Professional Documents
Culture Documents
'
#
LINEAR
CIRCUITS
TIME DOMAIN, PHASOR, AND LAPLACE
TRANSrORM APPROACHES
THIRD EDITION
Raymond A. DeCarlo
Purdue University
Pen-Min Lin
Purdue University
Kendall Hunt
p u b l i s h i n g c o m p a n y
o
n
o
Kendall Hunft
p u b l i s h i n g c o m p a n y
www.kendallhunt.cpm
Send all inquiries to:
4050 Westmark Drive
Dubuque, lA 52004-1840
ISBN 978-0-7575-6499-4
O
TABLE OF CO N TEN TS
Preface......................................................................................................................................................................vii
Chapter 2 • Kirchhoff’s Current & Voltage Laws and Series-Parallel Resistive C ircu its..............51
Chapter 15 * Time Domain Circuit Response Computations: The Convolution M ethod...... 763
Chapter 18 • Tw o-Ports...................................................................................................................................959
In d ex................................................................................................................................................................... 1119
o
n p ,
0
o
O
o
0
n -
o
o
0
0
0
■ 0
0
n
n ^
o
PREFACE
For the last several decades, EE/ECE departments o f US universities have typically required two
semesters o f linear circuits during the sophomore year for EE majors and one semester for other
engineering majors. Over the same time period discrete time system concepts and computer engi
neering principles have become required fare for EE undergraduates. Thus we continue to use
Laplace transforms as a vehicle for understanding basic concepts such as impedance, admittance,
fdtering, and magnetic circuits. Further, software programs such as PSpice, MATLAB and its tool
boxes, Mathematica, Maple, and a host o f other tools have streamlined the computational drudg
ery o f engineering analysis and design. MATLAB remains a working tool in this 3'''^ edition o f
Linear Circuits.
In addition to a continuing extensive use o f MATLAB, we have removed much o f the more com
plex material from the book and rewritten much o f the remaining book in an attempt to make the
text and the examples more illustrative and accessible. More importantly, many o f the more diffi
cult homework exercises have been replaced with more routine problems often with numerical
answers or checks.
Our hope is that we have made the text more readable and understandable by today’s engineering
undergraduates.
C H A P T E R
CHAPTER O U TLIN E
1. Introduce and investigate three basic electrical quantities: charge, current, and voltage,
and the conventions for their reference directions.
2. Define a two-terminal circuit element.
3. Define and investigate power and energy conversion in electric circuits, and demonstrate
that these quantities are conserved.
4. Define independent and dependent voltage and current sources that act as energy or sig
nal generators in a circuit.
5. Define Ohm’s law, v{t) = R i{t), for a resistor with resistance R.
6. Investigate power dissipation in a resistor.
7. Classify memoryless circuit elements by dieir terminal voltage-current relationships.
8. Explain the difference between a device and its circuit model.
ch ap ter 1 • Charge, Current, Voltage and O hm ’s Law
Are you curious about how fuses blow? About the meaning o f different wattages on Hght bulbs?
About the heating elements in an oven? And how is the presence o f your car sensed at a stoplight?
Circuit theory, the focus o f this text, provides answers to all these questions.
W hen you learn basic circuit theory, you learn how to harness the power o f electricity, as is done,
for example, in
• an electric motor that runs the compressor in an air conditioner or the pump in a dish
washer;
• a microwave oven;
• a radio, TV, or stereo;
• an iPod;
• a car heater.
In this text, we define and analyze common circuit elements and describe their interaction. Our
aim is to create a modular framework for analyzing circuit behavior, while simultaneously devel
oping a set o f tools essential for circuit design. These skills are, o f course, crucial to every electri
cal engineer. But they also have broad applicability in other fields. For instance, disciplines such
as bioengineering and mechanical engineering have similar patterns o f analysis and often utilize
circuit analogies.
W H A T IS A C IR C U IT ?
CH A R G E
Charge is an electrical property o f matter. Matter consists o f atoms. Roughly speaking, an atom
contains a nucleus that is made up o f positively charged protons and neutrons (which have no
charge). T he nucleus is surrounded by a cloud o f negatively charged electrons. Th e accumulated
charge on 6.2415 x 10’^ electrons equals -1 coulomb (C). Thus, the charge on an electron is
-1 .6 0 2 1 7 6 X 10-19 C.
Particles with opposite charges attract each other, whereas those with similar charges repel. The
force o f attraction or repulsion between two charged bodies is inversely proportional to the square
o f the distance between them, assuming the dimensions o f the bodies are very small compared
with the distance o f separation. Two equally charged particles 1 meter (m) apart in free space have
charges o f 1 C each if they repel each other with a force o f 10“^ c^ Newtons (N), where c = 3 x
10^ m/s is the speed o f light, by definition. The force is attractive if the particles have opposite
charges. Notationally, Q will denote a fixed charge, and q or q{t), a time-varying charge.
Exercise. Sketch the time-dependent charge profile q{t) = 3 (l-^ ^ 0 C, ? > 0, present on a metal
plate. M ATLAB is a good tool for such sketches.
A conductor refers to a material in which electrons can move to neighboring atoms with relative
ease. Metals, carbon, and acids are common conductors. Copper wire is probably the most com
mon conductor. An ideal conductor offers zero resistance to electron movement. Wires are
assumed to be ideal conductors, unless otherwise indicated.
Insulators oppose electron movement. Common insulators include dry air, dry wood, ceramic,
glass, and plastic. An ideal insulator offers infinite opposition to electron movement.
C U R R EN T
Current refers to the net flow o f charge across any cross section o f a conductor. T he net move
ment o f 1 coulomb (1 C) o f charge through a cross section o f a conductor in 1 second (1 sec)
produces an electric current o f 1 ampere (1 A). The ampere is the basic unit o f electric current
and equals 1 C/s.
The direction o f current flow is taken by convention as opposite to the direction o f electron flow,
as illustrated in Figure 1.2. This is because early in the history o f electricity, scientists erroneously
believed that current was the movement o f only positive charges, as illustrated in Figure 1.3. In
metallic conductors, current consists solely o f the movement o f electrons. However, as our under
standing o f device physics advanced, scientists learned that in ionized gases, in electrolytic solu
c h ap ter 1 • Charge, Current, Voltage and O hm ’s Law
tions, and in some semiconductor materials, movement o f positive charges constitutes part or all
o f the total current flow.
One Ampere
of Current "
One ; ; Cloud o f \
se co n d ^ ....... |---- 6.24x10’® 1
later i ; k electrons J
Boundary
FIG U RE 1.2 A cloud o f negative charge moves past a cross section of an ideal conductor from right
to left. By convention, the positive current direction is taken as left to right.
One Ampere
of Current
One
Coulom b One
of positive 'second
charge later
Boundary
FIGURE 1.3 In the late nineteenth cenmry, current was thought to be the movement of a positive charge
past a cross section of a conduaor, giving rise to the conventional reference “direction of positive current flow.”
Both Figures 1.2 and 1.3 depict a current o f 1 A flowing from left to right. In circuit analysis, we
do not distinguish between these two cases: each is represented symbolically, as in Figure 1.4(a).
The arrowhead serves as a reference for determining the true direction o f the current. A positive
value o f current means the current flows in the same direction as the arrow. A current o f negative
value implies flow is in the opposite direction o f the arrow. For example, in both Figures 1.4a and
b, a current o f 1 A flows from left to right.
1A -1A
> <
(a) (b)
FIG U RE 1.4 1 A of current flows from left to right through a general circuit element.
Chapter 1 • Charge, Current, Voltage and O hm ’s Law
In Figure 1.4, the current is constant. The wall socket in a typical home is a source o f alternating
current, which changes its sign periodically, as we will describe shortly. In addition, a current direc
tion may not be known a priori. These situations require the notion o f a negative current.
E X A M P L E 1.1.
Figure 1.5 shows a slab o f material in which the following is true:
1. Positive charge carriers move from left to right at the rate o f 0.2 C/s.
2. Negative charge carriers move from right to left at the rate o f 0.48 C/s.
A B
1 , © o
Connecting — 0 © 0 Connecting
wire wire
Sem iconductor iVlaterial
S o lu tio n
a) The current from left to right, due to the movement o f the positive charges, is 0.2 A. The
current from left to right, due to the movement o f the negative charges, is 0.48 A.
Therefore, /^, the total current from left to right, is 0.2 + 0.48 = 0.68 A. Since ly is the
current from right to left, its value is then -0 .6 8 A.
Exercise. In Example 1.1, suppose positive-charge carriers move from right to left at the rate o f 0.5
C/s, and negative carriers move from left to right at the rate o f 0.4 C/s. Find and
AN SW ER: /, = - 0 .9 A; ^ = 0.9 A
If a net charge crosses a boundary in a short time frame o f At (in seconds), then the approxi
mate current flow is
Aq
/=
At
( 1 . 1)
where I, in this case, is a constant. The instantaneous (time-dependent) current flow is the limit
ing case o f Equation 1.1, i.e.,
Chapter 1 • Charge, Current, Voltage and O hm ’s Law
dq{t)
dt
( 1. 2)
Here q{t) is the amount o f charge that has crossed the boundary in the time interval [tQ, t] . The
equivalent integral counterpart o f Equation 1.2 is
q{t) = J i{r)dr
(1.3)
E X A M P L E 1.2
The charge crossing a boundary in a wire is given in Figure 1.6(a) for ? > 0. Plot the current i{t)
through the wire.
(a)
(b)
FIG U RE 1.6 (a) Charge crossing a hypothetical boundary; (b) current flow
associated with the charge plot o f (a).
Chapter 1 • Charge, Current, Voltage and O hm ’s Law
S o lutio n
As per Equation 1.2, the current is the time derivative o f q{t). The slopes o f the straight-Une seg
ments o f q{f} in Figure 1.6(a) determine the piecewise constant current plotted in Figure 1.6(b).
■ ■ • • l-cos(co?)
Exercise. The charge crossing a boundary in a wire varies as q[t) = ---------------- C, for t >Q.
Compute the current flow.
A N SW ER: sin(cof) A, for f > 0
E X A M PLE 1.3
Find q{t), the charge transported through a cross section o f a conductor over [0, f], and also the total
charge Q transported, if the current dirough the conductor is given by die waveform o f Figure 1.7(a).
-l-*-t(se c)
FIG U RE 1.7 (a) Square-wave current signal; (b) q{t) equal to the integral of i{t) given in (a).
Chapter 1 • Charge, Current, Voltage and O hm ’s Law
S o lutio n
From Equation 1.3, for t>Q,
q{t)=p{T)clT
Thus, q{t) is the running area under the i{t) versus t curve. Since i{t) is piecevv'ise constant, the
integral is piecewise linear because the area either increases or decreases linearly with time, as
shown in Figure 1.7(b). Since q{t) is constant for ^ > 3, the total charge transported is Q = q{5) =
3 C.
Exercise. If the current flow through a cross section o f conductor is i{t) = cos(120jtf) A for ? > 0
and 0 otherwise, find q{t) for t>Qi.
Exercise. Suppose the current through a cross section o f conductor is given in Figure 1.8. Find
q{t) for t > 0 .
FIGURE 1.8
T Y P ES OF C U R R EN T
There are two very important current types: direct current (do) and alternating current (ac).
Constant current (i.e., dqldt = / is constant) is called direct current, which is illustrated graphi
cally m Figure 1.9(a). Figure 1.9(b) shows an alternating current, generally meaning a sinusoidal
waveform, i.e., current o f the form y4sin(w? + ()>), where A is the peak magnitude, co is the angu
lar frequency, and (|) is the phase angle o f the sine wave. W ith alternating current, the instanta
neous value o f the waveform changes periodically through negative and positive values, i.e., the
ch a p ter 1 • Charge, Current, Voltage and O h m s Law
direction o f the current flow changes regularly as indicated by the + and - values in Figure 1.9(b).
Household current is ac.
Lastly, Figure 1.9(c) shows a current that is neither dc nor ac, but that nevertheless will appear in
later circuit analyses. There are many other types o f waveforms. Interestingly, currents inside com
puters, C D players, TV s, and other entertainment devices are typically neither dc nor ac.
i(t) (A)
t(sec)
-H ----------------------1 - -I-----►
3
(a)
F IG U R E 1.9 (a) Direct current, or dc; i{t) = Iq\ (b) alternating current, or ac;
Because the value o f an ac waveform changes with time, ac is measured in different ways. Suppose
the instantaneous value o f the current at time t is A!sin(ci)i- + (j>). The term peak value refers to K
in K sin(co? + (j)). The peak-to-peak value is 2K. Another measure o f the alternating current,
indicative o f its heating effect, is the root mean square (rms), or effective value. The rms or effec
tive value is related to the peak value by the formula
A derivation o f Equation 1.4 with an explanation o f its meaning will be given in Chapter 11.
A special instrument called an ammeter measures current. Some ammeters read the peak value,
whereas some others read the rms value. One type o f ammeter, based on the interaction between
the current and a permanent magnet, reads the average value o f a current. From calculus, Fave!
the average value o f any function y(^), over the time interval [0, 7] is given by
(1.5)
For a general ac waveform, the average value is zero. However, ac signals are often rectified, i.e.,
converted to their absolute values, in power-supply circuits. For such circuits, the average value o f
the rectified signal is important. From Equation 1.5, the average value o f the absolute value o f an
ac waveform over one complete cycle with T = 2jt/co, is
K ^ 2.K
Average Value = —^\s,m{wt)\dt = ----- J sin(cot)clt
0 ^ 0
0.5T
2K -cos{(ot) 2,K
— = 0.636K
T (O jt ( 1.6)
Exercise. Suppose i{t) - 169.7 sin(50jtr) A. Find the peak value, the peak-to-peak value, the rms
value o f i{t), and the average value o f
AN SW ER: 169.7, 339.4, 120, and 107.93 A, respectively
3. VO LTAG E
W hat causes current to flow? An analogous question might be. W hat causes water to flow in a pipe
or a hose? W ithout pressure from either a pump or gravity, water in a pipe is still. Pressure from
a water tower, a pressured bug sprayer tank, or a pump on a fire truck will force water flow In
electrical circuits, the “pressure” that forces electrons to flow, i.e., produces a current in a wire or
a device, is called voltage. Strictly speaking, water flows from a point o f higher pressure— say,
p o in ts — to a point o f lower pressure— say, point 5 — along a pipe. Between the two points and
B, there is said to be a pressure drop. In electrical circuits, a voltage drop from point A to point B
Chapter 1 • Charge, Current, Voltage and O hm ’s Law 11
along a conductor will force current to flow from point A to point B; there is said to be a voltage
drop from point A to point B in such cases.
Gravity forces the water to flow from a higher elevation to a lower elevation. An analogous phe
nomenon occurs in an electric field, as illustrated in Figure 1.10(a). Figure 1.10(a) shows two con
ducting plates separated by a vacuum. O n the top plate is a fixed amount o f positive static charge.
On the bottom plate is an equal amount o f negative static charge. Suppose a small positive charge
were placed between the plates. This small charge would experience a force directed toward the
negatively charged bottom plate. Part o f the force is due to repulsion by the positive charges on
the top plate, and part is due to the attraction by the negative charges on the bottom plate. This
repulsion and attraction marks the presence o f an electric field produced by the opposite sets o f
static charges on the plates.
The electric field indicated in Figure 1.10 sets up an “electric pressure” or voltage drop from the
top plate to the bottom plate, which forces positive charges to flow “downhill” in the way that
water flows from a water tower to your faucet. Unlike water flow, negative charges are forced
“uphill” from the negatively charged bottom plate to the positively charged top plate. As men
tioned in the previous section, this constitutes a net current flow caused by the bilateral flow o f
positive and negative charges. The point is that current flow is induced by an electric pressure
called a voltage drop.
© © © © © © © © © © © © © © © ©
A 0 Positive A Force on
charge, q negative
charge
Electric Field Force on Electric Field
charge q Negative
B B © charge,-q
© © © 0 © © © © © © © © © © © ©
(a) (b)
FIGURE 1.10 (a) Positive charge in a (uniform) electric field; (b) negative charge in a uniform elearic field.
As mentioned, in Figure 1.10, the positive charge ^ at ^ tends to move toward B. We say, quali
tatively, that point A in the electric field is at a higher potential than point B. Equivalently, point
5 is at a lower potential than point A. An analogy is now evident: a positive charge in an electric
field “falls” from a higher potential point to a lower potential point, just as a ball falls from a high
er elevation to a lower elevation in a gravitational field.
Note, however, that if we turn the whole setup o f Figure 1.10(a) upside down, the positive charge
q still moves from point A to point B, an upward spatial movement. Similarly, if a negative charge
- q is placed at B, as in Figure 1.10(b), then the negative charge experiences an upward-pulling
force, moving from the lower potential, point B, to the higher potential, point A.
n
12 Chapter 1 ® Charge, Current, Voltage and Ohms Law —
----------------------------------------------------------------------------------------------------------------------- ^ n
Again, consider Figure 1. 10(a). As the charge q moves from point ^ toward B, it picks up veloci
ty and gains kinetic energy. Just before q hits the bottom plate, the kinetic energy gained equals
the (constant) force acting on q multiplied by the distance traveled in the direction o f the force. The
kinetic energy is proportional to q and to the “distance traveled.” Therefore,
The missing proportionality constant in this relationship is defined as the potential difference or
voltage between A and B, The term “voltage” is synonymous with “potential difference.”
Mathematically,
, . , energy converted
voltage = potential difference = ( 1.8)
magnitude of charge
The standard unit for measuring potential difference or voltage is the volt (V). According to
Equation 1.8, i f 1 joule {]) o f energy is convertedfrom one form to another when moving 1 C o f charge
from point K to point B, then the potential difference, or voltage, between A and B w i VTIn equation
form, with standard units of V, J, and C, we have
O
1V = 1 ^ (1.9)
The use of terms such as “elevation diflFerence,” “energy converted,” “potential difference,” or
“voltage” implies that they all have positive values. If the word “difference” is changed to “drop”
(or to “rise”), then potential drop and elevation drop have either positive or negative values, as the
case may be. The following four statements illustrate this point in the context of Figure 1.10:
. . . . ’ ^
This discussion describes the phenomena of “voltage.” Voltage causes current flow. But what pro
duces voltage or electric pressure? Voltage can be generated by chemical action, as in batteries. In
a battery, chemical action causes an excess of positive charge to reside at a terminal marked with
a plus sign and an equal amount of negative charge to reside at a terminal marked with a negative
sign. When a device such as a headlight is connected between the terminals, the voltage causes a
current to flow through the headlight, heating up the tiny wire and making it “Ught up.” Another
source of voltage/current is an electric generator in which mechanical energy used to rotate the ^
shaft of the generator is converted to electrical energy using properties of electro-magnetic fields.
All types of circuit analysis require knowledge of the potential difference between two points, say ^
A and B, and specifically whether point A or point 5 is at a higher potential. To this end, we speak ^
of the voltage drop from point A to point B, conveniently denoted by a double-subscript, as Vj^.
