent resistance, R eq . Ans. 9.4 Q
Chap. 2 Resistive Circuits 35

v
eq

EXERCISE 2.4.4

2.5 ANALYSIS EXAMPLES

For our first example, let us consider a single-loop circuit containing three resistors and
two independent voltage sources, as shown in Fig. 2.20(a). KVL and Ohm's law give

-10 20/ + 30; + 20 + 50/ =

Thus the reduced equation is

10 + 100/ =
for which = -0.1 A and y, = 30/ = — 3 V.
i

We see that the reduced equation is satisfied by the equivalent circuit of Fig. 2.20(b),
where the voltage sources are replaced by sum and the resistances by
their algebraic
their series equivalent. From this circuit, power absorbed by all resistors is
the

/'loon
= 100/ 2 = 1 W. Therefore the net power delivered by the two voltage sources
must be 1 W. We recall that the power absorbed by an element is the product of the
voltage and the current, where the current enters the positive voltage terminal. For the
10-V source, we see that /> 10V = (10)(-/) = (10)(0.1) = 1 W absorbed. Physically,
this means that current is flowing into the positive terminal and that the source
is receiving power, or is being charged. The absorbed power in the 20-V source is

p 2Q v = 20/ = (20)( — 0.1) = — 2 W. The minus sign indicates that the source is

delivering 2 W to the circuit. Hence the net power from the two sources is 1 W, which
is the power delivered to the resistors. Finally, let us find the potential of point a with

20 A 30 n,
'VW
+ v, -

,.v(*)

so a
© 20 V I ioo a

-w\, —
(a) (b)

FIGURE 2.20 (a) Single-loop circuit; (b) Equivalent circuit

36 Resistive Circuits Chap. 2

respect to point b. which we denote as v ab . From KVL, — 10 + 20/ + v ab = 0, or

v ab = 10 - 20/ = 10 - (20)( — 0.1) = 12 V. i~„ tO
For a second example, let us find /, and v(t) for the circuit of Fig. 2.21. In the
circuit, three conductances and two independent current sources are connected in
parallel. Applying KCL to the upper node yields

10 sin nt - 0.01 v{t) - 0.02r(/) - 5 - 0.07i>(/) =

or
(10 sin nt - 5) - 0.1 K0 =

+
10 s n*t(T) v(t)§.OI U 02 V 05A 07 U

FIGURE 2.21 Single-node-pair circuit

It is apparent that a current source of (10 sin nt — 5) A connected to a conductance of

0.1 15 (a 10-fi resistor) would be an equivalent circuit for the network [see Fig. 2.17(b)].
The total instantaneous power absorbed by the conductances is therefore

= (10 sin nt - 5)
2

Pab;
0.1

= 1000 sin nt(sm nt - 1) + 250 W

Solving for v{t) above and subsequently for /,, we have
v(t) = 100 sin nt - 50 V
i x
= 0.02(100 sin nt - 50) = 2 sin nt - I A
We observe v(0) = -50 V, r(0.5) = 50 V, and u(1.5) = -150 V.
The power delivered by the leftmost source is

p x
= 10 sin nt(\00 sin nt - 50) W
Similarly, the 5-A source delivers

p2 = —5(100 sin 71/ - 50) W

Thus the total power delivered by these sources is

p del = Pl -f p2 = 1000 sin ?r/(sin nt — 1) + 250 W

which equals that absorbed by the conductances.
Chap. 2 Resistive Circuits 37

As a third example, consider finding r. the power delivered by the source, and
/.

thepower absorbed by the 8-fi load of Fig. 2.22(a). We begin by obtaining successive
combinations of parallel and series resistor connections. The 4- and 8-fi resistances
(in series) add to give 12 fi. These 12 Q are in parallel with the 6-fi resistor, giving an
equivalent value of (12)(6)/(12 -4- 6) = 4fi [Fig. 2.22(b)]. We now add the 12- and
4-fi resistances, which are in parallel with the 16 Q, giving (16)(16)/(16 + 16) = 8 Q,
as shown in Fig. 2.22(c). This is the equivalent resistance as seen looking into the
circuit at .v-.v. Applying KVL to the simplified circuit, we have

-30e" 2/ + 8/ =
from which
i = 3<r 2 '
A

Therefore the power delivered by the source is

p, = (30e- 2 0(3e>- 2 ') = 90t>-

4 '
W
By voltage division we see that

Vl = (-Jr^30<r = 2'
24<r 2 'V

(o)

30e" 30e

lb) (c)

FIGURE 2.22 Circuit for analysis example using voltage division

which is the voltage across points x-x in the circuit. Proceeding to Fig. 2.22(b), we see
that v t
is the voltage across the series combination of the 12- and 4-fi resistors;
 —

38 Resistive Circuits Chap. 2

hence, again using voltage division, we find

U2 = t,1=6e
" 2'
v
(T2T4)

which is the voltage across points y-y in the circuit. In Fig. 2.22(a), it is obvious that
vz is the voltage across the series connection of the 4- and 8-fi resistors. Therefore
voltage division requires that

4e~ 2 'V
(wh) v >

The power to the 8-Q load is

p2 = V = 2e-*' W
As a final example, let us find the current i in Fig. 2.23(a). We first combine the two

6-Q resistors on the right of Fig. 2.23(a) and obtain the equivalent resistance which is
in parallel with the 4-fi resistor. The result, (4)(12)/(4 + 12) = 3 Q, is shown in
Fig. 2.23(b). If we now replace the two series resistors to the right of points x-x and
the parallel 3- and 6-fi resistors to the left of x-x by their respective series and parallel
equivalents, we obtain the circuit of Fig. 2.23(c). Using current division, we find

x12 = 3a
'HtT6)
A second application of current division to Fig. 2.23(a) reveals immediately that

x3 =
'-(fTTT,)'.- (1) TA
x 3n 6 A
H *AV ^A\

12 A &a 2>a 4 A 'J

>6A

6A

(b) (c)

FIGURE. 2.23 Circuit for analysis example using current division

Chap. 2 Resistive Circuits 39

EXERCISES
2.5.1 Find v, r ab , and the power delivered by the 5-V source.
Ans. -10 V, -8.33 V, -0.167 W
20.fl 60A
-• "vVv-

30 A b 40 a
EXERCISE 2.5.1

2.5.2 Find G and construct an equivalent circuit having one current source and a
single conductance. Ans. G = 0.03 15, i = 3 sin t A directed upward

2 sin t A
(| J
5 .01 IT |.02 V () 5 sin t A G550sintV

EXERCISE 2.5.2

2.5.3 Find the power absorbed by the 90-Q resistor. Ans. 3.6 W

24n.

EXERCISE 2.5.3

2.5.4 Find u and i. Ans. 6 V, -£ A

ion 2 A
-t —
'Wv -^Wv—

2A 4n: 6A 3A :6fi 2n

EXERCISE 2.5.4
40 Resistive Circuits Chap. 2

2.6 AMMETERS, VOLTMETERS, AND

OHM METERS*

A good example of the usefulness of current and voltage division is demonstrated in

the design of simple two-terminal measuring instruments, such as ammeters, volt-
meters, and ohmmeters. An ideal ammeter measures the current flowing through its

terminals and has zero voltage across its terminals. In contrast, an ideal voltmeter
measures the voltage across its terminals and has a terminal current of zero. An ideal
ohmmeter measures the resistance connected between its terminals and delivers zero
power to the resistance.
The practical measuring instruments that we shall consider only approximate the
ideal devices. The ammeters, for instance, will not have zero terminal voltages.
Likewise the voltmeters will not have zero terminal currents, and the ohmmeters will
not have zero power delivered from their terminals.
A popular type of ammeter consists of a mechanical movement known as a
D'Arsonval meter. This device is constructed by suspending an electrical coil between
the poles of a permanent magnet. A dc current passing through the coil causes a
rotation of the coil, as a result of magnetic forces, that is proportional to the current.
A pointer is attached to the coil so that the rotation, or meter deflection, can be
visually observed. D'Arsonval meters are characterized by their full-scale current,
which is the current that will cause the meter to read its greatest value. Meter move-
ments are common having from 10 juA to 10 mA.
full-scale currents

An equivalent circuit for the D'Arsonval meter consists of an ideal ammeter in

series with a resistance RM , as shown in Fig. 2.24. In this circuit, RM represents the
resistance of the electrical coil. Clearly, a voltage appears across the ammeter terminals
as a result of the current i flowing through RM RM
. is usually a few ohms, and the
terminal voltage for a full-scale current is nominally from 20 to 200 mV.
The D'Arsonval meter of Fig. 2.24 is an ammeter which is suitable for measuring
dc currents not greater than the full-scale current 7FS Suppose we wish, however, to
.

measure a current which exceeds /Fs It is apparent that we must not allow a current
.

greater than 7FS to flow through the device. A circuit to accomplish this is shown in
Fig. 2.25, where R p is a parallel resistance which reduces the current flowing through
the meter coil.

* -i M

"(D
'

<^>
FIGURE 2.24 Equivalent circuit for a FIGURE 2.25 Ammeter circuit

D'Arsonval meter
Chap. 2 Resistive Circuits 41

From current division, we sec that

R,
* FS
— 'fs
R. R

where iFS is the current which produces /, s in the D'Arsonval meter. (Clearly, this is

the maximum current the ammeter can measure.) Solving for R p we

, have

(2.30)
p i
'FS
— YI F

A dc voltmeter can be constructed using the basic D'Arsonval meter by placing a

resistance R s
in series with the device, as shown in Fig. 2.26. It is obvious that the full-

scale voltage, v = v FS , occurs when the meter current is 7FS Therefore, from
. KVL,

RJ
s'FS R.uhs —
1(2
from which

R — 7^ — Rm
s
t
V (2.31)

The current sensitivity of a voltmeter, expressed in ohms per volt, is the value
obtained by dividing the resistance of the voltmeter by its full-scale voltage. Therefore

Q/V rating = R. + R, R
(2.32)

(Note: "«k" means approximately equal to.)

A simple ohmmeter circuit employing a D'Arsonval meter for measuring an

unknown resistance Rx is shown in Fig. 2.27. In this circuit, the battery E causes a
current i to flow when R x is connected into the circuit. Applying KVL, we have

-E + (R, + RM + R x - )i

from which

R, (Rs + RM )

-SVW-

FIGURE2.26 Voltmeter circuit FIGURE 2.27 Ohmmeter circuit

42 Resistive Circuits Chap. 2

We select E and R s such that for Rx — 0, / = 7 FS . Therefore

R< + R,

Combining the last two equations, we find

R* =
(¥ ~ Y +
R l
* Rm) (2.33)

A very popular general-purpose meter which combines the three previously

described circuits is the VOM (voltmeter-ohmmeter-milliammeter). In the VOM,
provisions are made for changing R p and R s
so that a wide dynamic range of operation
is provided.

EXERCISES

2.6.1 A D'Arsonval meter has 7 FS = 1 mA and RM = 50 Q. Determine R„ in Fig.

2.25 so that /
FS is (a) 1.0 mA, (b) 10 mA, and (c) 100 mA.
Am. (a) infinite, (b) 5.556 fi, (c) 0.505 Q
2.6.2 In Fig. 2.26, determine R and the fi/V rating for a voltmeter to have a full-
s

scale voltage of 100 V using a D'Arsonval meter with (a) R M = 100 fi and
7FS = 50 fiA and (b) R M = 50 fi and 7FS = 1 mA.
Arts, (a) 2 Mfi, 20 kfi/V, (b) 100 kfi, 1 kfi/V

2.6.3 What voltage would each meter design of Ex. 2.6.2 measure in the circuit
below? Why are the two measurements different?
Am. 99.5 V, 90.9 V

10 mA

EXERCISE 2.6.3

2.6.4 The meter movement of Ex. 2.6.1 is used to form the ohmmeter circuit of Fig.
2.27. Determine R s and E so that /'
= / FS /2 mA when R x = 10 kfi.
Am. 9.95 kfi, 10 V
 .

Chap. 2 Resistive Circuits 43

2.7 PHYSICAL RESISTORS*

Resistors are manufactured from a variety of materials and are available in many
sizesand values. Their characteristics include a nominal resistance value, an accuracy
with which the actual resistance approximates the nominal value (known as tolerance).
a power dissipation, and a stability as function of temperature, humidity, and other
;i

environmental factors.
The most common type of found in electrical circuits is the carbon com-
resistor
position or carbon film resistor. The composition type is made of hot-pressed carbon
granules. The carbon film device consists of carbon powder which is deposited on an
insulating substrate. A typical resistor of this type is shown in Fig. 2.28. Multicolored

% Tolerance

4)iiD) y

a b c

FIGURE 2.28 Carbon resistor

°
bands, shown as a. b, c. and on the resistor body to indicate
tolerance, are painted
the nominal value of the resistance. The color code
bands is given in Table 2.
for the 1

Bands a. b. and c give the nominal resistance of the resistor, and the tolerance band
gives the percent that the resistor may deviate from the nominal value. Referring to
Fig. 2.28, the resistance is

R = (10a -r b)\0 c ± % tolerance (2.34)

TABLE 2.1 Color code for carbon resistors

Bands a, b, and c

Color Value Color Value

Gold* -2 Yellow 4
Silver* -1 Green 5

Black Blue 6
Brown 1 Violet 7
Red 2 Gray 8
Orange 3 White 9

% Toleranc e Band

Gold ±5%
Silver ±10%

These colors apply to band c only.

R^(\o* (- b) \b
c\-
44 Resistive Circuits Chap. 2

by which we mean that the % tolerance of the nominal resistance is to be added or

subtracted to give the range in which the resistance lies.

As an example, suppose we have a resistor with band colors of yellow, violet, red,
and silver. The resistor will have a value given by

R= (4 x 10 + 7) x 10 2 ± 10%
= 4700 ± 470 Q

Therefore the resistance value lies between 4230 and 5 1 70 Q.

Values of carbon resistors range from 2.7 to 2.2 x 10 7 Q, with wattages from
^ to 2 W. For resistance values less than 10 £2, we see from (2.34) that the third band
must be gold or silver. Carbon resistors are inexpensive but have the disadvantage of
a relatively high variation of resistance with temperature.
Another resistor type which is commonly used in applications requiring a high
power dissipation is the wire-wound resistor. These devices consist of a metallic wire,
usually a nickel-chromium alloy, wound on a ceramic core. Low-temperature-
coefficient wire permits the fabrication of resistors that are very precise and stable,
having accuracy and stability of the order of ±1 % to ±0.001 %.
The metal film resistor is another valuable and useful resistor type. These resistors
are made by vacuum-depositing a thin metal layer on a low-thermal-expansion
substrate. The resistance is then adjusted by etching or grinding a pattern through the
film. Accuracy and stability for these resistors approaches that of wire-wound types,

and high resistance values are much easier to attain.

In the future, the reader may be more likely to encounter resistors in integrated
circuits, which were developed in the late 1950s and came into their own in the 1960s.

An integrated circuit is a single monolithic chip of semiconductor (material with

conducting properties between those of a conductor and an insulator) in which active
and passive elements are fabricated by diffusion and deposition processes. Integrated
hundreds of elements are available on chips about ^ in. square. An
circuits containing
integrated-circuit resistor has the typical structure shown in Fig. 2.29.

RESISTIVE
LAYER

FIGURE2.29 Integrated- circuit resistor

Chap. 2 Resistive Circuits 45

EXERCISE
2.7.1 Find the resistance range of carbon resistors having color bands of (a) brown,
black, red, silver; (b) red, violet, yellow, silver; and (c) blue, gray, silver, gold.
Ans. (a) 900-1 100 Q, (b) 243-297 kQ, (c) 6.46-7.14 Q

PROBLEMS
2.1 A 5-V battery is connected to the ends of a 1000-ft length of conducting wire
and a 10- A current flows. What is the resistance per foot of the wire?

2.2 A 2-kQ resistor is connected to a battery and 10-mA flow. What current will
flow if the battery isconnected to a 500-Q resistor?

2.3 A 200-Q resistor is connected to a 12-V source. Find the current and the mini-
mum wattage of the resistor. If a 100-fi resistor is inserted in series in the circuit,
what is the voltage across it ?

2.4 Find v and /'.

PROBLEM 2.4

2.5 Several series resistors draw 100 mA from a 10-V source. If a 1-kfi resistor is

inserted in series in the circuit, find the new current.

2.6 Two resistors connected in series draw 25 mA from a 100-V source. The voltage
across one of the resistors is 25 V. Find the value of each resistor.
46 Resistive Circuits Chap. 2

2.7 Design a voltage divider to provide 1, 5, 10 and 20 V from a 25-V source. The
source is to deliver 25 mW of power to the divider.
2.8 A 1-A current source is connected to the parallel combination of 20- and 30-Q

resistors. Find the voltage, current, and minimum wattage of each resistor.

2.9 A 6-Q resistor is added in parallel with the combination of Prob. 2.8. Find the
voltage and current for this resistor.

2.10 Find the equivalent resistance of 10 1-kQ resistors connected in parallel.

2.11 A 4-fi and a 12-fi resistor are in parallel. The parallel combination is in
resistor
series with a 7-Q resistor and a voltage source of 10 sin It V. Find the current
in the 4-Q resistor and the instantaneous power from the source.

2.12 Two 1-kfi resistors are in series. When a resistor R is connected in parallel with
one of them, the resistance of the combination is 1750 Q. Find R.

2.13 Find R ea .

ea

6fl
l
eq irx> 2 n

-AAAr-
5A 7 A
PROBLEM 2.13

2.14 A voltage of 20e~' V is connected to the network of Prob. 2.13. Find v.

2.15 A current source is connected to the network of Prob. 2.13 and i = 3 cos t A.
What is the value of the current source?

2.16 Find i.

PROBLEM 2.16

2.17 Repeat Prob. 2.16 with the 6-V source replaced by a current source of 6 A.

2.18 Find R.
Chap. 2 Resistive Circuits 47

4a 5n
-'Wv/-
2A

20 V
© 3flS

PROBLEM 2.18

2.19 Find the power delivered to the 1-Q resistor.

I2A
12 V

PROBLEM 2.19

2.20 Find the potential of a with respect to b in Prob. 2.19.

2.21 Find R.

t V>A—
+ 30V
+
I5A( ii
3A< 4A v \ i2n

PROBLEM 2.21

2.22 Find v in Prob. 2.21.

2.23 Find il and i2 .

PROBLEM 2.23
48 Resistive Circuits Chap. 2

2.24 Find v(t) and the instantaneous power to the 3-ft resistor.

+
v(t)> in.
© I A
f
2a
G) e
"A 3A

PROBLEM 2.24

2.25 Find t<f).

4 n
4 A 4 A
-^\AA/ O-
+
4 A
-© V(t) Q4s,n t V

-AAA, O-
4A 4A
PROBLEM 2.25

2.26 A D'Arsonval meter has a full-scale current of 1 mA and a resistance of 4.9 Q.

If a 0.1 -ft parallel resistor is used in Fig. 2.25, what is z'
FS ? What voltage occurs
across the meter?

2.27 A 20,000-Q/V voltmeter has a full-scale voltage of 120 V. What current flows in
the meter when measuring 90 V?
2.28 Two 10-kft resistors are connected in series across a 100-V source. What voltage
will the voltmeter of Prob. 2.27 measure across one of the 10-kfJ resistors?
Repeat for two \-MQ resistors in series.

2.29 The D'Arsonval meter of Prob. 2.26 is used for the ohmmeter of Fig. 2.27.
What value of series resistance is required if E— 1.5V? What value of unknown
resistance will cause a one-quarter full-scale deflection?

2.30 Determine the color codes for resistors having the following resistance ranges:
(a)4.23-5.17 Q, (b) 6460-7140 ft, and (c) 3.135 -3.465 Mft.
 i#
^
3
DEPENDENT SOURCES

The voltage and current sources of Chapters 1 and 2 are independent sources, as
defined earlier in Sec. 1.4. We may also have dependent sources, which are very
important in circuit theory, particularly in electronic circuits. In this chapter we shall
define dependent sources and consider an additional circuit element, the operational
amplifier, which may be used to obtain dependent sources.
We shall also analyze a few simple circuits containing resistors and sources, both
independent and dependent. As we shall see. the analysis is very similar to that
performed in Chapter 2. and the results may be used to construct a number of impor-
tant circuits, such as amplifiers and inverters, which will be defined in the chapter.

3.1 DEFINITIONS

A dependent or controlled voltage source is one whose terminal voltage depends on, or
is some other place in the circuit. A
controlled by. a voltage or a current existing at
voltage-controlled voltage source(VCVS) is a voltage source controlled by a voltage,
and a current-controlled voltage source (CCVS) .is one controlled by a current. The
symbol for a dependent voltage source with terminal voltage v is shown in Fig. 3.1(a).
A dependent, or controlled current source, symbolized by Fig. 3.1(b), is one whose
current is dependent on a voltage or a current existing elsewhere in the circuit. A
voltage-controlled current source (VCCS) is controlled by a voltage, and a current-
controlled current source (CCCS) is controlled by a current.
Figure 3.2 illustrates the four types of controlled sources and shows the voltage or
current on which they are dependent. The ju and ft are dimensionless con-
quantities
stants, commonly referred to as the voltage and current gain, respectively. The
constants /- and g have units of ohms and mhos, respectively.

49
50 Dependent Sources Chap. 3

'+N
Kt

6 6

(a) (b)

FIGURE 3.1 (a) Dependent voltage

source; (b) Dependent current source

o-
+

V --
/iV, V = n,

(a) (b)

f J>
i = gv, f>..*l

(O (d)

FIGURE 3.2 (a) VCVS; (b) CCVS; (c) VCCS; (d) CCCS

As an example, in the circuit of Fig. 3.3 we have an independent source, a depen-

dent source, and two resistors. The dependent source is a voltage source controlled by
the current i v The value of r for the dependent source is 0.5 V/A.

3.2 CIRCUITS WITH DEPENDENT SOURCES

Circuits containing dependent sources are analyzed in the same manner as those
without dependent sources. That is. Ohm's law
and Kirchhoff's voltage
for resistors
and current laws apply, as well as the concepts of equivalent resistance and voltage and
current division.
As an illustration of the procedure, let us find the current /in the circuit of Fig. 3.4.
Chap. 3 Dependent Sources 51

SV,
i

t
2n
Vva^-
- V, +

6
0.5i.
6V V,?6A

FIGURE 3.3 Circuit containing a dependent FIGURE 3.4 Dependent source example
source

The dependent source is a voltage source controlled by the voltage v x

as shown.
Applying KirchhofTs voltage law around the circuit, we have

-v, + 3v l
+ v2 = 6 (3.1)
f*i 3-V
and by Ohm's law we have
Vy = —2i, v2 = 6i (3.2)

Using (3.2), we may eliminate i\ and v 2 in (3.1), which results in

2(-2z) + 6/ = 6

or / = 3 A. Thus the dependent source has complicated matters only to the extent of
requiring an extra equation, the first of (3.2).

As another example, let us find the voltage v in Fig. 3.5. Applying KirchhofTs
current law to the currents leaving the top node, we have

-4 + f, - 2/, + y = (3.3)

Also, by Ohm's law we have

v
'-=-6 (3.4)

Substituting (3.4) into (3.3) yields

-4-* + *-0
otv= 12 V.

zn
-V^l

FIGURE 3.5 Another dependent source example

52 Dependent Sources Chap. 3

EXERCISES
3.2.1 In Fig. 3.3, let R =
{
3 Q, R2 = I Q, and /, = 2 A. Find i
g . Ans. 7 A
3.2.2 In Ex. 3.2.1, find the resistance seen by the source looking in terminals a-b.
Ans. § Q.

3.2.3 In Fig. 3.3, let /?, = R2 = r\j Q. Find the resistance seen by the source.
Ans. — j'-jj Q

3.3 OPERATIONAL AMPLIFIERS

A logical question at this point might be, How do we obtain dependent sources? One
answer is that they arise as parts of equivalent circuits of electronic devices operating
under certain conditions. Another answer is that they can be deliberately constructed
by means of certain electronic devices in conjunction with passive elements.
We shall not be interested here in undertaking a study of electronic devices.
However, there is one such device that is extremely useful in the construction of
dependent sources and whose ideal mathematical model is both simple and elegant.
This is the operational amplifier, or op amp, the ideal model of which we shall consider
in this section.
The symbol that we shall use for an operational amplifier is shown in Fig. 3.6. The
op amp is a multiterminal device, but for simplicity we shall show only the three
terminals indicated. Terminal 1 (marked — ) is the inverting input terminal, terminal 2
(marked +) is the noninverting input terminal, and terminal 3 is the output terminal.
The purposes of the terminals that are not shown include, in general, dc power supply
connections, frequency compensation terminals, and offset null terminals. We shall
not be interested in discussing these other terminals here, but the interested student
may find their purposes and how they are used in any op amp user's manual.
Operational amplifiers are commonly available in integrated circuit form and are
normally fabricated in packages having to4opamps.
8 to 14 terminals and containing 1

The operational amplifier has many characteristics that are important to designers,
but the ideal model of the op amp has only two properties that the circuit analyst needs
to know: that the currents into both input terminals are zero and that the voltage
between the input terminals is zero.
As an example of a an operational amplifier, let us consider Fig. 3.7.
circuit with
It is and the voltage v 3 considering v g to be a known
desired to find the current /'
,

generator voltage. Let us write KVL around the loop abca through the source. Since
the voltage across terminals a and b is zero, we have

v 1
~v = g (3.5)
Chap. 3 Dependent Sources 53

+ v2

2A 9 A
in
-o3

FIGURE 3.6 Operational amplifier FIGURE 3.7 Circuit containing an op amp

or i\ = vg . Applying KCL at node b and noting that the current into the negative
terminal of the op amp is zero, we have

(3.6)
1

or v 2 = —2v = — 2vg 1
. Next, KVL around loop cbdc through the 9-fi resistor yields

— v, + v2 + v3 =
or

v3 = v i
-v = 2 3v g (3.7)

Finally. Ohm's law yields

i= v1 _3vt _v
9 " 9 "3a

Thus if vg = 12 cos 107, for example, we have i = 4 cos 10/. Also, we note that KCL
is not valid at the bottom node c. This is because node c (referred to as ground, to be
discussed in Chapter 4) is connected to the unshown power supply terminals.

As a by-product of this example, let us note that i

g
= (the current into an input
lead of the op amp) and v3 = Thus we may draw an equivalent circuit,
3v g from (3.7).

insofar as vg i g v 3 and
, , , i are concerned, as shown in Fig. 3.8. Analysis of the equiva-

lent circuit yields exactly the same v 3 and i for the same v g and i g as in the case of Fig.
3.7. Thus the op amp has been used to obtain a controlled source with a gain of 3. (In
this case the controlled source is a VCVS, in which v g controls v 3 .)

ln=0

9n

FIG U R E 3.8 Circuit equivalent to that of Fig. 3. 7

54 Dependent Sources Chap. 3

Before we leave this section, let us observe that the op amp in Fig. 3.7 is being
operated in a feedback mode. That is, the output v 3 at node d is fed back to the
inverting input terminal through the 2-Q resistor. A practical op amp is a very high-
gain device and is generally never used without feedback. In cases when the feedback
is to one input terminal rather than to both, it is always to the inverting terminal, for
the simple reason that otherwise the op amp will not work. We are not interested here
in the reason for this, which is a consequence of the op amp's design. The interested
student, in all probability, will have occasion to thoroughly study op amps and their
construction in a later course in electronic devices.

3.4 AMPLIFIER CIRCUITS

Equations (3.6) and (3.7), which were the basis for the VCVS circuit of Fig. 3.8, are

independent of the current /in the 9-fi load of Fig. 3.7. This is true in general of VCVS
circuits of this type. To see this, let us consider the circuit of Fig. 3.9. The voltage v 2 is

the output voltage of the op amp and, as we shall see, is a function only of the input
voltage v l
and the two resistors.
Since there is no voltage across the input terminals of the op amp, we have

vba = v,. Also KVL around the loop abca containing v 2 yields

Vba + Vbc + *>2 =

or

v hr = v,„ — v, = v, — v,

Therefore applying KCL at node b results in

*!l_I_
v j
- v*
=
^1 ^2
Solving for v 2 , we have
v2 = nv x
( Mr
where

H=\
is therefore a VCVS with gain //. Since it is a three-terminal network
Figure 3.9
(the output and input share a common terminal) and the currents are zero into the
input leads of the op amp, an equivalent circuit may be drawn as in Fig. 3.10. We note
that since i?, and R 2 are nonnegative, we have fi 1. >
A special case of Fig. 3.9 is the case R 2 = (short circuit) and R = oo (open {

circuit), shown in Fig. 3.1 1. This circuit has = 1, or v 2 = v u and is called a voltage
fj,

follower; that is, v 2 follows v v It is also called a buffer amplifier because it may be used
 +

Chap. 3 Dependent Sources 55

H
'
+
+ o-

v,
«v,
Ri

FIGURE 3.9 VCVS FIGURE 3.10 Circuit equivalent to Fig.

3.9

to isolate, or buffer, one circuit from another. (The voltages at the two pairs of
terminals are the same, but no current can flow from one pair to the other.)
Let us consider next the circuit of Fig. 3.12. Applying KCL at the inverting input
terminal of the op amp, we have

Ri Ri
or

(3.10)

(Rec all tha t_the_¥o ltage acros input terminals of the op amp
are zero.)

^\AA^

vw -Q +
,—

v,

-o
—

F I GURE 3 . 1 1 Voltage folio wer FIGURE 3.12 Inverter

This circuit is called an inverter because the polarity of v 1 is opposite that oiv x ,
as
seen in (3.10). It is also a VCVS, but in this case the input current i 1
is not zero, being
given by

(3.H)
*.

An equivalent circuit shown in Fig. 3.13.is

By(3.1 1) v RJi. and

x
=
thus we may eliminate v t
in Fig. 3.13, resulting in another
equivalent circuit, shown in Fig. 3.14. This circuit is evidently a CCVS, since the
voltage v 2 is controlled by the current i,.
56 Dependent Sources Chap. 3

+ ^ +

FIGURE 3.1 3 Equivalent circuit of the inverter

-o +

R-,i
Z'l v2

FIGURE 3.14 CCVS

We may also obtain dependent current sources from Fig. 3.12, which we redraw as
shown in Fig. 3.15. Since there is no current into the op amp terminals, we have

(3.12)
Ri

Insofar as terminals 1, 2, 3, and 4 are concerned, an equivalent circuit is then that of

Fig. 3.16, which is a CCCS with a gain of 1.

'i A AA^

+ o- vw

FIGURE 3.1 5 inverter circuit redrawn

R2

FIGURE 3.16 CCCS

Chap. 3 Dependent Sources 57

Finally, substituting for/', in (3.12), we may redraw Fig. 3.16 as a VCCS, shown in
Fig. 3.17. In this case, g = 1/7?,.

v,

— o-

FIGURE 3.17 VCCS

EXERCISES
3.4.1 If in Fig. 3.7 v g = 18 V, find v u v 2 and/. , Ans. 18 V, -36 V, 6 A
3.4.2 If in Fig. 3.12;, = 2 A, t- 2 = -10 V, and tf, = 2 ft, find R 2 and v x .

Ans. 5 ft, 4 V
3.4.3 If in Fig. 3.15 the terminals 3-4 are open, find i?, and ?;
43 if v t = \2V,i l
= 3 A,
and R2 = 2 ft. Ans. 4 ft, 6 V

PROBLEMS
3.1 Find the power delivered to the 8-ft resistor in the circuit shown.

10 V 5 V

PROBLEM 3.1

3.2 Find y in the circuit shown if R = 4 ft.

58 Dependent Sources Chap. 3

PROBLEM 3.2

3.3 Find R in the figure for Prob. 3.2 so that v = 2 V.

3.4 Find R in the figure for Prob. 3.2 so that ij = A. 1

3.5 Find vj and i in the circuit shown if R = 4 fi.

+ V, -

PROBLEM 3.5

3.6 Find i? in the figure for Prob. 3.5 so that v r = —4 V.

3.7 Find i in the circuit shown if i
g
= 5e~' A and R = 4Q.

PROBLEM 3.7

3.8 Find i
g in the figure for Prob. 3.7 if R = 2 Q and / = 3 cos / A.

3.9 In the circuit given, show that

*--*&+#
(This circuit is called a summer, since the output voltage is the negative of a
weighted sum of the input voltages. Note that the result is independent of the
output connections at terminals a-b.)
Chap. 3 Dependent Sources 59

PROBLEM 3.9

3.10 In the figure for Prob. 3.9, connect a resistance of R = 5 kQ across terminals
a-b and find the resulting current iab (from a to b) if u, = 10 V, v 2 = 5 V,
^0=2 kfi, /?! = 1 kfi, and Rz = 2 kfi.

3.11 Tn the circuit given, show that regardless of the load at terminals c-d, we have

A\
' 1= ^' 2

Ri

oC +

PROBLEM 3.11

3.12 In the figure for Prob. 3.1 1, let R =

t
/? 2 and connect a resistor 7? between ter-
minals c-d. Show that the resistance seen at input terminals a-b is R ab = —R.
(The figure for Prob. 3.11 thus converts positive resistance to negative resistance.)

3.13 Find i in the circuit shown if R2 = 2/?, = 1 kQ.

PROBLEM 3.13
60 Dependent Sources Chap. 3

3.14 Find R 2 /Ri in the figure for Prob. 3.13 so that i = 2 cos At mA.
3.15 Find i in the circuit shown if R = 6 Q and u g = 5 cos 10/ V.

R
-vw 3A< zni

PROBLEM 3.15

3.16 In the figure for Prob. 3.15, find v g so that R = 12 Q. and i = 100 mA.
 **'
4
ANALYSIS METHODS

In Chapter 2 we considered methods of analyzing simple circuits, which we recall are

those that may be described completely by a single equation. The analysis of more
general circuits entails the solution of a set of simultaneous equations, as we shall
see in this chapter. As an example, the reader may have noticed that Prob. 2.23
required two equations in its solution. Also, most of the circuits of Chapter 3 generally
involved more than one equation, but the equations were of a type that were easily
solved.
In this chapter we ways of formulating and solving the
shall consider systematic
equations that arise in the analysis of more complicated circuits. We shall consider two
general methods, one based primarily on KirchhofT s current law and one on Kirch-
hoffs voltage law. As we shall see, KCL generally leads to equations in which the
unknowns are voltages, whereas KVL leads to equations in which the unknowns are
currents.
It should be evident from our work in previous chapters that a complete analysis
of a circuit can be performed by finding a relatively few key voltages and/or currents.
For example, in a simple circuit consisting of a single loop, a key variable is the
current, for if we know the current, we may find every voltage around the loop, and,
of course, the current around the loop is the current in every element.
In Sec. 4.1 we shall discuss the case where the selected unknowns are voltages.
Quite naturally, our choice of voltages should lead to a set of independent equations.
This technique will be referred to as nodal analysis. In Sec. 4.5 we shall consider mesh
analysis, in which the unknowns are currents.
In this chapter we shall discuss the techniques for selecting the voltages or currents
to befound and the formulation of the circuit equations. The justification of the
methods in the general case will be left until Chapter 6.

61
62 Analysis Methods Chap. 4

4.1 NODAL ANALYSIS

In this section we shall consider methods of circuit analysis in which voltages are the
unknowns to be found. A convenient choice of voltages for many networks is the set of
node voltages. Since a voltage is defined as existing between two nodes, it is convenient

to select one node in the network to be a reference node or datum node and then
associate a voltage or a potential with each of the other nodes. The voltage of each of
the nonreference nodes with respect to the reference node is defined to be a node
voltage. It is common practice to select polarities so that the node voltages are positive
relative to the reference node. For a circuit containing N nodes, there will be TV — 1
node some of which may be known, of course, if voltage sources are present.
voltages,
Frequently the reference node is chosen to be the node to which the largest number
of branches are connected. Many practical circuits are built on a metallic base or
chassis, and usually there are a number of elements connected to the chassis, which is
often then connected to the earth. The chassis may then be called ground, and it
becomes the logical choice for the reference node. For this reason, the reference node
is frequently referred to as ground. The reference node is thus at ground potential or

zero potential, and the other nodes may be considered to be at some potential above
zero.
Since the circuit unknowns are to be voltages, the describing equations are
obtained by applying KCL at the nodes. The currents in the elements are proportional
to the element voltages, which are themselves either a node voltage (if one element
node ground) or the difference of two node voltages. For example, in Fig. 4.1 the
is

reference node is node 3 with zero or ground potential. The symbol shown attached to
node 3 is the standard symbol for ground. The nonreference nodes 1 and 2 have node

voltages t>i and v2 . Thus the element voltage v 12 with the polarity shown is

v X2 = v x
-v 2

The other element voltages shown are

v 13 = v, -0 = v,

and

hz = v2 — = v2

These equations may be established by applying KVL around the loops (real or

imagined).
The application of KCL results in an equation relating node voltages. Clearly,
simplification in writing the resulting equations is possible when the reference node is

chosen to be a node with a large number of elements connected to it. As we shall see,

however, this is not the only criterion for selecting the reference node, but it is fre-

quently the overriding one.

Since we are going to apply KCL, it seems that the simplest networks to consider
 £

Chap. 4 Analysis Methods 63

FIGURE 4.1 Reference and nonreference

nodes

are those whose only sources are independent current sources. This is not always true,
as we we shall begin with an example of this type. In the network shown
shall see. but
in Fig. 4.2(a), there are three nodes, dashed and numbered as shown. [This is
easier to see in the redrawn version of Fig. 4.2(b).] Since there are four elements
connected to node 3, we select it as the reference node, identifying it by the ground
connection shown.

r)—Wv— R2

Q Q>
c z--T.-i^rzrz=:
(a)

FIGURE 4.2 Circuit containing independent current sources

Before writing the node equations, consider the element shown in Fig. 4.3, where
v x
and v 2 are node voltages. The element voltage v is given by

v = v,

and thus by Ohm's law we have

= V\ -1>2
i
R R
or

i =G(v — l
v2)

where G= \jR is the conductance. That is, the current from node 1 to node 2 is the
difference of the node voltage at node I and the node voltage at node 2 divided by the
resistance R, or multiplied by the conductance G. This relation will allow us to rapidly
write the node equations by inspection directly in terms of the node voltages.
64 Analysis Methods Chap. 4

Now returning to the circuit of Fig. 4.2, the sum of the currents leaving node 1

must be zero, and this results in the equation

'i + ii ~ i
g i
=
In terms of the node voltages, this equation becomes

G ]
vl +G 2 (v 1
- v2) - i gi =
We could have obtained this equation directly using the procedure of the previous
paragraph. Applying KCL at node 2 in a similar manner, we obtain

-i 2 + i3 + i g2 =
or

G2 (v 2 — ^i) + G> + 2 i
g2 =
Again, the G"s are the conductances (reciprocals of the 7?'s).

We could have equated the sum of currents leaving the node to the sum of currents
entering the node. Had we done so, the terms i gl and i gl would have appeared on the
right-hand side:

G l
v l
+ G (v - v =
2 l 2)
i gl

G 2
{v 2 -v +Gv = x ) 3 2 -i g2

Rearranging these two equations results in

+ G )v, - G v =
(G, 2 2 2 i gl (4.1)

-G v + (G + G )v =
2 x 2 3 2 -i g2 (4.2)

These equations exhibit a symmetry that may be used to write the equations in
the rearranged form by inspection. In (4.1) the coefficient of v is the sum of con- x

ductances of the elements connected to node 1, while the coefficient of v 2 is the

negative of the conductance of the element connecting node to node 2. The same 1

statement holds for (4.2) if numbers 1 and 2 are interchanged. Thus node 2 plays
the
the role in (4.2) of node 1 in (4.1). That is, it is the node at which KCL is applied. In
each equation the right-hand side is the current from a current source which enters
the corresponding node.
In general, in networks containing only conductances and current sources, KCL
applied at the kth node, with node voltage v k , may be written as follows. In the left

member the coefficient of v k is the sum of the conductances connected to node k, and
the coefficients of the other node voltages are the negatives of the conductances
between those nodes and node k. The right member of the equation consists of the net
current flowing into node k due to current sources.
To illustrate the process, consider Fig. 4.4, which is a portion of a circuit. The
dashed lines indicate connections to nodes other than node 2 (labeled v 2 ). At node 2 we
Chap. 4 Analysis Methods 65

1
F I GURE 4 . 3 Single element FIGURE 4.4 Part of a circuit

have, using the shortcut procedure,

-<?!«, + (G, +G +G 2 3
)v 2 -G 2
v3 -G 3
i>
4
= i gl - i
g2 (4.3)

This may be checked by applying KCL in the usual way, equating currents leaving
node 2 to those entering node 2. The result is

G x
{v 2 - i\) +G 2 (v 2
- v3) +G 3
(v 2 - vt ) + /, 2 = / gi

Rearranging this result leads to (4.3).

4.2 AN EXAMPLE

To illustrate the method of nodal analysis, let us consider the circuit of Fig. 4.5. We
have taken the reference node as shown and labeled the nonreference nodes as v,, v 2 ,

and v 3 . We note that the conductances are specified for the resistors.
Since there are three nonreference nodes, there will be three nodal equations. At
node v u using the shortcut method, we have

4v, — V2 = 2 (4.4)

I
15
2U
^\r\r\j-

7A
© 313- 5A 3U IU 4U- 7A

FIGURE 4.5 Example of a circuit

66 Analysis Methods Chap. 4

The sum of the conductances at node w, is 3 + 1 = 4, the conductance between nodes

1 and 2 is 1, the conductance between nodes and 3 is 0, and the net current into the
1

node through the sources is 7 — 5 = 2. If we had used the conventional KCL method
at v i, we would have obtained

l(v, — v2) + 3v, + 5 = 7

which is equivalent to (4.4).

At nodes v 2 and v3 , we have

-v x + 6v — 2 2v 3 = 5
(4.5)
— 2v + 2 lv 3 = 17

We may solve (4.4) and (4.5) for the node voltages using any one of a variety of
methods for solving simultaneous equations. Two such methods are Cramer's rule,
which employs determinants, and Gaussian elimination. For the reader who is not
familiar with these two methods, a complete discussion is given in Appendices A
and B.
Using Cramer's rule, we first find the coefficient determinant, given by

4 -1
-1 6 145

-2
Then we have
2-1
5 6-2
v, = 17 -2 7 145
1 V
145 145

4 2
-1 5 -2
= 17 7 290
v, 2V
. 145 145

and

4 -1 2
-1 6 5

-2 17 435
= 3V
145 145

Now that we have the node voltages we may completely analyze the circuit. For
example, if we want the current i in the 2-15 element, it is given by

i = 2(v 2 - v3) = 2(2 - 3) = -2 A

Chap. 4 Analysis Methods 67

Equations (4.4) and (4.5) are symmetrical in a way that facilitates the writing of the
equations. This symmetry also is present in the general case. For example, the coeffi-
cient of v 2 in the first equation is the same as that of ?>, in the second equation. Also,
the coefficient of r, in the first equation is that off, in the third equation. Finally, the
coefficient of v 3 in thesecond equation is that of v 2 in the third equation. These
results follow from the fact that the conductance between nodes and 2 is that 1

between 2 and 1, the conductance between nodes and 3 is that between 3 and 1, etc. 1

This symmetry shows up also in the coefficient determinant A. Its diagonal ele-
ments. 4. 6. and 7. are the sums of the conductances connected to the three non-
reference nodes, and the off-diagonal elements are symmetrical about the diagonal.
The latter elements are, of course, the negatives of the conductances between nodes.

EXERCISES
4.2.1 Using nodal analysis, find v and v 2 in Fig. 4.2
x
(a) if R =
1
2CI, R2 = 2 Q,
R =4Q,
3 /„i = 2 A, and i <?2 -4 A. Ans. 7 V, 10 V
4.2.2 Repeat Ex. 4.2.1 with i g2 = 0. Check by using current division
Ans. 3 V, 2 V
4.2.3 Using nodal analysis, find v u v z and v } , . Ans. 8 V, -3 V, 5 V

EXERCISE 4.2.3

4.3 CIRCUITS CONTAINING VOLTAGE

SOURCES

At first glance it may seem that the presence of voltage sources in a circuit complicates
the nodal analysis. We can no longer write the equations using the shortcut method
because we do not know the currents through the voltage sources. However, nodal
analysis is no more complicated and in many cases is even easier to apply when
voltage sources are present, as we shall see.
To illustrate the procedure, let us consider the circuit of Fig. 4.6. We have labeled
the nonreference nodes as v u v 2 v 3 and v, and have taken the sixth node as
,

reference, as indicated. Again, the resistors are labeled by their conductances.

68 Analysis Methods Chap. 4

FIGURE 4.6 Circuit containing voltage sources

Since there are five nonreference nodes, we need five equations. Without writing
any KCL equations we may note that we have, by inspection,

(4.6)
v< — Vd = V «2

We thus need only three KCL equations. To be systematic and at the same time elimi-
nate the need to know the currents in the voltage sources, let us enclose the voltage
sources by dashed lines as shown in Fig. 4.6. We may think of these surfaces as
generalized nodes or, as some authors prefer, super nodes. We recall that KCL holds
for such a generalized node as well as for an ordinary node. We have, therefore, two
generalized nodes and two regular nodes, labeled v 2 and v 3 a total of four nodes. ,

Thus we need only three KCL equations, which together with (4.6) constitute the
required set of five equations in the node voltages.
To complete the formulation of the nodal equations, let us apply KCL at nodes
v2 and v 3 and the generalized node containing v g2 . The first two are obtained as
before, resulting in

(C?, + G + G )v — G v — G v — G v =
2 4 2 x x 2 3 4 s
(4.7)
{G + G + G )v — G v - G i\ =
2 3 5 3 2 2 5

Finally, equating to zero the currents leaving the generalized node, we have

G4 (v -
s v2) +G 5
(v 4 - v3) + G 6 v5 = (4.8)

The circuit is analyzed by simultaneously solving (4.6), (4.7), and (4.8).

As another example, let us consider the circuit of Fig. 4.7, which contains an
independent voltage source and a dependent current source. Anticipating that the
Chap. 4 Analysis Methods 69

FIGURE 4.7 Circuit with independent and dependent sources

presence of the voltage source reduces the number of unknowns by one, we have
node at the top
labeled the left v
g
. the reference node being chosen as indicated. The
unknown node voltages are i\. v2 , and v 3 . Applying KCL at these nodes, we have

(G, -f G2 + G — G v — G v - G,v =
3 )i'i 2 2 3 3 g

-G 2 v + (G + G )v + P(v -v ) =
} 2 5 2 t 3

-G v + (G + G4 )v = fi{v, - v
3 t 3 3 3 )

These equations may now be solved for the unknown node voltages (v g is con-
sidered to be known). We note that the presence of the dependent source has destroyed
the symmetry that was present in the equations of the previous examples. This is true
in the general case involving dependent sources.
As a final example, let us consider Fig. 4.8, which contains independent and
dependent current and voltage sources. To simplify matters we have chosen the
reference node as shown and labeled the nonreference nodes, which in this case may
all be expressed in terms of only the two unknowns v x and i y (We have used Ohm's

FIGURE 4.8 More complex circuit

70 Analysis Methods Chap. 4

law to obtain v 2 and v 3 was given by v 2

,
— 3/, = — 5/,.) Wetherefore need only two
equations. Since we have three nodes, the regular node and the two generalized nodes,
there are precisely two independent nodal equations.
At node ,
^ 4 we have

_^L4_
+ Q _i_
2 ~ v* ~ (~ 5i y) - o
2 ^ '

and at the generalized node containing the dependent voltage source we have

-2/ v —2— .. *,„

2v x + —5i
,
y
— (2 — vj
~5h-(2-v x = )
*,
1 1

These equations simplify to

•3^+ 10/, = -22

— v — 8/ = 4
r v

which have solution

^ = 4V, 1 A
We may now find all the currents and voltages in the circuit.

EXERCISES
4.3.1 Using nodal analysis, find yj if the element x is a 12-V independent voltage
source with the positive terminal at the top. Ans. —4 V

6V

2ft 3n ,Vo

6A 2A

EXERCISE 4.3.1

4.3.2 Repeat Ex. 4.3.1 if element x is a 0.5-A independent current source directed
downward. Ans. ~3V
4.3.3 Repeat Ex. 4.3.1 if element x is a dependent voltage source of 6 Wj V with the
positive terminal at the top. Ans. — V
1
Chap. 4 Analysis Methods 71

4.4 CIRCUITS CONTAINING OP AMPS

Nodal analysis very often is the best method of analysis when a circuit contains op
amps, because in electronic circuits the reference node is usually shown as grounded
and all other elements connected to the reference node are often shown individually
grounded. Thus the nodes are easily identified for the nodal method, but the loops are
not so easily visualized. Therefore a method based on loops, such as we shall consider
in the next section, is not so easy to apply. Also, quite often, only a relatively few nodal
equations arc required.
As an example, let us consider the VCVS of Fig. 3.9, redrawn as shown in Fig. 4.9.
The reference node is shown as grounded, so that the voltages v and v 2 of Fig. 3.9 are x

now node voltages, as shown. Since the voltage is zero between the input terminals of
the op amp and the currents are zero into the input terminals, we see that v 3 = «, and
that the node equation written at node v 3 is

i ,
t'i - v2
= (4.9)
R R,

From this we find

vz = (l + !*)», = flV, (4.10)

which is the result given earlier in Sec. 3.4.

FIGURE 4.9 VCVS

As another example, let us consider the circuit of Fig. 4.10. We shall take the

ground to be the zero voltage reference point so that the inverting input terminal of
the op amp is also at zero potential, as labeled. We shall consider the voltage ?;, to be a
known input voltage and solve for the output voltage v 2 At node v 3 KCL yields .
,

(G, +G +G 2 3 -f G 4 )v — G
3 l
v 1
—G 4v2 =
using the shortcut procedure of Sec. 4.1. The sum of the currents entering the inverting
input node of the op amp is given by

G 3 v 3 + G s v2 =
72 Analysis Methods Chap. 4

FIGURE 4.10 Circuit containing an op amp

Eliminating v 3 from these two equations results in

(G, +G 2

from which we have

-G G x 3 v x

G S
{G, G + G4 ) + G
3 3 G<

It is always fruitful to write nodal equations at the inverting input nodes of the op

amps, as we have done in this example. However, one generally avoids writing nodal
equations at output nodes of op amps because it is difficult to find the current out of
an op amp. There is no current into the input terminals, but because there are other
terminals not shown, the output terminal carries a current.

EXERCISES
4.4.1 Solve Ex. 3.4.1 using nodal analysis.

4.4.2 Solve Prob. 3.9 using nodal analysis.

4.4.3 Find / in the circuit shown. Ans. 6 cos At A

EXERCISE 4.4.3
Chap. 4 Analysis Methods 73

4.5 MESH ANALYSIS

In the nodal analysis of the previous sections we applied KCL at the nonreference
nodes of the circuit. We shall now consider a method, known as mesh analysis, or loop
analysis, in which KVL is applied around certain closed paths in the circuit. As we
shall see. in this case the unknowns generally will be currents.
We shall restrict ourselves in this chapter to planar circuits, by which we mean
circuits that can be draw n on a plane surface in a way such that no element crosses any
other element. In this case, the planeis divided by the elements into distinct areas, in

the same way that the wooden or metal partitions in a window distinguish the window
panes. The closed boundary of each area is called a mesh of the circuit. Thus a mesh is
a special case of a loop, which we consider to be a closed path of elements in the
circuit passing through no node or element more than once. In other words, a mesh
is a loop that contains no elements within it.

As an example, the circuit of Fig. 4.11 is planar and contains three meshes,
identified by the arrows. Mesh contains the elements R R 2 R 3 and v gi mesh 2
1
x
, , , ;

contains R 2 i? 4 v g2 and R s and mesh 3 contains R 5 v g2 R 6 and R 3

, , , ; , , , .

In the case of nonplanar circuits (that is, those that are not planar), we cannot
define meshes. Thus in the analysis using KVL, the closed paths are loops. The
procedure is the same, of course, but the equations are not as easily formulated. We
shall consider this more general case in detail in Chapter 6.

As an example in which KVL is applied, let us consider the two-mesh circuit of

Fig. 4.12.The element currents are /„ /,, and I 3 Applying KVL around the first .

mesh (containing v gi ), we have

/?,/, +R 3 I3 = v g\ (4.11)

Similarly, around the other mesh we have

R 2 I2 -R 3 I3 = -v g2 (4.12)

We define a mesh current as the current which flows around a mesh. The mesh

FIGURE4.11 Planar circuit with three meshes FIGURE 4.1 2 Circuit with two meshes
74 Analysis Methods Chap. 4

current may constitute the entire current in an element of the mesh, or it may be
only a portion of the element current. For example, in Fig. 4.12 the currents and i1

i2 are mesh currents, with the directions as shown. The element current is the mesh
current in R^ and R 2 , but the element current in R 3 is the composite of two mesh
currents.
In general, element currents are algebraic sums of mesh currents. This is illustrated
in Fig. 4.12 since the element current in R t
is

/i = h
that in R2 is

and that in R 3 , by KCL, is

h= It —I = 2 /, - i2

Using these results, we may rewrite (4.11) and (4.12) as

iJj/j +R 3 (i 1
- i 2) = v gl
(4.13)
R2 i2 —R 3 (i l
— i2) = —v g2
These are the mesh equations of the circuit.

There is also a shortcut method of writing mesh equations which is similar to the
shortcut nodal method of Sec. 4.1. Rearranging (4.13) in the form

(/?, +R 3 )h - R = v el 3i2

-R 3 i, + (R 2 + R )i = -v g2
3 2

we note that in the first equation, corresponding to the first mesh, the coefficient of the
first current is the sum of the resistances in the first mesh, and the coefficients of any
other mesh current is the negative of the resistance common to that mesh and the first

mesh. The right member of the first equation is the algebraic sum of the voltage
sources driving the first mesh current in its assumed direction. Replacing the word

firstby the word second everywhere it appears in these last two sentences will describe
the second equation, and so forth. This shortcut procedure is a consequence of
selecting all the mesh currents in the same direction (clockwise in Fig. 4.12) and
writing KVL as the meshes are traversed in the directions of the currents. Of course,
the method applies only when no sources are present except independent voltage
sources.
As another example, let us return to Fig. 4.1 and define /,, i 2 and i 3 as the mesh 1 ,

currents shown in meshes 1, 2, and 3, respectively. Then applying the shortcut method
to mesh 1, we have

(R 1
+R + 2 i? 3 )/, -R 2i2
-R 3i3
= v gi
Chap. 4 Analysis Methods 75

This result may be checked by applying KVL to mesh 1, resulting in

i?,/, + R 2 (i - t
f 2) + /? 3 (/"i - h) = v aX

The two results are evidently the same.

Applying KVL to meshes 2 and 3 yields, in the same manner,

-/? 2 f, { + R + * )»2 - *s*' =

(R 2 4 5 3 ~v,i
-R 3it
- R + (R + R + R =
5
i2 3 5 6 )i 3 v g2

The analysis is completed by solving the three mesh equations for the mesh currents.
The same symmetry is present in the mesh equations as was noted in the nodal
equations. If Cramer's rule is to be used for solving for the currents, then the coeffi-
cient determinant to be calculated is

R + R2 + R
t 3
-/?, -R,

A = Rl + ^4 + ^5 — ^5
-R, —R 5 R +R + 3 s ^6

The diagonal elements are the sum of the resistances in the meshes, and the off-
diagonal elements are the negatives of the resistances common to the meshes corre-
sponding to the row and column of the determinant. That is, —R 2 in row 1, column 2
or in row
2, column 1 is the negative of the resistance common to meshes 1 and 2, etc.
Thus the determinant is symmetric about its diagonal.

EXERCISES
4.5.1 Using mesh analysis, find fj and i 2 in Fig. 4.12 if /?j = 1 Q, R2 = 2 Q,
R 3 = 4 Q, v gl = 21 V, and v g2 = 14 V. Ans. 5 A, 1 A
4.5.2 Repeat Ex. 4.5.1 with v g i = 0. Check by using voltage division.
Ans. 9 A, 6 A
4.5.3 Using mesh analysis, find /'[ and ;'
2 . Ans. 8 A, 5 A

EXERCISE 4.5.3
76 Analysis Methods Chap. 4

4.6 CIRCUITS CONTAINING CURRENT

SOURCES

As in the case of nodal analysis of circuits with voltage sources, mesh analysis is

easier if there are current sources present. To illustrate this point, let us consider the
circuitof Fig. 4.13, which has two current sources and a voltage source.
With the mesh currents /,, i 2 and i 3 chosen as shown, it is clear that we need three
,

independent equations. Not all of these, however, have to be mesh equations. The
presence of the two current sources provides us with two constraints which we may
obtain by inspection:

h = —ie \

(4.14)

We need, therefore, only one more equation. Since it will have to come from
KVL, we need to select a closed path in which all the voltages are easily obtained.
That is, we need to avoid the current sources since their voltages are not readily
obtained.
If we imagine for a momenttwo current sources are removed, that is,
that the
opened, then we shall have two we already have two equations so
less meshes. But
there will still be enough meshes left for the required number of equations. Moreover,
the loops left (they may not be meshes) will have only resistors and voltage sources in
them, and therefore KVL is easily applied. We must stress that we are not taking the
current sources out. We are only imagining them out for a moment in order to locate
the loops around which KVL is to be applied.
Returning to Fig. 4.13 and imagining the current sources open for a moment, we
see that we have only one loop left, namely the loop containing v g3 /?,, R 2 and R 3 , , .

Applying KVL to this loop, we have our third equation,

/?,(/,- i2) +R 2 (i 3
- i2) +R 3i 3
= v e3 (4.15)

The analysis of the circuit can now be completed by solving (4.14) and (4.15).
As a final example, let us apply loop analysis to the circuit of Fig. 4.8, which was
analyzed previously by the nodal method. The circuit is redrawn in Fig. 4.14. There
are four meshes and two constraints to be satisfied by the controlling variables v x and
i
y We shall need, therefore, six equations.
We may choose mesh currents and obtain their relations to the two current sources
as we did earlier in Fig. 4.13. However, to illustrate the use of loop currents and also
we have chosen as the unknowns currents/,, i 2 i 3
to simplify the resulting equations, , ,

and 4 2 , as shown The selection we have made results in a single-loop

in the circuit.
current through the independent current source and through the element whose cur-
rent controls a dependent source. Thus the constraints are simple equations. Before
Chap. 4 Analysis Methods 77

FIGURE 4. 13 Circuit with two current sources FIGURE 4.14 More complex circuit
and one voltage source

applying KVL, we may write, by inspection of the circuit,

/, =9
h + *2 +U= 1v x

i3 = i
y

From these results we may express all the loop currents in terms of v x and i
y . The
result is

U =9

(4.16)
/, = z,

,
_5v£ _o

As yet we have not We need two more equations

written a single loop equation.
since there are basicallytwo unknowns, v x and iy By imagining for the moment the .

current sources as open, we see the two loops to which we shall apply KVL. They are
abecdo. and bceb. The respective loop equations are

2/ 2 + 2 + 2/ 3 + 3/, + 1(i'i + Q=
and
l(/ 3 + /4 ) +2+ 2/ 3 =
78 Analysis Methods Chap. 4

Substituting (4.16) into these equations, we have

3v
2
+ 5/,= -11

^4-3/ =7
The solution is given by

vx = A\, i=-lA
which checks with the result obtained earlier in Sec. 4.3. Clearly the nodal analysis
performed earlier was a simpler method.
The loop currents may now be found from (4.16), completing the analysis. Any
other current or voltage in the circuit may be readily obtained also.

EXERCISES
4.6.1 Using the methods of this section, find i u i2 and ; 3 , in Fig. 4.13 if R = t
1 Q,
R 2 = 3 Q, R 3 = 4 Q., i gl = 2 A, i gl = 8 A, and v g3 = 24 V.
Arts. -5 A, -2 A, 3 A
4.6.2 Repeat Ex. 4.6.1 if v g3 = 0. Ans. -8 A, -2 A,

4.6.3 Repeat Ex. 4.6.1 if i gl = 0. Ans. -A A, 0, 4 A

4.7 DUALITY'

The reader may have noticed a similarity in certain pairs of network equations which
we have considered so far. For example, Ohm's law may be stated as

v = Ri (4.17)

or

i=Gv (4.18)

In the second case we have solved for i, of course, and used the definition G = \jR.
Another way of looking at these equations is to note that the second may be obtained
from the first by replacing v by /, i by v, and R by G. In like manner, the first may be
obtained from the second by replacing i by v, v by /, and G by R.
Similarly, in the case of series resistances R R2 t
, , . . . , R n,
the equivalent resistance
was shown in Sec. 2.3 to be

R =
S
R, +R + 2 ... +R n (4.19)
Chap. 4 Analysis Methods 79

and for parallel conductances G,, G 2

G„, we saw in Sec. 2.4 that the equivalent
conductance was

Gp = G, +G + 2 . . . +G n (4.20)

It is one of these equations may be obtained from the other by interchanging

clear that
resistances and conductances and the subscripts and p. i.e., series and parallel. ,v

There is thus a definite duality between resistance and conductance, current and
voltage, and series and parallel. We acknowledge this by defining these quantities as
duals of each other. That is. R is the dual of G, i is the dual of v, series is the dual of
parallel, and vice versa in each case.
Another simple case of dual equations is

v = (4.21)

and its dual

i = (4.22)

In the general case, an element described by (4.21) is a short circuit, and one described
by (4.22) is an open circuit. Thus short circuits and open circuits are duals. There are
also other dual quantities, as we shall see later.
Every equation in circuit theory has a dual, obtained by replacing each quantity in

the equation by its dual quantity. If one equation describes a planar circuit, then the
other equation describes the dual of the circuit, or the dual circuit. (As the reader may
see in a later course, nonplanar circuits do not have duals.) For example, consider the
circuit of Fig. 4.15. The mesh equations are given by

+R -R
(/?, 2 )i l 2i 2
= vg
(4.23)
-R + (R + 7?
2 i, 2 3 )/ 2
=
To obtain the dual of (4.23), we simply replace the R's by G's, the fs by v's, and v
by /'.
The result is

((7, + G )v, - G v =
2 2 2 i
g
(4.24)
— G2v + {G + G )v =
t 2 3 2

These are the nodal equations of a circuit having two nonreference node voltages, v x

and v z three conductances, and an independent current source i g From our shortcut
, .

procedure for nodal equations we see that G^ and G 2 are connected to the first node,
G2 is common to the two nodes, G 2 and G 3 are connected to the second node, and i
g
enters the first node. Since (7, and i g are not connected to the second node and G 3 is

not connected to the first node, these elements are connected to the reference node.
Such a circuit is shown in Fig. 4.16 and is a dual of Fig. 4.15.
Figure 4.15 may be described as vg in series with R l
and the parallel combination
of R 2 and R 3 . Replacing the quantities in this statement by their duals, we see correctly
80 Analysis Methods Chap. 4

FIGURE 4.15 Two-mesh circuit FIGURE4.16 Dual of Fig. 4.15

that Fig. 4.16, a dual circuit, may be described as i

g
in parallel with G^ and the series
combination of G 2 and G 3
.

In this example we note that nodes are duals of meshes and vice versa. This is true
in general because of the duality between the mesh currents and the nonreference node
voltages. The reference node is the dual of the boundary of the region outside the
circuit, or what we might call the outer mesh.
This last duality (between nodes and meshes) gives us a procedure for finding a
dual of a given circuit. We may place one node of the dual network to be obtained
inside each mesh, thereby assuring us of the correspondence between each nonreference
node and a mesh. Then we may place the reference node outside the circuit (corre-
sponding to the outer mesh). Connecting these nodes together by elements drawn
through the elements of the original circuit assures us of correspondence between
elements. If these last elements drawn are the duals of those they cross, then the
circuit obtained is a dual of the original circuit.
As an example, the circuit of Fig. 4.15 is redrawn using solid lines in Fig. 4.17. The
dashed circuit, superimposed in accordance with the dual circuit procedure, is its dual.
It should be clear that the latter circuit is the same as that of Fig. 4.16.
In the case of independent or dependent sources, the polarity of the dual source is

specified when the dual circuit is obtained from the dual equations, as was the case in
Fig. 4.16. However, in the geometric method of Fig. 4.17, the polarity of i g cannot be
determined since the mesh current directions are not shown and thus are completely
arbitrary.

~<9 ^ / u3

FIGURE 4.17 Two dual circuits

Chap. 4 Analysis Methods 81

If the mesh currents or node voltages are shown in the original circuit, then the
polarity of a source obtained in the dual circuit may be determined from the concept
of driving a node or a mesh. We shall say that a mesh is driven by a source in it if the
polarity of the source is such as to drive the current in the direction of the mesh
current. A node is driven by a source connected to it if the polarity of the source is

such as to apply voltage to the node, i.e., send current toward the node. Thus the dual
source will drive or not drive a mesh or a node in accordance with whether the source
it crosses in the original circuit drives or does not drive the corresponding node or
mesh. This is consistent, as it must be, with the two sets of dual equations of the
circuits.

In general, when we have analyzed one circuit, the numerical values of its voltages
and currents are the same as those of their duals in the dual circuit. When we solve
one circuit, therefore, we have really solved two. As an illustration, the numerical
values of/', and i 2 in Fig. 4.15 are, respectively, the same as those of f, and v 2 in
Fig. 4.16.
As a final note in this chapter, we shall not attempt to give a dual of an op amp.
We may. of course, obtain duals of circuits containing dependent sources which are
equivalent to circuits with op amps.

EXERCISE
4.7.1 Find the dual of the circuit in the figure for Ex. 4.5.3.

PROBLEMS
4.1 Using nodal analysis, find the current / in the circuit shown.

4A

PROBLEM 4.1
 —

82 Analysis Methods Chap. 4

4.2 Using nodal analysis, find the voltage v in the circuit shown,

2A
-0-
in 2ft
-wv- ^\^

6A 2ft in:' 4fi

Q)

PROBLEM 4.2

4.3 Repeat Prob. 4.2 if the 6-A source is replaced by an independent voltage source
having a terminal voltage of 10 V with the positive terminal at the top.

4.4 Using nodal analysis, find v in the circuit shown.

2V
3ft
-0-

2n: (T)3A «vQ 5ft

| © 4A

PROBLEM 4.4

4.5 Using nodal analysis, find v in the circuit shown.

12V 3mA

PROBLEM 4.5
Chap. 4 Analysis Methods 83

4.6 Using nodal analysis, find v in the circuit shown.

2 mA

PROBLEM 4.6

4.7 Using nodal analysis, find v in the circuit shown.

18V

2a
PROBLEM 4.7

4.8 Using nodal analysis, find the power delivered to the 4-Q resistor in the circuit
shown.

PROBLEM 4.8
84 Analysis Methods Chap. 4

4.9 Repeat Prob. 4.8 for the circuit shown.

PROBLEM 4.9

4.10 Using nodal analysis, find i>, in the circuit shown.

12V

6mA 2kn
CD i
mA

PROBLEM 4.10

4.11 Solve Prob. 3.13 using nodal analysis.

4.12 Solve Prob. 4.1 using mesh analysis.

4.13 Solve Prob. 4.2 using mesh analysis.

4.14 Solve Prob. 4.6 using mesh analysis.

4.15 Solve Prob. 4.9 using mesh or loop analysis.

4.16 Using mesh or loop analysis, find v in the circuit shown. The sources x, y, and z
are independent current sources with arrows pointing upward and currents of
10, 15, and 5 A, respectively.

3n

20, 4ft
-A/W-

3fi
-V\A/-

PROBLEM 4.16
Chap. 4 Analysis Methods 85

4.17 Solve Prob. 4.16 if source z is changed to a 12-V voltage source with the positive
terminal at the top.

4.18 Solve Prob. 4.16 using any method if and z are independent voltage sources
x, y,
w nil the positi\e terminals at the top and terminal voltages of 39, 10, and 20 V,
respectively.

4.19 Find i in the circuit shown.

-0-
2mA

|3V
3KQ
|4V
I Kit

PROBLEM 4.19

4.20 Find v in the circuit shown.

4kU
-AA/V—

3mA
25V

4kQ

5mA
© •
3 k n
2kn
6 IOmA

PROBLEM 4.20

4.21 Find i in the circuit shown.

12V

PROBLEM 4.21
86 Analysis Methods Chap. 4

4.22 Find v t
in the circuit shown.

PROBLEM 4.22

4.23 In the circuit shown, find v 2 in terms of R and Ig and show

, that if R= 10 kQ,
then

»2 %-5x I0*lg

(Note: "sfc" means approximately equal to.)

PROBLEM 4.23

4.24 Find v in terms of the node voltages v u v2 and v 3 and the resistances in the
,

circuit shown. (This circuit is a summer, like that of Prob. 3.9, with an additional
voltage source.)

R
"I o-

n2
v 0-
2
-o v o

n3
"3 0-

PROBLEM 4.24
Chap. 4 Analysis Methods 87

4.25 Find v 2 in terms of v\ in the circuit shown.

IkO
-AAA—

2kf! Ikfi
v
i
o aaa^- AAA f
-OVg

I k
4 2kn
5 Ikn

PROBLEM 4.25

4.26 Find v2 in terms of v {

and the conductances in the circuit shown.

G„
Aaa-

"3
AAA^
G2
I O » AAA i
-O u 2

-Wfv-
G,

PROBLEM 4.26

4.27 Find the input resistance R ia , in terms of the other resistances, in the circuit
shown.

PROBLEM 4.27

'4.28 Draw the dual of the circuit of Prob. 4.1, and find v in the dual circuit corre-
sponding to the unknown i in the original circuit.
88 Analysis Methods Chap. 4

*4.29 Draw a circuit and its dual if the mesh equations of the circuit are given by

- 2/ 2 = 4
10/,

-2/, + 8/ 2 - f = 3

-iz + ll/ = -6 3

*4.30 Draw the dual of the circuit of Prob. 4.16. Identify i in the dual circuit corre-
sponding to v in the original circuit and show that they have the same numerical
values.
 5
NETWORK THEOREMS

In the previous chapters we have considered fairly straightforward methods of

analyzing circuits. However, in many cases the analysis can be shortened considerably

by the use of certain network theorems. For example, if we are interested only in what
happens to one particular element in a circuit, it may be possible by means of a
network theorem to replace the rest of the circuit by an equivalent and simpler
circuit.

In this chapter we shall introduce a number of network theorems and illustrate

their use in simplifying the analysis of certain circuits. The theorems are applicable, in
general, to circuits that are linear, a term which we also shall discuss.
Finally, we shall use the network theorems as a motivation for introducing
practical sources, as distinguished from the ideal sources considered thus far. As we
would expect, practical sources are capable of delivering only a finite amount of
power, and we shall devote a section to determining the maximum power that can be
delivered by a given source.

5.1 LINEAR CIRCUITS

In Chapter 2 we defined a linear resistor as one that satisfied Ohm's law

v = Ri

and we considered circuits that were made up of linear resistors and independent
sources. We defined dependent sources in Chapter 3 and analyzed circuits containing
both independent and dependent sources in Chapter 4. The dependent sources that we

89
90 Network Theorems Chap. 5

considered all had describing relations of the form

y = kx (5.1)

where A: is a constant and the variables x and y were either voltages or currents. Clearly,
Ohm's law is a special case of (5.1).
In (5.1) the variable y is proportional to the variable x, and the graph of >> versus x
is a straight line passing through the origin. For this reason, some authors refer to
elements which are characterized by (5.1) as linear elements.
For our purposes we shall define a linear element in a more general way, which
includes (5.1) as a special case. If x and y are variables, such as voltages and currents,
associated with a two-terminal element, then we shall say that the element is linear if
multiplying x by a constant K results in the multiplication of y by the same constant
K. This feature is called the proportionality property and evidently holds for (5.1) since

Ky = k(Kx)

Thus not only is an element described by (5.1) linear, but, in addition, elements
described by relations of the forms

$-«, y = »% (5-2)

are also linearif a and b are nonzero constants.

The op amp is a multiterminal element and is described by more than one

ideal
equation. However, we shall use the op amp only in a feedback mode, as stated in
Chapter 3, and in this case the equivalent circuits which result consist of linear
elements with describing relations like (5.1). Therefore we may add the ideal op amp
to our list of linear elements.
We shall define a linear circuit as one containing only independent sources and/or
linear elements. As examples, all the circuits we have considered thus far are linear
circuits.

The describing equations of a linear circuit are obtained by applying KVL and
KCL, and therefore they contain sums of multiples of voltages or currents. For
example, a loop equation is of the form

a v
x x
+ a2v2 + . . . + a n vn =f (5.3)

where/is the algebraic sum of the voltages of the independent sources in the loop, the
v's are the voltages of the remaining loop elements, and the o's are or ±1. To
illustrate, KVL around the loop shown in the circuit of Fig. 5.1 yields

v l
+ v2 - v3 = v gl - v g2

In this case, a x
= a2 = 1; a3 = —I; the other a's, if any, are all zero; and
Chap. 5 Network Theorems 91

/= v — vg2 We also have

g i
.

v x
= 2ij, v2 = 5/ 2 , v3 = 3/ 6

where /„ is a current elsewhere in the circuit.

The proportionality property of a single linear element also holds for a linear
circuit in the sense that if all the independent sources of the circuit are multiplied by a
constant A', then all the currents and voltages of the remaining elements are multiplied
by this same constant K. This is easily seen in (5.3), which multiplied through by K
becomes
a x
Kv x
+ a 2 Kv 2 + ... + a n Kv„ = Kf
The right member is the consequence of multiplying the independent sources by K. and
for equality to still hold all the v's are multiplied by K. Since the elements are linear,
multiplying their voltages by K multiplies their currents by K.
To illustrate the proportionality property, let us find the current i in Fig. 5.2.

Since by KCL the current to the right in the 2-Q resistor is /' — /„,,
l
«2' KVL around the
left mesh yields
2(/ - i
t2 ) + 4/ = v*i (5.4)
from which

I
—
= l£l-
^
I '_£2
(5.5)
3

If v gl = 18 V and i
g2 = 3 A, then / =3+ = 1 4 A. If we double v gl to 36 V and / Kl

to 6 A, then i is doubled to 8 A.

2fl

© 4a
© g2

FIG U R E 5.1 Loop of a linear circuit FIGURE 5.2 Linear circuit with two sources

As a final example, let us illustrate another use of the proportionality relation by

finding i\ in the circuit of Fig. 5.3. Such a circuit, because it resembles a ladder, is

sometimes called a ladder network.

We could write mesh or nodal equations, but to illustrate proportionality we shall
introduce an alternative method. Let us simply assume a solution,

1 V
92 Network Theorems Chap. 5

FIGURE 5.3 Ladder network

and see where it leads us. Referring to the figure, this assumption on v gives
x
/, =2A
and i2 = 1 A, and therefore

it = i, 3A
Proceeding down the ladder toward the source, we have, by KVL and KCL,

3/, 9V
V3 = V \ +V = 2 10 V
2A
5

is = i3 + /'

4 =5A
v4 = l(/ 5 ) = 5V
and, finally, if the guess that v x = V 1 is correct,

v. = v-. v* = 15 V
Our guess was not correct, since actually v g = 45 V, but in view of the laws of
probability this should not surprise us. However, by the proportionality relation, if a
15-V source gives an output v x
= 1 V, then our 45- V source will give three times as
much, so that the correct answer is

v, =3V
This method of assuming an answer for the output, working backwards to obtain
the corresponding input, and adjusting the assumed output to be consistent with the
actual input, by means of the proportionality relation, is particularly easy to apply to
the ladder network. There are other circuits to which it also applies, but the ladder
network is one of the most-often-encountered circuits, and the method is worthwhile
for it alone.
A nonlinear circuit is, of course, one that is not linear. That is, it has at least one
element whose terminal relation is not of the form of (5.1) or (5.2). An example is

given in Ex. 5.1.4, for which it is seen that the proportionality property does not
apply.
 /

Chap. 5 Network Theorems 93

EXERCISES
5.1.1 Show that in Ex. 4.5.3 if all the source voltages are doubled, then the loop
currents are doubled.

5.1.2 Show that in Ex. 4.4.3 if the source voltage is doubled, then the response / is

doubled.

5.1.3 Find /,, r2 , /3 , r3 , r4 , and /'

5 in Fig. 5.3.

Am. 6 A, 27 V, 9 A, 30 V, 15 V, 15 A
5.1.4 A circuit is made up of a voltage source v g a , 2-fi linear resistor, and a non-
linear resistor in series. The nonlinear resistor is described by

v = i
2

where v is the voltage across the resistor and ;', which is constrained to be non-
negative, is the current flowing into the positive terminal. Find the current flow-
ing out of the positive terminal of the source if (a) »g = 8V and (b) v g = 16 V.
Note that the proportionality property does not apply.
Am. (a) 2 A, (b) 3.123 A

5.2 SUPERPOSITION

In this section we shall consider linear circuits with more than one input. The linearity
property makes it possible, as we shall see, to obtain the responses in these circuits by
analyzing only single-input circuits.
To see how this may be accomplished, let us first consider the circuit of Fig. 5.2,

which was analyzed in the previous section. The output i satisfied the circuit equation
(5.4), which we repeat as

2(7 - i g2 ) + 4/ = vel (5.6)

From the solution,

/ = 5jl + if (5.7)

also given earlier, we see that i is made up of two components, one due to each input.
If/, is the component of /'due to v gX alone, i.e., with i gl = 0, then by (5.6) we have

2(7, - 0) + 4/, = vgl (5.8)

Similarly, if 2 is the component of/ due to i gl alone, i.e., with v gl = 0, then by (5.6)
we have
2(/ 2 - i gl ) + 4/ 2 = (5.9)
 94 Network Theorems Chap. 5

Adding (5.8) and (5.9), we have

2(/, - 0) + 4/, + 2(/ 2 - i g2 ) + 4/ 2 = vgl +

or
2[(/, + i2) - /.J + 4C/, + /a) = v tl (5.10)

Comparing this result with (5.6), we see that

i = h + i%

Also solving (5.8) and (5.9), we have

— v g\

t tol

which check with the two components given in (5.7).

— tgi
- *m (5.11)

Alternatively, we may obtain the components of/ directly from the circuit of Fig.
5.2. To find i we need to make the current source i gl zero. Since i g2 =
l
is the equa-

tion of an open circuit, this is accomplished by replacing the current source in the
circuit by an open circuit. This operation of making a source zero is sometimes
referred to, rather grimly, as "killing" the source, or making the source "dead." The
resulting circuit, in this case, shown in Fig. 5.4(a), from which it is easily seen, using
is

Ohm's law, that /,, the component of i due to v gl alone, is given by the first equation
of (5.11).

(al (b)

FIGURE 5.4 Circuit of Fig. 5.2 with (a) the current source dead and
(b) the voltage source dead

To find i2 , the component of/ due to i g2 alone, we must have v gl = 0, the equation
of a short circuit. Thus to kill a voltage source such as v gl we replace it by a short
,

circuit, as shown in Fig. 5.4(b). Using current division it is easy to see that i 2 is given by
the second equation of (5. 11).

The method we have illustrated with this example is called superposition, because
we have superposed, or algebraically added, the components due to each independent
source acting alone to obtain the total response. The principle of superposition
Chap. 5 Network Theorems 95

applies to any linear circuit with two or more sources, because the circuit equations are
linear equations (first-degree equations in the unknowns). This may be seen from (5.3)
and the nature of the v-i relations for the linear elements. In particular, in the case of
resistive circuits, the response may be found using Cramer's rule and is clearly a sum of
components, one due to each independent source alone. (Expanding the numerator
determinant by cofactors of the column containing the sources clearly reveals this.)

The linearity of the circuit equations enabled us to add equations (5.8) and (5.9)
and to see in (5.10) that the response was the sum of the individual responses. For
example, we are able to say that

2/, + 2i 2 = 2(/, + i2)

because these expressions are linear. However, we cannot say that

20, + y=i2 n\ + 2/1

because these are quadratic and not linear expressions.

Superposition allows us to analyze linear circuits with more than one independent
source by analyzing separately only single-input circuits. This is quite often advan-
tageous since we can use network reduction properties, such as equivalent resistance
and voltage division, in analyzing single-input circuits.
Formally, we state the principle of superposition as follows:

In any linear resistive circuit containing two or more independent sources,

any circuit voltage (or current) may be calculated as the algebraic sum of all
the individual voltages (or currents) caused by each independent source acting
alone, i.e., with all other independent sources dead.

As a second example, let us find the voltage v in the circuit with three independent
sources,shown in Fig. 5.5. To illustrate the earlier statement that the response is a sum

6V

en

FIGURE 5.5 Circuit with three sources

96 Network Theorems Chap. 5

of components, each due solely to an independent source, we shall first solve the
circuit by conventional means. Then to contrast the methods we shall use superposi-
tion.
To illustrate the role each source plays in the response we shall label the sources as

J
g\ 6V
'*2 (5.12)

Hi
= 18V
Assigning the mesh currents as indicated, the mesh equations are then given by

8/'i
— 2i 2 = —v *3
(5.13)
2/j + 5/ 2 = 3/ *:

and the response v is given by

»='— 3(i, a +i 2 ) (5-14)

Solving (5.13) for i2 by Cramer's rule we have

8 - v *i '

h = -2 v gi - 3/, 2 2
v *i
2 . 1
V* 3
8 -2 -
9
"
T8
-2 5

so that

v = Ki - ',2 + K3
Thus we see that v, as well as i2 , is a sum of components due to the individual sources.
Finally, substituting (5.12) into the expression for v, we have

w=4-2+3=5V (5.15)

Solving for v now by the method of superposition, we may write

v = *y, + v2 + v3 (5.16)

where ?',component due to the 6-V source alone (the 2- A and 18-V sources
is the
dead), v 2 component due to the 2 -A source alone (the two voltage sources dead),
is the
and i>is the component due to the 18-V source alone (the 6-V and 2- A sources dead).
3

The circuits showing v v 2 and v 3 are given in Figs. 5.6(a), (b), and (c), respectively.
r
, ,

In Fig. 5.6(a), killing the 18-V source ties nodes b and d together; in Fig. 5.6(b),
killing the two voltages ties nodes a and c together and b and d together; etc.
Chap. 5 Network Theorems 97

2A
e
a,c

en V^/\J\|

en

(c)

FIGURE 5.6 Circuit of Fig 5.5 with various sources killed

From Figs. 5.6(a), (b), and (c), it is almost trivial to obtain

v, =4V
r2 = -2V

which are the values given in (5.15).

To illustrate superposition when there is a dependent source present, let us
consider the circuit of Fig. 5.7, where it is required to find the power delivered to the
3-ft resistor. At the outset we should make it clear that power is not a linear combina-

6 A

FIGURE 5.7 Circuit with a dependent source

98 Network Theorems Chap. 5

tion of voltages or currents. In the case we are considering, we have

P =

which is a quadratic, and not a linear, expression. Therefore superposition will not
apply to obtaining power directly. That is, we cannot find the power due to each

independent source acting alone and add the results to obtain the total power. How-
ever, we can find v by superposition and subsequently obtain p.
Letting v x
be the component of v when the 12-V source is acting alone [Fig. 5.8(a)]
and v 2 be the component of v due to the 6-A source alone [Fig. 5.8(b)], we have, as
before,
v = v, + v2

We note that the current i driving the dependent source also has components, which

we have labeled /, and i2 , due to each source alone. Solving for v in Fig. 5.8(a) and v 2 x

in Fig. 5.8(b), we have

vx = 6 V, v2 =9V
and therefore v = 15 V. The power is then given by

= (15)^ = 75 w

'i 12 V '2

o—,

(o) (b)

FIGURE 5.8 Circuit of Fig. 5.7 with (a) the current source killed and
(b) the voltage source killed

There are at least three things that are illustrated by the last example. First, as

already mentioned, we cannot use superposition directly to calculate power, but we

may use it to get current or voltage, from which power is then found. Second, in
applying superposition, only the independent sources are killed, never the dependent
sources. Finally, superposition is often a poor method for solving circuits with
dependent sources, because each of the individual single-input circuits frequently is
almost as difficult to analyze as the original circuit. For example, there are two
 Chap. 5 Network Theorems 99

meshes in Fig. 5.7. the original circuit, and one mesh current is known. Therefore only
one application of KVL is needed, around the left mesh, to solve for the current in the
1-fi resistor. This is also exactly the case in the circuit of Fig. 5.8(b).

EXERCISES
5.2.1 Solve Ex. 4.2.3 using superposition.

5.2.2 Solve Prob. 3.9 using superposition.

5.2.3 Find the power delivered to the 3-fi resistors in Figs. 5.8 (a) and (b), and show
that their sum is not equal to the total power delivered to the 3-fi resistor in
Fig. 5.7. Am. 12 W, 27 W

5.3 THEVEMIN'S AND NORTON'S THEOREMS

In the previous section .of this chapter we saw that the analysis of some circuits could
be greatly simplified by applying the principle of superposition. However, as we saw
in the last example of Sec. 5.2, superposition alone may not reduce the complexity of
the problems, but its use may lead to additional work. In this section we shall consider
Thevenins and Norton's theorems, which will in many cases be applicable and greatly
simplify the circuit to be analyzed.
We shall assume that the circuit to be considered can be separated into two parts,
as shown in Fig. 5.9. The part denoted as circuit A is a linear circuit containing
resistors, dependent sources, or independent sources. Circuit B may contain nonlinear
elements. We also add the constraint that any dependent source in either circuit must
have its controlling element in the same circuit. That is, no dependent source in
circuit A can be controlled by a voltage or current associated with an element in
circuit B and vice versa. The reason for this will become clear in the development which
follows.
We shall show that we can replace circuit A by an equivalent circuit containing a
source and a resistor in such a way that the voltage-current relations at terminals a-b
remain the same. Since our aim is to maintain the same terminal relations at a-b,
clearly, from the standpoint of circuit A, we may obtain the same effect if we replace
circuit B by a voltage source of v volts with the proper polarity, as shown in Fig. 5.10.
Insofar as circuit A is concerned, it has the same terminal voltage, and since circuit A
itself is unchanged, the same terminal current must flow. We have now obtained a
linear circuit and can consequently make use of all the properties we have established
for such circuits.
100 Network Theorems Chap. 5

Linear + Linear
Circuit
Circuit V Circuit
A B
A
b

FIGURE 5.9 Partitioned circuit FIGURE 5.10 Replacement of circuit

B by a voltage source

In particular, applying superposition to the linear circuit which we have now

obtained, we see that the current /'
will be given by

i = ix + 4 (5.17)

where i x
is produced by the voltage source v with the network A dead (all its indepen-
dent sources killed) and iK is the short-circuit current produced by any sources inside
circuit A with v killed (replaced by a short circuit). These two cases are shown in Figs.
5.11(a) and (b).

__
Circuit Circuit

dead
A
A A
alive

R th

(a) (b)

FIGURE 5.11 Circuits obtained for applying superposition

Since the independent sources are dead in circuit A [Fig. 5.11(a)], from the termi-
nals of the source v we see only a resistive circuit, the equivalent resistance of which we
shall call R tb . By Ohm's law, we therefore have

v
(5.18)

By (5.17) the expression for the current i is then

1 = ~ir + ls (5.19)

A in the general case, it must hold for any condition

Since (5.19) describes network
at the terminals. Suppose the terminals are open. In this case, i = 0, and we shall
denote the voltage by v — v x the open-circuit voltage. Substituting these values into
,
Chap. 5 Network Theorems 101

(5. 19). we have

0- -#* + *.

or

(5.20)

Eliminating / sc in (5.19) and (5.20). we have

v = -R J + t
v, (5.21)

The relations given in (5.19) and (5.21) may be used to obtain two very useful
circuitswhich are equivalent to circuit A. Historically, the first of these was Thevenin s
equivalent circuit, named in honor of the French telegraph engineer Charles Leon
Thevenin (1857-1926), who published his results in 1883. [The circuit might have been
more appropriately named for the great German physicist H. S. F. Helmholtz (1821-
1894), who gave a restricted case of it in 1853.]
The Thevenin equivalent circuit is simply one which is described by (5.21) with
terminal voltage v and terminal current /, oriented as in Fig. 5.9. To draw the circuit
we note that v is the sum of two terms, which therefore must represent two elements in
series whose terminal voltages add up The first term evidently corresponds to a
to v.

resistance i? th called the Thevenin resistance, and the second term corresponds to a
,

voltage source with terminal voltage oc The result is shown in Fig. 5.12, where the
^' .

dashed lines represent the connections to the external circuit B of Fig. 5.9. Analysis of
the circuit will show that (5.21) is satisfied. The statement that Fig. 5.12 is equivalent
at the terminals to Fig. 5.9 is known as Thevenin s theorem.
Another circuit that is equivalent to circuit A
of Fig. 5.9 is obtained from (5.19).
This circuit is the duaFof the Thevenin circuit
and is called the Norton equivalent
circuit in honor of the American engineer E. L. Norton (1898- ), whose work was

published some 50 years after Thevenin's. From (5.19) we see that is the sum of two /'

terms, which must then represent two parallel elements whose currents add up to i. The
first term evidently arises from the Thevenin resistance R lh and the second term ,

corresponds to a current source /sc The result is shown in Fig. 5.13, and the statement
.

"th

6
FIGURE 5.12 Thevenin equivalent FIGURE 5.13 Norton equivalent circuit of
circuit of Circuit A of Fig. 5.9 Fig. 5.9
102 Network Theorems Chap. 5

of equivalence of Figs. 5.9 and 5.13 is Norton's theorem. Again, the dashed lines
represent the connections to the external circuit B.
As an example, let us find the Thevenin and Norton equivalent circuits for the
network to the left of terminals a-b in Fig. 5.14. Then using the results, let us obtain
the current i, as shown, in terms of the load resistance R.

6 V

4-Q-
FIGURE 5.14 Circuit with a variable load resistor

obtain the Thevenin circuit we need to find R th and vx The Thevenin resistance
To .

R lh found from the dead circuit (the two independent sources made zero), shown in
is

Fig. 5.15(a), from which we have

th

(o)

FIGURE 5.15 Circuits for obtaining the Thevenin circuit of Fig. 5.14

The open-circuit voltage v x is obtained from Fig. 5.15(b). Since the terminals a-b
are open, the voltage v oc is across the 3-Q resistor. Labeling the nodes as shown, with
node b as reference, and writing a nodal equation at the generalized node, shown
dashed, we have

"
3
or
^-tf = 6V J

The Thevenin equivalent circuit, with the load R connected, is shown in Fig. 5.16.

(a
t
Chap. 5 Network Theorems 103

We note that the polarity for i ^— 6 V is such that the correct voltage polarity
results at terminals a-b when they are opened.
The current i in Fig. 5. 14 is the same as that of Fig. 5.16, which in the latter case is

readily seen to be

1
R 1 4

We may use this result to find the load current for any load R that we choose.
To obtain the Norton equivalent circuit we use R lh = 4 Q, as before, and calculate
/sc . We may short terminals a and b and calculate isc from the resulting circuit, or we
may use the v x we already have and get /'

sc
from (5.20). In the latter case we have

The Norton equivalent circuit, with the load R connected, is shown in Fig. 5.17.
Using current division, we have, as before,

C M C

FIGURE 5.16 Loaded Thevenin equi- FIGURE 5.17 Loaded Norton equivalent of
valent of the circuit of Fig. 5. 14 the circuit of Fig. 5.14

The direction of the 1.5-A source is such that when R is replaced by a short,
/ sc = 1.5 A has the correct direction. In this case and in the Thevenin case of Fig. 5.16,
it is a simple matter to place the sources correctly, but in a complex example some care
needs to be exercised so that the polarities are correct.
Let us now consider an example containing a dependent source, such as the circuit
of Fig. 5.18(a). Suppose we want the Norton equivalent circuit at terminals a-b. We
shall need, in this case, RA defined for the dead circuit of Fig. 5.18(b) and / sc shown in

Fig. 5.18(c).
We may find /'

sc from Fig. 5.18(c) by noticing that

10 h — i*

and writing the two mesh equations,

-4(10 - /, - y- 2/, + 6/, =

-6/j + 3/5c =
 °

1 04 Network Theorems Chap. 5

2i, V

i <r +> T
- ;
— ^AA/
V °
-r—^VW
3n
—
4a 6n

(a) <th

(b)

10 A

(c)

FIGURE 5.1 8 fa,) Circuit to be analyzed, with (b) its source killed and
(c) its terminals short-circuited

Eliminating i u we have
L= 5 A (5.22)

We cannot find R th from Fig. 5.18(b) simply by calculating equivalent resistance,

because of the dependent source. However, we could excite the circuit with a voltage
v (or a current i) at its terminals and calculate the resulting / (or v). Then R lh = vji.
Alternatively, we could find v oc as for the Thevenin circuit, and
, obtain R lh from (5.20).
This is the method we shall use.
To find v oc we refer to Fig. 5.19, where we have

^oc = 6/!

FIG U RE 5.1 9 Circuit of Fig. 5. 18(a) with its terminals opened

Chap. 5 Network Theorems 105

and around the center mesh,

-4(10-/,) -2/, + 6/, =0

From these equations we find i'
K = 30 V, and thus

As a final example, let us find the Thevenin equivalent of the circuit of Fig. 5.20.
To begin with, we see by inspection that since there is no independent source present,
we must have
Voc =L= (5.23)

Also, the dead circuit is the given circuit itself, so that R th is simply the resistance seen
at the terminals of Fig. 5.20. In view of (5.23), we cannot use v oc = i? th / sc to get i? th , as
we did in the previous example. The only recourse we have is to excite the circuit at its

terminals and calculate R th from the results.

2i, V
v-2i,

4n 6n> , (D
IA

4fl

'th

FIGURE 5.20 Circuit with a dependent FIGURE 5.21 Circuit of Fig. 5.20 excited by a
source current source

For example, suppose we excite the circuit with a 1-A current source, as shown in
Fig. 5.21. Then we have
v
R t

where v is the resulting terminal voltage. Taking the bottom node as reference, the
nonreference node voltages are as shown. A nodal analysis yields

v — 2/,
= 1
6
where

From these equations we have r = 3V and thus i? = th 3 Q. The Thevenin equivalent

(as well as the Norton equivalent) is shown in Fig. 5.22.
106 Network Theorems Chap. 5

3n

FIGURE 5.22 Thevenin equivalent of

Fig. 5.20

EXERCISES
5.3.1 Replace the circuit to the left of terminals a-b by its Thevenin equivalent and
find i. Ans. v x = 24 V, R th = 6 Q, i =3A

24V

EXERCISE 5.3.1

5.3.2 Replace everything in the figure for Ex. 5.3.1 except the 3-£2 resistor by its

Norton equivalent and find v. Ans. isc = \ A, 7? th = 3 Q, v =2V

5.3.3 Find jR th in Fig. 5.18 (b) by exciting the circuit with a voltage (or current) and
finding the current (or voltage) which results. Ans. 6 Q,

5.4 PRACTICAL SOURCES

In Chapter 1 we defined independent sources and pointed out that they were ideal
elements. An ideal 12-V battery, for example, supplies 12 V between its terminals
regardless of the load connected to the terminals. However, a real, or practical, 12-V
battery supplies 12 V when its terminals are open-circuited and supplies less than 12 V
Chap. 5 Network Theorems 107

as current is drawn through the terminals. A practical voltage source thus appears to
have an internal drop in voltage when current flows through its terminals, and this
internal drop diminishes the voltage at the terminals.
We may represent a practical voltage source by the mathematical model of Fig.
5.23, consisting of an ideal source v g in series with an internal resistance R g The
.

voltage v seen at the terminals of the source now depends on the current i drawn from
the source. The relationship is easily seen to be

RJ (5.24)

A PRACTICAL
VOLTAGE
SOURCE

FIGURE 5.23 Practical voltage source connected to a load RL

Thus under open-circuit conditions (/' = 0), we have v = vg , and under short-circuit
conditions (v = 0), we have i = vjR g
. If Rg > 0, as it is for a practical source, the
source can never deliver an infinite current, as an ideal source can.
For a given practical voltage source (fixed values of v g and Rg in Fig. 5.23), the
load resistance RL determines the current drawn from the terminals. For example, in
Fig. 5.23 the load current is

/ = (5.25)
Rg + R L

DEAL SOURCE

FIGURE 5.24 Practical and ideal voltage source

characteristics
108 Network Theorems Chap. 5

Also, by voltage division we have

v — R,v
L"g
(5.26)
R g + Rl

Therefore as we vary RL both i and v vary. A sketch of v vs. RL is shown in Fig. 5.24,
along with the ideal case, which is dashed. For large values of RL relative to Rg , v is

very nearly equal to the ideal value of v g (If . RL is infinite, corresponding to an open
circuit, then v is vg .)
We may replace the practical voltage source of Fig. 5.23 by a practical current
source by means of Norton's theorem. The result is shown in Fig. 5.25 and is seen to
be an ideal current source in parallel with the internal resistance. The two sources in
Figs. 5.23 and 5.25 are equivalent at the terminals, of course, if Rg is the same in both
cases, and
vg =R i
g g (5.27)

By current division, we find, in Fig. 5.25,

RJ g (5.28)
R. + Rr

A PRACTICAL
CURRENT
SOURCE

FIGURE 5.25 Practical current source connected to a load RL

Therefore, for a given current source (fixed values of i

g
and v g ), the load current
depends on R L A . sketch of i vs. RL is shown in Fig. 5.26, along with the ideal case,
which is dashed.
Very often network analysis can be greatly simplified by changing practical voltage
sources to practical current sources, and vice versa, by means of Norton's and
Thevenin's theorems. For example, suppose we wish to find the current shown in /'

Fig. 5.27. We could solve the problem in a number of ways, such as replacing every-
thing except the 4-Q resistor by its Thevenin equivalent and then finding i. However,
we shall illustrate instead the method of successive transformation of sources.
Let us begin by replacing the 32-V source and internal 3-Q resistance by a practical
current source of a 3-fi internal resistance and a ^-A ideal source. Then let us replace
the 4-A source and the internal 2-Q resistance by a voltage source of 2(4) = 8 V and a
Chap. 5 Network Theorems 109

IDEAL SOURCE

FIGURE 5.26 Practical and ideal current source

characteristics

3A 4a in

32 V
Q 6A 6A 2H
© 4A

FIGURE 5.27 Circuit with two practical sources

2-Q internal resistance. We are applying, respectively, Norton's and Thevenin's

theorems, or equivalently, (5.27). The results of these two source transformations are
shown in Fig. 5.28.

32

FIGURE 5.28 Result of two transformations applied to Fig. 5.27

We may now combine and 6-Q resistances and the series l- and 2-Q
the parallel 3-
resistances, as shown and repeat the source transformation procedure.
in Fig. 5.29(a),
Continuing the process, as shown in Figs. 5.29(b), (c), and (d), we finally arrive at an
equivalent circuit (insofar as is concerned) which can be analyzed by inspection. In
/'

this case, from Fig. 5.29(d), the answer is

64 16
/ = 3 J- = 2A
2 + 4 + 2

This procedure may seem unduly long, but it should be observed that most of the steps
may be carried out mentally.
 — -

110 Network Theorems Chap. 5

(o)

¥v©
(b)

i
/
2A
WVL —
i-». 4A
j vw-

T*© 2a
on*
(c)

2n 2X1

i
—VW^
^V
© © S*
(d)

FIGURE 5.29 Steps in obtaining i in Fig. 5.28

As a final note in this section, we often may combine sources as we do resistors to

obtain equivalent sources. For example, ifwe are interested only in i in Fig. 5.29(d), we
may combine the three series resistors, as we know, but we may also combine the
series sources. They represent a net source of ^—^= 16 V with a polarity like that
Chap. 5 Network Theorems 111

of the larger source. Thus an equivalent circuit insofar as i is concerned is that of Fig.
5.30. Similarly, we may combine parallel current sources to obtain an equivalent
source.

FIGURE 5.30 Circuit equivalent to

Fig. 5.29 for finding i

EXERCISES
5.4.1 By source transformations, replace the network to the left of terminals a-b in
the figure for Ex. 5.3.1 by an equivalent network having only one resistor R.
From this network, find /'
as shown. Am. R = 6 Q, / = 3 A
5.4.2 Convert all the sources in the figure for Ex. 4.2.3 to voltage sources and find u 3 .

Arts. 5 V
5.4.3 In the figure for Prob. 4.5, convert the network to the left of terminals a-b to a
single practical current source by source transformations and find v. Ans. 10 V

5.5 MAXIMUM POWER TRANSFER

There are many applications in circuit theory where it is desirable to obtain the
maximum possible power that a given practical source can deliver. It is very easy, using
Thevenin's theorem, to see what maximum power a source is capable of delivering and
how to load the source so as to obtain this maximum power. This will be the subject of
this section.
Let us begin with the practical voltage source shown earlier in Fig. 5.23 with a load
resistance R L The
. power p L delivered to the resistor RL is given by

and it is this quantity which we wish to maximize.

Since the source assumed to be given. vg and R g are fixed, and thus/> L is a func-
is

tion of RL . To maximize p L we can make dpJdR L = and solve for R L From (5.29) .
112 Network Theorems Chap. 5

we obtain
+ R L - 2(R g + R L )R,
(R g )
2

dR L (R g + RlY
(R g - R L )vt = Q
(5.30)
(R g + RLy
which results in
RL = R g (5.31)

It may be readily shown that

d lp L
dR\ SRI

and therefore (5.31) is the condition which maximizes p L We see, therefore, that the .

maximum power is delivered by a given practical source when the load R L is equal to
the internal resistance of the source. This statement is sometimes called the maximum
power transfer theorem. We have developed it for a voltage source, but in view of
Norton's theorem it also holds for a practical current source.
The m aximum po wer th at the practical voltage source is capable of delivering to
the load is given by (5.29) and (5.31) to be

(5.32)

In the case of the practical current source, the maximum deliverable power is

Pi*. = ^ (5 - 33 )

This may be seen from Fig. 5.25 and (5.31) or by (5.32) and Norton's theorem.
We may extend the maximum power transfer theorem to a linear circuit rather
than a single source by means of Thevenin's theorem. That is, the maximum power is

obtained from a linear circuit at a given pair of terminals when the terminals are
loaded by the Thevenin resistance of the circuit. This is obviously true since by
Thevenin's theorem the circuit is equivalent to a practical voltage source with internal
resistance R th .

As an example, we may draw the maximum power from the circuit of Fig. 5.18(a)
if we load terminals a-b with the Thevenin resistance

RL = *,h = 6n
Since /sc — 5 A by (5.22), we may draw the Norton equivalent circuit with the required
RL as shown in Fig. 5.31. The power supplied to the load is given by

_ 90QR L
P
(Rl + 6)
2
Chap. 5 Network Theorems 113

r, = 6a

FIGURE 5.31 Equivalent circuit of Fig. 5.18(a) loaded

maximum power transfer
for

which for RL = 6 yields

= 37.5 W
Any other value of RL will result in a lower value of p. For example, if RL 5 ft,
thenwe have/? = 37.19 W, and for RL == 1 Q, we have/? = 37.28 W.

EXERCISES
5.5.1 Find the maximum power that can be transferred from the circuit to the left
of terminals a-b in the figure for Ex. 5.3.1. Ans. 24 W
5.5.2 Show two networks given are equivalent
that the at terminals a-h and find the
power dissipated in the 3-Q resistor in each case.
Ans. (a) 12W, (b)3 W

en. 6Ci

(o)

EXERCISE 5.5.2

5.5.3 Find the maximum power delivered to the load RL in Fig. 5.23 if v
g
and RL >
are fixed and R g is variable. Ans. vljR L when Rg =
114 Network Theorems Chap. 5

PROBLEMS
5.1 Solve Prob. 2.16 using the method applied to Fig. 5.3.

5.2 Solve Prob. 2.17 using the method of Prob. 5.1.

5.3 Solve Prob. 3.13 using the method of Prob. 5.1. (Suggestion: Assume i = mA,
1

etc.)

5.4 Find v in Prob. 4.5 using superposition.

5.5 Solve Prob. 4.10 using superposition.

5.6 Using superposition, find the power delivered to the 3-kfi resistor.

4V

2kA 3k.fl

PROBLEM 5.6

5.7 Find v using superposition.

PROBLEM 5.7

5.8 Find v using superposition.

Chap. 5 Network Theorems 115

4n

PROBLEM 5.8

5.9 Find v in Prob. 4.5 by replacing the network to the left of terminals a-b by its

Thevenin equivalent.
5.10 Find the power delivered to the 4-fi resistor in Prob. 4.9 by replacing the net-
work to the left of terminals a-b by its Norton equivalent.
5.11 Solve Prob. 4.21 by replacing the network to the left of terminals a-b by its

Thevenin equivalent.

5.12 Replace everything except the 6-kfi resistor in Prob. 3.13 by its Thevenin
equivalent and find i.

5.13 Find v by replacing the network to the left of terminals a-b by its Norton
equivalent.

PROBLEM 5.13

5.14 In the figure for Prob. 5.13, replace the network to the right of terminals c-d by
its Thevenin equivalent and find »,.

5.15 By successive source transformations, change the circuit shown to an equivalent

one at terminals a-b containing only one resistor. From this equivalent circuit,

find v.
116 Network Theorems Chap. 5

PROBLEM 5.15

5.16 Solve Prob. 4. 1 9 by using source transformations to obtain an equivalent circuit,

insofar as the 4-kQ resistor is concerned, containing only one source and one
resistor, in addition to the 4-kQ resistor.

5.17 Find the maximum power that can be transferred to R by the circuit to the left

of terminals a-b. (Suggestion: Use successive source transformations to find the

Thevenin or Norton circuit.)

3A 2d

PROBLEM 5.17

5.18 In the figure for Prob. 4.22, find the resistance R to replace the 1-Q resistance
between terminals a-b which will draw the maximum power from the rest of the
circuit. Find this maximum power.

5.19 Show that the 1-Q resistance between terminals c-d in the figure for Prob. 4.22
has the correct value to draw the maximum power from the rest of the circuit.
Find this maximum power.

5.20 Solve Prob. 4.23 by assuming v 2 = 1 V and using the proportionality principle.
 6
INDEPENDENCE
OF EQUATIONS

An electric network is determined by the type of elements it contains and the manner
in which the elements are connected. We have spent considerable time in the previous
chapters considering the elements themselves and their volt-ampere characteristics. In
this chapter we shall consider the manner in which the network elements are con-
nected, or, as it is called, the network topology. As we shall see, a study of
sometimes
the topology of the network provides us with a systematic way of determining how
many equations are required in the analysis, which ones are independent, and the best
set of equations to select for the most straightforward analysis.

6.1 GRAPH OF A NETWORK

To illustrate the problems involved in the analysis of more complicated networks, let
us consider the circuit of Fig. 6. 1 . The resistors are numbered 2, 9, with values
1 , . . . ,

of resistance, say. R R2
x
, , . . . , R 9 . Suppose we are required to perform a loop analysis,
in which case we need to write a set of independent KVL equations. (We note that the
circuit is nonplanar, and thus we cannot perform a mesh analysis. Anyone doubting

this is welcome to try redrawing the circuit in a planar fashion.)

There are 15 loops in the circuit, as may be verified by several means, one of which
is trial and error (much trial and more error). For the curious reader the 15 loops are
(1, 3, 4, 5), (1. 3, 7, 9), (2, 3, 5, 6), (I, 2, 8, 9), (1, 2, 4, 6), (4, 5, 7, 9), (2, 3, 4, 5, 8, 9),

(1, 2, 5, 6, 7, 9), (1, 3, 5, 6, 8, 9), (2, 3, 7, 8), (2, 3, 4, 6, 7, 9), (4, 6, 8, 9), (5, 6, 7, 8),

(1, 3, 4, 6, 7, 8), and (1, 2, 4, 5, 7, 8). That is, resistors 1, 3, 4, and 5 form a loop; 1, 3,

7, and 9 (with the source vg ) form a loop; etc.

To undertake a loop analysis of Fig. 6.1 we need to know which of these loops are
independent and how many are required. To answer these questions we need consider

117
118 Independence of Equations Chap. 6

only how the elements are connected; it is unimportant what kind of elements are
involved. To facilitate matters, then, we may retain the nodes of the network and
replace its elements by lines. The network topology thus is preserved in a much
simpler configuration.
The configuration of lines and nodes obtained by replacing the elements of a
network by lines is called the graph of the network. The lines of the graph are called
its and the nodes of the graph are, of course, the nodes of the network. As an
branches,
example, the graph of the network of Fig. 6.1 is shown in Fig. 6.2. It has nine branches
and six nodes. (We could consider a node between resistor 9 and v g but inasmuch as ,

this will not affect the number of loops, we have chosen to consider these two series
elements as one element, namely a nonideal voltage source.)

F I GURE 6 . 1 Nonplanar circuit FIGURE 6.2 Graph of a network

We shall say that a graph is connected if there is a path of one or more branches
between any two nodes. The graph of Fig. 6.2 is evidently connected. An example of a
graph that is not connected is shown in Fig. 6.3. There is, for instance, no path
between nodes a and d. For the present we shall consider only connected graphs.

FIGURE 6.3 Unconnected graph

Chap. 6 Independence of Equations 119

EXERCISES
6.1.1 Show that the given graph is planar.

EXERCISE 6.1.1

6.1.2 Show that the given graph is planar.

EXERCISE 6.1.2

6.2 TREES AND LINKS

We define a tree of a graph as a connected portion of the graph that contains all the
nodes but no loops. As an example, Fig. 6.4(b) is a tree of the graph of Fig. 6.4(a). The
tree is connected, has no loops, and contains all the nodes of the graph.
1 20 Independence of Equations Chap. 6

2 4
o o o

(a) (b) (c)

FIGURE 6.4 Graph and two of its trees

Generally, a graph has many trees. The configuration of Fig. 6.4(c) is evidently
another tree of the graph of Fig. 6.4(a), since it satisfies all the requirements. This
particular graph has 24 trees, which the reader may wish to try to discover. It will help
in enumerating the trees to notice that each one has exactly three branches, since it

takes at least three lines to connect four nodes and more than three lines will form a
loop. There are 35 ways to select seven branches three at a time, but 1 1 of these
combinations are not trees.

In the general case, let B be the number of branches and TV be the number of nodes
in a given graph. Then any tree of the graph contains TV nodes and TV — 1 branches.
The number of nodes follows from the definition of a tree, and the number of tree
branches may be established by the following construction argument. Let us build the
tree starting with one branch and the two nodes to which it is connected. Each
additional branch connected to build the tree adds one additional node. The number of
nodes is, therefore, one more than the number of branches, and since there are TV
nodes, there must be TV — 1 branches.
As examples, the graph of Fig. 6.4(a) has TV = 4, and thus the number of tree
branches in both Figs. 6.4(b) and (c) is TV — 1 = 3.

Once the branches of a tree are designated, the remaining branches of the graph are
called links. The number of links graph is then B — (TV — 1) or B — TV + 1. As an
in a

example, the tree of Fig. 6.4(b) is redrawn in Fig. 6.5, with the tree branches shown as
solid lines and the links as dashed lines. The number of links in this case, since B= 7,

is 7 -4+ 1 = 4.

FIGURE 6.5 Tree branches and links

of a graph
Chap. 6 Independence of Equations 121

EXERCISES
6.2.1 Find the number of tree branches and links in the graph of Fig. 6.2. Ans. 5, 4

6.2.2 Find all the trees of the graph shown.

Ans. (1, 2, 4), (1, 2, 5), (1, 3, 4), (1, 3, 5), (1,4, 5), (2, 3, 4), (2, 3, 5), (2, 4, 5)

EXERCISE 6.2.2

6.2.3 In the figure for Ex. 6.2.2, let i 2 and / 4 be link currents in elements 2 and 4,

directed to the right. Let /",,

/'

3 , and / 5 be tree currents directed downward. Find

the tree currents in terms of the link currents. Ans. — /*2, 1*2 ~ U, U

6.3 INDEPENDENT VOLTAGE EQUATIONS

In this section we shall consider the nodal analysis of a circuit whose graph has N
nodes and B branches. A tree of the graph then has N— 1 branches and, of course,
TV — 1 tree branch voltages. There are also B —N+ I link voltages in the circuit.
Let us imagine that all any tree are made zero by short-
the branch voltages of
circuiting the branches (i.e., the branches are replaced by short circuits). Then all the
nodes of the circuit, being in the tree, are at the same potential, and thus the link
voltages are all zero also. Therefore the link voltages depend on the tree branch
voltages, for if a link voltage were independent of the tree voltages, it could not be
forced to zero by short-circuiting the tree branches. On the other hand, if one tree
branch were not short-circuited, there would be one node at a different potential and
thus one tree branch voltage not dependent on the others. We may conclude, there-
fore, that the N — 1 branch voltages of any tree are independent and may be used to
find the link voltages.
Another way to see that any link voltage may be expressed in terms of tree voltages
is adding the link to the tree completes a circuit whose only other elements
to note that
are tree branches. Thus by KVL the link voltage is an algebraic sum of tree branch
voltages.
1 22 Independence of Equations Chap. 6

As an example, the graph of Fig. 6.6 has tree branch voltages v u v 2 and v 3 indi- , ,

cated by the solid lines. The dashed lines are the links whose voltages are v 4 v 5 , and ,

v 6 If the link labeled v 4

. is added to the tree, the circuit of v 2 v 3 , , v4 is formed. By KVL
around this circuit, we may obtain

v4 = v3 — v2

In like manner, adding first link v s and then link v 6 , we obtain

v5 = v x
— v2

v6 = Vi - v3

Thus the link voltages may be found from the tree branch voltages.

FIGURE 6.6 Tree and link voltages

A systematic way of writing equations involving tree branch voltages is to imagine

opening a tree branch, noting that this separates the tree into two parts. Currents flow
between the two parts through the tree branch imagined to be open and through links.
Thus by KCL the algebraic sum of these currents in a given direction is zero. This
procedure may then be repeated for the other tree branches.
To illustrate the use of tree branch voltages, let us consider the circuit in Fig.
6.7(a). A graph of the circuit is shown in Fig. 6.7(b) with the tree branches shown as
solid lines and the links as dashed lines. Since the tree branch voltages are independent,

we have included in the tree the 20-V source. Thus the number of unknowns is reduced

HA

(a)

FIGURE 6.7 Circuit and its graph

 —
Chap. 6 Independence of Equations 1 23

by one. B> K.VI we may find the link voltages in terms of the tree voltages, with the
results shown in Fig. 6.7(b).
[f we imagine the tree branch (a, b) labeled r, as open, then the tree is separated

into two parts. These two parts are connected by branch (cr, b) and links {a, c), (b, d),
and (</. c), as indicated by the line marked I. Summing the currents across the line in
the direction of the arrow, we have

—
v t + v2 + v t
20
2
v,
l -H=0n
, ,

Repeating the procedure for tree branch (b, c), labeled v 2 , leads to line II and the
equation
— (»i + v i) ~2v 2 + 11 =0
Solving these equations, we have r, = 8 V and v 2 = 1 V. All the link voltages, and
consequently all the link and tree currents, may now be found.
In any circuit analysis procedure where the unknowns are voltages, we need to find
only the N— 1 tree branch voltages which constitute an independent set. This means
that only N — 1 independent voltage equations are required in the analysis. Since any
independent set of equations will suffice, then any independent set of N— 1 voltages
constitutes a solution.
Another independent set of N — 1 voltages, other than the tree branch voltages, is
the set of nondatum node voltages, considered in the nodal method of Chapter 4. To
see this, we note that any nondatum node is in the tree and is connected through tree
branches to the datum node. Thus every nondatum node voltage is an algebraic sum of
tree branch voltages (the tree branches between the nondatum and datum nodes). On
the other hand, every tree branch voltage is the difference between its two node
voltages. In summary, the node voltages may be determined from the tree voltages and
vice versa. Thus the nondatum node voltages are also an independent set.
In the example of Fig. 6.7 we see that if d is the datum node, then the nondatum
node voltages va v b and v c are related to the tree voltages v u v 2 and 20 by
, , ,

va = 20

vb = 20 - »,

vc = 20 — t>i
— v2

Conversely, we have
*>i =%~ vt

v2 = Vb - Vc

20 =v a

The nodal method, in many cases, is easier to apply than the loop method because
the nodes are easy to find. In the loop method, as exemplified by the example of Fig.
6.1, the appropriate loops may be difficult to identify. In the next section we shall
1 24 Independence of Equations Chap. 6

consider a method based on graph theory of finding sets of independent loop

equations.

EXERCISES

6.3.1 In the graph of the circuit in the figure for Ex. 4.3. 1
, select the tree of the voltage
sources and the 3-fi resistor. Using the method of this section, write one KCL
equation and determine u lt Ans. —4 V
6.3.2 Select the tree of the voltage source, and the 3- and 6-fi resistors, and use the
method of this section to find v. Ans. 3 V

EXERCISE 6.3.2

6.3.3 Using an appropriate tree and the methods of this section, find v in Prob. 4.18.
(Note: The tree should contain the voltage v and the three sources as well as
one other branch.) Ans. 20 V

6.4 INDEPENDENT CURRENT EQUATIONS

As we have seen in the example of Fig. 6.1, it is not always easy to identify the indepen-
dent loops for a loop analysis of a circuit. To develop a systematic means of writing
loop equations, let us consider a general network with B branches and nodes. N
Corresponding to a given tree there are B — N
+ 1 links.
Suppose that all the link currents are made zero by open-circuiting the links
(replacing the links by open circuits). Since the tree contains no loops, then all the tree
branch currents are zero also. The tree currents thus depend on the link currents, i.e.,
they may be expressed in terms of the link currents, for if a tree current were indepen-
dent of the link currents it could not be forced to zero by open-circuiting the links.
Chap. 6 Independence of Equations 1 25

Moreover, if one link is not open-circuited, a loop is left in the graph, and a current
will flow in the link. A link current thus is not dependent on the other link currents. In
summary, the B —N+ l link currents are an independent set, and the loop analysis
of the circuit requires B —N l independent equations.
One systematic way to find B —N+ l independent loops is to start with the tree
and add one of the links. This determines the loop containing that link, since adding
the link to the tree closes a loop. Remove this link and add another link to the tree,
determining a second loop. Continue the process until the B — N -|- l loops are
found. The set is independent because each loop contains a different link.

As an example, let us reconsider the circuit of Fig. 6.1. One tree consists of
branches 1. 5. 7, 8. and 9. with corresponding links 2, 3, 4, and 6, as shown in Fig. 6.8.

Closing the links one at a time results in the four independent loops I, II, III, and IV
shown. Loop I contains link 2 and tree branches 8, 9, and 1 ; loop II contains 3, 7, 9,

and I ; loop III contains 4, 5, 7, and 9; and loop IV contains 6, 5, 7, and 8. These four
loops are sufficient for performing a loop analysis.
To illustrate the use of link currents in circuit analysis, let us return to the example
of Fig. 6.7(a). The graph is redrawn in Fig. 6.9, showing the link currents and /,, i 2 ,

1 1 A. We have chosen the current source as a link because the link currents are an
independent set. This reduces the number of unknowns by one. Generally, for this
reason, one should place voltage sources in the tree and current sources in the links.
The tree branch currents, as in the general case, may be found from the link
currents, as shown in Fig. 6.9. Closing the links labeled /', and i2 forms loops 1 and 2, as
indicated. Applying KVL to these loops yields, from Figs. 6.7(a) and 6.9,

2/, - 20 + /J
— i2 + 11 =
h -F l: 11=0

the solution of which is /', = 6A and ;'-

9 A.

i, - i,+ II loop 2

loop I

FIGURE 6.8 Circuits of Fig. 6.1 FIGURE 6.9 Graph of Fig. 6.7(a)
1 26 Independence of Equations Chap. 6

Since the link current 1 1 A is known, we needed only two loops involving link
currents i, and i2 . Incidentally, in this simple example the links were chosen so that the
link currents are also mesh currents. This is, of course, not the case in general.

The results obtained thus far in this chapter are valid for general networks, which
may be either planar or nonplanar. In the special case of planar networks, as we saw in
Chapter 4, a mesh analysis is possible. In the circuits of that chapter the mesh cur-
rents were independent and were sufficient in number to perform the analysis. We
shall now show that this is the case in general for planar networks.
Let us begin by taking apart the planar circuit with meshes and reconstructing it M
one mesh at a time. The first mesh in the reconstruction has the same number, say k u
of nodes and branches, for the first branch has two nodes, each additional branch
adds one new node, and the last branch adds no nodes since it is tied back to a node of
the first branch. This is illustrated by the graph of four meshes in Fig. 6.10(a). The
first mesh constructed, shown in Fig. 6.10(b), has the same number of branches and

nodes, namely four in this case.

D o

T 9 9

o 6 o

(a) (b) (c)

9 9

o o 9

6 -o o

(d) (e)

FIG U R E 6.1 Planar circuit and its meshes

After the first mesh, each subsequent mesh is formed by connecting branches and
nodes to previous meshes. Each time the number of nodes added is one less than the
number of branches, because each added branch adds one node, except for the last
branch, which is connected to a node of a previously added mesh. This process is
illustrated in Figs. 6.10(c), (d), and (e).
Chap. 6 Independence of Equations 1 27

Thus, if mesh adds k 2 branches, then it adds only k 2 —

the second nodes. l

Similarly, the third mesh adds k branches and k — nodes, and so forth. The last
3 2
l

mesh, the .Uth. adds k M branches and k M — nodes. If in the completed graph the l

number of branches is B and the number of nodes is N, we have

k,+k + 2 ... + kM = B (6.1)

and
k 1 + (k 2 -l) + ... + (k M -\) = N (6.2)

The latter equation may be written

k x
+ k2 + ... + kM -(M- \) =N
which by (6.1) becomes
B - {M - 1) =N
Solving for the number of meshes, we have

M=B-N+ I (6.3)

which is also the number of links in the graph.

The mesh currents, therefore, constitute an appropriate set of currents to com-

pletely describe a planar network. They are the same number as the independent set
in

of link currents, and they are independent since each new mesh contains at least one
branch not in the previous meshes.

EXERCISE
6.4.1 Show that (6.3) holds for the circuits of Probs. 4.19, 4.20, and 4.22.

6.5 A CIRCUIT APPLICATION

As a final illustration in this chapter we shall analyze a moderately complicated

circuit, shown in Fig. 6.1 1(a). Its graph is shown in Fig. 6.1 1(b), where we have selected
a tree shown by the solid lines.
From we see that there are four tree branches labeled v v 2 3v 2 and 10.
the graph l
, , ,

Thus if the tree branch voltage method is used, there will be only two unknowns, v x

and v 2 requiring two equations. There are five link currents in the graph, one of which
.

is known (the 6-A source) and another, the 2v source, which may be expressed in l
1 28 Independence of Equations Chap. 6

IOV

IOV

(a)

(b)

FIGURE 6.11 Network and its graph

terms of the tree branch current i be , and subsequently in terms of other link currents.
Thus if the link current method is used, we must have three equations. Accordingly, we
shall analyze the circuit using tree branch voltages.
To write thetwo necessary equations we first imagine tree branch (b, e) opened,
which separates the tree into two parts. These parts are connected by branch (b, e) and
links (b, c), (c, d), and (a, e) of the 6-A source and link {a, e) of the ^-Q resistor. These
elements are identified by line I. A similar procedure using branch (a, d) leads to line
II. Applying KCL to the currents crossing lines I and II in the direction indicated

yields

+ ^+ 2v,dc 3v„. = (6.4)

and
2v.dc 2v 1
3v 2 =0 (6.5)

Inspection of the graph shows that

v bc = Vi — 3v 2

v dc = v, - 2v 2 10

Vac = »! — 10

Substituting these values into (6.4) and (6.5) and solving for the tree branch voltages,
we have
i>, = -11 V, v2 = -20 V
We may note that nodal analysis is equally easy to apply. The node voltages may
Chap. 6 Independence of Equations 1 29

be expressed in terms of the two unknowns r, and ?\, and thus only two nodal
equations are required. We shall leave the details to the problems (Prob. 6.4).

EXERCISES
6.5.1 Using the graph described for the figure for Ex. 6.3.2, write one KVL equation
and find using the method of Sec. 6.4.
/', Ans. 3 A

6.5.2 Using an appropriate tree for the graph of Prob. 4. 1 6, find the current i flowing
to the right in the 4-fi resistor. (An appropriate tree should not contain the
current sources or the current i. Thus, using the methods of Sec. 6.4, only one
KVL equation is required.) Ans. —9.5 A

6.5.3 Solve Prob. 4.9 using the method of Sec. 6.4. Ans. 4 W

PROBLEMS
6.1 Find a tree, if possible, that contains all the voltage sources and the branches
whose voltages control dependent sources but does not contain current sources
or branches whose currents control dependent sources. Use this tree with an
appropriate graph theory method to find v x
.

PROBLEM 6.1

6.2 Select a tree as described in Prob. 6.1 and use an appropriate graph theory
method to find y t .
1 30 Independence of Equations Chap. 6

2i V

I5V

PROBLEM 6.2

6.3 Solve Prob. 4.21 selecting an appropriate tree and using graph theory methods.

6.4 Find v and v 2

t
in Fig. 6.11 using nodal analysis.

6.5 In the figure shown, all the resistances are 1 Q, element w is also a 1-Q resistor,
elements x and y are independent 1-A current sources directed upward, and
element z is an independent 3-A current source directed to the left. Selecting an
appropriate tree and using graph theory methods, find i. (Note that this circuit

is similar to that of Fig. 6.1 and is thus nonplanar. However, only one loop
equation is required in this case.)

PROBLEM 6.5

6.6 Find v in Prob. 6.5 using graph theory methods if element x is a 4-V source
with positive terminal at the top, y is a 2-V source with positive terminal at the
bottom, z is a 6V source with positive terminal at the left, and w is a 4-V source
with positive terminal at the left.

6.7 The figure for Ex. 6.2.2 and Figs, (a) and (b) shown are examples of graphs of
ladder networks. There is a theorem that states that the number of trees in a
Chap. 6 Independence of Equations 1 31

ladder graph of n branches is the Fibonacci number a n defined by a = a — 1,

, t

a2 = a + a t
= 2, o3 = a x
+ a2 = 3, a4 = a 2 + a 3 = 5, etc. That is, except
for a and a u each Fibonacci number is obtained by adding the previous two.
Verify that the theorem holds for the ladder graphs shown and also for the graph
in the figure for Ex. 6.2.2.

(a) (b)

PROBLEM 6.7

6.8 The graphs shown two basic nonplanar graphs [the one in (b)
in (a) and (b) are
is that in the figure for Prob. 6.5 redrawn]. Branch a-b in (a) is an ideal 6-V
voltage source with its positive terminal at the top and in (b) is a 4-A ideal cur-
rent source directed upward. All the other branches in both figures are 1-Q
resistors. Find i shown in each figure using graph theory methods.

(a) (b)

PROBLEM 6.8

6.9 Find y a and v 2 in Fig. 6.11 (a) using link currents as the unknowns.

6.10 Solve Prob. 4.22 using graph theory methods.

 7
ENERGY-STORAGE
ELEMENTS

Up to now we have considered only resistive circuits, that is, circuits containing
resistors and sources. The terminal characteristics of these elements are simple
algebraic equations which lead to circuit equations that are algebraic. In this chapter
we shall introduce two important dynamic circuit elements, the capacitor and the
inductor, whose terminal equations are differential equations rather than algebraic
equations. These elements are referred to as dynamic because, in the ideal case, they
store energy which can be retrieved at some later time. Another term which is used, for
this reason, is storage elements.
We shall first describe the property of capacitance and discuss the mathematical
model of an ideal device. The terminal characteristics and energy relations will then be
given, followed by derivations for parallel and series connections of two or more
capacitors. We shall then repeat this procedure for the inductor.
The chapter will be concluded with a discussion of practical capacitors and
inductors and their equivalent circuits.

7.1 CAPACITORS

A capacitor is a two-terminal device that consists of two conducting bodies that are
separated by a nonconducting material. Such a nonconducting material is known as a

As a result of the dielectric, charges are not permitted to move from one
dielectric.

conducting body to the other within the device. They must therefore be transported
between the conducting bodies via external circuitry connected to the terminals of the
capacitor. One very simple type, called a parallel-plate capacitor, is shown in Fig. 7.1.
The conducting bodies are flat, rectangular conductors that are separated by the
dielectric material.

132
Chap. 7 Energy-Storage, Elements 1 33

To describe the charge-voltage relationship for the device, let us transfer charge
from one plate to the other. Suppose, for instance, by means of some external circuit,
that we take a small number of electrons, say — Aq, from the upper plate to the lower
plate. This, of course, leaves a charge of Ar/ on the top plate and deposits a charge of
I

— Aq on the bottom plate. Since moving these charges requires the separation of unlike
charges (recall that unlike charges attract one another), a small amount of work is

performed, and the top plate is raised to a potential of say -\-Av with respect to the
bottom plate.
Each increment of charge — Aq that we transfer increases the potential difference
between the plates by Ar. Therefore the potential difference between the plates is
proportional to the charge being transferred. This suggests that a change in the
terminal voltage by an amount Av causes a corresponding change in the charge on the
upper plate by an amount Aq. Thus the charge is proportional to the potential differ-
ence, and we may write sr
q = Cv fatA (7.1)
Vo i

where C is the constant of proportioriaTity, known as the capacitance of the device, in

coulombs per volt. The unit of 1 C/V is known as the farad (F), named for the famous
British physicist Michael Faraday (1791-1867). Capacitors which satisfy (7.1) are
called linear capacitors since their charge-voltage relationship is the equation of a
straight line having a slope of C
It is interesting to note in the above example that the net charge within the capaci-
tor is always zero. Charges removed from the top plate always appear on the lower one
so that the total charge remains zero. We should also observe that charges leaving one
terminal enter the other. This fact is consistent with the requirement that current
entering one terminal must exit the other in a two-terminal device.
Since the current is defined as the rate of change of charge, differentiating (7.1), we
find that
,dv
(7.2)
dt

which is the current-voltage relation for a capacitor.

The symbol for the capacitor and the current-voltage convention which
circuit
satisfy (7.2) are shown in Fig. 7.2. It is apparent that moving a charge of — Aq in Fig.
7.1 from the upper to the lower conductor represents a current flowing into the upper

o-
dielectric +

^c

FIGURE 7.1 Parallel-plate capacitor FIGURE 7.2 Circuit symbol for a

capacitor
1 34 Energy-Storage Elements Chap. 7

terminal. The movement of this charge causes the upper terminal to become more
positive than the lower one by an amount Av. Hence the current-voltage convention of
Fig. 7.2 is satisfied. Obviously, if v is assigned so that the lower terminal is positive or
if the current is assumed to enter the lower terminal, but not both, then a minus sign
must be used in the right-hand side of (7.2). We recall that this is also necessary in the
case of a resistor.
In (7.2) we see that if v is constant, then the current i is zero. Therefore a capacitor
acts like an open circuit to a dc voltage. On the other hand, the more rapidly v changes,
the larger is the current flowing through its terminals. Consider, for example, a voltage
which increases linearly from to 1 V in a"
1
s, given by

v = 0, t<
= at, 0<t<a~ l

= 1, t>a-*

If this voltage is applied to the terminals of a 1-F capacitor (an unusually large value
which is convenient for illustrative purposes), the resulting current is

i = 0, / <
= a, < < t a- 1

= 0, txr 1

Plots of v and i areshown in Fig. 7.3. We see that i is zero when v is constant and
that it is equal to a when v increases linearly. If a is made larger, then v changes more
rapidly and i increases. It is apparent that if a
-1
= (a is infinite), v changes abruptly
(in zero time) from to 1 V.
In general any abrupt or instantaneous changes in voltage, such as in the above
example, require that an infinite current flow through the capacitor. An infinite

current, however, requires that an infinite power exist at the capacitor terminals, which

/\
//
1

l/a t

FIGURE 7.3 Voltage and current waveforms

for a 1-F capacitor
Chap. 7 Energy-Storage Elements 1 35

is a physical impossibility. Thus abrupt or instantaneous changes in the voltage across

a capacitor are not possible, and the voltage is continuous even though the current may
be discontinuous. (It is possible, of course, todraw on paper circuits that contradict
this statement. These circuits are mathematical models that do not sufficiently describe
the entire physical picture, as we shall see in Sec. 7.9.) An alternative statement
concerning abrupt changes in voltages in circuits containing more than one capacitor IX-
is that the total charge cannot change instantaneously (conservation of charge).
Let us now find v(t) in terms of /(f) by integrating both sides of (7.2) between times
/„ and f. The result is

= ±f'f<ft
C
+ t</ ) yJ (7.3)

where v(t ) = q(t )/C is the voltage on C at time f . In this equation, the integral term
represents the voltage that accumulates on the capacitor in the interval from t to /,
whereas r(/ ) is that which accumulates from -co to t The voltage v{— oo), of .

course, is taken to be zero. Thus an alternative form of (7.3) is

v(t) = -L \
idt

In applying this result, we obviously are obtaining the area associated with a plot
of/ from — oc to f. In Fig. 7.3, for example, since v(— oo) = 0, we have

v = -^ |
(0) dt + v(- >o) = 0, t <

Therefore v(0) = 0, and

v = ^\ adt + v(0) = at, 0<t<a~>

C

Similarly, v(\[a) = 1, which yields

.-fjjro* + .(i)-i. t > a'

(Recall that C= 1 F in this example.)

In this example, we see that v and /'
do not necessarily have the same shape.
Specifically, the maximum and minimum values of v and do not necessarily occur at
/

at the same time, unlike the case for the resistor. In fact, inspection of Fig. 7.3 reveals
that the current can be discontinuous even though the voltage is continuous, as
stated previously.
1 36 Energy-Storage Elements Chap. 7

EXERCISES
7.1.1 If a 100-//F capacitor has v(t) = f(t) V, as shown, find (a) /(-3), (b) i(-l), (c)
i(1.5), and (d) /(2.5). (e) Sketch i(f) for all t.

Ans. (a) 50 //A, (b) -50 //A, (c) 0, (d) 0.1 mA

EXERCISE 7.1.1

7.1.2 If a 1000- //F capacitor has i(t) = f(t) mA, where f(t) is given in Ex. 7.1.1, find
(a) u(-2), (b) v(0), and (c) v(2). (d) Sketch v(t) for all t. [Note: v(— 4) = 0.]
Ans. (a) 1 V, (b) 2 V, (c) 0.5 V
7.1.3 A 1-juF capacitor has a terminal voltage of 100 cos 1000/ V. Find /(?)
Ans. -0.1 sin 1000/ A
7.1.4 A 10-//F capacitor has a terminal current of lOe -100 ' mA. If v(0) = — 10 V,
find v(t) for / > 0. Ans. -lOe" 100 '
V

7.2 ENERGY STORAGE IN CAPACITORS

The terminal voltage across a capacitor is accompanied by a separation of charges

between the capacitor plates. These charges have electrical forces acting on them. An
electric field, a basic quantity in electromagnetic theory, is defined as the force acting
on a unit positive charge. Thus the forces acting on the charges within the capacitor
can be considered to result from an electric field. It is for this reason that the energy
stored or accumulated in a capacitor is said to be stored in the electric field.
The energy stored in a capacitor, from (1.6) and (7.2), is given by

Wc(t)= p vidt= r vfc^dt

=C [ vdv = \Cv\t)
Chap. 7 Energy-Storage Elements 137

Since r( — oo) = 0. we ma\ write

wc(t) = $Cv*(t) J (7.4)

From this result, we see that wc (t) >

0. Therefore, from (1.7), the capacitor is a
passive circuit clement. In terms of the charge on the device, (7.1) and (7.4) yield

«e>-i*P> (7.5)

The ideal capacitor, unlike the resistor, cannot dissipate any energy. The energy
which is stored in the device can thus be recovered. Consider, for instance, a l-F
capacitor which has a voltage of 10 V. The energy stored is

if, = \Cv 2 = 50 J

Suppose the capacitor is not connected in a circuit; then no current can flow, and
the charge, voltage, and energy remain constant. If we now connect a resistor across the
capacitor, a current flows until all the energy (50 J) is absorbed as heat by the resistor
and the voltage across the combination is zero. The analysis of such a network is
found in the next chapter.
As has been pointed out earlier, the voltage on a capacitor is a continuous func-
tion. Thus by (7.4) we see that the energy stored in a capacitor is also continuous. This
is not surprising since otherwise energy would have to be transported from one place

to another in zero time, which is an impossibility.

t =
AW-
+

C ZZ

FIGURE 7.4 Circuit illustrating continuity of capacitor

voltage

To illustrate let us consider Fig. 7.4, which

continuity of capacitor voltage,
contains a switch that opened at / = 0, as indicated. (Ideally, a switch transforms a
is

pair of terminals from an open circuit to a short circuit, or vice versa, in zero time.) To
discuss the effect of the switching action we first need to consider two different types of
time / = 0. We shall denote / = 0" as the time just before the switching action and
/ —
+
as the time just after the switching action. Theoretically, of course, no time has
elapsed between 0" and + but the two times represent radically different states of the
,

circuit. Thus v c (0~) is the voltage on the capacitor just before the switch is moved and

v c (0 + ) is the voltage immediately after switching. Mathematically, v c (0~) is the limit of

v c (t) as t approaches zero through negative (/ < +
0) values and v c (0 ) is the limit as /
1 38 Energy-Storage Elements Chap. 7

approaches zero through positive (/ > 0) values. The same notation applies to v l
across the resistor J?,.

Suppose that in Fig. 7.4 we have V = 6 V and v c (0~) = 4 V. Just prior to the
switching action (t = 0") we have v^O') = V — v c (0~) = 2 V. Immediately after the
switch is opened we have V!(0 + ) = 0, since no current is flowing in i?,. However, since
vc is continuous we have

v c (0+) = vc (0~) = 4V

Thus the voltage on the resistor has changed abruptly, but that on the capacitor has
not.
Obviously, one could consider circuits on paper in which capacitor voltages are
forced to change abruptly. For example, if two capacitors having different voltages
are suddenly connected in parallel by a switching action, their resulting common
voltage cannot be the same as both their previous, different voltages. We shall consider
such singular circuits in Sec. 7.9, where we shall see that stored energy has appeared to
change abruptly. The apparent change cannot be accounted for in the lumped circuit
models we are using, but it is a remarkable fact that the lumped models are valid
before and after (though not during) the switching action. Physical circuits, however,
have resistance associated with the capacitor (such as in the leads and the dielectric)
which precludes the infinite currents that must accompany discontinuous capacitor
voltages. These are the type circuits we shall be concerned with, in general.

EXERCISES

7.2.1 A 1-//F capacitor is charged to 100 V. Find the charge and energy.
Arts. 10" 4 C, 5 mJ
7.2.2 In Fig. 7.4, let C= £ F, J?, = R2 = 3 ft, and V= 9 V. If the current in R 2 at
/ = 0" is 1 A directed downward, find at t = 0" and at / = + (a) the charge
on the capacitor, (b) the current in R x
directed to the right, and (c) the current
inC directed downward. Am. (a) 1, 1 C, (b) 2, A, (c) 1, —1 A

7.3 SERIES AND PARALLEL

CAPACITORS

In this section we shall determine the equivalent capacitance for series and parallel
connections of capacitors. As we shall see, equivalent capacitance is in direct analogy
with equivalent conductance.

1/
Chap. 7 Energy-Storage Elements 1 39

Let us first consider the series connection of N capacitors, as shown in Fig. 7.5(a).
Applying K.VL. we find

v = «, + v2 + . . . + vN (7.6)

i
+ ' _ + < _

=^H{ If-

Q)

(a) (b)

FIGURE 7.5 (a) Series connection of N capacitors; (b) equivalent

circuit

From (7.3), this equation can be written

= 1 f 1 f
v -^ \ idt + w,(/ ) + £- / <fc + w 2 (/ )

+ ... + -J- \
idt + V N (t )

1 1
=
(^ + gr + • • • + ^)
J'/
<* + »i(f„) + ^('o) + • • • + tvOo)

or, by (7.6),

In Fig. 7.5(b), we see that

v = ^ I / dt + K'o)

where v(t ) is the voltage on Cs at t = t .

Suppose we require that the circuit of Fig. 7.5(b) be an equivalent circuit for that
of Fig. 7.5(a). Comparing the last two equations, we see that

Cf-> i si i • • i /-i ^ (-> (7.7)

from which we may find the equivalent capacitance C,.

1 40 Energy-Storage Elements Chap. 7

As an example of the utility of (7.6) and (7.7), consider the series connection of
1- and ^-F capacitors having initial voltages of 4 and 6 V, respectively. Then

&- 1

or

C. = 0.25 F
and
v(t ) = 4 + 6= 10 V

Let us now consider the parallel connection of N capacitors, as shown in Fig.

7.6(a). Application of KCL gives

/ = /, + iz + . . . + i N

Substituting from (7.2), we have

^ dv . „ dv ,

CN
dv
dt

dv dv
(C, + c2 + . . . + CN )
di (I
c-)
di

(a) (b)

FIGURE 7.6 Parallel connection of N capacitors; (b) equivalent circuit

In the circuit of Fig. 7.6(b), the current is

If we now require that this circuit be an equivalent circuit for that of Fig. 7.6(a), the
above equations give

C, = C, +C + 2 . . . + C„ = £ C„ (7.8)
n= 1

Thus the equivalent capacitance of TV parallel capacitors is simply the sum of the
individual capacitances. An initial voltage, of course, would be equal to that which is

present across the parallel combination.

Chap. 7 Energy-Storage Elements 141

It is interesting to notice that the equivalent capacitance of series and parallel

capacitors is analogous to the equivalent conductance of series and parallel conduc-
tances.

EXERCISES

7.3.1 Find the equivalent capacitance and the initial voltage. Ans. 0.1 F, —2 V

2V 3V

'/
2 F '/3F

'/5 F 7ZZ 7 V

EXERCISE 7.3.1

7.3.2 Find the equivalent capacitance. Ans. 6 F

o 1 p 1 1

-
v(t) CIF ~2F 3F

o 1
i -4 i

EXERCISE 7.3.2

7.3.3 Find the equivalent capacitance. Ans. 3 fiF

6(iF

EXERCISE 7.3.3
1 42 Energy-Storage Elements Chap. 7

7.4 INDUCTORS

In the previous sections, we found that the electrical characteristics of the capacitor
are the result of forces that exist between electric charges. Just as static charges exert
forcesupon one another, it is found that moving charges, or currents, also influence
one another. The force which is experienced by two neighboring current-carrying
wires was experimentally determined by Ampere in the early nineteenth century.
These forces can be characterized by the existence of a magnetic field. The magnetic
field, in turn, can be thought of in terms of a magnetic flux that forms closed loops

about electric currents. The origin of the flux, of course, is the electric currents. The
study of magnetic fields, like that of electric fields, comes in a later course on electro-
magnetic theory.
An inductor is a two-terminal device that consists of a coiled conducting wire. A
current flowing through the device produces a magnetic flux which forms closed <f>

loops encircling the coils making up the inductor, as shown by the simple model of
Fig. 7.7. Suppose that the coil contains N
turns and that the flux passes through each <f>

turn. In this case, the total flux linked by the turns of the coil is N
= N<f>

This total flux is commonly referred to as the flux linkage. The unit of magnetic flux is

the weber ( Wb), named for the German physicist Wilhelm Weber (1804-1891).
In a linear inductor, the flux linkage is directly proportional to the current flowing
through the device. Therefore, we may write

/ \ X=Li (7.9)

where L, the constant of proportionality, is the inductance in webers per ampere. The
unit of 1 Wb/A is known as the henry (H), named for the American physicist Joseph
Henry (1797-1878).
In (7.9) we see that an increase in i produces a corresponding increase in A. This
increase in A produces a voltage in the JV-turn coil. The fact that voltages occur with
changing magnetic flux was first discovered by Henry. Henry, however, repeating the
mistake of Cavendish with the resistor, failed to publish his findings. As a result, the
famous Michael Faraday is credited with discovering the law of electromagnetic
induction. This law states that the voltage is equal to the time rate of change of the
total magnetic flux. In mathematical form, the law is

dt

Thus, differentiating (7.9), we find

di
V V = ,
L-r- (7.10)
dt
Chap. 7 Energy-Storage Elements 1 43

Clearly, as i increases, a voltage is developed across the terminals of the inductor,

the polarity of which is shown in Fig. 7.7. This voltage opposes an increase in /',
for if
this were not the case, that is, if the polarity were reversed, the induced voltage
would "aid" the current. This physically cannot be true because the current would
increase indefinitely.
The circuit symbol and the current-voltage convention for the inductor is shown in
Fig. 7.8. Just as in the cases of the resistor and the capacitor, if either the current
direction or the voltage assignment, but not both, are reversed, then a negative sign
must be employed in the right-hand side of (7. 10).
An examination of (7. 10) shows that if i is constant, then the voltage v is zero.
Therefore an inductor acts like a short circuit to a dc current. On the other hand, the
more rapidly i changes, the greater is the voltage that appears across its terminals.
Consider, for instance, a current that decreases linearly from 1 to A in i" 1 s,

defined by

i = 1, t<0
= 1
-
bt, 0<t<b~ l

= 0, t> b~ l

A 1-H inductor having this terminal current has a terminal voltage given by

v = 0, t <
= -b, 0<t<b- 1

= 0, t>b~ l

Plots of/ and v for this case are shown in Fig. 7.9. We see that v is zero when / is

constant and is equal to —b when i decreases linearly. If b is made larger, i changes

\ "b '

'/b

FIGURE 7.7 Simple model of an FIGURE 7.8 Circuit FIGURE 7.9 Current and voltage wave-
inductor symbol for the inductor forms for a 1-H inductor
1 44 Energy-Storage Elements Chap. 7

more becomes more negative.

rapidly and v Clearly, if b~ x
= (b infinite), then /

changes abruptly from 1 to A.

In general, abrupt changes in the current, such as in the above example, require
that an infinite voltage appear across the terminals of the inductor. As described in the

case of the capacitor, this requires that an infinite power exist at the terminals of the
inductor, a physical impossibility. Thus instantaneous changes in the current through
an inductor are not possible. We observe that the current is continuous even though
the voltage may be discontinuous.
An alternative statement concerning abrupt changes of the currents flowing in
circuits containing more than one inductor is that the total flux linkage cannot change
instantaneously. That is,Lu L2 L N the sum
for a circuit containing inductors , . . . , ,

A[ + X2 + . . . +
X N cannot change instantaneously. If we compare (7.1) and (7.9),
we see that the flux linkage in an inductor is analogous to the charge on a capacitor.
Thus the sum of the flux linkages given above (conservation of flux linkage) is ana-
logous to conservation of charge. An example employing the conservation of flux
linkage is given in Sec. 7.9.

Let us now find the current i(t) in terms of the voltage v(t). Integrating (7.10) from
time t to / and solving for i(t), we have

1(0 = 4- fv(t)dt + i(t ) (7.11)

In this equation, the integral term represents the current buildup from time t to t,

whereas i(t ) is the current at t . Obviously, /'(/ ) is the current which accumulates
from t = — oo to t , where /(— °°) = 0. Thus an alternative expression is

i(')
H>
= xj V{t)dt

In the application of (7.11), we are obtaining the net area under the graph of v
from — co to /, since /(/„) represents the area from -co to t . In Fig. 7.9, for instance,
since /(0) = 1, we have, for L= 1 H,

i(t) = -j- \ (-b) dt + i(0) = -fa + 1, < < b't

1

Thus /( 1/6) = 0, and

i(t) = -£-(" (0) dt + i(j) = 0, b~ » < /

In the above example, we see that v and /Just as in the case of the capacitor, do not
necessarily have the same variation in time. Inspection of Fig. 7.9, for example, shows
Chap. 7 Energy-Storage Elements 145

that the voltage can be discontinuous even though the current is continuous, as
previously mentioned.

EXERCISES
7.4.1 If a 100-mH inductor has /(/) =/(/)mA, where /(0 is given in Ex. 7.1.1, find
(a) K-3),(b) K-l), (c) r(1.5), and (d) i>(2.5). (e) Sketch v{t) for all t.

Ans. (a) 50, (b) -50, (c) 0, (d) 100 juV

7.4.2 If a 10-mH inductor has v(t) =f(t) mV, where f(t) is given in Ex. 7.1.1, find (a)
i(-2), (b) ;'(0), and (c) i(2). (d) Sketch i(t) for all t. [Note: /(— 4) = 0.]
Ans. (a) 0.1 A, (b) 0.2 A, (c) 50 mA
7.4.3 A 1-mH inductor has a terminal current of 10 sin 100/ A. Find v(t).

Ans. cos 100/ V

7.4.4 Find the current for > t in a 100-mH inductor having a terminal voltage of
l(fe-ioo«vifi(0) = OA. Ans. I - e" 100 ' A

7.5 ENERGY STORAGE IN INDUCTORS

A current i flowing through an inductor causes a total flux linkage A to be produced

that passes through the turns of the coils making up the device. Just as work was
performed in moving charges between the plates of a capacitor, a similar work is
necessary to establish the flux <f>
in the inductor. The work or energy required in this

case is said to be stored in the magnetic field.

The energy stored in an inductor, employing (1.6) and (7.10), is given by

'

= L\' i di = %Li\t)
J-oo f=-oo

Recalling that /(— °o) = 0, we have

wL(t) = lLP{t)3 (7.12)

Inspection of this equation reveals that w L (t) > 0. Therefore, from (1.7), we see
that the inductor is a passive circuit element.
The ideal inductor, like the ideal capacitor, does not dissipate any power. There-
fore the energy stored in the inductor can be recovered. Consider, for example, a 2-H
 r (

1 46 Energy-Storage Elements
Ui- / 2- U* I
C/73P. 7

inductor that is carrying a current of 5 A. The energy stored is

^. V wL = ill = 2
25 J

Suppose the inductor, by means of an external circuit, is connected in parallel with

a resistor. In this case, a current flows through the inductor-resistor combination until
all the energy previously stored in the inductor (25 J) absorbed by the resistor and
is

the current is zero. Solutions of circuits of this type are found in the next chapter.
Since inductor currents are continuous, it follows that the energy stored in an
inductor, like that stored in a capacitor, is also continuous. To illustrate this let us

consider the circuit of Fig. 7.10, which contains a switch that is closed at t = 0, as
indicated. Suppose that /
L (0~)= 2A and / = 3 A. Then by KCL, i x
{0)~) = 3 —2=
1 A. After the switch is closed (t = +
), we have /](0
+
) = since a short circuit is

placed across R,. However, we have

/t (0+) = /L (0-) = 2A
Thus the resistor current has changed abruptly but the inductor current has not.

FIGURE 7.10 Circuit illustrating continuity of inductor

current

An example of a singular circuit for which inductor currents appear to be dis-

continuous is given in Sec. 7.9. As in the case of singular capacitive circuits, the

apparent discontinuity in the energy stored in the inductors cannot be accounted for in
the lumped circuit model. However, lumped circuit theory is valid before and after
(though not during) the switching action. Physical circuits having inductors contain
associated resistance which does not permit the infinite inductor voltages that must
accompany abrupt changes in inductor currents. We shall be concerned primarily
with circuits of this type.

EXERCISES
7.5.1 Derive an expression for the energy stored in an inductor in terms of the flux
2
linkage X and the inductance L. Ans. 2. /2L

7.5.2 A 10-mH inductor has a current of 100 mA. Find the flux linkage and the energy.
Ans. 1 mWb, 50 juJ
Chap. 7 Energy-Storage Elements 147

7.6 SERIES AND PARALLEL INDUCTORS

In this section we shall determine the equivalent inductance for series and parallel
connections of inductors. Let us first consider a series connection of TV inductors, as
shown in Fig. 7.11(a). Applying KVL, we see that

v =v +v +
t 2 . . . + vN

from which we may write

di di di
v=L t
+ L ** +
.

T ,

--- +
,

Lr
>Tt »li

= (L,+L + ...+L N
)f 2
(

dl
•fa) dt

(a) (b)

FIGURE 7.11 (a) Series connection of N inductors; (b) equivalent

circuit

For the circuit of Fig. 7. 1 l(b), the voltage is

Ls
dt

If we now require that this circuit be an equivalent circuit for the series connection,
then the above equations yield -.

V
L s
— Lj -r Li — . . . + LN — E 2-i Ln (7.13)
n= 1

Therefore the equivalent inductance of N series inductors is simply the sum of the
individual inductances. In addition, an initial current would clearly be equal to that
flowing in the series connection.
Let us now consider the parallel connection of N inductors, as shown in Fig.
7.12(a). Application of KCL gives

* =h+ it + . . . + i N (7.14)
148 Energy-Storage Elements Chap. 7

(a) (b)

FIGURE 7.12 (a) Parallel connection of N inductors; (b) equivalent

circuit

Substituting from (7.11), we have

i = i- f V dt + /,(/„) + £" f
« * + ^o) + + fifr + /*(/„)

= + zb)
(r + r+ 2
• •
•

/
v
*+ /l( ' o) + '
2( ' o) + • • • + /;v( ' o)

or, by (7.14),

<'(')=(£ £") JT »# + /(*„)

In Fig. 7.12(b), we see that

i =— v *+ /(* )

where ;(/ ) i s the current in Lp at? = /„. If this circuit is an equivalent network for the

parallel connection, then the above equations require that the equivalent parallel

s—Liii
inductance be given by

L L = + + ...
i^i->,
+ 7L = i:~
dt -

(7.15)
7 7 7

As an example of the application of (7.14) and (7.15), suppose we have two parallel
inductors of 2 and 3 H carrying initial currents of 2 and 1 A, respectively. The parallel
combination could be replaced by a single inductance Lp satisfying

1 1 1

*<, ' 2
or

^ = |H
carrying an initial current of

i{t ) = 2 + 1 = 3A
Chap. 7 Energy-Storage Elements 1 49

In the case of inductors, it is interesting to observe that the equivalent inductance

for series and parallel connections is analogous to the equivalent resistance of series

and parallel resistors.

EXERCISES

7.6.1 Find the equivalent inductance. Ans. 9 H

2H 4H
-/nsin-

i(0) =IA

v(t) 3H

EXERCISE 7.6.1

7.6.2 Find the equivalent inductance and the initial current. The currents shown are
initial values. Ans. 0.1 H, —2 A

''5H

EXERCISE 7.6.2

7.6.3 Find the equivalent inductance. Ans. 3 H

EXERCISE 7.6.3
1 50 Energy-Storage Elements Chap. 7

7.7 PRACTICAL CAPACITORS

AND INDUCTORS*

Commercially available capacitors are manufactured in a wide variety of types, values,

and voltage ratings. The capacitor type is generally classified by the kind of dielectric
used, and its value is determined by the type of dielectric and the physical geometry of
the device. The voltage rating, or working voltage, is the maximum voltage that can
safely be applied to the capacitor. Voltages exceeding this value may permanently
damage the device by destroying or breaking down the dielectric.
Simple capacitors are often constructed employing two sheets of metal foil which
are separated by a dielectric material. The foil and dielectric are pressed together into
a laminar form and are then rolled or folded into a compact package. Electrical
conductors attached to each metal-foil sheet constitute the terminals of the capacitor.
Practical capacitors, unlike ideal capacitors, generally dissipate a small amount of
power. This is due primarily to leakage currents that occur within the dielectric
material in the device. Practical dielectrics have a nonzero conductance which allows
an ohmic current to flow between the capacitor plates. This current is easily included
in an equivalent circuit for the device by placing a resistance in parallel with an ideal
capacitance, as shown in Fig. 7.13. In this figure, R c represents the ohmic losses of the
dielectric and C the capacitance. The leakage resistance R c is inversely proportional to
the capacitance C. Therefore the product of the leakage resistance and capacitance
R C C, a quantity often given by manufacturers, is useful in specifying the capacitor
loss.

Common types of capacitors include ceramic (barium titanate), Mylar, Teflon,

and polystyrene. These types are available capacitance values ranging typically from
in

100 pF to 1 /zF having tolerances of 3, 10, and 20%. Resistance-capacitance products

for these types range from 10 3 Q-F (ceramic) to 2 x 10 6 Q-F (Teflon).
Another type of capacitor which gives larger values of C is the electrolytic capac-
itor. This capacitor is constructed of polarized layers of aluminum oxide or tantalum
oxide and has values of 1 to 100,000 /uF. Resistance-capacitance products, however,
range from 10 to 10 3 Q-F, which indicates that electrolytics are more lossy than
nonelectrolytic types. Also, since electrolytic capacitors are polarized, they must be
connected into a circuit with the proper voltage polarity. If the incorrect polarity is

used, the oxide will be reduced, and heavy conduction will occur between the plates.
Practical inductors, like practical capacitors, usually dissipate a small amount of
power. This dissipation results from ohmic losses associated with the wire making up
the inductor coil and core losses due to induced currents arising in the core on which
the coil wound. An equivalent circuit for an inductor can be realized by placing a
is

resistance in series with an ideal inductor, as shown in Fig. 7.14, where R L represents
the ohmic losses and L the inductance.
Inductors are available with values ranging from 1 //H to 100 H. Large inductance
values are obtained by employing many turns and ferrous (iron) core materials;
hence as the inductance increases, the series resistance generally increases.
Chap. 7 Energy-Storage Elements 1 51

FIGURE 7.13 Simple equivalent FIGURE 7.14 Simple equivalent

circuit for a practical capacitor circuit for a practical inductor

Like the resistor and the operational amplifier, the capacitor can be fabricated in
integrated-circuit form. However, attempts at integrating the inductor have not been
very successful because of geometry constraints and because semiconductors do not
exhibit the necessary magnetic properties. For this reason, in many applications,
circuits are designed using only resistors, capacitors, and electronic devices, such as
op amps.

EXERCISE
7.7.1 Mylar capacitors have a resistance-capacitance product of 10 5 Q-F. Find the
equivalent parallel resistor in Fig. 7.13 for the following capacitors: (a) 100 pF,
(b) 0.1 n¥, and (c) 1 //F. Ans. (a) 10 15 Q, (b) 10 12 fi, (c) 10 11 Q

7.8 DUALITY AND LINEARITY

Let us now determine the dual relationships for the capacitor and the inductor. This
is easily done by considering the current-voltage relations of (7.2) and (7.10) for the
elements. Repeating these equations for convenience, we have

i = C dv (7.16)
dt

and

r d
dt
(7.17)

Comparing the equations, we see that replacing i by v, v by i, and C by L in the first

equation yields the second equation. Therefore it is and the
clear that the capacitor
inductor are dual elements. A similar comparison of the equations for charge and
flux, given by (7.1) and (7.9), shows that these are also dual quantities. A summary of
the dual quantities that we shall consider in the book is given in Table 7.1.
1 52 Energy-Storage Elements Chap. 7

TABLE 7.1 Dual quantities

Voltage Current
Charge Flux
Resistance Conductance
Inductance Capacitance
Short circuit Open circuit
Impedance Admittance*
Nonreference node Mesh
Reference node Outer mesh
Tree branch Link*
Series Parallel
KCL KVL
* Tree branch and link were considered in Chapter 6,

which the reader may have omitted, and impedance and

admittance will be considered in Chapter 1 1.

We are now able to construct dual circuits for networks containing the dual
quantities listed in Table 7.1. Consider, for example, the two-mesh network of Fig.
7.15(a). Using the geometric method of Sec. 4.7, the dual circuit is shown dashed in

Fig. 7.15(a) and is redrawn in Fig. 7.15(b). The solutions for the currents and voltages

(a)

(b)

FIGURE 7.15 (a) Two-mesh network; (b) dual network

Chap. 7 Energy-Storage Elements 1 53

in networks including combinations of resistors, capacitors, and inductors will be

considered in the coming chapters.
Let us now consider the property of linearity for capacitors and inductors, which
are defined by (7.16) and (7.17). respectively. Comparing these equations to that of
(5.2). we and that they are
see that the elements satisfy the proportionality property
therefore linear elements. Thus any combination of independent
circuits containing
sources, and linear dependent sources, resistors, capacitors, and inductors are linear,
and superposition and Thevenin's or Norton's theorems are applicable. These topics
will be considered in later chapters.

EXERCISE
7.8.1 Construct dual circuits for the networks of (a) Fig. 7.4, (b) Fig. 7.10, (c) Fig.

7.12. and(d) Fig. 7.14.

7.9 SINGULAR CIRCUITS*

A circuit inwhich a switching action takes place that appears to produce discon-
tinuities in capacitor voltages or inductor currents is sometimes called a singular
circuit. In this section we shall consider two such circuits, one containing capacitors

and one containing inductors.

Let us consider first Fig. 7.16, where C and C 2 have voltages of 1 and V, respec-
x

tively, prior to the closing of the switch. That is, ^(0") = V and v 2 (0~) = V. We 1

shall now determine n'i(0 ) and vv,(0 + ), the energies stored in C, and C 2 at / =
+ +
.

The energy stored in the circuit prior to closing the switch is

w 1
(p-) = $C v\(0-) = $J1

and
w 2 (0-) = $C2 vl(0-) = 03

The charge on each capacitor at this time is

qi (o-) = c^co-) = i c
and
q 2 (0~) =C 2v2
(0-) = C
After the switch closes, we see from KVL that

i<,(0
+
) = w2 (0 + )
1 54 Energy-Storage Elements Chap. 7

Therefore, from (7.1), the charge on each capacitor satisfies

?.(0
+
) = ? 2 (0
+
)

In addition, when the switch is closed, conservation of charge requires that the total
charge remain constant; hence

?i(0") + <7 2 (0") = ?,(0

+
) + q 2 (0
+
)

Combining these equations, we find that

?,(0
+
) + q z (0
+
) = 2^,(0
+
) d 2q 2 (0+) = C l

or
?>(0
+
) = ? 2 (0
+
) = iC
Substituting into (7.1), we find

v (o
l
+
) = v 2 (o + ) = ^\
Therefore the energy stored in each capacitor at / = +
is

Let us now compare the energy in the system before and after closing the switch.
At / = 0",

Wi(0-) + w2 (0-) = $J
At t = +
,

w:(0 + ) + w 2 (0+) = \ J

We know that capacitors do not dissipate power. What, then, has happened to the
\ from = 0" to =
+
J t ? Looking back over our work, we see that v
t changes x

abruptly from to V at = 0. As pointed out in Sec. 7.1, instantaneous changes in

1 Jr t

the voltage are not possible. Therefore during the infinitesimal time from = 0" to t

t = +
, our mathematical model is not valid. In what has happened is the
reality,

following. When the switch closes at time t = 0, a large current is produced as charges
are transferredfrom Cj to C 2 This rapidly changing current gives rise to an electro-
.

magnetic wave which radiates \ J of energy. The voltage v changes in a short, but l

nonzero, time from 1 to \ V. Our network during this interval does not behave as a
lumped-parameter circuit, and concepts from electromagnetic theory (a later course)
are required for the solution we have described.
Although our circuit model is not valid at the instant the switch closes, the solutions
for the voltages and energies before and after the closing of the switch are correct. This
is due entirely to the fact that the total charge did not change during this time interval.