If the value of is positive, then point ^ is at a higher potential than point B. On the other
hand, if is negative, then point 5 is at a higher potential than point A. Since stands for
the voltage drop from point B to point A,
o
n
Chapter 1 • Charge, Current, Voltage and O hm ’s Law
The double-subscript convention is one o f three methods commonly used to unambiguously specify
a voltage drop. Using this convention requires labeling all points o f interest with letters or integers so
that ’ KiO ^12’ ^13 sense. A second, more-common convention uses + and - markings
on two points, together with a variable or numerical labeling o f the voltage drop from the point
marked + to the point marked - . Figure 1.11 illustrates this second convention, where Vq denotes the
voltage drop from A (marked +) to B (marked - ) . If Vq is positive, then ^4 is at a higher potential than
B. O n the other hand, if Vq is negative, then 5 is at a higher potential than A. The value o f Vq, togeth
er with the markings + and stipulates which terminal is at a higher potential; neither alone can do
this. For a general circuit element, the (+, —) markings— that is, the reference directions— can be
assigned arbitrarily. A third method for specifying a voltage drop, using a single subscript, will be dis
cussed in Chapter 2.
-I-
V„
FIGURE 1.11 The + and - markings establish a reference direction for voltage drop. For accuracy,
always place the (+, - ) markings reasonably close to the circuit element to avoid uncertainty.
The following example illustrates the use o f the double subscript and the (+, - ) markings for des
ignating voltage drops.
E X A M P L E 1.4
Figure 1.12 shows a circuit consisting o f four general circuit elements, with voltage drops as indi
cated. Suppose we know that = 4 V, and = 9 V. Find the values o f V^q and
CD-
V -I-
3V
FIG U RE 1.12 Arbitrary circuit elements for exploring the use of (+, - ) for specifying a voltage drop.
14 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
S o lutio n
T he meaiiing o f the double subscript notation and the (+, - ) markings for a voltage imply that
'DA
^ 5 C = 3I V
^CZ> = - ^ Z )C = -(-2 ) = 2 V
Exercise. T he convention o f the (+, - ) markings is commonly used as described. Figure 1.13 shows
an old 12-V automobile battery whose (+, - ) markings cannot be seen because o f the corrosion o f
the terminals. A digital voltmeter (DVM ) is connected across the terminals, as shown. The display
reads -1 2 V. Figure out the (+, - ) marking o f the battery terminals.
A N SW ER: left terminal, right terminal, +
DVM
12V
battery
FIG U RE 1.13 Digital voltmeter connected to a 12-V (car) battery whose plus and
minus markings have corroded away.
One final note: As with current, there are different types o f voltages— dc voltage, ac voltage, and
general voltage waveforms. Figure 1.9, with the vertical axis relabeled as v{t), illustrates different
voltage types.
4. C IR CU IT ELEM EN TS
Circuits consist o f interconnections o f circuit elements. The most basic circuit element has two
terminals, and is called a two-terminal circuit element, as illustrated in Figure 1.14. A circuit eie-
Chapter 1 • Charge, Current, Voltage and O h m s Law 15
ment called a source provides either voltage, current, or both. The battery is a very common
source, providing nearly constant voltage and the usually small current needed to operate small
electronic devices. Car batteries, for example, are typically 12 volts and can produce large currents
during starting. The wall outlet in a home can be thought o f as a 110 -volt ac source. Figure 1.14(a)
shows a (battery) voltage across a general undefined circuit element. A current z(r) flows
through the element. Recall from our earlier intuitive discussion that voltage is analogous to water
pressure: pressure causes water to flow through pipes; voltage causes current to flow through cir
cuit elements. Total water into a pipe equals total water out o f the pipe. Analogously, the current
entering a two-terminal device must, by definition, equal the current leaving the two-terminal device.
Current
The circuit element o f Figure 1.14(a) has a specific labeling: the current i(f) flows from the plus
terminal to the minus terminal through the circuit element. Such a labeling o f the voltage-current
reference directions is called the passive sign convention. In contrast, the current iij) flows from
the minus terminal to the plus terminal through the battery; this labeling is conventional for
sources but not for non-source circuit elements.
For a resistor, the amount o f current flow depends on a property called resistance; the smaller the
resistance, the larger the current flow for a fixed voltage across the resistor. A small-diameter pipe
offers more resistance to water flow than a large-diameter pipe. Similarly, different types o f con
ductors offer different resistances to current flow. A conductor that is designed to have a specific
resistance is called a resistor. If the device is an ideal resistor, then v(f) = Ri{i), where i? is a con
stant o f resistance. More on this shortly.
The circuit elements called the capacitor and the inductor will be described later in the text. Also,
future chapters will describe the operational amplifier and the transformer that are circuit elements
having more than two terminals.
16 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
The relationship between voltage across and current through a two-terminal element determines
whether power (and, thus, energy) is delivered or absorbed. The heating element in an electric
oven can be thought o f as a resistor. The heating element absorbs electric energy and converts it
into heat energy that cooks, among other things, turkey dinners.
In Figure 1.14(a), a battery is connected to a circuit element. Figure 1.14(b) concretely illustrates
this with a 12-V car battery connected to a headlight. W ith reference to Figures 1.14(a) and
1.14(b), suppose v{t) = 12 V, and i{t) = 5 A: 5 A o f current flows through the headlight. The head
light converts electrical energy into heat and light. Power (in watts) is the rate at which the ener
gy is converted. At each instant o f time, the electrical power delivered to (absorbed by) the head
light is pit) = v[t)i{t) - 12 X 5 = 60 watts. Similarly, at each instant o f time, the battery can be
viewed as delivering 60 watts o f power to the headlight. Inside the battery, the stored potential
energy o f the chemicals and metals undergoes a chemical reaction that produces the electrical
potential difference and the current flow to the headlight: chemical energy is converted into elec
trical energy that is converted into light and heat.
Figure 1.15 depicts a more general scenario: a circuit element is connected to its surrounding cir
cuit at points A and B. (One, o f course, could imagine that the “remainder o f circuit” is a battery,
and circuit element 1 is a headlight.) Suppose there is a constant voltage drop from A to B, denot
ed by Also assume that a constant current flows from terminal A to terminal B through
circuit element 1, as shown.
For discussion purposes, assume > 0 and > 0. During a time interval o f T s, (V^g x T) C o f
charge moves through circuit element 1 from A to B. In “falling” from a higher potential, point A, to
a lower potential, point B, the charge loses electric potential energy. The lost potential energy is con
verted within element 1 into some other form o f energy— heat or light being two o f several possibil
ities. According to Equation 1.8, the amount o f energy converted {absorbed by the element) is
y. T) >Q. The power absorbedhj element 1 is, by definition, the rate at which it converts or absorbs
energy. This rate equals
^a b (^ab ^ T)
■^Vab I a b > 0 .
T
Chapter 1 • Charge, Current, Voltage and O h m s Law 17
Exercise. In Figure 1.15, the current ^AB - 5 niA, and = 400 V. W hat is the energy absorbed
by circuit element 1 in one minute? W hat is the power absorbed by circuit element 1?
AN SW ER: W = 120 J; P = 2 watts
W ith respect to Figure 1.15, for constant (direct) voltages and currents, we arrive at a very simple
relationship:
P\-V ab I ab 0 -1 0 )
where is the power (in W ) absorbed by the circuit element. Consequently, the energy, W , (in
J), absorbed during the time interval Tis
W^=P\xT (1-11)
Now, let us reconsider Figure 1.15. One can think o f-/ ^ g as flowing from A w B through the
remainder o f the circuit. In this case, -1 ^ ^ ^ ^ < 0 . This means that the remain
der o f the circuit absorbs negative power or equivalently delivers |^ 5 (— | = ^a ^AB circuit
element 1. As such, the remainder o f the circuit is said to generate electric energy. By definition,
the electric power generated by the remainder o f the circuit is the rate at which it generates elec
tric energy. From Equation 1.8, this rate equals
Observe that the rate at which the remainder o f the circuit generates power precisely equals the
rate at which circuit element 1 absorbs power. This equality is called the principle o f conserva
tion o f power: total power generated equals total power absorbed. Equivalently, the sum o f the
powers absorbed by all the circuit elements must add to zero, + Pq = y^gl^B ^AB^^^AB^ ~
Exercise. In Figure 1.15, -Pg = watts, i.e., the remainder o f the circuit absorbs - 1 0 watts o f
power. How much power does circuit element 1 absorb?
A N SW ER: 10 watts
In general, whenever a two-terminal general circuit element is labeled according to the passive sign
convention, as in Figure 1.15, then P = > 0 means the element absorbs (positive) power,
whereas P = V^b ^ab ^ absorbs negative power or delivers (positive) power to whatever it
is connected. As a general convention, non-source circuit elements are labeled according to the
passive sign convention. Usually, sources are labeled with the current leaving the terminal labeled
with “+”. For such labeling o f sources, if the product o f the source voltage and the current leaving
the “+” terminal is positive, then the source is delivering power to the network.
Chapter 1 • Charge, Current, Vohage and O hm ’s Law
+ V AB
FIGURE 1.16
Exercise. Compute the power absorbed by each o f the elements in Figure 1.17.
> < Z3
10V 10V 10V
(a) (b) (c)
FIGURE 1.17
AN SW ER: (a) 10 W; (b) - 2 0 W; (c) 20 W
As mentioned, power is the rate o f change o f work per unit o f time. T he ability to determine the
power absorbed by each circuit element is highly important because using a circuit element or
some device beyond its power-handling capability could damage the device, cause a fire, or result
in a serious disaster. This is why households use circuit breakers to make sure electrical wiring is
not overloaded.
Exercise. In Figure 1.18, a car heater is attached to a 12-volt D C voltage source. How much power
can the car heater absorb before the 20 -amp fuse blows.
20 Amp Fuse
FIGURE 1.18 Car heater connected to a 12-volt car battery through a 20-amp fuse.
As mentioned earlier, the calculated value o f absorbed power P may be negative. If the absorbed
power P is negative, then the circuit element actually generates power or, equivalently, delivers
power to the remainder o f the circuit. In any circuit, some elements will have positive absorbed
powers, whereas some others will have negative absorbed powers. If one adds up the absorbed
powers o f ALL elements, the sum is zero! This is a universal property called conservation o f
power.
The 2"*^ edition o f this text contains a rigorous proof o f this principle. For the present, we will
simply use it to solve various problems. The following example will help clarify the sign conven
tions and illustrate the principle o f conservation o f power.
E X A M P L E 1.5
Light bulbs come in all sorts o f shapes, sizes, and wattages. W a t t l e measures the power consumed
by a bulb. Typical wattages include 15, 25, 40, 60, 75, and 100 W. Power consumptions differ
because the current required to light a higher-wattage (and brighter) bulb is larger for a fixed out
let voltage: a higher-wattage bulb converts more electric energy into light energy. In Figure 1.19,
the source delivers 215 watts o f power. W hat is the wattage o f the unlabeled bulb?
7?
100V
’ watts
watts watts
S o lutio n
From conservation o f power, the total power delivered by the battery equals the total power
absorbed by all the bulbs. Therefore, the power absorbed by the unknown bulb is
10A
Electroplating
Apparatus
S o lutio n
Step 1. From Equation 1.10, the power consumed is
Step 2. According to Equation 1.11, the energy consumed per 12-h period is
Exercise. Suppose the electroplating apparatus o f Example 1.6 draws 12 A D C at the same volt
age. W hat is the cost o f operation for a single 12-h day? W hat is the cost o f operating for a 20
workday month?
AN SW ER: $3,168; $63.36
E X A M PLE 1.7
Each box in the circuit o f Figure 1.21 is a two-terminal element. Compute the power absorbed by
each circuit element. W hich elements are delivering power? Verify the conservation o f power prin
ciple for this circuit.
Chapter 1 • Charge, Current, Voltage and O hm ’s 21
S o lutio n
Step 1. Compute power absorbed by each element. Using either Equation 1.10 or the power con
sumption rule, the power absorbed by each element is
a) For element 1 P i = 4 X 1 = 4 W
b) For element 2 P l = 8 x 2 = 1 6 W
c) For element 3 ^ 3 = 10 X 1 = 10 W
d) For element 4 14 x (-1)=-14W
e) For element 5 P 5 = 2 x 2 = 4W
0 For element 6 Pe = 1 0 X ( - 2 ) = - 2 0 W
Step 2 . Verify conservation o f power. Since P 4 and Pg are negative, element 4 delivers 14 W, and
element 6 delivers 20 W o f power. T he remaining four elements absorb power. Observe that the
sum o f the six absorbed powers, 4 + 16 + 10 - 14 + 4 - 2 0 = 0, as expected from the principle o f
conservation o f power. Equivalently, the total positive generated power, (14 + 20) = 34 W, equals
the total positive absorbed power, (4 + 16 + 10 + 4) = 34 W.
FIG U RE 1.22
AN SW ER: 8 W, 20 W, - 2 8 W; element 3 equivalently delivers 28 W
22 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
Exercise. In Figure 1.22, suppose the current 2 A were changed to - 4 A. W hat is the new power
absorbed by element 3?
A N SW ER: 56 watts
If the power absorbed by a circuit element is positive, the exact nature o f the element determines the
type o f energy conversion that takes place. For example, a circuit element called a resistor (to be dis
cussed shortly) converts electric energy into heat. If the circuit element is a battery that is being
charged, then electric energy is converted into chemical energy within the battery. If the circuit ele
ment is a dc motor turning a fan, then electrical energy is converted into mechanical energy.
N O N -D C PO W ER A N D EN ER G Y C A LC U LA TIO N S
Consider Figure 1.23, where i{t) is an arbitrary time-varying current entering a general two-ter
minal circuit element, and v{t) is the time-varying voltage across the element. Because voltage and
current are functions o f time, the power p{t) = v{t)i{t) is also a function o f time. For any specific
value o f ^ = ?j, the value p{t^) indicates the power absorbed by the element at that particular
time— hence, the terminology instantaneous power for p{t).
i(t)
FIGURE 1.23 Calculation of absorbed power for time-varying voltages and currents for circuit ele
ments labeled with the passive sign convention; here, power is p{t) = v{t)i{t).
i.e., the instantaneous (absorbed) power p{t), in W, is the product o f the voltage v{t), in V, and
the current i{t), in A, with labeling according to the passive sign convention. This product also
makes sense from a dimensional point o f view;
Knowing the power p{t) absorbed by a circuit element as a function o f t allows one to compute
the energy W{tQ, t) absorbed by the element during the time interval [^q, t > Iq], W[tQ, t) (J) is the
integral o f p{t) (W) with respect to t over [?q, t], i.e..
Chapter 1 • Charge, Current, Voltage and O hm ’s Law 23
W(to,t)^ r p ir ) d r
where the lower limit o f the integral, could possibly be -oo. For the dc case, p{t) = P (a con
stant). From Equation 1.13,
t t
W(tQ,t) = f p ( r ) d T = P f d r = P(t-tQ) = P x T
where T = t - t^, as given in Equation 1.11. If, in Equation 1.13, tg = -oo, then W (-co, t) becomes
a function only o f t which, for convenience, is denoted by
t
W{t)= f p ( r ) d r
L (1.14)
W{t) = W{—00, t), in joules, represents the total energy absorbed by the circuit element from the
beginning o f time to the present time rwhen p{t) is in watts.
Exercise, a) Suppose the power absorbed by a circuit element over [0,oo) is p{i) = watts. Find
W (0, oo). b) Now suppose the absorbed power o f the circuit element is
p{t) =
j j >0 • for t > 0 .
A N SW ER: 4 J; (4+t) J
Since energy is the integral o f power, power is the rate o f change (derivative) o f energy.
Differentiating both sides o f Equation 1.14 yields the expected equation for instantaneous power.
dW(t)
v m o = P ( o = ^ (1.15a)
or, equivalently, for t > (q,
Exercise. Suppose that for t > 0 , the work done by an electronic device satisfies W{t) = 10(1 — J- If
the voltage supplied by the device is 10 V, then for t > 0, find the power and current supplied by the
device, assuming standard labeling, i.e., the passive sign convention.
FIGURE 1.24 (a) Current and (b) voltage profdes with respect to t for circuit o f Figure 1.23.
S o lutio n
A simple graphical multiplication o f Figures 1.24(a) and (b) yields the sketch o f the curves in
instantaneous power shown in Figure 1.25(a). From Equation 1.13 with = 0, we have, for 0 <
t< %
,2
p(T)dr - J — c/t = —
0
(b)
FIGURE 1.25 (a) Profile of the instantaneous power p{t) = v{t)i{t) for the current and voltage wave
forms of Figure 1.24; (b) associated profde of energy versus time.
Two-terminal circuit elements may be classified according to their terminal voltage-current rela
tionships. The goal o f this section is to define ideal voltage and current sources via their termi
nal voltage-current relationships.
The wall socket o f a typical home represents a practical voltage source. After flipping the switch
on an appliance plugged into a wall socket, a current flows through the internal circuitry o f the
appliance, which, for a vacuum cleaner or dishwasher, converts electrical energy into mechanical
energy. For modest amounts o f current draw (below the fuse setting), the voltage nearly maintains
its nominal pattern o f 120 / 2 sin(120 lit) = 169.7 sin(120 nt) V. This practical situation is ide
26 Chapter 1 • Charge, Current, Voltage and O h m s Law
alized in circuit analysis by the ideal voltage source symbol shown in Figure 1.26(a), a circle with
a ± reference inside. The symbol is more commonly referred to as independent voltage source.
FIGURE 1.26 Equivalent representations of ideal voltage source attached to a hypothetical circuit.
The waveform or signal v{t) in Figure 1.26 represents the voltage produced by the source at each
time t. The plus and minus (+, —), on the source define a reference polarity. T he reference polari
ty is a labeling or reference frame for standardized voltage measurement. T he reference polarity
does not mean that v(t) is positive. Rather, the reference polarity (+, - ) means that the voltage drop
from + to - is v{t), whatever its value/sign. Finally, the voltage source is ideal because it maintains
the given voltage v{t), regardless o f the current drawn from the source by the attached circuit.
voltage (V)
V,
1(A)
(b)
FIGURE 1.27 (a) Ideal battery representation of ideal voltage source; (b) v-i characteristic of ideal battery.
Figure 1.2 7 (a) shows a source symbol for an ideal battery. The voltage drop from the long-dash
side to the short-dash side is Vg, with Vjj > 0. In commercial products, the terminal marked with
a + sign corresponds to the long-dash side o f Figure 1.27(a). An ideal battery produces a constant
voltage under all operating conditions, i.e., regardless o f current drawn from an attached circuit
or circuit element, as indicated by the v-i characteristic o f Figure 1.27(b). Real batteries are not
ideal but approximate the ideal case over a manufacturer-specified range o f current requirements.