As pointed out earlier, most circuit models do not permit an infinite current in a
capacitance. Physical circuits normally have a finite value of resistance and inductance
which limit such currents. As a result, the capacitor voltages and energies are con-
Chap. 7 Energy-Storage Elements 1 55

tinuous functions. If. for example, a series resistance is included in the circuit of Fig.
7. 16. the voltage on each capacitor is continuous. That is,

u,(0-) = v,(0
+
)

and
v 2 (0-) = v 2 (0 + )

Analyses for circuits of this type are given in Chapter 8.

For a second example, consider the circuit of Fig. 7.17. Inductors L, and L 2 have
currents of and 1 A, respectively, before the switch is closed at time t = 0. Therefore
/,(0 ) =
1 A and / 2 (0") = A. We shall now determine w,(0 + ) and h' 2 (0 + ), the energy
stored in the inductors at t = +
.

FIGURE 7.16 Circuit containing two FIGURE 7.17 Circuit containing two inductors
capacitors which are switched at time t = which are switched at time t =

The energy stored in the network prior to the closing of the switch is

h^(O-) = iLj/ftO-) = 1 J
and
H' 2 (0-) = jL 2/ 2
(0-) = 0J

The flux linkage of each inductor at this time is

A lv0-) = £,i,(0-) = 2 Wb
and
A,(0-) =L 2/ 2
(0-) = Wb
After the switch is closed, we see from KCL that

i.CO*) = /
2 (o+)

Also, when the switch is closed, conservation of flux linkage requires that the total
flux remain constant; hence

L,/,(0-) + L 2 / 2 (0-) = LMO*) + £zi'a(0 + )

= (L, + L )/,(0 + 2 )
1 56 Energy-Storage Elements Chap. 7

or
2(1) + 1(0) = 3/ t (0 )
+

Therefore
/,(0
+
) = / 2 (0
+
) = fA
Thus the energy stored in each inductor at / = +
is

w I (0
+
) = iL 1 iK0
+
) = |J y

and
w 2 (o+) = ^L 2 m +
) = i j

If we now compare the total energy stored in the network, we see at / = 0" that

h',(0-) + W2 (oA = i j
and at / = +
that
h' 1 (0
+
) +w 2 (0
+
) = fJ

which indicates that % J has been lost by the circuit even though ideal inductors can
dissipate no power.
Looking back over the problem, we see that i (t) changes abruptly from 1 to \ A t

at t— 0. We know, however, that abrupt changes in the current are not possible.
Therefore during the infinitesimal time from t = 0" to t — + our mathematical model ,

is once again not valid. As pointed out previously, a rapidly changing current gives

rise to an electromagnetic wave which radiates energy. In this case ^ J is radiated, and

the current changes from 1 to ^ A in a short, but nonzero, time. Our circuit, of
course, does not behave as a lumped-parameter network during this interval of time.
As in the case of a capacitive circuit having infinite currents, most inductive circuit
models do not permit infinite voltages to occur across an inductor as a result of abrupt
currents. As discussed in the case of the capacitor, physical circuits having inductors
contain a finite value of resistance and capacitance which limit such voltages. The
currents and energies in these circuits are continuous functions. If, for instance, a
parallel resistance is included in Fig. 7.17, the currents are continuous at / = 0.

Circuits of this type are studied in Chapter 8.

EXERCISES

7.9.1 If in Fig. 7.16 C 2 is changed to i F and ^(O - ) to 1 V, find the voltage and the
total stored energy at t = +
. Ans. |V, jJ
7.9.2 If L z = 3 H in Fig. 7.17, find the current and the total energy after the switch
_
is closed if i,(0 ) = 1 A. Ans. \ A, \ J
Chap. 7 Energy-Storage Elements 1 57

PROBLEMS
7.1 Find the charge residing on each plate of a 2-//F capacitor that is charged to
100 V. If the same charge resides on a 1-//F capacitor, what is the voltage?

7.2 Determine the voltage required to store 50 fiC on a 0.25-//F capacitor. What
mA to deliver this charge?
time will be required for a constant current of 25

7.3 The voltage on a 0.01 -//F capacitor increases linearly from V at / = to 100 V
at / = 10 ms. It then decreases linearly to V at / — 30 ms. Find (a) </(5), (b)

/'(15), and (c)/>(15) (arguments in milliseconds), (d) Sketch q{t), i{t), and p(t).

7.4 Determine the current flowing in a 0.1 -jnF capacitor having the following ter-
minal voltages: (a) 10, (b) 100/, (c) 50^"'°', (d) 150 cos 100/, and (e) 100c"' sin /

V.

7.5 The current in a 10-//F capacitor decreases linearly from mA at = / to — 10

mA at 10 ms. It then increases linearly to mA at / = 20 ms. If v{0) = 10 V,
find (a) r(15) and (b)/>(15) (arguments in milliseconds), (c) Sketch v{t) and p(t).

7.6 Find the work required to charge a 0.1 -/zF capacitor to 100 V.

7.7 Determine w(t) for each case of Prob. 7.4.

_ _
= = = 0*
7.8 Determine rj(0 + ) if y lv ) 2r 2 (0 ) 2 V. Find the current at t

flowing to the right through R.

t =

I F 2F v (t)
2

PROBLEM 7.8

7.9 What are the maximum and minimum values of capacitance that can be obtained
from 10 1-yWF capacitors?

7.10 The capacitances shown are all in microfarads. Find Ceq at terminals a-b (see
figure below).

7.11 Derive an equation for current division between two parallel capacitors.

7.12 Derive an equation for voltage division between two series capacitors.

7.13 Determine the flux linkage of a 100-mH inductor carrying a current of 0.1 A.

7.14 The current in a 10-H inductor increases linearly from A at / = to 0.1 A

at / = 0.1 s. It then decreases linearly to A at / = 0.5 s. Find (a) A(0.05), (b)

t<0.30), and (c)/?(0.30). (d) Sketch A(/), v{t), and ;>(/).

 —
1 58 Energy-Storage Elements Chap. 7

1
1 i
\(
it K
24 12

I2P ^20 ~6

4 8
•^ 4
\(
\\
,

'
If
\\
1
2 ^

2^: ~4

1 1

i i 1 i

PROBLEM 7.10

7.15 Determine the terminal voltage for a 10-mH inductor having the following
terminal currents: (a) 10 A, (b) 50t 2 A, (c) 10 sin 377? A, and (d) 100/e'' A.

7.16 The terminal voltage of a 100-mH inductor increases linearly from V at

/ = 0tol0Vat/ = l ms. It then returns immediately to V and repeats this
pattern indefinitely. If /(0) = A, find (a) /(0.5), (b) X{2), and (c) />(0.5) (argu-
ments in milliseconds), (d) Sketch i(/) and p(t) for < < 2 ms.
t

7.17 Determine the terminal current for a 10-mH inductor having a terminal volt-
age of lOe" 100 '
V if i(0) =0A.
7.18 Find the work required to establish a current of 100 mA in a 10-mH inductor.

7.19 Determine w(t) for each case of Prob. 7.15.

+
7.20 Determine /i(0 ) and v(0 h
).

(0") = i A 3H 3 |
i

2
(0") = 2A

PROBLEM 7.20

7.21 Find the maximum and minimum values of inductance which can be obtained
from five 10-mH inductors.

7.22 Determine Z, eq (values shown are in henrys).

7.23 Derive an equation for voltage division between two series inductors.

7.24 Derive an equation for current division between two parallel inductors.

*7.25 A 400- and a 600-pF ceramic capacitor are connected in parallel. The resistance-
capacitance product for a ceramic capacitor is 10 3 fi-F. What is the equivalent
capacitance and parallel resistance for the combination?
Chap. 7 Energy-Storage Elements 159

PROBLEM 7.22

7.26 Determine a dual circuit for the network of (a) Fig. 7.13, (b) the figure for

Prob. 7.10, and (c) the figure for Prob. 7.22.

"7.27 (a) Make R = in Prob. 7.8 and find the energy lost during the switching
action due to radiation, (b) Replace R by an open circuit in Prob. 7.20 and
determine ^(O*).
 8
SIMPLE RC AND Rl CIRCUITS

In this chapter we shall consider simple circuits containing resistors and capacitors or
resistors and inductors, which we shall refer to for brevity as RC or RL circuits,
respectively. The application of Kirchhoff's laws to these networks gives rise to
differential equations that, in general, are more difficult to solve than the algebraic
equations encountered in the previous chapters. Several methods of solving these
equations will be presented.
We shall first concern ourselves with source-free RC and RL circuits, so-called

because they contain no independent sources. As we shall see, the source-free

responses result from energies stored in the dynamic circuit elements and are charac-
terized by the nature of the circuit itself. For this reason the response is known as the
natural response of the circuit.
Following our study of source-free circuits, we shall consider driven RC and RL

circuits in which the forcing or driving functions are constant independent sources
that are suddenly applied to the networks. We shall find that the responses in

Y. these networks consist of two parts, a natural response, similar in form to that of
the source-free case, and a forced response, which is characterized by the forcing
function.

8.1 SOURCE-FREE RC CIRCUIT

We shall begin our study of a source-free network by considering the series connection
of a capacitor and a resistor, as shown in Fig. 8.1. In this circuit, the capacitor is
charged to a voltage of V at time t = t . Since there are no current or voltage sources
in the network, the circuit response is due entirely to the energy which is stored in the

160
Chap. 8 Simple RC and RL Circuits 1 61

capacitor. The energy in this case, at time / = /„, is

w(t ) \CVU (8.1)

Let us now determine v(t) and /(/) for / > t . Applying KCL at the top node, we
find that

or

% + mv = ° < 8 -2 >

which is a. first-order differential equation. (The order of a differential equation is the

order of the highest-order derivative in the equation.)
Numerous methods are available for solving differential equations of the form of
(8.2). One straightforward method is to rearrange the terms in the equation so as to
separate the variables v and t. Then simply integrating the result leads to the solution.
In (8.2) the variables may be separated by first writing

*- ~
_L V
dt RC
from which we obtain
dv

We can now obtain the indefinite integral of both sides of this equation, given by

d
\ i=-TCc\ d, + K <8 4> '

where K is a constant of integration. Completing the integration, we find that

For the solution to be valid for / > / , we see that K must be selected such that the
initial condition of v(t Q ) = V is satisfied. Therefore, at t= we have
f ,

lnv(t ) = \nV = —^+K

or

K= '« v° +
m
 1

1 62 Simple RC and RL Circuits Chap. 8

Substituting the value of K into the solution yields

In v — In Vn = In —= -
t -t
RC
If we now recall the relationship

it is apparent that

(8.5)

In Fig. 8.1, we see that this is the voltage across R; therefore the current is

Vo
e -u-
to) IRC
R

Another method of solving the separated equation of (8.3) is performed by

integrating each side of the equation between its appropriate limits. In our case v has
a value of V at time / , and thus

L--icl* (8.6)

where the integrals in this equation are definite integrals. Performing the integrations,
we have

In v lnV =
RC
which is equivalent to (8.5).

A graph of (8.5) for t = is shown in Fig. 8.2. We see that the voltage is initially

VQ and that it decays exponentially toward zero as / becomes large. The rate at which
the voltage decays is determined solely by the product of the resistance and the
is characterized by the circuit elements
capacitance of the network. Since the response
and not by an external voltage or current source, the response is called the natural
response of the circuit.

i(t)

v(t)

FIGURE 8.1 Source-free RC circuit FIGURE 8.2 Graph of the voltage response for
fo = in the simple RC circuit of Fig. 8.
Chap. 8 Simple RC and RL Circuits 163

The energy stored network at t = r is given in

in the (8.1). As time increases, the
voltage across the capacitor, and hence the energy stored in the capacitor, decreases.
From a physical standpoint, it is apparent that all the energy stored in the capacitor
at t =
must be dissipated by the
t resistor as time becomes infinite. The instantaneous
power absorbed by the resistor is

Therefore the energy absorbed by the resistor as time becomes i

Wj,(oo) = f
p R (t) dt

'"?•-»- fo) ARC

rff

= -$CVl£
icvi r
which is, indeed, equal to the energy initially stored in the network.
-K~

EXERCISES
In Fig. 8.1, Iet/o = V = 10 V, R = kQ, and C= Find (a) w(10- 3 ),
8.1.1

(b) i(10" 4 ), and (c) wc

0,

W s
).
1

Arts, (a) 3.68 V, (b) 9.05

1 /zF.
mA, (c) 49.01 jxi

8.1.2 The capacitor is charged to a voltage of 100 V prior to the closing of the switch.
For t > 0, find (a) v(t), (b) /(/), (c)wc (t), and (d) the time at which v(t) = 50 V.
Ans. (a) lOOe" 100 ' V, (b) -O.Ole" 100 A, '
(c) 5e- 200! mJ, (d) 6.93 ms

1 =

10k

EXERCISE 8.1.2

8.1.3 The switch in the network opens at / = 1. For t> 1, find (a) t>(l
+
), (b) v(t),
and (c) w c (t). (d) Sketch i(t).
Ans. (a) 10 V, (b) lfor"" 1 '
V, (c) 0.5e- 2 "-" J
164 Simple RC and RL Circuits Chap. 8

10 v — 90 n

EXERCISE 8.1.3

8.2 TIME CONSTANTS

In networks that contain energy-storage elements it is very useful to characterize with

a single number the rapidity with which the natural response decreases. To describe
such a number, let us consider the network of Fig. 8.1 and for convenience take
t = in (8.5). The voltage response, in this case, is given by

where V is now the voltage at

Graphs of v{t) for RC — k
— 0.
c t
v (t) = V e~' /RC

= a constant, RC = 2k, and RC = 3k are shown in

D
Fig. 8.3. We see that the smaller the RC product, the more rapidly the exponential
function v(t) decreases. In fact, the voltage for RC = k decays to a specific value in
one-half the time of that required for RC = 2k and in one-third the time of that
required for RC = 3k. It is also clear that the voltage response remains unchanged if
R is increased and C decreased, or vice versa, such that the product RC is the same.
is

For instance, if we double R and halve C, the voltage response is unchanged.

The current in the network of this example is

/(/) = ^R e
-' /RC

v(t)

fcSs^ ^ RC= k

V\^><^ RC 2k =

\y ^^\?^\^ / RC = 3K

0.368 Vr

12
i

i —-
i i

3 t/k

FIGURE 8.3 Graphs of v(t) for various values of RC

Chap. 8 Simple RC and RL Circuits 165

Clearly, the current decreases in the same manner as the voltage. It should be noticed
that changing R and C such that the product of RC remains constant causes a change
in the initial current VJR. The current response, however, still decreases in the same
fashion because e~' RC is unchanged.
The time required for the natural response to decay by a factor of \/e is defined as
the time constant of a circuit, which we shall denote by t. In our case, this requires that

V e~ (' +r) RC = —^-e~

: RC -- v
V o-u+ro/rc
n e-

which yields

x = RC
The units of r are fi-F = (V/A)(C/V) = C/A = s\ In terms of the time constant, the
voltage response is

Go
K
v{f) V e-'^ (8.7)

The response at the end of one time constant is reduced to e' — 0.368 of its 1

initial value. At the end of two time constants it is equal to e'

1
= 0.135 of its initial
value, and at the end of five time constants it has become e~ = 0.0067 of its initial
5

value. Therefore, after four or five time constants, the response is essentially zero.
An interesting property of exponential functions is shown in Fig. 8.4. A tangent to
the curve at t = intersects the time axis at / = t. This is easily verified by considering
the equation of a straight line tangent to the curve at t = 0, given by

i>j = mt + V
where m is the slope of the line. Differentiating v{t), we have

dv
= _YjL e -th
dt x

FIG U R E 8.4 Graph illustrating the relation between a line tangent

to v(t) att = and x
166 Simple RC and RL Circuits Chap. 8

Therefore

dv
m=
di

and

__ ^_o
T

The line intersects the time axis at v = 0, which requires that t = t.

1
It can also be
shown that a tangent to the curve at a time /, intersects the time axis at /, + t (see
Prob. 8.6).

From Fig. 8.4, we see that an alternative definition for the time constant is the time
required for the natural response to become zero if it decreases at a constant rate
equal to the initial rate of decay.
As a second example, let us consider the more general RC circuit of Fig. 8.5(a).
Prior to the opening of the switch, the circuit is in a dc steady-state condition, by
which we mean that the currents and voltages are constant. As we shall see in Sec. 8.4

on driven circuits, a dc steady-state condition is established when a switch, such as

that in Fig. 8.5(a), has been closed for a long period of time.
We see by inspection that the capacitor voltage is equal to the voltage across the
equivalent resistance to the left of the capacitor terminals. The equivalent resistance is

given by

eq ^3+ 10Q
(2 + 4)

isn t = o
AAAr/^ 1

IOOV
%J

(a)

-°-

=z ioov

(b)

FIGURE 8.5 (a) More general RC circuit; (b) its equivalent at t = 0~

Chap. 8 Simple RC and RL Circuits 167

Since the capacitor is an open circuit to steady-state dc, we see in Fig. 8.5(b) that at
/ =
0" the capacitor voltage (by voltage division between R cq and the 15-fi resistor)
is simply

* ->=(totb) x100=40V

The capacitor voltage is continuous at / = 0, so that

y = V (0 + ) = W (0-) = 40 V

For t > the time constant for the network is simply the product of the capaci-
tance and the equivalent resistance, given by

t =R C= eq 10 s

Therefore, from (8.7), with / = and V = 40 V, the voltage is

v{t) = 40e-' /10

EXERCISES

8.2.1 In a series RC circuit, determine (a) T for R = 10 kfi and C= 10 //F, (b) C
for R= 100 kft and x = 10 ^s, and (c) R for v(t) to halve every 10 ms on a
1-jxF capacitor. Ans. (a) 0.1 s, (b) 100 pF, (c) 14.43 kfi

8.2.2 A series RC circuit consists of a 50-kfi resistor and a 0.02-//F capacitor. It is

desired to increase the current in the network by a factor of 4 without changing

the capacitor voltage. Find the necessary values of R and C
Ans. 12.5 kfi, 0.08 //F

8.2.3 The circuit is in a dc steady-state condition at t = 0". For / > 0, find (a)
+
v{0 ), (b) the time constant z, and (c) v,(?).

Ans. (a) 30 V, (b) 20 ms, (c) lOe^ 50 ' V

IOO v^

EXERCISE 8.2.3
1 68 Simple RC and RL Circuits Chap. 8

8.3 SOURCE-FREE RL CIRCUIT

In this section we shall study the series connection of an inductor and a resistor, as
shown in Fig. 8.6. The inductor, in this case, is carrying a current I at time / = t . As
in the case of the source-free RC circuit,
no current or voltage sources in the
there are
network, and the current and voltage responses are due entirely to the energy stored in
the inductor. The stored energy at / = t is given by

w L (t ) = \L1\ J (8.8)

Summing the voltages around the circuit, we have

L^- + Ri =
dt

or

f + |. = (8.9)

This equation is of the same form as that of (8.2) for the RC circuit. We may
therefore solve by separating the variables. Instead, however, let us introduce a
it

second, very powerful method, which we shall generalize in the next chapter. The
method consists of assuming or guessing (a perfectly legitimate mathematical tech-
nique) a general form of the solution based on an inspection of the equation to be
solved. In guessing a solution, we shall include several unknown constants and 4
determine their values so that our assumed solution satisfies the differential equation
and the initial conditions for the network.
A close inspection of (8.9) reveals that i must be a function that does not change its
form upon differentiation; that is, di/dt is a multiple of i. The only function which

satisfies this requirement is an exponential function of /. Realizing also that the

function must be multiplied by a constant to allow for a proper amplitude in the
response, we make an obvious guess that the solution is of the form

i{t) = Ae st
(8.10)

where A and s are constants to be determined. Substituting the solution into (8.9), we
obtain

{' + %)*" o

From this result, we see that our solution is valid if Ae s'

= or if s = —RjL. The
first case is disregarded since, by (8.10), it results in i = for all t and cannot satisfy

/(/„) = I - Thus we take the case s = —RjL, and (8.10) becomes

/(/) - Ae- R,!L

r
Chap. 8 Simple RC and RL Circuits 169

The constant A can now be determined from the initial condition i(t ) = I . This
condition requires that

i(t ) = /„ = Ae- R"'L

or
— V
J „Rlo;L

Therefore the solution becomes

i(t) = I e-«'-"
)/L
(8.11)

In terms of the time constant t. we may write this result as

i(t) = 7 e-"-'°
)/r

where t = L/R. = (V-s/A)/(V/A) = Increasing L,

Evidently, the units of x are H/fi s.

like increasing C RC circuit, increases the time constant. However, an increase

in the

in R, in contrast to the RC circuit, lowers the value of the time constant. A graph of a

typical current response for = is shown in Fig. 8.7.

tQ

The instantaneous power delivered to the resistor in Fig. 8.6 is

p(t) = Ri 2 (t) = Rp e- 2R <'-" )/L

Therefore the energy absorbed by the resistor as time becomes infinite is given by

u-(oo) = )
p(t) dt

= P RIle- 2R{f -" )/L

dt

= -^Hq J

Comparing this result to (8.8), we see that the energy initially stored in the inductor is

dissipated by the resistor, as expected.

i(t)

368 I

v(t)

FIGURE8.6 Source-free RL FIGURE 8.7 Current response of a simple RL circuit fort =

circuit
1 70 Simple RC and RL Circuits Chap. 8

Suppose we had chosen to find the inductor voltage v in the circuit instead of the

current /. Applying KCL, we find, for tn = 0,

v 1 r
,

dt + 7(0) =

which is an integral equation. Differentiating this equation with respect to time, we see
that

1 dv 1 A
+
.

-Rd7 Tv = °
or

This equation is a differential equation that we can solve using one of the methods
discussed previously. It is also interesting to note that if we replace v by iR, we have

di R = r\

Tt
+,

T l

which is equivalent to (8.9), that was obtained using KVL.

Let us now determine i(t) and v{i) in the more general RL circuit of Fig. 8.8. The
circuit is in a dc steady-state condition at t = 0~; therefore, recalling that an inductor
is a short circuit to dc, we have
i(0-) = W = 2A
Since the current in the inductor is continuous at / = 0, we have

j(0 )
+ = /'(()-) =2A
The time constant for the network for t > is clearly the ratio of the inductance
and the equivalent resistance as seen from the terminals of the inductor. The equivalent
resistance is

R eq = 50 + ^lf =
(
100ft

and hence the time constant is3

T = jp = 0.1s

Therefore, since I = +
z(0 )
t
= 2 A and t = 0, we have

/(0 = 2<r 10 'A

Chap. 8 Simple RC and RL Circuits 171

Summing the voltages of the inductor and the 50-Q resistor, the voltage v(t) is given by

V (t)= 10 j- + 50/

= -100e- 10 'K

As a final example, consider the network of Fig. 8.9, which contains a dependent
voltage source. The initial current is z'(0) = I . Summing the voltages around the loop,
we find that

L~
di
dt
+
^ Ri + ki =
or

£+m<=°

IOOV

FIGURE 8.8 More general RL circuit FIGURE 8.9 RL circuit

containing a dependent voltage
source

Comparing this equation to (8.9), we see that the equations are identical if R in (8.9)
is replaced by R— k. Thus, from (8.11) with t = 0, we have

i(t) = I e~
{R+k),/L

The time constant in this case, which is modified by the presence of the dependent
source, is given by

= L
x
R+ k

EXERCISES
8.3.1 In a series RL circuit, determine (a) T for R = 200 D. and L = 10 mH, (b) L
for R = 1 kQ and x = 100 ms, and (c) R for the stored energy in a 0.1 -H induc-
tor to halve every 10 ms. Ans. (a) 50 /is, (b) 100 H, (c) 3.47 ft
 n

172 Simple RC and RL Circuits Chap. 8

8.3.2 A series RL circuit has a time constant of 1 ms with an inductance of 1 H. It is

desired to halve the inductor voltage without changing the current response.
Find the new values of inductance and resistance required.
Arts. 0.5 H, 500 Q.

8.3.3 The circuit is in a dc steady-state condition at t = 0". Determine z'(0

+
) and v(t)
for/>0. Ans. -2 A, 100<r 100 'V

-AW
200 n

i(t) =

IH
~ m 7^—
t

— f

(t) £200
£200. :20on loon iooo 6A

EXERCISE 8.3.3

8.3.4 The circuit shown is in a dc steady-state condition at / = 0~. Find v(t) for t > 0.

Ans. -20e- l0 'V

^)< t =

ion
0.5v(f)
^ 10 V

EXERCISE 8.3.4

8.4 RESPONSE TO A CONSTANT FORCING

FUNCTION

In the preceding sections we have considered source-free circuits whose responses have
been the result of initial energies stored in capacitors and inductors. All independent
current or voltage sources were removed or switched out of the circuits prior to
finding the natural responses. It was shown that these responses, when arising in
circuits containing a single capacitor or inductor and an equivalent resistor, die out
with increasing time.
In this section we
shall examine circuits which, in addition to having initial stored
energies, are driven by constant independent current or voltage sources, or forcing
functions. For these circuits we shall obtain solutions which are the result of inserting
or switching sources into the networks. We shall find that the responses in these cases,

f'
Chap. 8 Simple RC and RL Circuits 173

unlike those of source-free circuits, consist of two parts, one of which is always a

Let us begin by considering the circuit of Fig. 8.10. The network consists of the
parallel connection of a constant current source and a resistor which is switched at
time ! across a capacitor having a voltage v(0~) = V V. For / > 0, a nodal
equation at the upper node is given by

V
c £+ = /o
~R

or

dv ,
I
_ h
dl+RC V C
(8.12)

t=0
X
I ..,
'•I +
R c~ ^ v(t)
) \
;

FIGURE 8.10 Driven RC network

Equations of this type that have constant forcing functions (7 in this case) can be
solved by the method of separation of variables. We may first write (8.12) in the form

dv Rh
di RC
Multiplying both sides by dt/(v — RI ) and forming indefinite integrals, we have

where A' is a constant of integration. Thus, performing the integrations, we obtain

ln(i'-i?/ )= ~ +K
This result can be written as

v - RI = e
(l IRC) + K

or, solving for v, we find

v = Ae~ Rh (8.13)

where we have taken A = e

K a
, constant to be determined by the initial condition of
the circuit.
From (8.13) we see that the general solution for the voltage response consists of
1 74 Simple RC and RL Circuits Chap. 8

two parts, an exponential function and a constant function. The exponential function
is of the identical form as that of the natural response in a source-free circuit composed
of R and C. Since this part of the solution is characterized entirely by the RC time
constant, we shall refer to it as the natural response v„ of the driven circuit. As in the
case of the source-free circuit, this response approaches zero as time increases.
The second part of the solution, given by RI , bears a close resemblance to the
forcing function I . In fact, as time increases, the natural response disappears, and the
solution is simply RIQ . This component is due entirely to the forcing function, and we
shall call it the forced response v f of the driven circuit.
Let us now evaluate the constant A in (8.13). As in the case of the source-free
circuit, its value must be selected so that the initial voltage in the circuit is satisfied. At
t = +
, we see that

v(0+) = v(0~) = V
Therefore, at t = +
, (8.13) requires that

V =.A + RI
or
A = V - RI

Substituting this value of A back into our solution yields

v(t) = RI + (V - RIQ )e-"* c (8.14)

We should observe in constant A is now determined not only by

this solution that the
on the capacitor but also by the forcing function I
the initial voltage (or energy) .

Graphs of v n v f and v are shown in Figs. 8.11(a) and (b). In (a) the natural
, ,

response v n for V — RI > and the forced response v f are shown. In (b) the com-
plete response is shown.
The current in the capacitor for t > is

M) R*0
r -t/RC
dt R
he resistor is

l
R
= *0 l C = /o
_|_ *<) ^"0 „-tlRC

It is interesting to note that the resistor voltage has changed abruptly from RI at
/ = 0" to V at t = +
. The capacitor voltage, as previously pointed out, is continuous.
The solutions that we have encountered so far in this chapter are often referred to
in other more descriptive terms. Two such terms that are very popular are the transient
response and the steady-state response. The transient response is the transitory portion
of the complete response which approaches zero as time increases. The steady-state
response, on the other hand, is that part of the complete response which remains after
the transient response has become zero.
Chap. 8 Simple RC and RL Circuits 1 75

n "' '«

v -Rl

(a)

Vo

RI,

(b)

FIG U R E 8.1 1 Graphs of voltage response for the driven RC

network of Fig. 8. 10. (a) Natural and forced responses; (b)
complete response

In our example, we see that the transient response and the natural response are
identical, as are those for the source-free circuits of the previous sections. The steady-
state response is therefore identical to the forced response. In our example this
response is RI , a dc value which we defined as a dc steady-state condition in Sec. 8.2.
We should not conclude from the above discussion that the natural and forced
responses are always the transient and steady-state responses. If the forcing function
is a transitory function, for instance, the steady-state response is zero, as we shall see
in the next chapter. In this case the complete response is the transient response.

EXERCISES
8.4.1 In Fig. 8.10 I = 100 mA and R = 1 kQ. For t > 0, determine (a) V such
that the natural response is zero, (b) /c (10
-4
) for V = 50 V and C= 1 /zF,
and (c) C for VQ = -50 V such that v(W~ 3 ) = 0.

Ans. (a) 100 V, (b) 45.2 mA, (c) 2.47 fiF.

1 76 Simple RC and RL Circuits Chap. 8

8.4.2 Find v(t) for t > if v(0~) = 5 V. Am. 10 - 5e- 100 'V

t =
ioi<n

iov I//F ^ v(t)

D l/f i A-

EXERCISE 8.4,2

8.5 THE GENERAL CASE

The equations describing the networks of the previous sections are all special cases of a
general expression given by

(8.15)

where y the unknown variable, such as v or

is and P and Q are constants. If, for /',

instance, we compare this equation to that of (8.2) for the source-free circuit of Sec.
8.1, we see that y = v, P = /RC, and Q = 0. The same relations are valid for the
1

forced RC circuit of Sec. 8.4 except that Q= IJC.

A solution of (8.15) can be found by separation of variables. However, let us.
introduce another method which also is applicable when Q is a function of time, an
important case in later chapters. This method, known as the integrating factor method,
consists of multiplying the equation by a factor which makes its left-hand side a
perfect derivative and simply integrating both sides.
Let us begin by considering the derivative of a product, given by

£0«") = $«" + !*"

Idy
(!+'>)<

We see that if we multiply both sides of (8.15) by e Pl the left-hand side becomes a ,

perfect derivative. Thus performing this multiplication, we have

(ye"') = Qe>
dt

Integrating both sides of the equation, we find

ye f = Qe p '
dt
J
Chap. 8 Simple RC and RL Circuits 177

where A is a constant of integration. Solving for y, we have

y = e
-p
'

I"
Qe p '
dt + Ae~ (8.16)

which is valid, of course, if Q is a function of time or a constant. Taking Q to be a

constant, however, we obtain
iv\

(8.17)
/\*s\ ft
where y„ =
Ae~ Pt and y f = Q/P are the natural and forced responses. We observe
tKaTTn has the same mathematical torm as the source-free natural response and that
y f is always a constant which is proportional to Q. In addition, \jP is the time constant
in the natural response.
To illustrate the method, let us find i2 for t > in the circuit of Fig. 8.12, given that
':(0) = 1 A. Although the somewhat complex combination of elements, the
circuit is a
solution of (8.17) is valid since the network contains a constant forcing function and a
single energy-storage element (the inductor). The loop equations for the circuit are

8i, - 4/ 2 = 10

a\
-4/, + 12/j =
dt

FIGURE 8.12 Driven RL circuit

Eliminating i t
from these equations, we find that

di 2
+ 10/ 2 = 5
dt

Comparing this equation with (8.15), we see that P= 10 and Q= 5. Hence (8.17)
yields

i2 = Ae~ 10 '
+ £

Applying the initial condition, we have

i
a (0)
= A + $=1
1 78 Simple RC and RL Circuits Chap. 8

Therefore A = £, and the solution is given by

ii = y- io,
+

EXERCISES
8.5.1 Find v(t) for t > if v(0) = 0. Ans. 12(1 -<?-"") V

EXERCISE 8.5.1

8.5.2 Find iz (t) in Fig. 8.12 if the source voltage is lOe"' V and / 2 (0) = 1 A.
Ans. \e-< + %e~ 10 '
A
8.5.3 Find /(?) for / > if v(0) = 12 V. Ans. 2 - <r 5 < A

3A

EXERCISE 8.5.3

8.6 A SHORTCUT PROCEDURE

Let us now introduce a shortcut procedure that is very useful for finding the currents
and voltages in many circuits. The technique involves formulating the solution by
merely inspecting the circuit. Consider, for instance, the example of the previous
section (Fig. 8.12). We know that
k = hn + Hf

where i 2n and i
2f are the natural and forced responses, respectively. Since i 2n has the
Chap. 8 Simple RC and RL Circuits 179

same form as the source-free response, we can look at the network in the absence of the
forcing function (i.e., make the 10-V source zero by replacing it by a short circuit), as
shown in Fig. 8.13(a). The natural response is then

'In

The forced response is constant; therefore, insofar as the forced response is

concerned, it does not matter at what time we look at the circuit. We may choose then
to look at the circuit in the steady state when i 2 „ is zero. At this time the inductor is a
short circuit, as shown in Fig. 8.13(b), and

'2/ \

Therefore

i2 = Ae- i0 '
+ £

The constant A is now determined as before from the initial condition, z'
2 (0)
= 1.

(a) (b)

FIGURE 8.13 Circuits for finding the response of Fig. 8.12. (a) Circuit
for finding ij. n ; (b) circuit for finding iif

A word of caution is appropriate at this point. When evaluating the constant A,

the student should always apply the initial condition to the complete response — never
to the natural response alone — because the initial condition is always given for the
current, not for a part of it.

As a second example, let us find i{t) for t > in Fig. 8.14, given v(0) = 10 V. The
current is given by

* = K +h
60 a

v(t)

FIGURE8.14 Driven RC circuit

1 80 Simple RC and RL Circuits Chap. 8

To obtain /„ we note that it has the same form as v„, the natural response of the
capacitor voltage. In fact, the natural response of every current or voltage in the
circuit has the same form as v„. This is true because all the other currents and voltages
in the source-free circuit may be obtained from v„ by applying one or more of the
operations of addition, subtraction (in KCL and KVL), differentiation, and integra-
tion, none of which changes the nature of the exponential e~"\ Examining the source-
free circuit (the current source open circuited), we see that the time constant for the
capacitor voltage is x — 100 s. Therefore

= Ae -(/100

In the steady state the capacitor is an open circuit, and the forced response is, by
inspection,

// = A 1

Therefore

i(t) = Ae-' /10 °

+ 1

To evaluate A, we must find the value of /(0 + ). Since v(0) = v(0 + ) = 10 V, summing
the voltages around the right-hand mesh, for t = +
, we have

-40/(0 + ) + 60[1 - /(0

+
)] + 10 =
or

/(0
+
) = 0.7

Substituting this initial current into our solution, we find that

0.7 = 1 + A
Therefore A = —0.3 and

/(/)= 1 -0.3<?-'/10 °

Before concluding this section let us, as a final example, determine /'(/) and v{t) in

the circuit of Fig. 8.15(a). The network is in a dc steady-state condition at / = 0~ with

the switch open; therefore the inductor and capacitor are a short circuit and an open
circuit, respectively, at this time. The capacitor voltage is equal to the voltage that
appears across the 20-fl! resistor, and the inductor current is equal to the current in the
15-fi resistor. By current division, the currents in the 15- and 20-f2 resistors are easily
shown to be 2 and 3 A, respectively. Thus

/(0-) =2A
and
v(0~) = 60 V
Chap. 8 Simple RC and RL Circuits 181

When the switch closes at / = 0, we observe that nodes a and b are short-circuited
together, and we can redraw the network as shown in Fig. 8.15(b). It should be
noticed that the 30-fl resistor need not be included in this circuit because the switch is

a short circuit across its terminals. The combination is equivalent to a 30-fi resistor in
parallel with a 0-Q resistor, which, of course, is Q or a short circuit.

io n
-/VW-^
f-

I H 3on:
*A
t = 3F v(t) 2on
© bA

(o)

(b)

FIG U R E 8.1 5 (a) Circuit containing an inductance and a capacitance;

(b) equivalent circuit for t >

Let us next consider the current i(t) leaving a in Fig. 8.15(b) through the 15-Q
resistor. From KCL, this same current must enter a through the 1-H inductor; hence
no current flowing in the circuit to the left of a can enter the other part of the circuit to
the right of o, and vice versa. Thus after the switch is closed, the network reduces to
two independent circuits, each of which can be solved individually.
The first circuit, consisting of the 1-H inductor and the 15-Q resistor, is simply a
source-free RL network having /(0 ) = z'(0") = 2A. Therefore
+

i{t) = 2e' is '

The second circuit, composed of all the elements to the right of a, is simply a driven
RC network with r(0 + ) = v(Q~) — 60 V. From our shortcut procedure, we find

V {t) = 40 + 20?-' /20

The shortcut procedure presented in this section is applicable also to circuits

containing dependent sources. However, no savings in time or effort usually result
because the circuit equations still have to be written for the source-free and dc steady-
state cases.
182 Simple RC and RL Circuits Chap. 8

EXERCISES
8.6.1 Find v(t) for / > if the circuit is in dc steady state at / = 0~.
5
Arts. 2.5(1 -e-'° ')V

10V 10mA

EXERCISE 8.6.1

8.6.2 Find /(/) and v(t) for / > if the circuit is in dc steady state at t = 0~.

Arts. -0.45e- 10 'mA, -45e- l0 "V

Aaa,
lOOkO

i(t)
r t
lOOmH
nnnp
v(t)
— »
l.5kn
vw-

3kO: t=0 3kO (l) 90V

\/jF
K

EXERCISE 8.6.2

8.7 THE UNIT STEP FUNCTION

In the previous sections we have analyzed circuits in which energy sources have been
suddenly inserted into the networks. At the instant these sources are applied the
voltages or currents, at the points of application, change abruptly. Forcing functions
whose values change in this manner are called singularity functions. There are many
singularity functions that are useful in circuit analysis. One of the most important is

the unit step function, so named by the English engineer Oliver Heaviside (1850-1925).
The unit step function is a dimensionless function which is equal to zero for all

negative values of its argument and which is equal to one for all positive values of its

argument. If we denote the unit step function by the symbol u{t), then a mathematical
description is

u(t) = 0, t<
(8.18)
= 1, t>0
Chap. 8 Simple RC and RL Circuits 183

From a graph of (8.18), shown in Fig. 8.16, we see that at t = 0, u{t) changes abruptly
from to 1. Some authors define u(0) to be 1, but we are leaving w(/) undefined at
( = 0.

u(t)

FIG U R E 8.1 6 Graph of the unit step function u(t)

Since the unit step function is a dimensionless quantity, a voltage or current

source can be obtained by multiplying u(t) by a voltage or a current. A voltage step of
V volts is represented by the product Vu(t). Clearly, this voltage is for t < and V
volts for t > 0. A voltage step source of V volts is shown in Fig. 8.17(a). A circuit
which is equivalent to this source is shown in Fig. 8.17(b). A short circuit exists for
t < and the voltage is, of course, zero. For t > 0, a voltage V appears at the
0,

terminals. We have assumed in our model that the switching action occurs in zero
time.

o t =o

J<
Vu(t)(+) V -^
Vu(l)

(a) (b)

FIGURE8.17 (a) Voltage step source of V volts; (b) equivalent circuit

Equivalent circuits for a current step source of /amperes are shown in Fig. 8.18.
An open circuit exists for / < 0, and the current is zero. For t > 0, the switching
action causes a terminal current of / amperes to flow.

Iu(t)
t=o

<
;u(,)
(t)
©
(a) (b)

FIGURE 8.18 (a) Current step source of I amperes; (b) equivalent

circuit
184 Simple RC and RL Circuits Chap. 8

The switching action shown in Fig. 8.17 can only be approximated in actual
circuits. However, in many cases, it is not necessary to require that the voltage source
be a short circuit for / < 0, as we shall see in the next section. If the terminals of a
network to which the source is to be connected remain at V for / < 0, then a series
connection of a source V and a switch is equivalent to the voltage step generator, as
shown in Fig. 8.19.

t =

r +
v(t)
Network
(v(t)=0 Vu(t)( Network
for t<0)

(a) (b)

FIGURE 8.19 (a) Network with V applied at t = 0; (b) equivalent circuit

Equivalent circuits for a current step generator in a network are shown in Fig.
.20. In each case the current in the network terminals must be zero for / < 0.

Network
w-o (

for
i(t) =
t <0)
Iu(t)(f Network

(o) (b)

FIGURE 8.20 (a) Network with I applied at t = 0; (b) equivalent circuit

Let us now return to our definition of the unit step function given in (8.18). We may
generalize this definition by replacing t by t — t in the three places that it occurs,
which results in

u(t - /„) = 0, t < t

(8.19)
= 1. t>t

The function u(t — t Q ) is the function u(t) delayed by t seconds, as shown in Fig. 8.21.

Multiplying (8.19) by Kor /gives us a voltage step source or a current step source
whose value changes abruptly at time t . Equivalent networks for these sources are
obtained in Figs. 8.17-8.20 by taking all actions related to switching to occur at
t = t .
Chap. 8 Simple RC and RL Circuits 1 85

u(t-t n )

FIGURE 8.21 Graph of the unit step function u(t - t )

Step functions are very useful in formulating more complex functions. Take, for
instance, the rectangular voltage pulse of Fig. 8.22(a). From this figure, we see that

/ <
= V, < < t t

= 0, t > t

Since u(t) becomes 1 for / > and — u(t — t ) becomes —1 for t > t , it is obvious
that

»,(/) = V[u(t) - u{t - /„)] (8.20)

To check this result, we see that, for t < 0,

Vl (t) = V(0 - 0) =
for < < t t ,

Vl (t) =V(l-0)=V
and for / > t ,

v t (f) = V{\ - 1) =
Now suppose that we wish to produce a train of these pulses with one occurring
every T seconds, where T> / , as shown in Fig. 8.22(b). Such a wave is called a
square wave. The first pulse is given by (8.20). The second pulse is simply the first pulse
delayed by T seconds. Therefore replacing t by t — T \n (8.20). we have

Pulse 2 = V{u(t -T)- u[t - (T + /„)]}

The (n ~ 1 )th pulse in the pulse train is the first delayed by nT, and therefore

Pulse n + = 1 V{u(t - nT) - u[t - n(T + /„)]

To obtain an expression for the square wave for all / > 0, we add the above expres-
sions and obtain

*>2(0 = V2 {u(t - nT) - u[t - n(T + f )]} (8.21)

J>=0
1 86 Simple RC and RL Circuits Chap. 8

v.(t)

(a)

v (t)
2
r
1

K T T+ t 2T 2T + t„

(b)

FIGURE 8.22 (a) Rectangular pulse; (b) square wave

Waveforms like those of (8.20) and (8.21) are very common in digital circuits such
as those in the digital computer.

EXERCISES
8.7.1 Using unit step functions, write an expression for the current i'(f) which satisfies

(a) i(t) = 0, t <

= —4 mA, t >
(b) i(t) = 0, t < -10 ms
= 2 A, -10 ms <t< 100 ms
= -1 A, 100ms < t
(c) i(0 = 4 M> / < 1 s

= 0, / > 1 s

Am. (a) -4«(/) mA, (b) 2u(t + lO"*) - lu{t - 10" 1

) A, (c) 4u(-t + 1) fiA

8.7.2 Sketch the voltage given by

v(t) = (t + !)«(/ - 1) - ltu{t) + Qt - l)u(t - 2)

Chap. 8 Simple RC and RL Circuits 187

8.7.3 Using unit step functions, write an expression for v(t) for — oo < < oo.
t

Ans. 10 sin 2nt\u(t) - u(t - 1)]

10 sin 2nt

8.8 THE STEP RESPONSE

The step response is the response of a circuit having only one input which is a unit
step function. The response and step input can, of course, be a current or a voltage.
The is due entirely to the step input since no initial energies are present
step response
in the dynamic circuit elements. This is the case because all the currents and voltages
in the network are zero at / = 0~ due to the fact that the step function is zero for
— oo < / < 0. In this section we shall consider step responses for which the step input
may or may not be a unit step.
As an example of a step response, let us find v(t) in the simple RC circuit of Fig.
8.23(a) having a voltage step input of Vu{t) volts. Applying KCL, we have

r dv v - Vu(t)
_ n
C
dt
+ "
R ~
U

or

dV V V
+ RC = RC u{t)
.
f.y.

dl

For t < 0, this equation become^

The solution of an equati

Applying the initial con

v{t) = 0, t <
188 Simple RC and RL Circuits Chap. 8

which confirms our assertion that the response is zero prior to the change in the
input.
For t > 0, our differential equation is

dv V
dt
r RC ~ RC
We know that

V = vn + v,
where

v, ,
= Ae~'/RC

and, by inspection,

vf = V
Therefore

v= V+ Ae' t/RC

The initial condition v(0 + ) = v(0~) = requires that A = —V, and therefore our
solution for all / is

v(t) = 0, t<0
= V[\ -t/RC
I t>0
This may be written more concisely, using the unit step function, as

v(t) = V{\ - e-"RC )u{t)

The voltage across the resistor and the capacitor is zero for t < 0. Therefore an
equivalent circuit for our network issatisfied by the circuit of Fig. 8.23(b) provided we
specify that v(Q~) — 0.

Vu(t)(jM N/(t) v -=- Ttz v(t)

(a) (b)

FIGURE 8.23 (a) RC circuit with a voltage step input; (h) equivalent circuit

As a second example, let us find v 2 {t) in the circuit of Fig. 8.24, consisting of a
resistor, a capacitor, and an op amp. A nodal equation at the inverting terminal of the
Chap. 8 Simple RC and RL Circuits 1 89

op amp is given by

7f +C f= °

since the node voltage and the current of the inverting terminal are both zero.
Therefore

FIGURE 8.24 Integrator

Integrating both sides of this equation, we find

»i(0= ~w\ Vldt + K

where K is a constant of integration. The response v 2 (t) is proportional to the integral
of the input voltage i',. Thus the circuit is called an integrator.
Let us now determine the step response of the network by taking i>, = Vu(t). In
this case,

V_
v 2 (0
RC
u(t) dt +K
V_
RC
tu(t) +K

The capacitor voltage at t = + is zero, which requires that v 2 (0 + ) = 0. Thus

K = 0, and our solution can be written

=-
v2
m tu(t)

A graph of v 2 is
shown in Fig. 8.25. This function is called a ramp function with a slope
of - Vj RC.
As a final example, let us find the voltage v{t) in the network of Fig. 8.26, given that
190 Simple RC and RL Circuits Chap. 8

FIGURE 8.25 Step response of an integrator

2H< v(t)

FIGURE 8.26 RL circuit driven by i g (t)

j'(0") = 0. The forcing function for the circuit is the current pulse,

i
g {t)= \0[u(t)-u(t- 1)]A

which is shown in Fig. 8.27(a).

Examining the circuit, we see that a zero initial condition [/(0") = 0] exists and
that at / = a current step of 10 A is applied. Thus the network response is identical
to the step response until time t — At this time, the forcing function g (t) becomes
1 s. i

zero, and the response is simply the source-free response resulting from the energy
which the inductor has accumulated during the interval of the current pulse. There-
fore we see that the solution of the problem involves finding the step response of the
circuit for / < 1 s and then finding the source-free response for t > 1 s.
We know that the step response is of the form \

v = v„

where
v„ = Ae-*">'
/L
= Ae~ s " s Ae

and, by current division and Ohm's law, that

„ _ 2
r (3)(io) n = 12
Chap. 8 Simple RC and RL Circuits 1 91

Combining these equations, we have

v = Ae~' + 12

Since /(0 + ) = i(0~) = 0, then v(0 + ) = v(0~) = and A = -12. Therefore

v = 0, / <
= 12(1 - e-'), < < / 1

For t > 1, we know that w is of the form

t; = Ar'

=
\
At t 1". our solution for the step response gives

w(l")= 12(1 -e" 1

)

Since the inductor current is continuous, f(l~) = v(\+) = 12(1 — e' 1

). Therefore

v(\
+
) = Be' 1
= 12(1 - <?">)

or

B= 12(1 -*-»)*

(t)

IOA

FIGURE 8.27 (a) Forcing (unction i

g (t); (b)
response of an RL cir-

cuit to i
g (t)
1 92 Simple RC and RL Circuits Chap. 8

and our solution becomes

v= 12(1 - r'K 1
'- 11
, t > 1

The solution for all t can be written as

v(t) = 12(1 - e-')[u{t) - u(t - 1)] + 12(1 - e-')e- -"u{t (,

- 1) (8.22)

A graph of this response is shown in Fig. 8.27(b).

EXERCISES

8.8.1 Find the step response i{t) to the voltage step v(t) = 20u(t) V.
Am. (1 -0.5£>- 100 'M0mA

</0) «,(t)

EXERCISE 8.8.1

8.8.2 Find the step response v 2 {t) to the voltage step v t (t) = I0u(t) V. Sketch v 2 (t)
for < t < 1(T 2 s. Ans. (10 + 100f)«(/)

i0kn

EXERCISE 8.8.2

8.8.3 Find v x
{t) in Ex. 8.8.1 if v(t) =
20[u(t) - u(t - 0.1)] V. Sketch Vl (t).
Ans. 10[1 -
e- i00 '][u(0 «(' ~ - 0.1)] + 10(e 10 - l)e- 100 'u(t - 0.1) V
Chap. 8 Simple RC and RL Circuits 193

8.9 APPLICATION OF SUPERPOSITION

In this section we shall consider the use of superposition for obtaining solutions of RC
and RL circuits containing two or more independent sources. As a first example, let us
consider the circuit of Fig. 8.26 of the previous section. The value of the independent
current source is given by

i, = I0u(t) - \0u(t - l)

This source is equivalent to a pair of independent current sources connected in

parallel. Thus if we let

', = 'i + h

where ;', = lCh/(/) and i2 = -\0u(t — 1), the circuit of Fig. 8.26 can be redrawn as
shown in Fig. 8.28.

FIGURE 8.28 Equivalent circuit ot Fig. 8.26

From the principle of superposition, we may write the output voltage as

v = v x
+ v2

where i^ and v 2 are the responses due to /, and 2 respectively. In our previous z ,
solu-
tion we found that the step response due to the current step /, was given by

v, = 12(1 - e-')u(t)

Next, we need the response v 2 due to i2 . We note that i 2 is simply the negative of /,

delayed 1 s in time. Therefore v 2 is obtained from v t by multiplying v t by —1 and

replacing t by / — 1. The result is given by

v2 = -12(1 - e- u -")u{t - 1)

Our solution is now given by

v = 12(1 - e-')it(t) - 12(1 - e- '-'>(/

(
- 1) (8.23)

which is equivalent to (8.22) (see Ex. 8.9.1).

194 Simple RC and RL Circuits Chap. 8

As a second example, let us consider the RC network of Fig. 8.29, which contains
two independent sources and an initial capacitor voltage v(0) — V Applying KVL .

around the left mesh, we find that

(R, +R 2 )i +
l

-^ idt+V =V 1
-R 2IX

FIGURE 8.29 RC network

Multiplying each term in this equation by a constant K, we have

(/?, +R 2 ){Ki) + -^ (Ki) dt + KV = KV - R X 1 (KI l )

Clearly, the current response becomes Ki when the independent sources and the
initial capacitor voltage are multiplied by the factor K, which demonstrates the
proportionality property for a linear network. This result is easily extended to any
linear circuit containing one or more capacitors. Thus initial capacitor voltages can be
treated as independent voltage sources. In a similar manner, it is easily shown that
initial inductor currents can be treated as independent current sources.
We shall now employ superposition to determine the voltage v by finding v u v 2 ,

and v 2 the responses due to V ,I U and V respectively. The circuit for finding v is
,
X
,
1

shown in Fig. 8.30(a). This is a simple driven RC circuit having a zero initial capacitor
voltage. The solution is given by

Vl = K,(l - e
- ,nRl+R ' )c
)

The circuit for finding v 2 is shown in Fig. 8.30(b). This again is a simple driven
circuit having a zero initial capacitor voltage, for which we find

v2 = -* 2 /iO - e
-t/{R + X i )C
l

In the circuit of Fig. 8.30(c), the voltage v z is simply the source-free response
resulting from the initial capacitor voltage. Since v 2 (0) = V , we find that

— V
V e -tHRi + Ri)C
Chap. 8 Simple RC and RL Circuits 195

(a) (b) (c)

FIGURE 8.30 Circuits for finding (a) V\, (b) v 2 . and (c) v 3 for an RC network

Therefore the entire response is given by

V = +v +v
i>, 2 3

= v - rj, +
x (/? 2 /, - v + r )*-'
x
/0, ' +JWC
(8.24)

Inspecting our solution, we see that it consists of a forced response v f and a

natural response which we could have anticipated. An alternative method of
v„,

finding the solution is to obtain v f and vn as described previously. Superposition can,

,

of course, be used in finding v f In our case, we know that

.

vf = v lf + v 2f

From Figs. 8.30(a) and (b) we see, by inspection, that

and

Vif = -RJx
Therefore the forced response is

vf = V, - i? 2 7,

The natural response, obtained from the source-free circuit [Fig. 8.30(c)], is given
by
y — U+Rt)C
^g-'/(i!l+«2)C

Therefore

v=V - x
R 2 I, + Ae-'/(* l+Rt)c (8.25)

Since v(0) = V , we have

A = RJ, Vo

which substituted into (8.25) gives (8.24).

 t->o

1 96 Simple RC and RL Circuits

EXERCISES
8.9.1 Reduce (8.22) to the form of (8.23).

8.9.2 Using superposition, repeat Ex. 8.8.3.

8.9.3 Use superposition to find i for t > 0. Assume the circuit is in a steady-state
condition at / = 0" Am. 5 + 10(1 - r s6o ')mA

P
Vji-WU 1st

EXERCISE 8.9.3

%A> H*
i ^ 7*
•>

*5^ rVt
,

PROBLEMS
»> -H-xi? 1*5
ft
-2
.1 The current i(10 ) = 3.68 mA. Determine (a) t<0), (b) v(r) for f > 0, and (c)

the energy dissipated by the 10-kQ resistor as t becomes large, (d) Sketch v(t)

forr>0.

ft!
t = o

v(t) io kn
i

PROBLEM 8.1

2
8.2 Determine the new value of resistance necessary in Prob. 8.1 for r(10 )

3.68 V, given v(0) = 50 V.

8.3 Find v{t) for / > 0, given v(0) = 15 V.

~
\
\A
Chap. 8 Simple RC and RL Circuits 1 97

PROBLEM 8.3

8.4 Find i(t) for t > in Prob. 8.3.

8.5 The circuit is in a steady-state condition at / = 0". Find i(t) for / > 0.

PROBLEM 8.5

8.6 Consider a source-free circuit that has a response v(t) = V e~' /r . Show that a
straight line that is tangent to a graph of this response at time t {
intersects the
time axis (abscissa) at time /] + T.

8.7 The circuit is in a steady-state condition at / = 0". Find /'(/) for / > 0.

i(t)
1 2 .ft 2H
OlfP-

n
V

PROBLEM 8.7

'V
8.8 Find v(t) for / > 0, given /(0) = 2 A.

-
i(t)
3ft
Aa/v 1

v(t)
-o
4H3 v(t)

PROBLEM 8.8

CO:) --
f\

;
1 98 Simple RC and RL Circuits Chap. 8

8.9 Find /(/) for / > 0, given /(0) = 10 A.

i(t)

O.I H 2i(t)A

PROBLEM 8.9

8.10 Find i(t) for t > 0, given «(0) = 6 V.

xl0 i(t)V
3 y v(t)

PROBLEM 8.10

8.11 Find v{t) for t > 0, given v(0) = 2 V.

\AA

6 V -=- 0.25 F^: 40 5 v(t)

PROBLEM 8.11

8.12 Repeat Prob. 8.11 with the 0.25-F capacitor replaced by a ^-H inductor. Find the
inductor current.

8.13 Repeat Prob. 8.11 with the 6-V source replaced by 6e~' V.

8.14 Find i(t) for t > 0, given i(0) = 0.

lOi.V

4A

PROBLEM 8.14
 ' *
3 1H /YO -
&?> ' V, f "L-rt-

~"^A— —
Chap. 8 i Simplte RC and RL Circuits 199

8.15 Repeat Prob. 8.14 with the 4-A source replaced by 6e~' A.

8.16 The circuit is in a steady-state condition at t = 0~. Find i for / > 0.

V.^ I mH 5 n 5 n

(Jjp
(l) 20v

i/ - £•.
'^^
PROBLEM 8.16 ^< +
8.17 Express r(/) in terms of unit step functions, where ^
= ©
'
P -
-
£ ft) ~ov, /<-io iic'-fZ^-rf'
10 V, 10 < <0 /

2-
= f
W K =
20V,
15V,
< < 10
10 <'
t

6U.
8.18 Express i(t) in terms of unit step functions.

^PP PROBLEM 8.18

8.19 Sketch the voltage given by

«(0 = S
n=0
(' - «)[«(' - « - 1) -u{t-n - 2)] ^
8.20 Find i for t > if v = 24«(r) V.

i2n

PROBLEM 8.20
 W
200 S/>77p/e RC and RL Circuits Chap. 8

8.21 Repeat Prob. 8.20 for v = 2A[u{t) - w(r - 1)] V.

8.22 Find v for f > 0.

^(c0 V

Ca ' '

_u(t)(p

r' ^ t/<LC^
PROBLEM 8.22

i/
c 8.23 Using superposition, find v for t > 0, if v(0) = 0.

V .* + * '

-v^

v c<a PROBLEM 8.23

U' 8.24 Find u for ? > 0. Assume the circuit is in a steady-state condition at t = 0"

>/(.oY
vu t =

2kH IKfl
\&
10 V IOmA
f^
1*

y% PROBLEM 8.24

* *
U>
 9
SECOND-ORDER CIRCUITS

In the case of linear circuits with energy-storage elements the describing equations
(those relating the outputs to the inputs) may be expressed as linear differential
equations, because the terminal relations of the elements are such that the terms in
the loop or nodal equations are derivatives, integrals, or multiples of the unknowns
and the source variables. Evidently a single differentiation of an equation will remove
any integrals that it may contain, so that in general the loop or nodal equations for a
given circuit may be considered to be differential equations. The describing equation
then may be obtained from these equations.
The circuits containing storage elements that we have considered so far were
first-order circuits. That is, they were described by first-order differential equations.
This is always the case when there is only one storage element present or when a
switching action converts the circuit into two or more independent circuits each having
no more than one storage element.
In this chapter we shall consider second-order circuits, which as we shall see,
contain two storage elements and have describing equations that are second-order
differential equations. In general, //th-order circuits, containing n storage elements,
are described by mh-order differential equations. The results for first- and second-
order circuits (n =
1 and n — 2) may be readily extended to the general case, but we

shall not do so here. However, a solution of a third-order differential equation is

outlined in Prob. 9.25, which may be used to solve a third-order circuit given in
Prob. 9.26. Higher-order circuits will be treated in more detail in Chapter 14.

9.1 CIRCUITS WITH TWO STORAGE ELEMENTS

To introduce the subject of second-order circuits, let us begin with the circuit of Fig.
9.1, where the output to be found is the mesh current z*
2 . The circuit contains two

201
202 Second-Order Circuits Chap. 9

storage elements, the inductors, and as we shall see, i 2 satisfies a second-order differ-
ential equation. Methods of solving such equations will be considered in later sections
of this chapter.
The mesh equations of Fig. 9.1 are given by

2^+13,-4/, = *,
, (9.1)
-4/, + % + 4/ = 2

From the second of these we have

which differentiated results in

4% + 4^
dt
" 4 V dt
2 (9.3)

Substituting (9.2) and (9.3) into the first of (9.1) to eliminate /,, we have, after multi-
plying the resulting equation through by 2,

^+10§+16/ = 2*, 2 (9.4)

The describing equation for the output i2 is thus a second-order differential equation.
That is, it is a differential equation in which the highest derivative is second-order.
For this reason we shall refer to Fig. 9.1 as a second-order circuit and note that,
typically, second-order circuits contain two storage elements.
There are exceptions, however, to the rule that two-storage-element circuits have
second-order describing equations. For example, let us consider the circuit of Fig.
9.2, which has two capacitors. With the reference node taken as indicated, nodal
equations at the nodes labeled v and v 2 are given by
x

£ + .,«.. (9.5)
$h. +
_i_ 2v 2 -2v *
dt

The choice of thenode voltages v and v 2 as the unknowns has resulted in two
x

first-order differential equations,each containing only one of the unknowns. When

this happens, we say that the equations are uncoupled, and thus no elimination proce-
dure is required to separate the variables. It was the elimination procedure which,
applied to (9.1), gave the second-order equation of (9.4). The equations of (9.5) may
be solved separately by the methods of the previous chapter.
Chap. 9 Second-Order Circuits 203

FIG U R E 9.1 Circuit with two inductors FIGURE 9.2 Circuit with two capacitors

Evidently Fig. 9.2, although it contains two storage elements, is not a second-

order The same voltage v g is across each RC combination, and thus the circuit
circuit.

may be redrawn as two first-order circuits. If the source were a practical source rather
than an ideal source, then the circuit would be a second-order circuit. (See Prob. 9.5.)

EXERCISES
9.1.1 Find the equation satisfied by the mesh current i2 .

dh\ dvg
Ans.
dt 2
+ 7f + 6;2
dt

EXERCISE 9.1.1

9.1.2 Let Vg = Ue- 2 'V, /,(0

+
) = 6A, and / 2 (0
+
) = 2 A in Ex. 9.1.1, and find
+
di 2 (0 )ldt (the value of di 2 /dt at / = +
). Ans. -4 A/s
+ +
9.1.3 For the values of i 2 (0 ), di 2 (0 )/dt, and v g given in Ex. 9.1.2, show that i2 in
Ex. 9.1.1 is given by

i2 = -Ae~' + le~ 2 ' - e~ 6 < A

(Suggestion: Substitute the answer into the differential equation, etc.)

9.2 SECOND-ORDER EQUATIONS

In Chapter 8 we considered first-order circuits in some detail and saw that their describ-
ing equations were first-order differential equations of the form
204 Second-Order Circuits Chap. 9

4* + a x=f(t) (9.6/

In Sec. 9.1 we denned second-order circuits as those having two storage elements with
describing equations that were secon d-order differential equa tions, given generally by

yA In -fU^ -ferr^
f^g + ^f + Oox=f(t)
} (9.7)

In (9.6) and (9.7) the a's are real constants, x may be either a voltage or a current, and
f{t) is a known function of the independent sources.
As an example, for the circuit of Fig. 9.1, the describing equation was (9.4).
Comparing this equation with (9.7), we see that a = 10, a Q = 16, fit) = 2vg and x ,

x = i2 .

From Chapter 8 we know that the complete response satisfying (9.6) is given by

x = x„ + xf (9.8)

where x n is the natural response obtained when/(/) = and x f is the forced response,
which satisfies (9.6). The forced response, in contrast to the natural response, contains
no arbitrary constants.
Let us see if this same procedure will apply to the second-order equation (9.7).
By a solution to (9.7) we shall mean a function x which satisfies (9.7) identically. That
is, when x is substituted into (9.7), the left member becomes identically /(/)• We shall

also require that x contain two arbitrary constants since we must be able to satisfy
the two conditions imposed by the initial energy stored in the two storage elements.
If x n is the natural response, i.e., the response when/(/) = 0, then it must satisfy
the equation
d 2 x„ dx
+ a ^t
,

, a xn = (9.9)
~dF

Since each term contains x„ to the same degree, namely 1 (the right member may be
thought of as = 0x n ), this equation sometimes called the homogeneous equation.
is

If x f is to satisfy the original equation, as it did in the first-order case, then by

(9.7) we must have
l

dt
*f.
2 + ai £?
'

dt
+a '
xf =fQ) (9.10)

Adding (9.9) and (9.10) and rearranging the terms, we may write

£(x + n Xf) + a, |(x„ + Xf) + a (x n + x f) = fit) (9.11)

The rearrangement is possible, of course, because the equations involved are linear.
Comparing (9.7) and (9.11) we see that (9.8) is our solution, as it was in the first- \

order case. That is, x satisfying (9.7) is made up of two components, a natural response
Chap. 9 Second-Order Circuits 205

.v„ satisfying the homogeneous equation (9.9) and a forced response x f satisfying the
original equation (9.10) or (9.7). As we shall see. the natural response will contain
two arbitrary constants and. as in the first-order case, the forced response will have
no arbitrary constants. We shall consider methods of finding the natural and forced
responses in the next three sections.
Of course, if the driving, or forcing, functions are such that/(/) = in (9.7), then
the forced response is zero, and the solution of the differential equation is simply the
natural response.

EXERCISES
9.2.1 Show that
x t
=A t
e '

and
xz = A 2 e~ 2 '

are each solutions of

& + 3$ + *-0
regardless of the values of the constants, A and A 2
x
.

9.2.2 Show that

x = x x + x2 = A x
e~< + A 2 e~ 2 '

is also a solution of the differential equation of Ex. 9.2.1.

9.2.3 Show that if the right member of the differential equation of Ex. 9.2. 1 is changed
from to 6, then
x = Aie~' + A 2 e~ 2 + '
3
is a solution.

THE NATURAL RESPONSE

The natural response x n of the general solution

x = x„ + xf

of (9.7) must satisfy the homogeneous equation, which we repeat as

^ + a^ + a x = (9.12)

Evidently the solution x = x„ must be a function which does not change its form when
206 Second-Order Circuits Chap. 9

it is differentiated. That is, the function, its first derivative, and its second derivative
must all have the same form, for otherwise the combination in the left member of the
equation could not become identically zero for all t.

We are therefore led to try

xn = Ae s '
(9.13)

since this is the only function which retains its form when it is repeatedly differentiated.
This is, of course, the same function that worked so well for us in the first-order case
of Chapter 8. Also, as in the first-order case, A and s are constants to be determined.
Substituting (9.13) for x in (9.12), we have

As 2 e s'
+ Asa e"
x
+ Aa e
s'
=
or

Ae s
'(s
2
+ a :
s + a ) =
Since Ae" cannot be zero [for then by (9.13) x„ = 0, and we cannot satisfy any initial

energy-storage conditions], we have

s
2
+ a,s + aQ = (9.14)

This equation is and is simply the result of replacing

called the characteristic equation
derivatives in (9.12) by powers of s. That is, x, the zeroth derivative, is replaced by
s°, the first derivative by s and the second derivative by s 2
1
,
.

Since (9.14) is a quadratic equation, we have not one solution, as in the first-order
case, but two solutions, say s and s 2 given by the quadratic formula as
x ,

— a ± Ja\ — x
4a
(9.15)
S 1 7.

have two natural components of the form (9.13), which we denote by

The coefficients A and A 2 are, of course, arbitrary. Either of the two solutions (9.16)
x

will satisfy the homogeneous equation, because substituting either into (9.12) reduces
it to (9.14).
As a matter of fact, because (9.12) is a linear equation, the sum of the solutions
(9.16) is also a solution. That is,

xn = xnl +x n2 (9.17)

is a solution of (9.12). To see this, we have only to substitute the expression for x„
into (9.12). This results in
Chap. 9 Second-Order Circuits 207

d2 d
x
jp( »i + x »i) + °i ^(x + »i *») + fl °(*»i + a »J

/d 2 x„, . dx„, . \ ,
(d 2 x. 2 , (/x n2 i \

-0+0=0
since both x nl and .v n2 satisfy (9.12).
By (9.16) and (9.17) we have
x„ = A<?
J
" + A ze s "
(9.18)

which is a more general solution (unless s, = s2) than either equation of (9.16). In
fact. (9.18) is called the general solution of the homogeneous equation if s t
and s 2 are
distinct (i.e., not equal) roots of the characteristic equation (9.14).
As an example, the homogeneous equation corresponding to (9.4) in the previous
section is given by

^+10^+16/
dt
+ *2 2
= (9.19)

and thus the characteristic equation is

j2 + 10s + 16 =
The roots are s — —2 and s = —8, so that the general solution is given by

i2 = A^- + 2'
A 2 e-*' (9.20)

The reader may verify by direct substitution that (9.20) satisfies (9.19), regardless of
the value of the arbitrary constants.
Because (9.18) is and s 2 are sometimes called
the natural response, the numbers s,

same role as the negative

the natural frequencies of the circuit. Evidently they play the
reciprocal of the time constants considered in Chapter 7. There are, of course, two
time constants in the second-order case as compared to one in the first-order case.
For example, the natural frequencies of the circuit of Fig. 9.1 are s = —2, —8, as
displayed in (9.20); the time constants of the two terms are then \ and \.
The unit of natural frequency, which is the inverse of that of the time constant, is

the reciprocal of seconds. That is, it is a dimensionless quantity divided by seconds.

Therefore st is dimensionless, as it must be in e".

EXERCISES
9.3.1 Given the linear differential equation

show that X[ = te~', x2 = 1, and x : + x 2 are all solutions.

208 Second- Order Circuits Chap. 9

9.3.2 Given the nonlinear differential equation

d 2x
dt 2
<§-°
show that xi = t
1 and x 2 = 1 are both solutions but that x x + x2 is not a
solution.

9.3.3 Given

<*>©+<§ 3x =
2
d x
(b)
dt
2§ + * =
2

find the characteristic equation and the natural frequencies in each case.
Am. (a) -1, -3, (b) -1, -1

9.4 TYPES OF NATURAL FREQUENCIES

Since the natural frequencies of a second-order circuit are the roots of a quadratic
characteristic equation, they may be real, imaginary, or complex numbers. The nature
of the roots is determined by the discriminant a 2 — 4a of (9.15), which may be posi-
tive (corresponding to real, distinct roots), negative (complex roots), or zero (real,
equal roots).
For example, consider the circuit of Fig. 9.3, where the response to be found is the
mesh current i2 The loop equations, written around the left mesh and the outer loop,
.

are given by

c*+4y,+£-»,-£=..
At, + 4 f i, dt + t<0) = v,
Jo

FIGURE 9.3 Second-order circuit

Substituting for /, from the second equation into the first, we have

1 (dv„ A . \ „. di

(H-V«-\j>*-«> +-H*-*-) 2
dt
Chap. 9 Second-Order Circuits 209

Differentiating to eliminate the integral and rearranging, we have

+ (/?+ l)% + CR + 4)/ 2

The natural component /'

2n
== /n satisfies the homogeneous equation

^ +^
dt
1 K "^ 1)^
(/?+
(/?+l)^f
>
l

dt
+ (/? + 4)/ n =

from which the characteristic equation is

s2 + (R + \)s +R+ 4 =
Using the quadratic formula, we have the natural frequencies, given by

s ^_ -(R+\)±^R
2
-2R- 15
(92])

If R= 6 Q in (9.21), the natural frequencies are real and distinct, given by

s U2 = -2, -5 (9.22)

If R= 5 Q, the natural frequencies are real and equal, given by

s U2 = -3, -3 (9.23)

Finally, if R= ID, the natural frequencies are complex numbers, given by

s 1>2 = -l±j2 (9.24)

where j = «J—\. (In electrical engineering we cannot use as the mathematicians do,
/',

for the imaginary number unit, since this would result in confusion with the current.
Complex numbers are considered in Appendix C for the reader who needs to review
the subject.)

Distinct Real Roots

If the natural frequencies 5, 2 are real and distinct, then the natural response is

given by (9.18). For example, in the case of (9.22), we have

in = A.e- 1 '
+ A 2 e~ 5 '

Complex Roots

If the natural frequencies are complex, then in general we have

210 Second-Order Circuits Chap. 9

where a and fi
are real numbers. For example, in the case of (9.24), a =— 1 and fi
=
2. By (9.18) the natural response in the general case is

xn = A.e^W' + A^-'V (9.25)

This appears to be a complex quantity and not a suitable answer for a real current
or voltage. However, because A and A 2
x
are complex numbers, it is mathematically
correct, although somewhat inconvenient.
To put the natural response (9.25) in a better form, let us consider Euler's formula,
given by
e" = cos0+./sin0 (9.26)

and its alternative form, obtained by replacing j by —j,

e-'« = cos 6 -jsinO (9.27)

These results are derived in Appendix D. They are named for the great Swiss mathe-
matician Leonhard Euler (pronounced "oiler"), who lived from 1707 to 1783. Euler's
greatness is by the fact that the number
attested to e, the base of the natural loga-
rithmic system, was chosen in his honor.
Using (9.26) and (9.27), we may write (9.25) as

x„ = <?«'(iV' + A e- ") z
Jl

= e« [^,(cos fit +j sin fit) + A (cos fit -j sin fit)]

r
2

= e"[(A + A cos fit + (jA, -jA sin fit)

1 2) 2)

Since A and A 2
x
are arbitrary, let us rename the constants as

A = B, x
+ A2
Mi —Mi = B 2

so that

x„ = a
e '(B l
cos fit +B 2
sin fit) (9.28)

As an example, in the case of (9.24) we have a = — 1 and fi

= 2, so that

in = e~'(B 1
cos It +B 2 sin It)

where B x
and B 2 are, of course, arbitrary.

Real Equal Roots

The last type of natural frequencies we may have are those that are real and equal,
say

j, =s = 2 k (9.29)
Chap. 9 Second-Order Circuits 211

In this case (9. 18) is not the general solution since both x ni and x„ 2 are of the form
Ae kt and
, thus there is only one independent arbitrary constant. For (9.29) to be the
natural frequencies, the characteristic equation must be

(s - k) 2 = s
2 - 2ks + k2 =
and therefore the homogeneous equation must be

^-2k^f+k x 2
n
= (9.30)

Since we know that Ae k '

is a solution for A arbitrary, let us try

xn = h{t)e
k<

Substituting this expression into (9.30) and simplifying, we have

#h *kt _
dt 2

Therefore /;(/) must be such that its second derivative is zero for all t. This is true
if /;(/) is a polynomial of degree 1, or

h(t) =A + i
A2t

where A and A 2
l
are arbitrary constants. The general solution in the repeated-root
case, s u2 = k, is thus

xu = {A +A % t)t» l (9.31)

which may be verified by direct substitution into the homogeneous equation (9.30).
As an example, in the case of (9.23) we have s l2 = —3, —3, and thus

iH = (A +A t 2 t)e-
3'

EXERCISES
9.4.1 Find the natural frequencies of a circuit described by

d 2x dx n
Wl +
.

a +.

a x =
x
Tt

if (a) a x
= 5, a = 6; (b) a x = 4, a = 13; and (c) a = 8, a = 16.
x

Arts, (a) -2, -3; (b) -2 ±J3; (c) -4, -4

212 Second-Order Circuits Chap. 9

9.4.2 Find x in Ex. 9.4.1 with the arbitrary constants determined so that x(0) = 1

and dx(0)/dt = 4.
Ans. (a) 7e~ 2 ' - 6e~ 3 ', (b) <r 2 '(cos 3t + 2 sin 30, (c) (1 + 8/)<r 4r

9.4.3 Find x if

d 2x
J+ 9X = °

Ans. x =A i
cos 3/ + A 2 sin 3?

THE FORCED RESPONSE^

The forced response x f of the general second-order circuit must satisfy (9.10) and
contain no arbitrary constants. There are a number of methods for finding x f but ,

for our purposes we shall use the procedure of guessing the solution, which has worked
so well for us in the past. We know from our experience with first-order circuits that
the forced response has the form of the driving function. A constant source results
in a constant forced response, etc. However, the response must satisfy (9.10) identi-
cally, which means that first and second derivatives of x
f as well as x f itself, will ,

appear in the left member of (9.10). Thus we are led to try as x f a combination of the
right member of (9.10) and its derivatives.
As an example, let us consider the case v g — 16 V in Fig. 9.1. Then by (9.4), for
i 2 = x, we have

^ T+ 10^+16*
dt
2 ^ =
dt
32 (9.32)'
v

The natural response was given earlier in (9.20) by

Xn = A x
e- 2 ' + A 2 e~" (9.33)

Since the right member of (9.32) is a constant and all its derivatives are constant
(namely zero), let us try

xf =A
where A is a constant to be determined. We note that A is not arbitrary but is a par-
ticular value that hopefully makes x f a solution of (9.32). Substituting xf into (9.32)
yields

16A = 32

or

xf =A= 2

Therefore the general solution of (9.32) is

*(/) = A,e- 2 '

+ A 2 e~*' + 2
Chap. 9 Second-Order Circuits 213

A knowledge of the initial energy stored in the inductors can now be used to evaluate
A and A 2
x
.

In the case of constant forcing functions we may often obtain x f from the circuit
itself. In the example just considered xf is the steady-state value of i 2 in Fig. 9.1 when
vg = 16 V.At steady state the inductors look like short circuits, as shown in Fig.
9.4, so that, from the figure, we have

FIGURE 9.4 Circuit of Fig. 9.1 in the

steady state

As another example, suppose in Fig. 9.1 we have

vg = 20 cos At V

Then by (9.4), again for i2 = x, we have

d2 x dx
.
1 + 10^+ 16x = 40 cos 4? (9.34)
dt dt

The natural response x„ is given by (9.33), as before. To find the forced response x f
we need to seek a solution which contains all the terms, and their possible derivatives,
in the right member of (9.34). The coefficients of these terms will then be determined
by requiring x f to satisfy the differential equation. In the case under consideration,
the only term is a cos At term and the trial,

A cos At + B sin At (9.35)

contains this term and all its possible derivatives (which are cos At and sin At).
From (9.35) we have

fy = -AA sin At + AB cos At

dt

^= dt
2
-16,4 cos At - 165 sin At

Substituting these values and (9.35) into (9.34) and collecting terms, we have

405 cos At — AOA sin At = 40 cos At

214 Second-Order Circuits Chap. 9

Since this must be an identity, the coefficients of like terms must be the same on both
sides of the equation. In the case of the cos At terms we have
405 - 40

and for the sin At terms we have

-40,4 =
Thus A = and B — 1, so that

xf = sin At (9.36)

The general solution of (9.34), from (9.33) and (9.36), is given by

i2 =x= A x
e- 2 '
+ A 2 e-*' + sin At (9.37)

This may be readily verified by direct substitution.

Some ofthe more common forcing functions f(t) which occur in (9.7) are listed
in the first column of Table 9.1. The general form of the corresponding forced response
is given in the second column, which may be useful for formulating the trial solution
xt .

TABLE 9.1 Trial forced responses

fit) xf

k A
t At +B
t* At 1 + Bt + C
C" Ae"'
sin bt, cos bt A sin bt + B cos bt
e"' sin bt, e"< cos bt e"'(A sin bt + B cos bt)

EXERCISES
9.5.1 Find the forced response if

£g* + 4 $ + 3,-/<0

where f{t) is given by (a) 6, (b) Ae' 1 ', and (c) 9t.

Ans. (a) 2, (b) -Ae~ 2 ', (c) 3t - A

9.5.2 If x(0) = 2 and dx(0)/dt = 4, find the complete solution in Ex. 9.5.1.
Ans. (a) 2e~' - 2e' 3 '
+ 2, (b) le~< - Ae' 1 - e~ (c) 9.5<r' - 3.5e" + 3/ - 4
' 3
',
3t
Chap. 9 Second-Order Circuits 215

9.6 EXCITATION AT A NATURAL FREQUENCY

Suppose the circuit equation to be solved is given by

dz x dx
w --(a + b)^ +
,
{a
, ,

b)
v

di
abx=f(t) (9.38)

where a and b =£ a are known constants. In this case the characteristic equation is

s2 - (a + b)s + ab =
from which the natural frequencies are

st = a, s2 = b

Therefore we have the natural response

x„ = A.e"' + A 2 eb '
(9.39)

where A :
and A 2 are arbitrary.
Let us suppose now that the excitation function contains a natural frequency, say

fit) = e
at
(9.40)

The usual procedure is to seek a forced response,

xf = Ae°' (9.41)

and determine A so that xf satisfies (9.38), which in this case is

^- (a + b) ^ + abx = e°' (9.42)

However, substituting x f into (9.42) yields

= e"

which is an impossible situation.

This difficulty could have been foreseen by observing that x f in (9.41) has the form
of one of the components of xn in (9.39), and thus x f will satisfy the homogeneous
equation corresponding to (9.38). That is, x f substituted into (9.38) makes its left
member identically zero. There is no point then in trying such a forced response as
(9.41).
Let us consider what happens if we multiply by t the part of xf that is duplicated
in x n That
. is, let us try
xf = Ate"' (9.43)
21 6 Second- Order Circuits

instead of (9.41). We then have

dx f
dt
-A(at + l)e°

d2 xf
2
= A(a 2 t + 2a)e a
dt

Substituting these values, along with (9.43), into (9.42), we have

Ae°'[a 2 t + 2a - (a + b)(at + 1) + abt] = e"'

which, upon simplification, becomes

A(a — b)e a ' = e

a'

Since this must be an identity for all /, we must have

a — b

The general solution of (9.38), using (9.39) and (9.43), is then

x = A.e"' + A 2 eb '
+ -^~
a — b

As an example, suppose the excitation in Fig. 9.1 is given by

vg = 6e~ 2 '
+ 32

Then if i 2 = x, we have, by (9.4),

<L£
^F +
, i
10
r\
^ + 16* = 12er
Ci-X

~di
2'
+ 64 (9.44)

The natural response, as before, is

xn =A t
e- 2 ' + A 2 e-»'

Noting that the right member of the differential equation has the term e~ 2 ' in com-
mon with x„, we try

xf = Ate~ 2 '
+B
The factor t has been inserted into the natural trial solution of xf to remove the dupli-
cation of the term e~ 2 '. Substituting xf into (9.44) and simplifying, we have

6Ae~ 2 '
+ 165= 12e" 2 ' + 64
Chap. 9 Second-Order Circuits 217

Therefore we have A = 2 and B= 4, so that

x, = 2te" 2 ' + 4

The general solution is now

;, = X = X„ + X,

As a final example, let us consider the case of (9.38) where b = a and/(f) is given
by (9.40). That is.

d2 x ~ dx
2a =£ + a 2x = e
a'
(9.45)
dt- dt

The characteristic equation is

s1 — 2as + a2 =
and thus the natural frequencies are

st = s2 = a

The natural response is then

x„ = (A, + A lt )e«

We know it is fruitless to try as the forced response x f given in (9.41) because it is

duplicated in the natural response. In this case, (9.43) will not work either because it,

too, is duplicated. The lowest power of / that is not duplicated is 2; thus we are led
to try
xf = At 2 e a '

Substituting this expression into (9.45) we have

2Ae a ' = e"'

so that A = \. The forced and complete responses follow as before.

EXERCISES
9.6.1 Find the forced response if

g+ 4 ^ + 3*=/(0
where /(/) is given by (a) 2e~ 3 ' — e~ 2 ' and (b) 4e~' + 2e~ 3 '.

Ans. (a) <r 2 ' - te~ 3

', (b) t(2e~< - e~ 3 ')
218 Second-Order Circuits Chap. 9

9.6.2 Find the forced response if

W2+A Tt + <*=/(')

where f(t) is given by (a) 6e~ 2 ' and (b) 6te~ 2t . [Suggestion: In (b), try
xf = At 3 e~ 2 '.] Am. (a) 3t 2 e~ 2 ', (b) Oe~ 2t
9.6.3 Find the complete response if

d 2x
-ifi + 4x = 8 sin It

and x(0) = dx(0)/dt = 0. [Suggestion: Try xy = /(/* cos 2/ + B sin 2/).]

-4/w. sin 2t — 2/ cos 2t

9.7 THE COMPLETE RESPONSE

In the previous sections we have noted that the complete response of a circuit is the
sum of a natural and a forced response and that the natural response, and thus the
complete response, contains arbitrary constants. These constants, as in the first-order
cases of Chapter 8, are determined so that the complete response satisfies specified
initial energy-storage conditions.
To illustrate the procedure, let us find x(7), for t > 0, which satisfies the system
of equations

^ + 2x +
dt
5 ['
Jo
xdt= 16<r 3 '

(9.46)

x(0) = 2

To begin, let us differentiate the first of these equations to eliminate the integral; this
results in

^ + 2^ + 5x=
dt
1
dt
-48<r 3f

The characteristic equation is

s
2
+ 2s +5=
with roots

Therefore the natural response is

x„ = e~\A cos x
2t +A 2 sin 2t)

Trying as the forced response

xf = Ae~ 3 '
Chap. 9 Second-Order Circuits 219

we see that

SAe~ 3 ' = -48<r 3 '

so that A = —6. The complete response is therefore

x(t) = e-'(A {
cos It +A 2
sin 2t) - 6c 3 '
(9.47)

To determine the arbitrary constants we need two initial conditions. One, x(0) = 2,
is given in (9.46). To obtain the other we may evaluate the first of (9.46) at t = 0,
resulting in
•o
d
-^ + 2x(0) + 5 f x dt = 16

Noting the value of a(0) and that the integral term is zero, we have

dx(0)
-12 (9.48)
dt

Applying the second of (9.46) to (9.47), we have

x(0) =A -6= ! 2

or A :
— 8. To apply (9.48) we may differentiate (9.47), obtaining

-T- = e''(—2Ai sin 2t + 2A 2 cos 2t)

-e~%A x
cos 2? +A 2 sin 2t) + 18e- 3 '

from which

^^ = 2A 2
-A + t
18 = 12 (9.49)

From this, knowing A u we find A2 = 1.

At this point let us digress for a moment to note a very easy way to get (9.49).
We may differentiate x(t) and immediately replace t by before we write down the
result. That is, in (9.47), the derivative of x at t = is e~' at t = (which is 1) times
the derivative of (A }
cos 2t +A 2
sin 2r) at f = (which is 2A 2 ) plus (^ cos 2/ +
^2 sin 2r)at t = (which is ^,) times the derivative at / (which is l)plus off = —
the derivative of — 6e~ 3 '
at t = (which is 18). These steps are written down in (9.49)
and can be done mentally, avoiding the intermediate prior step.
Returning to our problem, we now have the arbitrary constants, so that by (9.47),
the final answer is

x = e~'(S cos 2t + sin 2f) - 6e~ 3 '

As a last example, let us find v, t > 0, in the circuit of Fig. 9.5 if v^O) = v(0) =
and vg = 5 cos 2000? V. The nodal equation at node v {
is
220 Second- Order Circuits Chap. 9

o v

FIGURE 9.5 Example

2 X lO" 3 ^ - 10-X* - Iz X 10" 3 v + 10" 6 $i =

<//

or

4u, 1
- v + 2 x 10- 3 $i = 2v gK = 10 cos 2000/ (9.50)
dt

and the nodal equation at the inverting input of the op amp is

10" 6 dv
\ x 10- 3 v, +\ X ^=
dt
or

v - _i x 10" 3 ^ (9.51)

Substituting (9.51) into (9.50) and simplifying, we have

^ + 2x
dt
2
10 3 ^
dt
+2x 10 6 v = -2 X 10 7 cos 2000/

The characteristic equation is

i
2
+ 2x 10 3 s + 2 x 10 6 =
so that the natural frequencies are j 1>2 = 1000( — 1 ±yl). The natural response is

therefore
v. = e~ '{A x
cos 1000/ +A 2 sin 1000/)

For the forced response we shall try

vf =A cos 2000/ + B sin 2000/

which substituted into the differential equation yields

(-2/4 + AB) cos 2000/ + (-4 A - 2B) sin 2000/ = -20 cos 2000/
Therefore equating coefficients of like terms, we have
Chap. 9 Second-Order Circuits 221

-2/4 + 45 = -20
-4/4 - 25 =

from which A = 2 and B = —4. The complete response is then

v = e-
1000
'^, cos 1000/ j A2 sin 1000/) + 2 cos 2000/ - 4 sin 2000/ (9.52)

From (9.51), for / = +

. we see that

= -{x
r,(o
+
) io- 3
^p
and since r,(0 + ) = ^
,

1
(0~) = we have

**P = (9.53)

Like v u v is also a capacitor voltage (across the ^-/*F capacitor), so that

r(0 + ) = v(0-) = (9.54)

From (9.52) and (9.54) we have

A 1
+2=
or A t
= —2, and from (9.52) and (9.53) we have

100(X4 2 - 1000/4, - 8000 =

from which A2 = 6. Thus the complete response is

t> = e- I000 '(-2 cos 1000/ + 6 sin 1000/) + 2 cos 2000/ - 4 sin 2000/ V

EXERCISES
9.7.1 Find x, t > 0, where

Jo
/o

x(0) = -1

and (a)/(/) = 1 and (b)/(/) = -It.

Ans. (a) (-1 + 2t)e~', (b) (1 + 30c"' - 2
222 Second-Order Circuits Chap. 9

9.7.2 Find /, v, di/dt, and dv/dt at/ = +

. Ans. 0, 0, 2 A/s, 40 V/s

u(t,(f)

EXERCISE 9.7.2

9.7.3 For t > in Ex. 9.7.2, find (a) v and (b) /'.
(Suggestion: Because of Kirchhoff's
laws and the terminal relationships of the elements, i has the same natural
frequencies as v. Thus v„ is easily obtained after i„ is found; its forced response
is evident by inspection of the circuit.)
Ans. (a) e _2 '(-6 cos At + 7 sin At) + 6, (b) e
-2 '(-2
cos 4/ — { sin 4r) + 2

9.8 THE PARALLEL RLC CIRCUIT

One of the most important second-order circuits is the parallel RLC circuit, the source-
free version of which is shown in Fig. 9.6. We shall assume that at t = there is an
initial inductor current,
/(0) =h (9.55)

and an initial capacitor voltage,

v(0) = V (9.56)

and analyze the circuit by finding v for / > 0.

FIGURE 9.6 Source-free parallel RLC

circuit

The single nodal equation that is necessary is given by

.
+ + /l + c *_o (9.57)
^.J.*
Chap. 9 Second-Order Circuits 223

which is an integrodifferential equation that becomes, upon differentiation,

Since there is no forcing function, the natural response is the complete response.
The characteristic equation is

from which the natural frequencies are

*'' 2 " ~2RC ± ^\2RCJ ~LC (9,59)

As in the general second-order case already discussed, there are three types of
responses, depending on the nature of the discriminant, (\f2RC) 1 — (l/LC), in (9.59).
We shall now look briefly at these three cases.

Overdamped Case

If the discriminant is positive, i.e.,

(m;) -ic >0

or, equivalently,

L > 4R Z C (9.60)

then the natural frequencies of (9.59) are real and distinct negative numbers. In this
case, known as the overdamped ease, the response has the form

v = A.e 1 " + A 2e s *
(9.61)

From the initial conditions and (9.57) evaluated at / = +

, we obtain

dv(0 + ) _ V + RI
(9.62)
dt RC
which together with (9.56) can be used to determine the arbitrary constants.
As an example, suppose R = 1 Q, L — $ H, C — \ F, VQ = 2 V, and I = —3 A.
Then by (9.59) we have s, 2 = 1, —3, and hence —
v =A x
e~' + A 2 e~ 3'

Also, by (9.56) and (9.62) we have

224 Second- Order Circuits Chap. 9

v(0) = 2V
*£2 = 4v/s '
dt

which may be used to obtain A x

— 5 and A 2 = —3, and thus

v = 5e~' - 3e' 3!

This overdamped case is easily sketched, as shown by the solid line of Fig. 9.7, by
sketching the two components and adding them graphically.

FIGURE 9.7 Sketch of an overdamped response

The reason for the term overdamped may be seen from the absence of oscillations
(fluctuations in sign). The element values are such as to "damp out" any oscillatory
tendencies. Il is, of course, possible for the response to change signs once, depending
on the initial conditions.

Under damped Case

If the discriminant in (9.59) is negative, i.e.,

L < 4R C 2
(9.63)

then the circuit is said to be underdamped. In this case the natural frequencies are
complex, and the response contains sines and cosines, which of course are oscillatory-
type functions. In this case it is convenient to define a resonant frequency,

COn (9.64)
JLC
a damping coefficient,
1
(9.65)
2RC
 Second-Order Circuits 225

and a damped frequency.

COd = ^0)1 (9.66)

Each of these is a dimensionless quantity "per second." The resonant and damped
frequencies are defined to be radians per second (rad/s) and the damping coefficient
is nepers per second (Np/s).
Using these definitions, the natural frequencies, by (9.59), are

s lw2 = -a±ja>d
and therefore the response is

v = e~"(A 1
cos co d t + A2 sin co d t) (9.67)

which is oscillatory in nature, as expected.

As an example, suppose R= 5 Q, L = 1 H, C= -Jq F, V = 0, and I = —\ A.
Then we have
v = e~'(A l
cos 3t + Az sin 3?)

From the initial conditions we have r(0) = and dv(0 + )/dt = 15 V/s, from which
A = x
and A 2 = 5. Therefore the underdamped response is

v = 5e~' sin 3/

This response is readily sketched if it is observed that since sin 3/ varies between
+1, and —l,v must be a sinusoid that varies between 5e~' and — 5<r'. The response
is shown in Fig. 9.8, where it may be seen that it is oscillatory in nature. The response
goes through zero at the points where the sinusoid is zero, which is determined, in
general, by the damped frequency co d .

FIGURE 9.8 Sketch of an underdamped response

226 Second-Order Circuits Chap. 9

Critically Damped Case

The dividing line between the overdamped and the underdamped cases is the
damped
critically case, occurring when the discriminant of (9.59) is zero. That is,

L = 4R 2 C (9.68)

In this case the natural frequencies are real and equal, given by

5 lf2
= — cc, — oc

where a is given by (9.65). The response is then

v = (A l
+ A z t)e-« (9.69)

As an example, consider the case R — 1 Q, L — 1 H, C= \ F, V — 0, and I =

— 1 A. In this case we have a = 2, A = 0, and A 2 l
= 4. Thus the response is

v = Ate' 1 '

This is easily sketched by plotting At and e~ 2 and multiplying the two together. The
'

result is shown in Fig. 9.9.

For every case in the parallel RLC circuit, the steady-state value of the natural
response is zero, because each term in the response contains a factor e"', where a < 0.

I 2

FIGURE 9.9 Sketch of a critically damped response

EXERCISES
9.8.1 In a source-free parallel RLC circuit, R = 1 kfi and C = 1 /^F. Find L so that
the circuit is overdamped with s 1>2 = —250, —750 s" (b) underdamped
(a) 1
,

with co d = 250 rad/s, and (c) critically damped.

Am. (a) H, (b) -^ H, (c) 4 H ^
Chap. 9 Second-Order Circuits 227

9.8.2 (a) Find the differential equation satisfied by i in Fig. 9.6. (b) Use this result to
find i, for t > 0, if R= 5 CI, L= 1 H, C- 0.1 F, v(0) = 0, and ;(0) = 3 A.
d2i 1 di
Ans. (a) ^3 -f -g£ -jj +
1
jj^i = 0, (b) <r'(3 cos 3f + sin 30

9.8.3 The larger the value of R, the less damping there is in the underdamped case of
the parallel RLC circuit (because a = 1/2RC). Let R = oo (open circuit) and
show that
d 2v
dt 2
+ <olv =

For this case, find the general solution for v.

Ans. A x
cos CO t +A 2 sin co t

9.9 THE SERIES RLC CIRCUIT

The source-free series RLC circuit, shown in Fig. 9.10. is the dual of the parallel cir-
cuit, considered in the previous section. Therefore all the results for the parallel circuit
have dual counterparts for the series circuit, which may be written down by inspection.
In this section we shall simply list these results using duality and leave to the reader
their verification by conventional means.

FIGURE 9.10 Series RLC cir-

cuit

Referring to Fig. 9.10, the initial conditions will be taken as

v(0) = V
/(0) = /„

The single-loop equation necessary in the analysis is

L §+ m+:vi"h+v'- (9.70)

which is valid for t > 0. The resulting characteristic equation is

Ls 2 + Rs + -1 = (9.71)
228 Second-Order Circuits Chap. 9

and the natural frequencies are

2
(9-72)
*-=-ffi±V(I) -ie
The series RLC circuit is overdamped if

AL
C> R2
(9.73)

and the response is

i = A x
e "
s
+A 2e
s " (9.74)

The circuit is critically damped if

C = AL2 (9.75)
R

in which case sx =s = 2
—R/2L, and the response is

i={A x
+A % t)e>* (9.76)

Finally, the circuit is underdamped if

C<f 2
(9.77)

in which case the resonant frequency is

w = « -4=pr (9.78)
JLC
the damping coefficient is

a 4
=~2L ( 9 - 79 )

and the damped frequency is

co d = jToi—^ (9.80)

The underdamped response is

i = e~ M{A x cosgv + ^2 sinoojO (9.81)

Since the series and parallel RLC circuits are duals, we shall not consider any
examples in this section. In the following section we shall illustrate the series circuit

with a forcing function.

Chap. 9 Second-Order Circuits 229

EXERCISES
9.9.1 Let R = 4Q, L - 1 H, r(0) = 4 V, and ;(0) = 2 A in Fig. 9.10. Find i for
/ > if C is (a) J^ f, (b) | F, and (c) ->
F.

Am. (a) 2e" 2 '(cos At - sin 4?) A, (b) 2(1 - 40c 2 '
A, (c) -3e"' + 5c A 3'

9.9.2 Find v for / > if R = 40 fi, L = 10 mH, and C = 5 jj.¥.

Am. e - 2000 '(4 cos 4000/ - 3 sin 4000/) V

EXERCISE 9.9.2

9.10 THE COMPLETE RESPONSE

OF THE RLC CIRCUIT

In this section we shall consider the series RLC circuit excited by a source, an example
of which is shown in Fig. 9.1 1. The parallel RLC circuit is the dual circuit and we
shall leave examples of it to the problems.

2A

FIGURE 9.11 Driven series RLC cir-

cuit

Suppose it is required to find v, for t > 0, in Fig. 9.11, given that

i-(0) = 6 V, /(0) =2A

We know that
v = v„ + vf

where the natural response v„ contains the natural frequencies. The natural frequencies
of the current i are the same as those of v because obtaining one from the other, in
230 Second-Order Circuits Chap. 9

general, requires only KirchhofTs laws and the operations of addition, subtraction,
multiplication by constants, integration, and differentiation. None of these operations
changes the natural frequencies. Therefore, since the natural frequencies of i are easier
to get (only one loop equation is required), let us obtain them. Around the loop we

have

j- + 2i + 5 ( idt + 6= 10 (9.82)

The characteristic equation, following differentiation, is

s
2
+ 2s + 5 =
with roots
*i,2 = —1 ±j2
Thus we have
v„ = e~'(A l
cos It + A2 sin It)

[We could have obtained s ul from (9.72) directly, but it is probably easier to write
the characteristic equation, since it can be done by inspection.]
The forced response vf is a constant in this case and may be obtained by inspection
of the circuit in the steady state. Since in the steady state the capacitor is an open
circuit and the inductor is a short circuit, if = and vf = 10. Therefore the complete
response is

v = e~'(A 1
cos It +A 2
sin It) + 10

From the initial voltage, we have

v(0) = 6 =A + l
10

or A x
— —4. Also, we have

l*£2 = /(0) = 2
5 dt
or

2 1
dt

Therefore A2 = 3, and we have

v = <r'(— 4 cos It + 3 sin It) + 10

As a final example, let us find v in Fig. 9.11 if the source is

vg = 4 cos t V

In this case we shall need the differential equation for v, which we may get from (9.82)
and
Chap. 9 Second-Order Circuits 231

'-7$ (983>

We may also obtain it directly from the figure since the voltages across the inductor
and resistor are

&_ \_dH
1
dt 5 dt
and
2 dv
-,.
2i = -C--J-
5 dt

and that across the capacitor is v, of course. In either case, the result is

d2 v ~ dv c = 1rt
+
. .

diT+ dt
5v 20cost

The natural response is the same as in the previous example. To get the forced
response we shall try
vf = A cos t +B sin t

which substituted into the differential equation results in

(4A + 2B) cos t + (4B - 1A) sin t = 20 cos *

Equating like coefficients and solving for A and B, we have

A = 4, B= 2

Therefore the general solution is

v = e~'(A 1
cos 2t + A2 sin It) + 4 cos t + 2 sin t

From the initial voltage, we have

1,(0)= 6 = A x
+4
or A x
= 2. From the initial current and (9.83), we have

*W11= m = 2A 2z - A, '
+ 2
dt

or A2 = 5.

The complete response is therefore

v = e~'(2 cos 2t -j- 5 sin 2t) + 4 cos t -j- 2 sin r

In the case of a driven, parallel RLC circuit, the complete response is obtained in
the same manner as that of the series case. Examples are given in Ex. 9.10.1 and in
the problems at the end of the chapter.
232 Second- Order Circuits Chap. 9

EXERCISES
9.10.1 Find / for t > if (a) L = f H and (b) L = 2 H.
Ans. (a) -3e-' + e' + 3'
2, (b) (-2 - 4t)e~ 2 ' + 2

EXERCISE 9.10.1

9.10.2 Find v for / > if (a) C= | F and (b) C= ± F.

-' -
Ans. (a) -25e + e _s ' + 24, (b) e" 3 '(-24 cos / 32 sin /) + 24

EXERCISE 9.10.2

9.11 ALTERNATIVE METHODS FOR OBTAINING

THE DESCRIBING EQUATIONS*

We shall close this chapter by considering two methods of expediting the process of
obtaining the describing equation of the circuit. In the case of the parallel and series
RLC circuits, a single equation is required, which, after differentiation, is the describ-
ing equation. However, in many second-order circuits there are two simultaneous
circuit equations from which the describing equation is obtained after a tedious
elimination process.
As an example, let us consider the circuit of Fig. 9.12 for / > 0. Taking node b
as reference and writing nodal equations at nodes a and v we have x

v- v*
|
v — Vi i

"*"
_L* = o
"t"
4 '
" 6 "
4 dt
(9.84)
v
-^=^ + v, dt + /(0) -
J"
Chap. 9 Second-Order Circuits 233

FIGURE 9.12 Circuit with two storage elements

If we are interested in finding v we must eliminate r, and obtain the describing equation
in terms of v. The result, as the reader may verify, is

dt
1
dt
+ dt ^ °
(9.85)

In this case the process is not overly complicated but it can be shortened by the

methods we shall consider in this chapter.

The first method we shall discuss is a systematic way of obtaining a describing
equation, such as (9.85). from the circuit equations, such as (9.84). To develop the
method, let us first introduce the differentiation operator D, which is defined by

dt

That is, D.x = dxjdt, D{Dx) =D 2

x = d 2 x/dt 2 , etc. Also, we have, for example,

a -r- + bx = aDx + bx — (aD + b)x

dt

It is important here to note that x is factored out of the middle member and placed
after the operator expression, indicating that the operation is to be performed on x.

Otherwise the meaning is changed radically.

With these ideas in mind, let us rewrite (9.84) in operator form. This result, after
first differentiating the second equation, is

/ 1
n ,
5\ 1 1

-4i»+0+i)*,=o
which when cleared of fractions becomes

+ 5)v — 2v = 3v g
(3D l

(9.86)
-Dv + (D + 6)v, =0
To eliminate v u let us "operate" on the first equation with D+ 6, by which we mean
234 Second-Order Circuits Chap. 9

multiply it through by D+ 6 on the left of each term. Then let us multiply the second
equation by 2, resulting in

(D + 6)(3Z> + - 2(D + 6)w, =

5)v 3(Z> + 6)v g

-2Dv + 2(Z) + 6)i>! =

[We note that constants such as 2 commute with operators, i.e., (D + 6)2 = 2(Z> + 6),
but variables do not.] Adding these last two equations eliminates v and x
results in

[{D + 6)(3Z> + 5) - 2D]v = 3(D + 6)v. (9.87)

Multiplying the operators as if they were polynomials, collecting terms, and dividing
out the common factor 3, we have

(D 2 + 7D+ 10> = (D + 6)v g

which is the same as (9.85).

The procedure may be carried out in a more direct manner by using determinants.
For example, we may use Cramer's rule to obtain the expression for v from (9.86),
given by

(9.88)

where A is the coefficient determinant

3D + 5 -2
A
1\ — (9.89)
-D D+6
and Aj is given by
3v g -2
A,= = 3(ZM (9.90)
D+6
We note that in this last expression we must be careful to write v g after the operator.
Writing (9.88) as
Av = Aj

we see from (9.89) and (9.90) that we have the describing equation (9.87).
The second method we shall consider is a mixture of the loop and'nodal methods
in which we select the inductor currents and the capacitor voltages as the unknowns,
rather than the loop currents or the node voltages. We then write KVL around loops
which contain only a and KCL at nodes, or generalized nodes, to which
single inductor
only a single capacitor is connected. In this manner each equation contains only one
derivative, that of an inductor current or a capacitor voltage, and no integrals. The
equations are then relatively easy to manipulate to find the describing equation.
As an example using Fig. 9. 12 again, let i, the inductor current, and v, the capacitor
voltage, be the unknowns. (These unknowns are sometimes called the state variables
of the circuit.) Then the nodal equation at node a is
Chap. 9 Second-order Circuits 235

^i + ' + IJ-o (9.91)

and the loop equation around the right mesh is

v = 6/ +
di_
(9.92)
dt

It is a relatively simple matter to solve for i in (9.91), substitute its value into (9.92),
and simplify the result to obtain (9.85).
Some advantages of this method are that no integrals appear (thus no second
derivatives occur as a result of differentiation), one unknown is easily found in terms
of the others, and the initial conditions on the first derivatives are easily obtained for
use in determining the arbitrary constants in the general solution. For example, it

may be seen from Fig. 9.12 that z'(O) = A and

1 v(0) = 6 V. Thus from (9.91) and the
+
value of t'
g (0 ), we have

dv(0 + )
v g (0 + ) - 10
dt

This last method can be greatly facilitated, particularly in the case of complex

circuits,by using graph theory. Since we are looking for inductor currents and capac-
itor voltages (the state variables), we put the inductors in the links, whose currents
constitute an independent set, and the capacitors in the tree, whose branch voltages
constitute an independent set, as we recall from Chapter 6. (The tree should also
contain voltage sources, and the links should contain current sources, etc., if possible.)
Each inductor L is then a link with current which forms a loop whose only other/',

elements are tree branches. Therefore KVL around this loop will contain only one
derivative term, L(di/dt), and no integrals. This loop can easily be found since it is
the only loop in the graph if the only link added to the tree is L. For example, the
graph of Fig. 9.12 is shown in Fig. 9.13 with tree branches shown as solid lines and
links as dashed lines. The loop containing the 1-H inductor is a,v u b,a through the
branch labeled v, and KVL around it is (9.92).
Each capacitor is a tree branch whose current, together with link currents, consti-

FIGURE 9.13 Graph of Fig. 9.12

236 Second-Order Circuits Chap. 9

tutes a set of currents flowing out of a node or a generalized node, because if the capac-
from the circuit, the tree is separated into two parts connected only by links.
itor is cut
In the example of Fig. 9.13 the three currents, shown crossing the dotted line through
the capacitor, labeled v, and two links, add to zero by KCL. This is stated by (9.91).

EXERCISES
9.11.1 Find the describing equation for v, t > 0, in Ex. 9.10.2 by (a) the first method
of this section using nodal equations and (b) the second method of this section.
Ans. CidHldf1 ) + 6C(dvldt) + v = 24

9.11.2 Solve Ex. 9.1.1 using (a) the first method of this section applied to the mesh
equations and (b) the second method of this section.

PROBLEMS
9.1 Find z'i and i 2 , for t > 0, in Ex. 9.1.1 if vg = (a short circuit), j'i(0)
= 2 A,
and / 2 (0) = 3 A.

9.2 Find > 0, in Ex. 9.1.1 if ^ = 12 V, /,(0) = -1 A, and

i2 , for / i 2 (0) = 2 A.

9.3 Find for > 0, if w(0) = 4 V, i(0) = 2 A, and (a) L = H

i, / 1 and R = 1 Q,
(b) L = H and R = 3 Q, and (c) L = 2 H and R = 5 Q.
1

PROBLEM 9.3

9.4 The circuit is in steady state when the switch is opened at / = 0. Find v and i

for t > 0.

9.5 Insert a 1-fi resistor in series with v g in Fig. 9.2, thereby making the source a
practical rather than an ideal one. Show that in this case v 2 satisfies the second-
order equation

5^ + ii^l +4 =4^ +
dt< dt
,2
dt
4f ;
Chap. 9 Second-Order Circuits 237

t = o

16V

PROBLEM 9.4

9.6 Find i2 , for t > 0, if /,(0) = 3 A, ; 2 (0)

= -1 A, and v g = 0.

PROBLEM 9.6

9.7 Find v, for r > 0, if the circuit is in steady state when the switch is moved from
position 1 to position 2 at / = and L is (a) 2 H, (b) 4 H, and (c) J^2 H.

— 8V

PROBLEM 9.7

9.8 Find ;', for t > 0, if the circuit is in steady state when the switch is opened at
t = 0.

PROBLEM 9.8
238 Second-Order Circuits Chap. 9

9.9 Find i2 in Prob. 9.6 if (a) vg = 15 V, (b) v g = lOe" 2 V, and (c) v g =

'
5e~' V.

9.10 Find v, for t > 0, if (a) v g = 12k(0 V and (b) v g = \2tu{t) V.

PROBLEM 9.10

9.11 Find v, for t > 0, if the circuit is in steady state when the switch is moved from
position 1 to position 2 at t = 0.

I kn
—2 o \ o—
I

h
v "
t =

10 V O.I H 5 V
0.025 gF t

PROBLEM 9.11

9.12 Find v, > 0, if there

t is no initial stored energy and (a) i
g
= 2 mA and (b)
i
g
= 4 e -iooor mA _

PROBLEM 9.12

9.13 Find /, / > 0, if the circuit is in steady state when the switch is opened at t = 0.

t =o.

PROBLEM 9.13
 Second-Order Circuits 239

9.14 Find r, t > 0.

2u(t) A

€^
2A IH

l
g>
e-
6u(t) V
—
PROBLEM 9.14

9.15 Find v, t > 0, in the figure for Prob. 9.7 if L — 1 H and the switch is moved from
position 2 to position 1 at / = 0.

9.16 Find v, t > 0, if t<0) =4V and /(0) = 2 A.

_t
3e A

2H J_ 2H
M\ —TSV> *"
i

2A F;±: v

PROBLEM 9.16

9.17 Find the maximum value of the overdamped response of Prob. 9.7(c) and the
time at which it occurs.

9.18 Find /, t > 0, if there is no initial stored energy and (a) R = \ Q, // = 2; (b)
r = i.
Q, ix = 1 ; and (c) R = \ Q, fj.
= 2.

2Q

PROBLEM 9.18

9.19 Find /', / > 0, if there is no initial stored energy and (a) C= \ F, (b) C= \ F,
and (c) C= TV$ F.
240 Second-Order Circuits Chap. 9

v,/2 A

PROBLEM 9.19

9.20 Find v, t > 0, in the circuit of Fig. 9.12 if vg = 29 cos 2t V and the circuit is in
steady state when the switch is moved at t = 0.

9.21 Find v, t > if there is no initial stored energy and v g = 4 cos 2t V.

2d 3fl

PROBLEM 9.21

9.22 Find v, t > 0, if there is no initial stored energy and v g = 5 cos At V.

.I o v

PROBLEM 9.22

9.23 Find v, / > 0, if there is no initial stored energy and v g = 2 cos t V.

9.24 Find 0, if (a) ^(0) = 4 V, «(0) = 0; (b) y,(0) = 0, i>(0) = 2 V; and (c)
v, t >
i-,(0) 4 V, v(0) = 2 V. (Note that the response is an unforced sinusoidal
=
response. Such a circuit is called a harmonic oscillator.)
Chap. 9 Second-Order Circuits 241

in
JSA^r-

in
in

I F 1.
I '

3
n
in

PROBLEM 9.23

in
-\KAr-

m
in

1 I F

m
_+

PROBLEM 9.24

9.25 Higher-order differential equations may be solved in the same manner as second-
order equations. There are more natural frequencies and thus more terms in the
natural response. For example, if

d3x , d 2x dx .-
dF +
.

6
-d^
.,.
+ n ^t
+6x =
. ,
l2

show that the characteristic equation

s3 + 6s 1 + \\s + 6 =

has the natural frequencies

s= -1, -2, -3
242 Second-Order Circuits Chap. 9

as its roots. Thus show that the natural response is

xn =A x
e~' + A 2 e- 2 + '
A^<
Show also that the forced response is

Xf = 2

and that the general solution is

x = xn + xf

9.26 Using the results of Prob. 9.25, find i, t > 0, if there is no initial stored energy.

^o
PROBLEM 9.26
 10
SINUSOIDAL EXCITATION
AND PHASORS

In the two previous chapters we have analyzed circuits containing dynamic elements
and have seen that the complete response is the sum of a natural and a forced response.
The natural response for a given circuit is obtained from the dead circuit and therefore
is independent of the sources, or excitations. The forced response, on the other hand,
depends directly on the type of excitation applied to the circuit. In the case of a dc
source, the forced response is a dc steady-state response, an exponential input results
in an exponential forced response, and so forth.
One of the most important excitations is the sinusoidal forcing function. Sinusoids
abound in nature, as, for example, in the motion of a pendulum, in the bouncing of a
ball, and in the vibrations of strings and membranes. Also, as we have seen, the

natural response of an underdamped second-order circuit is a damped sinusoid and in

the absence of damping is a pure sinusoid.
In electrical engineering, sinusoidal functions are extremely important for a
number of reasons. The carrier signals generated for communication purposes are
sinusoids, and, of course, the sinusoid is the dominant signal in the electric power
industry, to name two very important examples. Indeed, as we shall see later in the
study of Fourier series, almost every useful signal in electrical engineering can be
resolved into sinusoidal components.
Because of their importance, circuits with a sinusoidal forcing function will be
considered in detail in this chapter. Since the natural response is independent of the
sources and can be found by the methods of the previous chapters, we shall concen-
trate on finding only the forced response. This response is important in itself since it is

the ac steady-state response that is left after the short time required for the transitory
natural response to die.
Since we are interested only in the ac steady-state response, we shall not limit
ourselves, as we did in Chapters 8 and 9, to first- and second-order circuits. As we shall

243
244 Sinusoidal Excitation and Phasors Chap. 10

see, higher-order RLC circuits may be handled, insofar as the ac steady-state response
is concerned, in the same way as resistive circuits.

10.1 PROPERTIES OF SINUSOIDS

Because of the importance of sinusoidal functions, we shall devote this section to a

review of some of their properties. Let us begin with the sine wave,

v(t) = Vm sin cot (10.1)

which is sketched in Fig. lO.l. The amplitude of the sinusoid is Vm which is the ,

maximum value that the function attains. The radian frequency, or angular frequency,
is co, measured in radians per second (rad/s).
The sinusoid is aperiodic function, defined generally by the property

v(t + 77= v(t) (10.2)

where T is the period. That is, the function goes through a complete cycle, or period,
which is then repeated, every T seconds. In the case of the sinusoid, the period is

2n
(10.3)
CO

as may be seen from (10. 1) and (10.2). Thus in 1 s the function goes through l/T
cycles, or periods. Its frequency f is then

co
/ =

T In
(10.4)
i
v(t)'

olt7
cycle

FIG U R E 1 0.1 Sinusoidal function

Chap. 10 Sinusoidal Excitation and Phasors 245

cycles per second, or hertz (abbreviated Hz). The latter term, named for the German
physicist Heinrich R. Hertz ( 1 857-1894), is now the standard unit for frequency. Some
older books use the former term, but it is being discontinued. The relation between
frequency and radian frequency is seen by (10.4) to be

to = 2nf (10.5)

A more general sinusoidal expression is given by

/•(/) = Fm sin(a>/ + 0) (10.6)

where is the phase angle, or simply the phase. To be consistent, since cot is in radians,
should be expressed in radians. However, in electrical engineering it is often con-
\enient to specify in degrees. For example, we may write

v = Vm sin (* +
*)
or

v = Vm sin (2/ + 45°)

interchangeably, even though the latter expression contains a mathematical incon-

sistency.
A is shown in Fig. 10.2 by the solid line, along with a sketch of
sketch of (10.6)
(10.1), shown dashed. The solid curve is simply the dashed curve displaced 0/a>
seconds, or radians to the left. Therefore points on the solid curve, such as its peaks,
occur rad, or co s. earlier than corresponding points on the dashed curve. Accord-
ingly, we shall say that Vm sin (cot + 0) leads Vm sin cot by rad (or degrees). In
general, the sinusoid

Vml sin (cot + a)

leads the sinusoid

Vm2 sin (cot + p)

by a — p. An equivalent expression is that v 2 lags v x by a — /?.

As an example, consider

v, = 4sin(2r + 30°)

and

v2 = 6sin(2r - 12°)

Then y, leads v2 (or v 2 lags v ) by 30 — (— 12) = 42°.

t

Thus far we have considered sine functions rather than cosine functions in defining
 A

246 Sinusoidal Excitation and Phasors Chap. 10

sinusoids. It does not matter which form we use since

cos {cot - Z) = sin cot (10.7)

or

in (cot
sin + y \ = cos cot (10.8)

The only difference between sines and cosines is thus the phase angle. For example,
we may write (10.6) as

v(t) = Vm cos (cot + - ^-\

To determine how much one sinusoid leads or lags another of the same frequency,
we must first express both as sine waves or as cosine waves with positive amplitudes.
For example, let

v x
=4 cos (It + 30°)

and

v2 = -2 sin (2t + 18°)

Then, since

—sin cot =s sin (cot + 180°)

we have
vz = 2sin(2r + 18° + 180°)
= 2 cos (It + 18° + 180° - 90°)

= 2cos(2r + 108°)

Comparing this last expression with v ,, we see that v x leads v 2 by 30° — 108° = —78°,
which is the same as saying that i\ lags v z by 78°.

The sum of a sine wave and a cosine wave of the same frequency is another
sinusoid of that frequency. To show this, consider

B
A cos cot + B sin cot = J z
+ B'
V^ + B
2 2 ^
cos cat ^
'

JA ,
,,1
+ B„,
,
2
sin cot

which by Fig. 10.3 may be written

A cos cot + B sin cot — */A z +B z

(cos cot cos 6 + sin cot sin 0)

By a formula from trigonometry, this is

A cos cot + B sin cot = JA 1

+B 2
cos (cot - 6) (10.9)
Chap. 10 Sinusoidal Excitation and Phasors 247

vlt ) 1
Vm sin (cott<J>)

//% //
Vm sin u) t

s&//
9>

A x

s\ 9

FIGURE 10.2 Two sinusoids with different phases FIGURE 10.3 Triangle useful
in adding two sinusoids

where, by Fig. 10.3.

= tan'4 (10.10)
A

A similar result may be established if the sine and cosine terms have phase angles
other than zero, indicating that, in general, the sum of two sinusoids of a given
frequency is another sinusoid of the same frequency.
As an example, we have
12
-5 cos + = J5 + -
3t 12 sin 3r 1
12 2 cos 3/ tan" 1

m.
= 13 cos (3/ - 112.6°)

since tan
-1
(12/— 5) is in the second quadrant, because A = —5 < and B — 12 > 0.

EXERCISES jCl, /
c
10.1.1 Find the period of the following sinusoids:
(a) 6sin(5r + 17°) f.
SOI
(b) cos (2/ + + 2 sin (it - |-).
-J)
(c) 4 cos In t. Ans. (a) 271/5, (b) n, (c) 1

10.1.2 Find the amplitude and phase of the following sinusoids:

(a) + 4 sin 2t.
3 cos 2t
(b) (4^T - 3) cos (2t + 30°) + (3^3~ - 4) cos (It + 60°).

[Suggestion: In (b), expand both functions and use (10.9).]

Ans. (a) 5, -53.1°; (b) 5, 36.9°

10.1.3 Find the frequency of the following sinusoids:

(a) 3 cos (4nt - 10°).

(b) 4 sin 377/. Ans. (a) 2, (b) 60 Hz

248 Sinusoidal Excitation and Phasors Chap. 10

10.2 AN RL CIRCUIT EXAMPLE

As an example of a circuit with a sinusoidal excitation, let us find the forced com-
ponent if of the current /'
in Fig. 10.4. The describing equation is

L^L + Ri = vm cos cot (10.11)

at

and, following the method of the previous chapter, let us assume the trial solution

if = A cos cot + B sin cot

Substituting the trial solution into (10.1 1), we have

L{— coA sin cot + coB cos cot) + R(A cos cot + B sin cot) = Vm cos cot
Therefore equating coefficients of like terms, we must have

RA + = Vm
coLB
-coLA + RB =
from which

= RV„
A
R2 + co
2
L2
coLVm
B=
R2 + co
2
L2
The forced response is then

if
f
=
R2 +
RV„
co
2
L 2 ^
cos ""
cot — -^-
i? + 2
—L co
2__
2 2
Si n at

which by (10.9) and (10.10) may be written as

V. ,coL
'/ cos cot — tan (10.12)
JR + co L 2 2
R)

Vm cos cot

FIGURE 10.4 RL circuit

Chap. 10 Sinusoidal Excitation and Phasors 249

The forced response is, therefore, a sinusoid like the excitation, as we predicted when
we chose the trial solution. We may write it in the form

if = Im cos (cot + (f>)

(10.13)

where

L=
V-R 2 + co
2
L 2

and

<t>
= -tan-^ (10.14)

Since the natural response is

i„ = A l
e'
R,L

it is clear that after a short time /„

— * 0, and the current settles down to its ac steady-
state value, givenby (10.12).
The method we have used is straightforward and conventional but, as the reader
might agree, is rather laborious for such a simple problem. For a second-order circuit,
the method is more tedious, as was illustrated by the example of (9.34) in the previous
chapter. For very high-order circuits the procedure is, of course, even more com-
plicated. Evidently we need a better method. One such method is developed in the
remainder of this chapter, and its use allows us to treat circuits with storage elements
in the same way that we treated resistive circuits in Chapters 2, 4, and 5.

EXERCISES
10.2.1 Find the forced response i
f in Fig. 10.4 if L = 4 mH, R = 3 kfi, Vm = 5 V,
and co = 10 6 rad/s. Ans. cos (10 6 r - 53.1°) mA
10.2.2 Find the forced component of v.

Ans. (RIJy/l - co 1 R2 C 1 ) cos (cot - tan" 1

coRC) V

EXERCISE 10.2.2
250 Sinusoidal Excitation and Phasors Chap. 10

10.3 AN ALTERNATIVE METHOD USING

COMPLEX NUMBERS

The alternative method of analyzing which we

circuits with sinusoidal excitations,
shall consider in the on the concept of complex
remainder of the chapter, relies heavily

numbers. The reader who is unfamiliar with complex numbers, or who needs to review
the subject, should consult Appendices C and D, where complex numbers and their
properties are discussed in some detail. For convenience, we shall list a few of these
properties in this section before considering the alternative method of analysis.
The complex number A is written in rectangular form as

A = a+jb (10.15)

where y = <J — 1 and the real numbers a and b are the real and imaginary parts of A,
respectively. Equivalently we may say that

a = Re A, b = Im A

where Re and Im denote "the real part of" and "the imaginary part of."
The number A may be written also in the polar form,

A = \A\e'* = \A\\aL (10.16)

where \A |
is the magnitude, given by

and a is the angle or argument, given by

a = tan"
a

These relations between rectangular and polar forms are illustrated in Fig. 10.5.
As an example, suppose we have A — A +7*3. Then \A |
= *J4
2
+ 3
2
= 5 and
a = tan
-1
| = 36.9°. Therefore the polar form is

A = 5 /36.9°

As another example, consider A = —5 — jl2. Since both a and b are negative, the
line segment representing A lies in the third quadrant, as shown in Fig. 10.6, from
which we see that

\A\ = V5 + 2
12 2 = 13

and
a= 180° + tan-'-^ = 247.4°

Thus we have A = 13/247.4°.

Chap. 10 Sinusoidal Excitation and Phasors 251

A= o + |b

FIGURE 10.5 Geometrical representation of FIGURE 10.6 Complex number with negative
a complex number A real and imaginary parts

Other useful results are

j= /90° 1

f= - = 1 1/180"

and so forth.
By Euler*s formula, which is discussed in Appendix D and which we have used
previously in Chapter 9, we know that

Vm cos cot + jVm sin cot = Vm e °" J

Therefore we may say that

Vm cos cot = Re(Vm e ia

") (10.17)

and

Vm sin cot = lm(Vm e ja

")

Returning to the RL circuit example of Fig. 10.4, we know that exponentials are
mathematically easier to handle as excitations than sinusoids. Therefore let us see
what happens if we apply the complex excitation

y e
j<»<
(10.18)

instead of the real excitation

ve = Vm cos cot = Re v x
(10.19)

We cannot duplicate such a complex excitation in the laboratory, but there is no

reason we cannot consider it abstractly. In this case the forced component of the
252 Sinusoidal Excitation and Phasors Chap. 10

current, which we shall call /,, satisfies

L dii + /?/, Vm e J °"

(10.20)
dt

To solve this equation, we try

i
f
= Ae' u

which, substituted into (10.20), yields

(jcoL + R)Ae im ' = Vm e io"

from which

A =
R + jcoL
-
j tan ' coL/R

JR. 1 + co 2
L2

using complex number division. Therefore we have

;'(cu(-tan-' 01L! R)

JR + co 2 2
L 2

Let us now observe that

Re/, = Re ,y(ajf-tan"' oiL/R)

WR 2
+ o>
2
L2
Vm —
JR. 2
+ co
-^-^ cos (cot
2
L2 \
tan 1

%R)
r )

which, by (10.12), is the correct forced response of Fig. 10.4. That is,

i
f
= Re /j (10.21)

We have established for this example the interesting result that if/, is the complex
response to the complex forcing function v u then i
f
= Re/, the response to is vg =
Re v1 . That is, v x
yields /, and Re v x
= v g yields Re /, = f The reason for this
i . is that
the describing equation (10.20) contains only real coefficients. Thus, from (10.20), we
have

di
Re^Ly + Ri\ = Rev,

or

L -^(Re /,) + R(Re /,) = Vm cos cot

Chap. 10 Sinusoidal Excitation and Phasors 253

and therefore, by (10.11),

i = i
f
= Re/, (10.22)

Thus we see that it is easier to use the complex forcing function i\ to find the
complex response /,. Then since the real forcing function is Re v u the real response is
Re /,. This principle holds for all our circuit analyses, since the describing equations
are linear with real coefficients, as may be seen in a development analogous to that
leading to (10.22).

EXERCISES
10.3.1 Replace the real forcing function Im cos cot in Ex. 10.2.2 by the complex forcing
function Im e Ja", find the resulting complex response v u and show that the real
response is v = Re vx .

10.3.2 Show that, for a real,

H % =a &
a
dx\ d
{K ' x)

and use this result to establish (10.22). {Suggestion: Let x = f + jg, where/
and g are real.)

10.4 COMPLEX EXCITATIONS

Let us now
generalize the results using complex excitation functions in the previous
section.The excitation, as well as the forced response, may be a sinusoidal voltage or
current. However, to be specific, let us consider the input to be a voltage source and
the output to be a current through some element. The other cases may be considered
in an analogous way.
In general, we know that if

Vg = Vm cos(tot + 6) (10.23)

then the forced response is of the form

/ = Im cos (ot + 0) (10.24)

as indicated in the general circuit of Fig. 10.7. Therefore if by some means we can find
Im and 0, we have our answer, since co, 9, and Vm are known.
 ,

254 Sinusoidal Excitation and Phasors Chap. 10

Vq = Vm COSlcot +9) j
i
= Im cos(cot +<P)

FIGURE 10.7 Genera/ circuit with input and output

To solve for / in Fig. 10.7, let us apply the complex excitation

v, = Vm e J{a" +e)
(10.25)

and find the complex response i u as shown in Fig. 10.8. Then we know from the
results of the previous section that the real response of Fig. 10.7 is

i = Re / (10.26)

This is a consequence of the fact that the coefficients in the describing equation are
real, as pointed out previously.

j (co1 1 )

v, = Vm e

FIGURE 10.8 General circuit with a complex excitation

The describing equation may be solved for the forced response by the method of
Chapter 9. That is, since we may write the excitation as

v x
= Vm e je
e i0
"
(10.27)

which is a constant times e ia", then the trial solution is

/, = Ae icot

Comparing (10.24) and (10.26), we must have

I m cos (at + <j>) = Re [Ae j0»]

which requires that

A = Im e j *
Chap. 10 Sinusoidal Excitation and Phasors 255

and hence
i, = Im e J +e JO
"
(10.28)

Taking the real part, we have the solution (10.24), of course.

As an example, let us find the forced response i f of

d2 i =
dT2 , +
,

^
r.
2
di_
^
dl
+ 8/ 12^/T cos (2f + 15°)

First we replace the real excitation by the complex excitation,

v, = l2^Je Jii2,+ iy)

where for convenience the phase is written in degrees. (This is, of course, an inconsis-
tent mathematical expression, but as long as we interpret it correctly it should present
no difficulty.) The complex response i, satisfies

+15t
^1 + 2^+8/,= 12 x/T^' (2
' >

and it must have the general form

i, = Ae i2 '

Therefore we must have

(-4 +y4 + S)Ae J2 ' = 12^/2 e i2 'e

jl5 °

or

A
n^/T^ _ 12VT/15_°
~
4+y4 4^T/45° ~ ^^-^
which gives
/, = (3 /-30V 2'

_ 3 e ;(2t-30°)
Thus the real answer is

i
f
= Re/, = 3 cos (2?- 30°)

EXERCISES
10.4.1 Find the forced response v using the method of this section.

Ans. 5 cos (16/ - 28.1°) V

256 Sinusoidal Excitation and Phasors Chap. 10

17 cosl6t V v

EXERCISE 10.4.1

10.4.2 Find the forced response i using the method of this section if vg = 6 sin At.
[Suggestion : sin At = Im e'*'] Ans. 3 sin At A

I H 2 n
->m-

© ZZ 1/16 F

EXERCISE 10.4.2

10.4.3 Repeat Ex. 10.4.2 if vg = 6 cos At. Ans. 3 cos At A

10.5 PHASORS

The results obtained in the previous section may be put in much more compact form
by the use of quantities called phasors, which we shall introduce in this section. The
phasor method of analyzing circuits is credited generally to Charles Proteus Steinmetz
(1865-1923), a famous electrical engineer with the General Electric Company in the
early part of this century.
To begin, let us recall the general sinusoidal voltage,

v = Vm cos {cot + 9) (10.29)

which, of course, is the source voltage v g of the previous section. If the frequency co is

known, then v is completely specified by its amplitude Vm and its phase 9. These
quantities are displayed in a related complex number,

V = Vn e» = VJ9 (10.30)

which is defined as a phasor, or a phasor representation. To distinguish them from other

complex numbers, phasors are printed in boldface type, as indicated.

The motivation for the phasor definition may be seen from the equivalence, by
Euler's formula, of

Vm cos (cot + 9) = Re (Vm e jee J °*)

(10.31)
Chap. 10 Sinusoidal Excitation and Phasors 257

Therefore, in view of (10.29) and (10.30), we have

w = Re(Ve**) (10.32)

As an example, suppose we have

v= 10 cos (4/ + 30°) V

The phasor representation is then

V= 10/30_° V
since Vm =
10 and 9 = 30°. Conversely, since co = 4 rad/s is assumed to be known,
v is readily obtained from V.
In an identical fashion we define the phasor representation of the time-domain
current

i = /„ cos (cot + 0) (10.33)

to be

\ = Im e* = lJ± (10.34)

Thus if we know, for example, that co = 6 rad/s and that I = 2/15° A, then we have

i= 2 cos (6/+ 15°) A

We have chosen to represent sinusoids and their related phasors on the basis of
cosine functions, though we could have chosen sine functions just as easily. Therefore
if a function such as

v = 8 sin (3? + 30°)

is given, we may change it to

v = 8 cos (3/ + 30° - 90°)

= 8 cos (3/ - 60°)

Then the phasor representation is

V= 8 /-60

To see how the use of phasors can greatly shorten the work, let us reconsider Fig.
10.4 and its describing equation (10.11), rewritten as

L^- + Ri = Vm cos cot (10.35)

at

Following our method, we replace the excitation Vm cos cot by the complex forcing
function

Vl = Vm e^'
 ,

258 Sinusoidal Excitation and Phasors Chap. 10

which may be written

Vl = \e Ja"

since 6 = 0, and therefore V = Vm [0_ = Vm . Substituting this value and i = i, into

(10.35), we have

L^i + Ri, '

= Ve Jat
dt

whose solution /, is related to the real solution i by

i = Re /

Next, trying

/j = Ie
/»>'

as a solution, we have
jcoLle im! + ja
/?Ie " = Ve- Q"
'

Dividing out the factor e ;wr we have the phasor equation

,

jcoLl + Rl =V (10.36)

Therefore

K- 1
'-tan-*R
/? + ;o>L Jri + co
2
L2 1

Substituting this value into the expression for /,, we have

it = ,Jloit-tan-' coLIR)

JR 1
+(o 2 L 2

Taking the real part, we have i = i f obtained earlier in (10.12).

,

It is important to note that if we can go directly from (10.35) to (10.36), there is a

vast saving of time and effort. Also, in the process we have converted the differential
equation into an algebraic equation, somewhat like those encountered in resistive
circuits. Indeed, the only difference is that the numbers here are complex, whereas in

resistive circuits they were real. With the hand calculator as commonly available as it
is today, even the complexity of the numbers presents little difficulty.

In the remainder of the chapter we shall see how to bypass all the steps between
(10.35) and (10.36) by studying the phasor relationships of the circuit elements and
considering Kirchhoff's laws as they pertain to phasors. Indeed, as we shall see, we
may go directly from the circuit to (10.36), bypassing even the step of writing down the
differential equation.
In general, the real solutions are time-domain functions, and their phasors are
frequency-domain functions; i.e., they are functions of the frequency a. This is illus-

trated by the phasor I of the last example. Thus to solve the time-domain problems, we
Chap. 10 Sinusoidal Excitation and Phasors 259

may convert to phasors and solve the corresponding frequency-domain problems,

which are generally much easier. Finally, we convert back to the time domain by
finding the time function from its phasor representation.

EXERCISES
10.5.1 Find the phasor representation of (a) 10 cos {It -f 45°), (b) 4 cos 2/ + 3 sin
It, and (c) -6 sin (5f - 75°).

Arts, (a) 10/45% (b) 5/ -36.9° ,

(c) 6/15°

10.5.2 Find the time-domain function represented by the phasors (a) 10 / — 12° , (b)
6 +j8, and (c) —jA. In all cases co = 3.
Ans. (a) 10 cos (3* - 12°), (b) 10 cos (3? + 53.1°), (c) 4 cos (3f - 90°)

10.6 VOLTAGE-CURRENT RELATIONSHIPS

FOR PHASORS

In this section we show that relations between phasor voltage and phasor current
shall
for resistors, inductors, and capacitors are very similar to Ohm's law for resistors. In
fact, the phasor voltage is proportional to the phasor current, as in Ohm's law, with
the proportionality factor being a constant or a function of the frequency co.

We begin by considering the voltage-current relation for the resistor,

v = Ri (10.37)

where

v = Vm cos (cot + 6)
(10.38)
i = Im cos (cot + 0)
+e)
If we apply the complex voltage Ym e Jico '
, the complex current which results is
ibat+
Im e *\ which substituted into (10.37) yields

+ B) __ +
y e
J(cot
ft J e
J(cot 4>)

Dividing out the factor e Ja" results in

Vm e je = RIm e» (10.39)

which, since Vm e' e and lm e'* are the phasors V and I, respectively, reduces to

V = R\ (10.40)
260 Sinusoidal Excitation and Phasors Chap. 10

Thus the phasor or frequency-domain relation for the resistor is exactly like the time-
domain relation. The voltage-current relations for the resistor are illustrated in
Fig. 10.9.

v = Ri V = RI

(a) (b)

FIG U R E 1 0.9 Voltage-current relations for a resistor R in the (a) time

and (b) frequency domains

From (10.39) we have Vm RIm and 6 = <f>.

=
Thus the sinusoidal voltage and
current for a resistor have the same phase angle, in which case they are said to be in
phase. This phase relationship is shown in Fig. 10.10, where the voltage is represented
by the solid line and the current by the dashed line.
As an illustration, suppose that the voltage

v= 10 cos (100/ + 30°) V (10.41)

is applied across a 5-Q resistor, with the polarity indicated in Fig. 10.9(a). Then the
phasor voltage is

and the phasor current is

v,i

FIGURE 10.10 Voltage and current waveforms for a resistor

Chap. 10 Sinusoidal Excitation and Phasors 261

Therefore in the time domain we have

i = 2 cos (100/ + 30°) A (10.42)

This is. of course, simply the result we would have obtained using Ohm's law.
In the case of the inductor, substituting the complex current and voltage into the
time-domain relation,

di
v =L ,

dt

gives the complex relation

Vm e Ha" +e) = L^-[/m ^ to ' + *>]

Again, dividing out the factor e ico and identifying the phasors, we obtain the phasor
'

relation

\=jcoLl (10.43)

Thus the phasor voltage V, as in Ohm's law, is proportional to the phasor current I,

with the proportionality factor jcoL. The voltage-current relations for the inductor are
shown in Fig. 10.1 1.

i I

O 1 o
+ +

v=Ldi/dt <3 L V = jloLI 3 jujL

(a) (b)

FIGURE 10.11 Voltage-current relations for an inductor L in the (a)

time and (b) frequency domains

If the current in the inductor is given by the second of (10.38), then by (10.43), the
phasor voltage is

v = (M)(/J)
= coLIJtj) + 90°

since j = 1 /90° Therefore in the time

. domain we have

v = coLIm cos (cot t <f> -r 90°)

Comparing this result with the second of (10.38), we see that in the case of an inductor
262 Sinusoidal Excitation and Phasors Chap. 10

the current lags the voltage by 90°. Another expression that is used is that the current
and voltage are 90° out of phase. This is shown graphically in Fig. 10.12.

V,l

FIGURE 10.12 Voltage and current waveforms for an inductor

Finally, let us consider the capacitor. Substituting the complex current and voltage
into the time-domain relation.

i = C dv
di

gives the complex relation

J
dt

= jcoCVm e Ho" +m
Again dividing by e
j0J '
and identifying the phasors, we obtain the phasor relation

I=jo>CV (10.44)

or

(10.45)
jcoC

Thus the phasor voltage V is proportional to the phasor current I, with the propor-
tionality factor given by 1/jcoC. The voltage-current relations for a capacitor in the
time and frequency domains are shown in Fig. 10.13.
In the general case, if the capacitor voltage is given by the first equation of (10.38),
then by (10.44), the phasor current is

/ = (jcoC)(VJ9)
= coCVje + 90°
Chap. 10 Sinusoidal Excitation and Phasors 263

i
= Cdv/dt 1= jujCV

jooC

(a) (b)

FIGURE 10.13 Voltage-current relations for a capacitor in the (a) time

and (b) frequency domains

Therefore in the time domain we have

/ = coCVm cos (cot + 9 + 90°)

which, by comparison with the first equation of (10.38), indicates that in the case of a

capacitor the current and voltage are out of phase with the current leading the voltage
by 90°. This is shown graphically in Fig. 10.14.

v,i

FIG U R E 1 0.1 4 Voltage and current waveforms for a capacitor

As an example, if the voltage of (10.41) is applied across a 1-//F capacitor, then by

(10.44) the phasor current is

I=y(100)(10-«)(10/30°)A
= 1 /120° mA
The time-domain current is then

/ = cos (100/ + 120°) mA

and therefore the current leads the voltage by 90°.
264 Sinusoidal Excitation and Phasors Chap. 10

EXERCISES
10.6.1 Using phasors, find the ac steady-state current iif v = 12 cos (1000/ + 30°) V
in (a) Fig. 10.9(a) for R = 4kft, (b) Fig. 10.11(a) for L = 20 mH, and (c)

Fig. 10.13(a) for C= 1 jnF.

Am. (a) 3 cos (1000? + 30°) mA, (b) 0.6 cos (1000/ - 60°) A,
(c) 12 cos (1000/ + 120°) mA
10.6.2 In Ex. 10.6.1, find i in each case at / = 1 ms.
Ans. (a) 0.142 mA, (b) 0.599 A, (c) -11.987 mA

10.7 IMPEDANCE AND ADMITTANCE

Let us now consider a general phasor circuit with two accessible terminals, as shown in
Fig. 10.15. If the time-domain voltage and current at the terminals are given by
(10.38), then the phasor quantities at the terminals are

v = vje
— (10.46)
I = UA
We define the ratio of the phasor voltage to the phasor current as the impedance of
the circuit, which we denote by Z. That is,

(10.47)
i

which by (10.46) is

Z= \Z\/0 z = ^f /d-4> (10.48)

where |
Z |
is the magnitude and 6 Z the angle of Z. Evidently,

|z| = 7=» ez = e-<t>

* m

Impedance, as is seen from (10.47), plays the role, in a general circuit, of resistance in
resistive circuits. Indeed, (10.47) looks very much like Ohm's law; also like resistance,
impedance is ohms, being a ratio of volts to amperes.
measured in
It is important to stress that impedance is a complex number, being the ratio of two

complex numbers, but it is not a phasor. That is, it has no corresponding sinusoidal
time-domain function of any physical meaning, as current and voltage phasors have.
The impedance Z is written in polar form in (10.48); in rectangular form it is
Chap. 10 Sinusoidal Excitation and Phasors 265

generally denoted by

Z = R + jX (I0.49)

where R = Re Z is the resistive component, or simply resistance, and X = Im Z is the

reactive component, or reactance. In general, Z= Z(ja>) is complex function ofy'a),
a
but 7? = /?(co) and X = A'(co) are real functions of co. Both R and A', like Z, are
measured in ohms. Evidently, comparing (10.48) and (10.49) we may write

|Z| JR + X Z 1

ez tan -i
*
A'

and
7? = Z | |
cos Z

X= |Z|sin6> z
These relations are shown graphically in Fig. 10.16.
As an example, suppose in Fig. 10.15 that V = 10 /56.1° V and I = 2 /20° A. Then
we have
10/56.1°
= 5 / 36
jo
n
2/20°

In rectangular form this is

Z = 5(cos 36.1° +j sin 36.1°)

= 4+;3Q
The impedances of resistors, inductors, and capacitors are readily found from their
V-I relations of (10.40), (10.43), and (10.45). Distinguishing their impedances with
subscripts R, L, and C, respectively, we have, from these equations and (10.47),

ZR = R
Z L = joiL = coL/90°
(10.50)

PHASOR
V
CIRCUIT

FIGURE 10.15 Genera/ phasor circuit FIGURE 10.16 Graphical representation

of impedance
266 Sinusoidal Excitation and Phasors Chap. 10

In the case of a resistor, the impedance is purely resistive, its reactance being zero.
Impedances of inductors and capacitors are purely reactive, having zero resistive
components. The inductive reactance is denoted by

XL = coL (10.51)

so that

ZL =jXL

and the capacitive reactance is denoted by

z'--£s < 10 52 >

-

and thus

Z c =jXc (10.53)

Since co, L, and C are positive, we see that inductive reactance is positive and that
capacitive reactance is negative. In the general case of (10.49), we may have X— 0, in

which case the circuit appears to be resistive; X> 0, in which case its reactance
appears to be inductive; and X < 0, in which case its reactance appears to be capaci-
tive. These cases are possible when resistance, inductance, and capacitance are all
present in the circuit, as we shall see. As an example, the circuit with impedance given
by Z— 4 + y'3, which we have just considered, has reactance X = 3, which is of the
inductive type. In all cases of passive circuits, as we shall see in Chapter 12, the

resistance R is nonnegative.
The reciprocal of impedance, denoted

Y = -L (10.54)

is called admittance and is analogous to conductance (the reciprocal of resistance) in

resistive circuits. Evidently, since Z is a complex number, then so is Y, the standard
representation being

Y = G+jB (10.55)

The quantities G = Re Y and B = Im Y are, respectively, called conductance and

susceptance and are related to the impedance components by

Y = e +>* = ( '°' 56)

-Z-IT7*
The units of Y, G, and B are all mhos, since in general Y is the ratio of a current to a
voltage phasor.
Chap. 10 Sinusoidal Excitation and Phasors 267

To obtain the relation between components of Y and Z we may rationalize the last
member of (10.56). which results in

R-jX
Rz + X 2
Equating real and imaginary parts results in

R
G
R1 + X 2

(10.57)
D _ X
R + X1 2

Therefore we note that R and G are not reciprocals except in the purely resistive case
{X — 0). Similarly, X and B are never reciprocals, but in the purely reactive case
(R = 0) they are negative reciprocals.
As an example, if we have
Z = 4+y3
then

v- ^__ 4-y3 _
42 + 2
4
25
J
J
3_
1?
25
;3 3

Therefore G— -^ and B =- —
__!_
^
25-
Further examples are

YR = G
YL --L
~jcoL
Y c = jcoC
which are the admittances of a resistor, with R— \]G, an inductor, and a capacitor.

EXERCISES
10.7.1 Find the impedance seen at the terminals of the source in Fig. 10.4 in both
rectangular and polar form.
Ans. R +j(OL, ^R + co
2 2
L 2 ltan~ coL/R
l
268 Sinusoidal Excitation and Phasors Chap. 10

10.7.2 Find the admittance seen at the terminals of the source in Fig. 10.4 in both
rectangular and polar form.
. R coL 1
/ _ coL
Ans Ri J
l

-
+ Oi'-U- Ri + C0 2 L 2 '

JR* + CQiL ! 2 R
10.7.3 Find the conductance and susceptance if Z is (a) 5 +7 12, (b) 0.8 — y'0.6, and
(c) ^2/45°. Ans (a)^,
hb, - J&,
--&, (b) 0.8, 0.6, (c) i,
{,
- -\

10.8 KIRCHHOFF'S LAWS AND IMPEDANCE

COMBINATIONS

Kirchhoffs laws hold for phasors as well as for their corresponding time-domain
We may see this by observing that if a complex excitation, say
voltages or currents.
Vm e il0J,+e \ is applied to a circuit, then complex voltages, such as V 1
e Mo" +Bl) ,
V2 e Jlo" +e,-\ etc., appear across the elements in the circuit. Since Kirchhoffs laws hold
in the time domain, KVL applied around a typical loop results in an equation such as

y e
j(ait+e,) _j_ y e
j(mt+e 2 ) _i
, -\- V e
iio" +BN) =
Dividing out the common factor e io", we have

V, +v + 2 . . . + v„ =
where

VJd„ n = l,2,...,N

are the phasor voltages around the loop. Thus KVL holds for phasors. A similar
development will also establish KCL.
In circuits having sinusoidal excitations with a common frequency co, if we are
interested only in the forced, or ac steady-state, response, we may find the phasor
voltages or currents of every element and use Kirchhoffs laws to complete the
analysis. The ac steady-state analysis is therefore identical to the resistive circuit
analysis of Chapters 2, 4, and 5, with impedances replacing resistances and phasors
replacing time-domain quantities. Once we have found the phasors, we can convert
immediately to the time-domain sinusoidal answers.
As an example, consider the circuit of Fig. 10.17, which consists of N impedances
connected in series. By KCL for phasors, the single phasor current I flows in each
element. Therefore the voltages shown across each element are

V,=Z,I
v =Z I
2 2

Vjv — LN i
Chap. 10 Sinusoidal Excitation and Phasors 269

+ o-
*l
L
2 ^
+ + +
V — v ,
V2 VN

FIGURE 10.17 Impedances connected in series

and by KVL around the circuit,

V = V, + V 2 + . . . + VN
= (Z, +Z + 2 ... +Z JV )I

Since we must also have, from Fig. 10.17,

V = Z eq I

where Z eq is the equivalent impedance seen at the terminals, it follows that

Z. eq = Zq +Z 2 -p . . . -p ZN (10.58)

as in the case of series resistors.

Similarly, as was the case for parallel conductances in Chapter 2, the equivalent
admittance Y eq of A r
parallel admittances is

Y2 + . . . + Y„ (10.59)

In the case of two parallel elements (N = 2), we have

1 1 Z]Z;
(10.60)
Y, +Y 2 Z, +Z 2

In like manner, voltage and current division rules hold for phasor circuits, with
impedances and frequency-domain quantities, in exactly the same way that they held
for resistive circuits, with resistances and time-domain quantities. The reader is asked
to establish these rules in Ex. 10.8.2.
For example, let us to return to the RL circuit considered in Sec. 10.2. The circuit
and its phasor counterpart are shown in Figs. 10.18(a) and (b), respectively. By KVL

in the phasor circuit we have

ZJ + RI= VJ0
270 Sinusoidal Excitation and Phasors Chap. 10

(a) (b)

FIGURE 10.18 (a) Time-domain circuit; (b) equivalent phasor circuit

or

(j(oL + R)I = VJQ_

from which the phasor current is

VJO
R + jcoL
iCoL
-tan
V-R 2
+ co
2
L2 ~R

Therefore in the time domain we have, as before.

cos ( -tan-^)
JR + a2 2
LI

An alternative method of solution is to observe that the impedance Z seen at the

source terminals is the impedance of the inductor, jcoL, and the resistor, R, connected
in series. Therefore
Z = jcoL + R
and

"
Z R + jcoL
as obtained earlier.

EXERCISES
10.8.1 Derive (10.59).

10.8.2 Show in (a) that the voltage division rule,

Z 1
+Z 2
Chap. 10 Sinusoidal Excitation and Phasors 271

and in (b) that the current division rule,

_ Y2 _ Z
Y,+Y I.=
¥ T 1

*« Z,+Z 2 2

are valid, where Z, = 1/Y, and Z2 = 1/Y 2 .

+ o- o t

o •

(o) (b)

EXERCISE 10.8.2

10.8.3 Find the steady-state current i using phasors.

Ans. 2 cos (4/ — 45°) A

3 n I H
AW 'TRRP-
+

6/2 cos 4t (tj 1/4 F

EXERCISE 10.8.3

10.8.4 Find the steady-state value of v in Ex. 10.8.3 using voltage division. Check by
using the fact that the current leads the voltage by 90° in a capacitor.
Ans. 2 cos (4/ - 135°) V

10.9 PHASOR CIRCUITS

As the discussion in the previous section suggests, we may omit the steps of finding the
describing equation in the time domain, replacing the excitations and responses by
their complex forcing functions and then dividing the equation through by eJa" to
obtain the phasor equation. We may simply start with the phasor circuit, which is the
time-domain circuit with the voltages and currents replaced by their phasors and the
elements identified by their impedances, as illustrated previously in Fig. 10.18(b). The
describing equation obtained from this circuit is then the phasor equation. Solving
this equation yields the phasor of the answer, which then may be converted to the time-

domain answer.
272 Sinusoidal Excitation and Phasors Chap. 10

The procedure from starting with the phasor circuit to obtaining the phasor
answer is identical to that used earlier in resistive circuits. The only difference is that
the numbers are complex.
As an example, let us find the steady-state current i in Fig. 10.19(a). The phasor
circuit, shown in Fig. 10.19(b), is obtained by replacing the voltage source and the
currents by their phasors and labeling the elements with their impedances. In the
phasor circuit the impedance seen from the source terminals is

7 _ i .
(3+73)(-j3)

= 4-j3Q
Therefore we have

5/0 5/0
I, 1/36.9°
4 - ;3 5/-36.9
c

and by current division,

In the time domain, the answer is

/ = v^cos(3/ + 81.9°) A

1/9 F

(a)

5ZQ V ^ -j3 n

(b)

FIGURE 10.19 RLC time-domain and phasor circuit

Chap. 10 Sinusoidal Excitation and Phasors 273

In the case of a dependent source, such as a source kv x volts controlled by a volt-

age i\. it will appear in the phasor circuit as a source k\ x where V* is the phasor
,

representation of rx . because vx <f>)

= Vm domain will be-
cos (cot -| in the time
come Vm e JUot ** )
when
complex excitation is applied. Then dividing e j0" out
a
of the equations leaves v x represented by its phasor Vm e'*. In the same way,
kr x = kVm cos (cot + (f>) is represented by its phasor kVm e J *, which is k times the
phasor of >\.
As an example of a circuit containing a dependent source, let us consider Fig.
10.20(a). in which it is required to find the steady-state value of/. The corresponding
phasor circuit is shown in Fig. 10.20(b). Since phasor circuits are analyzed exactly like
resistive circuits, we may write a nodal equation at node a in Fig. 10.20(b), resulting in

14 3/0 (10.61)
-J2
By Ohm's law we have V, = 41. which substituted into (10.61) yields

-j21 4 ^(41) = -J6

or

-j6 6 /-90
I
- )2 /-45°A
2 2y r 2~ /-45
Therefore we have
3
cos (4r - 45°) A

1/8 F

3 cos 4 t 1/2 v

(a)

]2 n

3/0 A 1/2 V,

(b)

FIGURE 10.20 (a) Circuit containing a dependent source; (b)

corresponding phasor circuit
274 Sinusoidal Excitation and Phasors Chap. 10

In the case of an op amp. the phasor circuit is the same as the time-domain circuit.
That is. an ideal op amp in the time-domain circuit appears as an ideal op amp in the
phasor circuit, because the time-domain equations

i = 0, v =
which characterize the current into and the voltage across the input terminals retain
the identical form,

1 = 0, V=
in the phasor equations.

EXERCISES
10.9.1 Solve Ex. 10.2.2 by means of the phasor circuit.

10.9.2 Find the forced response v in the figure for Ex. 10.4.1 by means of the phasor
circuit.

10.9.3 Find the forced response i in the figure for Ex. 10.4.2 by means of the phasor
circuit.

PROBLEMS
10.1 Given
vi = 10 cos At

v2 = 20(cos At + +/T sin At)

determine if u t leads or lags v 2 and by what amount.

10.2 Find z4 , using only the properties of sinusoids, if i{ = 2 cos 3t,

iz = A cos (3/ - 30°), and i3 = 6 cos (3/ - 60°).

PROBLEM 10.2
Chap. 10 Sinusoidal Excitation and Phasors 275

10.3 In the figure of Ex. 10.2.2, if the source is 2 cos 5000/ mA and the output is

v = 4 cos (5000/ 45 ) V, find R and C.

10.4 A voltage Vm and an inductor L are all in scries.

sin cot V, a resistor /?,

Find the forced response two ways, (a) Noting that

i in the following
Vm sin cot = Vm cos {cot — 90), use the complex excitation Vm e ito"~ 90) and take
the real part of the current response, (b) Noting that Vm sin cot = Im ( Vm c'""),
use the complex excitation Vm e Jo" and take the imaginary part of the current
response.

10.5 A complex voltage source 4^' (3 ' f3 °

» V produces a current response 0.5c-'' 3
'" 10 '

A. Find the current response to a voltage source of (a) \6e J{i "* s) V, (b)
8 cos (3/ + 35 ) V, and (c) 2 sin (3/ - 60 ) V.

10.6 If co = 4 rad/s, find the time-domain functions represented by the phasors

(a) -6 -jS, (b) -6 +/8, (c) -ylO, and (d) -10.

10.7 Find the impedance of the circuit of Fig. 10.15 if the time-domain functions
represented by the phasors V and I are
(a) -5 cos It + 12 sin It V, = 1.3 cos (2/ + 40°) A.
v = i

(b) = RefyW V, = Re [(1 +j)e'^'^o mA

v 2 ']
i )]

(c) v = aVm cos (COt + 8) V, = Vm cos (cot + 9 — a) A. i

10.8 Three admittances, 5 + j6, 7 — j], and 17 /taiT -^ are connected 1

, in parallel.

Find the equivalent impedance of the combination.

10.9 Show that if Re Z =R is positive, then Re Y = Re [1/Z] = G is also positive.

10.10 A circuit has an impedance

16(2 +7<P)(8 -6) 2 -;2qj)

Z= - Q
co 4 \5co 2 + 64

Find the reactance at CO = 1 rad/s, 2 rad/s, and 3 rad/s. If the time-domain

voltage applied to the circuit is 64 cos cot V, find the steady-state current i

produced in each case.

10.11 Find the steady-state response v.

0.05 H

4 cos 20,000 1 I kO
mA

PROBLEM 10.11

10.12 Find the power being delivered by the source in Prob. 10.11 at / = 7T/20 ms.

10.13 Show that if the capacitance in the figure for Prob. 10.1 1 is changed to 0.025 /zF,
then the impedance seen at the terminals of the source is real. In this case, find
the power being delivered by the source at / = fl/20 ms.

10.14 Find the steady-state value of v using phasors.

 —
276 Sinusoidal Excitation and Phasors Chap. 10

I/3 H

(P
4 cos 6t (tj I/I2 F

PROBLEM 10.14

10.15 Find the steady-state value of i using phasors.

^IHF

PROBLEM 10.15

10.16 Find the steady-state value of v using phasors and voltage division.

3 0. I H i n
-A/vv nsw-

1/2 H
9 cos 2
V
t
© io
1/4 F

PROBLEM 10.16

10.17 Find the steady-state value of / if (a) co = 4 rad/s and (b) co = 2 rad/s.Note that
in the latter case the impedance seen at the terminals of the source is purely
resistive.

I H 4 n

10 cos oot

V
© 1/4 F

PROBLEM 10.17

10.18 Find the steady-state current i.

 i

Chap. 10 Sinusoidal Excitation and Phasors 277

I H I H

-w-

10 cos

V
t

© 3 n

PROBLEM 10.18

10.19 Find the steady-state values of i and v.

_!_^ 1/2 H H
—nm^- I

^rkp —
2 H
4 cos
V
t
© I F
1/4 F

PROBLEM 10.19

10.20 Find the steady-state value of i.

I/2(jF

PROBLEM 10.20

10.21 Find the steady-state value of v.

I kO 3 kO.

-A/W J VW^

6 cos 40,000 t 0.025 uF. 5v, O.IHoj v

PROBLEM 10.21

10.22 Find the steady-state value of v if vg = 5 cos 10,000/ V.

278 Sinusoidal Excitation and Phasors Chap. 10

PROBLEM 10.22

10.23 Find the steady-state current i if vg = 2 cos 4000/ V.

"g o

2kQ

PROBLEM 10.23

10.24 Find the forced response ;'.

[Suggestion: Note that I = V/Z fails to work since
Z(;'l) = 0. Solve the describing equation by the method of Sec. 9.6]

I H

2 cos

V
t
© ^ I F

PROBLEM 10.24
Chap. 10 Sinusoidal Excitation and Phasors 279

10.25 Find the complete response / if/ (0) — 2 A and ?(0) = 6 V.

4 n h
_j_^
—\W
i

n5W —
20 cos
V
3 t

© 1/3 F

PROBLEM 10.25

10.26 Determine ;'(0) and r(0) in Prob. 10.25 so that the natural component vanishes
and i is simply the forced component.
 11

AC STEADY-STATE
ANALYSIS

In the previous chapter we have seen that in the case of circuits with sinusoidal inputs
we may obtain the ac steady-state response by analyzing the corresponding phasor
circuits. The circuits encountered were relatively simple ones which could be analyzed
by the use of voltage-current relations and current and voltage division rules.
It should be clear, because of the close kinship between phasor circuits and resistive
circuits, that we may extend the methods of Chapter 10 to more general circuits using
nodal analysis, loop analysis, Thevenin's and Norton's theorems, superposition, etc.

In this chapter we shall formally consider these more general analysis procedures,
limiting ourselves to obtaining the forced, or ac steady-state, response.

11.1 NODAL ANALYSIS

As we have seen, the voltage-current relation

V= ZI

for passive elements is identical in form to Ohm's law, and KVL and KCL hold in

phasor circuits exactly as they did in resistive circuits. Therefore the only difference
in analyzing phasor circuits and resistive circuits is that the excitations and responses
are complex quantities in the former case and real quantities in the latter case. Thus
we may analyze phasor circuits in exactly the same manner in which we analyzed
resistive circuits. Specifically, nodal and mesh, or loop, analysis methods apply. We
shall illustrate nodal analysis in this section and loop analysis in the following section.
To illustrate the nodal method, let us find the ac steady-state voltages v x
and v 2 of
Fig. 11.1. First we shall obtain the phasor by replacing the element values by
circuit
their impedances for co = 2 rad/s and the sources and node voltages by their phasors.

280
Chap. 11 AC Steady-State Analysis 281

5 cos 2t f+\ cos 2t

FIGURE 11.1 Circuit to be analyzed by the phasor method

This results in the circuit of Fig. 1 1.2(a). Since we are interested in finding V, and V2 ,

the node voltage phasors, we may replace the two sets of parallel impedances by their
equivalent impedances, resulting in the simpler, equivalent circuit of Fig. 1 1.2(b).

(a

2 ° V |
v2

•jin

+ J2

fU
I

5 :&) a 5/0.
Q)

(b)

FIGURE 11.2 Two versions of the phasor circuit corresponding to

Fig. 7 77

e nodal equations, from Fig. 1 1.2(b), are

V V —V
2(V, - 5/0) + -±j- + l

_j{
2
= °
{

v -v,
2 ,

+ J2)/5 1
282 AC Steady-State Analysis Chap. 11

which in simplified form are

(2+y2)V -yiv 2 1
= io

-/lV 1 + (l-./l)V2 = 5
Solving these equations by determinants, we have

10 -jl

V:
5 1 - j\ 10-75
= 2-;iv
2 + j2 -jl
-jl 1-/1
2+j2 10

-Jl 5 10 +j20 = 2+./4V

5 5

In polar form these quantities are

V, = yT/-26.6°V
2VT /63.4° V
Therefore the time-domain solutions are

T cos
- 26.6°) V
(2?