Practical sources (i.e., non-ideal); voltage sources, such as commercial dc and ac generators; and
real batteries deviate from the ideal in many respects. One important respect is that the terminal
voltage depends on the current delivered by the source. The most common generators convert
mechanical energy into electrical energy, while batteries convert chemical energy into electrical
Chapter 1 • Charge, Current, Voltage and O hm ’s Law 27
energy. There are two general battery categories: nonrechargeable and rechargeable. A discussion
o f the dramatically advancing battery technology is beyond the scope o f this text.
Besides batteries and ideal voltage sources, devices called ideal or independent current sources
maintain fixed current waveforms into a circuit, as illustrated in Figure 1.28. T he symbol o f an
ideal current source is a circle with an arrow inside, indicating a reference current direction. An
ideal current source produces and maintains the current i{t) under all operating conditions. O f
course, the current i{t) flowing from the source can be a constant (dc), sinusoidal (ac), or any other
time-varying function.
FIGURE 1.28 Equivalent ideal current sources whose current i{t) is maintained
under all operating conditions o f the circuit.
Independent sources have conventional labeling, as shown in Figure 1.29, which is different from
that o f the passive sign convention. Here the source delivers power if p{t) = v{t)i{t) > 0 and would
absorb power i f p{t) = < 0. A complicated circuit called a battery charger can deliver ener
gy to a drained car battery. T he car battery, although usually a source delivering power, exempli
fies a source absorbing power from the charger.
Another type o f ideal source is a dependent source. A dependent source or a controlled source
produces a current or voltage that depends on a current through or voltage across some other ele
ment in the circuit. Such sources model real-world devices that are used in real circuits. In the text,
the symbol for a dependent source is a diamond. If a ± appears inside the diamond, it is a depend
ent voltage source, as illustrated in Figure 1.30. If an arrow appears inside the diamond, it is a
dependent current source, as illustrated in Figure 1.31. In Figure 1.30, the voltage across the dia
mond-shaped source, v{t), depends either on a current, labeled through some other circuit
device, or on the voltage across it. If the voltage across the source depends on the voltage v^,
i.e., v{t) = p then the source is called a voltage-controlled voltage source (VCVS). If the volt
age across the source depends on the current z^, i.e., v{t) = then the source is called a cur
rent-controlled voltage source (CCVS).
FIGURE 1.30 The right element is a voltage-controlled voltage source (VCVS) if v{t) = (p is here
dimensionless), or a current-controlled voltage source (CCVS) if v(t) = r^i (r^ here has units of ohm).
Exercise. The voltage across a particular circuit element is = 5 V, and the current through the
element is 0.5 A, using the standard labeling.
a) If a V CV S (Figure 1.30) with p = 0.4 were associated with the controlled-source branch,
fmd vit).
b) If a CCV S (Figure 1.30) with = 3 £2 were associated with the controlled branch, fmd
v{t).
ANSW ER: a) 2 V; b) 1.5 V
There is dual terminology for dependent current sources. The configuration o f Figure 1.31 shows
a voltage-controlled current source (VCCS), i.e., i{t) = g^v^, or a current-controlled current
source (CCCS), for which i{t) =
r~\
Chapter 1 • Charge, Current, Voltage and O hm ’s Law 29
i(t) = or
Pi
Q V
X
or
Pi:
-o
FIG U RE 1.31 The right element is a voltage-controlled current source (VCCS) if i{t) = (g^ has
units o f siemens) or a current-controlled current source (CCCS) if i{t) = |3/^ ((3 is dimensionless).
Source voltages or currents are called excitations, inputs, or input signals. A constant voltage will nor
mally be denoted by an uppercase letter, such as V, Vq, V^, and so on. A constant current will typi
cally be denoted by /, /g, /p and so on. The units are volts, amperes, and so on. Smaller and larger
quantities are expressed by the use o f prefixes, as defined in Standard Engineering Notation Table 1.1.
Exercise. The voltage across a particular circuit element is = 5, and the current through the ele
ment is = 0.5 A using the standard labeling.
a) If a VCCS (Figure 1.31) with ^^ = 0.1 S were associated with the controlled-source branch, find
i{i).
b) If a CCCS (Figure 1.31) with P = 0.5 were associated with the controlled-source branch, find i{i).
AN SW ER; a) 0.5 A; b) 0.25 A
femto f 10-15
i
pico P 10-12
nano n 10-9
micro P 10-6
milli m 10-3
w kilo k 103
mega M lO^’
g‘ga G 109
tera T 1012
30 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
Different materials allow electrons to move from atom to atom with different levels o f ease.
Suppose the same dc voltage is applied to two conductors, one carbon and one copper, o f the same
size and shape. Two different currents will flow. T he current flow depends on a property o f the
conductor called resistance: the smaller the resistance, the larger the current flow for a fixed volt
age. The idea is similar to water flow through different-diameter pipes (analogous to electrical con
ductors): for a given pressure, a larger-diameter pipe allows a larger volume o f water to flow and,
therefore, has a smaller resistance than a pipe with, say, half the diameter.
A conductor designed to have a specific resistance is called a resistor. Hence, a resistor is a device
that impedes current flow. Just as dams impede water flow and provide flood control for rivers,
resistors provide a means to control current flow in a circuit. Further, resistors are a good approx
imate model to a wide assortment o f electric devices such as light bulbs and heating elements in
ovens. Figure 1.32(a) shows the standard symbol for a resistor, where the voltage and current ref
erence directions are marked in accordance with xhie.passive sign convention. Figure 1.32(b) pic
tures a resistor connected to an ideal battery.
I R
+ V -
(a)
FIG U RE 1.32 (a) Symbol for a resistor with reference voltage polarity and current direction
consistent with the passive sign convention; (b) resistor connected to an ideal battery.
In 1827, Ohm observed that for a connection like that o f Figure 1.32(b), the direct current
through the conductor/resistor is proportional to the voltage across the conductor/resistor, i.e., I
= V. Inserting a proportionality constant, one can write
1 = — V —GV (1 .1 6a)
or, equivalently, ^
V = R1
The proportionality constant R is the resistance o f the conductor in ohms. The resistance R meas
ures the degree to which the device impedes current flow. For conductors/resistors, the ohm (Q)
is the basic unit o f resistance. A two-terminal device has a 1-Q resistance i f a 1-V excitation causes
1-A o f current to flow. In Equation 1.16(a), the proportionality constant is the reciprocal o f R, i.e.,
G = HR, which is called the conductance o f the device. T he unit for conductance according to
the International System o f Units (SI) system is the siemen, S. In the United States, the older term
for the unit o f conductance is the mho ^5, that is, ohm spelled backward, which is still widely
used. In this text, we try to adhere to the SI system. If a device or wire has zero resistance {R = 0)
or infinite conductance {G = t»), it is termed a short circuit. On the other hand, if a device or
wire has infinite resistance (zero conductance), it is called an open circuit. Technically speaking,
Chapter 1 • Charge, Current, Voltage and O hm ’s Law 31
a resistor means a real physical device, with resistance being the essential property o f the device.
In most o f the literature on electronic circuits, resistor and resistance are used synonymously, and
we will continue this practice.
O H M 'S LAW
Ohm’s law, as observed for constant voltages and currents, is given by Equation 1.16(b), with
its equivalent form in Equation 1.16(a). However, it is true for all time-dependent waveforms
exciting a linear resistor. Thus, we can generalize Equation 1.16 as
v (0 = « W
or
i(t) ^
AO ----- OB
+ v(t) -
FIG U R E 1.33
If either the voltage or the current direction is reversed, but not both, then Ohm’s law becomes
v(t) = -Ri{t). As an aid in writing the correct v-i relationship for a resistor. Ohm’s law is stated
here in words:
For a resistor connected between terminals A and B, the voltage drop from A to B is equal to the
resistance multiplied by the current flowing from A to B through the resistor.
Exercise. Find the resistance R for each o f the resistor configurations in Figure 1.34.
AN SW ER: (a) 12 Q ; (b) 3 Q; (c) 6 Q
-1A R 4A R -2A R
FIG U RE 1.34
Once the voltage and the current associated with a resistor are known, the power absorbed by the
resistor is easily calculated. Assuming the passive sign convention, then combining Equation 1.12 for
32 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
power and Ohms law (Equation 1.17), the instantaneous absorbed power is
9
p{t) = v{t)i{t) = i' ^ { t ) R ^ ^ ^ (1.18a)
R
which for the dc case reduces to
Exercise. Find the power absorbed by each o f the resistors in Figure 1.35.
80 4A R 90
FIGURE 1.35
AN SW ER: (a) 18 W; (b) 160 W; (c) 16 W
Equations 1.18(a) and (b) bring out a very important property; a resistor always absorbs power,
dissipating it as heat. Intuitively speaking, electrons that flow through the resistor collide with
other particles along the way. The process resembles the action in a pinball game: the pinball sue-
cessively collides with various pegs as it rolls from a higher to a lower elevation. W ith each colli
sion, part o f the electron’s kinetic energy is converted into heat as the voltage pressure continues
to reaccelerate the electron.
Electrical energy that is converted to heat or used to overcome friction is usually called a loss. Such
losses are termed /-squared-i? {f-R) losses because o f the form o f Equation 1.18. On the other
hand, a stove’s heating element purposely converts to heat as much electric energy as possible, in ^
which case, the P-R loss is desirable. This heating effect also proves useful as the basis for the oper
ation o f fuses. A fuse is a short piece o f inexpensive conductor with a very low resistance and a
predetermined current-carrying capacity. When inserted in a circuit, it carries the current o f the
equipment or appliances it must protect. W hen the current rises above the fuse rating, the gener
ated heat melts the conducting metal inside the fuse, opening the circuit and preventing damage
to the more-expensive appliance. Oversized fuses or solid-wire jumpers circumvent safe fuse oper
ation by permitting unsafe operation at overload currents, with consequent electrical damage to
the appliance that may cause overheating and fire.
Resistance o f a conductor depends on the material and its geometrical structure. For a specific
temperature, R is proportional to the length I o f a conductor and inversely proportional to its
cross-sectional area A,
R= p^ (1.19)
Chapter 1 • Charge, Current, Voltage and O hm ’s Law 33
S o lutio n
The resistivities o f aluminum and nickel wire relative to copper are 1.6 and 5.1, respectively.
Hence, 100 feet o f aluminum/nickel wire has a resistance o f
(aluminum) 1.6 x 0.4094 = 0.655 Q
(nickel) 5.1 X 0.4094 = 2.088 Q
Given a 10-A current flowing through 100 feet o f copper, aluminum, and nickel wire, Ohm’s law
implies
(copper) V = /?/ = 0 .4 0 9 4 x 10 = 4 .0 9 4 V
Finally, from Equation 1.18(b), the absorbed power given off as heat is
(copper) P = V I ^ R I - = 0 .4 0 9 4 x 100 = 4 0 .9 4 W
Notice that every 100 feet o f 16 AWG aluminum wire would absorb 65.5 - 4 0.9 = 2 4 .6 W more
power than copper. And nickel wire absorbs even more power:
^ ^ 208.8
■ “ 4 0.94
times more power than copper per unit length. This absorbed power, given off as heat, is why
nickel wire is used for heating elements in toasters and ovens.
34 Chapter 1 • Charge, Current, Vokage and O hm ’s Law
Exercise, (a) If a constant current o f 10 A flows through 1,000 feet o f (16 AWG) copper wire, how
many watts o f heat are generated by the wire?
(b) If the wire o f part (a) were changed to (16 AWG) aluminum, how many watts o f heat would
be generated?
Temperature also affects resistance. For example, light bulbs have a “cold” resistance and a “hot”
resistance o f more importance during lighting. For most metallic conductors, resistance increases
with increasing temperature— except carbon, which has a decrease in resistance as temperature
rises. Since resistors absorb power dissipated as heat, they should have adequate physical dimen
sions to better radiate the heat or there must be some external cooling to prevent overheating.
-O -
-O -
S o lutio n
Step 1. From Ohm’s law. Equation 1.16(a),
V 90
/ = - = ----- = 0.75 A
R 120
Step 2 . By Equation 1.18(b), the power absorbed by the lamp is
Step 3 . C/?eck conservation o f power. T he power delivered by the source is 90 x 0.75 = 67.5 W.
Therefore, the power delivered by the source equals the power absorbed by the resistor. This ver
ifies conservation o f power for the circuit.
Exercise. In Example 1.10, suppose the battery voltage is cut in half to 60 V. W hat is the power
absorbed by the lamp? W hat is the power delivered by the battery? Repeat with the battery volt
age changed to 120 V.
The following example illustrates power consumption for a parallel connection o f light bulbs.
S o lutio n
(a) From Equation 1.18(b), P - V^IR,
12^ 144
^ 3 5 W = ^ - 4 .1 1 4 Q
Rxiw - = 5.33D
27
144 144
^50W = = 2 .88 Q - ' 60 = 2.4 Q
50
(b) The power delivered by the source equals the sum o f the powers consumed by each bulb,
which is 172 W.
(c) Since the current supplied by the source has dropped to 11.417 A, then the power delivered
by the source drops to P;„urcenew ~ ^ 11.417 = 137 watts, which is 35 watts less than the ear
lier-delivered power o f 172 watts. Hence, the 35-watt bulb has gone dark.
Exercise. Repeat Example 1.11 (a) with the battery voltage changed to 48 V and a new set o f light
bulbs whose operating voltage is 48 V.
AN SW ER; 85.333 Q; R^c,^= 65.83 Q; R^q^ = 4 6 .0 8 Q; R(^^^= 38.4 Q..
E X A M PLE 1.12
W hen connected to a 120-volt source, halogen light bulb number 1 uses 40 watts o f power. W hen
similarly connected, halogen light bulb 2 uses 60 watts o f power.
(a) Find the hot resistance o f each bulb.
(b) If the two bulbs are connected in a series, as in Figure 1.38 and placed across the 120-V
source, find the power absorbed by each bulb and the power delivered by the source,
assuming the hot resistances computed in part (a) do not change.
36 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
b u ib l
120 V ^ = i- 120V
bulb 2
(a) (b)
FIGURE 1.38 Series connection o f two light bulbs and equivalent resistive circuit model.
S o lu tio n
Step 1. Find the hot resistances. The hot resistances o f each bulb are given by
Step 2. Find the current through each bulb, the power absorbed by each bulb, and the power delivered
by the source. The circuit o f Figure 1.38(a) has the equivalent representation in terms o f resistanc
es in Figure 1.38(b). By definition, in a nvo-terminal circuit element, the current entering each
resistor equals the current leaving. Therefore, the current through each resistor in the series con
nection is the same, and is denoted /. So the new power dissipated by each bulb/resistor is
To calculate these values, we need to know I. By conservation o f power, the power delivered by
the source is the sum o f the absorbed powers, i.e.,
Therefore,
;=™=o.2A
600
Hence,
Vl =7?,7 = 72 V and V2 = ^ 2 ^ = 48
Although involved, the solution o f this problem uses the definition o f a two-terminal circuit ele
ment and conservation o f power to arrive at the result in a roundabout way. In Chapter 2, we can
more directly arrive at the answers by using Kirchhoff’s voltage and current laws.
A potential problem with series connections o f light bulbs is circuit failure. If one bulb burns out,
i.e., the filament in the bulb open-circuits, then all other lights are extinguished. Parallel circuits
continue to operate in the presence o f open-circuit failures and are easier to fix: only the unlit bulb
must be replaced.
V
+
V ( + ) V / Circuit •V-
-►I
(a) (b)
FIG U RE 1.40 (a) Constant source attached to circuit; (b) v-i characteristic
is a constant horizontal line in the v-i plane.
Analogously, an ideal current source maintains the given current, irrespective o f the voltage
requirements o f the attached circuit. For constant-current sources, as in Figure 1.41(a), this prop
erty is depicted by a constant vertical line (infinite slope) in the v-i plane (Figure 1.41(b)). This
means that an ideal current source has infinite “internal” resistance. Further, if = 0, the current
source looks like an open circuit because no current will flow, regardless o f any voltage generated
by the rest o f the circuit. Again, we must be content with this brief discussion until Chapter 2 reit
erates and expands on the ideas.
V
+
Y / Circuit /
(a) (b)
FIG U RE 1.41 (a) Constant source attached to circuit;
(b) v-i characteristic is a constant vertical line in the v-i plane.
9. SUM M ARY
Building on a simplified physics o f charge (coulombs), electric fields, and charge movement, this
chapter set forth the notions o f current, i{t) or / for dc, and voltage, v{t) or V for constant volt
ages. A rigorous treatment would require field theory and quantum electronics. More specifically,
the notions o f current, current direction, voltage, and voltage polarity, a two-terminal circuit ele
ment (the current entering equals the current leaving), the passive sign convention, power con
sumption [pit) = v{t)i{t) assuming the passive sign convention], and dissipated energy (the inte
gral o f power) were all defined. In general, we can say that every circuit element does one o f the
following:
• Absorbs energy
• Stores energy
• Delivers energy, or
• Converts energy from one form to another
Chapter 1 • Charge, Current, Voltage and O hm ’s Law 39
T he chapter subsequently introduced ideal independent and dependent voltage and current
sources: the voltage-controlled voltage source (VCVS), the current-controlled voltage source
(CC VS), the voltage-controlled current source (VCCS), and the current-controlled current source
(C C C S). A dependent source produces a voltage or current proportional to a voltage across or a
current through some other element o f the circuit. The various types o f dependent sources are
summarized in Table 1.4.
VCVS
-I-
(Voltage-Controlled
Voltage Source, p is dimen-
sionless)
ccvs
(Current-Controlled
Current Source, is in
ohms)
V CC S
(Voltage-Controlled
Voltage Source, is in S)
CCCS
(Current-Controlled
Current Source, P is
dimensionless)
40 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
The chapter keynoted a special two-terminal element, called a resistor, whose terminal voltage and
current satisfied Ohms law, v(t) = Ri{t), where v{t) is the voltage in volts, R is the resistance in
ohms, and i{t) is the current in amperes. The resistor, as defined in this chapter, is a passive ele
ment, meaning that it always absorbs power,/>{;■) = v{t)i{t) = i?-{t)IR = Rp-{t) > 0 since R>Q. This
absorbed power is dissipated as heat. Hence, the (passive) resistor models the heating elements in
a stove or toaster oven quite well. In addition, the resistor models the hot resistance o f a light bulb.
Throughout the text, the resistor will often represent a fixed electrical load. In a later chapter, we
will discover that it is possible to construct a device with a negative resistance, R<Q, which can
generate power. However, such a device is rather complex to build and requires such things as the
operational amplifier covered in Chapter 4.
The various quantities defined and used throughout the chapter have various units. The quanti
ties and their units are summarized as follows:
Throughout this chapter, a number o f examples illustrated the various concepts that were intro
duced. Some simple resisrive circuits were analyzed. To analyze more complex circuits, one needs
Kirchhoff’s voltage and current laws, which specify how circuit elements interact in a complex cir
cuit. These basic laws o f circuit theory are set forth in the next chapter.