2,/y cos (2? + 63.4°) V

As an example involving a dependent source, let us consider Fig. 1.3, in which 1

it is Taking the ground node as shown, we have

required to find the forced response /'.

the two unknown node voltages v and v + 3000/, as indicated. The phasor circuit in
its simplest form is shown in Fig. 11.4, from which we may observe that only one

nodal equation is needed. Writing KCL at the generalized node, shown dashed, we
have
30001
+ =
^(io 3 )Tf(i-;2)(i03) '

(2-./i)(io 3 )

v+30001
AAAr
3000i 2k.fl
V
4 cos 5000 1 f* J 2kn *»F*
-i" F
T
FIGURE 11.3 Circuit containing a dependent source
 V

Chap. 11 AC Steady-State Analysis 283

U 2
vw
k
° /'
'
V V+3000I\
V
\ 3000 I

«zQ (l-j2)kA (2- jl)kn

FIG U R E 1 1 .4 Phasor circuit of Fig. 1 1.3

Also, from the phasor circuit we have

I
4-
A(IO')

Eliminating V between these two equations and solving for I, we have

1 = 24 x 10- 3 /53.1° A
= 24 /53.1° mA
Therefore in the time domain, we have

i = 24 cos (5000/ + 53.1°) mA

As a final example illustrating nodal analysis, let us find the forced response v in
Fig. 11.5 if

vg = Vm cos oil V

We note first that the op amp and the two 2-kfi resistors constitute a VCVS with

IpF

Vg O A^A

FIGURE 11.5 Circuit containing an op amp

284 AC Steady-State Analysis Chap. 11

gain 1 + |££|j = 2 (see Sec. 3.4). Therefore v = 2v 2 or v 2

,
= v[2, as indicated in the
phasor circuit of Fig. 1 1.6.

Writing nodal equations at the nodes labeled V, and V/2, we have

V, - Vm /0 V, - (V/2) ,
V. -V _ n
^
,

(1/VTX10V yT(10 3
) -yl0 6 /co
(V/2) V/2
+ =
2(10 3 )
!

-j\0 6 /co

Eliminating \ x
and solving for V results in

2Vm
[1 -((D 2 /\0 6 )]+j(^f2eo/l0 3 )
In polar form this is

y_ 2Kw /fl
(11.1)
VI + (co/ 1000) 4

where

2~co/1000
9 = -tan" 1

- 2
(11.2)
.1 (co/1000) J

In the time domain we have

= 2K.
v cos (cor + 0) (11.3)
VI + (co/ 1000)'

We might note in this example that for low frequencies, say < co < 1000, the
amplitude of the output voltage v is relatively large, and for higher frequencies, its

amplitude is relatively small. Thus the circuit of Fig. \. 5 filters out higher frequencies 1

and allows lower frequencies to "pass." Such a circuit is called a filter and will be
considered in more detail in Chapter 15.

-
j iooo/ go ka

I A/2 k A
Vm /Oo —AAA

-j lOOO/co kfl

FIGURE 11.6 Phasor circuit of Fig. 11.5

Chap. 11 AC Steady-State Analysis 285

EXERCISES
11.1.1 Find the forced response v using nodal analysis. Ans. 2 sin 2t V

in

8 cos 2t
A © 4sm 2t

EXERCISE 11.1.1

11.1.2 Find the forced response v. Ans. 4.8 cos (2/ - 53.1°)

3cos2t V

/"T\4sm2t

EXERCISE 11.1.2

11.1.3 Find the amplitude of v in (11.3) if Vm = 10 V for (a) CO = 0, (b) CO = 1000

rad/s, (c) co = 10,000 rad/s, and (d) co = 100,000 rad/s.
Ans. (a) 20, (b) 14.14, (c) 0.2, (d) 0.002 V

11.2 MESH ANALYSIS

To illustrate mesh analysis of an ac steady-state circuit, let us find v in Fig. 11.1, x

which was obtained, using nodal analysis, in the previous section. We shall use the
phasor circuit of Fig. 11.2(b), which is redrawn in Fig. 1.7, with mesh currents I, and
1
286 AC Steady-State Analysis Chap. 11

jl A
b[0_
*fO in
\a A

FIGURE11.7 Circuit of Fig. 11 .2 redrawn for mesh analysis

I,, as indicated. Evidently, the phasor voltage V! may be obtained from

I, = 2(5 - V,)

or

V, *-* (11.4)

The two mesh equations are

^1,-71(1,-1,) =5
(11.5)
-;l(I 2 -I )-;H + (i^)(I
1 2 2 5) =

Solving these equations for I,, we have

I, =6 +jl A
which substituted into (11.4) yields

v = 2-yi v
1

This is the same result that was obtained in the previous section and may be used to
obtain the time-domain voltage v x
.

The same shortcut procedures for writing loop and nodal equations, discussed
in Sees. 4.1 and 4.5 for resistive circuits, apply to phasor circuits. For example, in
Fig. 11.7 if I = —5 is the mesh current in the right mesh in the clockwise direction,
3

the two mesh equations are written down by inspection as

(i-yi)i,-(-7i)i 2 = 5

_(_„„, + (_,, -;, + l+i?), 2 - (i+i?) I = 3

These are equivalent to (1 1.5) and are formed as in the resistive circuit case. That is,

in the first equation the coefficient of the first variable is the sum of the impedances
around the first mesh. The other coefficients are the negatives of the impedances
Chap. 11 AC Steady-State Analysis 287

common mesh and the meshes whose numbers correspond to the currents.
to the first
The right member the sum of the voltage sources in the mesh with polarities consis-
is

tent with the direction of the mesh current. Replacing "first" by "second" applies to
the next equation, and so forth. The dual development, as described in Sec. 4.1, holds
for nodal equations.
As a final example, let us consider the circuit of Fig. 1 1 .8(a), where the response is

the steady-state value of ?•,. The phasor circuit is shown in Fig. 1 1.8(b), with the loop
currents as indicated.

sin 2t A

4/0 + j2a

(b)

FIGURE 11.8 (a) Time-domain circuit; (b) phasor counterpart

Applying KVL around the loop labeled I, we have

-V, -M-jl + I) + (l +j2)(l + 2V,) =

Also, from the figure we see that

V,=./l(4-I)
288 AC Steady-State Analysis Chap. 11

Eliminating I from these equations and solving for V,, we have

~ 4+
V, = -/3
= 1 /143.1° V

Therefore, in the time domain, the voltage is

«, = cos (2/+ 143.1°) V

EXERCISES
11.2.1 Find the forced response i in Fig. 11.3 using mesh analysis.

11.2.2 Solve Ex. 11.1.1 using loop analysis.

11.2.3 Find the steady-state current i using loop analysis. Ans. cos (t 36.9°) A

en

10 cos +\
t
(

EXERCISE 11.2.3

11.3 NETWORK THEOREMS

Because the phasor circuits are exactly like the resistive circuits except for the nature
of the currents, voltages, and impedances, all the network theorems discussed in
Chapter 5 for resistive circuits apply to phasor circuits. In this section we shall illustrate
superposition, Thevenin's theorem, Norton's theorem, and the proportionality prin-
ciple, as applied to phasor circuits.
In the case of superposition, if a phasor circuit has two or more inputs, we may
find the phasor currents or voltages due to each input acting alone (i.e., with the others
dead) and add the individual corresponding time-domain responses to obtain the total.

In the case of a circuit like Fig. 11.1, we may solve the corresponding phasor circuit
of Fig. 1 1.2 by mesh or nodal analysis or by superposition because both sources are
operating at the same frequency, namely to = 2 rad/s. If the sources have different

frequencies, we must use superposition, because the definition of Z(jco) allows us to

use only one frequency at a time, and thus we cannot even construct a phasor circuit.
Chap. 11 AC Steady-State Analysis 289

To illustrate superposition, let us find the forced response i in fig. I 1. 9. There are
two sources, one an ac source with co 2 rad/s and one a dc source with o) 0.

Therefore

/ /,

FIGURE 11. 9 Circuit with an ac and a dc source

where /, due to the voltage source acting alone and

is is due to the current source
/',

acting alone. Using phasors, we may find and i 2 by finding their respective phasor
/',

representations I, and I 2 which are the phasor currents shown in Figs.

,
1. 0(a) and 1 1

(b). respectively. Figure II. 10(a) is a phasor circuit representing the time-domain

circuit with the current source killed and <x> = 2 rad/s. Figure 1.10(b) is also a phasor 1

circuit, with the voltage source killed and co = 0.

3+j2n 2a

-jin in <

D
—
*V <
o

(a )

4Z0 A

(b)

FIGURE11.10 Phasor circuits representing Fig. 11.9

290 AC Steady-State Analysis Chap. 11

In the latter case, co = 0, the inductances are short circuits (Z L = jO = 0) and the
capacitance is an open circuit (Z r = \/jcoC, which becomes infinite as w— 0). Also,
since the current source is

i
g
= 4 cos (0/ + 0) A
its phasor representation is

I, = 4/01 A

In short, this is just the dc case considered earlier, as is evident from Fig. 11.10(b).
From Fig. 11.10(a) we have
5/0°
I,
3 +J2 + [(1 +y2)(-;l)/(l + J2 - jl)]
T/-8.1
from which

/, = /I cos (2? - 8.1°) A

From Fig. 11.10(b) we have, by current division,

*--(rb)« = -'£ A
from which, since co = 0,

/2 = - A 1

Therefore the total forced response is

* =h +H
= ^Tcos(2/ - 8.1°) - 1 A

For an example of a circuit with two sinusoidal sources with nonzero frequencies,
the reader is referred to Ex. 11.3.1. The procedure is, of course, exactly the same as
in the preceding example.
In the case of Thevenin's and Norton's theorems the procedure is identical to that
for resistive circuits. The only change is that v oc and iK the time-domain open-circuit ,

voltage and short-circuit current, are replaced by their phasor representations, VK

and and R th the Thevenin resistance, is replaced by Z th the Thevenin impedance
I sc , , ,

(of the dead circuit). There must be only a single frequency present, of course; other-
wise we must use superposition to break the problem up into single-frequency prob-
lems, in each of which Thevenin's or Norton's theorem may be applied.
In general, the Thevenin and Norton equivalent circuits in the frequency domain
are as shown in Fig. 11.11. There is, of course, a close similarity with the resistive case.
As an example, let us use Thevenin's theorem to find the forced response v of Fig.
11.12. We shall find its phasor representation V using the Thevenin equivalent of the
phasor circuit to the left of terminals ci-b. The open-circuit phasor voltage VK is found
from Fig. 1 1.13(a), and the short-circuit phasor current is found from Fig. 11.13(b).
 —

Chap. 11 AC Steady -State Analysis 291

6 <b
7 th

(a) (b)

FIGURE11.11 (a) Thevenin and (b) Norton equivalent phasor circuits

If

a
* o-

+ 2v, A
2 cos

A
3
•0 iF±,

FIGURE 11.12 Circuit with a dependent current source

Then the Thevenin impedance, as for resistive circuits, is

V PC
T
th — ~j (1 1.6)

In Fig. 11.1 3(a), since terminals a-b are open, the current 2/0° flows in the resistor
and the current 2\ flows
x
in the capacitor. Therefore by KVL we have

V« = V, - (-;1)(2V 1 )

where

V, = 2(1) = 2 V
Thus we have
V = 0C 2+ 7-4V
In Fig. 11.13(b), the two nodal equations needed are

V, V,
2V, = 2
1 '
-j\
and
lK = -V, + 2
292 AC Steady-State Analysis Chap. 11

-jio
11

a
/T\
+ A

K
2V,

^ in <

< •
o

(b)

FIGURE 11.13 Phasor circuits for use in Thevenin's theorem

From these we find

I sc = 3+7lA
The Thevenin impedance, by (11.6), is therefore

The Thevenin equivalent circuit is shown in Fig. 11.14, with the ^-F capacitor, corre-
sponding to —j\ Q, connected to terminals a-b.
It is a simple matter now to see, by voltage division, that

-n (2 + /4)
ui +;i) + (-7ij

4-/2
2 /V/T/-26.6°V
Chap. 11 AC Steady-State Analysis 293

i+ji n

2 + j4

V
6 - j m -p v

FIGURE 11.14 Thevenin equivalent phasor circuit of Fig. 11.12

Therefore, in the time-domain, we have

v = 2J1>~ cos (3* - 26.6 ) V

As a final topic in this section, let us consider the ladder network of Fig. 11.15
and use the proportionality principle to obtain the steady-state response V. This
requires, as in the resistive case, that we assume V
to have some convenient value like
V= 1 and work backward to find the corresponding \ g Then the correct value of .

V is found by multiplying the assumed value by an appropriate constant.

Let us begin by assuming

V= 1 V
Then from the circuit we have
v y_ 1 + JIA
Continuing, we have

V, =yll, +V=7l(l +j\)+ 1 =jl\

I2 --Xl- I +(l+yl)=ylA
~ -jl '
'
1

\g = ii 2 + v,=yi +;i=/2v
Therefore V= 1 is the response to \ g =j2. If we multiply this value of \g by 6/J2
to get the correct value of \g , then by the proportionality principle, we must multiply

FIG U R E 1 1 .1 5 Phasor ladder network

294 AC Steady-State Analysis Chap. 11

the assumed response of 1 by the same factor, 6//2, to get the correct value of V.
Therefore we have

V=(£)(1)=-/3V

EXERCISES
11.3.1 Find the forced response /'.

Ans. 2.88 cos (It + 19.4°) + 1.8 cos (3^ + 111.9°)

MM6cos(3t-l5°)

EXERCISE 11.3.1

11.3.2 Solve for v in Fig. 11.12 using Norton's theorem.

11.3.3 Find V„ I„ and I 2 in Fig. 11.15. Ans. 3 V, 3 - ;3 A, 3 A

11.4 PHASOR DIAGRAMS

Since phasors are complex numbers, they may be represented by vectors in a plane,
where operations, such as addition of phasors, may be carried out geometrically.
Such a sketch is called a phasor diagram and may be quite helpful in analyzing ac
steady-state circuits.
To illustrate, let us consider the phasor circuit of Fig. 11.16, for which we shall
draw all the voltages and currents on a phasor diagram. To begin with, let us observe

R L

Vo - v,

v
°© Vr -T- c

FIGURE11.16 RLC series phasor circuit

Chap. 11 AC Steady-State Analysis 295

that the current I is common to all elements and take it as our reference phasor,
denoting it by

I I|/0°

We have taken the angle of I arbitrarily to be zero, since we want I to be our reference.
We may always adjust this assumed value to the true value by the proportionality
principle discussed in the previous section.
The voltage phasors of the circuit are

V R = R\- R\\\
V L = jcoLl = coL\I |/9CT

and
V, = VR + \L + \ c

These are shown in the phasor diagram of Fig. 11.17(a), where it is assumed that
|VJ > |V C |. The cases, |V L |
< |V C and |VJ
|
= |VC |, are shown in Figs. 11.17(b)
and (c), respectively. In all cases the lengths representing the units of current and
voltage are not necessarily the same, so that for clarity I is shown longer than V*.
In case (a) the net reactance is inductive, and the current lags the source voltage
by the angle 6 that can be measured. In (b) the circuit has a net capacitive reactance,
and the current leads the voltage. Finally, in (c) the current and voltage are in phase,
since the inductive and capacitive reactance components exactly cancel each other.
These conclusions follow also from the equation

(11.7)
Z R j[coL - (1/wC)]

\*
vR -v
g

v
c V

Vet Vc t

(o) (b) (c)

F I GURE 1 1 .1 7 Phasor diagrams for Fig. 11.16

296 AC Steady-State Analysis Chap. 11

Case (c) is characterized by

coL =
coC
or

co = (ll.8)
.s/LC

If the current in Fig. 1 1. 16 is fixed, then the real component of the voltage V^ is

fixed, since it is R\ 1 1. In this case the locus of the phasor \g (its on

possible location
the phasor diagram) is the dashed line of Fig. 11.18. The voltage phasor varies up and
down this line as co varies between zero and infinity. The minimum amplitude of the
voltage occurs when co = Xj^LC, as seen from the figure. For any other frequency,
a larger amplitude of voltage is required for the same current.
As a last example illustrating the use of phasor diagrams, let us find the locus of
I as R varies in Fig. 11.19. The current is given by

1
l„ Vm (R-jcoL)
R + jcoL R 2 + co L 2 2

Therefore if

I = X .+ jy (11.9)

we have
RV,
x = Re I (11.10)
R2 I co
2
L 2

— m ~coLVm (11.11)
y I I
R2 co
2
L2

The equation of the locus is the equation satisfied by x and y as R varies ; thus we need
to eliminate R between these last two equations.

FIGURE 11.18 Locus of the voltage FIGURE 11.19 RL phasor circuit

phasor for a fixed current

 v

Chap. 11 AC Steady-State Analysis 297

[f we divide the first of these two equations by the second, we have

\ _R_
y coL

from which
coL.x
R

Substituting this value of R into (1 1.1 1) we have, after some simplification.

x2 +y = - coL
2

This result may be rewritten as

: !
vm y= ( m \ (11.12)
(;•
2coL} \2coL)

which is the equation of a circle with center at [0, — {VJ2coLj\ and radius VjloiL.
The circle (11.12) appears to be the locus, as R phasor I = x
varies, of the jy.

However, by (11.10), x >

0; thus the locus is actually the semicircle shown dashed
on the phasor diagram of Fig. 11.20. The voltage VJO, taken as reference, is also
shown, along with the phasor I. If R = 0, we have, from (11.10) and (11.11), x =
and v -VJcoL. If ,/? co, then .v
> * —
and v » 0. Thus as R varies from to oo,
the current phasor moves counterclockwise along the circle.

R=
FIGURE 11.20 Locus of the phasor I
298 AC Steady-State Analysis Chap. 11

If Iis as shown in Fig. .20, the current phasor may be resolved into two compo-
1 1

nents, one having amplitude I m cos 6 in phase with the voltage and one with amplitude
l m sin 6, which is 90 out of phase with the voltage. This construction is indicated by

the dashed and dotted vertical line. As we shall see in the next chapter, the in-phase
component of the current is important in calculating the average power delivered by
the source. Thus the phasor diagram gives us a method of seeing at a glance the maxi-
mum in-phase component of current. Evidently this occurs at point a, which corre-
sponds to 6 = 45°. This is the case x = —y, or R = coL.

EXERCISES
11.4.1 Eliminate coL in (11.10) and (11.11) and show that, as coL varies, the locus of
the phasor I = x + j'y is a semicircle.
K.V +
Ans.(,-^y y> = (^)\y<0
2R>

11.4.2 Find coL in Ex. 1.1.4.1 so that Iml has its largest negative value. Also find I for
this case. Ans. R, (VJ^/2 R)l -45°

PROBLEMS
11.1 Solve for the steady-state response / in Fig. 10.19(a) using nodal analysis.

11.2 Solve for the steady-state response i in Fig. 10.20(a) using nodal analysis.

11.3 Solve Prob. 10.18 using nodal analysis.

11.4 Solve Prob. 10.19 using nodal analysis.

11.5 Find the steady-state value of v if v g = Vm cos cot and find the amplitude of v
for the cases CO= (a) 0, (b) 1000 and (c) 10 6 rad/s. Compare the results for this
circuit with those for Fig. 11.6.

PROBLEM 11.5
Chap. 11 AC Steady-State Analysis 299

11.6 Repeat Prob. 11.1 using mesh analysis.

11.7 Repeat Prob. 11.3 using mesh analysis.

11.8 Find the forced response v.

—^w^ 2H ia
^vVv-

—
1

nm^-
IH in I H

4 cos
*6
V
I F ^ v

2F^~ in

PROBLEM 11.8

11.9 Show that if ZjZ 4 = Z 2 Z 3 in the "bridge" circuit shown, then I = V =

and therefore all the other currents and voltages remain unchanged for any
value of Z 5 Thus it may be replaced by an open circuit, a short circuit, etc.
.

In this case, the circuit is said to be a balanced bridge.

o —— 1*

Z 1
z 3

I
m
z5
+ V
Z2 z4

o -4 l —

PROBLEM 11.9

11.10 Show that the circuit of Prob. 1 1 .8 is a balanced bridge, with the series combina-
tion of 1 Q.and H constituting Z 5
1 in Prob. 1 1 .9. Replace Z 5 by a short, and
show that the same v is obtained as before.

11.11 Find the steady-state response

 1

300 AC Steady-State Analysis Chap. 11

2v, A

PROBLEM 11.11

11.12 Find the steady-state response /'.

6 cos 3000 1 V? ±

PROBLEM 11.12

11.13 Find the steady-state response v if i gX = 6 cos It A and i gl =2 cos 2? A.

IF 2a
-A/W "
+

4H 4n > v

•g
2
CD r
2D 8

PROBLEM 11.13

11.14 Find the steady-state response v.

Chap. 11 AC Steady-State Analysis 301

4n

e
4 cos 2t V
IH

6 cos 2
V
•6 Ad
2a
2F -

PROBLEM 11.14

11.15 Find the steady-state response r if r K 2 cos 2000/ V.

"go

PROBLEM 11.15

11.16 Find the steady-state response ;• if r

g
= 5 cos / V.

'g »

, o v

PROBLEM 11.16

11.17 Find the forced response u if v. = 2 cos t V.

302 AC Steady-State Analysis Chap. 11

•g o o v

PROBLEM 11.17

11.18 Find the forced response r if (a) v g = 2 cos 2/ V and (b) v g = 2 cos / V.

"go

PROBLEM 11.18

11.19 Find the forced response i if r

g
= 4 cos lOOOf V.
Chap. 11 AC Steady-State Analysis 303

PROBLEM 11.19

11.20 Find the steady-state value of r if

vg =9 \ 10cos(f-30°) I 18 cos (3/ + 15°) V

8a 2H
>A/V 'TJoTT 1

PROBLEM 11.20

11.21 Find the steady-state value of v.

PROBLEM 11.21

11.22 Find the steady-state response v.

304 AC Steady-State Analysis Chap. 11

6 cos (3t + 30°)

PROBLEM 11.22

11.23 For the phasor circuit corresponding to Prob. 11.12, replace the part to the
left of terminals a-b by its Thevenin equivalent and find the steady-state value
of/.

11.24 Replace the phasor circuit of the figure for Prob. 11.13, except for the 4-fl
resistor, by its Thevenin equivalent, and find the forced response v.

11.25 Replace the circuit to the left of terminals a-b by its Thevenin equivalent, and
find V.

j3A
-nsw

^-j2a

PROBLEM 11.25

1 1.26 Use the principle of proportionality on the corresponding phasor circuit to find
the steady-state value of v.

PROBLEM 11.26

11.27 Solve Prob. 11.16 by the proportionality principle.

11.28 Solve Prob. 11.19 by the proportionality principle.

Chap. 11 AC Steady-State Analysis 305

11.29 Kind 1. the phasor representation of i, using a phasor diagram. Show the
phasors of iR .
c and
f , /,. with the phasor of the source voltage as reference.

PROBLEM 11.29

11.30 Show that the locus of c in Fig. 11.16 is a circle if R = 2 Q, C = { F,

V
Vg = 1/CT V, (O = 2 rad/s, and L varies from to oo. Select the value of L that
gives the maximum amplitude of the time-domain voltage vc and find r c
, in

this case.

11.31 Kind the locus of Vc

in Prob. 1.30 if L = H, C = \
1 1 F, \g = 2/tT V, w = 2
rad/s, and /? from to oo. Select the value of R
varies for which 1m Vr has its

largest negative value, and find vc for this case.

 12
AC STEADY-STATE POWER

In this chapter we shall consider power relationships for networks which are excited
by periodic currents and voltages. We shall concern ourselves primarily with sinu-
soidal currents and voltages since nearly all electrical power is generated in this form.
Instantaneous power, as we now well know, is the rate at which energy
is absorbed

by an element, and it varies as a function of time. The instantaneous power is an

important quantity in engineering applications because its maximum value must be
limited for all physical devices. For this reason, the maximum instantaneous power,
or peak power, is a commonly used specification for characterizing electrical devices.
In an electronic amplifier, for instance, if the specified peak power at the input is

exceeded, the output signal will be distorted. Greatly exceeding this input rating may
even permanently damage the amplifier.
A
more important measure of power, particularly for periodic currents and volt-
ages, that of average power. The average power is equal to the average rate at which
is

energy is absorbed by an element, and it is independent of time. This power, for

example, is what is monitored by the electric company in determining monthly
electricity bills. Average powers are encountered which range from a few picowatts,
in applications such assatellite communications, to millions of watts, in applications

such as supplying the electrical needs of a large city.

Our discussion will begin with a study of average power. We shall then consider
superposition once again and introduce a mathematical measure for characterizing
periodic currents or voltages, known as effective or rms values. We shall then consider
the power factor associated with a load and present a complex power. Finally, we shall
describe the measurement of power.

306
Chap. 12 AC Steady-State Power 307

12.1 AVERAGE POWER

In linear networks which have inputs that are periodic functions of time, the steady-
state currents and voltages produced are periodic, each having identical periods.
Consider an instantaneous power

p = vi (12. 1)

where v and /'

are periodic of period T. That is, v(t + T) = v(t), and /(/ + T) = i{t).

In this case,

p{t + T) = v(t + T)i(t + T)

= v(t)i(t)
= p(t) (12.2)

Therefore the instantaneous power is also periodic of period T. That is, p repeats
itself every T seconds.
The fundamental period T of p x
(theminimum time in which p repeats itself) is not
1

necessarily equal to T, however; but J must contain an integral number of periods T v

In other words,

T=nT x
(12.3)

where n is a positive integer.

As an example, suppose that a resistor R carries a current = i Im cos at with period
T= 2n/co. Then
p = Ri 1

= RIl, cos
2
cot

= ^(1 + cos2g>0

Evidently, T, = n/co, and, therefore, T — 27V Thus, for this case, n = 2 in (12.3).
This is illustrated by the graph of/? and i shown in Fig. 12.1(a).
If we now take i — IJA -f cos cot), then

p = RPm {\ + cos cot)

2

In this case, 7\ = T= Injco, and n = 1 in (12.3). This may be seen also from the
graph of the function in Fig. 12.1(b).
Mathematically, the average value of a periodic function is defined as the time
integral of the function over a complete period, divided by the period. Therefore the
average power P for a periodic instantaneous power/? is given by

t, + T,

P=
Y pdtW (12>4)
\

where r, is arbitrary.
 i

308 AC Steady-State Power Chap. 12

i(t),p(t)

(a)

T, --T

i(t),p(t)

(b)

FIGURE 12.1 Graphs of p and

A periodic instantaneous power p is shown in Fig. 12.2. It is clear that if we

integrate over an integral number of periods, say ml, (where m is a positive integer),
then the total area is simply m times that of the integral in (12.4). Thus we may write

1
p dt (12.5)
ffl7",

If we select m such that T = mT l

(the period of v or /), then

P= p dt (12.6)

Therefore we may obtain the average power by integrating over the period off or /.
Chap. 12 AC Steady -State Power 309

FIGURE 12.2 Periodic instantaneous power

Let us now
consider several examples of the average power associated with sinu-
soidal currentsand voltages. A number of very important integrals which often occur
are tabulated in Table 2.1. Verification of these integrals is left as an exercise (see
1

Ex. 12.1.1).

TABLE 1 2.1 integrals for Sinusoidal Functions and Their Products

f 2 jt/o
f(0 f(t)dt, co^O
Jo

1. sin (cot + a), cos (cat + a)

2. sin (ncot + a), cos (ncot + a)*
3. sin 2 (cot + a), cos 2 (cot + a) n/co
4. sin (mcot + a)cos (ncot + a)*
5. cos (mcot + a)cos (ncot + /?)* 0,w ?t /;

r. cos (a — P)jco, in = n

* in and n are integers.

First, let us consider the general two-terminal device of Fig. 1 2.3, which is assumed
to be in ac steady state. If, in the frequency domain,

Z = |Z|/0
is the input impedance of the device, then for

v = V,„ cos(w/ - 0) (12.7)

two-
V terminal
device

FIGURE 12.3 General two-ter-

minal device
310 AC Steady -State Power Chap. 12

we have
/ = Im cos (cot + - 0) (12.8)

where

|Z|

The average power delivered to the device, taking t {

= in (12.6), is

y , fin >co
-^ cos (cot + 0) cos (co + <f)
-9)dt

Referring to Table 12.1, entry 5, we find, for m= n = 1, a = 0, and /? = — 0,

p = KJnlcos (12 .9)

Thus the average power absorbed by a two-terminal device is determined by the

amplitudes Vm and lm and the angle by which the voltage v leads the current /.

In terms of the phasors of v and i,

\= VJ± = \\\I±
1 = I J<f> - = \l\ /<f,
-
we have, from (12.9),

P = ^|V||I|cos(angV - angl) (12.10)

where
ang V = 0, ang 1 = —
are the angles of the phasors V and I.

If the two-terminal device is a resistor R, then = and Vm = RIm , so that (12.9)

becomes

Pr = \RJl

It is worth noting at this point that if / = /dc a constant (dc) current, then
, co = =
= 0, and l m = /dc in (12.8). In this special case, we have, by (12.6),

Pr = RIl

In the case of an inductor, = 90°, and in the case of a capacitor, = —90°,

and thus by (12.9), P = 0. Therefore an inductor or a capacitor, or
for either one,
for that matter any network composed entirely of ideal inductors and capacitors, in
any combination, dissipates zero average power. For this reason, ideal inductors and
capacitors are sometimes called lossless elements. Physically, lossless elements store
 1

Chap. 12 AC Steady-State Power 311

energy during part of the period and release it during the other part, so that the aver-
age delivered power is zero.
A very useful alternative form of ( 2.9) may be obtained by recalling that

Z = ReZ +./ ImZ = \Z\/9_

and therefore

cos 9
ReZ

Substituting this value into (12. 9) and noting that Vm = |Z|/m we have ,

P=£/'ReZ (12.11)

Let us now consider this result if the device is a passive load. We know from the
definition of passivity in (1.7) that the net energy delivered to a passive load is non-

negative. Since the average power is the average rate at which energy is delivered to
a load, it follows that the average power is nonnegative. That is, P > 0. This requires,
by (12.11), that

Re Z{jco) >
or, equivalently,

If 9 = 0. the device is equivalent to a resistance, and if9 = n/2 (or —n/2), the
device is equivalent to an inductance (or capacitance). For — n/2 < 9 < 0, the device
is equivalent to an RC combination, whereas for < 9 < n/2, it is equivalent to an
RL combination.
Finally, if |0| > n/2, then P < 0, and the device is active rather than passive.
In this case the device is delivering power from its terminals and, of course, acts like
a source.
As an example, let us find the power delivered to each element of Fig. 12.4. The
impedance across the source is

Z = 100+./100
= lOOv^T /45° Q

ioon

ioo cos loot

v

FIGURE 12.4 RL circuit in the ac steady state

312 AC Steady-State Power Chap. 12

The maximum current is

^
lm
izi ~^T
Therefore, from (12.9), the power delivered to Z is

P= yyy cos 45° = 25 W

Alternatively, from (12.1 1),

P= -j(-7j) (100) = 25 W
The power absorbed by the 100-Q resistor is

P fi
= ^(ioo)(i /v^-)* = 25W
2

This power, of course, is equal to that delivered to Z since the inductor absorbs no
power.
The power absorbed by the source is

Pg = -£ffi£-cos0= -25 W
where the minus sign is used because the current is flowing out of the positive terminal
of the source. The source therefore is delivering 25 W to Z. We note that the power
flowing from the source is equal to that absorbed by the load, which illustrates the
principle of conservation of power.

EXERCISES
12.1.1 Verify the integrals of Table 12.1.

12.1.2 For a capacitor of C farads carrying a current / = Im cos cot, verify that the
average power is zero from (12.6). Repeat this for an inductor of L henrys.

12.1.3 Find the average power delivered to a 100-Q resistor carrying a current given
by (a) 10 cos 10/ mA, (b) a square wave for which one cycle consists of 10 mA
1 1

for 1 s followed by —10 mA for 1 s, and (c) a triangular wave for which one
cycle consists of a current that increases linearly from to 10 mA in 1 s.
Ans. (a) 5 mW, (b) 10 mW, (c) 3.33 mW
12.1.4 Find the average power absorbed by each resistor, the capacitor, and the
source. Ans. 20 mW, 10 mW, 0, -30 mW
Chap. 12 AC Steady-State Power 31 3

10 COS lOt I
mF
mA

EXERCISE 12.1.4

12.1.5 lf/i(0 is periodic of period 7", and/2 (0 is periodic of period T2 show that ,

fiU) + fiit) is periodic of period T if positive integers m and exist such that /;

T = mTi = nT2

Extend this result to the function (1 + cos O)?) 2 considered in this section, to
,

find its period, T= 27r/&).

12.2 SUPERPOSITION AND POWER

In this section we shall consider the power in networks containing two or more
sources, such as the simple circuit of Fig. 12.5. By superposition, we know that

i = U +h
where /, and /
:
are the currents in R due to v gi and vg2 , respectively. The instantaneous
power is

p = R{i + Y {
i2

= Ri\ + Rij -f 2Ri 1

i2

= Pi + p + 2Ri 2 1
i2

where /?! and p 2 are the instantaneous powers, respectively, due to v gl acting alone
and to vg2 acting alone. In general, 2RiJ 2 0; thus p ^
p + p 2 and superposition ^ x
,

does not apply for instantaneous power.

FIGURE12.5 Simple circuit with two sources

314 AC Steady-State Power Chap. 12

In the case of/? periodic with period T, the average power is

-iripdt
'T
1 f
j (Pi +^2 + 2Ri x
i
2 )dt

P 1
+P + 2R
^T 2 | hi a2
/,/ A
JO

where Pj and P2 are the average powers from v gl and v g2 , respectively, acting alone.
(We are assuming, of course, that each component of p is periodic of period T.)
Superposition for average power applies when

P= P X
+P 2 (12.12)

Clearly, this condition holds if

T
[
ij 2 dt = (12.13)
Jo

The most important case in which this equation is satisfied is when i(t) is composed
of sinusoidal components of different frequencies. Suppose, for instance, that

/, = Iml COS (cO.t + <f>i)

and

h = Imi cos (co 2 t + (f) 2)

Since we are assuming that / = i l

+ i2 is periodic of period T, we must have

/.i cos [co.it + T) + fa] + Im2 cos [a> a (* + T) + 2 ]

= /„, COS (CO,/ + 0,) + Im2 COS (© 2 / + fa)

which requires that

co.T^lnm, co 2 T= Inn

where w and « are positive integers. Therefore if co is a number such that T = 2n/co,
then co l = mco and co 2 = nco. In this case the integral in (12.13) becomes, using Table
12.1,

rT p2 n/oi
ij 2 dt = Im Jm2 cos (mcot + $.) cos (ncot -f $2) rfr
Jo Jo

_ 7 wl /m2 ^cos(0 1
- 2) .
m= n
CO

= 0, /w ^ «
Chap. 12 AC Steady-State Power 315

Thus if m n (co, = co 2 ), superposition does not apply. However, if m =£ n,

superposition does apply. We may generalize this result to the case of a periodic
sinusoid with any number of sinusoidal components of different frequencies. The
average power due to the sum of the components is the sum of the average powers due

to each component acting alone.

It can be shown that superposition of average power holds for sinusoids whose
frequencies are not integral multiples of some frequency co, provided we generalize
the definition of average power to

= r —]
P pdt
—
lim
T Jo

This generalization applies to the periodic case just considered as well as to the case
i = *! + i2 , where

f, = cos t

U = cos nt
In this case i is not even periodic (the ratio coJco 2 = \jn is not a rational number
mjn), but

lim —T dt =
T —oo
r«
| /,/,.

As an example, suppose in Fig. 12.5 that v gl = 100 cos (377/ + 60°) V, v g2 =

50 cos 377? V, and R = 100 ft. Since w, = co 2 , we cannot apply superposition to
power. However, we can use superposition to find the current and subsequently
find the power. The phasor currents due to the respective sources are

I, == A
1/60°

I2 = -0.5 A
Therefore

1 = 1, +i =yo.866A
2

so that I m = 0.866 A. From (12.11), we find

p= ^-(100)(0.866)
2
= 37.5 W
Let us now repeat the above example with v g2 = 50 V. Since v gl and vg2 are sinu-
soids with co l
= 377 and co 2 = rad/s, respectively, superposition for the average
power is applicable. Proceeding as before, we find

I, = 1/60^ for co = 377

I2 = -0.5 for co =
 .

316 AC Steady-State Power Chap. 12

where I ; is now a dc current. Therefore

P
tl W "'ml
2
= 50

Pi = RIli = 25 W
and the average power is

p = p + p2 = i
75 W
This example illustrates a very important case for electronic amplifiers with
sinusoidal inputs. These amplifiers contain dcpower supplies that produce dc currents
which provide the energy for the amplified ac signals. Thus superposition is very
useful in finding the average power associated with each frequency, including co = 0.
Extending the above procedure to a periodic current which is the sum of N + 1
sinusoids of different frequencies,

i = /dc + Imi cos (av + 0i) + Li cos (*»»* + 2)

+ . . - + ImN cos {<o N t + <f> N) (12.14)

we find the average power delivered to a resistance R is

P = RIl + !(/*, + Pm2 + . . + HN ) (12.15)

The first term, RIl,,, in which the factor £ is missing, is the special case of zero
frequency and must be considered separately, as in Sec. 12.1. That is, by (12.6),

Y RhJmi cos {co + (

t 0,) dt = 0, / = 1, 2, 3, . .
.
, N
J

and
T
z
RI dc dt

EXERCISES
12.2.1 Find the average power delivered to the resistor in Fig. 12.5 if R= 10 Q. and
(a) v = 10 cos 100/ and v g2 = 20 cos (100/ + 30°) V.
>

(b) V tl
<) = 100 cos (/ + 60°) and v gl = 50 cos (2/ - 30°) V.
(c) Vy t i = 50 cos (/ + 45°) and v g2 = 100 sin (/ + 30°) V.
(d) V gl = 20 cos (/ + 25°) and v g2 = 30 sin (5/ - 50°) V.
Arts, (a) 7.68 W, (b) 625 W, (c) 754.4 W, (d) 65 W
 1

Chap. 12 AC Steady-State Power 31 7

12.2.2 Miul the average power absorbed by each resistor and each source.
Am. 3.5 W, 3.5 W, -5 W, -2 W
v^W ? >

1 kfl

100 V ~
T l>4F

1 1 *

EXERCISE 12.2.2

12.3 RMS VALUES

We have seen and voltages deliver an

in the previous sections that periodic currents
average power to resistive loads. The amount of power that delivered depends on is

the characteristics of the particular waveform. A method of comparing the power

delivered by different waveforms is therefore very useful. One such method is the use
of rms or effective values for periodic currents or voltages.
The rms value of a periodic current (voltage) is a constant that is equal to the dc
current (voltage) which would deliver the same average power to a resistance R. Thus,
if / rms is the rms value of i, we may write

RIL Ri 1 dt

from which the rms current is

/rn
M" dt (12.16)

In a similar manner, it is easily shown that the rms voltage is

vrms = y.i fVA

The term rms is an abbreviation for root-mean-square. Inspecting (12.16), we see
that we are indeed taking the square root of the average, or mean, value of the square
of the current.
From our definition, the rms value of a constant (dc) is simply the constant itself.
The dc case is a special case (a> — 0) of the most important type of waveform, the
sinusoidal current or voltage.
Suppose we now consider a sinusoidal current / = Im cos (cot + <j>)- Then, from
318 AC Steady-State Power Chap. 12

(12.16) and Table 12.1, we find

^ J
COS
2
(cot + 0) (ft

Thus a sinusoidal current having an amplitude / m delivers the same average power
to a resistance R as does a dc current which isequal to IJ^J~1 We also see that the .

rms current is independent of the frequency co or the phase <j) of the current Similarly, /'.

in the case of a sinusoidal voltage, we find that

Substituting these values into the important power relations of (12.9) and (12.1 1) for
the two-terminal network, we have

P= ^ ms /rms cos0 (12.17)

and

P= I;ms Re Z (12.18)
\ SfcCs or>\*)
In practice, rms values are usually used in the fields of power generation and^
For instance, the nominal 15-V ac power which is commonly used for
distribution. 1

household appliances is an rms value. Thus the power supplied to our homes is
provided by a 60-Hz voltage having a maximum value of 115,^/T «= 163 V. On the
other hand, maximum values are more commonly used in electronics and com-
munications.
Finally, let us consider the rms value of the current in (12.14), which is made up
of sinusoids of different frequencies. In terms of rms currents, (12.15) becomes

P = R(ll + /?,„, + /*,„, + • • • + /JL.) (12.19)

Since P= 2
i?/ r ms , we see that the rms value of a sinusoidal current consisting of
different frequencies is

"rms — -V 'dc i
1 1 , m, + * 2,*, ~1~ • • \ * f/, m .

Similarly,

Vtm = V^lc + V\„, + nZ + . . . + nZ (12.20)

These results are particularly important in the study of noise in electrical networks,
a subject of later courses.
Chap. 12 AC Steady -State Power 319

EXERCISES
12.3.1 Find the rms value of a periodic current for which one period is defined by
(a) i /. 0< /</,
= 0, /, < <t (T=t z
t 2 ).

(b) i = It, 0<t<T=t 1

.

(c) i = I m sin cot, 0<t< njw

- 0, k/co < < 2n/co (T =
t 2n/co).

Am. (a) >/TJtll, (b) 7/,/,/T, (c) IJ2

12.3.2 Find the rms values of (a) / = 10 sin cot + 20 cos (cot + 30°), (b) i =
10 sin tot - 20 cos (2tot + 10°), and (c) / = 15 (1 + cos 377/).
Arts, (a) 12.25 A, (b) 15.81 A, (c) 18.37 A
12.3.3 Find / rms . Arts. 0.5 A

EXERCISE 12.3.3
"^

12.4 POWER FACTOR /

= V, ws T
The average power delivered to a load in the ac steady state, repeating (12.17), is

P = ^rms'rms COS d
^^
The power is thus equal to the product of the rms voltage, the rms current, and the
cosine of the angle between the voltage and current phasors. In practice, the rms
current and voltage are easily measured and their product, Krms /rms , is called the
apparent power. The apparent power is usually referred to in terms of its units, volt-
amperes (VA) or kilovoltamperes (kVA), in order to avoid confusing it with the unit
of average power, the watt. It is clear thai the average power can never be greater than
the apparent pow er.
The ratio of the average power to the apparent power is defined as the power
factor. Thus if we denote the power factor by pf, then in the sinusoidal case

Pf= V I
cos 9 (12.21)
320 AC Steady-State Power Chap. 12

which, of course, is dimensionless. The angle 0, in this case, is often referred to as

the pf angle. We also recognize it as the angle of the impedance Z of the load.
In the case of purely resistive loads, the voltage and current are in phase. Therefore,
= 0, pf =
and the average and apparent powers are equal. A unity power factor
1,

(pf = 1) exist for loads which contain inductors and capacitors if the reac-
can also
tances of these elements are such that they cancel one another. Adjusting the reactance
of loads so as to approximate this condition is very important in electrical power
systems, as we shall see shortly.
In a purely reactive load, = ±90°, pf — 0, and the average power is zero.
In this case, the equivalent load is an inductance (9 = 90°) or a capacitance (9 =
—90°), and the current and voltage differ in phase by 90°.
A load for which —90° < 9 < is equivalent to an RC combination, whereas
one having < 9 < 90° is an equivalent RL = cos — 9),
combination. Since cos 9 (

it is evident that the/?/ for an RC load having = — where < 0, < 90°, is equal 15

to that of an RL load with = P To avoid this difficulty in identifying such loads,

the pf is characterized as leading or lagging by the phase of the current with respect to
that of the voltage. Therefore an RC load has a leading pf and an RL load has a lagging
pf. For example, the series connection of a 100 Q resistor and a 0.1-H inductor at

60 Hz has Z = 100 +;37.7 = 106.9 /20.66° fl and has a pf of cos 20.66° = 0.936
lagging.
In practice, the power factor of a load is very important. In industrial applications,
for instance, loads may require thousands of watts to operate, and the power factor
greatly affects the electric bill. Suppose, for example, a mill consumes 100 kW from
a 220-V rms line. At a/?/ of 0.85 lagging, we see that the rms current into the mill is

P 105
534.8 A
rms
K ms pf (220)(0.85)

which means that the apparent power supplied is

Vrm Jrms = (220)(534.8) VA = 1 17.66 kVA

Now suppose that the pf by some means is increased to 0.95 lagging. Then

10 s
/
rms
= — 478 S A
(220)(0.95)

and the apparent power is reduced to

Krms / rms = 105.3 kVA

Comparing the latter case with the former, we see that 7 rms was reduced by 56.3 A
(10.5%). Therefore the generating station must generate a larger current in the case
of the lower/;/. Since the transmission lines supplying the power have resistance, the
generator must produce a larger average power to supply the 100 kW to the load.
Chap. 12 AC Steady-State Power 321

If the resistance is 0.1 Q. for instance, then the power generated by the source must be

/> =10 + 5
0.1/?n

Therefore we find

Pg = 128.6 kW, pf = 0.S5

= 122.9 kW, pf = 0.95

which requires that the power station produce 5.7 kW (4.64 "J more power to supply
the lower/'/' load. It is power companies encourage apf exceeding
for this reason that
say 0.9 and impose a penalty on large industrial users who do not comply.
Let us now consider a method of correcting the power factor of a load having an
impedance

Z = R + jX

We may alter the power factor by connecting an impedance Z, in parallel with Z, as

shown in Fig. 12.6. For this connection, it is clear that the load voltage does not
change. Since Z is fixed, I does not change, and the power delivered to the load is

not affected. The current I, supplied by the generator, however, does change.

z T — LJ z. z R + ,X

FIGURE 12.6 Circuit for correcting the power factor

Let us denote the impedance of the parallel combination by

— zz,
Mwp
z,

In general, we select Z so that (1) Zj absorbs zero average power, and (2) Z T satisfies
x

the desired power factor pf = PF. The first condition requires that Z, be purely
reactive. That is,

j*i

The second condition requires that

/Im Z
cos tan
,

PF
VRe Z T )]
322 AC Steady-State Power Chap. 12

Substituting Zr in terms of R, X, and X x

into this equation, we find that (see Prob.
12.14)

7?2 _L V2
v (12.22)
1

k iRtun {cos- l
PF)-X
where &i = 1 if the power factor is lagging upon correction and k^ =— 1 if the power
factor is leading upon correction. In each case, tan (cos -1 PF) is taken as positive.
As an example of the application of (12.22), let us change the power factor for the

circuit of Fig. 12.4 to 0.95 lagging. We have already found that

Z = 100+y'100= 141.4/45°

Therefore, before a parallel reactance is added across Z, the power factor is

pf = cos 9 = cos 45° = 0.707 lagging

Since we desire a power factor of 0.95 lagging, A', = 1 in (12.22), and

100 2 + 100 2 _ _ 2Q7 Q2 o

1
100 tan (cos" 1
0.95)- 100

Since X < x
0, the reactance is a capacitance C = —\/coX = l
33.6 ^F. The load
impedance now becomes

7 _ (100 +7l00)(-/297.92) _

Therefore the power to the corrected load is

P= cos(18 2 -
°
)
= 25W
2TTW)

which is the same as that delivered to Z in Fig. 12.4. The current, however, is

as compared to that of Fig. 12.4, given by

-0.5 A
2

We see, therefore, that the current has been reduced by 0.128 A or 25.6%.
 A

Chap. 12 AC Steady-State Power 323

EXERCISES
12.4.1 Find the apparent power for (a) a load that requires 30 Arms from
a 230-V mis line and (b) a load consisting of a 100-Q resistor in parallel with
a 25-//F capacitor connected to a 1 10-V rms 60-Hz source.
Ans. (a)6.9kVA, (b) 166.3 V
12.4.2 Find the power factor for (a) a load consisting of a series connection of a 100-Q
and a 20-//F capacitor operating at 60 Hz, (b) one that is capacitive
resistor
and requires 50 A rms and 5 kW at 10 V rms, and (c) one consisting of
1

a 5-kW load at a power factor of 0.85 leading, connected in parallel with

a 10-kW load at a power factor of 0.9 lagging.
Ans. (a) 0.602 leading, (b) 0.909 leading, (c) 0.993 lagging

12.4.3 A load consists of a 200-Q resistor in series with a 0.1-H inductor. Find the
parallel capacitance necessary to adjust to (a) a p/of 0.95 lagging and (b) a pf
of 0.95 leading if CO = lOOOrad/s. Ans. (a) 0.685 p.¥, (b) 3.31 //F

12.5 COMPLEX POWER

We shall now introduce a complex power in the ac steady state which is very useful
for determining and correcting power factors associated with interconnected loads.
Let us begin by defining rms phasors for general sinusoidal voltages and currents.
The phasor representations for (1 2.7) and (12.8) are

V = Vm e*
\ — i P m-e)

The rms phasors for these quantities are defined as

V = V
V rms =
* /~0~ rms e'* 1

(12.23)
I
*rms
= r~*r
— J/ rms c
ptt4>-&)

Let us now consider the average power given in (12.17). Using Euler's formula,
we may write

P = yrm J rm s
cos 6 - Re( VTm ,Irmi e»)

Next, inspecting (12.23), we see that

V I* = V 1 e'
e
324 AC Steady-State Power Chap. 12

*
where I r ms is the complex conjugate of Irms . Thus

P=Re(V rms I*ms ) (12.24)

and the product V rms I*ms is a complex power whose real part is the average power.
Denoting this complex power by S, we have

S = VrmsI *ms = i>+yg r (12.25)

where Q is the reactive power. Dimensionally, P and Q have the same units; however,
the unit of Q is defined as the var (voltampere reactive) to distinguish it from the watt.
The magnitude of the complex power is

I
SI
kJ
I
= IV
I
I*
T rms*rms I
I
= IV
I
T rras III*
I
J rras
I
I
I
= Vrms'rms
1 r

which, of course, is equal to the apparent power.

From (12.25), we see that

e= ImS= Krms /rms sin0 (12.26)

For an impedance Z, we know that sin 9 = (Im Z)/| Z |, so that

Q=y jrms
ImZ
\L ' rras 1 i
7 i

Therefore, since Frms /|Z| = /rmi , we see that

Q= 2
/ r ms Im Z

= ^tw (12.27)

A phasor diagram of V rms and I rms is shown in Fig. 12.7. We see that the phasor
current can be resolved into the two components I!ms cos and /rms sin 6. The com-
ponent 7rms cos 9 is in phase with V rms and it produces the real power P. In contrast,
,

7rms sin 9 is 90° out of phase with V rras and it causes the reactive power Q. Since 7rms
,

sin 9 is 90° out of phase with V rms it is called the quadrature component of I rms As a
, .

consequence, the reactive power is sometimes referred to as the quadrature power.

It is often convenient to view the complex power in terms of a diagram such as

that of Fig. 12.8. It is evident that for an inductive load (lagging pf), < 9 90°, <
Q is positive, and S lies in the first quadrant. For a capacitive load (leading pf),
— 90° 9 < 0, Q is negative, and S lies in the fourth quadrant. A load having
<
a unity power factor requires that Q — since 9 = 0. In general, we see that

9 = tan (12.28)
'(f)
Chap. 12 AC Steady-State Power 325

Re K Re

FIGURE 12.7 Phasor diagram of V rr FIGURE 12.8 Diagram for the

and l
rms complex power

Let us now consider the complex power associated with a load consisting of two
impedances Z and Z 2 as shown in Fig. 12.9. The complex power delivered to the
l
,

combined impedances is

V rms*rms
T I* V rms\(Ix
T
1 rms l

V rms*T*
T
1 rr
i_
1
v rms*2
*
I* rms

Therefore the complex power delivered by the source to the interconnected loads is

the sum of that delivered to each individual load, and the complex power is thus con-
served. This statement no matter how many individual loads there are or how
is true
they are interconnected, because it depends only on Kirchhoff's laws and the defini-

tion of complex power. This principle is known as conservation of complex power.

Conservation of complex power may be used in a straightforward manner to
correct the power factor. As an example, let us consider the circuit of Fig. 12.6 once
again. The complex power to the uncorrected load Z is

S = P+jQ
Connecting a pure reactance Z, in parallel with Z results in a complex power to Z, of

S, =JQ t

l,„.

FIGURE 12.9 Load consisting of Z x

and
7.1 in parallel
326 AC Steady-State Power Chap. 12

Therefore, from conservation of complex power, for the composite load we have

= P+KQ+Qr)
It is evident that the addition of Z x
does not affect the average power P delivered to
the load. It does, however, affect the net reactive power. We can therefore select Q 1

to obtain our desired power factor, which, of course, is usually raised. This causes
a reduction in the current required to produce P, as discussed previously.
Let us again consider the circuit of Fig. 12.4 and change the power factor to
PF = 0.95 lagging. The complex power of the uncorrected load is

c
*-*
__ v rms'rras
T T* — P+jQ = 25+y'25
since

V rms
T = 70.7 V
—_ V rms
T
= 0.3535(1 -yi)A
-*rms
Z
From (12.28), we see that QT = Q + Q l
must satisfy

= tan-(^)
Therefore

cos^ = PF=cos(tan-'^
and

QT = P tan (cos' 1
PF)
= 25 tan 18.2° = 8.22 vars

The required Q x
is

Qi = Qt - Q= 8.22 - 25 = -16.78 vars

Since 0, = ^UlmZ.VIZJ = V\JX U 2

then

X = x

^W = -297.9 Q

which represents a capacitance C— —\/coX l

= 33.6 /iF. This is identical to our
previous result in Sec. 12.4.
As a final example, let us find the power factor of two loads connected in parallel,
as shown in Fig. 12.9. Suppose Z represents a 10-kW load with a power factor
x

pfi = 0.9 lagging and Z 2 a 5-kW load with pf2 = 0.95 leading. For Z, we have

s, =P +jQ i i
Chap. 12 AC Steady -State Power 327

where P =
i
10 4 W, 0, = cos" 1

pf, = 25.84°, and, from (12.28),

(?, = P, tan 0, = 4843 vars

Similarly, for Z 2 we have

S2 = P2 + ;&
where P =
2
5 x 10 3 W, 9 2 = -cos" pf2 1
= -18.2°, and

Q2 = 5 x 10 3 tan^ 2 = -1643 vars

The total complex power is

ST = P + P2 +j(Q + Q
1 1 2)

= 1.5 x 10 4
-fy'3200

Therefore, for the combined loads,

°
= an 04
* '
-'(rOT) = 12 -

and

/,/ = 0.978 lagging

EXERCISES
12.5.1 Find the complex power delivered to a load which has a 0.91 lagging power
factor and (a) absorbs 1 kW, (b) 1 kvar, and (c) 1 kVA.
Arts, (a) 1099 /24.5° (b) 2412 /24.5°
, ,
(c) 1000/24.5°

12.5.2 Find the impedance of the loads in Ex. 12.5.1 if Vtms = 115 V.
Ans. (a) 12.03 /24.5° (b) 5.48 /24.5°
, ,
(c) 13.23 /24.5°

12.5.3 Repeat Ex. 12.4.2 (c) using the concept of complex power.

12.6 POWER MEASUREMENT

A wattmeter is a device which measures the average power that is delivered to a load.
It contains a rotating high-resistance voltage, or potential, coil, connected in parallel
with the load and a fixed low-resistance current coil which is connected in series with
the load.The device has four terminals, a pair to accommodate each coil. A typical
connection is shown in Fig. 12.10. We see that the current coil responds to the load
328 AC Steady-State Power Chap. 12

CURRENT
COIL

VOLTAGE
COIL

FIGURE 12.10 Typical connection of a wattmeter

current, whereas the voltage coil responds to the load voltage. For frequencies above
a few hertz, the meter movement responds to the average power. Ideally, the voltage
across the current coil and the current in the voltage coil are both zero, so that the
presence of the meter does not influence the power it is measuring.
One terminal on each of the coils is marked ± so that if the current enters the ±
terminal of the current coil and the ± terminal of the voltage coil is positive with
respect to its other terminal, then the meter gives a positive, or upscale, reading. In
Fig. 12.10 this corresponds to the load absorbing power. If the terminal connections
of either the current coil or voltage coil (but not both) are reversed, a negative, or
downscale, reading is indicated. Most meters cannot read downscale —the pointer
simply rests on the downscale stop. Thus such a reading requires reversing the connec-
tions of one of the coils, usually the voltage coil. Reversing the connections of both
coils does not affect the reading.
The wattmeter of Fig. 12.10, represented by the rectangle and the two coils, is

connected to read

P = |V| • |I|cos0

where V and I are as indicated and is the angle between V and I (or, equivalently,
the angle of the load impedance). This, of course, power delivered to the load.
is the
Other types of meters are available for measuring the apparent power and the
reactive power. An apparent power or VA meter simply measures the product of the
rms current and rms voltage. The varmeter, on the other hand, measures the reactive
power.

EXERCISE
12.6.1 Determine the power reading of each wattmeter after assigning the terminal
markings required for a positive reading. Ans. 87.5 W, 9.375 W, W
Chap. 12 AC Steady-State Power 329

ioon
—v>

6n;

ioo rnrn
100
n
©J I A

V rms
Q)

EXERCISE 12.6.1

PROBLEMS
12.1 One cycle of a periodic current is given by

i = 0.1(l - e~'), 0< < t 1

If the current flows in a 10-Q resistor, find the average power.

12.2 Determine the average power for p(t) of Fig. 12.1 (b).

12.3 Find the average power absorbed by the 10-Q resistor.

->Wv-
ion

10 COS 5t
V
© 0.5 H 0.1 F _~

PROBLEM 12.3

12.4 Repeat Prob. 12.3 with the Jg-F capacitor replaced by a current-controlled
voltage source of value v = 10/ V.

12.5 For a Thevenin equivalent circuit consisting of a voltage source \ g and an

impedance Z g = R g + jXg (a) show that the circuit delivers maximum
,

average power to a load Z L = R L + jXL when R L = R g and XL = g

—X ,

and (b) show that the maximum average power is delivered to a load R L when
330 AC Steady-State Power Chap. 12

RL = |Z g |. This is the maximum power transfer theorem for ac circuits. (In

both cases, Vg and Zg are fixed, and the load is variable.)

12.6 Find the average power delivered to R.

~ 0.25 F
2 COS 2t

I n

PROBLEM 12.6
o
12.7 Repeat Prob. 12.6 with the 2-A source replaced by a source having i = 2 cos A t

in the same direction.

12.8 We have defined the average power for an instantaneous power p(t), which is

not necessarily periodic, by

P= lim
1

Pit) dt

Show for v and i of (12.7)— (12.8) that this definition yields the same result as
(12.6).

12.9 Given: i = /', + i2 , where

/, = I lm cos co x
t

'2 = hm cos (o 2 t
is the current flowing in a resistor R. Using the definition of Prob. 12.8, show
for cx>! =£ co 2 that

P= /», + P2
where P, and P2 are the average powers associated with /', and i2 , respectively,
acting alone. Note that this includes the case of pit) being nonperiodic.

12.10 Find the rms value of the voltage (a) v = 100 cos (lOOt + 35°) V and (b)
v = 8 cos It + 4 cos It + b^TL cos (3t - 45°).

12.11 Determine the rms value of a periodic current for which one cycle is given by

(a) i = ftA, 0<t<7l

= 0, < t<2n.
7i

(b) i = Im e~'A, 0</ < 1.

12.12 Find the power factor of the parallel connection of a 1-kfi resistor, a 2-H
inductor,and a 10-//F capacitor at 60 Hz. At what frequency does a unity
power factor occur?
Chap. 12 AC Steady -State Power 331

12.13 Repeat Prob. 12.12 for the elements connected in series.

12.14 Derive (12.22).

12.15 Determine the power factor for the circuit. What parallel reactance will change
the power factor to 0.95 lagging?

30X1

PROBLEM 12.15

12.16 Determine the power factor as seen from the terminals of the independent
voltage source.

I .a

us Z0( ,21 2]

V rms

PROBLEM 12.16

12.17 Determine the reactive element needed to be added in parallel to the source to
power factor of Prob. 12.13 to 0.95 lagging. Assume a frequency of
raise the
100 Hz.

12.18 A load Z L requires 10 kW at a power factor of 0.7 lagging for a load voltage of
230 V rms at 60 Hz. The transmission line connecting the generator to the load
has a series impedance of 0.1 +/0.4D. Determine the generator voltage and
power factor.

12.19 A 230-V rms 60-Hz load draws 500 kW and 400 kvars. Compute the parallel
capacitance necessary to give a power factor of (a) 1.0, (b) 0.9 lagging, and (c)

0.9 leading.

12.20 Two loads in parallel draw a total of 3 kW at a 0.9 lagging pf from a 115-V
rms 60-Hz line. One load is known to absorb 1000 W at a 0.8 lagging power
factor. Find (a) the power factor of the second load and (b) the parallel reactive
element value necessary to correct the power factor to 0.95 lagging for the
combined load.

12.21 A load Z L requiring 5 kW at a 0.8 lagging power factor is supplied from two
60-Hz generators, as shown. Generator Vj has a terminal voltage of 230 V
rms and delivers 2.5 kW at a 0.8 lagging power factor through a transmission
332 AC Steady -State Power Chap. 12

line having an impedance Z =1.4 +yl.6fi. The transmission line connect-

{

ing V2 has an impedance Z 2 = 0.8 + /1 Q. Find (a) the load voltage V L (b) ,

the generator voltage V 2 , and (c) the complex power supplied by V2 .

z, 7
"~z

T
z L vL
O
PROBLEM 12.21
 13
THREE-PHASE CIRCUITS

As we have already noted, one very important use of ac steady-state analysis is its

application to power systems, most of which are alternating current systems. One
principal reason for this is that it is economically feasible to transmit power over long
distances only if the voltages involved are very high, and it is easier to raise and lower
voltages in ac systems than in dc systems. Alternating voltage can be stepped up for
transmission and stepped down for distribution with transformers, as we shall see in
Chapter 16. Transformers have no moving parts and are relatively simple to construct,
whereas with the present technology, rotating machines are generally needed to raise
and lower dc voltages.
Also, for reasons of economics and performance, almost all electric power is
produced by polyphase sources (those generating voltages with more than one phase).
In a single-phase circuit, the instantaneous power delivered to a load is pulsating, even
if the current and voltage are in phase. A polyphase system, on the other hand, is

somewhat like a multicylinder automobile engine in that the power delivered is

steadier. Consequently, there is less vibration in the rotating machinery, which, in
turn, performs more efficiently. An economic advantage is that the weight of the
conductors and associated components required in a polyphase system is appreciably
less than that required in a single-phase system that delivers the same power. Virtually

all the power produced in the world is polyphase power at 50 or 60 Hz. In the United

States 60 Hz is the standard frequency.

In this chapter we shall begin with single-phase, three-wire systems, but we shall
concentrate on three-phase circuits, which are by far the most common of the poly-
phase systems. In the latter case the sources are three-phase generators that produce a
balanced set of voltages, by which we mean three sinusoidal voltages having the same
amplitude and frequency but displaced in phase by 120°. Thus the three-phase source
is equivalent to three interconnected single-phase sources, each generating a voltage

333
334 Three-Phase Circuits Chap. 13

with a different phase. If the three currents drawn from the sources also constitute
a balanced set, then the system is said to be a balanced three-phase system. This is

the case to which we shall restrict ourselves, for the most part.

13.1 SINGLE-PHASE, THREE-WIRE SYSTEMS

Before proceeding to the three-phase case, let us digress in this section to establish
our notation and consider an example of a single-phase system that is in common
household use. This example will also serve to give us some practice with a single-
phase system, with which we are already familiar and which will serve as an introduc-
tion to the three-phase systems to be considered next.
In this chapter we shall find extremely useful the double-subscript notation
introduced in Chapter 2 for voltages. In the case of phasors, the notation is \ ab for the

voltage of point a with respect to point b. We shall also use a double-subscript notation
for current, taking, for example, l ab as the current flowing in the direct path from point
a to point b. These quantities are illustrated in Fig. 13.1, where the direct path from a
to b is distinguished from the alternative path from a to b through c.

V Qb

FIG U R E 1 3.1 Illustration of double-subscript notation

Because of the simpler expressions for average power that result, we shall use rms
values of voltage and current throughout this chapter. (These are also the values read
by most meters.) That is, if

V = |V|/0° Vrms
(13.1)
1 = |I|/-fl Arms

are the phasors associated with an element having impedance,

Z = |Z|/0 Q (13.2)
 1

Chap. 13 Three-Phase Circuits 335

the average power delivered to the element is

P = |V| •
1
1 1 cos <9

1
2
ReZW (13.3)

In the time domain the voltage and current are

v = «/2 V | |
cos cot V
i = »/2 1
1 1 cos (cot — 0) A

The use of double subscripts makes it easier to handle phasors both analytically
'

and geometrically. For example, in Fig. 13.2(a), the voltage V ob is

Van : |0 °

V rms
£© V Qb

-0-
°
/„„= l00 /- |20
V rms

(o) (b)

FIGURE 1 3.2 (a) Phasor circuit; (b) corresponding phasor diagram

This is evident without referring to a circuit since by KVL the voltage between two
points a and b is the same regardless of the path, which in this case is the path a,n,b.
Also, since \ nb = —\ hn , we have

100 - 100/- 120°

which, after simplification, is

\ ab = 100.^3730° V rms

These steps are shown graphically in Fig. 13.2(b).

A single-phase, three-wire source, as shown in Fig. 13.3, is one having three output
terminals a, b, and a neutral terminal n, for which the terminal voltages are equal. That
is,

V =V v, (13.4)
336 Three-Phase Circuits Chap. 13

This is a common
arrangement in a normal household supplied with both 115 V and
230Vrms, since if |V J = |V,|= 115 V, then |V
fl
= |2V,| = 230 V. fl6 |

Let us now consider the sourceof Fig. 13.3 loaded with two identical loads, both
having an impedance Z,, as shown in Fig. 13.4. The currents in the lines aA and bB
are

aA "
z, z,
and

l„B
z, 'z,

a a

Von© 2 I

Vnb(^ 1

FIGURE 13.3 Single-phase, three- FIGURE 13.4 Single-phase, three-

wire source wire system with two identical loads

Therefore the current in the neutral wire, nN, by KCL is

U = -Q.A + 1„b) =
Thus the neutral could be removed without changing any current or voltage in the
system.
If the lines aA and bB are not perfect conductors but have equal impedances Z 2 ,

then I nN is still zero because we may simply add the series impedances Z, and Z 2 and
have essentially the same situation as in Fig. 13.4. Indeed, in the more general case
shown in Fig. 13.5, the neutral current I nN is still zero. This may be seen by writing the
two mesh equations

(Z, + z + Z )I^ + Z I 6B - Z,I = V,

2 3 3 3

ZJ aA + (Z, + Z + Z )I + z^ = -V, 2 3 fcB

and adding the result, which yields

(Z, +Z +Z 2 3
)(I a/1 + l bB ) +Z 3 (I a ^ + i bB ) = o
Chap. 13 Three-Phase Circuits 337

Vi

9 O z3 N 1

".© ' hP

FIGURE 13.5 Symmetrical single-phase, three-wire system

or
Ka + hB = o (13.5)

Since by KCL the left member of the last equation is —l nN , the neutral current is zero.
This is, of course, a consequence of the symmetry of Fig. 13.5.
If the symmetry of Fig. 13.5 is destroyed by having unequal loads at terminals
A-N and N-B or unequal line impedances in lines aA and bB, then there will be a
neutral current. For example, let us consider the situation in Fig. 13.6, which has two
loads operating at approximately 115 V and one at approximately 230 V. The mesh
equations are

431, - 2I 2 -40I =115 3

21, 631, - 601, =115

-401, - 60I 2 + (110 +ylO)I 3 =
Solving for the currents, we have

I A

ion

jioo

b B

FIGURE 13.6 Unsymmetrical single-phase, three-wire system

338 Three-Phase Circuits Chap. 13

I, = 16.32 /-33.7° Arms

12 = 15.73 /-35.4° Arms
1 3
= 14.46 /-39.9 A rms

Therefore the neutral current is

\nN = I2 - I, = 0,76 /184.3° A rms

and, of course, is not zero.

EXERCISES
13.1.1 Derive (13.5) by superposition applied to Fig. 13.5.

13.1.2 Find the power P 40 P 60 and P l0 +j l0

, , delivered to the loads, 40 Q, 60 Q, and
10 +yl0fi, respectively, of Fig. 13.6. Am. 249, 181, 2091 W
13.1.3 Find the power PaA PbB , , and PnN lost in the lines in Fig. 13.6.

Am. 266.3, 247.4, 1.2 W

13.1.4 Find the power Pan and Pnb delivered by the two sources in Fig. 13.6. Check
the results in Exs. 13.1.2 and 13.1.3 for conservation of power.
Am. 1561.4, 1474.5 W

13.2 THREE-PHASE Y-Y SYSTEMS

Let us consider the three-phase source of Fig. 13.7(a), which has line terminals a, b,

and c and a neutral terminal n. In this case, the source is said to be Y-connected
(connected in a Y, as shown). An equivalent representation is that of Fig. 13.7(b),
which is somewhat easier to draw.
The voltages \ an Y bn
, , and V c„
between the line terminals and the neutral terminal
are called phase voltages and in the cases we shall consider are given by

y an = vP iw
\ bn = K. /-12Q (13.6)

v c„
= r,/120°
or
v an = v,/or
v,l<r

v 6n = VJ120° (13.7)

v „ = VJ-120
c

In both cases, each phase voltage has the same rms magnitude Vp and , the phases are
Chap. 13 Three-Phase Circuits 339

(o) (b)

FIGURE 13.7 Two representations of a Y -connected source

displaced 120°, with \ an arbitrarily selected as the reference phasor. Such a set of
voltages is called a balanced set and is characterized by

(13.8)

as may be seen from (13.6) or (13.7).

The sequence of voltages in (13.6) is called the positive sequence, or abc sequence,
while that of (13.7) is called the negative, or acb, sequence. Phasor diagrams of the two
sequences are shown in Figs. 13.8(a) and (b), where we may see by inspection that

(o) (b)

FIGURE 13.8 (a) Positive and (b) negative phase sequence

 ^

340 Three-Phase Circuits Chap. 13

(13.8) holds. Evidently, the only difference between the positive and negative sequence
is the arbitrary choice of the terminal labels, a, b, and c. Thus without loss in generality
we shall consider only the positive sequence.
By (13.6), the voltages in the abc sequence may each be related to V a „.
The relation-
ships, which will be useful later, are

ybn = van /-i2o° ,

(13.9)
V en = V an /120°
T — I
l

The line-to-line voltages, or simply line voltages, in Fig. 13.7 are Y ab Ybc
, , and V ca ,

which may be found from the phase voltages. For example,

Y ab = V +
a„ Y„ b
= v,l& - vp /-\20°
= J^v p
\w_
In like manner,
y„ c = */TVP l- -90°
yca = «/?v,i- -210°

If we denote the magnitude of the line voltages by VL . then we have

Vt = */JV. (13.10)
and thus
V ab =VL /W, Vbc = VJ-9Q° , V ca = VL /-2\0° (13.11)

These results also may be obtained graphically from the phasor diagram shown in
Fig. 13.9.
Let us now consider the system of Fig. 13.10, which is a balanced Y-Y, three-phase,
four-wire system, if the source voltages are given by (13.6). The term Y-Y applies since
both the source and the load are Y-connected. The system is said to be balanced since
the source voltages constitute a balanced set and the load is balanced (each phase
impedance is equal — in this case to Z p The
). fourth wire is the neutral line n-N, which
may be omitted to form a three-phase, three-wire system.
The line currents of Fig. 13.10 are evidently

T vY an

V,„ V a „/-120°
bB = W-120° (13.12)
zP z,

V„„ V an /120°
cC ~ z, "
Z,
WI20°
Chap. 13 Three-Phase Circuits 341

FIGURE 13.9 Phasor diagram showing phase and line voltages

O
O^
'bn

e FIGURE 13.10 Balanced Y-Y system

The last two results are a consequence of (13.9) and show that the line currents also
form a balanced set. Therefore their sum is

-Ii* = I.* + Im + Ice =

Thus the neutral carries no current in a balanced Y-Y four-wire system.
In the case of Y-connected loads, the currents in the lines aA, bB, and cC are also
the phase currents (the currents carried by the phase impedances). If the magnitudes of
the phase and line currents are Ip and I L , respectively, then I L = Jp , and (13.12)
becomes
342 Three-Phase Circuits Chap. 13

hB = h l-0 ~ 120° = IJ-6 - 120° (13.13)

i cC = ij-e + no = ij-e + i2o°

where 6 is the angle of Zp .

The average power Pp delivered to each phase of Fig. 13.10 is

P. = Vj p cos 6
(13.14)
= I, Re Z p
2

and the total power delivered to the load is

P=3PP
The angle 9 of the phase impedance is thus the power factor angle of the three-phase
load as well as that of a single phase.
Suppose now that an impedance Z L is inserted in each of the lines aA, bB, and cC
and that an impedance Z N not necessarily equal to Z L is inserted in line nN. In other
, ,

words, the lines are not to be perfect conductors but are to contain impedances.
Evidently, the line impedances, except for ZN , are in series with the phase impedances,
and the two sets of impedances may be combined to form perfect conducting lines aA,
bB, and cC with a load impedance Z p + Z L in each phase. Therefore, except for Z N in
the neutral, the equivalent system has perfect conducting lines and a balanced load. If
the impedance Z N were not present in the neutral, it would be a perfect conductor,
and the system would be balanced as in Fig. 13.10. In this case, as we have seen,
points n and N are at the same potential, and there is no neutral current. Thus it does
not matter what is in the neutral line. It may be a short circuit or an open-circuit or
contain an impedance such as Z N and still no neutral current would flow and no
,

voltage would appear across nN. Obviously, then, the presence of equal line imped-
ances in aA, bB, and cC and an impedance in the neutral does not change the fact that
the line currents form a balanced set.
As an example, let us find the line currents in Fig. 13.11. We may combine the 1-Q
line impedance and (3 + j3)-Q phase impedance to obtain

Zp = 4+j3 = 5 /36.9° Q

as the effective phase load. Since by the foregoing discussion there is no neutral
current, we have
10 °/0° - 0/ , o
I fiQ A

The currents form a balanced, positive sequence set, so that we also have

\ bB = 20/- 156.9° A, I cC = 20/-276.9 A

 o A

Chap. 13 Three-Phase Circuits 343

100/0° V rms
o
in

O- 3+j3a

100

-Q
/-1

—
20° Vrms

w 3+j3n

3n
100 /120°

-0 — Vrms
S
I

w\a 3+j3a

FIGURE 13.11 Balanced system with line impedances

This example was solved on a "per-phase" basis. Since the impedance in the
neutral is immaterial in a balanced Y-Y system, we may imagine the neutral line to be a
short circuit. We may do this if it contains an impedance or even if the neutral wire is

not present (a three-wire system). We may then look at only one phase, say "phase A"
consisting of the source \ an in series with ZL and Zp , as shown in Fig. 13. 12. (The line
nN is replaced by a short circuit.) The line current \ aA the , phase voltage l aA Zp , and the
voltage drop in the line I^Z^ may all be found from this single-phase analysis. The
other voltages and currents in the system may be found similarly, or from the previous
results, since the system is balanced.

a A
zL

Zp

p n N
1
o o

FIG U R E 1 3.1 2 Single phase for a per-phase analysis

As another example, suppose we have a balanced Y-connected source, having line

voltage VL = 200 V rms, which is supplying a balanced Y-connected load with
P= 900 W at a power factor of 0.9 lagging. Let us find the line current lL and the
phase impedance Z.
p
. Since the power supplied to the load is 900 W, the power supplied
to each phase is Pp = 2M = 300 W, and from

VJB cos 9
we have
300= 0W) 7,(0.9)
 ° 5

344 Three-Phase Circuits Chap. 13

i

Therefore since for a Y-connected load the phase current is also the line current, we
have

'• - •> = wl) = 2 89- A

The magnitude of Zp is given by

| Z '' I
Vp 200/yT _ ^
1

/, 3 s A3"/(2)(0.9)

and since 6 = cos

_1
0.9 = 25.84° is the angle of Z p we , have

Zp = 40 /25.84° Q

If the load is unbalanced but there is a neutral wire which is a perfect conductor,
then we may still use the per-phase method of solution for each phase. However, if
this is not the case, this shortcut method does not apply. There is a very useful method
employing so-called symmetrical components which is applicable to unbalanced systems
and which the reader may encounter in a course on power systems. It is important to
note, in any case, that a three-phase circuit is still a circuit and, balanced or unbalanced,
may always be solved by general analysis procedures.

EXERCISES

13.2.1 Given: \ ab = 100/0° V rms is a line voltage of a balanced Y-connected three-

phase source. Jf the phase sequence is abc, find the phase voltages.
Ans. 57.7 /-30 , 57.7 /- 150° 57.7 /-27Q , V rms
13.2.2 In Fig. 13.10 the source voltages are determined by Ex. 13.2.1, and the load in
each phase combination of a 40-Q resistor, a 100- fj.¥ capacitor, and
is a series
a 0.1 -H inductor. The frequency is co = 500 rad/s. Find the line currents and
the power delivered to the load.
Ans. 1 .15 / -66.9 , 1.1 5 /- 186.9° , 1.1 / -306.9° A rms, 160 W
13.2.3 Show that if a balanced three-phase, three-wire system has two balanced
three-phase loads connected in parallel, as shown, then the load is equivalent
to that of Fig. 13.10 with

j Z Z
t 2
p
z, + z 2

13.2.4 If in Ex. 13.2.3 Z, draws 10 kW at a 0.8 power factor lagging, Z 2 draws 15 kW

at a 0.9 power factor leading, and the line voltage is VL = 300 V rms, find the
current I L in each line. Ans. 144.3 A rms
Chap. 13 Three-Phase Circuits 345

A o

B o

C o

EXERCISE 13.2.3

13.3 THE DELTA CONNECTION

Another method of connecting a three-phase load to a line is the delta connection, or A

connection. A balanced A-connected load (each of its phase impedances equal) is

shown in Fig. 13.13(a), in a way that resembles a A, and in an equivalent way in Fig.
1 3. 1 3(b). If the source is Y- or A-connected, then the system is a Y-A or a A-A system.
An advantage of a A-connected load over a Y-connected load is that loads may be
added or removed more readily on a single phase of a A, since the loads are connected
directly across the lines. This may not be possible in the Y connection, since the
neutral may not be accessible. Also, for a given power delivered to the load the phase
currents in a A are smaller than those in a Y. On the other hand, the A phase voltages
are higher than those of the Y connection. Sources are rarely A-connected, because if

the voltages are not perfectly balanced, there will be a net voltage, and consequently a
circulating current, around the delta. This, of course, causes undesirable heating
effects in the generating machinery. Also, the phase voltages are lower in the
Y-connected generator, and thus less insulation is required. Obviously, systems with
A-connected loads are three-wire systems, since there is no neutral connection.
From Fig. 13.13 we see that in the case of a A-connected load the line voltages are

the same as the phase voltages. Therefore if the line voltages are given by (13.11), as
before, then the phase voltages are

Vab = Vl/w, \ bc ----

yj-90\ \ CA = vj-210 (13.15)
where
VL =V, (13.16)
346 Three-Phase Circuits Chap. 13

a o a o-

bo- , B Zp

(o) (b)

FIGURE 13.13 Two versions of a ^-connected load

If Zp = \Z P \/ 9 , then the phase currents are

-^ = / p /30° - 9

IJ -90° - 9 (13.17)

Ica = ^= I J-2\0°-9
where
Vl
(13.18)

The current in line aA is

*~aA — *AB *C

which after some simplification is

ia„ = yr j p /-e

The other line currents, obtained similarly, are

i bB = A/jij-i2o°-e
l cC = /Jl /-240°-9
/S p

Evidently the relation between the line and phase current magnitudes in the
A case is
Chap. 13 Three-Phase Circuits 347

and the line currents are thus

i,< fj-e, \ bB rj -no e. i 7,7-240" -0 (1 3.20)

Thus the currents and voltages are balanced sets, as expected. The relations between
line and phase currents for the A-connected load are summed up in the phasor diagram
of Fie. 13.14.

Vab

FIGURE 13.14 Phasor diagram for a A-connected load

As an example of a three-phase circuit with a A-connected load, let us find the line
current I L in Fig. 1 3. 1 3 if the line voltage is 250 V rms and the load draws l .5 kW at a
lagging power factor of 0.8. For one phase, Pp = '
5
3
00 = 500 W, and thus

500 = 250/(0.8)
or
Ip = 2.5 A rms
Therefore we have
IL = J3 ip = 4.33 A rms

Finally, in this section, let us derive a formula for the power delivered to a balanced
three-phase load with a power factor angle 6. Whether the load is Y-connected or
A-connected, we have
P=3Pp = 3Vp Ip cos6
348 Three-Phase Circuits Chap. 13

In the Y-connected case, Vp = VL / />/T and Ip = IL , and in the A-connected case, Vp

= VL and Ip = IJ+/7. In either case, then,

/>=3-£=cos0
or
P = +/TVJL cos6 (13.21)

As a check on the previous example, (13.21) yields

1500 = v^T(250)/L (0.8)

or, as before,

IL = 4.33 A rms

EXERCISES
13.3.1 Solve Ex. 13.2.2 if the source and load are unchanged except that the load is
A-connected. [Suggestion: Note that in (13.15), (13.17), and (13.20) 30° must

be subtracted from every angle.]

Ans. 2vTT /-66.9°, 2^/T /- 186.9°, 2^T /- 306.9°, 480 W
13.3.2 A balanced A-connected load has Xp = 4 +j3Q, and the line voltage is

VL = 200 V rms at the load terminals. Find the total power delivered to the
load. Ans. 19.2 kW
13.3.3 If the lines in Ex. 13.3.2 each have a resistance of 0.1 D, find the power lost in
the lines. Ans. 1.44 kW

13.4 Y-A TRANSFORMATIONS

In many power systems applications it is important to be able to convert from a

Y-connected load to an equivalent A-connected load and vice versa. For example,
suppose we have a Y-connected load in parallel with a A-connected load, as shown in
Fig. 13.15, and wish to replace the combination by an equivalent three-phase load. If
both loads were A-connected, this would be relatively easy since corresponding phase
impedances would be in parallel. Also, as we saw in Ex. 13.2.3, if both loads are
Y-connected and balanced, the phase impedances may also be combined as parallel
impedances.
To obtain Y-to-A or A-to-Y conversion formulas, let us consider the Y and A
connections of Fig. 13.16. To effect a Y-A transformation we need expressions for
Chap. 13 Three-Phase Circuits 349

FIGURE 13.1 5 Y -connected and A-connected had in parallel

A B

1 !* 1
I ;

(a) (b)

FIGURE 13.16 (a) Y and (b) A connection

Y ab Y bc
, , and Y fa of the A in terms of Y a Y b and Y
, ,
c
of the Y so that the A connection
is equivalent to the Y connection at the terminals A, B, and C. That is, if the Y is

replaced by the A, the same node voltages V,,, YB , and Vc will appear, and the same
currents I, and I 2 will flow. Conversely, a A-Y transformation is an expression of the Y
parameters in terms of the A parameters.
Let us begin by writing nodal equations for both circuits. If node C is taken as
reference, in the case of the Y network we have

Y.V^ -YY a fl

Y VB
6
Y,V = I2

Y V^ - Y.V,, + (Y
a a + Y + Y )VD
b C

Solving for YD in the third equation and substituting its value into the first two equa-
350 Three-Phase Circuits Chap. 13

tions, we have, after simplification,

/ Y a Y, + Y a Y c
\ / Y a Y, \ _
\\. + \b + \ c r A
\\ a +\ +\ b c r B ~ li

(13.22)
(
Ya Y 6 W / Y a Y, + Y,Y c \ _

The nodal equations for the A circuit are

+ Y JYA - Y \ B =
(Y fli ab I,

-Y V, + (Y + Y )\ B =
flfc a6 bc I2

Equating coefficients of like terms in these equations and (13.22) and solving for the
admittances of the A circuit, we have the Y-A transformation

v*ab YaY 6

Y a + Y4 + Y c

v*bc Y4 Y C
(13.23)
Y a + Y, + Y c

Y1 ca Y Ya c

Ya + Y + Y b c

If we imagine the Y and A circuits superimposed on a single diagram, then Y a and

Y 6
are adjacent to Y aA Y and Y are adjacent to Y6c etc. Thus we may state (13.23) in
,
b c ,

words, as follows:

The admittance of an arm of the A is equal to the product of the admittances

of the adjacent arms of the Y divided by the sum of the Y admittances.

To A-Y transformation we may solve (13.23) for the Y admittances, a

obtain the
difficult task, we may write two sets of loop equations for the Y and A circuits. In
or
the latter case we shall have the dual of the procedure which led to (13.23). In either
case, as the reader is asked to show in Ex. 13.4.4, the A-Y transformation is

Z„6 +z +z ic ca

z =
z z fcc lb
fc

z fl6 + z, c + Zca

z fa z be

Z a6 + Z bc + Z ca

where the Z's are the reciprocals of the Y's of Fig. 13.16. The rule is as follows:

The impedance of an arm of the Y is equal to the product of the impedances

of the adjacent arms of the A divided by the sum of the A impedances.
Chap. 13 Three-Phase Circuits 351

(By adjacent here we mean "on each side of and terminating on the same node as." For
example, in the superimposed drawing of the Y and A, Z lies between Z ah and Zca and
all three have a common terminal A. Thus Z ab and Z ca are adjacent arms of Z a .)

As an example, let us find the input impedance Z of Fig. 3. 7(a). This is a problem
1 1

which would have required us to write loop or nodal equations in the past, because we
cannot simplify the circuit by combining series and/or parallel impedances. Replacing
the l-. I-. and ^-Q resistors, which constitute a Y, by their equivalent A, as shown in
Fig. 13.17(b). however, enables us to solve the problem readily.

60 6H

4n 4fl

FIGURE 13.17 Two equivalent circuits

Comparing Figs. 13.17(a) and 13.16(a), we see that Y a = 1, Yb = 1 + £ = 2, and

Y =
c
1 -f- •} = 3 U. Therefore, from (13.23),we have

1(2)
1+2 + 3

Y bc 2(3)
= 1
1+2 + 3

Y, „ = 3(1)
1+2 + 3

Therefore in Fig. 13.17(b) we have

3 a i a 2fi

Thus Fig. 13.17(b) may be simplified by combining parallel and series resistors to
obtain
Z=3 lQ g

As another example, suppose we have a balanced Y-connected load with phase

impedance Z y and wish to convert it to an equivalent A-connected load. By (13.23),
 l

352 Three-Phase Circuits Chap. 13

since Ya Y ,
4 , and Y c
are all equal to Y^,, the reciprocal of Zy , then the equivalent
A-connected load is also balanced because

YM ab = Y be = Y ca = t\/ Y
Yy = — 2

-5

Thus if Zd is the phase impedance of the equivalent balanced A-connected load, then

Z, = 3Z„ (13.25)

which may be used to convert from Y to A and vice versa.

EXERCISES

13.4.1 Find the input impedance seen by the source in the figure for Prob. 1 1.8 using
a Y-A or A-Y transformation to simplify the circuit. From this result find the

average power delivered by the source. Ans. (1 +j2)/5 Q, 8 W

13.4.2 A balanced three-phase source with VL = 200 V rms is delivering power to a
balanced Y-connected load with phase impedance Zj =6 +/8fi in parallel

with a balanced A-connected load with phase impedance Z2 = 9 +j\2Q.

Find the power delivered by the source. Ans. 7.2 kW
13.4.3 Show that the Y-A transformation of (13.23) is equivalent to

Z_ Z
ab
a + Z^Z
Zj
z
Z Za
c
r
-)-
c

Z„Zj + z„z + z z a c c
Zbc
za
za z + z*z + z z„
Z ra ft c c

or in words:

The impedance of an arm of the A is equal to the sum of the

products of the impedances of the Y, taken two at a time,
divided by the impedance of the opposite arm of the Y.

13.4.4 Derive (13.24).

Chap. 13 Three-Phase Circuits 353

13.5 POWER MEASUREMENT

It appears to be a simple matter to measure the power delivered to a three-phase load

by using one wattmeter for each of the three phases. This is illustrated for the three-
wire Y-connected load in Fig. 13. 8. Each wattmeter has its current coil in series with
1

one phase of the load and its potential coil across one phase of the load. The connec-
tions are theoretically correct but may be useless in practice because the neutral point
\ may not be accessible (as, for example, in the It would
case of a A-connected load).
be better, in general, to be able to make
measurements using only the lines a. h, and
the
c. In this section we shall show that this is, in fact, possible and that, moreover, only

two wattmeters are required, instead of three. The method is general and is applicable
to unbalanced as well as balanced systems.

n
Z|
/ \

b o z2
/ \

c o Z 3
f 1

'

FIGURE 13.18 Power measurement using three wattmeters

Let us consider the three-wire Y-connected load of Fig. 13.19, which has three
wattmeters connected so that each has its current coil in one line and its potential coil

between that and a common point x. If T is the period of the source voltages and
line

i
B ,and ic are the time-domain line currents, directed into the loads, then the total
/',,,

power P x indicated by the three meters is

354 Three-Phase Circuits Chap. 13

a o-

rrcr\±

b o- < N

c o-

FIGURE 13.19 Three wattmeters connected to a common point

Px = y\ ( v "* ia + Vbx ' b + v "' c) dt (13.26)

Regardless of the point jc, which is completely arbitrary, we have

V ax — V aN I
V Nx

V bx = V bN + V Nx

v cx = v cN + V Nx

Substituting these results into (13.26) and rearranging, we obtain

Px = y\ ( v °»'° + v M't + v cnO dt + -1 v Nx (/. + ib + Q dt

By KCL we have
i„ + it + ic =
so that

T (*W' a + vbN i b + v eN i c ) dt (13.27)

Thus the sum of the three wattmeter readings is precisely the total average power
delivered to the three-phase load, since the three terms in the integrand of (13.27) are
the instantaneous phase powers.
Chap. 13 Three-Phase Circuits 355

Since the point v in Fig. 13.19 is arbitrary, we may place it on one of the lines. Then
the meter whose current coil is in that line will read zero because the voltage across its

potential coil is zero. Therefore the total power delivered to the load is measured by
the other two meters, and the meter reading zero is unnecessary. For example, the
point x is placed on line b
in Fig. 13.20. and the total power delivered to the load is

P = PA+ Pc

where P A and Pc are the readings of meters A and C. It is important to note that one or
the other o\~ the two wattmeters may indicate a negative reading, and thus the sum

of the two readings is the algebraic sum.

A
a o Z|
1 k °

+
A

! 1
B
bo
Do z2
i I

JUUU+
C "
+ nnnr

) '
o Z3

FIGURE 1 3.20 Two wattmeters reading the total load power

The proof of the two-wattmeter method of measuring the total three-phase power
has been carried out for a Y-connected load. However, it holds also for a A-connected
load, as the reader is asked to show in Prob. 13. 19.
As an example, in Fig. 13.20 let the line voltages be a balanced abc sequence with

\ ab = lOO^T [0_° V rms

and phase impedances given by

z, =z =z = 2 3
io -fyion
Then we have
-V 6c
= -10(^/57- 120°:= 100-yT/60 o V rms
:

(100^J/^y)/-30 3

z, ioyr/45 c 5^/T /-75 Arms

\ cC = 5^/T7-315° A rms
356 Three-Phase Circuits Chap. 13

The meter readings are thus

Pa = V I aA cos (ang \ - ang l aA

I aft || |
ab )

= (100VT)(5 AT) cos (0° + 75°) /V

= 317W
and

Pc = V I c6 1|
I cC |
cos (ang \ cb - ang I cC )

= (100^3")(5v^2") cos (60° + 315°)

= 1183 W
with a total of

P= 1500 W
As a check, the power delivered to phase A is

PP = Van I || IaA cos I

(ang \ AN - ang \ aA )

= 100^/T
( >j )(
5 vr2")cos(-30° + 75°)

= 500 W
Since the system is balanced, the total power is

p=3Pp = 1500 W
which agrees with the previous result.

EXERCISES
13.5.1 In Fig. 13.18, let Z, = Z 2 = Z = 10/30^ fi, and let the line voltages be a
3

balanced abc sequence set, with \ ab = 200 /CT V rms. Find the reading of each

meter. Arts. 2/^/T kW

13.5.2 If the power delivered to the load of Ex. 13.5.1 is measured by the two watt-
meters A and C connected as shown in Fig. 13.20, find the wattmeter readings.
Check for consistency with the answer of Ex. 13.5.1.
Ans. 2/^/T, 4/^3" kW
13.5.3 Find the wattmeter readings PA and Pc and the total power P if the line volt-
ages are as given in Ex. 13.5.1 and Zj =Z =Z =
2 3
10/75° Q. Check the
answer by using P = 3PP . Ans. -598, 1633, 1035 W
Chap. 13 Three-Phase Circuits 357

PROBLEMS
13.1 It ym = \ nb = 100/0_° V rms, Z, = 10/30° Q, in /f-N. and Z 2 = 20/60° Q,
in A'-/?, find the neutral current \ nN in Fig. 13.4.

13.2 In Fig. 13.5, let V, = 100 /(T V rms, Z, = 5 + ;5 ft, Z 2 = 0.5 Q, Z 3 = 1 ft,
and Z4 = 10 — y"5 Q. Find the average power absorbed by the loads, lost in the
lines, and delivered by the sources.
13.3 In Fig. 13.10, the line currents form a balanced, positive sequence set with
\ aA = 10/0° A rms. If Zp = ^J [W Q, find the power delivered to the three-
phase load and the line voltages.

13.4 A balanced Y-connected load is present on a 240-V rms (line-to-line) three-

phase system. If the phase impedance is 4 /60° Q, find the total average power
delivered to the load.

13.5 Repeat Prob. 13.4 if the load is A-connected and the phase impedance is

12/60° Q.

13.6 A balanced Y-Y system has a positive sequence source with \an = 100/0_° V
rms and Zp = 10 /30° Q. Find the line voltage, the line current, and the power
delivered to the load.

13.7 A balanced Y-connected source, \an = 240 /(T V rms, positive sequence, is
connected by four perfect conductors (having zero impedance) to an unbal-
anced load, Z AN = 10 -y'5fi, Z BN = 10 Q, and Z CN =j20Q. Find the four
line currents.

13.8 A balanced Y-Y three-wire, positive sequence system has Van = 200/0^ V
rms and Zp = 3 - y'4 O,
and the lines each have an impedance of 1 ft. Find
the line current and the power delivered to the load.

13.9 A balanced three-phase Y-connected load draws 6 kW at a power factor of

0.8 lagging. A balanced Y of capacitors is to be placed in parallel with the load
so that the power factor of the combination is 0.9 lagging. If the frequency is

60 Hz and the line voltages are a balanced 200-V rms set, find the capacitances
required.

13.10 In Fig. 13.10, the source is balanced, with positive phase sequence, and
\ an = 100 /(F V rms. If the source provides 3 kVA at pf = 0.8 lagging, find
Zp .

13.11 Repeat Prob. 13.10 if the load is a balanced A.

13.12 Repeat Prob. 13.11 if each line contains an impedance of 1 fi.

13.13 A balanced Y-A system with V an = 120/CF V rms, positive phase sequence,
has Zp = 6 — j9Q. and an impedance of 1 Q in the lines. Find the power
delivered to the load.

13.14 A balanced three-phase, positive-sequence source with \ab = 200 [0^ V rms
is supplying a A-connected load, ZAB = 50 Q, Z BC = 20 -f-/'20fi, and
Z CA = 30 — y'40 Q. Find the phasor line currents. (Assume perfectly conducting
lines.)
358 Three-Phase Circuits Chap. 13

13.15 Repeat Prob. 13.14 if the impedance of each line is 1 Q, Z AB = 10Q,

ZBC =ylOft, and ZCA = 10 -jlOQ.
13.16 A balanced three-phase positive-sequence source with V = fl6 240 /0° V rms
is supplying a parallel combination of a Y-connected load and a A-connected
load. If the Y and A loads are balanced with phase impedances of 8 +y'8 Q
and 12 — j24Cl, respectively, find the line current lL and the power supplied
by the source, assuming perfectly conducting lines.
13.17 Find the power delivered by the source in Prob. 13.16 ifthe A load is replaced
by an unbalanced A load of Z AB = 6 — y'24 Q, Z BC = Z CA = —j24 Q.
13.18 For a balanced three-phase system, if the phase voltages are

v a {t) = Vm cos cot

v b (t) = Vm cos (cot - 120°)
Vc (t) = Vm cos (cot - 240°)
then the phase currents are

ia (t) = Im cos (cot -9)

i b (t) = Im cos (cot -9 - 120°)

ic (t) = Im cos (cot -9 - 240°)

Show that the total instantaneous power,

p(t) = v a ia +v b ib + vjc
is a constant, given by

/>(') = y VJm cos9

which is also P, the total average power. [Suggestion: Recall that cos a +
cos (a - 120°) + cos (a - 240°) = 0.]

13.19 Show that Px given by (13.26) is equal to the total average power delivered to
a A-connected load.

13.20 In the system of Fig. 13.20, the line voltages are a balanced, positive-sequence
set, with Va6 = 100/(r V rms, and Z, = Z 2 = Z = 3 10/3CT Q. Find the power
delivered to the load (a) by finding the readings of the two wattmeters and (b)
by P = 2PP .

13.21 Repeat Prob. 13.20 if Z = Z 2 = Z 3 = 10/60° Q. Note that

x
in this case watt-
meter C reads the total power since wattmeter A reads zero.

13.22 Show that in Fig. 13.20 if Va6 = VJjt, VAc = VJa - 120° ,V ca = VL /a - 240°
Vrms, and Z = Z 2 = Z 3 = |Z| /60° then wattmeter A reads zero and watt-
x ,

meter C reads the total average power delivered to the load,

T/2
P=
2|Z|

This is a generalization of the result of Prob. 13.21.

Chap. 13 Three-Phase Circuits 359

13.23 Show that if a balanced, positive-sequence source is connected by three per-

fectly conducting wires to a balanced load having phase impedance X p — Z, p \/9, \

then the two wattmeters, connected as in Fig. 13.20, read, respectively, PA and
Pc where
,

P* _ cos (30° + 9)
Pc cos (30° - 9)

From this result show that

tan 9 = +/T
Pc+Pa
Thus we may find the power factor, cos 9, of the load from the two wattmeter
readings.

13.24 InProb. 13.23, find the pf of the load if (a) PA =Pc ,(b)P.4 = -Pc ,(c)PA = 0,
(d)Pc = 0, and (e) PA = 2PC .
 14
COMPLEX FREQUENCY
AND NETWORK FUNCTIONS

In the previous chapters we have considered resistive circuit analysis, natural and
forced responses of circuits containing storage elements, and, in particular, ac steady-
state analysis.The excitations we have considered were, for the most part, constants,
exponentials, and sinusoids. In this chapter we shall consider an excitation, the damped
sinusoid, which includes all these excitations as special cases. From this function we
shall develop generalized phasors and general network functions which include, as
special cases, the phasors and impedances of Chapters 10-13.
The network functions will be expressed in terms of a complex frequency that
includes the frequency jco as a special case. The concepts of complex frequency and
general network functions will enable us to combine all of our earlier results into one
common procedure. Both the natural and forced responses of a circuit may be found
from its excitation and its network function, as we shall see. In addition, the network
function may be used to determine the frequency-domain properties of the circuit,

which will be the subject of Chapter 15.

14.1 THE DAMPED SINUSOID

In this section we shall consider the damped sinusoid,

v = Vm e°' cos (cot + 0) (14.1)

which is a sinusoid, like those of the previous chapters, multiplied by a damping

factor e at . The constant a (the Greek letter sigma) is real and is usually negative or
zero, which accounts for the term damping factor.
The damped sinusoid contains, as special cases, most of the functions we have

360
Chap. 14 Complex Frequency and Network Functions 361

considered thus far. For example, if ct 0. we have the pure sinusoid

v= Vm cos(cot I 0) (14.2)

of the previous chapters. If co 0. we have the exponential function

v= V e" (14.3)

and if a = oa = 0, we have the constant (dc) case

v=V (14.4)

where, in both (14.3) and (14.4). V = Vm cos <f>.

These functions are sketched in Fig.
14.1 for the various cases of a and co.

As we see from Fig. 14. 1. a >

represents growing oscillations (b) or a growing
exponential (e). and a < represents decaying oscillations (a) or a decaying expo-
nential (d). Finally, a = represents ac steady state (c) or dc steady state (f).

The units of co are radians per second and of <f>

are radians or degrees, as before.
Since at is dimensionless. a is in units of I per sec (1/s). This unit was encountered
earlier, in connection with natural frequencies, in Chapter 9, where the dimensionless
unit, neper (Np), was used for at. Thus a is the neper frequency in nepers per second
(Np/s). The neper was chosen in honor of the Scottish mathematician John Napier
(1550-1617), who invented logarithms.
As an example of a circuit having a damped sinusoidal excitation, let us find the
forced response i in Fig. 14.2. The loop equation is

2^-
dt
+ 5/ = 25e~' cos It (14.5)

Trying as a forced response

/ = e -'(A cos It + B sin It)

which is the excitation and all its possible derivatives, we have

2e-'(—2A sin It - A cos It + IB cos It —B sin It)

+ 5e~'(A cos It +B sin It) = 25e~' cos It

Since this must be an identity, we have

+
3A AB = 25

-AA + 35 =

or A = 3 and B= A. Therefore the forced response is

i = e-'{2> cos It + 4 sin It) A

 Complex Frequency and Network Functions Chap. 14
362

(o) a <0 (d) a< o, co =

(b) a >o (e) a >0, co=0

(c) a = ( f l <r = CO=

FIGURE 14.1 Various cases of (14.1)

A/W

25 e"' cos 2t
V © 2H

FIGURE 14.2 Circuit excited by a damped sinu-

soid
Chap. 14 Complex Frequency and Network Functions 363

or, cquivalently,

i = 5e~' cos (2t- 53.1°) A (14.6)

As expected, the forced response is. like the excitation, a damped sinusoid.

EXERCISES
14.1.1 Excite the circuit of Fig. 14.2 by the complex function

i>! = 25e-'eJ 2 '

= 25e (
- 1+ J' 2 "

and show that the forced response is

/, =(5 /-53.1° )g (
- 1+ ^"

14.1.2 Show that if fj is the response to v t

in

~di\ ..
2^+5,, =«,

then Re ij is the response to Re v\. Apply this result to the functions of Ex.
14.1.1 to obtain /'given by (14.6).

14.1.3 Find the forced response v using the methods of Exs. 14.1.1 and 14.1.2 if

d3v ,d v 2
dv
+ 6 dF + U dt +6v =
, , ,
,
, ,

l
diT

where
i = 4e~ 2 cos<
(t - 60°)
Ans. 2e~ 2 cos'
(t + 30")

14.2 COMPLEX FREQUENCY

AND GENERALIZED PHASORS

In Chapter 10 we considered circuits with sinusoidal excitations such as

v x
= Vm cos (cot + 0) (14.7)

which we also wrote in the equivalent form

v x
= Rc(Vm e J
^e j0")
364 Complex Frequency and Network Functions Chap. 14

Using the phasor representation of v u

V= V„e'* = Vm [$ (14.8)
we may also write

t>, = Re(Ve*»') (14.9)

The phasor representation was extremely useful in solving ac steady-state circuits in

which the voltages and currents were sinusoids of the form of (14.7).
In this chapter we wish to consider circuits in which the excitations and forced
responses are damped sinusoids, such as

v = Vm e°' cos (cot + <f>)

(14.10)

A good question at this point is, Can we define phasor representations of damped
sinusoids that will work as well for us as the phasors, such as (14.8), of undamped
sinusoids? It seems plausible that we can, because the properties of sinusoids that
made the phasors possible are shared by damped sinusoids. That is, the sum or differ-
ence of two damped sinusoids is a damped sinusoid, and the derivative, indefinite
integral, or constant multiple of a damped sinusoid is a damped sinusoid. In all these
operations only Vm and <f>
may change. This is exactly the case with undamped sinu-
soids.
Let us see what follows if we write (14.10) in the form analogous to (14.9). That is,

v = Vm e"' cos (cot + 0)

= RefF £>
CT
'e-'
1"
(' + *)
l

If we define the quantity

s = o+jco (14.11)
then we have
v=Re(Ve") (14.12)

where V is the phasor of (14.8).

Evidently, since V is the same for both, the undamped sinusoid given by (14.9) is

identical in form to the damped sinusoid of (14.12). The only difference is that the
numbery'co is used in one case and the number s is used in the other. Obviously, then,
we may do anything with the damped sinusoids that we did with undamped sinusoids
as long as we use s instead of jco. We may define the phasor V of (14. 8) to be the phasor
representation of v in (14.10) and use it for damped sinusoidal circuit problems in
exactly the same way that we used it for the sinusoidal problems. In the damped
sinusoidal case we may wish to write the phasor as \(s) to distinguish it from the
undamped sinusoidal case, \(jco).
As an example, the damped sinusoid

v = 25<? -' cos It V (14.13)

Chap. 14 Complex Frequency and Network Functions 365

of Fig. 14.2 has the phasor representation

V = V(» = 25/(F (14.14)

where s = I
j2. Conversely, if V is given by (14.14) and s is as specified, then v
is given by (14.13).
Some authors prefer to call V(.v), corresponding to v(t) in (14.10), a generalized
phasor. even though it is identical to the phasors of sinusoidal functions. It is, how-
ever, a function of a generalized frequency, namely s, given by (14.11). Since s is a
complex number, it is more often called a complex frequency. Its components are

a = Re s Np/s
co = Im .s- rad/s

which have units of frequency, and, indeed, s is the coefficient of / in an exponential

function, as was the case in Chapter 9 for the natural frequencies of a circuit. The
units of ,v are 1 /second and are sometimes called complex nepers per second or complex
radians per second.
It might be worth noting at this point that a function which can be written in the
form
f(t) = A>'" + K2 e°« + + K„e"<

where the AT,, and s, are independent of /, may be said to be characterized by the
complex frequencies s u s 2 s n For example, writing (14.10) in the form
. . . . , .

v=Vm ^—-ti )

results in

v =K l
e
(a+Ja' u
+K 2e
{
°- Ja> ) '

where AT, = and K2 = Vm e~ J *f2 = Kf (the complex conjugate). Thus v pos-

Vm e J
*/2

sesses not one but two complex frequencies, namely Si = a +jco and s 2 = s* = a —jeo.
This concept of complex frequency is consistent with that of the natural frequencies
considered in Chapter 9.

EXERCISES
14.2.1 Find the complex frequencies associated with (a) 6 + 4e 2 ', (b) cos cot,
(c) sin (cot + 6), (d) 6e~ 3 sin (At + 10°), and (e) e~'(\ + cos It).
'

Ans. (a)0, -2;(b) ±jco;(c) ±jco;(d) -3 ±/4;(e) -1, -1 ±j2

366 Complex Frequency and Network Functions Chap. 14

14.2.2 Show that if / is a damped sinusoid,

i = Im e<" cos (cot + 0)

then v, defined by

v=L*+Xt
is also a damped sinusoid of the same complex frequency.
L
Arts, v = Im s/(R + oL) 2
+ co 2 L 2 e°' cos (cot + + tan"" 1 °?
<f>
gL )
14.2.3 Find j and V(j) if w(f) is given by (a) 6, (b) 6e"', (c) 6e~' cos (2/ + 10°), and
(d) 6 cos (It + 10°).

/Ira. (a) 0, 6/CT; (b) -1, 6/0^; (c) -1 +j2, 6/10_°; (d);2, 6 /10°

14.2.4 Find v(t) if (a) V= 4/0°, 5 = -3; (b) V= 5/15°, ,y=/"2; and
(c) V= 6 / -30° s ,
= -2 +/4.
/Ira. (a) 4e~ 3 ', (b) 5 cos (2/ + 15°), (c) 6e~ 2t cos (At - 30°)

14.3 IMPEDANCE AND ADMITTANCE

Because of the identical form of the phasor representations for sinusoids and damped
sinusoids, we may find forced responses to damped sinusoidal excitations using
phasors precisely as we did in the previous chapters. All the concepts and rules, such
as impedance, admittance, KCL, KVL, Thevenin's theorem, Norton's theorem, super-
position, etc., carry over to the damped sinusoidal case exactly. We need only to use
s = a -\-jco rather than/co.
It follows that, in the s domain, the phasor current l(s) and voltage \(s), associated
with a two-terminal device, are related by

\(s) = Z(s)l(s)

where Z(s) is the generalized impedance, or simply impedance, of the device. We may
obtain Z(s) from Z(jco), the impedance in the ac steady-state case, by simply replacing
jco by s.

For a resistance R, the impedance is, therefore,

Z*(*) =R
In the case of an inductance L, the impedance is

Z L (s) = sL
and for a capacitance C, it is
Chap. 14 Complex Frequency and Network Functions 367

In a similar manner the admittances are. respectively,

Y*(*)
R
G, Yl(s) = IL>
Yf(, ° y<

Impedances in series in exactly the same way as in the

or parallel are combined
ac steady-state case, since we merely replace joi by s. For example, let us reconsider
the circuit of Fig. 14.2, whose phasor circuit is shown in Fig. 14.3, where s = — +j2. 1

The impedance seen from the source terminals consists of the impedances 2s and
5 in series and thus is given by

Z(s) = 2s + 5 ft

I(s)
5

Vg (S) f-\ 2s

FIGURE 14.3 Phasor circuit of Fig. 14.2

Therefore since the input voltage phasor is

Vg (s) = 25((F\
we have
V,(J) 25/0°
Ks) 5/-53.TA
2s + 5 2(-l+j2) + 5

Thus in the time domain the forced response, as before, is

i(t) = 5e~' cos (2/ - 53. 1 ) A

As another example, let us consider the time-domain circuit of Fig. 14.4(a), where
it is required to find the forced response v (t) for a given damped sinusoidal input
v g (t). The phasor circuit is shown in Fig. 14.4(b), from which we have the nodal
equations,

(y + 1 + |*)V, - |v - V - ±.sY
g 2 =

+ -H v - v, = o
>

We note that V 2 = V /2 since the op amp and its two connected resistors constitute
a VCVS with a gain of 2. Therefore eliminating V, and V we have 2 ,
 f

368 Complex Frequency and Network Functions Chap. 14

2A
v o—VW —VW— *

<i o "0

(o)

v
—w—^vw^
,, oV„

(b)

FIGURE 14.4 (a) Time-domain circuit; (b) its phasor circuit

16
V (s) = V.(J) (14.15)
s
2
+ 2s + 8
'

If we have
v g (t) = e~ 2 cos At
'
V

then s = -2+j4 and V g

(s) = 1 (0°. Thus from (14.15) we have

\ (s) = s/T/\35°V
and therefore
Vo (t) = J~2e~ 11 cos (4/ + 135°) V
Chap. 14 Complex Frequency and Network Functions 369

As a final example in this section let us find the forced response i in Fig. 14.5 if

FIGURE 14.5 Circuit with two sources

v gi = 8e"' cos / V
and
i
f2
= 2e-"A

Since the complex frequencies of v gl and i

g2 are different, we must use superposition.
That is.

/ = /,+/,

where i, is due to vel acting alone and i 2 is due to ig2 acting alone.
The phasor circuits for i g2 = and v gl — are shown in Figs. 14.6(a) and (b),

(o) (b)

FIGURE 14.6 Phasor circuits associated with Fig. 14.5

respectively. The phasor currents are I, and I2, as shown. In Fig. 14.6(a), using current
division, we have
8/0° 4/s
I,
1 - {[(2s - 4)(4/j)]/(2* - 4 + 4/s)} 2s -r 4 + 4/s

which, since s = —1 -j-jl for this case, becomes

I, - 2 va~/-45°A
Therefore the forced component due to v gl alone is

z !
= 2 /VA2V' cos (t - 45°) A
370 Complex Frequency and Network Functions Chap. 14

In Fig. 14.6(b), again by current division, we obtain

t
2
_ / [K4/J)]/(1 + 4/s) \
(2/0
o.

M[1(4/j)]/(1 + 4/s)} + 2s + 4r'— }

which, since s = —5, is

I2 = 0.8/0^ A
Therefore the forced component due to i g2 alone is

i2 = 0.8e" 5 'A

The complete forced response of Fig. 14.5 is, therefore,

i = 2 v r 2V' cos {( -
/ 45°) + 0.8e" 2 '
A

EXERCISES
14.3.1 Change the voltage source in Fig. 14.5 to

v gl = 6e- 3 'cos2?V

and find the forced response /'t due to v gl alone. Arts. —4e~ 3 '
cos 2t A
14.3.2 Replace all the circuit of Fig. 14.6 (a), except the inductor, by its Thevenin
equivalent, and find I,.
32
Voc = ^74, „ 4(s + 5)'
= -1 +yl
th =
\r v
a
Arts. Z ,
,

where s 1 , -i

^ 4
14.3.3 Assume Ij = A 1 in Fig. 14.6 (a), and use the method of proportionality to
find the true I,.

^'- ''
=
,»+%'+ 10
= 2 ^=4g A

14.4 NETWORK FUNCTIONS

A generalization of impedance and admittance is the so-called network function,

which, in the case of a single excitation and response, is defined as the ratio of the
response phasor to the excitation phasor. For example, if V^) and I(V) are the voltage
and current phasors associated with a two-terminal network, then the input impedance

is the network function if I(s) is the excitation and V(s) is the response. On the other
Chap. 14 Complex Frequency and Network Functions 371

hand, if V(.v) is the input and \{s) is the output, then the input admittance

\( S ) - M
is t lie network function.
The foregoing examples are special cases m thai both the input and output are
measured at the same pair of terminals. In general, for a given input current or voltage
at a specified pair of terminals, the output may be a current or voltage at any place in
the circuit. Thus the network function, which we designate in general as H(.v), may be
a ratio of a voltage to a current (in which case its units are ohms), a current to a voltage
(with units of mhos), a voltage to a voltage, or a current to a current. In the last two
cases, H(.v) is a dimensionless quantity.
As an example, consider the circuit of Fig. 14.4(b), which was analyzed in the
previous section. If V„(i) is the output phasor and V g (s) is the input phasor, then, by
(14.15), the network function is

\ g (s) s
2
+ 2s H 8

As another example, if the source \g = 8/0° V is the input and I, is the output of
Fig. 14.6(a). then from Ex. 14.3.3 we see that

HM = = (I4I6)
ffi) ,' + fe + io

In the former example H(s) is dimensionless and in the latter case its units are mhos.
The network function H(.s) is independent of the input, being a function only of
the network elements and their interconnections. Of course, when the input is specified
this in a given application. From a knowledge of
determines the value of s to be used
the network function and the input function, we may then find the output phasor and
subsequently the time-domain output. To be specific, suppose V,(^) is the input and
\ {s) is the output. Then

»w-v$ < I4I7 >

from which
V.(5) = Ws)Yfc) (14.18)

In general, since s is complex, H(j-) is complex. Therefore we may write, in polar

form,
H(j) = |H(s)|/0 (14.19)

where H(s) | |
is the amplitude and 6 the phase of H(s). Therefore if

Vfc)= VJ± (14.20)

then
V.(*)= Vm \H{s)\/<f> + 6 (14.21)
372 Complex Frequency and Network Functions Chap. 14

Thus the amplitude of V is Vm H(s) and its phase is <f> 9. That is, the amplitude
\ |, +
of the output is that of the input times that of the network function, and the phase
of the output is that of the input plus that of the network function.
As an example, I,(.y) in (14.16) is given by

I,(j) = H(j)V,(j)
where V g (s) = 8/0° V and s = —1 + y I, as given in the previous section. Therefore

H(-l +jl) =
(-1 +yi) 2 + 6(-i +j\)+ 10

T7-45
and

*' = (^)(8) /0 - 45°

= 2 <N /27-45° A
as was obtained earlier.

As a final note in this section, if there are two or more inputs, we may use super-
position to define a network function relating the output to each individual input,
with the other inputs made zero. For example, in the circuit of Fig. 14.5, we may find,
from Fig. 14.6(a) (with I 2
= 0), the network function

H,W - y
si

where s = -1 +./1 and V„ = 8/TV. Then from Fig. 14.6(b) (with V„ - 0), we
have the network function

H (j) =
2 ^
where s = —5 and \ gl = 2/0° A. We may then find Ij and I 2 and subsequently the
corresponding time-domain functions, /, and /,. These are, by superposition, the com-
ponents of i, as before.

EXERCISES
14.4.1 Given the network function

2( + 4)
HW- s2 +
' V

2^ + 2
and the input
y.(s) = 4/0°
 —
Chap. 14 Complex Frequency and Network Functions 373

find the forced response r (t) if (a) s = — 1 +73, (b) s = —3 + j\, and (c)

s= -1. Ans. (a) 3VT <r' cos (3? - 135°), (b) -2e" 3 'sin /, (c) 24<r'

14.4.2 Find H(.v) if the response is (a) !,(.?), (b) \ (s), and (c) \ (s).

Ans. (a)
J2 + 3s + 1
(b)
1 1

vw
(c)
(s + l){s + 2)(s I
;,
3)'
v "'
(^ | l)(s + 2)(s + c,
3)' (^ + l)(j + 2)

3A IH
—vw
i
tw^
+

v
»©

EXERCISE 14.4.2

14.5 POLES AND ZEROS

In general, the network function is a ratio of polynomials in s with real coefficients that
are independent of the excitation. To illustrate this, let us consider the example of
Fig. 14.2, described by

2^ + 5/ =v
dt

where i is the output and v is the input. Using the same technique with complex forcing
we used Chapter we note if v = We then the output must 1
functions that in 10, that ',

have the same form, namely i — le", where, of course, V = \(s) and I = l(s) are the
phasor representations of v and /'.
Substituting these values into the differential equa-
tion, we have
(25 + 5)Ie" = \e"

from which the network function is

V 2s

In the general case, if the input and output of the circuit are v,(t) and v (t), respec-
tively, then the describing equation is

'

a.
d"v
cit"
a .-1
dn
dt
n-l
]
v
a^ + a v Q

dm v l i\
+ *.-. T *=r + l

i6,f + V, (14.22)

The tf's and fe's, of course, are real constants and are independent off
374 Complex Frequency and Network Functions Chap. 14

As before, if v t
= \e t
s
', then the output must have the form v =V e", where
V,(^) and V (5) are the phasor representations of v t
and v . Substituting these values
into (14.22), we have

(a n s
n
+ flB -iJ
B_1
+ • • •
+ OiS + a )\ o e"

= {b m s
m
+ 6m . 1
5- + 1
• • •
+ b s
x
+ & )V,e" (14.23)

From this we obtain the network function

Vote)
VXj)
_ HM " sm + 6
w - b«„*" + a„. "
"--'-y

1
'""

5"- 1
1

+
+ •
•

•
'

'

+
+ ^5 +
a lS +
bQ
a
Kl ^
(W)A\
V
which is a ratio of polynomials in s.

We may also write the network function (14.24) in the factored form

hw = ^"
- Pi)(s —p
i?:"?!
,,,
— p„)
!rn"l
tf n (s
zi

2) . . . (s
(i4 25)
-

In this case thenumbers z 1 ,z2 ,...,zm are called zeros of the network function because
they are values of s for which the function becomes zero. The numbers p lt p 2 ,p„ ,

are values of ^ for which the function becomes infinite and are called poles of the
network function. The values of the poles and zeros, along with the values of the
factors a„ and b m uniquely determine the network function.
,

As an example, the network function

m K
v
}
_ 6(5
s(s
+ l)(s + 2^ + 2)
+ 2)(s + 4^+13)
2
2

_ 6(5+ l)(s+ 1 +jl)(s+ 1 -jl) n42M

s(s + 2)(s + 2 + 7'3)(5 + 2 - ;3) ^ '
;

has zeros at —1, —1 and poles at 0, —2, —2 +y'3, and —2 — y"3.

-{-jl, and —1 — jl
As is generally the case, because the a's and Z>'s of (14.24) are real, complex poles or
zeros always occur in conjugate pairs. Since the network function is a ratio of a third-
degree to a fourth-degree polynomial, it approaches zero as s becomes infinite. Thus
we also have a zero at s — oo. If the numerator were higher in degree than the denomi-
nator, s = oo would be a pole.
The poles and zeros of a network function may be sketched on a pole-zero plot,
which is simply the s-plane, consisting of a- and 7'cu-axes, with the poles and zeros
located on it. Zeros are represented by a small circle and poles by a small cross.
As an example, the pole-zero plot of (14.26) is shown in Fig. 14.7. The values
of a are on the horizontal {a) axis and those of co are on the vertical (so-called jco)
axis.

As we shall see in Chapter 15, pole-zero plots are useful in considering frequency-
domain properties of circuits.
Chap. 14 Complex Frequency and Network Functions 375

jOO
2+ j3 x -3

2
-i + i"o I

o
-2 -I
-l-jlO -I

-2

-2-j3 x -3

FIGURE 14.7 Pole-zero plot

EXERCISES
14.5.1 If the zeros of H(s) are s = —1, —1 ±/'l, the poles are s = —2, —1 ±j2,
and H(0) = 4, find H(s).

Ans
20(s + 1)(^ + 2s + 2)
-
{s + 2)(j 2 + 25 + 5)
14.5.2 Draw the pole-zero plot of the network function of Ex. 14.5.1.

14.6 THE NATURAL RESPONSE

FROM THE NETWORK FUNCTION

As we know from our previous experience, an output of a circuit consists of the sum
of a natural and a forced response. In the last few chapters we have been concerned
entirely with finding the forced response, and we have seen that the phasor technique
enables us to do this in a very easy, straightforward way in the cases of sinusoidal,
damped sinusoidal, and exponential excitations. In power systems studies the forced
response is, of course, an ac steady-state response and is always present. Therefore the
forced response is usually of more interest than the natural response, which is transient
and gone after a very short time.
With damped sinusoidal excitations, on the other hand, both the forced and
natural responses are transients. (In an actual circuit the natural response must be
a transient, for otherwise we would have either a sustained or a growing response
without an external excitation.) Therefore the natural response assumes more impor-
tance, relative to the forced response, than it does in ac steady-state cases. In finding
376 Complex Frequency and Network Functions Chap. 14

natural responses in Chapters 8 and 9 we considered only first- and second-order

circuits, for the simple reason that our methods were applied to the describing differen-
tialequations and became more difficult to use as the order of the circuit increased.
However, as we shall see in this section, the natural response may be obtained relatively
easily from the phasor representation.
The results of the previous section illustrate that we may obtain the network
function quite readily from the describing equation. For example, if (14.22) is the
describing equation, then the network function (14.24) is

K } ~ V,(.y)
~ D(s)
(14.27)

with the numerator

N(s) = ^ b m s m + b m ^s m
~ l
+ • • •
+b s + x
b (14.28)

and the denominator

D(s) = = a n s" + ff„_,.s"~

I
+ • • •
+ a t
s + a (14.29)

The last two expressions are merely the right member and left member, respectively,
of (14.22) with the derivatives replaced by powers of s. For example, if the describing
equation is

fe + A d-^- +
1
3v = 2* +v t
(14.30)
dt dt dt

then the network function is

H(KS}
S) = Y£?} = 2s + l
(14
K 31)}
V,(s) s
2
+ 4s + 3 •

The process is reversible if there has been no cancellation of common terms in

the network function. For example, it is a simple matter to write down (14.30), given
(14.31).
In general, from (14.28) and (14.29) we may reconstruct the describing equation
(14.22). In the special case of (14.29) equated to zero

D(s) = (14.32)

replacing powers of s by corresponding derivatives of v results in the homogeneous

equation of the system

1
d"v„ ,
d"~ v . . dv, .
„, a

as discussed in Chapter 9. Therefore (14.32) is the characteristic equation, and its

roots are the natural frequencies of the circuit. Since these roots, by (14.29), are also
Chap. 14 Complex Frequency and Network Functions 377

the poles of the network function, we see that the natural response of the circuit is

vn = A x
e** + A 2 ep " + • • •
+ A n e" (14.33)

where the natural frequencies/),. />, pn are the poles of the network function and
A lt A 2 4„ are arbitrary constants. Modifications, of course, must be made, as
described in Chapter 9, if the natural frequencies are not distinct.
We now have a very simple method, based on phasors, for finding the complete
response of a circuit. All we need to find is the network function from which by (14.27)
we may obtain the output phasor. The forced response is found from the phasor

response in and the natural response is given by (14.33), where the

the usual way,
natural frequencies are the poles of the network function.
If the input phasor is of the form

where Vm and are constants, then, by (14.27),

V.(j) = (Fw /#H(j)

Therefore since V,(.v) has no poles, the poles of H(s) are the poles of V (^). Thus the
complete response v (t) may be obtained from its phasor representation \ (s). The
forced response is obtained as before, and the natural frequencies are the poles of
y„(s), from which the natural response may be constructed.
As an example, let us find the complete response /'
of the circuit of Fig. 14.8. Using
current division, the phasor I(s) of i is given by

=
\ e {s) 3s
!(*)
12 + {{3s{2s + 6)]/(3j + 2s + 6)} 3s + 2s + 6
(s + \)(s+ 12)

since Vg = 2/0°. Since the poles of I are s = — 1, — 12, the natural response is

i n (t) = A,e~< + A 2 e~ 12 '

A

i2n

2t
2e" cos t ^4>\ 3H 2H

FIGURE 14.8 Second-order circuit

378 Complex Frequency and Network Functions Chap. 14

For 5 = —2 + j\, the phasor representation is

1 = 0.16 /12.7°
so that the forced response is

ijf) = 0A6e- 21 cos (t + 12.7°) A

The complete response is therefore

/(,) =A x
e-' + A 2 e~ xlt + 0.16e" 2 cos '
(t + 12.7°) A
The arbitrary constants may be evaluated, as in Chapter 9, if we know the initial
energy conditions.

EXERCISES
14.6.1 Find the natural response in Ex. 14.4.1, assuming that there is no cancellation
in the network function. Ans. e~\A cosx
t + A2 sin t) V
14.6.2 Find the complete response in Ex. 14.4.1 (a) if the natural response is as given
in Ex. 14.6.1 and

Wo(0+)==
«n =
Ans. 3e"'[cos / - 3 sin + «/2 cos (3r - 135°)] V
/

14.6.3 Find the complete response v (t) in Ex. 14.4.2 if v {t) = e~' cos V and t
g

lo{K)
'
~ dt ~ dt 2 ~ u

(Note that the natural frequency s = —3 has been cancelled in the network
function. Thus the corresponding term must be added to the natural response.)
Ans. -le~' + le'
1' - \e~*' + (1/V2>~ r
cos (/ - 135°) V

14.7 NATURAL FREQUENCIES*

As we have seen, the natural frequencies are the poles of the network function if

there is no cancellation of a common pole and zero. Also, as discussed in Chapter 9,

the natural frequencies are the same
any response in a given circuit unless one
for
portion of the circuit is from the rest. Thus if we are looking only
physically separated
for the natural frequencies, we may consider any response, and it is obviously better
to choose one which is easier to find.
 9

Chap. 14 Complex Frequency and Network Functions 379

For example, let us consider the circuit of Fig. 1 4.8 with the source killed (since
the natural response corresponds to a zero source). The natural frequencies are the
poles of any network function. Therefore we may some proper
excite the circuit in
manner and, some chosen output, determine the network function. Figure 14.
for
illustrates the two proper means of applying an excitation to the dead circuit. We may
insert a voltage source in series with an eicment as in x-x' of Fig. 14.9(a), or we may
place a current source across an element, as across y-y' of Fig. 14.9(b). The first entry
into the circuit is sometimes called the pliers entry since we cut a wire and insert a
voltage source. The second entry is a soldering entry since we solder the source across
two nodes.

(a)

Figure 14.9 Circuit with two possible entries

In the pliers entry, if V^s) is the voltage source inserted and I x (s), the current into
the circuit at the source, is the response, then the network function is

Y,(J)
Us)
V x (s)

Thus the natural frequencies are the poles of \ x {s) or the zeros of its reciprocal, Z^),
the input impedance at x-x'. In this case we have

12(3j)
Z x (s) = 2s + 6 + 12 + 35

2(s + l)(s+ 12)

s + 4

Therefore the natural frequencies are —I, —12, as before.

In the case of Fig. 14.9(b), if a current source l y(s) is placed across y-y' and the
voltage V y(s) across the source is the output, then the network function is

Z,(*) yx*)
Us)

Therefore the natural frequencies are the poles of Z y (s) or the zeros of its reciprocal,
Y y (s), given by
380 Complex Frequency and Network Functions Chap. 14

(s + \)(s + 12)
125(5 + 3)
Again, the natural frequencies are — 1, —12.

EXERCISES
14.7.1 Find the natural frequencies of the circuit of Fig. 14.5 by killing the sources
and using (a) a pliers entry in series with the capacitor and (b) a soldering entry
across the capacitor. Arts. —3 ±jl
14.7.2 Find the natural frequencies of the circuit of Ex. 14.4.2 by killing the source
and using (a) a pliers entry in series with the capacitor and (b) a soldering entry
across the capacitor. (Note: Do not cancel the common pole and zero,
s = —3.) Ans. —1, —2, —3

14.8 TWO-PORT NETWORKS

One of the most important applications of the network function concept
is to networks

for which the input and output measured at different pairs of terminals.
signals are
The simplest and most-often-encountered circuit for which this is possible is the two-
port network, a port being defined as a pair of terminals at which a signal may enter
or leave. A general two-port network is symbolized by Fig. 14.10(b), as contrasted
with the one-port network of Fig. 14.10(a).
In general, as seen in Fig. 14.10(b), a two-port network has four terminals. It is
possible, of course, for two of the terminals to be the same, in which case we have a
three-terminal network. A general example of this case is shown in Fig. 14. 1. 1

We may associate two pairs of currents and voltages with a general two-port
network, as shown in the frequency-domain case of Fig. 14.12, with variables V^s),
I,(s), \ 2 {s), and l 2 (s), as indicated. In case the network is linear, these variables may

be related in a number of ways. For example, if l and I 2 are inputs and V! and V 2
l

are outputs, we know by superposition that V, and V 2 each have components pro-

o o o
o o

o o o

(o) (b) o * o

FIGURE14.10 (a) One-port and (b) two-port network FIGURE 14.11 Three-terminal
two -port network
 ;

Chap. 14 Complex Frequency and Network Functions 381

portional to 1, and I : . That is.

V, z 1
1
' i i
z i 2 '
(14.34)
V, *2|I| z 22':

where the z's are proportionality factors, which in general are functions of s.

Since the z's multiply currents to yield voltages, they must be measured
in ohms.

Therefore they are impedance functions. We may

from the network by find the z's
open-circuiting either port (with variables V, and I,) or port 2 (with V 2 and I 2 ). For
1

example, if port 2 is open-circuited (I, = 0), then from (14.34) we have

_V,
I, I2 =
(14.35)

i,

Similarly, if port 1 is open-circuited (I, = 0), we have

_V,
I, =
(14.36)

Accordingly, the z's are called open-circuit impedances, or open-circuit parameters,

or simply z-parameters. In any case, they are examples of network functions.
As an example, let us consider the three-terminal network of Fig. 14.13. Because

II I
2

z Z2
+ l

v l
z3

— i

FIGURE 14.12 General two-port FIGURE 14.13 J network

network

of its shape, it is sometimes called a T network. Evidently, it is simply a Y network,

as discussed in Chapter 13. The reader may easily show, using (14.35) and (14.36), that

i
— Z,

*2I Z3 (14.37)

z,

If in Fig. 14.12 Vj and V 2 are inputs and I, and I2 are outputs, we may obtain, in
an analogous manner,
 1

382 Complex Frequency and Network Functions Chap. 14

(14.38)
*2 == y21* 1 ~T~ V 22'2
where, evidently,

y.r

y !2
v 2

I2
y 2 i-y^

I2
y22 — tt
v, = o

Accordingly, the y's are short-circuit admittances (V, = or V2 0), or short-circuit

parameters, or simply y-parameters.
As an example, let us find the ^-parameters of the three-terminal circuit of Fig.
14.14. Because of its shape, it is sometimes called a n network. Evidently, it is simply
a A network, as considered in Chapter 13. Using the expressions for the >'-parameters,

we may write, by inspection,

yiI =Y +Y a 4

yi2 =y 2 (14.39)

Y +Y 6 c

The z- and >'-parameters are but two sets of parameters associated with a two-port
network. Others include the hybrid parameters and the transmission parameters, which
willbe defined and considered in the exercises and problems.
The two-port parameters may be used to obtain various network functions. For
example, if port 2 is open (I 2 = 0), then from (14.34) we may find the voltage ratio
function

Xa = hi (14.40)
v 1
z i 1

As another example, let us find the voltage ratio function for the two-port loaded
with 1 f2, as shown in Fig. 14.15. Evidently, we have V = —
2 2, which substituted into

T 'b T °

Ya Yc

o • * o

FIGURE 14.14 n network FIGURE 14.15 Two-port loaded with a 1-Q.

resistor
Chap. 14 Complex Frequency and Network Functions 383

the second equation of (14.38) yields, after some rearrangement,

v 2= -y 2

V, " + 1 y2
,

(14.41)

Another use for the two-port parameters is in the construction of equivalent cir-
cuits. For example, in the first equation of 14.34), V, is a sum of two terms, z, ,1, and
(

z !2 I : The first may be obtained by an impedance z M carrying a current I,, and the
.

second ma\ be obtained by a dependent voltage controlled by I 2 In a similar way we .

may interpret the second equation of (14.34) and draw the equivalent circuit. The
result, as may be verified by inspection, is shown in Fig. 14.16. We say that this circuit
is equivalent (that is, at the terminals) to that of Fig. 14.12, because they both have
the same two-port parameters. Thus if the port currents are the same, then the port
voltages are the same, and vice versa, for both networks.

FIG U R E 1 4.1 6 Equivalent circuit ol Fig. 14. 12

If z 12 = z 21 (or, equivalently, y
l2
= y 21 ), the network is said to be a reciprocal
network. If the elements comprising the network are resistors, inductors, capacitors,
or independent sources, then the network is always reciprocal. An equivalent circuit
for a general reciprocal three-terminal network shown in Fig.
is 14.17. There are, of
course, many other equivalent circuits that may be drawn using the various network
parameters.

Z II" Z I2 Z 22~ Z I2
+

Z|2 v2

o < I

FIGURE 14.1 7 Equivalent circuit ol a general reciprocal three-

terminal network
384 Complex Frequency and Network Functions Chap. 14

EXERCISES
14.8.1 Draw an equivalent circuit for a two-port network with parameters z u = 6 Q,
z i2 = z 2i = 4 Q, and z 22 = 10 Q.

14.8.2 Show that the ^-parameters may be obtained from the z-parameters by

Z^ > Zi i

Zi i Zi i

where A = z n z 22 — Zi2 z 2i- Use this result to find the ^-parameters of the
circuit of Ex. 14.8.1. [Suggestion: Solve (14.34) for the voltages, and compare
the result with (14.38).] Ans. y n = ^, y 12 = y 2 i
= - 77, Y22 = ^U
.8.3 Two sets of hybrid parameters, h n , h 12 , h 21
2I , h 222 and g u g 12) g 21 g 22
, , , are
defined by
V, =^,1, +h 12 V 2

I2 = h 2 ili + h 22 V 2

and
Ii =guVi + g ia I a
V 2 =gaiV, + g 22 I2

Find the A-parameters and the ^-parameters of the network of Ex. 14.8.1.

[Suggestion: Solve (14.34) for V\ and I2 in terms of \ x

and V2 , etc., and com-
pare results.] Ans. h's: ffi, -fo,
— -fe, -jL

„-„. 1 _4 4 44

PROBLEMS
14.1 Find the phasor representation \(s) of (a) v(t) = be'' cos {At + 30°), (b)

u(0 = 8e~ 9 ', and (c) «(f) = e~ 2r (4 cos 3? + 3 sin 3/).

14.2 Find v(t) if (a) \(s) = 6/10° with j = -3 + j 8, (b) VQy) = 5/0° with s = -10,
and (c) V(s) = 4 + /3 with 5 = -1 1/2.

14.3 Find the complex frequencies which characterize (a) v = e~'(cos It + sin 2/)
and (b) v = e~ l <
+ e~ 4 ' + 2 cos t.

14.4 Find the impedance Z{s) and locate its poles and zeros. If an excitation
vg = 6e~' cos It V is applied at the terminals, find the forced component of
the current it delivers.
Chap. 14 Complex Frequency and Network Functions 385

in I H
AMA-

Z(s) F -±- i a

PROBLEM 14.4

14.5 Repeat Prob. 14.4 for the circuit shown.

60
—W^~
Z(s)- -4 F ~7 F = Fi
F

PROBLEM 14.5

14.6 Find in Prob. 9.6 the network functions

his) U(s)
H,
V,(j)
H 2 (.v)
~ v '
'
V ? U)

Use these results to find the forced responses /, and i2 if

v g (t) = \0e-' cos 5/ V

14.7 If the current source in Prob. 9.13 is i

g
(t) = 9e~' cos 3t A, find for / > the
network function l(s)/l g (s) and the forced response i(t). (Suggestion: Use the
proportionality principle.)

14.8 If the voltage source in Prob. 9.18 is r e (t) = 6e~' cos It V, find the network
function l(s)/V g (s) and the forced response i in each case.

14.9 In Prob. 14.7, replace everything in the circuit except the 5-H inductor by its
Thevenin equivalent circuit. Use the result to find the network function and
the forced response.

14.10 In Prob. 9.19, let the current source be i g (t) = Ae' 1 cos At
'
A and C= \ F.
Replace the corresponding phasor circuit, except for the series combination of
1 H and 1 Q, by its equivalent Thevenin circuit. Use the result to obtain the
network function I(s)/l g (s) and the forced response /(/).
14.11 Find the network function \(s)/\ g (s) in Prob. 9.21. Use the result to find the
forced response dt).

14.12 Find the network function \(s)IV g (s) in Prob. 9.22. Use this result to find the
forced response v(t) if v g (t) = 6e~ 2 ' cos At V.

14.13 Find the network function and the forced response v(t) if v g U) = 6e~ 2 V.
'
386 Complex Frequency and Network Functions Chap. 14

in h h
—VW—
i
i

Offiff
v-T rffWv
i

«.">© 2F;±: a v(t)

PROBLEM 14.13

14.14 Given the network function

H(*)
V (s) 4s(s + 2)
V,(s) (* + !)(* + 3)

If no cancellation has occurred, find the complete response v (t), for / > 0, if
v t (t) = 6e~' cos It V and v (0 + ) = dv o (0 + )/dt = 0.
14.15 Repeat Prob. 14.14 if

H(s)
4s(s + 2)
s2 + 2s + 2

14.16 Repeat Prob. 14.14 if i>,-(f) = 6e ' V. (Suggestion: Observe the describing differ-
ential equation.)

14.17 Find the complete response ?>(/), for t > 0, if there is no initial stored energy.

PROBLEM 14.17

14.18 Find the network function V (j)/V,-(*) and the forced response if (a) v t (t)
= 6e~ 2t cos / V and (b) v (t) = 6 cos / V. t

14.19 Find the complete response v(t), for t > 0, if v(0) = 0.

14.20 Find the network function, the natural frequencies, and the forced response
v if i
g
= 5e~ 2 cos 2/ A.
'

14.21 Find the network function \(s)/\ g (s) in Prob. 9.23 and the forced response
v(t) if v g (t) = 6e~ 2 cos
' / V.

14.22 Find the complete response v(t), for / > 0, in Prob. 14.21 if there is no initial

stored energy.

14.23 Considering the figure for Prob. 14.17 as a two-port network with terminals as
shown, find the z-parameters as functions of s.
Chap. 14 Complex Frequency and Network Functions 387

PROBLEM 14.18

PROBLEM 14.19

PROBLEM 14.20

14.24 Show that the hybrid parameters defined in Ex. 14.8.3 may be obtained from
the z-parameters by

A *12
h h 12 =
z2 2

h 21 = -m. h 22 = J_
*2 2 Z2 2

gi; gl2 =

o
62 1
-5*1,
—~Z ll
' 822 = —A
z ll

where A f= Z11Z22 ~ z i2 z 2i-

14.25 Find the y-, h-, and ^-parameters of the two-port network of Prob. 14.23.
 .

388 Complex Frequency and Network Functions Chap. 14

14.26 A set of transmission parameters, A, B, C, and D, is defined for the two-port

network of Fig. 14.12 by

V, = AV 2 - BI 2

I, = CV - 2 DI 2
Show that

B
Z21

C= —
Z:i
D Z 22
Z2I

where A is given in Prob. 14.24. Note that the transmission parameters relate
the output variables V2 , I 2 to the input variables V,, I,.

14.27 Find the transmission parameters of the two-port network of Prob. 14.23.

14.28 Derive the Y-A transformation of Chapter 13 by making Figs. 14.13 and 14.14
equivalent at the terminals (that is, equal two-port parameters).

14.29 Find the ^'-parameters of the network shown. Ferminate the output port with
a 1-fi resistor, and find the resulting network function V 2 /Vj.

i,/2 V
in
J VW-

I F

PROBLEM 14.29

14.30 Show that for Fig. 14.15, in terms of the transmission parameters we have

V, 1

V, A +B

14.31 Check the result for the terminated network of Prob. 14.29 by using the result
of Prob. 14.30.
 15
FREQUENCY RESPONSE

Frequency-domain functions are very useful, as we have seen, for finding corre-
sponding time-domain functions. The frequency response of a circuit, however, is

extremely useful in its own right, as we shall see in this chapter. For example, if we are
interested in which frequencies are dominant in an output signal, say Y(jco), then we
need only consider the amplitude |V(yo))|. The dominant frequencies correspond to
relatively large amplitudes, and frequencies that are virtually suppressed correspond to
relatively small amplitudes.
There are many applications in which frequency responses are important. One
very common application is in the design of electric filters, which are networks that
pass signals of certain frequencies and block signals of other frequencies. That is, if the
output signal of the filter has amplitude |
V(y'co) |, then co, passes if |
V(y'«,) |
is relatively
large and is blocked if V(y"a>,) is relatively small. There are many examples of electric
| |

filters in our modern society, some of the more common being those in our television
sets, which allow us to tune in a certain channel by passing its band of frequencies
while filtering out those of the other channels.
In this chapter we shall consider frequency responses, both amplitude and phase.
We shall also define resonance and quality factor and show how they are related to the
frequency responses. Finally, we shall consider methods of scaling networks to yield a
given frequency response with practical circuit element values.

15.1 AMPLITUDE AND PHASE RESPONSES

A network function H(jco), as well as any phasor quantity, is in general a complex

function, having a real and an imaginary part. That is.

389
 )

390 Frequency Response Chap. 15

H(j(o) = Re HO©) + ./ Im H(jco) ( 1 5. 1

As we know, we may also write the network function in the polar form

H(;co) = |HO'cu)|^'* to '

(15.2)

where H(jco) | |
is the amplitude, or magnitude, response and 0(co) is the phase response,
given, respectively, by

HO'cj)| = VRe 2
H(j(o) + Im 2 H(j<o) (15.3)
and

= Im H(jco)
0(a)) tan i
(15.4)
Re H(jco)

The amplitude and phase responses are, of course, special cases of frequency responses.
As an example, suppose the network function of the RLC parallel circuit of Fig.
15.1 is the input impedance

FIGURE15.1 RLC parallel circuit

H(s)
v 2 (.)
Z(n) _ l

Us) -
"*> sC
(15.5)
(\/R) I | (IM)
or

Ks)-
1
'?* '',_ (15.6)
, , "i

For s = y'o) we have, from (15.5),

HO) =
(l/R)+j[coC-(l/coL)]
I

(15.7)

so that the amplitude and phase responses are

Hijco) (15.8)
V(l/fl 2 H [a)C-(l/a)L)] 2

and

0(co) -tan" r(coC 1

- -~\ (15.9)

Since R, L, and C are constants, the maximum amplitude occurs at the frequency
to = a) for which the denominator in ( 1 5.8) is a minimum. Evidently this occurs when
Chap. 15 Frequency Response 391

coC - -V =
or

COn (15. 10)

JLC
Thus
|H(yo))| ni . x = |HO© () )| = /?

Also, it is clear that |H(/co)| —O as co — and co — oo. Therefore the amplitude

response has the form shown in Fig. 15.2(a). In a like manner we may sketch the
phase response, shown in Fig. 15.2(b). since <f>(o) ) = 0, 0(co)— n/2 as co - 0, and —
(f>(co) -— —n/2 as co — oo.

H(ju>)

/lc
(q) (b)

FIGURE 15.2 (a) Amplitude and (b) phase response of (15.7)

If the input of Fig. I5.l is the time-domain function

i{t) = I m cos cot

then the input phasor is I = /m /0°, and the output phasor is V2 = Im Z = / m H. Thus
the amplitude of the output is simply that of the network function multiplied by a
constant. Therefore we may obtain as much information from the network function
response as from the output response. For this reason and the reason that the network
function depends only on the network and not on how it is excited, we shall usually
consider the frequency response of the network function.

EXERCISES
15.1.1 Let R = 2Q, L = \ H, and C = \ F in Fig. 15.1, and find the maximum
amplitude and the point where it occurs. Also, sketch the amplitude and phase
responses. Ans. |H| max = 2, co
392 Frequency Response Chap. 15

15.1.2 For the RLC series circuit with a voltage source v g , let the network function
be H = I/Vg, where I is the phasor current. Show that the amplitude and phase
responses are similar to those of Fig. 15.2 with |H| max = l//?andco = l/*/LC.

15.1.3 Let the network function of Fig. 15.1 be H= IJli, where I L is the inductor
phasor current directed downward. Show that

\\LC
H(.r)
s* +(\IRC)s + (1/IC)

and that |H| 1, occurring at O) = 0, provided

2R 2 C<L

15.2 FILTERS

With reference to Fig. 15.2(a) we see that frequencies clustered around co = \/*/LC
rad/s, or/ = \/(2n^/LC)Y\z, correspond to relatively large amplitudes, while those
near zero and larger than co correspond to relatively small amplitudes. Thus Fig. 15.1
is an example of a bandpass filter, which passes the band of frequencies centered around
eo . In the general amplitude case, shown in Fig. 15.3, we say that co , the frequency at

'I
' 2

FIGURE 15.3 Genera/ bandpass amplitude response

which the maximum amplitude occurs, is the center frequency. The band of frequencies
that passes, or the passband, is defined to be

co Cl <co < co t ,

where and co are called the cutoff points and are defined as the frequencies at
co Cl c ,

which the amplitude is l/^/T = 0.707 times the maximum amplitude. The width of
Chap. 15 Frequency Response 393

the passband, given by

B= o) c <°c, (I5.ll)

is called the bandwidth.

As we shall see in Sec. 15. 4. the bandwidth in the case of Fig. 15.1 is B = \jRC.
\s another example, let us consider the circuit of Fig. 15.4. Analyzing the circuit,

2A
v.o V\AA

FIGURE 15.4 Low-pass filter

we may readily obtain the voltage-ratio function.

V 2 (j) _ 2
His)
V,(J) 2s + 2

Letting s — jco and calculating the amplitude response, we have

H(;w)|
V(2 — co
2
)
2
+ 4cu
2

or, after simplification,

H(;co)j (15.12)
VI -(co* 14)

The amplitude function continuously decreases as co increases, because its

numerator is constant and its denominator continuously increases with frequency.

Therefore the amplitude response attains its maximum of |H| max = 1 at co — and
thus has the shape of Fig. 15.5(a). From this we see that the circuit of Fig. 15.4 is a
low-pass filter. That is, it passes low frequencies (relatively large amplitudes) and
rejects high frequencies (relatively small amplitudes).
There is only one cutoff point, as indicated in the figure. This follows from the
definition
394 Frequency Response Chap. 15

H(joo)

(a) (b)

FIGURE 15.5 Low-pass frequency responses

1
H(M)I HOa>)|

(1)

VI +(co 4 /4) c

whose only real answer is co c 2. Thus the band of frequencies which passes is the
low-frequency band
< (o < ^fl
The phase response for Fig. 15.4 may be easily shown from H(jco) to be

2co
<f>(co) = — tan" z
co

which is sketched in Fig. 15.5(b).

An example of a passive, low-pass filter is that of Fig. 15.1, where the network
function is as defined in Ex. 1 5. 1 .3. In fact, if R= 1 Q, L = 1 H, and C = £ F, the
network function is the same as that of Fig. 15.4.
There are many types of filters other than low-pass and bandpass. Two of the more
common are high-pass filters, which pass high frequencies and reject low frequencies,
and band-reject filters, which pass all frequencies except a single band. Typical

FIGURE 15.6 (a) High-pass and (b) band-reject amplitude response

Chap. 15 Frequency Response 395

amplitude responses are shown in and examples are considered in Exs.

Fig. 15.6,
15.2.2 and 15.2.3. In Fig. 15.6(a), co c is the cutoff point, and the passband is a co c .

In Fig. 15.6(b). co is the center frequency of the rejected band of bandwidth

B = (o c.
- <» fl
.

EXERCISES
15.2.1 Show that

2v
H(.v)
s2 + 0.2s + 1

is the network function of a bandpass filter and find co , co Ct , co CQ , and B.

Ans. 1, (T0.2 + */4tt)ll = 0.905, 1.105; 0.2
15.2.2 Show that
Is 2
His)
s ! 0.5

is the network function of a high-pass filter, and find |

H(/a>) max and CO c
|
.

Ans. |H| max =2, co c = l/^TZ

15.2.3 Show that

H(.v)
3(.v
2
+ 25)
s + 25

is the network function of a band-reject filter, and find |H(/co)| max , co , co Cl ,

and co, .

Ans. |H| max = 3, co = 5, w c „ r! = (=F 1 + \/T0T)/2 = 4.525, 5.525

15.3 RESONANCE

A physical system which has a sinusoidal type of natural response reacts vigorously,
and sometimes violently, when it is excited at, or near, one of its natural frequencies.
This effect may have been noticed by the reader in Sec. 9.6, particularly in the case of
Ex. 9.6.3. The system in this respect is somewhat like all of us. When urged to do what
it naturally wants to do, it responds with enthusiasm.
phenomenon is known as resonance, and its side effects may be good or they
This
may be bad. As an example, a singer may break a crystal goblet with his voice alone by
properly producing a note at precisely the right frequency. Also, a bridge may be
destroyed if it is subjected to a periodic force with the same frequency as one of its
natural frequencies. This is why no thoughtful troop commander will march his men
in step across a bridge. On the other hand, without resonance there could be no
electric filters.
We shall define a sinusoidally excited network to be in resonance when the ampli-
tude of the network function attains a pronounced maximum or minimum value. The
396 Frequency Response Chap. 15

frequency at which this occurs is called the resonant frequency. As an example, the
RLC parallel circuit of Fig. 15.1 is in resonance when the frequency of the driving
function is co = \\J LC. This was shown in Sec. 5. where it was demonstrated that 1 1

the maximum network function amplitude occurred at co The amplitude response of .

Fig. 15.2(a) is typical, with its peak at the resonant frequency. The
relatively high
parallel RLC circuit is so important that the term parallel resonance is reserved for its
resonant condition.
The reader may recall encountering the term resonant frequency earlier in Sec. 9.8

in connection with the underdamped case of the parallel RLC circuit. The two fre-

quencies, there and here, are exactly the same.

The natural frequencies of the parallel RLC circuit are the poles of the network
function, given from (15.6) by

•Ti.2 = —a ±Jco d (15.13)

where

(15.14)
2RC
and

co, 'col or (15.15)

From the pole-zero plot, shown in Fig. 15.7, we see that the resonant frequency &> ,

or s = jco . is very near the natural frequency s = —a + jco d, since, by (15.15), co is

the radius of the dashed semicircle. If R is made larger, then a is made smaller and the

— jCO

jw d
j^o

' l\
\
/
/ V \
/ \
\ (* t
1
f

1
-a a

\
\
M-~:
FIGURE 15.7 Pole-zero plot of the parallel RLC circuit
Chap. 15 Frequency Response 397

resonant frequency is even closer to the natural frequency. In this case, as is clear in
peak is more pronounced. Of course, if R were infinite (open-circuited),
Fig. 15.2(a). the
then the resonant frequency would coincide with the natural frequency and the
amplitude would be infinite.
Before leaving the parallel RLC circuit, let us note that the network function
is actually the input impedance, as stated in (15.5). The resonant frequency
co = co = 1/+/LC is the frequency for which the input impedance is purely real, as
seen in (15.7). Indeed, many authors define the resonant frequency precisely this way in
the case of a two-terminal network. In the general case, maximum amplitude does not
always occur exactly at the frequency of real impedance, but usually there is very little

difference.
In the case of the RLC series circuit excited by a voltage source \g , if the phasor
current I is the output, then

HO"*,) = #M
V
=
g
(jco)
Y(Jto)
KJWJ
~
R + j[coL -
l

(1/©C)]

where Y is the input admittance seen at the source terminals. Evidently series resonance
occurs at co = 1/+/LC, yielding a maximum amplitude of \jR. As in the parallel
resonance case, the resonant frequency is also the frequency of real input impedance or
admittance. At resonance the effect of the storage elements exactly cancels, and the
source sees only the resistance.

EXERCISES
15.3.1 Show that the resonant frequency in the case of the function of Ex. 15.2.1
coincides with the frequency for which the function is real.

15.3.2 Find the resonant frequency for the parallel RLC circuit described by (a)
R= 2 kQ, L = 4 mH, and C = 0.1 //F; (b) co d = 5 rad/s and a = 12 Np/s;
and (c) a = 1 Np/s, J? = 4 Q, and L = 2 H.
Am. (a) 50,000, (b) 13, (c) 2 rad/s

15.3.3 In Ex. 15.3.2 (a), find the amplitude of the voltage across the combination if

the current source is i

g
= cos cot mA and co is (a) 10 4 , (b) 5 x 10 4 and , (c)

25 x 10 4 rad/s. Am. (a) 0.0417, (b) 2, (c) 0.0417 V

15.4 BANDPASS FUNCTIONS

AND QUALITY FACTOR
The network function (15.6) is a special case of the general second-order bandpass
function

H(5 > = s+t + i « 1516)

398 Frequency Response Chap. 15

where K, a > 0, and b > are real constants. (The function is called second-order
because of the quadratic denominator.) To see that the function is of the bandpass
type, let us consider its amplitude

A'col
HO) |

J(b -
I

(o
1
)
1
-I- a 2 co 2

\K\

V« 2
I Kb - 2
(o )/coY

Evidently the maximum value is

HOeo) m .* = J —
a
l

occurring at the center, or resonant, frequency co , satisfying

b = a>l (15.17)

At a cutoff frequency co c , we must have

HO'COj = H ( •/<*>) Lax

I |

-jY I

or
\K\ \K\
J a1 + [(b - 2
co )/co c Y J2a
which evidently holds if

b — co
2

= ±o ,

Thus we have
co
2
± aco c - b = (15.18)

which, because of the double-sign possibility, has four solutions. Using the positive
sign, we have, by the quadratic formula.

J
Q. t
= ~° +
f+ 4b
(15.19)

We have discarded the negative sign on the radical because this gives a negative cutoff

frequency. Using the negative sign in (15.18) and again suppressing the negative root,
we have the other cutoff frequency,

- U + V + 4b
(15-20)
<o ct
f
Evidently, from (15.19) and (15.20) we have the bandwidth,

B = co c , - co cl = a (15.21)
Chap. 15 Frequency Response 399

Thus in view of (15.17) and (15.21) we may write the network function as

S
H (*) = o 2 ( ' 5-22)
s
2
z
+,

Bs + g>o

which is the general network function of a second-order bandpass filter having center
frequency co and bandwidth B. The amplitude response is shown, of course, in Fig.

15.3.
Another result worth noting from (15.17), (15.19), and (15.20) is

col = co Cl co c , (15.23)

which demonstrates that the center frequency co is the geometric mean ^/co Cl co c , of
the cutoff points.
A good measure of selectivity or sharpness of peak in a resonant circuit is the so-
called quality factor Q, which is defined as the ratio of the resonant frequency to the
bandwidth. That is.

2= ^ (15.24)

(The letter Q is symbol for reactive power, as the reader will recall. However,
also our
the two quantities will in the same context so there should be no confu-
never be used
sion.) Evidently, since B = co /Q, a low Q corresponds to a relatively large bandwidth,
and a high Q (sometimes arbitrarily taken as 5 or more) indicates a small bandwidth,
or a more selective circuit.
With this definition of Q we may write (15.22) in the form

HOO = --*£- , ( 1 5.25)

s
2
+ (coJQ)s + col

so that we see at a glance the center frequency, the quality factor, and the bandwidth.
As an example, the filter described in Ex. 5.2. 1 1 is a bandpass filter with co = 1 rad/s,
B= 0.2 rad/s, and Q = 5. Another example is the parallel RLC circuit with the
network function given in (15.6). In that case, co Q = \J«JLC, B = l/RC, and Q=
co IB, which has the equivalent values

Q- co RC
= R^ (15.26)

=~ _R_
oi a L

We should mention that the quantity we have called Q is defined as selectivity by

some authors, who reserve for Q the definition
 = — —
400 Frequency Response Chap, 15

~ Total energy stored at resonance ,,.„,

Q= t
2n c
Energy ;r r^i per cycle
dissipated
!
— i i
at resonance
(15.27)

In the examples we have considered, the two definitions are the same, as the reader is
asked to show in Exs. 1 5.4. and 5.4.2. However, in general, there is a slight difference.
1 1

Finally, let us consider the effect of Q on resonance. Incorporating (15.19) and

(15.20) into one equation and replacing a and b by their values, we have

».....=
^!+V^ + (i)
or

».„.. (t^q + V(5J + wo (i5 28)

-

Evidently if Q is high, then we may neglect (\J2Q) Z in comparison with 1, and write

©cc^T^+Wo
2Q
or, approximately,

co Cl = co - B ,

(15.29)
S

Thus as £> increases, the amplitude response approaches arithmetic symmetry. That is,

bandwidth above and below the center frequency.

the cutoff points are half the
In the example of Ex. 15.2.1, we have o) = 1 and Q = 5, which we consider as
high. Thus by (15.29), the approximate cutoff frequencies are co = 0.9 and cl

co C2 =1.1 rad/s. The exact values are

a, cuc2 = 0.905, 1.105

EXERCISES
15.4.1 For the RLC parallel circuit in resonance, show that the total energy stored is

w(t) = yR 2
CIi cos 2 o> t + ^| sin 2 co t

= ^ C/ 2 2

where the excitation is i = Im cos co t.

Chap. 15 Frequency Response 401

15.4.2 Show thai for the circuit of Ex. 15.4.1 the energy dissipated per cycle is

nRIl
RJ;„ cos 2 av dt a»>,

and thus by definition (15.27) and the result of Ex. 15.4.1 that

Q cOqRC

15.4.3 For the RLC series circuit with excitation vg = Vm cos cot V, show that
co = 1/v/IC and = co L//?.

15.5 USE OF POLE-ZERO PLOTS

The pole-zero plot of a network function may often be used to readily sketch the
frequency responses. To seehow this may be done, let us write the network function in
the form

His)
K(s - *,)(j -z 2
)...(s ,)
(15.30)
(s - p,)(.v —p 2 ). . .(s --/>„)

where, as before, the z's and and poles. Each of the factors is a
/>'s are the zeros
complex number of the form s —
and may be represented in the .v-planeby a vector,
5,

drawn from ,v, to as shown .v. in Fig. 15.8. This is true since by vector addition the

vector s is clearly the sum of the vectors s, and s — s v

FIGURE15.8 Vector representation of s - S\

The typical vector s — .?,, drawn from s l

to s, may be written in polar form where
its magnitude is its length and its phase is the angle it makes with the positive real axis.
If s = jco. the point s in Fig. 1 5.8 is on theyco-axis, and the factors of (15.30) may be
402 Frequency Response Chap. 15

written

j(o — z, = Nie J<", i = I, 2, . .

.
, m
jco —p — k Me k
iPt
, k = 1,2,...,/?

Therefore the network function is

H(jco) = \H(jco)\e^'" )

where the amplitude is

H(/©)| = KN N *
2 • •
N™ (15 31)

and the phase is

0(co) = (a, + «2 + .. . + aj - (fi t + )? 2 + . . . + fi„) (15.32)

both of which may be measured directly from the pole-zero plot.

Thus any point jco we simply draW vectors from all the poles and zeros to jco,
for
measure their lengths and angles, and calculate the amplitude and phase from (15.31)
and (15.32). Quite often we need only a few points to sketch the responses, as we shall
see.

As an example, suppose we have

H(j)
s
2
+ 2s + 401

which has a zero at and poles at —1 ±./20. These are shown on the pole-zero plots
of Fig. 15.9. By (15.31) and (15.32) we have

H(yo»)'
I

MM 1 2

<f>(w) = a - (/?, + fi t )

+
whose components are identified in Fig. 5.9(a). First we note that 1 if co varies from
to +oo, then a = 90°. For w = 0we have, from Fig. 15.9(b),

|M|-
1111 ^°>_ -Q
v/40V401
<j> = 90° - (-tan" 20 1

+ tan" 1
20) = 90°

For co = 20 [Fig. 15.9(c)] we have

_j*(20)_
'"'
(1)V'601
= 90° (0 | tan 1
40)^0

This point is in the region where the amplitude changes the fastest, since Af , = 1 is a
Chap. 15 Frequency Response 403

J20
y*.

MA

-I
f
^
M2 /

2
h\
-J20

(b)

(c) (d)

FIGURE 15.9 Steps in the construction of the frequency responses

(Figures not drawn to scale)

minimum and is changing percentage-wise much faster than N or M 2

. The amplitude
therefore will reach its peak near this point. Actually we know from our previous work
that co = v40l yields the peak amplitude of 2.

If co = co h , a very high value, say. such as 10 6 then , all three vectors will be essen-
tially vertical, so that

N ^ M ^ M ^ co
t 2 h

a ^ ^ £ ^ 90
/?, 2

Therefore we have |H| «= Ajco h a very small value, and

.
<f>
^ 90°.
404 Frequency Response Chap. 15

Sketching the functions, we have the forms of Fig. 1 5.2, as expected, with co «= 20
and |H| max ^ 2.
We may rough idea of the cutoff points from Fig. 15.9(d). In this figure
get a
M x
= J~1
/ is ', which
,J~1 times its value at the approximate peak, represented by
Fig. 5.9(c). Since
1
2 M
and N have changed by much smaller percentages, the amplitude
is then approximately \\^fl times its peak, so that co Cl 19. By a similar argument at ^
co = 21, we have co C2 «=21. The exact values, by (15.19) and (15.20), are
ffl CliC = 19.05, 21.05. In this example, by comparing the network function with
.

(15.25), we see that Q = 10.01. Thus we have a high Q so that the poles are very
near the j'co-axis. Thus cod is very near a> and the approximations we have made are
,

very near the exact values.