Alternating current: a sinusoidally time-varying current signal having the form A'sin(co?+(j)).
Battery: a device that converts chemical energy into electrical energy, and maintains approxi
mately a constant voltage between its terminals.
Charge: an electric property o f matter, measured in Coulombs. Like charges repel, and unlike
charges attract each other. Each electron carries the smallest known indivisible amount o f
charge equal to - 1.6 x 10“ '^ Coulomb.
Conductance: reciprocal o f resistance, with siemens (S) (or formerly, mhos) as its unit.
Conductor: a material, usually a metal, in which electrons can move to neighboring atoms with
relative ease.
Conservation o f power (energy): the sum o f powers generated by a group o f circuit elements is
equal to the sum o f powers absorbed by the remaining circuit elements.
Current: the movement o f charges constitutes an electric current. Current is measured in
Amperes. One Ampere means movement o f charges through a surface at the rate o f 1
Coulomb per second.
Current source: a device that generates electrical current.
Dependent (controlled) current source: a current source whose output current depends on the
voltage or current o f some other element in the circuit.
Chapter 1 ®Charge, Current, Voltage and Ohms Law 41
'w '
Dependent (controlled) voltage source: a voltage source whose output voltage depends on the
voltage or current of some other element in the circuit.
Direct current: a current constant with time.
Ideal conductor: offers zero resistance to electron movement.
Ideal insulator: offers infinite resistance to electron movement.
Independent (ideal) current source: an ideal device that delivers current as a prescribed function
of time, e.g., {2 cos(/) + 12}A, no matter what circuit element is connected across its ter
minals.
Independent (ideal) voltage source: an ideal device whose terminal voltage is a prescribed func
tion of time, e.g., {2 cos{t) + 12}V, no matter what current goes through the device.
Instantaneous power: the value of p{t) = at a particular time instant.
Insulator: a material that opposes easy electron movement.
Mho: historical unit of conductance equal to the reciprocal of an ohm.
Ohm: unit of resistance. One ohm equals the ratio of IV to lA.
Ohm’s law: for a linear conductor, the current through the conductor at any time t is proportional
to the voltage across the conductor at the same time.
Open circuit: connection of infinite resistance or zero conductance.
Passive sign convention: voltage and current reference directions, indicated by +, - , and an arrow,
which conform to that shown in Figure 1.15.
Peak-to-peak value: equals 2 K 'm K sin(co^ + (()) of the ac waveform.
Peak value: refers to K m K sin(cor + (|)) of the ac waveform.
Power: rate of change of work per unit of time.
Resistance: for a resistor, v{t) a i{t). The proportionality constant R is called the resistance, i.e.,
v{i) = Ri{t). Resistance is measured in ohms: 1 ohm means the voltage is 1 V when the
current is 1 A.
Resistivity: the resistance of a conductor is proportional to its length and inversely proportional
to its cross-sectional area. The proportionality constant p is called the resistivity of the
material. The resistivity of copper at 2 0 ^C is 1.7 x 10~^ ohm-meters.
Resistor: physical device that obeys Ohms law. There are commercially available nonlinear resis
tors that do not obey Ohms law. Resistors convert electric energy into heat.
Root mean square (rms) or eflfective value: measure of ac current, which is related to the peak
value by the formula rms = 0.7071if, where K sin(o)^ + (|)) is the ac waveform.
Short circuit: connection of zero resistance or infinite conductance.
Siemens: unit of conductance (formerly, mho) or inverse ohms.
v-i characteristic: graphical or functional representation of a memoryless circuit element.
Voltage (potential difference): positive charge, without obstruction, will move from a higher
potential point to a lower potential point, accompanied by a conversion of energy.
Voltage is measured in volts; 1 volt between two points A and B means that the energy
converted when moving 1 Coulomb of charge between A and B is 1 joule.
V olt^e source: device that generates an electric voltage or potential difference.
Wattage: measure of power consumption.
42 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
i(t) (A)
2--
-t (secs)
- 1- -
Figure P I.7
- 2- -
CHECK: (c) For ,6 > ? > 3 , g (0 = j - 4 t + 12
Figure P I.5
6 . A plot o f the current flowing past point A is 8 . Find i(t) when the charge transported across
a surface cutting a conductor is shown in Figure
shown on the graph o f Figure P I . 6 . Find the
net positive charge transferred in the direction P I.8.
o f the current arrow during the interval 0 < ^<
6 sec, in Coulombs.
i (amps)
0.1
-t (sec)
10
Figure P I .8
Figure P I .6
44 Chapter 1 • Charge, Current, Voltage and O hm ’s Law
9. (a) Find the average value o f the voltage, 12.(a) W hich elements in Figure P I. 12 are
v{t) = K cos(cof) over one period, labeled with the passive sign conven
2k tion'
T = - (b) In the circuit o f Figure P I. 12, volt
(0
ages, currents, and powers o f some ele
Hint: See Equation 1.6.
ments have been measured and indi
(b) Find the average value o f the absolute
cated in the diagram.
value o f the voltage, v(f) = K cos(cot)
over one period, T = 2u/cl). 5A
+ 3V - 2V -
+ 5V -
C _ 3
6V
1A + circuit
circuit 6V element 2
20 V 3A lOV
(iii) element 1 16v Q ) 2 A
lOV
Figure PI. 11
9A
14A|
Figure P I. 15
circuit + + 10V-
15V 30
element 1 CHECK: Re 3.334, Rfj = 20 in ohms.
4A
circuit
5A ( ^ ^ 2 0 V
element 2 lOV 16. In the circuit shown in Figure P I. 16, i(t) =
lOA 6 5V - 5V +
5A
100(1 - e * ) mA for t > 0.
(a) How much energy does the element A
(a) absorb for the interval [0 , t] ?
(b) I f element B is a 5 Q resistor, deter
Figure P I. 14a mine the power absorbed at time t,
and the energy absorbed for the inter
circuit
val [0 , t],
element 2
I (c) W hat is the energy delivered by the
-v,(t) +
2A
source over the interval [0 , i\?
4V
(b) ti
25V
©
Figure P 1.14b
RESISTA N CE
18. Using Equation 1.19 and Table 1.2 , find
the resistance o f a nickel ribbon having these
dimensions:
length: 40 m
width: 1.5 cm
thickness: 0.1 cm
Figure P I.21
24. Repeat problem 23 when the switch is 28. In Figure P1.28, F q = 125 V.
repeatedly closed for 3 ms and opened for 1 ms. (a) Suppose bulb A and bulb B each use
100 watts o f power. Find /q and the
CH ECK: = 1.875 m W hot resistance o f each bulb.
(b) Suppose bulb A uses 40 watts o f power
25. In Figure P1.25, Vq = 10 V, and the switch S and bulb B 60 watts o f power. Find /q
alternately stays at position A for 4 ms and at and the hot resistance o f each bulb.
position B for 1 ms. Find the average value o f i(tj.
5kn
Figure P I.25
Figure PI.32b
C H EC K : 1 A, 20 Q or ?????
Figure P1.37
value of J?2 so that the voltage gain Gy 41. For the circuit o f Figure P 1.41, suppose
= 4. = 1 0 V.
(c) Given your answer to (b), determine (a) Find the output voltage and output
the power gain, which is the ratio of current.
the power delivered to divided by (b) Find the voltage gain
the power delivered by the source. ^out
Gv = , and the power gain
Gp = ^
Pin ■
(c) Find the power delivered by each
source.
Figure P I.38 (d) Suppose the power absorbed by the 2
k Q output resistor were 80 watts.
39. For the circuit of Figure P I .39, suppose Find the power delivered by the input
source, and the voltage
- 100 mA, 50 Q, 10 Q, and
100 Q.
(a) Find the output voltage and out 0.21 0.1V.
put current. v,^
Rb
(b) Find the current gain,
G p = ^
P- 42. For the circuit of Figure P 1.42, suppose
(c) Findthepowerabsorbedbyeachresistor. ^ 3= 10i?i. Find the resistor values
. I I so that Gy = —^ = 1000
Vin
2001
Figure P I.39
^2.
Figure P I.42
CHECK; 4 . 1 mA
C H EC K : i?,= 5 k Q
n
C H A P
One use o f resistors in electronic circuits is to control current flow, just as dams control water flow
along rivers. Ohm’s law, V = RI, gauges the ability o f resistors to control this current flow: for a
fixed voltage, high values o f resistance lead to small currents, whereas low values o f resistance lead
to higher currents. This property underlies the adjustment o f the blower (fan) speed for ventila
tion in a typical car, as represented in the following diagram.
In this diagram, three resistors are connected in series, and their connecting points are attached to
a switch. As we will learn in this chapter, the resistance o f a series connection is the sum o f the
resistances. So with the switch in the low position, the 12-V car battery sees three resistors in series
with the motor. Th e series connection o f three resistors represents a “large” resistance and heavily
restricts the current through the motor. W ith less current, there is less power, and the fan motor
speed is slow. W hen the switch moves to the Med-1 position, a resistor is bypassed, producing less
52 Chapter 2 • K irchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
resistance in the series circuit and allowing more current to flow. More current flow increases the
fan motor speed. Each successive switch position removes resistance from the circuit, and the fan
motor speed increases accordingly.
Analysis o f such practical circuits builds on the principles set forth in this chapter.
CHAPTER O U TLIN E
CHAPTER OBJECTIVES
Define and utilize Kirchhoff’s current law (KCL), which governs the distribution o f cur
rents into or out o f a node.
Define and utilize Kirchhoff’s voltage law (KVL), which governs the distribution o f volt
ages in a circuit.
Introduce series and parallel resistive circuits.
Develop a voltage division formula that specifies how voltages distribute across series con
nections o f resistors.
Develop a current division formula that specifies how currents distribute through a par
allel connection o f resistors. '
Show that a series connection o f resistors has an equivalent resistance equal to the sum o f
the resistances in the series connection.
Show that a parallel connection o f resistors has an equivalent conductance equal to the
sum o f the conductances in the parallel connection.
Explore the calculation o f the equivalent resistance/conductance o f a series-parallel con
nection o f resistances, i.e., a circuit having a mixed connection o f series and parallel con
nections o f resistors.
Explore the calculation o f voltages, currents, and power in a series-parallel connection o f
resistances.
Revisit the notion o f a dependent source and use a V C C S to model an amplifier circuit.
Describe a practical battery source and look at a general practical source model.
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 53
An important property o f the series connection o f Figure 2.2a is that all the rwo-terminal elements
carry the same current, in this case because the input current for each two-terminal element
must equal the exit current. Similarly, in a parallel connection, such as Figure 2.2b, the same volt
age, in this case, Vj^, appears across every circuit element.
node 1
+ 0
-1-
node
Vr
-
d- b node 2
(a) (b)
FIG U RE 2.2 (a) Series connection of resistors with the property that each resistor carries
the same current; (b) parallel connection of resistors with the property that
the same voltage appears across each resistor.
Sources interconnected with circuit elements produce currents through the elements and voltages
across the elements. For example, a voltage source connected across Figure 2.2a would generate a
current 2^ and the voltages through v^. K irchhoff’s voltage law (KVL) governs the distribu
tion o f voltages around loops o f circuit elements, as shown in Figure 2.2a. Similarly, a current
source connected across the circuit o f Figure 2.2b would produce the voltage and the currents
Z] through K irchhoff’s current law (KCL) governs the flow o f currents into and out o f a com-
54 Chapter 2 • KirchhofPs Current & Voltage Laws and Series-Parallel Resistive Circuits
mon connection point or node, as in the top and bottom connections o f Figure 2.2b. This chap
ter sets forth precise statements o f these laws and illustrates their application.
A proper statement o f KVL and KCL requires the additional notion o f branch. A branch o f a cir
cuit is a generic name for a two-terminal circuit element and is denoted by a line segment, as in
Figure 2.3. T he endpoints o f a branch (the terminals o f the circuit element) are called nodes, as
in Figure 2.3a. Ordinarily, however, node means a common connection point o f two or more cir
cuit elements (branches), as shown in Figure 2.3b.
node A
(a) (b)
FIG U RE 2.3 (a) Single branch representing a circuit element with terminals labeled as
nodes A and B; (b) interconnection of branches (circuit elements) with common
connection points labeled as nodes A through D.
The voltage polarity and current direction for the branches in Figures 2.2 and 2.3 are labeled in accor
dance with the passive sign convention; the arrowhead on a branch denotes the reference current direc
tion, which is from plus to minus. Recall that the + to - does not mean that the voltage is always posi
tive if measured from the plus-sign to the minus-sign. In general, reference directions can be assigned
arbitrarily. The conventional assignment o f voltage polarity and current direction to voltage and current
soiurces is given in Figure 2.4, which is different from the passive sign convention. Note that with these
conventional assignments, the (instantaneous) power delivered by a source is/^/^) = th^
power absorbed by a source is = -pjeff)-
Circuit Circuit
' 'J O ©
(a) (b)
FIG U RE 2.4 Conventional labeling o f (a) voltage, and (b) current sources.
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 55
Imagine a number o f branches connected at a common point, as at node A o f Figure 2.3b. The
current through each branch has a reference direction indicated by an arrow. If the arrow points
toward the node, the reference direction o f the current is entering the node; if the arrow points
away from the node, the reference direction o f the current is leaving. If a current is referenced as
leaving a node, then the negative o f the current enters the node, and conversely.
The two statements o f KCL are equivalent because the negative o f the sum o f the currents enter
ing a node corresponds to the sum o f the currents leaving the node. Further, from physics we know
that charge is neither created nor destroyed. Thus, the charge transported into the node must equal
the charge leaving the node because charge cannot accumulate at a node. KCL expresses the con
servation o f charge law in terms o f branch currents. Moreover, KCL specifies how branch currents
interact at a node, regardless o f the type o f element connected to the node.
Referring to Figure 2.3b, KCL at nodey4 requires that i^{t) + - i^ii) = 0 for all t. KCL at node
B requires that Finally, KCL at node D requires that 25(f) =
E X A M P L E 2.1
For the node shown in Figure 2.5, find
S o lutio n
By KCL, the sum o f the currents entering the node must be zero. Hence, the current z^(^) =
9cos(2r) - 3cos(2?) - cos(2r) - 2cos(2z) = 3cos(2^) A.
Exercise. 1. Suppose the current through the voltage source in Figure 2.5 is changed to - 2 cos(2?).
Find
AN SW ER: - 4 cos(2^) A.
2. Three branches connect at a node. All branch currents have reference directions leaving the
node. If /j = /2 = 2 A, then find ly
A N SW ER; - 4 A
Two implications o f KCL are o f immediate interest. First, as a general rule, KCL forbids the series
connection o f current sources. Figure 2.6a shows an invalid connection o f two arbitrary current
sources i,(?) and i^i), where z,(?) i^t). It is invalid because KCL requires that i^{t) = i2 {t). On
the other hand, a parallel connection o f two current sources can be combined to form an equiva
lent source, as in Figure 2.6b, where = i^{t) + i2 {t).
(a)
-O
(b)
A second immediate consequence o f KCL is that a current source supplying zero current [i{t) = 0
in Figure 2.7] is equivalent to an open circuit because the current through an open circuit is zero.
An open circuit has infinite resistance, or zero conductance. This means that a current source has
infinite internal resistance. From another angle, a constant current source is represented by a ver
tical line in the iv plane (see Figure 1.4 lb ). The slope o f the vertical line, which is infinite, deter
mines the internal resistance o f the source.
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 57
-O
+
v(t)
-O
E X A M P L E 2.2
In the parallel resistive circuit o f Figure 2.8, the voltage across each resistor is 6 cos(z) V. Find the
current through each resistor and the current, supplied by the voltage source.
'm
6cos(t) V
IQ <20 <3Q
S o lu tio n
By Ohm ’s law,
/^ l ( 0 = 6 c o s ( 0 A
. 6 c o s (/)
'R l W -------- = 3 c o s ( 0 A
, „ 3 , „ - 5 i 2 5 W = 2 c o s ( ,) A
By KCL,
Exercise. 1. In Figure 2.8, suppose the source voltage is changed to a constant, labeled Find
Iin in terms o f V-^.
AN SW ER:
58 Chapter 2 • KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
2. Suppose the source voltage in the circuit o f Figure 2.8 were changed to - 1 2 cos(2?) V. Find
AN SW ER: - 2 2 cos(2?) A
Kirchhoff’s current law holds for closed curves or surfaces, called Gaussian curves or surfaces. A
Gaussian curve or surface is a closed curve (such as a circle in a plane) or a closed surface (such as
a sphere or ellipsoid in three dimensions). A Gaussian curve or surface has a well-defined inside and
outside. Figure 2.9 illustrates the idea o f a Gaussian curve for three (planar) situations.
.............
Two
Terminal
Circuit
Elem ent /
For the two-terminal circuit element o f Figure 2.9a, KCL for Gaussian curves implies that i^{t) =
i2 {t), which is precisely the definition o f a two-terminal circuit element. For the three-terminal
device o f Figure 2.9b, KCL for Gaussian curves implies that i^{t) = i^{t) + i2 {t). Finally, for Figure
2.9c, i^ —i^ + if^ = 0. From these illustrations, one might imagine that the use o f Gaussian surfaces
might simplify or provide a short cut to certain branch current computations. T he general state
ment o f KCL for Gaussian surfaces is next followed by an example that demonstrates its use for
computing branch currents.
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 59
E X A M P L E 2 .3
This example shows how the use o f a Gaussian curve or surface can sometimes simplify a calcula
tion. Figure 2.10 portrays a complicated circuit whose branch currents and voltages are not solv
able by methods learned so far. Our objective is to find the current without having to solve a
set o f complex circuit equations.
FIG U RE 2.10 Circuit for Example 2.3, showing a Gaussian surface to compute directly.
S o lutio n
Using KCL for the indicated Gaussian curve, - 1 .1 5 + / ^ - 0.3 + 0.95 = 0. Equivalently, /^ = 1.15
+ 0.3 - 0.95 = 0.5 A.
In the next chapter, circuits such as the one in Figure 2 .10 are analyzed using a technique called
nodal analysis.
Exercise. 1. Draw a Gaussian surface on the circuit in Figure 2.10 that is different from the sur
face given but still allows one to compute /^.
AN SW ER; One choice is a circle enclosing the bottom node.
60 Chapter 2 • KirchhofF’s Current & Voltage Laws and Series-Parallel Resistive Circuits
2. Draw an appropriate Gaussian curve to find / in the graphical circuit representation in Figure
2 . 11.
AN SW ER: 2 A
Kirchhoff’s voltage law (KVL) specifies how voltages distribute across the elements o f a circuit.
Before conveying four equivalent versions o f KVL, we first set forth several necessary background
concepts. The first is the notion o f a closed path. In a circuit, a closed path is a connection o f
two-terminal elements that ends and begins at the same node and which traverses each node in
the connection only once. Figure 2.12 illustrates several closed paths. One closed path is A-B-C-
D-E-A, i.e., it begins at node A, moves to node B, drops to node C, moves through element 4 to
node D, down through element 6 to reference node E, and back through the voltage source to A.