EXERCISES
15.5.1 Use the method of this section to sketch the amplitude and phase responses
for

S* + 4s + 2504

Ans. co ^ 50, Q ^ 12.5, 5 = 4, C0 C „ C! =fe 48, 52

15.5.2 Find the exact values for the answers given in Ex. 15.5.1.
Ans. 50.04, 12.51, 4, 48.08, 52.08

15.6 SCALING THE NETWORK FUNCTION

The reader may have noticed that in many of our examples we have used network
elements such as 1 CI, 1 F, 2 H, etc., which are extremely nice numbers to have when we
are analyzing a network. For example, in Fig. 15.4 we had elements of 2 Q, \ F, and
\ F, and the network was a low-pass filter with a cutoff frequency of ^J~2 rad/s, or
0.23 Hz. However, such element values as these are not very practical, to say the least,
and there is very little demand for filters that pass frequencies between and 0.23 Hz.
(As an illustration, a 1-F capacitor, constructed of two parallel plates 1 cm apart with
air as the dielectric, would require a face area of 1.13 • 10 9
m 2
.)

It would be an ideal situation if we analyze or design contain simple

the circuits
element values such as 1F, etc., while the circuits we build have practical element
values such as 0.047 //F, and useful characteristics such as cutoff points of 100 Hz. As
we shall see in this section, network scaling allows us to have it both ways.
We shall consider two types of network scaling, namely impedance scaling and
frequency scaling. To illustrate the former, let us suppose that the network function is
Chap. 15 Frequency Response 405

an impedance given by

Z'(s) = sL' + R! + ±
The network has been impedance-scaled by an impedance scale factor k t
if the imped-
ance Z(s) of the scaled network is Z(.v) = k,Z'(s). In other words, we must have

Zis)=k,(sL' + R'+±^ 1

= s(k L')l + k R' i + ^T }

C5.33)

If the impedance of the scaled network is

Z(s) = sL +R+ -^

then we see by comparison with (15.33) that

L --
= k tL' U
R =- k,R' (15.34)

C =
C
'
k,

summary, to impedance-scale a network by the factor A'., we multiply the L's

In
and by A-, and divide the C's by A,. We have illustrated this for a special case, but it
7?'s

may be shown to hold in general. Also, if there are dependent sources, the scaling is
accomplished by multiplying gain constants having units of ohms by A,, dividing those
with units of mhos by k and leaving unchanged those that are dimensionless. The
t,

scaling multiplies by A,., divides by A,, or leaves unchanged network functions that have
units of ohms or mhos or are dimensionless.
To illustrate, let us = 5. The 2-fi
impedance-scale the network of Fig. 15.4 by k t

become 2 x 5 = 10 Q, the £-F capacitor becomes £ ~ 5 = 0.1 F, and the

resistors
^-F capacitor becomes ^ -i- 5 = 0.05 F. The op amp, an infinite gain device, remains
an infinite gain device, and thus is unchanged. The network function V 2 (j)/V (j), 1

being dimensionless, is also unchanged.

To frequency-scale a network function by a. frequency scale factor k f we simply
,

replace s by s/k f That is. if the unsealed network has the network function H'(5), then
.

the scaled network function H(.s) is obtained by letting 5 s/k f resulting in = ,

H(j) = H'(£.) (15.35)

Thus if the unsealed network had a property, such as center frequency, when S =jl,
then the scaled network has this property at s = jk f This is clear from (15.35), which
.
406 Frequency Response Chap. 15

gives

HC/M = H'(yi)

Another way to consider frequency scaling is to note that s = k f S, so that if

s = jco corresponds to S = j Q, then co = k f Q. Thus the values on the frequency axis

have been multiplied by the scale factor A7, without affecting values on the vertical
axis of a frequency response.
The scaling of the network to effect the transformation of (15.35) is quite simple.
If Z'(S). given by

Z'(5) = SV + R' + ^L

is any impedance in the unsealed network, then the corresponding impedance in the

scaled network is

Z(s) = sL +R J_
sC

= s iv) + R + '

57c7*7)

Comparing these results, we have

-I
=
R R' (15.36)

Therefore, to frequency-scale a network by a factor A7, we divide the L's and C's by k f
and leave the 7?'s unchanged. If there are dependent sources with constant gains, these
also are left unchanged.
To illustrate, the circuit of Fig. 15.4 is a low-pass filter with a cutoff frequency of
co c =
»J~2 rad/s, as we have previously seen. Suppose we want to frequency-scale the
network so that the cutoff frequency is 2 rad/s. Then the scale factor k f is given by

y/2k f = 2

or k f =
*J~2 Dividing the Cs in Fig. 15.4 by A7, we have, in the scaled network,
.

capacitances of lf2^/~2 F and 1/4^/T F. The rest of the circuit remains unchanged,
there being no inductors to scale.
Of course, we may perform both impedance and frequency scaling on a network.
To obtain a network with a practical property such as center or cutoff frequency, we
may first apply frequency scaling with the proper factor k f Then to make the resulting .

element values more practical we may impedance-scale by a factor k For example, if t

.
Chap. 15 Frequency Response 407

the parallel RLC network of Fig. I5.l contains a l-Q resistor, a 2-H inductor, and a

i-F capacitor, then for the network function of (1 5.5) we have an amplitude response
as in Fig. 15.2(a), with a resonant frequency ofrad/s and a peak amplitude of Ci.
I I

Suppose we amplitude- and frequency-scale the network to obtain a resonant fre-

quency of I0 6 rad/s using a capacitor of nF. Then we have k f — 10 5 and the new
I ,

capacitance is

1/2 _ 1
10
k,k f ~ 2k X (
I0 6

Therefore k, = 500. and the new inductance and resistance are

._ 2k,-
,
10 '
H 1 mH
R I A, 500 fi

The scaled network is shown in Fig. 15.10(a), and its amplitude response, a scaled
version of Fig. 15.2(a), is shown in Fig. 15.10(b).

(o)

500--

(b)

FIGURE 15.10 (a) Network and (b) its amplitude response

408 Frequency Response Chap. 15

EXERCISE
15.6.1 Frequency- and impedance-scale the circuit of Fig. 15.4 to obtain co c = 200071
rad/s (fc = 1000 Hz), using capacitors of 0.01 and 0.005 fiF.
Ans. k f = (lOOOv/Dn, k, = 10 5 \2J~1 n, R's = 22.5 kfi each

15.7 THE DECIBEL

If the output of a network is used to provide some quantity that is to be sensed by a

human being, the function is sensed in a noncontinuous manner. For example, if the
network provides sound, as in the case of the telephone, the listener cannot detect
continuous changes in intensity. Moreover, if the sound intensity is (on some l

arbitrary scale) and it must be increased to l.l before the listener detects any change,
then the same listener can detect no change, if the original level is 2, until it is
increased to 2.2. In other words, the ear is not a linear device but more like a loga-
rithmic device, since the differences,

log l.l — log l = log l.l

and

log 2.2 - log 2 = log (-j-\

in the two cases are the same.

For this reason, among others, the amplitude response.

VzUco)
H(jco)\
V,Oco)|

is more commonly expressed in decibels, abbreviated dB and defined by

dB = 20log 10 |H(;co)| (15.37)

Historically, the logarithmic unit, now known as the bel, was defined originally by
Alexander Graham Bell (1847-1922), who, of course, invented the telephone, as the
power unit,

bels = log 10
p
-^

However, this proved to be a large unit, so that the unit decibel (yL bel) became
common. That is.

dB 10 log
^
Chap. 15 Frequency Response 409

It" the two average powers />, and /', are referred to equal impedances, the last
expression ma\ be given in terms of the corresponding voltages by

dB 10 log,.
V2OC0)
V,Oo>)

which, of course, is equivalent to (15.37). In any case. (15.37) is taken as the standard
definition.
In practice, frequencies are not ideally blocked or attenuated'in the filtering process,

as may be seen in the low -pass response of Fig. 15.5(a). A zero amplitude would
correspond to absolute, or infinite, attenuation, and any approach to such an ideal
situation would be difficult to appreciate on a linear sketch. For example, if \W(j(x>)\
- 0.001 in Fig. 15.5(a) (jL of 1 ",, of its peak value of I), would correspond
this to
— 60 dB (or 60 dB below its peak value of dB). The latter figure means much more to
a telephone engineer than the linear figure.
Even more often in practice it is of interest to consider the attenuation or loss,

defined bv

a(co) = — 20 log , I
H( jco) I

-201og 10
y 2 Uco)
V,(»
V,(./co)
20 log 10 dB (15.38)
VzUco)

In this case, a frequency a> ]

passes if a(a>,) is relatively small and isattenuated if a(a>,)
is relatively large. The decibel units enable us to tell with some standard precision the
degree to which the frequency is attenuated.
As an example, suppose

H(5)
s
2
+ J~ls + I

As the reader may verify, the amplitude response is

1
HOw)
VI +co 4
I

which is that of a low-pass filter with co c = 1 rad/s. The linear sketch looks somewhat
like that of Fig. 15.5(a). The attenuation is given in decibels by

a(<w) = 201og 10 J\ + w 4
- I01og, (l I- a/) (15.39)

and is shown in Fig. 1 5. 1 1 for < co < 5.

410 Frequency Response Chap. 15

I 2 3 4

FIGURE 15.11 Attenuation of a low-pass filter

EXERCISES
15.7.1 Let the amplitude response |H(y'cu)| be such that |H0'w)| max = |H(jO)| = K,
so that at cutoff |H(yco c )| =
K\«J~2 Find the loss, given by (15.38), at co =
'.

and at co = co c . Note that co c corresponds to the "3-dB point," meaning that

at CO = co c the loss is approximately 3 dB more than at the point of minimum
loss, co = in this case.
Ans. <X(0) = -20 log K, tX,(CO c ) = -20 log K+ 3

15.7.2 Given the bandpass filter function

0.2s
His)
s2 + 0.2s + 1

find the loss in decibels at co = 0.0001,0.5,0.905, 1.0, 1.105, 10, and 100 rad/s.
Ans. 94, 18, 3, 0, 3, 34, 54

PROBLEMS
15.1 For the circuit shown, R = 2 Q, L =
H, and C = 0.01 F. If the input and1

output are Vj and V2 network function and sketch the

, respectively, find the
amplitude and phase responses. Show that the peak amplitude and zero phase
occur at co = 10 rad/s.
Chap. 15 Frequency Response 411

PROBLEM 15.1

15.2 For the circuit shown, R= 1 ft, L= 0.01 H, and C = 1 F. If the input and
output are V, and V2 , respectively, find the network function and sketch the
amplitude and phase responses. Show that the peak amplitude and zero phase
occur at co = 10 rad/s.

PROBLEM 15.2

15.3 For the circuit shown, R = R2 =

x 0.5 ft and R 3 = 0.01 ft. If the input and
output are \ x
and V,, respectively, find the network function and sketch the

PROBLEM 15.3
412 Frequency Response Chap. 15

amplitude and phase responses. Show that the peak amplitude and zero phase
occur at co = 10 rad/s.

15.4 For the circuit shown, find the network function, H(.v) = V2 (.
s)[Vi(s )i and
sketch the amplitude and phase responses. Show that the peak amplitude and
zero phase occur at CO — 0.

PROBLEM 15.4

15.5 Show that in the general case the network function in the figure for Prob. 15.1
is given by

H(.s)
\ 2 (s) (R/L)s
V,(s) s2 + (R/L)s +
{l/LC)

Thus by comparison with ( 1 5.22) and ( 1 5.25), show that the circuit is a bandpass
filter with resonant, or center, frequency <o = l/s/LC, bandwidth B= R/L,
and quality factor Q = co /B - {\jR)^TjC. Show also that the gain K of the
filter, defined as its peak amplitude value, is given in this case by K = 1.

15.6 Show that in Prob. 15.5 if L = Q, C= l/Q, and R= I, then

(MQ)s
H(j)
s2 + (llQ)s + 1

which is a bandpass network function with quality factor Q, center frequency

co = 1 rad/s, bandwidth B = \jQ rad/s, and gain K= 1.

15.7 Show that in the general case the network function in the figure for Prob. I 5.2
is given by

\ 2 (s)_ (\/RC)s
H(.v)
V,(j) .9- +[(/? + \)IRC]s + {\/LC)

Thus show that the circuit is a bandpass filter with center frequency, bandwidth,

quality factor, and gain given, respectively, by

co
«JLC
R + I

B
RC
R
Q R 1 1

K
* + 1
 3

Chap. 15 Frequency Response 41

15.8 Show thai in Prob. 15.7 if L (1 K)/Q, C (?/( I A), and

R (1 K)JK, where A I. then

(K/Q)s
H(.v)
s2 + (\IQ)s + 1

Thus the circuit lias resonant frequency co 1 rad/s, quality factor Q, gain
A. and bandwidth fi \!Q rad/s.

15.9 Show that in the general case of the figure for Prob. 15.3

V,(.v) (1/A,).v
H(.v)
V,(J) .v- :
<1/A\).v : (l//? 3 )

Thus show that the circuit is a bandpass filter with center frequency l/\/A 3
rad/s, bandwidth \jR z rad/s, gain R Z IR\, and quality factor /? 2 / v//? 3 Let .

A5 ,
= Q/K, A\ -= Q, and A\, = to obtain the network function of Prob.
1

15.8.

15.10 Using the results of Prob. 1 5.9, obtain a bandpass filter having the form of the
figure for Prob. 15.3 with co I rad/s, Q = 10, and A = 2. Find the approxi-
mate cutoff points and the bandwidth.
15.11 Show that the network function V 3 /V, in the figure for Prob. 15.1 is given by

V 3 \ILC
V, s2 - (RIDs + (1/Z.C)

Let R = 1 Q, L = 1/V 2 H, and C s 2 F, and show that the result is a

low-pass filter with co c = 1 rad/s by finding the amplitude response.

15.12 Show that the circuit of Prob. 15.4 is a low-pass filter with a) c
= 1 rad/s. Com-
pare its amplitude response with that of Prob. 15.11 and note the improvement
around the cutoff point.

15.13 One type of low-pass filter is the //th-order Butterworth filter, whose amplitude
response is

\HUa>)\ =-
rT
J^, n = 1,2,3,...

where A' is filter would pass all frequencies in its passband

a constant. Ideally, a
equally well (\H\>
would be constant) and perfectly block all other fre-
quencies (\H would be zero). As the accompanying figure shows, for /; = 2, 3,
|

and 8, the Butterworth filter improves (approaches ideal) as the order /;

increases. Show that (x) c = 1 rad/s for any /; and that the filters of Probs. 15.11
and 15.12 are Butterworth filters of second and third orders, respectively.
Finally, sketch a fourth-order Butterworth response and compare it with the
second, third, and eighth orders.

15.14 Show that for the figure for Prob. 15.3

1/A,
Hfr)-V.M_
V,U) s2 + (l/R 2 )s r (1/A 3 )
41 4 Frequency Response Chap. 15

|H(jw)|

PROBLEM 15.13

Choose values of the resistances so that the circuit is a low-pass Butterworth

filter with co c = 1 rad/s and H(0) — 2. [Suggestion: ByProb. 15.1 1, the denomi-
nator of H(s) is required to be s 1 + *J~2s + 1.]
15.15 If the gain of a low-pass filter with the network function H(s) is defined to be
K= |H(0)|, compare and 15.14,
the gains in the general cases of Probs. 15.11
which are, respectively, passive and active circuits that perform low-pass filter-
ing. Note that gains higher than 1 are possible with active elements present.
This feature and the absence of inductors, which are undesirable at lower
frequencies, are advantages of active filters over passive filters.

15.16 Show that in the figure for Prob. 15.1

V*_ s^
V, s2 + (R/L)s + (XjLC)

Let R = 1 CI, L = 1/vT H, and C = 2 F, and show that the result is a

high-pass filter with cutoff (o c = 1 rad/s.

15.17 Apply impedance and frequency scaling to the circuit of Prob. 15.1 to obtain
a bandpass filter with a center frequency f = 10 5 Hz, with a quality factor
6= 5, and using a capacitor of 1 nF. (Suggestion: See Prob. 15.6.)

15.18 Obtain an active bandpass filter using the configuration of Prob. 15.3 with
f = 1000 Hz, 6 = 10, and K= 2, using capacitors of 0.01 //F. (Suggestion:
See Prob. 15.9.)

15.19 Show that the network function of the given circuit is

 .

Chap. 15 Frequency Response 415

K{s' I)
Him
V, (i/t».v i

I bus the circuit is a band-reject filter with center frequency (rejected) co I

rad/s. Also, as in the bandpass case, is the quality factor, and B (o„IQ \jQ
is the bandwidth. Note that the gain is H(0) A', where K 1

l-K
K

Q(l-K)

© I

Q(l-K)
J_
O

PROBLEM 15.19

15.20 Scale the network in the figure for Proh. I 5. 1 9 so that co = 10 6 rad/s, Q 5,

K= 0.5, and the capacitance is 10 nF.

15.21 Show that the network of Prob. 11.18 is a band-reject filter with a gain of I,

= 1, and a center frequency w a = 1 rad/s. Scale the network to obtain a

center frequency f = 60 Hz, using capacitors of 1 and 2 nF.

15.22 The given circuit is a fourth-order bandpass filter (the denominator of the
network function is fourth degree), with co = */2 rad/s and = 5. Verify
this by finding

25'
H(.s)
N -i 2f3 T 102
r
5 5

and so forth. [Suggestion: Find the amplitude response and verify that
o> = peak point and that (D C[ C =& co =p (a> /10) satisfy the
-y/2" yields the ,
„

appropriate equation. Since |H[ >

0, its maximum occurs when the maximum

of [H 2 occurs. Also |H| and H 2 are functions of co 2 thus the maximum

|
|
|
;

occurs when
^|H(/O»p=0
Since (d/dco) \
H |
2 = (d/da> 2 ) \
H |
2
(d(o 2 ld(o) = 2o> (d/da) 2 ) |
H =
|
2
0, we
must check co = separately.]

PROBLEM 15.22
 16
TRANSFORMERS

In our study of inductance in Chapter 7, we found that a changing current produces

a changing magnetic flux which induces a voltage in a coil. In this chapter, we shall
consider the effect of a changing magnetic flux that is common to two or more distinct
coils. Neighboring coils which share a common magnetic flux are said to be mutu-
ally coupled. In mutually coupled circuits, a changing current in one coil winding
produces an induced voltage in the remaining mutually coupled windings. The induced
voltage is characterized by a mutual inductance which exists between the neighboring
coils.

A system of mutually coupled coils which are wound on a composite form, or

core, is commonly called a transformer. Transformers are available in a wide variety
of sizes and shapes that are designed for numerous applications. Devices as small as
an aspirin tablet, for instance, are common in radios, television sets, and stereos for
connecting various amplifier stages of the systems. On the other hand, transformers
are designed for 60-Hz power applications which range in size from that of aping-pong
ball to those which are larger than an automobile.
We shall restrict ourselves to linear transformers (those whose constituent coils are
linear) and begin by considering the properties of self- and mutual inductances. Energy-
storage and impedance properties are then analyzed, and an important special case of
the linear transformer, the ideal transformer, is introduced. We shall conclude the
chapter with a discussion of equivalent circuits that are useful for representing linear
transformers.
The circuit analyses presented in the chapter will include both time-domain and
frequency-domain cases. Particular emphasis will be given to the frequency-domain
case because the most important transformer applications occur in the ac steady
state.

416
 ; 7

Chap. 16 Transformers 41

16.1 MUTUAL INDUCTANCE

In See. 7.4 we found that the inductance /. of a linear inductor, as shown in Fig. 16. 1,

is related to the flux linkage X by the expression

A = N<f> = Li

From Faraday's law. the terminal voltage is given by

dk j di
dt dt

FIGURE 16.1 Simple inductor

It is evident from Faraday's law that a voltage is induced in a coil which contains
a time-varying magnetic flux, regardless of the source of the flux. Let us therefore
consider a second coil, having N z turns, positioned in the neighborhood of the first
coil, having N=N x
turns, as shown in Fig. 16.2(a). In this case, we have formed a
simple transformer having two pairs of terminals in which coil 1 is referred to as the
primary winding and coil 2 as the secondary winding.
To introduce several important inductive quantities, let us begin by examining the
open-circuited secondary case of Fig. 16.2(a). The current /', produces a magnetic
flux 0,, given by

011 = <f>Ll + 0;i

where LI is the flux of/', that links coil 1 and not coil 2. called the leakage flux, and
;
is the flux of/', that links coil 2 and coil 1, called the mutual flux.
We shall assume, as in the case of the linear inductor, that the flux in each coil
links all turns of the coil. Since the secondary is an open circuit, no current flows in

coil 2, and the flux linkage of this coil is

A2 = ^ : 21
 v 9 (open)

(a)

(c)

FIGURE 16.2 Mutually coupled coils in which (a) i2 = 0. (b) i\ = 0,

and (c) i\ and iz are nonzero

418
 — 9

Chap. 16 Transformers 41

Therefore the voltage r. is given by

,.
• *
_ dX
— —2 -. -— /V N —dt
d<f> 2
;
,

2
~
dt

(We shall justify later that the polarity is as shown.)

In a linear transformer, the flux 0,, is linearly proportional to/,. Thus we may write

/V,0 : ,
M Z J, (1 6. 1)

where A/ ;i is a mutual inductance in henrys (H). In terms of this mutual inductance.

the open-circuit secondary voltage becomes

Let us now find the primary voltage y,. We know that

'"'
dX t

St

where X t ,
the flux linkage of coil I. is gi\en by

A, -
A/,0,,

As in the case of a single linear inductor, the inductance L, of the primary winding,
sometimes called the self-inductance of the primary to distinguish it from the mutual
inductance, is defined by

L,/, = yv.0,, (16.2)

Hence we see that

, di ,

'"' L
'7/T

in the case under consideration.

Next let us consider Fig. 16.2(b) in which the primary is an open circuit and a
current /2 flows in the secondary. Proceeding as before, we have

0:: 0/: 0i:

where (j) l2 is the leakage flux of/', that links coil 2 but not coil I. and i: is the mutual
flux of/' : that links coil 1 and coil 2. The flux linkage of coil I is

A, = yv,0 12

Therefore the primary voltage is given by

(//, ., dd>,,
dt dt
420 Transformers Chap. 16

If we define

Nrf 12 = M X2 i 2 (I6.3)

where A/ 12 is a mutual inductance, then the open-circuit primary voltage is

In the next section we shall show that the mutual inductances Af 12 and M 2X are
equal; therefore we shall write

M=M =M 12 lx (16.4)

and refer to M
as the mutual inductance.
The secondary voltage is given by
dX^
V2 = —J
dt
where the flux linkage is

X2 =N 2 <f> Z2

We now define the relation

LJ 2 = N 2 <t> 22 (16.5)

where L 2 is the self-inductance of coil 2 in henrys (H). Therefore

2 ~
dt

in the open-circuit primary case.

Let us now consider the general case of Fig. 16.2(c) in which both i 1
and i2 are
nonzero. The fluxes in coil 1 and coil 2. as shown in the figure, are

01 = <f>Ll + 021 + 012 = 011 + 012

02 = 0L2 + 012 + 021 = 021 + 022

respectively. Therefore the flux linkages of the primary and secondary coils are

X x
- A/,0 n 4- N i <p l2

X2 = N 2{ + N
2 (f> 2 (f> 22

Substituting from (16.1)-(16.5), we find upon differentiation that the primary and
secondary voltages are

V, L x ^ + M dt
(16-6)

2
dt dt
Chap. 16 Transformers 421

h is clear that the voltages consist of self-induced voltages due to the inductances / ,

and /.. and mutual voltages due to the mutual inductance A7.

In the preceding discussion, the coil windings in Fig. 1 6.2 are such that the algebraic

sign o\' the mutual voltage terms \l <//, dt and M di 2 dt arc positive for the terminal
voltage and current assignments as shown. In practice, it is, of course, undesirable to
show a detailed sketch of the windings. This is avoided by the use of a dot convention
which designates the polarity of the mutual voltage. Equivalent circuit symbols lor
the transformer o\~ Fig. 6.2 are shown in Fig. 6.3. The polarity markings are assigned
1 1

so that a positivel) increasing current flowing into a dotted (undotted) terminal in

one winding induces a positive voltage at the dotted (undotted) terminal of the other
winding.

'2

(a) (b)

FIG U R E 1 6.3 Circuit symbols for the transformer of Fig. 16.2: (a) dots
on upper terminals; (b) dots on lower terminals

For the purpose of writing the describing equations, we may state the following
rule:

A current i entering a dotted (undotted) terminal in one winding induces a

voltage M di/dt with positive polarity at the dotted (undotted) terminal of the
other winding.

We note that with this rule it is unimportant whether the current i is increasing or
not. since the sign of the induced voltage is accounted for by di/dt. That is, if /'
is

increasing, the induced voltage M di/dt is positive, and if / is decreasing, the induced
voltage is negative. If / is a dc current, then, of course, the induced voltage is zero.
As an example, let us write the loop equations for Fig. 16.3(a). We see that /,

enters a dotted terminal. Thus the mutual voltage M di 2 /dt has positive polarity at
the dotted terminal of the primary. Similarly, since /', enters a dotted terminal, the
mutual voltage M dijdt has positive polarity at the dotted terminal of the secondary.
Application of KVL around the primary and secondary circuits gives (16.6).
In a like manner, identical statements apply to the undotted terminals of Fig.
16.3(b), and KVL once again yields (16.6).
Let us now establish a method for placing polarity markings on a transformer.
We begin by arbitrarily assigning a dot to a terminal, such as terminal a of Fig. 16.4(a).
422 Transformers Chap. 16

(a) (b)

FIGURE 16.4 fa) Model of a transformer to determine polarity mark-

ings; (b) circuit symbol

A current into this terminal produces a flux (f>

as shown. (The direction of is deter-
mined by the right-hand rule, which states that if the fingers of the right hand encircle
a coil in the direction of the current, the thumb indicates the direction of the flux.)
The dotted terminal in the remaining coil is the one which a current enters to produce
a flux in the same direction as (This statement, which may be used to obtain the
<j>.

polarity of the induced voltage, is a consequence of a rule known as Lenz's law. The
reader may study this law, named for the German scientist Heinrich F. E. Lenz, in a
later course in electromagnetic field theory.) We note that a current flowing into b
in Fig. 16.4(a) flux in the same direction as 0. Thus terminal b receives the
produces a
polarity dot. The symbol for this transformer is shown in Fig. 16.4(b).
circuit
Evidently the polarity markings on a transformer are independent of the terminal
voltage and current assignments. Consider, for instance, the assignment of Fig. 16.5(a).
Applying KVL around the primary and secondary loops, we find

/.
dt
M clL
dt

di,
M didt1 + L 2
dt

In Fig. 16.5(b), the voltage and current assignments have been changed. Application
of KVL for these assignments yields

(a) (b)

FIGURE 16.5 Transformer with different terminal voltage and current

assignments
Chap. 16 Transformers 423

'
i
— L,—r- — M—r
dt dt

v2
,di
M^i
~d7
+ {
,

L^
,
Ll
di :
~dt

The above results can also be obtained by an alternative method for selecting the
sign of the mutual voltages, as follows:

If both currents enter (or leave) the dotted terminals of the coils, the mutual
and self-inductance terms for each terminal pair have the same sign; otherwise,
they have opposite signs.

The student can easily verify the use of this method for the cases of Figs. 6.3 and 6.5. 1 1

As an example, let us find the open-circuit voltage v 2 in the circuit of Fig. 16.6,
given that i\(0~) = 0. For / > 0, since no current flows in the secondary, applying
KVL around the primary yields

4dt1 + 10/, '

- 20

This is a first-order differential equation with a general solution of

/, = 2 + Ae- i0!

Since ?(0 + ) = /(0 ) = 0. A = -2 and

i, = 2(1 - e~ ]0:
) A
Therefore, in the secondary.

v2 = -0.2541
dt

= -5e- 'V 10

Let us now consider the transformer of Fig. 16.3(a) in the case of an excitation
having a complex frequency s. In this case, the phasor equations corresponding to
(16.6) are

Y l
(s) = sL 1
l l (s) + sMU(s)
V z (s) = sMl^s) + sL 2 l 2 (s)

The phasor circuit for this network is shown in Fig. 16.7. In the case of a purely
sinusoidal excitation, we simply replace s by jco.
For a final example, consider finding the phasor voltage V 2 (5) in the network of
Fig. 16.8 due to the complex forcing function V,(s). The loop equations in the primary
424 Transformers Chap. 16

0.25 H sM

20 V^ V, Sl_, Sl_,

FIGURE 16.6 Circuit with an open-circuit secondary FIGURE 16.7 Phasor circuit for a

transformer

FIGURE 16.8 Circuit with a complex excitation

and secondary are

v,= + 2)1, + sl
is 2

= si, + (2.9 + 3)I 2

Therefore
s + 2 V,

r
s
l
2
s + 2 s

s 2s + 3

-*v,
s
2
+ Is + 6

and if H{s) is the voltage ratio function, then

_2 31^ -3s
H(s)
V, s
2
+ Is -

-35
(s + 1)0? + 6)

A pole-zero plot and a sketch of |H(/to)| are shown in Fig. 16.9. We may show that
the maximum ac steady-state response occurs for co = */~6 rad/s.
Suppose, as an example, that t>, = 100 cos 10? V. Then

V, = 100 V, s=yl0 rad/s

Chap. 16 Transformers 425

I
co

-6
*— 9 a

(a)

FIGURE 16.9 (a) Pole-zero and (b) \

H(j(o) \
plot for the network of
Fig. 16 8

and
V, = 25.6 /126.7° V

Thus the ac steady-state response is

r z {t) 25.6 cos (10/ + 126.7 ) V

EXERCISES
16.1.1 In the circuit of Fig. 16.3(a), L = L =0.1 H and A/ - lOmH. Find r, and
x 2

v 2 if (a) /, =10 mA and i2 = 0, (b) = and i2 = 10 sin 100/ mA, and (c)
/,

/, =0.1 cos / A and i2 = 0.3 sin (/ + 30 A. )

Arts, (a) 0, 0; (b) 10 cos 100/ mV, 0.1 cos 100/ V;

(c) -10 sin / + 3cos(/ + 30°) mV,
-sin / + 30 cos (/ + 30 mV )

16.1.2 Verify that the coil windings of Fig. 16.2 are consistent with the polarity
markings of Fig. 16.3.

16.1.3 Find I 2 in Fig. 16.8 for a sinusoidal voltage of V, = 10/CF V having CO = 10

rad/s if the 2-fi resistor is replaced by the series combination of a 10-Q resistor
and a 1-F capacitor. Ans. 0.417/163.5 A

16.2 ENERGY STORAGE

We have previously shown that the energy stored in an inductor at time / is

»(/) = \Li\i)

Evidently, for a given inductance L. the energy is completely specified in terms of /'(/)
426 Transformers Chap. 16

Let us now determine the energy stored in a pair of mutually coupled inductors, such
as those of Figs. 16.2-16.4. In so doing, we shall show that A/ 12 = M 2l = M and
establish limits for the magnitude of M.
In Fig. 16.3, the stored energy is the sum of the energies supplied to the primary
and secondary terminals. The instantaneous powers delivered to these terminals,
from (16.6), are

* = *'. = (^£ + "»£>'

(16.7)

Pi = v iH =
\
M ^-^ + ^"jfV*
where M has been replaced by M X2 and M 2i in the appropriate terms.
Let us now perform a simple experiment. Suppose we start at time t with /,(f )

— '2(^0) — 0- Since the magnetic flux is zero, no energy is stored in the magnetic
field; that is, w{t ) = 0. Next, assume, beginning at time t , that we maintain i2 =
and increase /, until, at time and i 2 {t ) = 0. During
/,, /,(?,) = /, x
this interval, i2 =
and dijdt = 0. Thus the energy accumulated during this time is

f" f" di
if, (Pi +Pi)dt= M,j<fi

As a final step, let us maintain /, = l and increase i 2 until, at time t 2 i(t 2 ) = I2

x ,

During this time interval, dijdt = 0, and the energy accumulated, using (16.7), is

Wi Hm^ + l^Y
-I2
= (M 12 /, + L 2 i
2 )di 2

= M X2 IJ2 + \L 2 1\

Thus the energy stored in the transformer at time t2 is

w(t 2 ) = w(t ) + iv, + w2

Let us now repeat our experiment but reverse the order in which we increase/, and
i2 . That is, in the interval from t to t u we increase /, so that i 2 {t x ) = I 2 while holding

/j = 0. Finally, we maintain i 2 = I2 while increasing /, so that /,(/ 2 ) = /,. Using the

same steps as before, we find in this case that

w(t 2 ) = \LJ\ + M 1X IJZ \

\L 2 l\
Chap. 16 Transformers 427

Since /,(M /, and i

:
{t :
) / : in both experiments, then w(t 2 ) should be the same in
both cases. Comparing our results, we see that this requires

M 12 = A/ 21 = M
If we repeat our experiment with the polarity markings of Fig. 16.4, the sign of
the mutual voltage is negative, which causes the mutual term in the energy M/,/ : to
be negative. Thus a general expression for the energy at any time / is given by

w{t) U,/f fc Mi\i 2 !

\L 2 i\ (16.8)

\\ here the sign in the mutual term is positive if both currents enter dotted (or undotted)
terminals; otherwise, it is negative.
The coefficient of coupling between the inductors indicates the amount of coupling
and is defined by

k = <l6 9)
7lk -

Clearly, k = if no coupling exists between the coils, since M= 0. We can establish

the upper limit of A- by substituting from (16.1 )-( 16.5), which yields

M = JM, M
V0 M V0 22 -<
k = 1 1 ,
= fat I
/02_,
,

V/L,L 2 JL,L 2

since 12 /0n < ' and 21 /0 22 < 1. Thus we must have

0< k< 1

or, equivalently.

0< M <^L X
L2

If A- == 1, all of the flux links all of the turns of both windings, which is a unity-coupled
transformer.
The value of A (and hence M ) depends on the physical dimensions and number of
turns of each coil, their relative positions to one another, and the magnetic properties
of the core on which they are wound. Coils are said to be loosely coupled if A < 0.5,
whereas those in which A > 0.5 are tightly coupled. Most air-core transformers are
loosely coupled, in contrast to iron-core devices for which A can approach I.

Let us now examine the values of i 1 and z\ in (16.8) for which w(t) is zero. From the
quadratic formula, we may write

/, = T j^ ± f-JW- - L L l
X Z

For real values of i1 and /\. we see that for w(t) = we must have
428 Transformers Chap. 16

0, M < JL L X 2

M.
±—'2, M = ^/TJT
(16.10)
2

where the sign of the mutual term is negative if both currents enter dotted (or undotted)
terminals; otherwise, it is positive. The second equation of (16.10) shows that w(t)
can be zero in a unity-coupled (k = 1) transformer even though /, and i 2 are nonzero.

EXERCISE

16.2.1 Determine the energy stored in each case of Ex. 16.1.1 at / = 0.

Ans. (a) 5 //J, (b) 0, (c) 1.775 mJ

16.3 CIRCUITS WITH LINEAR TRANSFORMERS

A two-winding transformer is in general a four-terminal device in which the reference

potential in the primary can be different from that of the secondary without altering
the values of v x
, v2 , i u or /',. Consider, for example, the circuit of Fig. 16. 10. KVL

FIGURE 16.10 Circuit showing different reference potentials in

primary and secondary

around the primary and secondary circuits gives

»,(0 = 7?,/, I
L,
dt
M di.

M^\R 2i 2
1 di 2
Chap. 16 Transformers 429

Evidently, the voltages and currents are not affected byV For this reason, the secon- .

dai a of the transformer from the primary. Point a, of course,

is said to have dc isolation
is at an absolute potential of /,/?, volts with respect to the ground reference. If
I -|

we now let V = 0. it is seen that the bottom terminals of the transformer are con-
nected and that the primary and secondary circuits have a common reference point.
The transformer, in this case, is a three-terminal device.
As a second example, let us find the complete response /, for / in Fig. 16. 1 1,

2a

FIGURE 16.11 Switching circuit containing a transformer

given M=
\\*f~2 and /,(()") / 2 (0~) H =
0. Since the forced response for /, is a dc =
current (by inspection), no steady-state voltage is produced in the secondary; thus
i
2f
= 0. To find the natural response, let us first obtain the network function for / > 0.

Application of KVL for a complex excitation V,(s) yields

3 „\ ¥
V,(5) = (-|j +
,

2)l , - -jjsl
l

1
= sh (s + 2)I 2

from which we find

H(J) = I

^ 2 (s + l)(s + 4)

The poles of H(s) are the natural frequencies. — 1, —4, of the natural response.
Thus, for / > 0.

h = 1
2/ i iH = +A 1
e-' + Aie-
From (16.8), w{0~) = 0; therefore since the energy cannot change instantaneously
in the absence of infinite forcing functions, w(0 + ) = 0. From (16.10), noting that
M< JL L 1
A we1 ,
+
see that /](0 ) = = To obtain a second initial condition for
/ 2 ( u+ ) u -

finding A and A 2 we apply

l
. KVL around the primary and secondary at = + which t ,

yields

3 J/ 1 (0 + )
+
12 = 1 di 2 (0 )

2 dt dt

+ +
rf/,(0 ) di 2 (0
= 1
)

dt dt
430 Transformers Chap. 16

Solving, we find c//\(0

+
)/rf/ = 6^/T A/s. Evaluating A i
and A 2 , using z'
2 (0
+
) and
+
di 2 (0 )/dt, the solution becomes

Suppose we now repeat this example for M= /L L 1

/S X
= ^/3/2. In this case the
network function for t > is

KS}
5s + 4
and the complete response is

i2 = Ae-°*'

We know that /,(0~) = z'

2
(0") = 0; however, if we take /2 (0
+
= (0") as before,
) /2 then
the unreasonable solution /, = results. Also, applying KVL gives

I? - ?i -4-
3 dl '
/
3 di 2

0= _7Z^l
V2(/f +
+^
f 2/,2t A
Multiplying the latter equation by v/3/2 and adding it to the former, we have

12 = 2/, +-v/F/ 2 (16.11)

This result contradicts /,(0 + ) = /2 (0>) = 0.

Recalling (16.10), we see that the energy is zero in the unity-coupled case when

/,(0
+
) = ^/ 2
(0+) = ^i 2 (0 )
+

Combining the last result and (16.11) gives /

2 (0
+
) = 2.94 A, and therefore

i2 = 2.94c- - 8 '

Thus the current in a unity-coupled transformer can change instantaneously as a

result of the application of a finite forcing function.
As another example, let us find the steady-state response v 2 in Fig. 16.12. Applying
KVL for a complex frequency s, we have

v = (,+ i+l),, _(, + !),,

1

°- -( s + t) i + '
(
2i + 1+ t)
for which

H(s)
V _ h. -2
S2 + l

V, V! s3 + 3s
2
+ 2s + 2
Chap. 16 Transformers 431

v, = 10 cos 2t

FIGURE 16.12 Circuit containing a linear transformer

Substituting s = ./2 rad/s and V, 10 V. we find V = 2

2.79 /-21.8 V. Thus

r, -- 2.79 cos (2/ - 21. 8 ) V

As a final us find the network function V 2 /V, for the phasor circuit
example, let

of Fig. 16.13. Let us examine the voltage V^ which appears across winding A in
first

loop 1. We see that the voltage due to I, is the self-inductance term (2.vl ) and the ,

mutual term of winding C (2sl,). The voltage due to I 2 arises from the mutual terms
of winding B — .d,) and winding C — 2.si 2 ). Thus
( (

V,, = (2.v + 2.9)1, -(s + 2s)\ 2

Similarly, we find

VB = -(.v- 35)1, -f (3.v+ 3.v)I 2

Vc = (4s + 2.v)I, - (4.9 + 35)I 2

FIGURE 16.13 Circuit containing a three-winding transformer

Applying KVL in loops I and 2 gives

V,= 3I, + V, + V C -2I 2

0= -21, - V c + \ B +5\ 2

or
V,= (lO.v + 3)1, - (10s L 2)I 2

0= -(10.9 + 2)1, + (13.9+ 5)1,

432 Transformers Chap. 16

Solving these equations for the ratio I 2 /V,, we find the network function to be

H(j)-
3I 2 3(10j+ 2)
V, ~ V, ~
30j 2 + 49^+11

EXERCISES

16.3.1 Determine i l
for / > in the network of Fig. 16.11, given M = l/*fT H.
Assume the circuit is in steady state at / = 0~. Ans. 6 — 4e~' — 2e~ 4! A
16.3.2 Determine the impedance seen from the terminals of the voltage source v\ in
Fig. 16.12. Ans. 1.48 /37.75 Q.

16.3.3 Determine the network function V 2 /V, for Fig. 16.13 if the polarity dot of coil
B is placed on the other terminal. Ans. 6(5 + l)/(6s 2 + 45^ + 11)

16.4 REFLECTED IMPEDANCE

In this section we shall develop several important impedance relationships for the ac
steady-state case. Let us begin by considering the phasor circuit of Fig. 16.14 having
a practical source \s with an impedance Z, connected in the secondary.

L
2
V2
D
l2 I

FIGURE 16.14 Circuit for deriving impedance relationships

Applying KVL at the primary terminals of the transformer we find

V, =jcoL 1 l 1
-jcoMl 2
(16.12)
= -jcoMI l + (Z 2 +jo>L 2 )l 2

Eliminating I
2
from these equations we have

jcoMijcoM)'
V, JcoL ">
1
Z 2 + jcoL 2 _
 —L
Chap. 16 Transformers 433

so that the input impedance seen at the primary terminals of the transformer is

Z == : -; Q,L '+ 16 13 )
'

l7 Z 2 +./o>L 2 ( -

The first part. jcoL^ depends entirely on the reactance of the primary. The second
part is due to the mutual coupling, and it is called the reflected impedance, given by

7 or A/ 2
£jR
-Z 2 +jcoL 2

It may be thought of as the impedance inserted into, or reflected into, the primary by
the secondary.
The input impedance as seen by the source Vg is evidently

z in
= zg + z,

Also the secondary-to-primary current ratio I 2 /I, may be found from the second
equation of (16.12). and the voltage ratio V 2 /V, may be found from

V
T, ~Z >«)
using (16.13) The result is

jcoM
I, Z 2 |
jcoL 2
(16.14)
v 2
jcoMZ 2
v, ja>L (Z 2 x + jcoL 2)
-\- a, 2 M 2

It is interesting to note that ZR is independent of the dot locations on the transform-

er. If either dot in Fig. 16.14 is placed on the opposite terminal, the sign of the mutual
term in each equation of (16.1 1) changes, which is equivalent to replacing M by —M.
Since ZR varies as M 2
, its sign is unchanged. A second important property is illus-

trated by rationalizing Z R which

, gives

z" = - jtt+<0li)l
5iTTO? [S!

where we have used the relation Z 2

=R 2
-\- jXi for the load impedance. It is seen that
the sign of the imaginary part of ZR is minus. Therefore the reflected reactance is

opposite that of the net reactance X + 2 a>L 2 of the secondary. In particular, if X 2 is

a capacitive reactance whose magnitude is less than a>L 2 or if it is an inductive react-

ance, then the reflected reactance is capacitive. Otherwise, the reflected reactance is

either inductive or it is zero. In the latter case, X 2 must be a capacitive reactance

— 1/cuC with a resonant frequency f = O) /27i = z ,. _ . In this case, for co = co ,

ln*J 2 C
434 Transformers Chap. 16

and the reflected impedance is purely real.

we see that if the polarity dot on either winding of
Inspecting (16.13) and (16.14),
Fig. 16.14 occurs on the opposite terminal, then the current and voltage ratios require
a sign change, whereas the impedance relations are unaffected.

EXERCISES
16.4.1 Given: In Fig. 16.14, \g = 100/0° V, Zg = 10 ft, L, = L2 = 0.5 H,
M= 0.1 H, and co = lOOrad/s. If Z2 = 10 - O'1000/o>)ft, find (a) Z in ,

(b) I„ (c) I 2 , (d) V„ and (e) V2 .

Arts, (a) 48.81 /77.47° ft, (b) 2.05 / -77 .47° A, (c) 0.50 /-63.43 A,
(d) 97.62 /11.83° V, (e) 7.07 /- 108.43° V
16.4.2 Repeat Ex. 16.4.1 if the polarity dot is on the lower terminal of the secondary.
Ans. (a) 48.81 /77.47° ft, (b) 2.05 / -77.47° A, (c) 0.50 /116.57° A,
(d) 97.62 /11.83° V, (e) 7.07 /71.57° V
16.4.3 Determine the resonant frequency in Ex. 16.4.1 for which the reflected imped-
ance is real. Ans. 7.12 Hz

16.5 THE IDEAL TRANSFORMER

An ideal transformer is a lossless unity-coupled transformer in which the self-induct-

ances of the primary and secondary are infinite but their ratio is finite. Physical
transformers which approximate this ideal case are the previously mentioned iron-core
transformers. The primary and secondary coils are wound on a laminated iron-core
structure such that nearly all of the flux links all of the turns of both coils. The react-
ances of the primary and secondary self-inductances are very large compared to
moderate load impedances, and the coupling coefficient is nearly unity over the fre-
quency range for which the device is designed. The ideal transformer is thus an approxi-
mate model for well-constructed iron-core transformers.
An important parameter that is necessary in describing the characteristics of an
ideal transformer is the turns ratio n, defined by

=
n
£ (16.15)

where N and N 2 are the number of turns on the primary and secondary, respectively.
x

The flux produced in a winding of a transformer due to a current in the winding is

Chap. 16 Transformers 435

proportional to the product of the current and the number of turns on the winding.
Thus, in the primary and secondary windings,

0ii = a/V,/,
<f>22 = aW 2
/
2

where a is a constant of proportionality which depends on the physical properties of

the transformer. Substituting these relations into (16.2) and (16.5), we see that

L, = a/V
2
, L2 = o.N\

Therefore

£-(&)'—* (16.16)

In the case of unity coupling, M = ^/L L l 2 , the second of (16.14) becomes

jcoZ 2 ^/L ]
L 2

V, y'coL,(Z 2 + jcoL 2 ) + (jJ

2
L L2
x

= n

For the ideal transformer, in addition to unity coupling, the inductances L, and L 2
tend to infinity in such a way that the ratio of (16.16) is the constant n 1 . From the first

equation of (16.14) we see that in this case we have

h = Joi^/L L2
ij m s

Ij l,.l,-~ Z 2 + jcoL 2

- lim
JQ>*JLJL 2
LuLr^jCO + (Z 2 /L 2 )
= ]im MV")
L,,L,—>j<Q + (Z 2 /L 2 )

J_
n

Thus the primary and secondary voltages and currents of an ideal transformer are
by
related simply to the turns ratio

v 2

rr- (16.17)
h _ l

n
436 Transformers Chap. 16

Replacing n by N 2 /N u we have

N 2l2

Therefore the voltages are in the same ratio as the turns, and the ampere turns (NI) are
the same for both primary and secondary.
The symbol for an ideal transformer is shown in Fig. 16.15, connected to a secon-

©9 |
•

V2

-
s^\
i')
Z
2

FIGURE 16.15 Circuit containing an ideal transformer

dary load Z2
and a source V g with impedance Z g The vertical lines are used to sym-
.

bolize the iron core,and 1 n denotes the turns ratio. If one or the other, but not both,
:

of the polarity dots is placed on the opposite terminal, then the ratios of (16.17)
become negative.
The primary impedance Z, of (16.13), in the case of the ideal transformer, is given
by
_ V, _ VJn _ V 2 /I 2
7

or

7 _Z 2 .
1 — »2 (16.18)
z,

where Z2 is the load impedance of Fig. 16.15. Thus the input impedance viewed from
the terminals of the voltage source is

Z in g
(16.19)
n2

The lossless property of an ideal transformer is easily demonstrated using (16.17).

Taking the complex conjugate of the current ratio, we have, since n is real,

If = n\%
from which
V 2 If
If

or
V,If _V 2 If
Chap. 16 Transformers 437

Thus the complex power applied to the primary is delivered to the load Z 2 , and hence
the transformer absorbs zero power.
In anal\ zing networks containing ideal transformers, it is often convenient to
remove the transformer before performing the analysis. Let us consider, for example,
replacing the transformer and load impedance Z 2 of Fig. 16.15. Clearly, the input
impedance seen by the generator V, is Z in given by (16.19), so that an equivalent circuit
insofar as V s is concerned is shown in Fig. 16.16. The voltages and currents can now
be easily determined from the single-loop circuit.

Zg

Z?/n V, =

FIGURE 16.16 Equivalent circuit for Fig. 16.15

obtained by replacing the secondary

Let us next consider replacing the primary circuit and the transformer of Fig.
16.15 by its Thevenin equivalent. By (16.17) we have

I, = „I 2 . V2 = A?V,

so that for V oc , I2 = and thus I, = 0. Therefore

V oc = V 2 = #iV, = n\ g
For I sc , we have V2 = and thus V, = 0. Therefore

V.
T
*sc
sc
=
— I
*2
-*2
I,
—- ii t
and

th — j— - n ^ g

The resulting equivalent circuit is shown in Fig. 16.17.

27
n Zg

Z2 V 2 =nV,
" = I,/n
p
FIG U R E 16.17 Thevenin equivalent circuit of Fig. 16.15
438 Transformers Chap. 16

Inspecting the results for the Thevenin equivalent circuit, we see that each primary
voltage is multiplied by n, each primary current is multiplied by \jn, and each primary
impedance is multiplied by n 1 when we replace the primary circuit and the transformer.
It may be shown that this statement holds in general whether the result is the Thevenin
circuit or not. On the other hand, if we replace the secondary circuit and transformer
by an equivalent circuit, as in Fig. 16.16, we simply multiply each secondary voltage,
current,and impedance by 1//?, n, and l//? 2 respectively. If either dot on the transform-
,

er is reversed, we simply replace n by —n.

In applying the above-described procedure, the student is cautioned that the tech-
nique is valid if the transformer divides the circuit into two parts. When external
connections exist between the windings, the method in general cannot be used. The
equivalent circuits to be discussed in the next section are often useful for networks of
this type.
As an example, let us find V 2 in the circuit of Fig. 16.18(a), which contains an
ideal transformer and a voltage-controlled current source. In Fig. 16.18(b) an equiva-

io^o 300 + j400a

(a)

iooZO° 300+1400 a

(b)

FIGURE 16.18 (a) Example circuit; (b) equivalent circuit

primary circuit and the transformer is shown. It should be noted

lent circuit for the
that a minus sign has been used on primary voltages and currents to account for the
polarity dots shown on the transformer. The nodal equation for V, gives

V + ioo/o_°
2
+ io 3
v2 +
10 3 300 + /400
from which
V2 = 27.95/206.6° V
Chap. 16 Transformers 439

So far we have considered only the ideal transformer in the ac steady-state case.
In the general case, we see from (16.6) that

Ll
~dT \L 2 L2 dt )

Thus since M =L 2
l
L 2 and ^'LJL^ — /?, we have

Next, let us rearrange the first equation of (16.6) in the form

t\ _ di ]

+
M di 2

L "~dJ
x
L[ dt

di, di\
=
^+
,

,1
-dT

Taking the limit as L, becomes infinite, we have

dh „dh
dt' dt

Integrating both sides, we have

/', = —n\ 2 + C,

where C x
is a constant of integration. Since dc currents produce no time-varying
magnetic flux, they do not contribute to the induced voltages or currents in the ideal
transformer. Therefore if we neglect the constant C,, then

i, = -ni2

where the minus sign arises due to the direction assigned i 2 in Fig. 16.3(a). Thus the
same current and voltage relationships are valid in the time domain as were found in
the frequency domain if we neglect any dc currents.

EXERCISES
16.5.1 In Fig. 16.15, V, = 100/0^ \,Z g = 100 ft, and Z2 = 90 kQ. Find // such that
Z, = Zg , and then find the power delivered to Z2 . Ans. 30, 12.5 W
440 Transformers Chap. 16

16.5.2 Using Norton's theorem, show that the primary circuit and transformer of
Fig. 16.15 is equivalent to a constant current source of V g /nZ g A in parallel
with an impedance of n 2 Z g fi.

16.5.3 Find V2 in Fig. 16.18 (a) by first replacing the secondary circuit and the trans-
former by an equivalent circuit.

16.6 EQUIVALENT CIRCUITS

Equivalent circuits for linear transformers are easily developed by considering the
equations for primary and secondary currents and voltages. In Fig. 16.19(a) we have

1
'
dt
~ dt

M^i + L
dt
2
^
di :
Tt

It is evident that the circuit of Fig. 16.19(b) satisfies these equations. The dependent
voltage sources, however, are controlled by the time derivatives of the primary and
secondary currents. In the frequency domain, these sources can be considered as
current-controlled voltage sources.

°—TiRP 1
—^WT^—
i
o

&<£> 6 dt

(o) (b)

FIGURE 16.19 Linear transformer; (b) equivalent circuit

Let us now rearrange the equations in the form

idi, ,
,,/di, ,
dL

m( di, ,
di-

) + (i
>
- M >£
These equations are satisfied by the T network of Fig. 16.20(b). Since this circuit is
a three-terminal network, it is equivalent to the transformer connection of Fig.
16.20(a). If either polarity dot is changed to another terminal, we must replace by M
—M in the equivalent circuits.
Chap. 16 Transformers 441

o—nnnr - 1

+ L,-M L 2 -M +

(o) (b)

FIGURE 16.20 (a) Linear transformer with common terminals; (b)

equivalent T network

In the case of an ideal transformer as shown in Fig. 16.21(a), the currents and
voltages are given by

i2 = —
n
!
>
21
=
v2 nv,

Clearly, the circuits of Figs. 16.21(b) and (c) satisfy these relations. If either of the dots
are reversed, of course, we must replace n by —n in the equivalent circuits.

><*> <$>! "a

(a) (b)

'2

V ni nv V
l

2<J> <d> l
2

(c)

FIGURE 16.21 (a) Ideal transformer; (b) and (c) equivalent circuits
442 Transformers Chap. 16

EXERCISE
16.6.1 Employing the T equivalent circuit for the linear transformer in Fig. 16.12,
determine the phasor current Ij. Ans. 6.11 /— 31.15° A

PROBLEMS
16.1 In Fig. 16.2 (a), 7V2 = 1000 turns and <f> 2l =50 ^Wb when /, = 1 A. Deter-
mine v 2 if i\ = 10 cos 100/ A.

16.2 Determine the polarity markings for the coupled circuits.

-oc

k><>

(a) (b)

PROBLEM 16.2

16.3 In Fig. (a) of Prob. 16.2 found that the inductance measured between ter-
it is

minals a and d is 0.2 H when

terminals b and c are connected. When terminals
b and d are connected, the measured inductance between terminals a and c is
0.6 H. Determine the mutual inductance M.

16.4 Find v if i = 2u(t) A.

PROBLEM 16.4

16.5 Repeat Prob. 16.4 if i CO = 2<r'w(/) A.

 1

Chap. 16 Transformers 443

16.6 Find the ac steady-state value of?'.

0.5 H

I A< v

PROBLEM 16.6

16.7 Repeat Prob. 16.6 if a 1-F capacitor is connected in parallel with the \-Q
secondary resistor.

16.8 Given that Af 12 = M 2 \ in (16.7), show that the integral

w{t) = [' (Pi +p 2 )dt

yields (16.8).

16.9 Substituting (16.9) into (16.8), show that a transformer satisfies the passivity
condition w(t) > 0.
16.10 Repeat Prob. 16.4 if a 3-Q resistor is connected to the secondary terminals.
(Hint: Replace the current source and 2-fi resistor by their Thevenin equiva-
lent.)

16.11 Repeat Prob. 16.4 if the mutual inductance is «/2 H and a 3-Q resistor is

connected to the secondary terminals.

16.12 Find the steady-state value of i1 .

I3A 4D

47 cos 1000
V 6 2mH ImH 3mH

PROBLEM 16.12

16.13 Find i(/) for / > 0, given i(0) = and r(0) = 4 V.

i(t)
2H

v(t) Z^±F 4H

VW^
en
PROBLEM 16.13
444 Transformers Chap. 16

16.14 Find his).

I.(s)
OIF

in

V(s)
6 2H 5H

PROBLEM 16.14

16.15 Find the network function, H{s) =\ 2 {s)l\ 1

(s).

IA
2
H IH m
H
v,(s)r + IH 2

h ia< v,( S )

PROBLEM 16.15

16.16 Find the steady-state value of ix .

F
I" 1:2
i

cos 2t
V © 2n

PROBLEM 16.16

16.17 Find the steady-state value of v.

PROBLEM 16.17
 '

Chap. 16 Transformers 445

16.18 Find the power delivered to the 4-fi resistor.

AA/*

2A

100 V
rms © 4n

PROBLEM 16.18

16.19 In Fig. 16.20 (a), determine the r-parameters for the transformer for complex
frequency s.

16.20 Repeat Prob. 16.19 for the >'-parameters.

16.21 Find u using the T equivalent circuit for the transformer.

VvV

10X1

20 cos 100 t (
V

PROBLEM 16.21
 .

17
FOURIER METHODS

Thus far, with few exceptions, we have considered excitations which were exponentials,
sinusoids, or damped sinusoids, and the reader may have been impressed by the
relative ease with which we can handle such functions. In particular, the concept of
impedance allowed us to treat circuits with these excitations as if they were resistive
circuits and find both their forced and natural responses.
However, as we all know, there are many functions that are important in engi-
neering which are not exponentials or sinusoids. A few examples are square waves,
sawtooth waves, and triangular pulses. Indeed, a function may be represented by a set
of data points and have no analytical representation given at all. In this chapter we
shall consider these additional functions and show how we may represent them in
terms of our familiar sinusoidal functions. The techniques we shall use were first given
in 1822 by the great French mathematician and physicist Jean-Baptiste Joseph
Fourier, who lived from 1768 to 1830.

17.1 THE TRIGONOMETRIC FOURIER SERIES

Let us begin by considering a function f(t) which is periodic of period T; that is,

f(t)=f(t+T)

As Fourier showed, if /(/) satisfies a set of rather general conditions, it may be

represented by the infinite series of sinusoids

f(t) = -j + a x
cos co t + a 2 cos 2co t + . .

+ bi sin eo t + b 2 sin 2co t + . . .

446
 7

Chap. 1 Fourier Methods 447

or. more compactly,

fit) = ^r
Z
+ 2 ( fl n cos nc<V + *n sin nco t) (17.1)
n - 1

where co = 2^/r. This series is called the trigonometric Fourier series, or simply the
Fourier series. of/(/). The as and 6's are called the Fourier coefficients and depend, of
course, on/(r).
The coefficients may be determined rather easily by the use of Table 12.1, given
earlier in Chapter 12. Let us begin by obtaining a , which may be done by integrating
both sides of (17.1) over a full period; that is,

•r t>T

I fif)dt=. (^-dt
Jo Jo L

t f (a. cos nco + b„ sin nco

"=' Jo
Qt t) dt

Since 7" = 2n/co , every term in the summation is zero by entry 2 in Table 12.1 (a = 0),
and therefore we have

°o =^
r.f fit)dt (17.2)

Next, let us multiply (17.1) through by cos mco Q t, where m is an integer, and
integrate. This yields

fit) cos ma) Q t dt = -y cos ma> t dt

+ L« n cos mco t cos «c<V dt

+ Y^bn \
cos /mgv sin nco t dt
«=i Jo

By entries 2, 4, and 5 of Table 12.1 (for a = /? = 0), every term in the right member is

zero except the term where n =m in the first summation. This term is given by

T rp

am I cos
2
tncoct dt = — a m = -^a„
o ^o
t»n

so that

°m = 2 C fit) cos mco m=

T Qt dt, \,2, 3, ... (17.3)

Finally, multiplying (17.1) by sin ww r, integrating, and applying Table 12.1, we have

"m rp fit) sin mco t dt, m= \,2, 3, ... (17.4)

448 Fourier Methods Chap. 17

We note that (17.2) is the special case, m = 0, of (17. 3) (which is why we used
a /2 instead of a for the constant term). Also, as the reader may easily show, we may
integrate over any interval of length T, such as t to t T, for arbitrary t and the + ,

results will be the same. Therefore we may summarize by giving the Fourier coefficients
in the form
['°
+T
_2
T fit) cos nm Q t dt, n = 0,1,2,
(17.5)

6„ = 4r f(t) sin noj t dt, n = 1, 2, 3,

We have replaced the dummy subscript m by n to correspond to the notation of (17.1).

The term a n cos nco Q t + b„ sin nco t in (17.1) is sometimes called the A?th harmonic,
by analogy with music theory. The case n = 1 is the first harmonic, or fundamental,
with fundamental frequency co . The case n = 2 is the second harmonic with frequency
2co , and so forth. The term a /2 is the constant, or dc, component.
The conditions that (17.1) is the Fourier series representing f(t), where the Fourier
coefficients are given by (17.5), are, as we have said, quite general and hold for almost
any function we are likely to encounter in engineering. For the reader's information
they are as follows:

In any finite interval /(0 has at most a finite number of finite discontinuities,
a finite number of maxima and minima, and

\
\f(t) dt < co (17.6)
Jo

As an example, suppose /(r) is the sawtooth wave of Fig. 17.1, given by

f(t) = t, —7i < <t n

(17.7)
f(t + 2n)=f(t)

37T t

FIGURE17.1 Sawtooth wave

Chap. 17 Fourier Methods 449

Since T 2n, we have eo = 2n/T = l . If we choose t = —n, then the first equation
of (17. 5) for n = yields

a = — I C"
t dt =

For /; 1. 2. 3, . . . , we have

t cos »* <//
n

~ n 2 n (cos nt + n* sin »0

=
and

/ sin /»/ r//

—5— (sin /tf

- /J/ COS M?)

2 COS «7I
«

2(-l)" +1

The case /? = had to be considered separately because of the appearance of/; 2 in the
denominator in the general case.
From our results the Fourier series for (17.7) is

/(,) = 2 (™i _ ^* + ™* _ . . .
)
(17.8)

The fundamental and the second, third, and fifth harmonics are sketched for one
period in Fig. 17.2. If a sufficient number of terms in (17.8) are taken, the series can be
made to approximate /(r) very nearly. For example, the first 10 harmonics are added
and the result sketched in Fig. 17.3.
As another example, suppose we have

fit) = 0, -2<t<- -1
= 6, -Kt< 1

-0, 1 < t< 2

fit + 4)= f(t)

Evidently, T= 4 and co = 2nfT = n/2. If we take t = in (17.5), we must break
each integral into three parts since on the interval from to 4/(0 has 0, 6, and
450 Fourier Methods Chap. 17

FIGURE 17.2 Four harmonics of (17.8)

FIGURE 17.3 Sum of the first 10 harmonics of (17.8)

 '

Chap. 1 7 Fourier Methods 451

values. If t I, we only have to di\ ide the integral into two parts since/(f) 6 on
I to I and /(f) on I to 3. Therefore let us choose this value of/ and obtain

a 5 l" 6dt I f Odt 6

J- 1 J I

Also we ha\e

2 I , nnt ,, 2 f ft /;7r/ .

a„ -j- 6 cos -s- at - . cos -=- <V/

4 J 2 4 J, :

12 . im
sin .
tm 2
and. finally.
3
2 f nt
/77T7 2 n /77H
/77W
,

o„ = -r
£
6 sin -=-
2" <at
,,
/?
,.

+
-•
,
,

-r
"4
f
°
..

s,n
sin —=- at
"2"
,

4 ' - 1
*- j i

Thus the Fourier series is

\2i nt 3nt 5nt

r/t \
/(/ ) = 3+-(cos
-,

T
.

-- 1

T cos- 5 -
. ]

T co St - -...)\

Since there are no even harmonics, we may put the result in the more compact form
+1
/(,) = 3 + 11 v (-l)" cos{[(2;7-
—
\)nt}/2}
71 „= i 2/7 1

EXERCISES
17.1.1 Find the Fourier series representation of the rectangular wave

f(t)=4, < < / 1

= -4, 1 < < t 2

h< 2)= f(t)

.
—16 „ sin (2/7
~
— \)nt
Ans. >.
71 ~, 2/7 — i
1

17.1.2 Find the Fourier coefficients for

f{t) = 3, 0</<l
= -1, 1 < t <4
/(' + 4)=/(0
^/7i-. a = 0, «„ = —
/77T
sin -=-, o„
2 ' "
= —
nn
I

V
1 — cos -=-
2 /
, n = 1, 2, 3, . .
452 Fourier Methods Chap. 17

17.1.3 Find the Fourier series for

f(t) = 2, < t < -§

= 0, \<t<n
fit + n) = fit)
Ans . l + ± ± sin 2(211-1)/
—
n „=
I
2« 1

17.1.4 Find the Fourier series for

/(0 = 0, -1 </< -i

4, -1 < / < I

= 0, } < <
f 1

fit + 2)= fit)

8 y (-l)" +1 cos(2/;-l);r/
2 ,

17.2 SYMMETRY PROPERTIES

If a function has symmetry about the vertical axis or the origin, then the computation
of the Fourier coefficients may be greatly facilitated. A function fit) which is sym-
metrical about the vertical axis is said to be an even function and has the property

f(r)=f(-t) (17.9)

for all /. That is, we may replace / by — / without changing the function. Examples of
1
even functions are t and cos f, and a typical even function is shown in Fig. 17.4(a).

Evidently we could fold the figure along the vertical axis, and the two portions of the
graph would coincide.
A function /(/) which is symmetrical about the origin is said to be an odd function
and has the property
/(')= -/(-/) (17.10)

In other words, replacing t by — tchanges only the sign of the function. A typical odd
function is shown in Fig. 17.4(b), and other examples are / and sin /. Evidently we can
fold the right half of the figure of an odd function and then rotate it about the /-axis
so that it will coincide with the left half.

Now let us see how symmetry properties can help us in determining the Fourier
coefficients. Evidently, in Fig. 17.4(a) we may see by inspection that for /(/) to be an
even function we have

f
" f{t) dt = 2 \"fit) dt (17.11)
J-o JO
Chap. 17 Fourier Methods 453

f(t)

(o) (b)

FIGURE 17.4 (a) Even and (b) odd function

This is true because the area from —a to is identical to that from to a. This result
may also be established analytically by writing

\' f(t)dt = C f(t)dt+ (°f(t)dt

J-a J -a Jo

= - Cf(-T)dT+ Cf«)dt
Ja Jo

= \'f(T)dt+ \"f{t)dt
Jo Jo
a
= 2 f{t)dt
Jo
\

In the case of/(?) an odd function, it is clear from Fig. 17.4(b) that

f f(t)dt = (17.12)
J -a

This is true because the area from —a to is precisely the negative of that from to a.
This result may also be obtained analytically, as was done for even functions.
In the case of the Fourier coefficients we need to integrate the functions

g(0 = f(t) cos nco t (17.13)

and
KO = /(0 sin "0) Qt (17.14)
If/(?) is even, then

g(-t) =/(-Ocos(-to»oO
= f(t) cos nco t

= g(t)
 . 7

454 Fourier Methods Chap. 1

and
K—0 =f(-0 sin (-nco t)

= —fit) sin nco t

= -hit)
Thus g(() is even and h(t) is odd. Therefore taking t = — 772 in (17.5), we have, for
fit) even,
•772
/z

a„ = 4
-= r
\ f(t) cos nay <#, n = 0,1,2, ...
* Jo (17.15)

6„ = 0, n = 1,2, 3, ...

By a similar procedure we may show that if/(/) is odd, then

an = 0, n = 0,1,2,...
2
(17.16)
6„ = —4 f<
/(/) sin nco t dt n = 1, 2, 3, . .

In either case one entire set of coefficients is zero, and the other set is obtained by
taking twice the integral over half the period, as described by (17.15) and (17.16).
In summary, an even function has no sine terms, and an odd function has no
constant or cosine terms in its Fourier series. As examples, the functions given earlier
in (17.7) and Ex. 17.1.1 are odd functions, and their Fourier series have only sine terms.
The function of Ex. 17.1 .4 is even so that its series contains only cosine terms (including
the dc case, n = 0). The function of Ex. 17.1.2 is neither even nor odd, and conse-
quently both types of terms are present.
To illustrate the procedure, let us find the Fourier coefficients of

fit) = 4. < < / 1

= -4, 1 < < t 2

f(t + 2)= fit)

which was given earlier in Ex. 17.1.1. Evidently T = 2, and thus co = n. The function
is odd, and therefore by (17.16) we have

<'„ = 0, n = 0,1,2,...

K -»r 4 sin nnt dt

= -5-P -(-1)"]

Therefore b„ = for n even and b n = 1 6/nn for n odd, as indicated in Ex. 17.1.1.
It often happens that the function /(/) is not periodic, and we shall consider an
alternative method later for this case. However, as is often the case, we may be
 n

Chap. 17 Fourier Methods 455

interested only in /(f) on some finite interval (0, T). in which case we can consider it as
periodic of period T. and find its Fourier series as before. Of course, what we have is

not the Fourier series of /(/) but of its periodic extension. We may, however, use the
result on the interval of interest and not be concerned with its periodic behavior
elsewhere. Alternatively, we may consider (0, T) to be halfxhe period and expand /(/)
as either an even or an odd function using (17. 15) or (17.16). The result, called a half-
range sine or cosine series, is easier to obtain and is equally valid on the interval of
interest.

EXERCISES

17.2.1 Find the Fourier series for the function of (17.7) using symmetry properties.

17.2.2 Find the Fourier series for the half-wave rectified sinusoid

/(/) = 0,
^ l
2 4

= 4 cos 2t,
4 4

= 0, ""
4 2

/(f + *)=/(/)

This function, shown in the accompanying figure, is a sinusoid with its nega-
tive amplitudes suppressed. Thus it is the output of a circuit that changes ac to
pulsating dc.

im
Am. a = — ;a, = 2;a„ =— 8
71(1 — n
=-\
2
)
cos -t> "
2
'
_ , A
= 2,'
3, 4,
' '

0, n = 1,2, 3, ..

37T -jr IT 37_r

4 2 2 4

EXERCISE 17.2.2
 = +

456 Fourier Methods Chap. 17

17.2.3 Find the Fourier series for the full-wave rectified sinusoid

/(/) = |4sin2/|

This function, shown in the accompanying figure, is a sinusoid with its nega-
tive amplitudes changed in sign.

Ansl
4{\ + %T^W C0 * An*)

EXERCISE 17.2.3

17.3 THE EXPONENTIAL FOURIER SERIES

Replacing the sinusoidal terms in the trigonometric Fourier series by their exponential
equivalents,

cos nco Q t = ~-(e> nw '

+ e~ inmo ')

and

sin nco t = -=(e Jno'°' — e~ Jnco ")

(see Appendix D), and collecting terms, we have

[(£^Ay-.. + (^±j^y
Kt)
f i (17.17)

If we define a new coefficient c n by

an - jb n
(17.18)

and then substitute for a n and b n from (17.5), with t = — T/2, we have
Chap. 1 7 Fourier Methods 457

T 2

H r i
f(t)(cos nco t — j sin nco Q t) dt

or. by Eulcr's formula.

if f(t)e-
Jna «
dt (17.19)

We observe also that c* (the conjugate of c„) is given by

r* _ an + jb„
2

= i r 2

-jt /(/)(cos «cy ? + j sin «a) /) t/r

J-r/2

which is evidently c_ n (c„ with /? replaced by —n). That is,

c_„ = «-+A (17.20)

Finally, let us observe that

f =4 jTy»*
which by (17.19) is

^=c (17.21)

Summing up, (17.18), (17.20), and (17.21) enable us to write (17.17) in the form

/(/) = c + f] c n e
in "°'
+ f] c-jr
ia»*

= /BO
I] cne "" + 2 c„e
;ncu °'

We have combined c with the first summation and replaced the dummy summation
index n by —n in the second summation. The result is more compactly written as

/<o= i; c**~* (17.22)

n= — oo

where c n is given by (17.19). This version of the Fourier series is called the exponential
Fourier series and is generally easier to obtain because only one set of coefficients
needs to be evaluated.
As an example, let us find the exponential series for the rectangular wave of Ex.
17.1.1, given by
458 Fourier Methods Chap. 17

f(t) = 4, < < t 1

= -4, 1< < 2 t

with T= 2. We have co = 2^/r = n, and thus by (17.19)

c„ = ij_ f(f)e-""dt

For a? ^ this is

c„ = A
f
(-4)e--""" cfr + A f
4<r-""
1
'
<#
J- 1 Jo

= 4-[i -(-i)i
Also we have

dt

r i

4rf 4 dt
= ir, '-il'

Since c„ = for « even and c„ = 8//'«7i for n odd, we may write the exponential
series in the form

f{t) = ~ t ^—re^"-^'
JTl n^=-° in — 1

The reader may verify that this result is equivalent to that obtained in Ex. 17.1.1.

EXERCISES
17.3.1 Find the exponential Fourier series for

f{t) = \, -\<t<\
= 0, 1 < 1
1 1
<2
fit + 4) = /(/)
Ans l+1 v sin (fflt/2) /2

n*0

17.3.2 The function

„ , ,
= sin x
Sa (x)
x

called the sampling function, occurs frequently in communication theory.

Show that if we define
Chap. 17 Fourier Methods 459

c = lim c„

then the result of Ex. 17.3.1 may be given by

17.3.3 Generalize the results of Ex. 17.3.2 by finding the exponential series of a train
of pulses of width 5 < T, described by

/(0 = 1, -|-</<4
=0, 4<i,i<-£
f(t + T)=f(t)
Ans.± ^Sa(^)^-/r

17.4 RESPONSE TO PERIODIC EXCITATIONS

We are now in a position to find the forced response of a circuit to any periodic
excitation that has a Fourier series. The excitation is expressed as a Fourier series,

and the response to each term of the series may be found by the phasor method. Then
the total response is, by superposition, the sum of the individual responses.
To illustrate, suppose we have an RL series circuit, with R = 6 Q and L = 2 H,
excited by a source

v (t) = 4 V, < < s t 1

= -4V, 1 < t<2s

v(t + 2) = v(t)
This excitation is a square wave, and its Fourier series was found earlier in Ex. 17.1.1
to be
sin( 2 "~ "v
Ko = -x;
1)7
n „= i
o
In — 1

Therefore we may write

<t) = ±v n

where

°-
= (2„' 6 sin(2 '| -' ) '"
l)*

= (5ri ^ c ° sK2 "- 1) '"- 90 °)

460 Fourier Methods Chap. 17

is the (2// — l)th harmonic of v(t), with frequency co„ = (2n — \)n rad/s.
The phasor voltage is

V = — /-90°

and the impedance seen at the terminals of the source at co n is

Z(jco„) = 6 + j2(2n - \)n

= 2^/9 + n\2n- lWtan" 1

(2/7 ~ 1)71

The phasor current is, therefore,

I. = |I,| /-90°-g.
where
o
II = -
(17 231
1

"'
|

(2« - 1)71^9 + n\2n - If

K
>

and

0„ = tan- 1 (2" "7 1)7r

(17.24)

The time-domain current is given by

/n (O = |I„|cos(aV-90 -0„)
= |I„| sin(ov ~ On)

Therefore, by superposition, the forced component of the current is

M»=i;|l,,|sin(av-0 n) (17.25)
n= 1

We may obtain the natural response, and thus the complete response, as we have
done in the past. A difficulty with the procedure is the necessity for finding the arbi-
trary constant in the natural response by summing an infinite series. This, of course,
also presents difficulties in finding Very often, however, a reasonable
i(t) for a given t.

number of terms can be used, with the remaining terms ignored, to approximate the
series.

The input v(t) in the example we have just considered contains not just one
frequency, as was the case for sinusoidal excitations, but an infinite number of
frequencies. This is, of course, also true of the output /(/). Since the Fourier series
displays the amplitude, as well as the phase, corresponding to each frequency, it is a
simple matter to which frequencies are playing an important role in the shape of
tell

the output and which are not. This, of course, is one of the important applications of
Fourier series and will be considered in more detail in the next section.
Chap. 17 Fourier Methods 461

EXERCISE
17.4.1 Obtain (17.25) using the exponential Fourier series. [Suggestion: Recall that

the series for v(t) was obtained in Sec. 17.3.]

17.5 FREQUENCY SPECTRA

We may write the /?th harmonic in the trigonometric Fourier series in the form of a
single sinusoid.

a„ cos nco t + b n sin nco t =A n cos (nco t + 0„) (17.26)

with amplitude

A„ = Jal + bl (17.27)

and phase

0„=-tan->^ (17.28)

In terms of the exponential series, the coefficients are

_ a„ - jb n

or, in polar form,

c. = ^ fl
-+*-/-tan-^
2 / an

Therefore we have

cn = 4?/<f>„ (17.29)

An output function /(/), given in the time domain, may be represented in the
frequency domain by A n and 0„, n = 0, 1, 2, For example, if we wish to know if
. . . .

the harmonic is dominant or not, we need only look at A n This information is

/;th .

readily obtained from a plot of A„ versus frequency. Since the frequencies are discrete
values such as 0, co 2<w etc.. the plot will be a set of lines whose lengths are pro-
. ,

portional to A„. Such a plot is called a discrete amplitude spectrum or a line spectrum
and is analogous to the amplitude response considered in Chapter 15 for the con-
tinuous case. A similar line plot of 0„ is a discrete phase spectrum and is analogous to
the phase response in the continuous case. The dc component, of course, is given by
A = a /2 and = 0, or if a <0, A = a„/2and O = 180°.
<f> {) —
462 Fourier Methods Chap. 17

As an example, the rectangular pulse of Ex. 17.3.1 had Fourier coefficients given by

nn
1

nn l
n= ±1,±2, ±3,...

Also, from Ex. 17.3.2, we may write this result in terms of the sampling function as

sin (nnjl)
A = 2\c n \

nnjl

Evidently, A„ = 0, n even, and 2/nn

1
|, n odd, while 0„ takes on only the value 0° for all
n. The amplitude spectrum is shown in Fig. 17.5 by the vertical solid lines. The dotted
curve is the absolute value of the sampling function, which is also the envelope of the
spectrum.

.-r^rvrvTV V IL VTVT\-TV-r-
5-3-10 1 3 5 7 n(co = n7772)

FIGURE 17.5 Amplitude spectrum

EXERCISES
17.5.1 Find the amplitude and phase spectra A„ and (f)„ of the sawtooth wave of (17.7).
Ans.A n =\2/n\,n = ±1, ±2, ±3,...
0„ = -90°, n = 1, 3, 5, -2, -4, -6, . . . , . . .

= 90°,« = 2,4,6,..., -1, -3, -5,...

17.5.2 Find the amplitude spectrum of

/(f) = 1, -a<t<a
= 0, a < | / 1 < 772

f(t + T)=f(t)
Chap. 17 Fourier Methods 463

where 772 > a. Note that this function is a generalization of that of Ex.
17.3.1 and that its spectrum envelope reaches zero at \/a, 2/a, 3/a, which . . . ,

are independent of T. That is, the "width" of the envelope is independent of

the pulse period. Arts. A„ = (4a/T) Sa (InanjT) |
\

17.6 THE FOURIER TRANSFORM

If a is not periodic and is defined on an infinite interval, we cannot

function f(t)
represent by a Fourier series. It may be possible, however, to consider the function
it

to be periodic with an infinite period and extend our previous results to include this
case. In this section we shall consider this case in a nonrigorous way, but the results
may be obtained rigorously if fit) satisfies the conditions of Sec. 17. 1 with (17.6)
replaced by

f(t) \dt<oo (17.30)

J" |

Let us begin with the exponential series for a function fT (t) defined to be/(/) for
-T/2 < <t T/2. The result is

MQ= 2 c n e»«" tn (17.31)

where

c„ = -^ fT {x)e-
iUnx/T
dx (17.32)

We have replaced co by 2^/Tand are using the dummy variable x instead of / in the
coefficient expression. Our intention is to let T —* oo, in which case/r (0 * f{t). —
Since the limiting process requires that a> = 2n/T — 0, for emphasis we replace

2n/Tby Aco. Therefore substituting (17.32) into (17.31), we have

M0= L-« n= ^j '

JAx)e-"»*» dx 7 jtn&co

t ^ I
fT(x)e- Jlx
-' )n/uo
dx Aco (17.33)

If we define the function

- ,)
g(co, = 2^ I Mx)e- jalx dx (17.34)

then clearly the limit of (17.33) is given by

fit) = lim
T->°° n=
2 gin^O), t)Aa) (17.35)
(Aeu-0)
464 Fourier Methods Chap. 17

By the fundamental theorem of integral calculus the last result appears to be

f(t)= fj(co,t)d(D (17.36)

But in the limit, fT —'/and 7— > oo in (17.34) so that what appears to be g(co, t) in
(17.36) is really its limit, which by (17.34) is

lim g(co, =
•*-££
t)
f( x ) e
-J^-» dx

Therefore (17.36) is actually

)e-jMx-t) dx dm
fit)
lit
£[£ f(x (17.37)

As we said, this is a nonrigorous development, but the results may be obtained

rigorously.
Let us rewrite (17.37) in the form

W> = 5 x jo
f(x)e-'™ dx e " da> (17.38)

Now, let us define the expression in brackets to be the function

F(;co) = P /(/>->' dt (17.39)

where we have changed the dummy variable from x to t. Then (17.38) becomes

/(0 = nJco)e J°" dco (17.40)

^J
The function F(y'co) is called the Fourier transform of f(t), and /(/) is called the
inverse Fourier transform of F(jco). These facts are often stated symbolically as

f(/g>) = nnt)]
(17.41)
/(0 = ff- 1
PFC/Q))]

Also, (17.39) and (17.40) are collectively called the Fourier transform pair, the sym-
bolism for which is

/(*)«-> F(/a>) (17.42)

The expression in (17.37), called the Fourier integral, is the analogy for a non-
periodic /(?) to the Fourier series for a periodic /(/). Equation (17.40) is, of course,
another form of (17.37). The Fourier transform corresponds, in this case, to the
Chap. 17 Fourier Methods 465

Fourier coefficients in the previous cases. Another description for these two analogies
is to say that the Fourier transform is a continuous representation (co being a con-
tinuous variable), whereas the Fourier scries is a discrete representation (/7co , for n an
integer, being a discrete variable).
As an example, let us find the transform of

fit) = e-°'u(t)

where a > 0. By definition we have

$[e-"'u(t)] = F e-"'u{t)e- ia" dt

= P e- {a+M 'dt
Jo

or

$[e-°'u(t)] = \ . e
-e+j»)i
-(a + jco)
The upper limit is given by

lim e""'(cos cot — j sin cot) =

t—oo

since the expression in parentheses is bounded while the exponential goes to zero. Thus
we have
5[e-"'u(t)] = —+-^-
a jco
(17.43)

or
1
e-"'u(t)
a +jco

As another example, let us find the transform of the single rectangular pulse

m=* -4 < ' < 4

=o, i,i>4

which is shown in Fig. 17.6. By definition we have

T(jco) = j"~
f(t)e-^'dt

rt/2
= Ae~'°" dt
J-S.2

__ 2A /e
icoi 2
— -
e iw62 \
- co \ —pr )
466 Fourier Methods Chap. 17

f(t)

2 2

FIGURE 17.6 Finite pulse of width 5

or

Alternatively, we may write

where Sa (co5/2) is the sampling function defined in Ex. 1 7.3.2.

EXERCISES
17.6.1 Find SFfc-"!'!], where a > 0. /1/w. 2a/(C0 2 + a2)

17.6.2 Find F(;©) if

fit) =t+U -1 </ <0

/ + 1, < < t 1

0, elsewhere

Ans. [2(1 - cosw)]/co 2 = Sa 2 (co/2)

17.6.3 Find /(r) if

F(jco) = 1, -1 <C0 < 1

= 0, |co|> 1

^/w. (1/71) Sa (0

17.7 FOURIER TRANSFORM OPERATIONS

Since the operation of obtaining the Fourier transform is that of performing an

integration, it is a linear operation. That is, the transform of the combination

v(t) = fc,/,(0 + kj 2 {t)

Chap. 17 Fourier Methods 467

is the combination of the transforms

V(/eo) = fcjF.C/o)) + k^Ajm)

where V. F,. and F : are, respectively, the Fourier transforms off,/,, and f2 and fc,
and fe 2 are constants. This concept of linearity enables us to find readily transforms of
relatively complicated functions from transforms of simpler functions. For example,
the transform of the function

fit) = 6(e~ 2 ' - e~ 3 ')u(t) (17.44)

is, by linearity and (17.43),

6 6
F(yco)
2
Z. + jco
-|-
JUJ J
3 + jo
JUJ
-f-

6
(2 + Jco)(3 + jco)
Another operation involving Fourier transforms that we shall find useful is that of
time differentiation. Suppose we wish to find the Fourier transform of the derivative of
a function /(/). By definition, if

f(t) < > F(jco)

then

fit) = 2^ f ViJco)e J°" dco

from which we obtain

Therefore we have

tiM^jcoFiJoo) (17.46)

That is. the transform of the derivative of/is found by simply multiplying the trans-
form off by jco. This result may be extended readily to the general case

^M<->iJcoy¥iJco) (17.47)

where // = 0, 1, 2 We are assuming that the derivatives involved exist and that
the interchange of operations of differentiation and integration is valid.
From the results represented by (17.47), we may show that Fourier transforms can
be used to extend the concept of network functions, considered in Chapter 14, to the
 .

468 Fourier Methods Chap. 17

case of nonsinusoidal excitations. For example, taking the Fourier transform of both
sides of (14.22) and making use of (17.47) and linearity, we have

[a n (j(o)
n
+ a^iijco)"-
1
+ . . . + a Jco + a ]\ o (jco)
-
--
[bjLjcoy I
b m ^Uco) m 1
+ . . . + bjco + boWMco)

The functions V and V, are the Fourier transforms of the output and input functions,
v and v From this result we may write
f
.

^>> - HUto)
V,Oco)

= bjjcoy + b m ^(j(or i

+ . . . + b
(17.48)
a n (jco)" + a n ^(jco) ni + . . +a
where, by (14.24), H(jco) is the network function of Chapter 14, evaluated at s =ja>.
This development indicates that the network function, defined previously as the
ratio of the output phasor to the input phasor, is precisely the same as the ratio of the
output transform to the input transform, and in the latter case the functions involved
do not have to be sinusoids. This generalization may be used also to find the time-
domain responses. For example, suppose the input v {t) is given by t

v t (t) = e-i<u(t)

and is related to the output v (t) by the equation

^+ 2v = 6» t

Transforming the equation, we have

(./co + 2)V (/co) = 6V,(./co)

or
yjijco) _ 6
H(yco)
\£]<o) jco + 2

Since by (17.43) we have

3 + jco
the transform of the output is

V (jco) = HUcoyVtUco)
6
(2 + jco)(3 + jco)
By (17.45) this is
Chap. 17 Fourier Methods 469

v -(
»^(r^)- 6
(rns)
so that by linearity, or by (17.44). we have

v (t) = 6(e" 2'

- e~ 3t)u(t)

Since the network function concept implies that there is only one input, the time-
domain responses obtained in this manner are responses of initially relaxed circuits.
(No initial energy is stored.) This is, of course, a disadvantage of the Fourier transform
technique.
We could develop general methods for using Fourier transforms to find time-
domain solutions, but there are many common functions which do not have Fourier
transforms, and the methods, as we have just noted, are limited to finding responses of
initially relaxed circuits. We prefer, therefore, to consider a more general transform

method, involving Laplace transforms, in the next chapter. As we shall see, Laplace
transforms are more versatile than Fourier transforms, in the sense that more func-
tions possess Laplace transforms. In addition. Laplace transforms may be used to
obtain the complete response, taking into account initial conditions.

EXERCISES

17.7.1 If x(t) is the input and y(t) is the output, find the network function, using
Fourier transforms, where

y" + Ay' + 3y = 4.v

Ans. 4/(3 - co 2 + jAco)

17.7.2 If in Ex. 17.7.1

x = e' 2 'u{t)

find 5[y(t)l Ans. 4/[(/eo + l)(j(0 + 2)(jC0 + 3)]

17.7.3 Verify in Ex. 17.7.2 that

jco + 1 JO) + 2 jco + 3

and find y(t) for the case y{0) = /(0) = 0.

Ans. (2e-' - 4e~ 2 '

+ 2e" 3 ')«(0
 )

470 Fourier Methods Chap. 17

PROBLEMS
17.1 Find the Fourier trigonometric series for the functions
(a) fit) = 1, < t < a
= 2,n<t<2n
/(/ + 2a) =/(/).
(b) fit) = e', —a < <n t

fit + 2n) =/(/).

(c) f{t) = t\ -\<t<\
fit + 2)= f(t).
17.2 Find the Fourier exponential series for the functions
(a) fit) = e~>, -1 <t < 1

fit + 2)= fit).

(b) /(/) = -1 </<
/, 1

=2- < t < 3.

/, 1

/(r + 4)=/(0
(c) /(r) = < <
/, r 1

/(/+ 1) =/(/).

17.3 Obtain the exponential series of the full-wave rectified sinusoid of Ex. 17.2.3
from its trigonometric series.

17.4 Obtain the trigonometric series of fit) given in Ex. 17.3.1 from its exponential
series.

.1
Ans. -=- + ^
—n2 fi
2j tt
—-
(-1)"
£—
+1
tx cos
(2/i -=—1)71/
-—
2 B (2a; 1 2

17.5 Find the first three terms of the forced response /(/) if v g it) is the function of
Prob. 17.4.

PROBLEM 17.5

17.6 Find the forced response i(t) in Prob. 17.5 if v g it) is the function of Prob.
17.1 (c).

17.7 Find the discrete amplitude and phase spectra of the functions of Prob. 17.1.

17.8 If fit) in (17.22) is the voltage across, or the current through, a 1-Q resistor,
then the instantaneous power is/ 2 (/), and the average power is
Chap. 17 Fourier Methods 471

Show that the average power is also given by

P(t)dt= £ |f n P
Jo

where cn is the exponential Fourier coefficient. This result is known as ParsevaVs

theorem. [Suggestion: Write f 2 {t) in the form

f 2 C) =22 n= — «> m=— c"°

c„c m eJ'" "e J "" ''

and integrate, noting that

I
e Jnwot e Jm<„o' dt = 0, 171, W = 0, ±1, ±2, . . .

Jo

unless m = — «.]
17.9 From Prob. 17.8 the quantity |c„| 2 is
the average power associated with the
frequency na> The plot of |c„| 2 versus frequency is a line spectrum known as
.

the discrete power spectrum of f(t). Plot this function in the cases in Prob.
17.2.

17.10 Find the Fourier transforms of the following functions:

(a) f{t) = u(t) - u{t - \).

(b) f{t) = e-°' cos btu(t), a > 0.

(c) /(/) = e-"[u(t) - u(t - 1)].

17.11 Show that

(a) fit - t) ^ e-^FiJco).
(b) FUt)<-*2nf(-co).
(c) f(-t)^¥(-jco).
[Suggestion: Substitute appropriately into the relations for F( jco) or fit).]

17.12 Find the Fourier transforms of

(a) /(/) = 1

a-

(b) /(O = ^-,«>0.

[Suggestion: Use the results of Prob. 17.11 (b), Ex. 17.6.1, and (17.43).]

17.13 If the transform of Prob. 17.10 (a) is a network function, find the amplitude and
phase responses.

17.14 If the output is /(/) and the input is v g (t) in the circuit of Prob. 17.5, find the
network function.

17.15 If in the circuit of Prob. 17.5 the output is v c (t) and the input is v e it), find (a)
the network function H(jCO) and (b) v c (t) if r c (0)= and v g it) = e~'uit) V.
[Suggestion : Note that
472 Fourier Methods Chap. 17

(A- + l)(.v + 2) x + l x + 2]

17.16 Find the amplitude and phase response of a network having input x(t) and out-
put vC) related by
y" + 5/ + 4y = 2x'

17.17 If /(/) is the voltage across, or the current through, a l-Q. resistor, then f 2 (t)
is the instantaneous power delivered to the resistor. Show that if /(f) is given
by (17.40), then the total energy delivered to the resistor is

w= f p (r) dt = JL f I
F(jco) fdco

where F(jco) = ST/COL This is the continuous version of Parseval's theorem

(see Prob. 17.8). Suggestion: Write

Pit) = fit) — | F(jCO)e'°" dCD

and integrate, changing the order of integration.)

17.18 Since in Prob. 17.17 |F(;'cy)| 2 is multiplied by frequency to yield energy, it is

sometimes called the energy density. The energy in an output/(/) in the band of
frequencies co = a to co = b is then

1
"*

Find the total energy (a = —oo,b = oo) in the output of an ideal low-pass
having input e~'u(t) and amplitude response
filter

|H(y"co)| = 1, -1 <(0 < 1

= 0, elsewhere

17.19 If the output transform is

find the energy in the output in — <

1 co < 1 and in — oo < co < co.

17.20 Find the network function if the output is y(t), the input is x(t), and they are
related by other variables q it q z> and q 3 by

dqi
-jf
= -« - 9. + i

y =Qz
 18
LAPLACE TRANSFORMS

As we have seen in the previous chapter, the Fourier transform may be used to
extend the concept of network functions to cases where the excitation is nonsinusoidal.
It may also be used in certain applications to find the forced response to an input
function which possesses a Fourier transform. With the addition of the Fourier
transform to our arsenal we are able to add networks with certain nonsinusoidal,
nonperiodic excitations to our set of solvable circuits. We have thus made considerable
progress since we first solved resistive circuits with dc excitations in Chapter 2. Along
the way we have developed differential equation techniques for circuits containing
storage elements, phasor techniques for sinusoidal excitations, generalized phasors for
damped and exponential Fourier series methods for periodic
sinusoids, trigonometric
excitations, and now Fourier transforms for certain nonperiodic excitations.
In this last chapter it remains for us to consider a most elegant procedure, the
Laplace transform method, which can be applied to a much wider variety of excitation
functions than the Fourier transform method and which gives at once both the natural
and forced responses for a given set of initial conditions. In the process we shall define
a special and highly useful function, known as the impulse function, and use its
properties to obtain general time-domain solutions to any linear circuit whose excita-
tion possesses a Laplace transform.
The Laplace transform may be obtained as a generalization of the Fourier trans-
form and is developed in this way by many authors. However, to make this chapter

self-contained and independent of the previous chapter, we shall give a completely

independent treatment of the Laplace transform. A treatment based on the Fourier
transform will be left to the problems to give the reader an optional alternative
development.

473
474 Laplace Transforms Chap. 18

18.1 DEFINITION

The Laplace transform, named for the French astronomer and mathematician Pierre
Simon, Marquis de Laplace (1749-1827), is defined by

£[/(*)] = F(5)= \~ f{t)e->'dt (18.1)

>0

The function F(s) is the Laplace transform of /(f) and is a function of the generalized
frequency, s + = a jco, We might note that our
considered earlier in Chapter 14.

definition sometimes referred to as the one-sided or unilateral transform, as dis-

is

tinguished from the two-sided or bilateral transform. The latter is defined by (18.1)
with the lower limit replaced by — oo. However, we shall not have need for this
generalization since for the circuits we shall consider the function /(f) is of interest only
for t > 0. In general, if/(0 has an infinite discontinuity at t — 0, we shall understand
that the lower limit is 0".

Comparing (18.1) with the definition of the Fourier transform of the previous
chapter, we see that

ne-"'f(t)u(t)] = £[/(/)]

The presence of the factors e~ at and u{t) accounts for the greater versatility of the
Laplace transform. For example, the Fourier transform of a function f(t) may not
exist because the condition

(18.2)
J" |/(0I<*<°°

fails to hold. It is possible, however, that for the same function /(r) a value of a may
be found so that

H \e-°'f(t)u(t)\dt = P e-°'\f(t)\dt < oo (18.3)

As a general rule, if/(/) = Ofor t < and ^[/(r)] exists, then the factor u{t) in (18.3) is

redundant, and the factor e~ at is not necessary for the existence of the integral. In this
case, the Laplace transform with s = jco is the Fourier transform.
As mentioned earlier, the Laplace transform can be used to take into account the
initial conditions prevailing in the circuit. To gain some insight into how this is

accomplished, let us consider the transform of the derivative f'(t), given by (18.1) as

£[f'0)]= rf'(t)e-"dt
Jo

Integrating by parts, we have

£[/'(')] =/(')*-" s [' f(t)e-«dt

Jo
Chap. 18 Laplace Transforms 475

[f /(f) and s are such that the integrated part vanishes at the upper limit, as will be the
case for virtually all functions we shall encounter in circuit theory, then in view of
(1 8. 1), this result becomes

-sF(s)-f(0) (1 8.4)
(it

[We are considering the case/(r) to be a continuous function on < <t oo.] Thus the
condition /(0) is automatically built into the transform of the derivative. This
initial

was not the case for the Fourier transform, as seen in (17.46).
To illustrate the computation of a Laplace transform, let us consider

/(/) = e -"u{t)

where a > 0. Then by (18.1) the transform is

¥(s) = F e-"'e-"dt
Jo

1 -f(+,

If Re s — o > —a, then the value at the upper limit is zero, and we have

£[e°'u(t)] = 1
(18.5)
s + a

In general, we shall define /(r) in (18.1) as the inverse Laplace transform of F(.v),

with the symbolism

f{t) = JB-'[F(j)] (18.6)

As in the Fourier transform, we can express the inverse transform explicitly as a

function of the transform. (See Prob. 18.6.) However, for our purposes we prefer to
obtain the inverse transform by a simpler procedure, which we shall outline in Sec.
18.3. As an example. (18.5) may be written in the form

e-"'u{t) (18.7)
s + a_

As another, simpler, but highly useful example, the transform of the step function
u{t) is given by

£[u(t)]=- j"e-"dt =± (18.8)

provided Re s > 0. An alternative statement is, of course,

£" uit)
476 Laplace Transforms Chap. 18

This function further illustrates the versatility of the Laplace transform. The step
function does not have a Fourier transform since for it (18.2) does not hold. (That is,

there is no Fourier transform in the conventional sense. It is possible by the use of

generalized functions to derive a Fourier transform of the step function, expressed in
terms of an impulse function, to be defined in Sec. 18.4. However, we shall not consider
transforms of this type here.)

EXERCISES
18.1.1 Find the Laplace transform of (a) sin ktu(t), (b) cos kt u(t), and (c)

(1 + 3e- 2 ')u(t)-

Ans. (a) p^, (b) -rj-p (c)

1 + -^
18.1.2 Find £[/«(/)] by using (18.4). Ans. 1/s 2

18.2 SOME SPECIAL RESULTS

As was the case for the Fourier transform, the Laplace transform is an integral and is

therefore a linear operation. Thus we may say that

£[cjAO + cMt)\ = r,F,(.v) + c 2 ¥ 2 (s) (18.9)

where F and F 2 are the Laplace transforms of/, and/2 and c and c 2 are constants.
:
,

Thus we may obtain a variety of transforms from two known ones. For example, the
function of Ex. 18.1.1(c) may be easily transformed by writing

£[u(t) + 3e-
z
'u(t)] = £[u(t)] + 3£[e' 2 'u(t)]

= ^-
s
+ s^-
'

+2
As another example, since we have

£[e Jk, u(t)] = — jk

s

and

£[e- Jk 'u(r)]
s -!- jk

we may write, using (18.9),

Chap. 18 Laplace Transforms 477

' gikt I g-jkl

£[cos kt u{t)] =£ u(t)

-j- £[e
ik
'u(t)] + -j- £[e- Jk 'u(t)]

_s — TT.+
jk s + jk.

s
(18.10)
s
2
+ kl

In a similar manner we may show that

£[sin kt u(t)] = (18.11)

Both of these results were previously given in Ex. 1 8.1.1.

As our next example, let us consider the transform

£>[e-"'f(t)] = P e-"f(t)e-" dt

u+a)
--
I' f(t)e- ' dt
Jo

By comparing this result with (18.1), we see that

£[<?-<"/(')] = F(j + a) (18.12)

where F is the transform of/ As a consequence of ( 18.10), (18.1 1), and (18.12), we may
write the useful transforms

s a
£[e-"' cos bt u(t)} = -\-
(18.13)

and
b
£[e' a sin
'
bt u(t)] (18.14)
(s + a)
2
+ b
2

as well as many others.

To illustrate these results, suppose it is required to find/(r), given that its transform
is

F(( 5 )} = ^
s
2
+ 2s + 5

We may complete the square in the denominator, obtaining

65
F(s)
(s \y--2 2
 1

478 Laplace Transforms Chap. 18

which we may write as

F(s)
6Q -\- l)-6
(s + \y + 1

cV s+ \

l(s+ l)
2
+ 22 . is + l)
2
+ 22

Finally, by (18.13) and (18.14). we have

f{t) = (6e~' cos It - 3<r' sin 2f)w(0 (18.15)

The justification for this last step is given in Ex. 18.2.1.

As a final example, let us find the transform of the function /(/ — x)u{t — t),
which is the function f(t) made zero for / < and delayed by x units, as shown in
Figs. 18.1(a) and (b). For this case we have

£[/(* - r)u(t - t)] - f

f(t - z)u(t - x)e-» dt
Jo

-- [° fit -T)e J
'
dt

f(t-T)u(t-t)

(b)

FIGURE 18.1 (a) Function f(t) and (b) f(t) made zero for t < and
translated x units to the right

Making the change of variable t = z -\- x, we have

£[f(t - x)u(t - t)] - [" f(x)e-* u+x) dx

Jo

= e~" f(x)e~
iX
dx
Jo
or
£[/(/ - r)u(t - t)] = e-"F(s) (18.16)

where, again, F is the transform of/.

To illustrate the use of this result, let us apply it to the problem of finding \>it), the
inverse of a product of two transforms, given by
Chap. 18 Laplace Transforms 479

y(t) =£->[G(*)F(j)]

By definition we may write

G(s) \ gix)e~"dx
Jo

so that

G(sWs)= [' g(x)[F( s )e -"] t h

Jo

By (1 8. 16) this becomes

G(s)¥(s) = f g(x)£[f(t - x)u{t -

JO
t)] dx

g(T)[j "/(f-T) W(f-T>-'*

£ o
rfT

Interchanging the order of integration and noting that u(t — x) = for x > t, we
have

G(s)FU)-- fV"|Ts(T)/(f-T)<fr dt
Jo L^o

But by definition of the transform this is

G(j)F(5) = 'g(T)/(/-T)A
£[Jo
Thus the function r(r) is given by

£-'[GU)F(5)]= \' g(x)f(t - x) dx (18.17)

JO

This result is known as the convolution theorem, and the integral involved is the
convolution of g and/, symbolized by

S(0 */(')= \' g(x)f(t - x) dx (18.18)

Jo

We shall find the convolution theorem very useful later in the chapter.

EXERCISES
18.2.1 Show that the inverse transform is also a linear operation. That is,

£" I
[c 1 F 1 (5) + c 2 F 2 (s)] = c 1 £-i[F 1 (j)] + c 2 £"'[F 2 (*)]
480 Laplace Transforms Chap. 18

Therefore the last step leading to (18.15) is justified. [Suggestion: Invert both
sides of (18.9).]

18.2.2 Find the Laplace transform of (a) te~ 2 'u(t), (b) u(t — 3), and (c) f(t) = 1,

< / < 2, and /(/) = 0, elsewhere.

Ans. (a)
(ppjjj,
(b) ^, (c) |(1 - e~ 2 ')

18.2.3 Show that the operation of convolution is commutative; that is,

f(t)*g{t) = g(t)*f(t)

[Suggestion: Make a change of variable, x = — x, t in (18.18).]

-1
18.2.4 Find <£ [l/(s 2 + l) 2 ]. [Suggestion: Use the convolution integral, where
F(s) = G(s) = \/(s 2 + 1).] Ans. ^(sin f - t cos t)u(t)

18.3 THE INVERSE TRANSFORM

It is evident from the two sections of this chapter that we may compile a lengthy
first

table of functions and Laplace transforms by repeated application of (18.1) for

their
various cases of f{t). We could then obtain a wide variety of inverse transforms by
matching entries in the table. For example, in Table 18.1 we have collected most of the

TABLE 18.1 Short table of Laplace transforms

fU) F(s)

1. S(t) 1

2. u(t)

3. e-"'iiU)
s + a
k
4. sin ktu(t)
s- + k2
s
5. cos kt u(t)
s2 + k2
h
6. e~ at sin bt ti(t)
(s + a) 2 + b2

7. e-'> cos bt«(t)

V + a
{s + g)2 + b2

8. MO h

common functions /(/) and their transforms F(^) which we are likely to encounter in
circuit theory. We have derived all of these previously except the first entry, a function
designated as 8(t) with transform F(s) = 1, which we shall define and consider in

Sec. 18.4.
To illustrate the use of the table, suppose we are required to find the inverse of the
Chap. 18 Lap/ace Transforms 481

transform

FO) -\
2
s 4 '

s I
9 s

Since, as discussed in Ex. 1 8.2. 1, the inverse transform is a linear operation, we have

+ +c-
l

/(/)= 6£"' 3£-

i

_s
2
+y
By entries 3, 4. and 2 of the table we have

f(t) = (6c*' + | sin 3* - 3)w(0

As another example, let us invert the transform

20 I- 10)
FO) (18.19)
+ 1)0 + 4)

There is no direct entry in Table 18.1 which we can use to obtain f{t) in this case;

however, we may obtain a partial fraction expansion of F(s) and apply the table to each
term in the expansion. Obtaining a partial fraction expansion is the opposite operation
of getting a common denominator. That is, we ask ourselves what simple fractions add
together to yield F(s). Since in this case the transform is a proper fraction (the numera-
tor is of lower degree than the denominator), the partial fractions will be proper
fractions and must therefore be of the form

20 + 10) s
FO)
+ 1)0 + 4) 5+1 +
The constants A and B are determined so as to make the second and third members
identities in s.

The simplest means of determining A and B is to note that

(s + ms) =XL±m=A + »k±£

and

c+ 4) FW _ 2^±m _ db±p + B

Since these must be identities for all s, let us evaluate the first at s = — 1, which
eliminates B, and the second at s = —4, which eliminates A. The results are

2(9)
,4 = 0+ 1)F0)
3
= 6

2(6) -4
B = + 4)F(s)
-3 =
482 Laplace Transforms Chap. 18

Therefore we have

F(.)
s + I s + 4
and, by Table 18. 1,

/(/) = (6e- - Ae~*')u{t)

In the case of a proper fraction with a denominator composed of distinct linear

factors, such as (18.19), the partial fraction expansion is a sum of terms which are
constants divided by the various linear factors. This is evidently the most general set of
terms whose common denominator is the denominator of the transform. If the
transform is not a proper fraction, then long division is required to obtain a proper
fraction. We shall consider only the case of the numerator not exceeding the denomina-
tor in degree, so that at most the long division results in a constant plus a proper
fraction. Higher-degree numerators require derivatives of the function 5(t) in Table
18.1.

As an example of this type, let us invert

9s 3
F(s) =
(s + l)(s
2
+ 2s+ 10)

We note that the degree of the numerator is the same as that of the denominator, so
that long division is required. It is not necessary to actually perform the division since
it is evident that it will result in 9 plus a proper fraction. Therefore let us write

9c 3
F(*) =
(s + \)(s
2
+ 2s+ 10)

= 9-1 A
+ s+ +I
Bs + C (18 20)}
1 s
2
+ 2s + 10
K
'

We know that the linear denominator term has a constant numerator and that the
quadratic denominator term has, in the most general proper fraction case, a linear
numerator, as indicated. We could factor the quadratic into two linear terms and
proceed as in the example of (18.19). However, this will result in complex coefficients,
so we choose to leave the quadratic factor intact.
We may find A, as in the previous example, from

~9
A = (s + l)F(s)
"1-2+10 -1
,«-,

To obtain B and C, let us multiply (18.20) by the denominator of F(^), obtaining

9s 3 = 9(5 + l)(s
2
+ 2s + 10) - (s
2
+ 2s -f 10) + (Bs + C)(s + 1)

(We have substituted in the value of A.) Since this is an identity in s, the coefficients of
like powers must be equal in both members of the equation. Thus the coefficients of
Chap. 18 Laplace Transforms 483

Z
5 arc
0=18 + 9-1 + 5
from which fi = — 26. The coefficients of 5° are

= 90 - 10+ C
yielding C — —80. Therefore we have

F( _ 1 __26, + 80
5
s + 1 5
s
2
+ 2s + 10
which we rearrange as

F(s) = 9 26,,
s +]
1 L(J + 1) + 3
2
_ (*+ I)
2

Therefore, by Table 18.1, we have

f{f) = 9(5(f) - e~'G + 26 cos 3/ + 18 sin 3t)u(t)

If the transform has repeated poles, such' as

4
F(5) =
(5 + 1) (5
2
+ 2)

then the procedure is slightly more complicated. Evidently in this case the partial

fraction expansion is

B
F(5)
(s + 1) (5
2
+ 2) (s + l)
2 +s+
'
1 '
s + 2
(18.21)

since the terms on the right combine to yield the proper common denominator. There
is no need to use a linear numerator for the second-degree denominator term, since

K,s +K _ 2 K,(s + 1) +K -K 2 x

(s + l) 2 "
(s + l) 2

_ A", K —K 2 x

5+1 '

(S + l)
2

We may evaluate /4 and C from

/* = (* + l)
2
F(s)| =4
U=— i

C= (s + 2)F(j) = 4

To obtain 5, let us clear (18.21) of fractions, obtaining

4 = 4(5 + 2) + B(s + \)(s + 2) + 4(5 + l)

2
484 Laplace Transforms Chap. 18

1
Equating coefficients of s yields

5 + 4 =
or B = — 4. The transform is then

Hs)
(s +
4
l)
2
s
4,4+
+ 1 '
s 2

The first term may be inverted by means of entry 8 in Table 18.1 and (18.12) and the
other terms by entry 3. The result is

f(t) = 4(te~' - e~' + e' 2 ')u(t)

As observed earlier, an explicit relation for the inverse transform is given in Prob.
18.6, but for our purposes we may generally use partial fractions and Table 18.1 to
obtain the inverse transforms.

EXERCISE
18.3.1 Find the inverse transform of

(a)
(s !- \){s + 2)(s + 3)
s nnA
(b)
(s 2 + \){s 2 + 4)'
(c)
s* + 5s + 21 s 2 + Als +
3
78
{s 2 + 9)(s 2 +2^ + 5)

Ans. (a) (<?"' — 2<r 2 '

+ e- ')u(t), (b) ](cos - cos 2t)u{t),
}
t

(c) 8{t) + (cos 3r + sin 3/ + 2e~' cos 2/)m(/)

18.4 THE IMPULSE FUNCTION

The function S{t), indicated in Table 1 (entry I) to have a Laplace transform of 1, is

known in the literature as the impulse function. It is not a function in the conventional
sense, but it can be defined and
its properties established by using generalized, or

distribution, function theory. In any case, it is a very useful function in circuit theory,
and we shall give a plausibility argument for its existence in this section.
Let us begin by considering the finite pulse of Fig. 8.2, centered about the origin, 1

of width a and height 1/a, defined by

/W =i-> -f«<i
= 0, elsewhere
Chap. 18 Laplace Transforms 485
f(t)

/a

-a/2 a/2 t

FIGURE 18.2 Finite pulse

The area under the pulse is 1 and remains fixed regardless of the value of a. If a is

made smaller, then the base of the pulse shrinks and its height increases, maintaining
the constant area of 1. toward zero, the pulse approaches an
In the limit as a tends
infinite pulse occurring over zero time but still associated with an area of 1. Such a
function is called an impulse function, or a unit impulse function, to indicate its rela-
tionship to the unit area, and is symbolized by 5{t). Graphically it is represented by an
arrow erected at t — 0, as shown in Fig. 18.3(a). A more general function is 8{t — t),

which is an impulse occurring at / = t, as indicated in Fig. 18.3(b).

6(t)

(a) (b

FIGURE 18.3 Impulses (a) 5(t) and (b) S(t - t)

Mathematically, the impulse function is defined by

5(t) - 0, I ^
(18.22)
8(t) dt = 1

J
or more generally by
8(t — t) = 0, t^x
(18.23)
d\t - t) dt = 1
J

Thus the impulse is zero everywhere except at its point of discontinuity, at which it

has an area of 1 concentrated. Physically, the impulse describes very well a force of
very large magnitude exerted for an extremely short time and in spite of its abstract
mathematical definition is quite a useful function.
An important property associated with the impulse function is the sampling
486 Laplace Transforms Chap. 18

property, described by

\" f(t)8{t-x)dt=f(x) (18.24)

where a < x < b and/(?) is continuous att = r. This property may be made plausible
by noting that since 8{t — x) is zero except at t — x, we may write

f
Ja
f(t)d(t-z)dt=- r
Jr — e
f

-x)dt
f(t)d(t-x)

Thus iff is sufficiently small, since /(r) is continuous at t, then/(/) is approximately

/(t) between x — e Thus we may
and x + e. factor /(r) out of the integral, and by the
nature of the impulse function we have

b t + f

\
f(t)d(t -x)dt = f(x) \
8(t - x) dt
Ja Jx-€

= m
The sampling property is very useful in determining the Laplace transform of the
impulse function, given by

£[d(t)]= \'e-"d(t) clt

Jo-

(Recall that in the event of an infinite discontinuity of the integrand at 0, we are using
-
as the lower limit.) By the sampling property, with x = 0, we have

£[<5(0] = e° F 8{t)dt = 1 (18.25)

This is entry 1 of Table 18.1.

Another useful result concerning the impulse function is

f(t)d(t) = f{0)S(t)
where /(/) is continuous at / = 0. This is plausible since 5(t) = for / ^ 0. It arises,

for example, when we differentiate functions that contain u(t) as a factor, such as in
Ex. 18.4.3.

EXERCISES
18.4.1 Evaluate the integral

)
(/
2
+ 3 cos 2t)5{t) dt
Ja

where (a) a = — 1 and (b) a = 1. Ans. (a) 3, (b)

Chap. 18 Laplace Transforms 487

18.4.2 Show that the pulse of Fig. 18.2 is given h\

u[t (o/2)] - u[t - (a/2)]

f() =
1

a
so that
.. ... . chid)
,im /(')
-0
= —JT
«'
a

This is a plausibility argument for establishing

<5(/) =
^ )

18.4.3 Use the results of Ex. 18.4.2 to obtain formally

^ U(t)uU)] =fVM0 '

fit)5{t)

= f'(t)u{t) + /"((>)<$(/)

18.4.4 Use the result of Ex. 18.4.3 and (18.4), which, because of the discontinuity, we
write as
£[/'(/)] = *£[/(/)] -/(0-)

to derive £[5(/)] = 1, where /"(/) = e~"'u(t).

18.5 APPLICATIONS TO DIFFERENTIAL

EQUATIONS

The Laplace transform, like the Fourier transform, may be used to solve differential
equations. However, the Laplace transform method has the advantage of yielding the
complete solution with the initial conditions accounted for automatically, as we shall
see in this section.
In Sec. 18.1 we have already established the result of (18.4), which we repeat here
as

£[/'(/)] = jF(5)-/(0) (18.26)

If we formally replace/ by /'. we have

£.[f"(t)] = s£[f'(t)]-f'(0)
or. by (18.26).
£[/"(')] - 2
s ¥(s) - sf(0) - /'(0) ( 1 8.27)

We may replace /by/' again in (18.27) to obtain £[/'"(()], and so forth, with the
general result being

£[/
(fl,
(0] = s"F(s) - s" </(0) - 5"" 2/'(0) -- • • • -/"-"(0) (18.28)

where (n)
The functions//', ,/'" n are assumed to be
/ is the /?th derivative. . . .
488 Laplace Transforms Chap. 18

(n)
continuous on (0, oo), and f is continuous except possibly for a finite number of
finite discontinuities. As in Ex. 18.4.4, if there is an infinite discontinuity present in

any of the functions at t = 0, then is replaced by 0".

Evidently, if we transform both members of a linear differential equation with
constant coefficients by (18.28), we see that the result will be an algebraic equation in
the Laplace transform. Also, the initial conditions will be accounted for. We may then
solve for the transform and invert it to yield the time-domain answer.
As an example, let us find the solution x(t), for / > 0, of the system of equations

x" + 4.v' + 3.y = e' 2 '

.y(0) = 1 , a-'(O) = 2

Transforming by means of (18.26) and (18.27), we have

s
2
X(s) - s - 2 + 4[sX(s) - 1] + 3X(j) = 1

from which
2
+ 8s + 13
x( W_ v

(s +
s
V)(s + 2)(s + 3)

The partial fraction expansion is

Y 3 1 1

Ms)
, ,

s + 1 s + 2 s + 3

so that the time-domain answer, for / > 0, is

x(t) = 3e~' - e~ 2 ' - e~ 3 '

(Since we are interested in / > 0, we have not needed to consider the step function
involved in the transforms.)
We may also handle certain integrodifferential equations directly without differ-
entiating to remove the integrals. Toward that end we need the transform of g(t),
given by

g(0= Jo[7(0* (18.29)

Differentiating, we have
g\t)=fit) (18.30)

so that by (18.26), since g(0) = 0, we have

£[g'(t)} = s£[g(')}

or

£[*(')] =- y £[g'(0]
Chap. 18 Laplace Transforms 489

Therefore, by (18.29) and (18.30), we have

FC£)
£ t)dt (18.31)
s
Jo

To illustrate the use of (18.31), suppose we have an RLC series circuit with
R = 2Q, L = 1 H, and C= | F, excited by a unit step function, v g (t) = u{t) V.
Then the current i{t) satisfies the equations

|- + 2H 5 f'/<ft = tt(0

1(0) =
Transforming, by (18.26) and (18.31), we obtain

s\{s) + 21(5) + ±\{s) = -

or
1

«*) = =r
2
+ + 2
+
s 2.y 5 "
2L(^+ l) 4.

Therefore the step response is

/(?) = U-' sin 2? A

Suppose in the general case that the input v t and output v are related by the
differential equation

d"v,
'dr [
d"'v
dt
-• +«i^ + flo«.

-{-b^ + boV;
fife

and that the initial conditions are all zero; that is,

t'o(0)
rfy.(0) tf-'^O) __ v{0) _ dvjjO)
dt dt"' 1
^ dt"'-'

Then, transforming, we have

5"-
(a„s" + fl B - 1
1
- •
«,* -: fl )V„(j)

= (b m s m +&„_,*" 6,5 + 6 )V,(5)

or

HUj - - U8 JZ;
a n . lS - +a + '

V,(5) a n s" + 1
+ lS aQ
490 Laplace Transforms Chap. 18

Comparing this result with (17.48), obtained in the previous chapter, we see that H(s)
is the network function, with/a> generalized to s. Or, going back to (14.24) of Chapter
14. H(.v), the ratio of the Laplace transforms of the output to the input, is precisely the
network function as it was Moreover, since admittances and
originally defined.
impedances are network functions, they are precisely the same in the Laplace trans-
form domain as they were in the case of the generalized phasors. With Laplace
transforms, however, the situation is more general. We may have any function that is
Laplace transformable as the input or output, instead of being restricted to damped
sinusoids.
To illustrate with the last example, if the input of the RLC series circuit is v g (t)

and the output is /(/), then we have

HW-^--
V g (s)
-J-
Z(s)
or

H(j) = I

sL +R+ (\/sC)

_ 5
~ (18.33)
s
1
+ 2s + 5
Then

1(5) = U(s)V g (s) = ^

s
2
+ 2s + 5
as before.
In the general case of (18.32), the output transform is given by

V (s) = H(5)V,(.v) (18.34)

Therefore if ?',(/) = (5(0. in which case V,(^) = 1, we have

\ (s) = H{s)

In this case v (t) is called the impulse response (response to an impulse) and is denoted
by h(t). That is,

v {t) = h(t) = £~ 1
[H(s)} (18.35)

The impulse response is thus very important since its transform is the network func-
tion itself.

Another point should be made in this section. Since, in general, \ (s) is by (18.34)
a product of two transforms, we may write immediately the output v (t) by means of
the convolution theorem of (18.17). The result is

v o (t) = h{t) * v£t) = P h(x)v t (t - t) dz (18.36)

Jo
Chap. 18 Laplace Transforms 491

Therefore if we know the network function and the input, we may find the forced
response directly.
As an example, for the RLC series circuit considered earlier in this section, the

network function given by (1 8.33) has the inverse transform

h(r) e '(cos It — \ sin 2t)u(t)

and the input is u(t). Therefore the output i(t) is given by

i(/) = A(0*"(0

I e
r
(cos 2t — \ sin 2x)u(r)u(t — x) dx
•'o

I e
T
(cos 2t — £ sin 2t) f/r
Jo

= £e~' sin It

as before.
The impulse response of this circuit is. of course, /?(/). As in the case of the step
response, with an impulse input there is no initial stored energy at / = 0~. Thus /;(/)

is the complete solution.

In the case of a set of simultaneous differential equations, we have only to trans-
form each equation, solve the resulting algebraic equations for the transforms of the
unknowns, and find the time-domain solutions by inverting the transforms. For
example, suppose we have the system, valid for / > 0,

2x' + / - y= \

x' +x+y= o

x(0) = 2, r(0) = -
Transforming, we have

2s\(s) - 4 + sY(s) + l - \(s) = —

s

sX(s) - 2 + X(5) + sY(s) +1=0

from which

v, v 2^ + 2

Therefore, for / > 0, we have

x(t) = 2(cos t + sin t)

In a similar manner we obtain

y(t) = —1 — 4 sin /
492 Laplace Transforms Chap. 18

EXERCISES
18.5.1 Solve for .v(r), for / > 0, by using Laplace transforms:
(a) x" + 3.v' + 2.y = <5(/).
(b) x" + Ax' + 3.v = 4e~'

.v(0) = ,v'(0) = 4. Ans. (a) e~' - e~ l >, (b) e~'{2t + 7) - 3e~ 3 '

18.5.2 Given the network function

H(.) -
2( * + 2)
{S + 1)(5 + 3)

find (a) //(/), (b) the step response, and (c) the forced response to an input e~ 2 '.
Ans. (a) (<?"' + e~ 3,
)u(t), (b) j(4 - 3e~' - <T ')«(0,
3

(c) (e~' - e- )u(t) 3!

18.6 THE TRANSFORMED CIRCUIT

When dealing with phasors we were able to omit the steps of writing the differential
equations by going directly to the phasor circuit. The circuit was analyzed by resistive
circuit techniques, the phasor output was found, and from this the time-domain
solution was obtained. We may perform a similar type of analysis with Laplace
transforms by first obtaining a transformed circuit. The rest of the procedure is similar

to that of the phasor method except that with Laplace transforms we are able to deal
with more general functions and obtain the complete solution with the initial condi-
tions satisfied as well.
To see how this may be done, let us consider first a resistance R, with current i R and
voltage v R for which
,

vR = Ri R

Transforming this equation, we have

y^s) = ia^s) (18.37)

which is represented by the transformed resistor element of Fig. 18.4(a). Next, let us
consider an inductor L for which

vL = A dt
Transforming, we have
VL(s) = sLIL (s) - LiL (0) (18.38)

This result is represented by the transformed inductor element of Fig. 18.4(b). The
initial condition is taken into account by the included voltage source. In the case of a
Chap. 18 Laplace Transforms 493

I (s)
R

ys)

— 6

(a) (b) (C)

FIGURE 18.4 Transformed circuit elements

capacitor C we have

i c dt '
r r (0)

Transforming, we obtain

V c(.v) -L lc( s ) + ±vJ0) (1 8.39)

sL S

since for t > the transform of a constant A' is K/s. The last result is represented by
the transformed capacitor of Fig. 18.4(c). The initial condition is provided for by the
included voltage source.
If we have dependent sources, we may transform them in the same manner. For
example, if we have a controlled voltage source defined by

f, = Kv 2

then in the Laplace transform notation we have

V,(.v) = K\ 2 (s)
which in the transformed circuit is a source controlled by a transformed variable.
We may write the time-domain equations of a circuit and solve them by Laplace
transform methods, or we may replace the circuit by the transformed circuit. In this
case the passive elements are replaced by transformed elements such as those of Fig.
18.4. the controlled sources are appropriately transformed, and the independent
sources are labeled with the transforms of their time-domain values. Then we may
solve the transformed circuit using resistive circuit methods.
As an example, suppose we require /'(/), for / > 0, in Fig. 18.5(a), given that
;(0)= 4A and r(0) = 8 V. The transformed network is shown in Fig. 18.5(b), from
which we have
+ 3)] + -
W_
Ir
[2/(s
3 +s+
4
(2/s)
(8/5)
(18.40)

This may be written

13 20
Us)
j+1 t j+2 5+3
494 Laplace Transforms Chap. 18

i(t)

3n I H

2e" 3t V
© v(t)

Ks)
3
-nm^ —Qy

hO s

FIGURE 18.5 (a) Circuit and (b) its transformed counterpart

so that, for / > 0,

i(0 = - 13f + '

20e" 2 ' - 3e~ 3
'

As a check, the time-domain equations for Fig. 18.5(a) are

di
+ 3/ i di + 8 = 2e~
dt

/(0) = 4
Transforming, we have

from which 1(5) is

sl(s) - 4

given by (18.40).
+ 31(5) + j-I(s) + A=
^
Jf we are interested in nodal analysis, we need to solve (18.37), (18.38), and (18.39)
explicitly for the currents. The results are

I*(*)
R

I c (5) - sCY c(s) - Cv c {0)

As the reader may verify, the transformed elements describing these equations are
shown in Fig. 18.6. The labels on the passive elements are impedances, as in Fig. 18.4.
Chap. 18 Laplace Transforms 495

+
T IIrIs) I (s) I (s)
L c

VR (s) \ R )!j
r r v (s)
Cp Cvc (0)
cii c

(o) (b) (c)

FIGURE 18.6 Transformed elements useful for nodal analysis

All the network theorems that apply to resistive and phasor circuits apply to the
transformed circuit. An added advantage in the latter case is that initial conditions may

be incorporated. As an example, the reader may verify that the elements of Fig. 18.6
are the Norton equivalents of those of Fig. 18.4.

EXERCISES
18.6.1 Solve Prob. 14.19 using the transformed circuit method if r<0) = 10 V.
Ans. 10(2 - e~') + {e~' sin It V
18.6.2 Find the voltage <•(/), for / > 0, across the terminals of the RLC parallel circuit,
excited by a current source i
g (t)
= e' 1 A,
'
using the transformed circuit
method. The passive elements are R = 4fi, L = 5 H, and C == ^o f the ^

initial inductor current is / L (0) = 1 A; and v(0) = 2 V. (The current i g is

directed into the positive terminal, and the current iL is directed out of the
positive terminal of v.) Ans. —\4e~' + 20p~ 2 ' — 4e~*' V

PROBLEMS
18.1 Find the Laplace transform of the functions
(a) «(/) - //(/ - 1).

(b) e -•[u(t) - u(t - 1)].

(c) cosh kt u(t\.

(d) sinh kt u(t).

18.2 Find the inverse Laplace transforms of

s
(a) , b j± a.
is + a)(s + b)
496 Laplace Transforms Chap. 18

(b)
(s + \\s + 2)

(c)
s + 1

(d)
(s + \)(s + 2)

18.3 Show that if /(f + T) = /(/), then

[Suggestion : Write the transform as an infinite series of integrals over (0, T),
(T, 2T), . . . , and sum.]
18.4 Use the results of Prob. 18.3 to obtain the transforms of the functions
(a) /(f) = 1, <t< 1

= 0, < < 1 f 2
fit + 2) = /(f).
(b) /(*) = sin |
/|.

18.5 Note that if />(/) is a pulse of some shape of finite duration, occurring on
< < t T[p(t) = elsewhere], and that if/(/) is defined by

fit) = p(t), 0<t<T

f(t + T)=f(t)

then the result of Prob. 18.3 is

£[p(t)]
£[/(/)] =
1 - e-

Use this fact to find the inverse transform of

(I — e ~s\l
s(\ - e' 2s )

18.6 In the Fourier integral (17.37), replace /(f) by e~"'f(t), and show that if/(f) = 0,
for t < 0, then

'« = Tn f(x)e~ sx dx dCO

o

In the second integral, replace co by (s — a)\j and obtain

fit) - ^7 e "F(s) ds

This is the inverse Laplace transform analogous to (17.40) in the Fourier trans-
form case and is valid if a is sufficiently large to ensure that (18.3) holds.

18.7 Find/(/) * g(t), where /(/) = tu{t) and g(t) = e~'u(t).

18.8 Solve Prob. 18.7 by finding F(s) and G(V) and using the convolution theorem.
Chap. 18 Laplace Transforms 497

18.9 Using Laplace transforms, solve the following for / > 0:

(a) x' x 0, .v(0) I. \ (0) I.

(b) x' 2 v Zx 0, .v(0) = 0, v'(0) I

(c) A - 2.v" ] - 2.v' 0. .\(0) \ (0) 1,.v"(0)=2.

(d) x' '
! 4a-' i 3a- 4 sin t I
8 cos t, .v(0) - 3, v'(0) = -1.
(e) x' 4a' I 3v 4e~ 3 ', a(0) a'(0) =0.
(f) x' 4a' I
3x 4c' |
8«r 3r a(0) , c'(0) 0.

18.10 Show by means of the convolution theorem that

Use this result to solve, for / 0.

x" ^\ 6 cos 3/, a(0) = .v'(0) =

18.11 If x(f) is the input and >(/) is the output, find the step response r(t) and the
impulse response h(t) for the following:

(a) y" + 6y' + 5y = 20a.

(b) /' +4/ + \3y = 13.v.
18.12 In terms of the network function H(.v), the impulse response hit) and the step
response r(t) are given by

His)!
hit) =£-'[H(.v)]. r(0 = £-'|^]
so that, by (18.31),

•(/) = I" h(t)dt

Jo-
Therefore
dr(t)
h{t)
dt

Verify the last result for the examples of Prob. 18.1 1.

18.13 Solve for x, for / > 0, if

X (/) <{f) |'.V(T)£- "<

T
dx = 0, A(0) = 1
Jo

18.14 If a network has an impulse response

//(/) = te~ 2 'u(t)

find the forced response to an input

Jit) = c~ l! cos t u{t)

18.15 Solve for x and y, valid for / > 0:

2a-' + 4a- +/ + 7 y =-

x x +y + ly = d(t)
498 Laplace Transforms Chap. 18

18.16 Find the network function and the impulse response if the output is /'(/).

ilt)

PROBLEM 18.16

18.17 Find the response of a circuit with the network function

His)
s + 1

and no initial stored energy to the input /(f) of Prob. 18.4(a). Suggestion:
Find the inverse transform involved by the procedure of Prob. 18.5, but note
that if p(t) ^ 0, for t > T, then the inverse is not periodic but is

pit) + pit - T) + pit -IT) + ...

18.18 Solve Prob. 9.9 using transformed circuits.

18.19 Solve Prob. 9.21 using a transformed circuit.

18.20 Find the network function of Prob. 9.22, and show that the circuit is a band-
passfilter. Find also (o and Q.
18.21 With the presence of impulse currents or voltages in a circuit it is possible,
theoretically, to change inductor currents and capacitor
instantaneously
voltages. Demonstrate this by finding v(i) and /'(/) for — oo < / < oo in
the given circuit. The switch is closed at / = and v(0~) = 0.

i(t)

6V
PROBLEM 18.21

18.22 The impulse response of a circuit is

h{t) = vTr'^ 1 sin

si "it)
V^2

Find the network function and the amplitude response, and thus show that
the circuit is a second-order Butterworth low-pass filter with co c = 1 rad/s.
 APPENDIX

A
DETERMINANTS AND
CRAMER'S RULE

The solution of simultaneous equations, such as those often encountered in circuit

theory, may be obtained relatively easily, in many cases, by the use of determinants.
We define a determinant as a square array of numbers having a numerical value, such
as

(A.l)
a,,
'21 a,.
'22

In this case the determinant is a 2 x 2 array, with two rows and two columns and a
value A defined to be

a, ,a, C/pfl, (A.2)

In the second-order, or 2 x 2, case of (A.l). the method of obtaining A in (A.2)

may be thought of as a diagonal rule. That is,

«I1«22 - "12^2! (A.3)

or A is a difference of the product a ,a 22 of elements

]
down the diagonal to the right
and the product a 12 a 2l of elements down the diagonal to the left.

As an example, consider

1 2
-3 4

which is given by

A = (1)(4) - (2)(-3) = 10

499
500 Determinants and Cramer's Rule App. A

A third-order, or 3 x 3, determinant, such as

12 0.3

22 «23 (A.4)

32 #33

has three rows and three columns. It may also be evaluated by a diagonal rule, given by

aX 13

^a><A
= («llO22033 + 01202303! + 01 3 03202l)

- (tf 13 022031 + 023032011 + a 33 a 2l Cl l2 ) (A.5)

Thus the value of the determinant is a difference of products of elements down the
diagonals to the right and products of elements down the diagonals to the left.

An example of a third-order determinant and its evaluation is given by

1 1 1

2 -1 1

-1 1 9

= [(1)(-1)(2) + (1)(1)(-1) + (1)(1)(2)]

-[(1)(-1)(-1) + (1)(1)(1) + (2)(2)(1)]

= -7 (A.6)

A general definition of determinants may be given and used to derive a number of

evaluation procedures. This is the technique usually given in elementary algebra books.
However, for our purposes we shall use the diagonal rules that we have considered for
second- and third-order determinants and evaluate higher-order determinants by the
method of expansion by minors, or cofactors.
The minor A u of the element a u in the /th row and /'th column of a determinant is
the determinant left after the /th row and the /th column are removed. For example,
in (A.6) the minor A 2{ of the element o 21 = 2 (second row, first column) is

1 1

A 2l
1 2

The cofactor Cu of the element a (j is given by

C,j = {-l)<+JA IJ (A.7)

In other words, the cofactor is the signed minor, the minor multiplied by ± 1 with the
App. A Determinants and Cramer's Rule 501

sign depending on whether the sum of the row number and column number is even
or odd.
The value of a determinant is the sum of products of the elements in any row or
column and their cofactors. For example, let us expand the determinant of (A. 4) by
cofactors of the first row. The result is

«nCii + a xl C v ,

or. by (A. 7).

A = fl I1 (-l)' + M +a ll 12 (-iy+*A 12

Writing out the minors explicitly, we have

A = au — a. + «u
a-n a,

By the diagonal rule this may be written

A = a n {a22 a 33 - a23 a 32 ) - a 12 (a 21 a 33 - a 2i a n ) + a 13 (a 21 a 32 - a 22 a 31 )

which may be simplified to (A. 5)

To illustrate expansion by minors, let us evaluate the determinant of (A. 6) by
applying the technique to the third column. We have

2 - +
H-l) 1

-1
H-D 2 :
2(— 1) 3+;

(1 + + 2(- -7

A method of obtaining solutions of simultaneous equations by determinants,

referred to earlier, is called Cramer's rule. We shall illustrate the method by applying it

to two systems of equations, a second-order and a third-order system. The examples

will be given in such a way that the generalization to higher-order systems should be
evident. The second-order system that we shall consider is the set of equations

x, - 2.v = 2 5
(A.8)
6x, + x = 2 4

We define the determinant of the system as the determinant A whose first column is the
coefficients of x u whose second column is the coefficients of .y 2 , etc. In the case of the
system of (A.8) we have

1 -2
13 (A.9)
6 1
502 Determinants and Cramer's Rule App. A

Cramer's rule states that

A
(A. 10)
x, =
etc.

where A, is A with column replaced by the constants in the right members of

its first

the system of equations, A2

is A with the second column so replaced, etc. In the

example of (A. 8), the right members of the equations are 5 and 4 so that

5 -2
A, = 13
4 1

1 5
A, = -26
6 4

By Cramer's rule the solution of (A. 8) is given by

x, = A 1 _13_ 1

13

-26
2
= -2
2
"A 13

As an example of a third-order system, let us consider the equations

x t
+ x2 + x3 = 6

2xj — x2 + .v 3 = 3 (A.ll)

x2 + 2x 3

The solution, by (A. 10), is

6 1

3 -1 1

7 1 2 -7
1 1 1 -7
2 -1 1

-1 1 2

1 6 1

2 3 1

x, = -1 7 2 -14
-7
App. A Determinants and Cramer's Rule 503

and
I 6

1 3

7 -21
-7 =
I

3
-7
 APPENDIX

B
GAUSSIAN ELIMINATION

Simultaneous equations, such as (A. 8) in Appendix A, may be solved by a method of

successively eliminating one unknown at a time, in a certain systematicmanner. One
such procedure, known as Gaussian elimination, will be given here for the special case
of (A.l 1), repeated as
.v r
+ x2 + x3 = 6

2.y, -.v 2 + x3 = 3 (B.l)

-.v, + x2 + 2x 3 = 7

It will be clear from the procedure how to generalize it to any system.

We may systematically eliminate .v, from all the equations of ( B. 1 ) except the first

equation by subtracting twice the first equation from the second and adding the
first equation to the third equation. We shall then have the first equation and the
two new equations neither of which contains .v,. The result is

.y, + x2 -\- x3 = 6

-3x 2 - x3 = -9 (B.2)

2x 2 + 3.y 3 = 13

Let us now divide the second equation by —3, resulting in

.Y, + x2 + x3 = 6

x2 + jx 3
= 3 (B.3)

2.y 2 + 3.y 3 = 13

We next eliminate x 2 from all the equations except the first two. In the simple case
of (B.3), this means eliminating x 2 from the third equation. This may be done by

504
App. B Gaussian Elimination 505

subtracting twice the second equation from t he third equation, resulting in

•v, + x2 x3 =6
xm [
jX 3 — 3 (B.4)

7
J*3

We now divide the third equation by I, yielding

.V, + a- 2 + x3 = 6

v, \x 3 = 3 (B.5)

x3 = 3

In a higher-order system we would repeat the procedure. That is, eliminate .v 3 from
all the equations except the first three. Eventually we reach the point, as in (B.5),

where the last unknown is found (x 3

in this case). We then back-substitute the last

unknown unknown. These two answers

into the next-to-last equation to find another
are then back-substituted into the next equation to get another unknown, and so forth.
To illustrate the back-substitution process, from the second equation of (B.5) we
have
x2 = 3 - $(x 3)

3 - j(3)

= 2 (B.6)

The known values of x2 and x 3 are then substituted into the first equation of (B.5),
yielding
•v, = 6 — x2 — x3
=6-2-3
= 1 (B.7)

In this example, this last step completes the process, since all the unknowns have been
found.
The steps in the Gaussian procedure may be put in compact form by leaving out
the symbols for the unknowns, the addition operation signs, and the equal signs. That
is, (B.l) may be represented by

2-113 (B.8)

1 12 7

Such an entity is sometimes called a matrix. It has rows and columns like a deter-
minant, but it is is 3 x 4) and has no
not necessarily a square array (the example given
number associated with it. The matrix merely represents the equations. That is, its
first row means one .v, plus one x 2 plus one x 3 equals six, etc.
506 Gaussian Elimination App. B

We may eliminate unknowns in (B.l) by manipulating rows in (B.8). Leaving the

first row intact, subtracting twice the first row from the second, and adding the first

row to the third, we have the new matrix

1 1 1 6

-3 -1 -9 (B.9)

2 3 13

This represents the system of (B.2).

Dividing the second row of (B.9) by —3 yields

1 1 1 6

1
i 3 (B.10)

2 3 13
This represents (B.3).
Subtracting twice the second row of (B.l 0) from the third row gives

1116
(B.ll)

.0 $ 7.

which represents (B.4). Dividing its last row by ^ gives

1116
1
\ 3 (B.12)

13
The last result is equivalent to (B.5) and may be used, as in (B.6) and (B.7), to obtain
the solution.
Alternatively, we may continue the elimination process as follows. In (B.12),
subtract from the second row ^ times the last row, and subtract the last row from the
first row. This results in

110 3

10 2
13
Finally, subtract the second row from the first, obtaining

1 1

1 2 (B.13)

1 3_

The last result represents x x

— I, x2 — 2, and x 3 = 3, which is the solution.
This procedure of continuing the elimination process to the form (B.13) is known
as the Gauss-Jordan method.
 APPENDIX

C
COMPLEX NUMBERS

From our earliest training in arithmetic we have dealt with real numbers, such as 3,
— 5, $, 7i, etc.. which may be used to measure distances in one direction or another
from a fixed point. A number such as x that satisfies

x1 = -4 (C.l)

is not a real number and is customarily, and ijnfnrtnnatply, rai led an imaginary
number. To deal with imaginary numbers, an imag inary unit, denoted by /', is defined

j = J~\ (C2)

Thus we have j 2 = — \,j = —j,j4 = 1, etc. (We might note that mathematicians use
3

the symbol i for the imaginary unit, but in electrical engineering this might be confused
with current.) An imaginary number is defined as the product of/ with a real number,
such as x = /2. In this case x 1 = (y'2) 2 = — 4, and thus .y is the solution of (C. 1 ).

A complex number is the sum of a real number and an imaginary number, such as

A = a + jh (C3)

where a an d b are real. The complex number A has a real part, a, and an imaginary
which are sometimes expressed as
part, d,

a = Re A
b = Im A

It is important to note that both parts are real, in spite of their names.
The complex number a - jb may be represented on a rectangular coordinate
plane, or a complex plane, by interpreting it as a point (a, b). That is, the horizontal

507
508 Complex Numbers App. C

coordinate is a and the vertical coordinate is b, as shown in Fig. C.l, for the case
4+y"3. Because of this analogy with points plotted on a rectangular coordinate
system, (C.3) is sometimes called the rectangular form of the complex number A.

IMAGINARY AXIS

3 • 4 + 3 ,

12 3 4
REAL AXIS
FIGURE C.1 Graphical representation of a complex number

The complex number A — a + jb may also be uniquely located in the complex

plane by specifying its distance r along a straight line from the origin and the angle 6
which this line makes with the real axis, as shown in Fig. C.2. From the right triangle

A = a + j b

FIGURE C.2 Two forms of a complex number

thus formed, we see that

r = V« + b2 2

(C.4)
= tan-' —
and that
a = rcosO
(C.5)
b = r sin 9
App. C A Complex Numbers 509

We denote this representation of the complex number by

(C.6)

is sometimes denoted b\
—
which is called the polar form. The number

r
/• is called the amplitude, or magni tude, and

The number 9 is the angle or argument and is often denoted by

.v = ang A - arg A

FIGURE C.3 Two forms of the complex

number A

One may easily convert from rectangular to polar form, or vice versa, by means of
(C.4) and (C.5). For example, the number .4, shown in Fig. C.3, is given b>

A = 4 +./3 = 5/36.9

since by (C.4)
= J¥ + 3 = 5
r
2

Q = tan | = 36.9 1

The conjugate of the complex number A = a -\- jb is defined to be

A* = a - jb (C.7)

That is, / is replaced by —j. Since we have

|
A* |
--=
Ja 1
+ (-b) 2 = Ja 2
+ b1 = \A\
and

arg A* - tan" 1
(—\ = — an -arg A

we may write, in polar form.

(r/0)* = r/-9 (C.8)

510 Complex Numbers App. C

We may note from the definition that if A* is the conjugate of A, then A is the
conjugate of A*. That is, (A*)* = A.
The operations addition, subtraction, multiplication, and division apply to
complex numbers exactly as they do to real numbers. In the case of addition and
subtraction, we may write, in general,

(a + jb) + (c + jd) = (a+ c) + j(b + d) (C.9)

and
(a+jb) - (c +jd) = {a-c) +j(b - d) (CIO)

That add (or subtract) two complex numbers, we simply add

is, to (or subtract) their
and their imaginary parts.
real parts
As an example, let A = 3 +y'4 and B = 4 — j\. Then
i 4 + J B = (3 + 4)+y(4- D = 7+./3

This may also be done graphically, as shown in Fig. C.4(a), where the numbers .4 and

B are represented as vectors from the origin. The result is equivalent to completing the
parallelogram or to connecting the vectors A and B in head-to-tail manner, as shown
in Fig. C.4(b), as the reader may check by comparing the numbers. For this reason,

complex number addition is sometimes called vector addition.

A + 8

(a) (b)

FIGURE C.4 Two methods of graphical addition

In the case of multiplication of numbers A and B given by

A =
+ jb = i\a cos 0, +./>i sin 0,
(C.ll)
B = C + jd = r 2 cos 9 2 + ji'i sin Qz
we have
AB = (a +jb)(c +jd) = ac + jad + jbc + j 2
bd
= {ac - bd) +j(ad+ be) (C12)

Alternatively we have, from (C.ll),

App. C Complex Numbers 51

AB = (r, cos#, f./''i sin #,)(>; cos 0, I jr 2 sin 9 Z )

= r,r 2 [(cos 0, cos 9 Z sin 0, sin 9Z) '

/(sin 0, cos 2 -f cos 0, sin 0,)]

/,,-.[cos(0, ;
; ) i
./sin «9, + : )]

Therefore in polar form we have

(rjejXrJjh) = r,r 2 /9,+9 2 (C.13)

and hence we may multiply two numbers by multiplying their magnitudes and adding
their angles.
From this result we see that

AA* = (r/9)(r/-9) = r
2
/0=\A\ 2 /0
2 2
Since |
/J
|
/0 is the real number |
/I
|
, we have

|/*|
2
= /*/** (C.14)

Di\ision of a complex number by another, such as

A_
= a + jb
B ~
TV
c + jd
an irrational denominator, since /
results in ^J = '— l . We may rationalize the denomi-
nator and display the real and imaginary parts of N by writing

N = d&l ^ a+Jb c-jd

BB* c +jd c - jd
'

which is

(ac + bd) + j(bc - ad)

We may also show by the method used to obtain (C.13) that

^f = ^/^ (C6)

As examples, let A = 4 + y3 = 5 /36.9° and 5= 5 +./12 = 13 /67.4° . Then we

have
AB = (5)(13) /36.9° + 67.4° = 65 /104.3°

and

4 = A /36.9° - 67.4° = 0.385 /-30.5"

51 2 Complex Numbers App. C

Evidently, it is easier to add and subtract complex numbers in rectangular form

and to multiply and divide them in polar form.
If two complex numbers are equal, then their real parts must be equal and their

imaginary parts must be equal. That is, if

a + jb == c -f jd
then
a — c = j(d — b)
which requires
a = c, b — d

Otherwise we would have a real number equal to an imaginary number, which, of

course, is impossible. As an example, if

1 +x+j(S-2x) = 3+jy
then
1 +x = 3

8 - 2x = y
or x = 2, y = 4.
 APPENDIX

D
EULER'S FORMULA

To derive Euler's formula, an important result, let us begin with the quantity

g = cos 9 + jsir\0 (D.I)

where 6 is real and j = *J—\. Differentiating, we have

^ = V(cos0^ ;sin0)-./g

as may be seen from (D.I). The variables in the last equation may be separated,
yielding

g
Integrating, we have
lng = j»H K (D.2)

where A' is a constant of integration. From (D.l) we see that g = 1 when t) — 0,

which must also hold in (D.2). That is,

In 1 = = +K
or K— 0. Therefore we have

In g = j6
or

g = e'
e
(D.3)

513
 —
514 Eu/er's Formula App. D

Comparing (D.l) and (D.3). we see that

e
je
= cos0 + y sin 9 (D.4)

which is known as Euler's formula. An alternative form found by replacing 9 by —9 is

e -><> = cos (-9) +./'sin {-9)

which is, equivalently,

e~ je = cos 9 —jsind (D.5)

Evidently. (D.4) and (D.5) are conjugates.

Euler's formula provides us with a means of obtaining alternative forms of cos 9
and sin 9. For example, adding (D.4) and (D.5) and dividing the result by 2, we have

pie ~j6
i

p
cos 9 = - l_£_ (D.6)

Similarly, subtracting (D.5) from (D.4) and dividing the result by 2j, we have

pie
e _ p-ie
e
sin 9 = (D.l)
2

We may use Euler's formula also to clarify the polar representation

A = r[9 (D.8)
of the complex number
A=a+jb (D.9)

This was considered in Appendix C, where by (C.5) we had

a = /• cos 9
(D.IO)
b = r sin 9

By Euler's formula we may write

re'
6
= r(cos 9 + j sin 9)
= r cos 9 + j'f sin 9
which, by (D.IO), is

re'
s
= a+jb (D.I l)

Therefore, comparing (D.l l) with (D.8) and (D.9), we have

A = r[l=re ie
(D.l 2)

This result enables us to easily obtain the multiplication and division rules, given

in Appendix C in (C.13) and (C.16). Clearly, if

A = rJ9, = r e ie
x
'
App. D Euler's Formula 515

and

then
AB -(rJOMrz/Oz)

r l r z e ne,+6 ' )

-- r r 1 /B
l l
\9 1

Similarly, we may obtain

^/g.-fla
8

Euler's formula is illustrated graphically in Fig. D.I. A unit vector is rotating

around a shown, with an angular velocity of wrad/s. Therefore
circle in the direction
in t seconds it has moved through an angle cot as shown, and thus the vector may be

specified by /cot or e iu". Its real part is the projection on the horizontal axis, given by
l

cos cot, and its imaginary part is the projection on the vertical axis, given by sin cot.
That is.
e'°" = cos cot + j sin cot
which is Euler's formula. The projections trace out the cosine and sine waves, as
shown, as the vector rotates.

FIGURE D .1 Graphical illustration of Euler's formula

INDEX
 INDEX

Bandwidth, 393
Battery, 1

Abe sequence, 339 Bell, A. G., 408

Ac (alternating current), 6 Bilateral Laplace transform, 474

Acb sequence, 339 Branch, 118

Ac steady-state analysis, 280 tree, 120

Active element, 10 Bridge circuit, 299

Admittance, 264, 266, 366 British System of Units, 2
short-circuit, 382 Buffer amplifier, 54

Alternating current (ac), 6 Butterworth filter, 413

Ammeter, 40
Ampere, A.M., 5, 142
Ampere, 3, 5
Ampere turn, 436
Amplifier circuit, 54 Candela, 2

Amplitude of sinusoid, 244 Capacitance, 132, 133

equivalent, 138-140
Amplitude response, 389, 390
Angular frequency, 244 Capacitance reactance, 266

Apparent power, 319 Capacitor, 132

characteristics of, 133
meter, 328
Attenuation, 409 energy storage in, 136

Average power, 306, 307 impedance of, 265

Average value of periodic function, 307 linear, 133

practical, 150
Capacitors:
in parallel, 138, 140
in series, 138, 139
Cavendish, H., 15, 142

Balanced bridge, 299 Center frequency, 392

Balanced three-phase system, 334 Characteristic equation, 206
Bandpass filter, 392 Charge, 4
Band-reject filter, 394 conservation of, 135

519
 11

520 Index

Circuit, 1 Coupling, coefficient of, 427

analysis of, 12, 13 Cramer's rule, 66, 499, 501
dual, 79 Critically damped case, 226
linear, 89, 90 Current, 2, 4, 5,

nonlinear, 92 coil, 327

nonplanar, 73, 117 conventional, 5
planar, 73 division, 30, 31
simple, 25 electron, 5
singular, 138, 146, 153 leakage, 150
synthesis of, 13 link, 124
Circuit element, 1, 2 mesh, 73
active, 10 ohmic, 150
linear, 90 sensitivity, 41

passive, 10 Current source, 1

Circuit synthesis, 13 dependent, 49

Coefficient of coupling, 427 ideal, 12, 108

Cofactor of determinant, 500 independent, 1

Coil, 142 practical, 108

current, 327 Cutoff point, 392

potential or voltage, 327 Cycle, 244
Color code for resistor, 43 Cycle per second, 245
Complete response, 218, 229
Complex excitation, 253
Complex frequency, 360, 363, 365
plane, 374
Complex number, 250, 507, 509
amplitude of, 509 Damped frequency, 225
angle of, 509 Damped sinusoid, 360
argument of, 509 Damping coefficient, 224
conjugate, 509 Damping factor, 360
magnitude of, 509 D'Arsonval meter, 40
polar form, 509 Datum node, 62
rectangular form, 508 dc (direct current), 6
Complex plane, 374, 507 dc steady state, 166
Complex power, 306, 323 Decibel, 408
conservation of, 325 Degree Kelvin, 2
Conductance, 17, 266 Delta-connected load, 345
Conductor, 7, 16 Delta connection, 345
Conjugate of complex number, 509 Dependent source, 12, 49
Conservation of charge, 135 Describing equation, 201
Conservation of complex power, 325 Determinant, 66, 499
Conservation of flux linkage, 144 Dielectric, 132
Conservation of power, 27 Differential equation, 160, 161
Continuous frequency spectrum (see Amplitude re- solution by Laplace transform, 487
sponse or Phase response) Direct current (dc), 6
Controlled source, 49 Discrete amplitude spectrum, 461
current, 49 Discrete phase spectrum, 461
voltage, 49 Discretepower spectrum, 471
Convolution theorem, 479 Dot convention, 421
Core loss, 150 Dual, 79
Coulomb, 2-4 Duality, 78, 79, 151, 152
Coulomb, C. A. de, 3 Dynamic element, 132
 Index 521

inverse, 464
pairs, 464
Four-wire system, 340
Effective value, 306, 3 17 Frequency, 244
Electric field, 135 angular, 244
Electric filter (see Filter) complex, 360, 363, 365
Electromotive force, 7 cutoff, 392
Electron, 4 domain, 258
Element (see Circuit element) fundamental, 448
Energy, 7, 8 generalized, 365
in capacitor, 136 natural, 378
density, 472 neper, 361
in inductor, 145 radian, 244
in mutual inductors, 425 resonant, 244
storage element, 132 response, 389, 390
Envelope of spectrum, 462 scaling, 404
Equivalent circuit of transformer, 440 spectrum, 461
Euler, L.,210 continuous (see Amplitude response or Phase
Euler's formula, 210, 513, 514 response)
Even function, 452 discrete, 461
Excitation, 12 line, 461
Exponential current, 6 Full-wave rectified sinusoid, 456
Exponential Fourier series, 456, 457 Fundamental harmonic, 448
Fundamental period, 307

G
Farad, 133
Gain of a filter, 412, 414, 415
Faraday, M., 133, 142
Gaussian elimination, 66, 504
Feedback mode, 54
Gauss-Jordan method, 506
Fibonacci number, 131
Generalized node, 68
Filter, 284, 389, 392
Giga-, 3
bandpass, 392
Graph, 117, 118
band-reject, 394
connected, 118
gain of, 412,414, 415
Ground, 62
high-pass, 394
low-pass, 393
Flux linkage, 142
Force, 3 H
Forced response, 174, 212
Forcing function, 172 Half-range series, 455
complex, 251, 253 Half-wave rectified sinusoid, 455
damped sinusoidal, 360, 364 Harmonic, 448
sinusoidal, 243 Harmonic oscillator, 240
Fourier, J.-B. J., 446 Heaviside, O., 182
Fourier coefficients, 447 Helmholtz, H.S. F., 101

Fourier integral, 464 Henry, J., 142

Fourier series, 447 Henry, the, 142
exponential, 456, 457 Hertz, 245
trigonometric, 446, 447 Hertz, H.R., 245
Fourier transform, 463, 464 High-pass filter, 394
 1

522 Index

Homogeneous equation, 204 Kilogram, 3

solution of, 207 Kirchhoff, G., 19
Hybrid parameter, 382, 384 Kirchhoff's laws, 15, 19
current law (KCL), 20, 21, 23
for phasors, 268
voltage law (KVL), 21,22

Ideal source, 12, 108

Ideal transformer, 434
Impedance, 264, 366
equivalent, 269 Ladder network, 91 92, 293 ,

generalized, 366 Lagging, 245

reflected, 432, 433 Laplace, P. S., Marquis de, 474

scaling, 404, 405 Laplace transform, 469, 473, 474
Thevenin, 290 application to differential equations, 487

Impulse function, 473, 484, 485 bilateral, 474

Impulse response, 490 inverse, 475, 480
Independent source, 1
one-sided, 474

current, 11 of periodic function, 496

voltage, 11 table of, 480
Inductance, 142 two-sided, 474

equivalent, 147, 148 unilateral, 474

mutual, 416, 417, 419 Laplace-transformed circuit, 492
self-, 419 Leading, 245
Inductor, 142 Leakage flux, 417
characteristics of, 142 Lenz, H.F. E.,422

energy storage in, 145 Lenz's law, 422

impedance of, 265 Linear circuit, 89, 90

linear, 142
element, 90
Linearity, 151, 153
practical, 150

Initial condition, 161 Linear resistor, 16

Line current, 340
Input, 12
Line spectrum, 461
Integrating factor, 176
Line terminal, 338
Integrator, 189
International System of Units (SI), 2
Line voltage, 340
Link, 119, 120
Inverse Fourier transform, 464
Inverse Laplace transform, 475, 480
Load, 30
Inverter, 55 Locus of phasor, 296
Inverting input terminal, 52
Loop, 73
analysis, 73, 280
Loss, 409
Low-pass filter, 393

Joule, J. P., 3
M
Joule, the, 3

Magnetic field, 142

Magnetic flux, 142
Magnitude scaling (see Impedance scaling)
Matrix, 505
Kelvin, degree, 2 Maximum power transfer theorem, 1 12
Kilo-, 3 for ac circuits, 330
 Index 523

Mega-, 3 One-port network, 380

Mesh, 61, 73 One-sided Laplace transform, 474
analysis. 73, 280, 285 Open circuit impedance parameter, 381
current, 73 Op amp (see Operational amplifier)
Meter, 2 Operational amplifier, 52
Mho, 17 Output, 12
Micro-, 3 Output terminal, 52
Milli-,3 Overdamped case, 223
Minor of determinant, 500
Mutual flux, 417
Mutual inductance, 416, 417, 419

Parallel connection, 30
N
Parallel resonance, 396
Parameters:
Nano-, 3
admittance, 382
Napier, J., 361
hybrid, 382, 384
Natural frequency, 207, 378
impedance, 381
Natural response, 160, 174, 205
transmission, 382, 388
from network function, 375
Parseval's theorem 47 1 , , 472
Negative phase sequence, 339
Partial fraction expansion, 481
Neper, 361
Passband, 392
frequency, 361
Passive element, 10
Network, 1
Period, 244
function, 360, 370, 371
Periodic extension, 455
one-port, 380
Periodic function, 244
theorems, 89, 288
effective value of, 317, 318
topology, 117
Laplace transform of, 496
two-port, 380
Per phase analysis, 343
Neutral terminal, 335, 338
Phase, 245
Neutral wire, 336
angle, 245
Newton, Sir I., 3
current, 341
Newton, the, 3
impedance, 340
Nodal analysis, 61 , 62, 280
response, 389, 390
Node, 19, 118
voltage, 338
generalized, 68
Phasor, 243,256
reference, 62
circuit, 269-71
voltage, 62
diagram, 294
Noninverting input terminal, 52
generalized, 363, 365
Nonlinear circuit, 92
reference, 295
Nonlinear resistor, 16
Physical resistor, 43
Norton, E. L., 101
Pico-, 3
Norton's theorem, 99, 102, 288, 290
Planar circuit, 73
Pliers entry, 378
Polar form of complex number, 250
O Pole, 373, 374
Pole-zero plot, 374, 401
Odd function, 452 Polyphase source, 333, 339
Ohm.G.S., 15 Port, 380
Ohm, the, 15 Positive-phase sequence, 339
Ohmmeter, 40 Potential coil, 327
Ohm's law, 15, 16 Potential difference (see Voltage)
 1

524 Index

Power, 7, 8 Response, 12
apparent, 319 complete, 204, 218, 229
average, 306, 307 forced, 160, 174,204,212
complex, 306, 323 impulse, 490
factor, 306, 319 natural, 160, 174,204,205
of three-phase load, 342 Rms value, 306, 317
instantaneous, 8 Root-mean-square value, 317
measurement, 327, 353, 355 steady-state, 174
peak, 306 transient, 174
quadrature, 324
reactive, 324

Practical source, 89, 106

current, 108
voltage, 107
Primary winding, 417
Proportionality property, 90 s plane, 374
Pulse, rectangular, 465, 466 Sampling function, 458
Sampling property, 486
Sawtooth wave, 6
Scaling, 404
frequency, 405, 406
impedance, 404, 405
Q (see Quality factor)
Second, 2, 3
Quadrature component, 324
Second-order circuit, 201 , 202
Quadrature power, 324
Secondary winding, 417
Quality factor, 389, 397, 399
Selectivity, 399
Self-inductance, 419
Series connection, 25
K resonance, 396
Short circuit admittance parameter, 382
Radian frequency, 244 SI (International System of Units), 2
Ramp function, 189 Single-phase, three-wire system, 334
Reactance, 265 Singular circuit, 153
Reactive power, 324 Singularity function, 182
Reciprocal network, 383 Sinusoid, 243,244-247
Rectangular form of complex number, 250 Sinusoidal excitation, 243, 244
Rectangular pulse, 465, 466 Soldering entry, 378
Reference node, 62 Source:
Reflected impedance, 432, 433 controlled, 49
Relaxed circuit, 469 current, 1 1, 49
Resistance, 15, 16 dependent, 12, 49
equivalent, 26 ideal, 12, 108
Thevenin, 101 independent, 1

Resistive circuit, 15 practical, 108

Resistor, 15 transformations, 108
linear, 16 voltage, 1 1 , 49
nonlinear, 16 Source-free circuit, 160
physical, 43 Spectrum, frequency, 461
Resonance, 389, 395 Square wave, 185
parallel, 396 State variable, 234
series, 396 Steady-state response, 174
Resonant frequency, 224, 396 Step response, 187
 Index 525

Superposition, 93-95, 193. 288, 313 Unit step function, 182

and power. 313 Unity-coupled transformer, 427
Susceptance, 266
Switch. 137
Symmetrical components, 344
Symmetry property. 452
Var, 324
Varmeter, 328
Volt, 7
Volta, A.G. A. A.,7
Voltage, 2, 7
Tellegen's theorem, 27
Voltage division, 25, 27
Thevenin, C. L., 101
Voltage source, 11, 106, 107
The\ enin impedance, 290
Thevenin resistance, 101 dependent, 49
ideal, 12, 106, 107
Thevenin's theorem, 99, 101 , 288, 290
independent, 1
Three-phase circuit, 333
practical, 106, 107
balanced, 340
Voltmeter, 40
A-connected load, 345
unbalanced, 344
VOM,42
y-connected source, 358
y-y connection, 338
\Y
Time constant, 164, 165, 169
Time-domain function, 258
Topology, 117 Watt, 3

Transformed circuit, 492 Watt, J., 3

Transformer, 333, 416-41 Wattmeter, 327

ideal, 434 Weber, 142
unity-coupled, 427 Weber, W., 142
Transient response, 174 Work, 3

Transmission parameter, 382, 388

Tree, 119
branch, 120
Turns ratio, 434
Two-port network, 380 Y-connection, 338
Two-sided Laplace transform, 474 y parameter, 382
Two-terminal element, 1 Y- A transformation, 348
Y-Y system, 340

Unbalanced three-phase circuit, 344

Underdamped case, 224 z parameter, 381
Unilateral Laplace transform, 474 zero, 373, 374