A second closed path is A-B-C-E-A, and a third is B-D-C-B.
A second concept pertinent to our KVL statements is that o f a node voltage with respect to a ref
erence. A node voltage o f a circuit is the voltage drop from a given node to a reference node. The
reference node is usually indicated on the circuit or is taken as ground. The circuit o f Figure 2.12
has branches labeled 1 through 6 and nodes labeled A through E, with node E taken as the refer
ence node. The associated node voltages are denoted by Vg, Vq and v^. The voltage denotes
the voltage drop from node A to node E-, denotes the voltage drop from node D to node E, and
similarly for the remaining node voltages. Node E, being the reference node, has zero as its node
voltage.
Chapter 2 • K irchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits 61
+ V. v„
BD
FIG U RE 2.12 Circuit diagram illustrating (i) three closed paths {A-B-C-D-E-A)\
(ii) the concept of node voltages with respect to a given reference node E, and Vj^-,
(iii) the concept of branch voltages and
The concept o f a closed path and the concept o f a node voltage allow us to state our first two ver
sions o f Kirchhoff’s voltage law.
Statem ent 1: The algebraic sum o f the voltage drops around any closed path is zero at every
instant o f time.
Statem ent 2 : For any pair o f nodes j and k, the voltage drop Vjj^ from node j to node k is
given by
at every instant o f time, where Vj is the voltage at node j with respect to the reference and
is the voltage at node ^ with respect to reference. Herey and k stand for arbitrary node indices.
For example, in Figure 2.12, j, k can be any o f the nodes A, B, C, D, or E.
62 Chapter 2 • KirchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
Referring back to Figure 2.12, for the closed A-B-C-D-E-A, statement 1 o f KVL requires that
^AB ^ ^BC ^CD + ^DE ^EA ^ O' for Figure 2.12, from statement 2 o f KVL, the branch
voltages = v^ - Vg and Vqj^ = ^D- Hence, = -v^. Thus, by knowing the node
voltages o f a circuit, one can easily compute the branch voltages.
Exercise. 1. Find Vy^g, VgQ and V^^-for the circuit o f Figure 2.1 3 in which we have introduced
the ground symbol at node E to identify the reference node.
AN SW ERS: = - 3 V, Vg(^=2\ V, = 18 V
2. Again, with reference to Figure 2.13, find the node voltages V^, Vg, Vq and
AN SW ERS: 2 V, 5 V, - 1 6 V, - 6 V
FIG U RE 2.13
A third concept needed for two further equivalent statements o f KVL is that o f a closed node
sequence. A closed node sequence is a finite sequence o f nodes that begins and ends at the same
node. A closed node sequence generalizes the notion o f a closed path. Finally, we define the notion
o f a connected circuit. In a connected circuit, each node can be reached from any other node by
some path through the circuit elements. Figures 2 .12 and 2.14 show connected circuits. However,
in Figure 2.14, the sequence o f nodes A-B-C-D-E-A is a closed node sequence but not a closed
path because there is no circuit element between nodes B and C.
+ 2.5V - + 10V -
D
+
V,
F IG U R E 2 .1 4 Simple dependent source circuit for illustrating the concepts o f a connected cir-
cuit and a closed node sequence.
Chapter 2 • KirchhofF’s Current & Voltage Laws and Series-Parallel Resistive Circuits 63
Statem ent 3 ; For connected circuits and any node sequence, say A -D -B-... -G-P, the volt
age drop
^AP = ^AD + ^DB + - + GP
at every instant o f time.
S tatem ent 4 : For connected circuits, the algebraic sum o f the node-to-node voltages for any
closed node sequence is zero for every instant o f time.
FIG U RE 2.15 Circuit with nodes labeled A through E. Node E is taken as the reference node.
64 Chapter 2 • K irchhoff s Current & Voltage Laws and Series-Parallel Resistive Circuits
Two further implications o f the KVL are o f immediate interest. First, as a general rule, KVL for
bids the parallel connection o f two voltage sources— say, and V2 (^)— for which Vj(?)
as illustrated in Figure 2.16a. O n the other hand, two voltage sources in series can be combined
to form a single source, as illustrated in Figure 2.16b, where {t) = v-^{t) + V2 {t).
FIGURE 2.16 (a) An improper connecdon of voltage sources when v^{t) ^ i>2 (i);
(b) an equivalent representation of two voltage sources connected in series in which = Vj{i) +
-O -o
+
ov ov
Finally, note that all four KVL statements can be justified using the definition and the notation
for “voltage” drop presented in Chapter 1. The justification is more readily comprehended via the
analogy o f the gravitational field, also developed in that section. Also, observe that KVL holds for
all closed node sequences, independent o f the device represented by each branch o f the connect
ed circuit. The distribution o f voltages around closed paths can be viewed as a special case o f this
general statement.
During holidays, one often sees strings o f lights hanging between poles or trees. Sometimes these
strings consist o f a series connection o f light bulbs. Each light bulb contains a filament, a coil of
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 65
wire, that gives o ff an intense light when hot. In a circuit’s perspective, the filament acts as a resis
tor and has an equivalent hot resistance. The series connection o f bulbs can be modeled by a series
connection o f resistors, with each resistor paired with a specific bulb. Computing the voltage
across each light (a very important type o f calculation) would then be equivalent to finding the
voltage across each o f the resistors in the equivalent circuit model. It is quite common to model
electrical loads, such as a light, by resistors.
E X A M P L E 2 .4
Figure 2.18a shows a voltage source v-JJ) connected to three resistors in series. The objectives o f
this example are to compute the voltages Vj{t),j = 1, 2, 3, across each resistor, and the
equivalent resistance seen by the voltage source.
ijt )
(a) (b)
FIGURE 2.18 (a) Three series resistors connected across a voltage source. By the
definition o f a two-terminal resistor or by the KCL, the current through each resistor is
(b) equivalent resistance = R^ ■¥Rj + R^ seen by the source, i.e., v-^{t) = Reqii„{t)-
So l u t io n
Step 1. Express the voltage across each resistor in terms o f the input current. For the circuit o f Figure
2.18a, the current through each resistor is i^JJ) by KCL. From Ohm’s law, the voltage across each
resistor is
fo t j = 1, 2, 3.
Step 2. Express v-J^t) in terms ofi-JJ), solvefor i-^<^t), and then compute an expression for v^t) in terms
o f the Rj and v-J^t). By KVL, the source voltage equals the sum o f the resistor voltages, i.e..
(2 . 1)
where we have substituted Rjii„{t) = Vj{t). Dividing Equation 2.1 by (i?j + Rj + -^3) yields
,2 2^)
Equation 2.2a is a volt^e division formula for a three-resistor series circuit. This formula imphes
that if a resistance R. is small relative to the other resistances in the series circuit, then only a small
portion o f the source voltage develops across it. O n the other hand, if a resistance R. is large rela
tive to the other resistances, then a larger portion o f the source voltage will develop across it. One
concludes that the voltage distributes around a loop o f resistors in proportion to the value o f each
resistance. The proportion is simply the ratio o f the branch resistance R- to the total series resist
ance.
Step 3. Compute the equivalent resistance R^^ seen by the voltage source. The equivalent resistance
seen by the voltage source for a resistive circuit is implicidy defined by Ohm s law, i.e., Vi^i) =
^eqhri^^- nonzero currents, the equivalent resistance is defined as
Figure 2.18b illustrates the idea o f the equivalence. By Equation 2.1, v^^i) = R^,^ij„{t) = (7?^ +
+ implies that the equivalent resistance is R^^ = R-^ + R^ + Ry This means that from the
perspective o f the voltage source, the series connection o f resistors is equivalent to a single resistor
o f value equal to the sum o f the resistances. A formal discussion o f equivalent resistance and its
generalization (the Thevenin resistance) is taken up in Chapter 6 .
Example 2.4 suggests some generalizations. Consider Figure 2.19. The first is that the equivalent
resistance R^^ seen by the source is the sum o f the resistors. This means that resistances in series add,
i.e., resistors in series can be combined into a single resistor whose resistance is the sum o f the indi
vidual resistances.
Req = R \ + R 2+ "' + Rn
Further, since vi^t) = Rjii„{t), a general voltage division formula can be derived as
Ri
R\ + + ■■' + Rn
(2 .2 b)
fory = 1, ... , n.
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 67
Exercise. In Figure 2.19, suppose each resistor has value Rq. Find the equivalent resistance seen by
the source and the voltage across each resistor in terms o f the source voltages.
A N SW ER: nR^^, v jt)ln
So l u t io n
From the preceding discussion, R^^ is defined by Ohm’s law, i.e., v-J^t) = Rg^i-Js)-
Step 1. Express v-^ in terms o f the remaining branch voltages. From KVL,
Step 2 . Express the branch voltages in terms o f and substitute into Equation 2.3. To express and ^2 in
terms of i^^, observe that i^^ is the current through each resistor (KCL or definition o f two-terminal circuit
element) and use Ohms law: = 2i-^ and Vj = 14/^^. Subsrimting into Equation 2.3 yields
^ in = 20//„ = R e q iin
Therefore, = 20
Notice that the dependent source increases the resistance o f the two series resistors by 4 Q.
Dependent sources can increase or decrease the resistance o f the circuit. W ith dependent sources,
it is even possible to make the equivalent resistance negative.
Step 3. Find the power absorbed by the 14-Q. resistor. To find the power absorbed by the l4-£2 resis
tor when = 2 V, first compute i^^ via Ohm s law; i^^ = v J R = 2/20 = 0.1 A. It follows that P
= = 0 .0 1 x 1 4 = 0.14 W.
Exercise. Suppose the dependent source in the circuit o f Figure 2 .20 has its value changed to 2[v-^
+ V2 ). Find R^q.
AN SW ER: 48 Q
Many o f the electrical outlets in the average home are connected in parallel. W hen too many
appliances are connected to the same outlet or set o f oudets on the same fused circuit, a fuse will
blow or a circuit breaker will open. Although each appliance uses only a portion o f the maximum
allowable current for the (fused) circuit, together, the total current exceeds the allowable limit.
Because o f this common occurrence, an engineering student ought to know how current distrib
utes through a parallel connection o f loads (resistors).
To keep the analysis simple, consider a set o f three parallel resistors driven by a current source.
E X A M P L E 2 .6
Figure 2.21a shows a circuit o f three parallel resistors driven by a current source. Our objectives are to
find expressions for v-Jyf), ij^t) in terms o f the input current i-JJ) and the circuit conductances (the recip
rocal o f the circuit resistances) and the equivalent resistance seen by the current source.
-o
0 -I-
- f-
ijt ) v jt ) v Jt)
ijt )
-O
(a) (b)
So l u t io n
Step 1. Find expressions for ij{t) in terms o f v-J^t). The variable that Hnks the branch current ij{t) to
the input current i^J^t) is the voltage which by KVL appears across each resistor. Since v-J^t)
appears across each o f the resistors, Ohm’s law implies that each resistor current is
ij{t) = ^
^ (2.4)
Step 2. Compute v^^i) in terms ofi-J^t). Applying KCL to the top node o f the circuit yields
Using Equation 2.4 to substitute for each ij{t) and then solving for v-^ yields
Step 3. Compute ijyi) in terms ofi-J^t). To obtain a relationship between i-J,t) and ij{t), substitute
Equation 2.5 into Equation 2.4 to obtain
1
. G, . G, ,
^ ± + + G 1 + G 2 + G3
Rl R2 R3
(2 .6)
Equation 2.6 is called a current division formula. It says that currents distribute through the
branches o f a parallel resistive circuit in proportion to the conductance o f the particular branch
G. relative to the total conductance o f the circuit G^^ = G^ + G2 + Gy The greater the conduc
tance, i.e., the smaller the resistance, the larger the proportion o f current flow through the associ
ated branch.
Step 4. Compute the equivalent resistance seen by the source. As in Example 2.5, Ohm’s law, i-^
= G^ v-^, defines G^^ or, equivalently, From Equation 2.5, G^^ = G j + G2 + G3 is the equiva
lent conductance o f the parallel circuit, and the equivalent resistance is
1 1 1
R.
± + G 1 + G2 + G 3 G ,,
R, R2 R3
Exercise. In Figure 2.21a, suppose = 1 Q, = 0-5 and = 0.5 O.. Find the current
through if = 1Oe~‘ A.
A N SW ER: z, (?) = 2e“^A
These ideas generalize to n resistors in parallel, as illustrated in Figure 2.22. In particular, the
equivalent resistance R o f the parallel set o f resistors in Figure 2 .22 is
Ren = - 1 1
R, R2 R.,
(2.7)
^eq = + *^2 + ••• + equivalent conductance. Further, the current through each
branch satisfies the general current division formula
R, G/ . G, .
1 1 1 rj A
G +. r A-•■•j-+ rG „
G2 + ~ Gr^ q
+ ---- -!-••• +
^1 ^2 R„
(2 .8 )
■ o
-I-
ijt ) ( f ) V, (t)
-o
Exercise. Consider Figure 2.22. Suppose ten 10 Q resistors are in parallel. Find and the cur
rent through each resistor.
AN SW ER: \ Q. and each current is Q.\i-J^t)
E X A M P L E 2 .7
Consider the circuit o f Figure 2.23 exhibiting a current source driving two parallel resistors. Show
that
R Ry
■o
©
-O
So l u t io n
Step 1. Find the equivalent resistance seen by the current source. From Equation 2.7, with « = 2, it
follows that
1 R^R2
Ri ^ Ri
This formula, called the product over sum rule, is quite useful in many calculations.
Step 2. Find i^{t) and i^it)- From Equation 2.8, with w = 2, it follows that
Gi R^ ■
■ X 7X ■
Ry
and
/2 (0 =
Gi 4- G 2 ,__ Ri + R'^
Ri R2
72 Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits
E X A M P L E 2 .8
For the circuit o f Figure 2.24, find the input voltage the current ijit) through i?2> and the
instantaneous power absorbed by Rj when
5e ‘
^ 0 t< 0
o
y r i,(t) U i3 (t) l|i,( t )
ijt ) ^2 SG 3 <G ,
So l u t io n
Step 1. Compute the equivalent conductance and equivalent resistance o f the circuit. Since conduc
tances in parallel add,
^ e q - G j + G 2 + G 3 + G 4 = 0.25 S
and
Step 2 . Compute v^JJ). From Ohm’s law, the voltage across the current source is
2 0 e~ 'V t^ 0
0 t< 0
Step 3. Compute the current i2 (t). Using the current division formula o f Equation 2.8 yields
3e~'A t > 0
Cjeq 0.25 0 r<0
Chapter 2 • K irchhoff’s Current & Voltage Laws and Series-Parallel Resistive Circuits 73
Step 4. Compute the power absorbed by Rj- To compute the power absorbed by for t >0,
Exercises. 1. For the circuit o f Figure 2.24, find i/^(t) and the power absorbed by i?4 .
A N SW ER: 0.6^-^ A, 12 W
2. In the circuit o f Figure 2.24, suppose each conductance is doubled and i-J^t) = 100 mA. Find
R , V- (?), and the power absorbed by the new Gy
AN SW ER; 2 Q, 200 mV, and 1.6 mW
The last two sections covered series and parallel resistive networks. Suppose we take a series circuit
and connect it in parallel with another series circuit; this is a parallel connection o f two series cir
cuits. Alternately, we could take two parallel circuits and connect them in series. This would result
in a series connection o f parallel circuits. We could also put a series connection o f two parallel sub
circuits in parallel with a replica o f itself or some other series or parallel circuit. Many other inter
connections are possible. Arbitrary series and parallel connections o f such subcircuits are called
series-parallel circuits. This section explores the calculation o f the equivalent resistance o f series-
parallel circuits by repeated use o f formulas for series and parallel resistance computation. Related
voltage and current computation is also explored. Example 2.11 presents a practical application o f
series-parallel concepts.
EXA M PLE 2 .9
Find the equivalent resistance, R^^, and the voltage across the source, the voltages V2 , the
power absorbed by the 6 kQ resistor, and the power delivered by the source for the circuit o f
Figure 2.25, when = 20 mA.
So l u t io n
Step 1. Compute To compute first compute R^^y and
8 x 4.8 38.4
Kq\ = ------= — = 3 kQ
1+ 4 .8 12.e
and
^^■^2 1 1 1 ^ 6
---- 1---- 1--- -------
12 6 4 12
Finally,
= Reqhn = 20 X 1.5 = 30 V
^source ~ ^inhn — 30 x 0 .0 2 = 0 .6 W
Exercise. 1. W hat is the current through the 2 k£2 resistor firom top to bottom?
AN SW ER: 50 mA
2 . In Example 2.9, suppose the resistance o f each resistor is doubled. Find the new R^ and the
power delivered by the source.
AN SW ER: 3 k n , 1.2 watts
This example points out a very interesting fact: finding the equivalent resistance o f a series-paral
lel connection o f resistors requires only two types o f arithmetic operations no matter the network
complexity: adding two numbers and taking the reciprocal o f a number. A hand calculator easily
executes both operations. Such is not the case with a non-series-parallel network. To find the
equivalent resistance o f a non-series-parallel network, one usually must write simultaneous equa
tions and evaluate determinants, a topic detailed in Chapter 3 .
Chapter 2 • KirchhofF’s Current & Voltage Laws and Series-Parallel Resistive Circuits 75
It is then important to recognize when a problem belongs to the series-parallel category in order to take
advantage o f the simple arithmetic operations. In the previous series-parallel examples, one— and only
one— independent source was specified on the circuit diagram. This is part o f the definition o f a series-
parallel network. The independent source must be indicated, or, equivalendy, the pair o f input termi
nals to which the source is connected must be specified. The specification o f the input terminals deter
mines whether or not a network is series-parallel. The following example illustrates the effect of differ
ent input terminal designations on the computation o f equivalent resistance.
E X A M P L E 2 .1 0
For the circuit o f Figure 2.26a, determine whether or not the network is series-parallel as seen
from each o f the following terminal pairs:
1. C a s e l;( A , B)
2. Case 2; (A, C)
3. Case 3: (C, D)
If the answer is affirmative, give an expression and compute the numerical value for the equivalent
resistance, using the notation // (double slash) for combining resistances in parallel, i.e., means
and are parallel, and R^IIR2 llR^ means is in parallel with vvhich is in parallel with Ry
(a)
>R,
R1
A<
D'
(0
FIG U RE 2.26 (a) From terminals (C, D) the network is not series-parallel. However, from terminals
{A, B) the network is a series-parallel one. (b) Redrawing of the network of (a) as seen from terminals
(A, Q; the resulting network is series-parallel.(c) Non-series-parallel network seen from (C, D).
76 Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits
So l u t io n ^
Case 1. Find equivalent resistance seen at (A, B). W ith an independent source connected to nodes
A and B, the source sees a series-parallel network. By inspection o f Figure 2.26a, the equivalent
resistance is
Case 3. Find equivalent resistance seen at (C, U). W ith (C, D) as the input terminal pair, the net
work is not series-parallel, as can be garnered from Figure 2.26c. T he calculation o f for this
case requires methods to be discussed in Chapter 3 and is omitted.
In electrical engineering laboratories, a student often uses a meter to measure voltages associated
with a piece o f electronic equipment. In older laboratories, or when using an inexpensive meter,
the voltage reading will sometimes differ from what the student calculated or expected to meas
ure. Typically, this results from the loading effect o f the meter. Using the concept o f series-paral
lel resistances, the following example explores the phenomenon o f loading.
E X A M P L E 2 .il
Suppose the circuit in Figure 2.27a is part o f a laboratory experiment to verify voltage division. In
this experiment, you calculate the expected voltage Vq and then measure the circuit voltage using
an inexpensive voltmeter.
(a) Calculate the expected voltage Vq in Figure 2.27a.
(b) A voltmeter with a 1-kQ/V sensitivity is used to measure V q.You use a 0 -10-V range. In
this range, the meter is represented by a 10-kQ resistance, i.e., 10 kD = full-scale reading
X meter sensitivity = 10 V x 1 kQ/V. W hat voltage will the meter read?
(c) A better-quality voltmeter with a 2 0 - k H / V sensitivity is used to measure the same volt-
^0’ ^ 0 -10-V scale. This better-quality meter is represented by a 2 0 0 -k tl resist
ance. W hat new voltage will the meter read?
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 77
" h 20 ko 20 kO >10 kO 20 kQ
-o
FIG U RE 2.27 Three circuits for exploring the effect of loading on a circuit: (a) circuit for validating
voltage division; (b) circuit o f (a) with an attached voltmeter having an internal resistance o f 10 k£2;
(c) circuit o f (a) with an attached voltmeter having an internal resistance of 200 k^2.
So l u t io n
(a) Voltage division on the circuit o f Figure 2.27a yields
20
Vo = -15 = 10 V
(b) O n the 0-10-V range, the voltmeter internal resistance between the probes is 10 kD, as
stated. This represents a 1G-Id2 load connected in parallel with the 20-kQ resistance, as
shown in Figure 2.27b. The voltage Kg will now change because the 15-V source no
longer sees 10 ld2 in series with 20 kO. Rather, the source sees 10 kQ in series with 6 .67
k ii = 20 kX2//10 kQ. By voltage division,
6 .67
Vo = - -15 = 6 V
10 -H6 .67
(c) O n the 0 -10-V range with the better voltmeter, the internal resistance between the probes
is 200 kD. As before, this represents a 200-kQ load connected in parallel with the 2 0 -k 0
resistance, as shown in Figure 2.27c. 20 kQ//200 klQ = 18.18 kO. By voltage division,
this yields
18.18
Vb = -15 = 9 .6 7 7 V
10 + 18.18
Example 2.11 demonstrates the effect o f loading due to a measuring instrument, emphasizing the
importance o f choosing a good voltmeter with adequate sensitivity. Although modern-day volt
meters typically have sensitivities better than 20 kQ/V, a meter with a sensitivity o f 1 kQ./Y is used
in the example to dramatize the effect o f loading.
2. Th e circuit o f Figure 2.28 shows a voltage divider whose voltage Kq is to be measured by a volt
meter having an internal resistance o f 80 kO. Find Kq without the meter attached, and then find
the value o f Vq measured by the meter.
20 V
AN SW ER: 15 V, 13.71 V
Chapter 1 introduced the notion o f a dependent or controlled source whose voltage or current
depends on the voltage or current in another branch o f the network, i.e., each source has a con
trolling voltage or current and an output voltage or current. Figure 2.29 depicts the four types o f
controlled sources designated by a diamond containing either a ± or an arrow:
1. Voltage-controlled voltage source (VCVS)
2. Voltage-controlled current source (VCCS)
3. Current-controlled voltage source (CCVS)
4. Current-controlled current source (CCCS)
An arrow inside the diamond indicates a controlled current source having the reference current
direction given by the arrow. A ± inside the diamond specifies a controlled voltage source, with
the reference voltage polarity given by the ± sign. A parameter value completes the specification
o f a linear controlled source. In Figure 2.2 9 the (constant) parameters are fx, g^, r^, and |3. These
symbols are common to many electronic circuit texts and have useful physical interpretations to
practicing engineers and technicians. For consistency, a ^^-type controlled source is a V C C S and
a jO,-type source is a V C V S, and so on.
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 79
O
+
a ................... o
(a) VCVS or |j-type (b) VCVS or g -type
O o o O
Pi,
>r
<;
6- ■o o - ......... .........o
(c )C C V S o rr -type (d) CCCS or P-type
In practical controllecl sources, the controUing voltage (t>j in Figure 2.29a and b) or current (z'j in
Figure 2.29c and d) is ordinarily associated with a particular circuit element, but not always. For
generality, the controlling voltage in Figure 2.29a and b is shown across a pair o f nodes. Also,
in Figure 2.29c and d, the controlling current Zj is shown to flow through a short circuit. (Strictly
speaking, neither an open circuit nor a short circuit is a circuit element.) In a real circuit, the cur
rent may be flowing through an actual circuit element, such as a resistor or even a source.
In Figure 2.29b, once the controlling voltage v-^ is known, the right-hand source behaves as an
independent current source o f value Since the unit for is amperes and the unit for
is volt, the unit for^^ is amperes per volt, or siemens. Since has units o f conductance, and the
controlling and controlled variables belong to two different network branches, is called a trans
fer conductance, or transconductance. The other controlled sources have a similar interpretation.
The parameter has the unit o f resistance, ohms, and is called a transfer resistance. The param
eter |i is dimensionless because the controlling voltage has units o f volts and the output vari
able must have units o f volts. Similarly, the parameter (3 is dimensionless. The units and asso
ciation are set forth in Table 2.1.
80 Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits
Transfer
VCCS Sm siemens
conductance
Figure 2.29 portrays each controlled source as a four-terminal device. In practical circuits, the
great majority o f controlled sources have one terminal or node in common, making them three-
terminal devices. The dashed lines joining the two bottom nodes in Figure 2 .29 suggest this quite
common configuration.
The controlled sources as defined in Figure 2.29 have linear v-i relationships. Controlled sources
may also have a nonlinear v-i relationship. In such cases, the element will be called a nonlinear
controlled source. This text deals only with linear controlled sources.
The next few examples describe some o f the unique features o f controlled sources.
Exercise. Find v^, and the power delivered by each source in Figure 2.30.
FIG U RE 2.30
AN SW ER: 4 V, 0.05 A, 1.6 W, 0.05 W
E X A M P L E 2 .1 2
This example analyzes the circuit o f Figure 2.31. The independent voltage source in series with
the 3-Q. resistor represents a practical source discussed at greater length later in this chapter. The
circuit within the box o f Figure 2.30 approximates a simplified amplifier circuit by a V C C S. The
8 -Q resistor is considered a load and could, for example, model a loudspeaker. Two important
quantities o f an amplifier circuit are voltage gain and power gain, which are computed here along
with various other quantities.
(a) Find the equivalent resistance seen by the independent voltage source.
(b) Compute / .
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 81
FIG U RE 2.31 Practical source (ideal independent voltage source in series with a resistor)
driving a simplified VCCS approximation of an amplifier circuit loaded by an 8 -Q resistor.
So l u t io n
(a) Since resistances in series add, = 3 + 47 = 50 Q.
(b) By Ohm’s law, /•„ = = 0.8 mA.
(c) To compute one must first compute K j. Here one can use Ohm’s law directly, since
we know /-^, or one can use voltage division. Doing the calculation by voltage division,
V = — 4 0 X 10'^ = 3 7 .6 X 10'^ V
‘ 50
Using this value o f and current division on the right half o f the circuit yields
0.125
79.8 X 3 7.6 X 1 0 '^ = 2 A
“ 0.125 + 0 .0 6 2 5
K,«.= 2 x 8 = 16V
^ = — = 400
0.0 4
82 Chapter 2 • K irchhofFs Current & Voltage Laws and Series-Parallel Resistive Circuits
(h) Pout is simply the product o f voltage and current delivered to the load
(i) The resulting amplifier power gain is the ratio o f the power absorbed by the 8 -Q load to
the power delivered to the amplifier, P-^,
= 1.064 X 10^
^ Pin 3 0 ,0 8
Exercise. Suppose the 8-f2 load resistor in Figure 2.30 is changed to 16 Q. Compute
and the power gain.
AN SW ER: 1.5 A, 24 V, 1.197 X 10<^
The analysis in Example 2.12 required only KCL, KVL, and simple voltage divider and/or cur
rent divider formulas. More complicated linear circuits necessitate a more systematic approach. To
see this need, add a resistor between the top o f the 47-Q resistor and the top o f the dependent
current source in Figure 2.31. The methods o f solution used in the example immediately break
down because the circuit is no longer series-parallel; hence, one cannot use voltage division to
compute V j. Chapter 3 will explain more systematic methods called nodal and loop analysis.
Unlike a passive element such as a resistor, which always dissipates power as heat, a controlled
source may generate power as computed in part (g) o f Example 2 . 12, or may dissipate power in
other cases. Since a controlled source has the potential o f generating power, it is called an active
element.
In Example 2.12, the practical voltage source delivers 30.08 pW o f power to the circuit, which is
easy to accept because the source could have been a small battery On the other hand, the con
trolled source generates 48 W. This seems a litde puzzling. Where does the power come from?
W hy not purchase a controlled source at a local electronics store and use it to power, say, a lamp?
Here it is important to recognize that a controlled source is not a stand-alone component picked
o ff the shelf like a resistor. A controlled source is usually constructed from one or more semicon
ductor devices and requires a dc power supply for its operation. The power delivered by the con
trolled source actually comes from the power supply. Here, we use the controlled source to math
ematically model an amplifier and facilitate analysis o f the circuit.
Chapter 2 • K irchhoff s Current & Voltage Laws and Series-Parallel Resistive Circuits 83
W ith simple series-parallel connections o f resistors, the equivalent resistance is always positive.
When controlled sources are present, a strange result may happen, as illustrated in the next exam
ple.
E X A M P L E 2 .1 3
Find the equivalent resistance for the circuit o f Figure 2.32 when (a) p = 0.5 and (b) p
= 2.
•O^
FIG U RE 2.32 Calculation of for a circuit with controlled source for two values o f p.
So l u t io n
W ith p unspecified, we can apply KVL to the single loop, noting that = V^. Here,
i-x
For p = 0.5, R^q = 2R, which means that the dependent source acts like a resistor o f R Q.. In this
case, it absorbs power. O n the other hand, for p = 2, R,^ = -R , a negative equivalent resistance. In
this case, the dependent source acts like a -2R-Q. resistor and, in fact, delivers power to the inde
pendent source. An important conclusion can be drawn from this example: in the study o f linear
circuit analysis, controlled sources allow the possibility o f negative resistances. Since a negative
resistance generates power, it is also an active element.
Exercise. In Figure 2.32, find the values o f p so that R^q - 0 .5 R and R^q - 2R.
A N S W E R :-!, 0.5
84 Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits
Exercise. For Figure 2.33, find for the following three values of^^: 0.5 mS, 1 mS, and 2 mS.
FIG U RE 2.33
AN SW ER: 2 k£l, open circuit, -1 kQ
The ideal battery o f Figure 1.30, repeated in Figure 2.34a, delivers a constant voltage regardless
o f the current drawn by a load. T he i-v plane characteristic is a horizontal line through V^, as
shown in Figure 1.40b and repeated in Figure 2.34b. Ideal batteries do not exist in the real world.
The terminal voltage always depends on the supplied current. A more accurate representation o f
a practical battery, but by no means a fully realistic one, is an ideal battery in series with a resist
ance, say, R^, as shown in Figure 2.34c. R^ is termed an internal resistance, which crudely models
the effects o f chemical action and electrodes inside the battery.
>
Vs
E X A M P L E 2 .1 4
This example shows the effect o f the internal resistance o f a battery on the terminal voltage.
Suppose a nickel-cadmium battery has an open circuit terminal voltage o f 6 volts. W hen con
nected across a 2-Q. resistor, the voltage drops to 5.97 V. Find the internal resistance o f the bat
tery.
So l u t io n
Figure 2.35 illustrates the situation. Here, the dashed box represents the battery model with inter
nal resistance R^. In Figure 2.35a, no load is connected to the battery. Hence, no current flows
through the internal resistance, in which case, the terminal battery voltage is 6 V.
(a) (b)
FIG U RE 2.35 Battery model with internal resistance; (a) open circuited (Is = 0); and
(b) connected to a 2 -0 load.
Figure 2.35b shows the battery connected to the 2-Q resistive load. The measured voltage is 5.97
V. By KVL, the voltage across the internal resistance, is Vj^ = 6 - 5-97 = 0.03 V. From Ohm’s
law, the current through is = (5.97/2) = 2 .985 A. Again, by Ohm’s law.
0.03
= 0 .0 1 0 0 5 Q
2.985
Exercise. In Example 2.14, suppose the internal resistance is known to be R^ = 0.005 Q and
although the load resistance is unknown, the load current is 4 A. W hat is the voltage across the
load resistance, and what is the load resistance?
AN SW ER; 5.98 V and 1.495 ^
Ideal voltage sources have zero internal resistance. Real voltage sources, such as batteries, have an
internal resistance. The value o f this resistance may change with the current load. There may also
86 Chapter 2 • KirchhofF’s Current & Voltage Laws and Series-Parallel Resistive Circuits
be other effects. However, for our purposes, a more realistic model o f a voltage source contains a
series internal resistance, as illustrated in Figure 2.36a.
L .(t)
(b)
FIG U RE 2.36 (a) A non-ideal voltage source as an ideal voltage source with an internal series resist
ance; (b) a non-ideal current source as an ideal current source with a parallel internal resistance.
Ideal current sources have infinite internal resistance. Real current sources have a finite, typically
large, internal resistance. Figure 2.36b depicts a more realistic current source model where the
internal resistance is in parallel with the ideal current source.
In the case o f constant voltage and current sources, ideal and non-ideal source models have a
graphical interpretation. The i-v (current-voltage) characteristic o f an ideal constant voltage
source {v^{t) = 1^) is a horizontal straight line. This means that the voltage supplied by the source
is fixed for all possible current loads. An ideal constant current source (z^(z) = has a vertical
straight line characteristic, which means that the current is constant for all possible voltages across
the source. Figure 2.3 7 illustrates these relationships graphically.
Vout
(a)
FIG U RE 2.37 v-i characteristics o f (a) an ideal constant voltage source, and
(b) an ideal constant current source.
Chapter 2 • KirchhofF’s Current & Voltage Laws and Series-Parallel Resistive Circuits 87
The non-ideal case is quite different. Because o f the internal resistance a non-ideal constant volt
age source i-v characteristic satisfies the linear relationship
hut — ^s'^oul ( 2 . 11 )
Equations 2.10 and 2.11 are illustrated by the graphs in Figure 2.38 when v^{t) = for the non
ideal voltage source and i^{t) = for the non-ideal current source. For a voltage source, if the value
o f R; is very small in comparison with potential load resistances, as ordinarily expected, then the
hne in Figure 2.38a approximates a horizontal line, the ideal case. O n the other hand, for a cur
rent source, the line in Figure 2.38b approximates a vertical line whenever is much much larg
er than a potential load resistance. This would then approximate the ideal current source case.
FIGURE 2.38 v-i characteristics of (a) non-ideal constant voltage source, and (b) non-ideal current source.
In a similar way, non-ideal dependent voltage sources are a connection o f an ideal dependent
source with a series resistance. A non-ideal dependent current source is a connection o f an ideal
dependent current source with a parallel resistance.
E X A M PLE 2 .1 6 Figure 2.39 shows the measured voltages o f a dc power supply found in an old
laboratory. Assuming a non-ideal model o f Figure 2.38a, find and the internal resistance R^ of
the power supply.
88 Chapter 2 • K irch h offs Current & Voltage Laws and Series-Parallel Resistive Circuits
Vout (V)
' '
FIG U RE 2.39 Graph of measured voltages and currents for a dc power supply.
So l u t io n
From Equation 2.10, we know that = -R^ + V^. From the graph, when = 0,
10 V = I/. Further, R^ = - (9.8 - 10)/(0.5 - 0.0) = 0.4 Q.
SUM M ARY
This chapter has presented the essential building blocks o f linear lumped circuit theory, beginning
with the two fundamental laws for interconnected circuit elements: KVL and KCL. KVL states
that for lumped circuits, the algebraic sum o f the voltages around any closed node sequence o f a
circuit is zero. Similarly, KCL says that for lumped circuits, the algebraic sum o f the currents enter
ing (or leaving) a node is zero.
These laws in conjunction with Ohm’s law allowed us to develop a voltage division and a current
division formula. The voltage division formula applies to series-resistive circuits driven by a volt
age source. The voltage developed across each resistor was found to be proportional to the resist
ance o f the particular element relative to the equivalent resistance seen by the source. For exam
ple, in a two-resistor series circuit, Rj in series with we found that
Vi =
7^1 + /?2
T he current division formula applies to parallel-resistive circuits driven by a current source. Here,
the current through each resistor with conductance Gi was found to be proportional to G/ divid
ed by the equivalent conductance seen by the source. Since conductance is the reciprocal o f resist
ance, the idea can also be expressed in terms o f the resistances o f the circuit. For example, in a two-
resistor parallel circuit, Rj is parallel with /?2>
G,
/i =
G 1 + G2 R, +Rn
In deriving the voltage division formula, we learned that the resistances o f a series connection o f
resistors may be added together to obtain an equivalent resistance, prompting the phrase “resistors
in series add.” Analogously, the derivation o f the current division formula for parallel circuits led
Chapter 2 • Kirchhofif’s Current & Voltage Laws and Series-Parallel Resistive Circuits 89
US to conclude that a parallel connection o f resistors has an equivalent conductance equal to the
sum o f conductances. This is sometimes expressed in terms o f resistances as the inverse o f the sum
o f reciprocal, i.e., for n resistors in parallel,
p .
7^1 R„
which leads to the very special formula for two resistors in parallel,
R =
Dependent sources, first introduced in Chapter 1, were re-examined in greater detail. Some prac
tical points were described.
All o f the above ideas were applied to the analysis o f series-parallel networks that are interconnec
tions o f series and parallel groupings o f resistors. Our analysis showed us how to compute the
equivalent resistance o f series-parallel circuits. An example was given that described the applica
tion o f these ideas to voltage measurement. This was followed by a discussion o f battery models
and battery usage. Finally, battery modeling ideas were used to describe non-ideal source models.
Kirchho£F’$ voltage law (KVL); for lumped circuits, the algebraic sum of the voltage drops
around any closed path in a network is zero at every instant of time. In general, for
lumped connected circuits, the algebraic sum of the node-to-node voltages for any closed
node sequence is zero for every instant of time.
Node: the common connection point between each element; in general, a node is a connection
point of one or more circuit elements.
Node voltage: the voltage drop from a given node to the reference node.
Parallel circuit: a side-by-side connection of two-terminal circuit elements whose top terminals
are wired together and whose bottom terminals are wired together.
Series circuit: a sequential connection o f two terminal circuit elements, end-to-end.
Voltage division: each resistor voltage in a series connection is a fraction o f the input voltage equal
to the ratio of the branch resistance to the total series resistance.
// (double-slash): notation for combining resistances in parallel, i.e., R^UR2 means and i ?2 are
in parallel, and R^IIR2 llRj^ means R^ is in parallel with which is in parallel with Ry
r\
r\ .
r\
n
Chapter 2 • K irchh off s Current & Voltage Laws and Series-Parallel Resistive Circuits 91
PROBLEMS
KIR C H H O FF'S C U R R EN T LAW
1. (a) Find the value o f /j for each o f the
node connections in Figure P2.1a and
P2.1b given that 1^ = 2 A, Figure P2.3
1^ = 3 A, and = 4 A. A N SW ER: ( b ) 4 A
(b) Repeat part (a) when l 2 = I^ = 1^ = 2 A.
4. (a) Find the value o f /j in the circuit o f
Figure P2.4.
(b) Find the value o f in the circuit o f
Figure P 2.4 by a single application o f
KCL.
Figure P2.1
A N SW ERS: (b) 0, 2 A
.1 0 0
40V
+ V, - N V, -
lOOV
6 , lO O v C ”^
Figure P2.2 + +
VI 5A
AN SW ERS: (b) (scrambled) 200, -300, 40V
-200, -300 mA
Reference Node Reference Node
K C L A N D KVL
9. (a) Consider the circuit o f Figure P2.9a. Use
KCL and KVL to find the voltage across
each current source from the arrow head
Figure P2.6 to the arrow tail and the current through
A N SW ERS: (a) 180 V, 50 V, -110 V, 10 V each voltage source from minus to plus.
Finally, find the power delivered by each
7. (a) Find the values o f the voltages Vp Kj, source and verify conservation o f power,
and Vj in the circuit graph o f Figure (b) For the circuit o f Figure P2.9b, find
P2.7, where each branch represents a the voltages and V^.
circuit element.
(b) Now determine the node voltages in
K l’ ^B’ respect
iUV
6^ 4A
to the reference node indicated by 30V 3A 10
the ground symbol.
(c) Compute and
(a)
(b)
Figure P2.9
AN SW ERS: ( b ) - 1 7 V 14 V
Figure P2.10
A N SW ERS: - 4 , - 9 , - 3 6 . 35, 10, 0 , 10, - 6 W
Chapter 2 • KirchhofF’s Current & Voltage Laws and Series-Parallel Resistive Circuits 93
11. Four circuit elements and a dependent (b) Supposing that /j = 4 A, /, = 2 A, and
voltage source are shown in the circuit o f Figure a = 0.25, determine and the power
P 2 .11. The current through and the voltage delivered to Rj^.
across each element are identified on the dia
gram. However, one— and only one— voltage
(or current) value is labeled incorrectly. Mark
the incorrect voltage (or current) on the circuit
diagram and give the correct value for this volt
Figure P2.14
age (or current).
A N SW ERS: (b) 80 Q, 20 W
Figure P2.15
SC RA M BLED A N SW ERS: 21, 16, 336, 56
R
>J .
10 k n ,
'+
6kO ■V,
Figure P 2.17
AN SW ER: (c) 12.5 watts and - 7 .5 watts (a)
18. Find the power absorbed by the unknown Figure P2.20
circuit element x and the voltage in the cir 6x 10"
cuit o f Figure P2.18. ANSWERS: (b) Vi = 3V-
'in
4 8 V 4 kQ i 6xl 0' -i-/?-i-6axl0
500 21. The circuit o f in Figure P2.21 is a blower
50 V motor control for a typical car heater. In this circuit,
6 0.8 A + resistors are used to control the current through a
<----
motor, thereby controlling the fan speed.
14V
44 V
6
'r 3R
© 10V
(b) W ith the switch in Med-1 position, delivered is 1250 W. How many possible medi
determine: um wattages are there and what are they?
(i) T h e voltage drop across each C H EC K : 10 ohms, 40 ohms
resistor.
(ii) The current delivered by the battery. 24. Consider the circuit o f Figure P2.24.
(iii) The relative efficiency of the circuit. (a) Suppose /? = 20 Q, find the power
(c) Repeat part (b) with the switch in delivered by the current source.
position Med-2. (b) Suppose the power delivered by the
(d) The switch is in the high position. A current source is 120 watts. Find the
winding in the motor shorts out. The value o f R.
fuse blows. W hat is the largest equiva
lent resistance o f the motor that will
cause the fuse to blow?
A N SW ERS:
(a) (i) Rj^g = 2.7 Q., =0.6 0., RcD “ (t) ^^^lOOV
= 0 .2 5 Q ,
1.25 a Figure P2.24
(ii) = 16.875 W Pbc = 3.75 W
1.5625 W P , , „ , , = 7.8125 C H E C K : (b) 8 </?< 15
W
(iii) 26% 25. Given that 4 W is absorbed by the 100-Q
(b) (i)K^5=0> ^s c = 3 -4 3 V , = resistor, find V} and the power delivered by the
1-43 = 7.14 V voltage source in the circuit o f Figure P2.25.
(ii) 5.71 A
(iii) 59.5% 150 Q 2o on
(c) (i) =0. Vcn = 2V,
Vmotor = 10 V 300n' lo o n
(ii) B A
(iii) 83.3%
Figure P2.25
(d) 0.4 Q
500
0.8 A
20V
6 C H EC K : (a) 45 < < 65; (b) -85 < 1^2 < '6 5 ;
200
(c) 8 0 < i? ^ < 125
Figure P2.27
SC RA M BLED A N SW ERS: 59.5, 9, 8, 49. 30. Consider the circuit o f Figure P2.30.
0.5, 45, 18 (a) Write an equation for in terms o f a
and 4
28. For the circuit o f Figure P2.28 with the (b) If 1/ = 40 V and a = 0.5, find the value
indicated currents and voltages, find o f the current
(a) Currents /j through (c) How much power is delivered by the
(b) Voltages Vj through independent source? How much
(c) Power delivered by each independent power is delivered by the dependent
current source source? Verify the principle o f conser
vation o f power for this circuit.
5 mA
500 0
2000
Figure P2.30
Figure P2.33
> ;i,
60kn lokn
Figure P2.39
Figure P2.37
Figure P2.38
i 2on >4on
1
CH ECK: = 40/,„ and
Chapter 2 • KirchhofF’s Current & Voltage Laws and Series-Parallel Resistive Circuits 99
Figure P2.43
A N SW ERS: 0.5R , 5 kQ, 2.6 kQ
O-
Figure P2.46
Figure 2.44 (a)
SC RA M BLED A N SW ERS: 60 kn ,
1.5R 22.5 kO, 135 k n
(a) > K ql
(b) 50. Consider the circuit o f Figure P2.50.
(a) Suppose = 320 V, = 256 V, R^ =
Kill = ^eq2
(d) There is no general relationship R^ = 800 n . Find R, and the
between R , and R^^j- relation resulting R^^.
ship depends on the value o f R. (b) Suppose = 320 V, V^= 192 V, R^ =
Explain your reasoning. 400 Q, R^ = 800 Q. Find R, and
the resulting R
C ircuit 1
Figure P2.50
SC RA M BLED AN SW ERS: 500, 1000, 1600,
C ircuit 2
400, 170.67, 128
1 .2 kn ■ 52Sn
O- o-
(c) (d)
Chapter 2 • K irchhoff s Current & Voltage Laws and Series-Parallel Resistive Circuits 101
Figure P2.52
N U M ERICA L AN SW ERS: 10.512, 14.177
Figure P2.54
53. Consider the circuit o f Figure 2.53a in
55. Consider the circuit o f Figure P2.55.
which = 0.5 £^. Suppose each AA-bulb rep
(a) Find max[ and the average
resents a 12-watt fluorescent bulb at approxi
value o f Zj(^).
mately 12 volts, having an internal resistance o f
(b) Find ijit), max[ i 2 {t)], and the average
12 a
value o f i2 {t).
102 Chapter 2 • K irch h off’s Current & Voltage Laws and Series-Parallel Resistive Circuits
o 3cost(2t)V
•2kn
is on, you turn the ignition to start the engine.
You noticed a momentary silence o f the radio.
12V The following circuit analysis explains this
effect quantitatively. Assume that with the car
Figure P2.55 engine not running, the 12-V car battery is rep
resented by the model shown in Figure P2.58.
56. Consider the circuits o f Figure P2.56. In The load due to the car radio is represented by
Figure 2.56, = 120 sin(377?) V and =5 an equivalent resistance o f 240 Q. The starter
(a) Find and the instantaneous motor draws 150 A o f current when the igni
ou f out'
power absorbed by 30 D, resistor. tion is turned on and before the engine starts.
(b) If is replaced by a current source, Find at the moment when the ignition
= 120 sin(377?) mA, pointing switch is turned on. Compare this to the volt
up, find and the instanta age before the ignition switch is turned on.
neous power absorbed by 30 Q, resis W hy do you think the radio goes silent
tor. Does affect the current through momentarily?
the other resistors in the circuit?
ignition
model for
car battery
with engine
not running
Chasis
ground
Figure P2.58
C H EC K : /*3oq = 37.97sin^(377r) watts
MATLAB PROBLEMS
(a) Find the output voltage, the output
current (what is its direction), and the
power absorbed by the load (8-Q resis
tor) for the circuit o f Figure P2.66.
Figure P2.63
C H EC K : = 10 kO.
© 8kO<
should insert
number or formula.
the proper
RL = 8:1:64;
Figure P2.65
% This command generates an array o f
numbers for RL beginning at 8 and
ending at 64 in increments o f 1. I f you
do not end it with a semicolon, it will
list every entry o f the array.
V2 = ??
% This value should be precomputed
I L = ??;
PL = RL .* I L . ^ 2 ;
Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits 105
R^= 180
---
R =90 R =20
20A
© R,= 150
R =80
R =40
Figure P2.67
106 Chapter 2 • K irchh off’s Current & Voltage Laws and Series-Parallel Resistive Circuits
200 mA ^3,2kn
(D
IkO 2.2kO 3kO 1.6kn
^ -
Figure P2.68
A N SW ERS: 591.2 Q, 8.869 V
300 300n
2on ion
eon
i3on
i5on
10V '4
6 4on 2oon > > V
135Q 500*
-O-
Figure P2.69
Figure P2.70
A N SW ER: 50.53 II, 133.8 mA
C H A P
For a network consisting o f resistors and independent voltage sources, one can apply KCL to the
nodes, KVL to the various loops, and Ohm’s law to the elements to construct a large set o f simul
taneous equations whose solution yields all currents and voltages in the circuit. In theory, this
approach completely solves the basic analysis problem. In practice, this approach proves imprac
tical because large numbers o f equations are required even for a small network. For example, a 6-
branch, 4-node network, with each node connected to the other nodes through a single element,
leads to a set o f 12 equations in 12 unknowns: 3 equations from KCL, 3 equations from KVL,
and 6 equations from the element v -i relationships. The 12 unknowns are the 6 branch currents
and 6 branch voltages.
Before the advent o f digital computers, engineers solved simultaneous equations manually, possi
bly with the aid o f a slide rule, or some primitive mechanical calculating machines. Any technique
or trick that reduced the number o f equations was highly treasured. In such an environment.
Maxwell’s mesh analysis technique (1881) received much acclaim and credit. Through the use o f
a fictitious circulating current, called a mesh current, Maxwell was able to greatly reduce the num
ber o f equations. For the above-mentioned network, the number o f equations drops from 12 to 3
equations in the unknown mesh currents.
An alternate KCL-based technique (now called nodal analysis) appeared in literature as early as
1901. The method did not gain momentum until the late 1940s, because most problems in the
early days o f electrical engineering could be solved efficiendy using mesh equations in conjunc
tion with some network theorems. W ith the invention o f multi-element vacuum tubes having
interelectrode capacitances, some compelling reasons to use the node method appeared; primari
ly, the node method accounts for the presence o f capacitances without introducing more equa
tions, and secondly, those vacuum tubes that behave very much like current sources are more eas
ily accommodated with nodal equations. By the late 1950s, almost all circuit texts presented both
the mesh and node methods.
Since the 1960s, many digital computer software programs (SPICE being the most ubiquitous)
have been developed for the simulation o f electronic circuits that otherwise would defy hand cal
108 Chapter 3 * Nodal and Loop Analyses
culation. These software packages use a node equation method over the mesh equation approach.
One o f several reasons is that a node is easily identifiable, whereas a set o f proper meshes is diffi
cult for a computer to recognize.
For resistive networks driven by current sources, writing node equations is straightforward.
Certain difficulties arise in writing node equations for circuits containing independent and
dependent voltage sources. During the 1970s, a modification o f the conventional node method
by a research group at IBM resulted in the “modified nodal analysis” (MNA) technique. W ith the
M N A method, the formulation o f network equations, even in the presence o f voltage sources and
all types o f dependent sources, becomes very systematic.
This chapter discusses the writing and solution o f equations to find pertinent voltages and cur
rents for linear resistive networks.
CHAPTER O U TLIN E
CH APTER O BJECTIVES
1. Describe and illustrate the method o f node analysis for the computation o f node voltages
in a circuit. Knowledge o f the node voltages o f a circuit allows one to compute all the
branch voltages and, thus, with knowledge o f the element values, all the branch currents.
2. Define the notion o f a mesh or loop current and describe and illustrate the method o f
mesh or, more generally, loop analysis for the computation o f loop currents in a circuit.
Knowledge o f all the loop currents o f a circuit allows one to compute all the branch cur
rents. Thus, in conjunction with the knowledge o f the branch element information, one
can compute all the branch voltages.
3. Formulate the node analysis and loop analysis equations as matrix equations and use
matrix methods in their solution emphasizing the use o f existing software for the gener
al solution.
4. Describe and illustrate the modified nodal approach to circuit analysis. This method
underlies the general software algorithms available for computer simulation o f circuits.
Chapter 3 • Nodal and Loop Analyses 109
Chapter 1 introduced basic circuit elements, Ohm’s law, and power calculations. Chapter 2 intro
duced the important laws o f circuit theory, K Y L and KCL, and investigated series, parallel, and
series-parallel circuits. Recall from Chapter 2 that a node voltage is the voltage drop from a given
node to a reference node. As a brief review, consider Figure 3.1, which portrays a circuit labeled
with nodes A through D having associated node voltages, V^, Vg, Vq V^, and eight branches,
one for the current source and one for each o f the seven conductances, ... , Gj. (Since this
chapter deals almost exclusively with dc, the uppercase notation for voltages and currents is com
monplace.)
FIG U RE 3.1. Diagram of a circuit with labeled node voltages, V^, Vg, Vq V^,
with respect to the given reference node.
KVL states that every branch voltage is the difference o f the node voltages present at the terminals
o f the branch: for circuits in this text and all pairs o f nodes, j and k, the voltage drop from
n o d ej to node k, is
at every instant o f time, where VJ- is the voltage at node j with respect to the reference and is
the voltage at node k with respect to reference. Here, j and k stand for arbitrary indices and could
be any o f the nodes. A, B, C, or D , in Figure 3.1. These statements mean that knowledge o f all
node voltages in conjunction with device information paints a rather complete picture o f the cir
cuit’s behavior. This chapter develops techniques for a systematic construction o f equations that
characterize a circuit’s behavior.
One last introductory point: Throughout this chapter and in many subsequent chapters, software
programs such as MATLAB facilitate calculations. Constructing sets o f equations that character
110 Chapter 3 • Nodal and Loop Analyses
ize the voltages and currents in a circuit is often a challenge. Solving such sets o f equations with
out the use o f software tools presents a much greater challenge. Yet facilitated by MATLAB or
equivalent, the calculations reduce to a hit o f the return key. MATLAB and the circuit simulation
program called PSpice or Spice (utilized in Chapter 4) are but two o f the many modern and
important software tools available to engineers.
Nodal analysis is an organized means for computing ALL node voltages o f a circuit. Nodal analy
sis builds around KCL, i.e., at each node o f the circuit, the sum o f the currents leaving (entering)
the node is zero. Each current in the sum enters or leaves a node through a branch. Each branch
current generally depends on the branch conductance, a subset o f the circuit node voltages, and
possibly source values. After substituting this branch information for each current in a node’s KCL
equation, one obtains a nodal equation.
As an example, the nodal equation at node A in Figure 3.1 is /^-^ = /j + /y = G j (V ^- V^) + Gy (Vj
- V^). The nodal equation at node C is -/2 + I^ + 1^^- Ij = Vj- + G j ( V ^ -
+ Gy ( K ( j- V^) = 0. Writing such an equation at each circuit node (except the reference node) pro
duces a set o f independent equations. O f course, one can substitute a KCL equation at the refer
ence node for any o f the other equations and still obtain an independent set o f nodal equations.
T he solution o f such a set o f nodal equations yields all circuit node voltages. Knowing all node volt
ages permits us to compute all branch voltages. Knowing each branch voltage and each branch con
ductance allows us to compute each branch current using Ohm’s law. The reference node may be
chosen arbitrarily and can sometimes be chosen to greatly simplify the analysis.
A set o f nodal equations has a matrix representation. The matrix representation permits easy solu
tion for the node voltages using MATLAB or an equivalent software package. A variation o f the
nodal analysis method, termed modified nodal analysis, relies heavily on matrix methods for
constructing and solving the circuit equations. The basic principles o f this widespread analysis
technique are illustrated in Section 4.
Because computer-based circuit analysis packages build on a matrix formulation o f the circuit
equations and because o f the widespread use o f matrices in circuits, systems, and control, we will
stress a matrix formulation o f equations throughout this chapter. The student unfamiliar with
matrix methods might look through a calculus text or a linear algebra text for a good explanation
o f their basic properties and uses.
T he counterpart to nodal analysis is loop analysis. In loop analysis, the counterpart o f a node
voltage is a loop current, which circulates around a closed path in a circuit. A loop or closed path
in a circuit is a contiguous sequence o f branches that begins and ends on the same node and touch
es no other node more than once. For each loop in the circuit, one defines a loop current, as illus
trated in Figure 3.2, that depicts three loops or closed paths having corresponding loop currents
/p Ij, and ly O f course, one can draw other closed paths or loops for this circuit and define other
loop currents.
Chapter 3 • Nodal and Loop Analyses 111
90
FIG U RE 3.2. Simple resistive circuit showing three closed paths (dotted lines) that represent
three loop currents, /j, Ij, and 1^; the branch current - which is a difference
o f the two loop currents through the resistor.
Using a fluid flow analogy, one can think o f loop currents as fluid circulating through closed sec
tions o f pipe. The fluid in different closed paths may share a segment o f pipe. This segment is anal
ogous to a branch o f a circuit on which two or more loop currents are incident. The net current
in the branch is analogous to the net fluid flow. Note that each branch current can be expressed
as a sum o f loop currents with due regard to direction. For example, in Figure 3.2, the branch cur
rent 7^3 = ^\ - Using loop currents, element resistance values, and source values, it is possible
by KVL and Ohm’s law to express the sum o f the voltages around each loop in terms o f the loop
currents. For example, the first loop, labeled in Figure 3.2, has the loop equation
^ « = 9 / i + 3(/i -/ 2 )+ 6 (/ ^ -/ 3 )
We will explore this concept more thoroughly in Section 5. Here we see that loop analysis builds
on KVL, whereas node analysis builds on KCL.
As mentioned earlier, nodal analysis is a technique for finding all node voltages in a circuit. W ith
knowledge o f all the node voltages and all the element values, one can compute all branch volt
ages and currents, and thus the power absorbed or delivered by each branch. This section describes
nodal analysis for circuits containing dependent and independent current sources, resistances, and
independent voltage sources that are grounded to the reference node (see Figure 3.3). Floating
independent or dependent voltage sources (those not directly connected to the reference node) are
covered in Section 4.
For the class o f circuits discussed in this section, it is possible to write a nodal (KCL) equation at
each node not connected to a voltage source. A node connected to a voltage source grounded to
the reference node has a node voltage equal to the source voltage. The other node voltages must
112 Chapter 3 • Nodal and Loop Analyses
be computed from the set o f nodal equations. Each nodal equation will sum the currents leaving
a node. Each current in the sum will be expressed in terms o f dependent or independent current
sources or branch conductances and node voltages. The set o f these equations will have a solution
that yields all the pertinent node voltages o f the circuit. Examples 3.1 and 3.2 illustrate the basic
techniques o f nodal analysis.
EX A M P L E 3 .1 .
The circuit o f Figure 3.3a contains an independent voltage source, an independent current source,
and five resistances whose conductances in S are G j through Gy The nodes other than the refer
ence are labeled with the node voltages V^, and V^, which respectively denote nodes a, b, and
c. T he analysis o f this circuit illustrates the process o f nodal analysis to find the node voltages V^,
y,,andK =
FIG U RE 3.3A. Resistive circuit for Example 3.1. Note that node voltage
is specified by the voltage source.
So l u t io n .
Step 1. Consider node c. A voltage source ties node c to the reference node. Hence, the node volt
age is fixed at V-^, i.e., Because it is not necessary to apply KCL to this node
unless the current through the voltage source is required, for example, when determining the
power delivered by the source.
Step 2. Sum the currents leaving node a. From KCL, the sum o f the currents leaving node a is
zero. As per the partial circuit in Figure 3.3b, this requires that
Grouping the coefficients o f and and moving the source values to the right side o f the equa
tion yields our first nodal equation
Chapter 3 • Nodal and Loop Analyses 113
(3.1)
Step 3. Sum the currents leaving node b. Applying KCL to node b, reproduced in Figure 3.3c,
yields the equation
G 2 ^y b- y a) ^G ,V b^ G, {V ,- VJ = Q
G4(v ,- v j
FIG U RE 3.3C
Step 4 . Write set o f nodal equations in matrix form. Equations 3.1 and 3.2 in matrix form are
Gi + G 2+ G5 -G 2 •
(3.3)
-G 2 G 2 -I- G 3 -I- G 4
Matrix equations organize relevant data into a unified framework. Because many calculators do
matrix arithmetic, because o f the widespread availability o f matrix software packages such as
MATLAB, and because equation solution techniques in circuits, systems, and control heavily uti
lize matrix methods, the matrix equation formulation has widespread and critical importance.
Step 5 . Solve the matrix equation 3.3: For this part, suppose that the conductance values in S are
Gj = 0.2, G2 = 0.2, G3 = 0.3, G 4 = 0.1, G 5 = 0.4, that = 2.8 A, and that = 24 V. After sub
stitution, equation 3.3 simpUfies to
114 Chapter 3 • Nodal and Loop Analyses
Solving using the inverse matrix method leads to the node voltages (in volts):
-1
■ 0.8 - 0 .2 ' T 2 .4 - 1 0 .6 0 .2 ' ■12.4' 18-
V
- 0 .2 0.6 2.4 0.4 4 0.2 0.8 2.4 10
Alternately, one could have solved equation 3.4 via MATLAB, its equivalent, or the age-old hand
method o f adding and subtracting equations. For example, in MATLAB
»M =[0.8 -0.2;-0.2 0.6];
»b= [12.4 2.4]';
>>NodeV = M\b
NodeV =
1.8000e+01
l.OOOOe+01
»% O R EQU IVA LEN TLY
»NodeV = inv(M )*b
NodeV =
18
10
Exercises. 1. Utilize the solution o f Example 3.1 to compute the current leaving and the power
delivered by the independent voltage source.
AN SW ER: 3.8 A and 91.2 watts
2. Referring to Figure 3.3a and the values set forth in Step 5 o f Example 3.1, suppose the value o f
is cut in half, the value o f V-^ is 24 V, and the value o f each o f the conductances is also cut in
half W hat are the new values o f the node voltages?
AN SW ER: All node voltages are the same.
EX A M PLE 3.2.
Consider the circuit o f Figure 3.5a. Similar to Example 3.1, the objective is to find the node volt
ages V^,Vf^, and . However, in the circuit o f Figure 3.5a, an independent current source has
replaced the independent voltage source o f Figure 3.3a. This change unfreezes the constraint on
the value o f present in the circuit Figure 3.3a. There will result three nodal equations in the
three unknowns , and V^.
0.4 U
So l u t io n .
Step 1. Sum currents leaving node a. This step is the same as Step 2 o f Example 3.1. By inspec
tion o f node a,
0 . 2 + 0.2(K^ - V^) + 0.4(V^ - V J - 2 = 0
Step 2. Sum currents leaving node b. This step is the same as Step 3 o f Example 3.1. Again, by
inspection,
0.2 - VJ + 0.3 + 0.1 (V^ - V;) = 0
Simplification yields
0 . 4 ( 1 / - K J + 0.1 ( K ^ - K ^ ) - 1 = 0
Step 4. Write equations 3 .5 -3 .7 as a matrix equation and solve. The matrix form o f our nodal
equations 3 .5 -3 .7 is
Solving equation 3.8 using MATLAB or equivalent, using a calculator that does matrix operations,
or solving via some form o f row reduction, one obtains the solution (in volts)
-1
■0.8 - 0.2 - 0 .4 ' ■2 ‘ '2 .9 1.4 2.6 ■2 ‘ ■8.4'
= - 0.2 0.6 - 0.1 0 = 1.4 2.4 1.6 0 = 4 .4 V (3.8b)
Specifically, in MATLAB
» M = [0.8 -0.2 -0.4;-0.2 0.6 -0.1;
-0.4 - 0.1 0.5];
>>b = [2 0 1]';
»NodeV = M\b
NodeV =
8.4000e+00
4.4000e+00
9.6000e+00
Exercises. 1. Suppose the values o f the current sources in Figure 3.5a are doubled. W hat are the
new values o f the node voltages? Hint: Consider the effect on equation 3.8.
AN SW ER: All node voltages are doubled.
2 . Suppose the conductances in the circuit o f Figure 3.5a are cut in half, i.e., the resistances are
doubled. W hat are the new node voltages?
A N SW ER: Node voltages are doubled.
3. Suppose the conductances in the circuit o f Figure 3.5a are cut in half W hat happens to the
magnitudes o f the branch currents? Hint: Express the branch current in terms o f the branch con
ductance and its terminal node voltages.
AN SW ER: The magnitudes o f the branch currents
stay the same.
I
4. Find two node equations characterizing the cir
cuit o f Figure 3.6.
FIG U RE 3.6
AN SW ER: (G j + G^) -V„~G^^Vy = and
Chapter 3 • Nodal and Loop Analyses 117
The matrices in equations 3.3, 3.4, and 3.8a are symmetric. A symmetric matrix, say A, is one
whose transpose equals itself, i.e., = A; this means that ifA = [a-^ is an n x n matrix whose i-j
entry is a-, then A is symmetric if a-j = a^j^. In words, the off-diagonal entries are mirror images o f
each other. For example.
When only resistances, independent current sources, and grounded independent voltage sources are
present in the circuit, the value o f the entries in the coefficient matrix o f the nodal equations can be
computed by inspection. The 1-1 entry o f the matrix is the sum o f the conductances at node a (or
1); the 2 - 2 entry is the sum o f the conductances at node b (or 2). In general, the i-i entry o f the
coefficient matrix is the sum o f the conductances incident at node i. Further, the 1 -2 entry o f the
matrix is the negative o f the sum o f the conductances between nodes a and b (or between nodes 1
and 2), and the 2 -1 entry has the same value. In Example 3.2, the 1 -2 entry o f - 0 .2 S is the nega
tive o f the sum o f the conductances between nodes a and b; the 1 -3 entry o f - 0 .4 S is the negative
o f the sum o f the conductances between nodes a and c (or between 1 and 3, if the nodes were so
numbered). Thus, whenever the circuit contains no dependent sources, the node equations can be
written by inspection. Further, if independent voltage sources are absent, then the right-hand side o f
the nodal matrix equation can also be written by inspection: the i-th entry is simply the sum o f the
independent source currents injected into the i—th node at which KCL is applied.
W hen controlled sources are present in the circuit, the resultant nodal matrix is generally not sym
metric, as illustrated in the following two examples.
The amplifier circuit model contains a current-controlled current source (CC CS) and a voltage-
controlled current source (VCCS). These two dependent sources have currents that depend on
other circuit parameters and require some special handling when constructing node equations.
Our objective is to set forth the methodology for writing the node equations when dependent cur
rent sources are part o f the circuit and to compute the magnitude o f the voltage gain, | V-^ |
= \y 2 |y^n\■
Note that the source voltage, V-^, specifies the voltage at the node at the bottom o f Gp hence, a
nodal equation at this node is unnecessary. Nodal equations must be written at the remaining
118 Chapter 3 • Nodal and Loop Analyses
nodes, which are labeled with the voltages Vj, Kj (= and V^. (Numbering and labeling is
often a matter o f personal preference. In this example, we have chosen 1, 2, and 3 as node labels,
in contrast to the previous two examples, where we used a, b, and c.)
G,
So l u t io n .
Step 1. Sum the currents leaving node 1. Summing the current leaving node 1 leads to
Step 2. Substitute fo r i^ in equation 3 .9 and simplify. In equation 3.9, |3 i^ accounts for the effect
o f the C C C S at node 1 and is not given in terms o f the circuit node voltages. To specify this term
in terms o f the circuit node voltages, observe that in Figure 3.7, i^ is the current from node 2 to
node 3 through G^. Hence,
Substituting equation 3.10 into 3.9, again grouping like terms, one obtains the first nodal equation,
- G 3 K1 + (G 3 + Gg 4. G,)V^ - G4 K3 + = 0 ( 3 . 12)
Chapter 3 * Nodal and Loop Analyses 119
Substituting o f equation 3.13 into equation 3.12 leads to our second nodal equation,
Step 5. Sum the currents leaving node 3. Applying KCL to node 3 yields,
G 2 ( V ,- V , ) ^ G , i V , - V , ) ^ G , V , - p i ^ - g ^ v ^ = 0 (3.15a)
0 = G , { V , - V ,) + G , { V , - V ,) + G ^ V ^ - ^ G , { V , - V , ) - g ^ { V , - V , )
Grouping like terms leads to our third equation in the three unknowns V^, V^, and V^:
Step 6. Put nodal equations in matrix form. The three nodal equations 3.11, 3.14, and 3.15b
have the matrix form
Step 7. Substitute values and solve. Suppose that the various circuit conductances have the fol
lowing values in [xS: G j = 1,000, G j = 2.0, G j = 1.0, G^ = 10, G j = 2 0 ,100, and Gg = 200 .
Suppose further that = 2.1 V, (3 = 4/1010 and^^ = 21,112 [xS. This allows us to generate the
following M ATLAB code for the solution:
in which case,
■V,- ■ 2 ■
V^2 = - 1 0 0 V
V3 1
Step 8. Compute the voltage gain. The voltage gain o f the amplifier is given by
Kut V2 -1 0 0
= 4 7 .6 2
Vin Vin 2.1
Exercises. 1. Suppose V-^ in the circuit o f Figure 3 .7 is doubled. W hat are the new node voltages?
Hint: Consider the matrix equation o f Step 6.
A N SW ER: Node voltages are doubled.
2. Suppose all conductances in the circuit o f Figure 3.7 are cut in half (resistances are doubled)
and (3 is held constant. How must^^ change for the node voltages to remain at their same values?
AN SW ER: must double.
Realistic problems do not permit hand solutions. For hand solutions, the smallest number o f
equations is generally desired. For matrix solutions using software packages such as MATLAB,
more variables with more equations may often be easier to construct and may often result in more
reliable numerical calculations. This can be illustrated using the equations o f Example 3.3. All the
pertinent basic equations o f the circuit o f Figure 3 .7 can be written down as follows: from equa
tions 3.9 and 3.10 we have
and
However, in contrast to the example, we do not substitute 3.10 into 3.9 to obtain 3.11. Rather,
we just let them be two independent equations. Further, from equations 3.12, 3.13, and 3.15a,
we have
- G 3 K, + (G 3 + G4 + G ,) K2 - G 4 K3 + g^v^ = 0
and
By not substituting for and v^, we avoid unnecessary hand calculation, and if there is an error,
it is easier to find. The resulting equations have the matrix form where i and now appear as
additional unknowns, easily handled by a software program:
Chapter 3 • Nodal and Loop Analyses 121
G] 4- G 2 + G 3 -G 3 -G 2 13 0 1 •V,-
0 G4 -G 4 -I 0 ^2 0
-G 3 G 3 -1- G 4 -HGg -G 4 0 8m = 0
1 0 -1 0 -1 iy 0
-G 2 -G 4 G2 + G 4 + G5 -/? ~Sm 0
As a general rule, we would reorder the equations so that rows 1, 3, and 5 came first, as they cor
respond to the three nodal equations at Kj, V2 , and . Then we would write the constraint
equations for i and . Such a reordering leads to certain symmetry properties discussed earlier.
Exercise. Solve the above matrix equation in MATLAB or equivalent, using the numbers o f
A and „ - 1 V.
Example 3.3 to verify that = -1 .0 1 x 10^-3■ A
Matrix methods as used in the above examples and in the ones to follow necessitate the power o f
a calculator or a software program such as MATLAB for easy solution. Such programs permit a
straightforward calculation o f the required answers and are not prone to arithmetic errors.
The next example illustrates how to write node equations for circuits containing a voltage-con
trolled voltage source (VCVS) grounded to the reference node. The analysis o f CCVSs grounded
to the reference node is similar. The more challenging analysis o f circuits containing floating
dependent or independent voltage sources is taken up in the next section.
EX A M P L E 3 .4 .
The circuit o f Figure 3.8 models a poor operational amplifier circuit' in which the output voltage
1/^^ = V2 approximates For the analysis, let |i = 70. The adjective “poor” arises because ^
should have a value much larger than 70.
R3= lO k O
V. =
So l u t io n .
The circuit contains two nodes labeled and (equivalently nodes 1 and 2) not constrained
by voltage sources. The goal o f our analysis is to find these node voltages by writing two equations
in these voltages and solving. As is commonly the case, resistances are in ohms and will be con
verted to conductances in S for convenience in writing the node equations.
/~N,
Step 1. Compute conductance values in S. Conductances are the reciprocal o f resistances, i.e., G-
= ^IRj- Hence,
Step 2 . Write a node equation at node 1. Summing the currents leaving node 1 yields
(3.16)
Step 3. Sum currents leaving node 2 . Summing the currents leaving node 2 yields
The dimensionless coefficient |i is placed with the conductance, while grouping like terms to
obtain
(HG4 - + (G 3 + G4+ ^ 5)^2 = 0
Step 4. Write equations 3 .1 6 and 3 .1 7 in matrix form and solve. In matrix form
■3.3333 -1 ■ -V f ■2'
0 .6 9 9 9 O .O Ill .^2. 0
Using the formula for the inverse o f a 2 x 2 matrix (interchange the diagonal entries, change the
sign on the o ff diagonals, and divide by the determinant), one obtains
•Vf 1 0.0111 1 0 .0 3 0 1 2 6
Exercises. 1. Write MATLAB code to solve the above example. Check that your code works. Hint:
See Example 3.3.
2. If R2 is changed to 100 k£2 in Example 3.4, show that V2 = -1 .9 0 6 3 V.
Chapter 3 • Nodal and Loop Analyses 123
A floating voltage source means that neither node o f the source is connected to the reference
node. When a floating dependent or independent voltage source is present with respect to a given
reference node, a direct application o f KCL to either terminal node o f the voltage source is unfruit
ful. There are several ways to handle this situation. One fruitful method is to enclose the source
and its terminal nodes by a Gaussian surface, i.e., a closed curve, to create what is commonly
called a supernode, as illustrated in Figure 3.9. One would then write KCL for the supernode as
is done in a number o f circuit texts. However, there is a conceptually more straightforward
approach, which is often called the modified nodal analysis, or MNA. In MNA, we add an addi
tional current label to each floating voltage source. In Figure 3.9, we have added the current label
l^y. This additional current becomes an unknown in a set o f nodal equations generated by apply
ing KCL to each node. At this point, further explanation is best done by an example, but the con
cept is similar to the discussion following Example 3.3.
E X A M P L E 3.5.
Find the node voltages V^, Vy, and the unknown current in the circuit o f Figure 3.9, when
the bottom node is taken as reference.
FIG U RE 3.9. Resistive circuit containing a floating voltage source for the given reference;
generally, the reference node may be chosen arbitrarily.
So l u t io n .
Step 1. Write a node equation at node a. Summing the currents leaving node a yields
8 + 0.15 + 3 + 0.2 ( K ^ -K .) = 0
(0.15 + 0 .2 )K ^ - 0.15^ -^ - 0 .2 V ;+ 8 + 3 = 0
or, equivalently,
- 3 + 0 .1 5 ( K ^ - K ; + 0.05 =0
or equivalently,
- 0 .1 5 (0 .1 5 + 0.05) n - / ^ ^ = 3
- 0 .1 5 K , + 0.2K^-/^^ = 3 (3.19)
Step 4. Write the node voltage relationship fo r the terminal nodes o f the floating voltage source,
i.e., between the voltages and V^. The voltages and are constrained by the voltage source.
Mathematically, this constraint is 1/ - = 440, i.e.,
K ^ = 440 (3.21)
Step 5. Write thefour equations 3.18, 3.19, 3.20, and 3.21 in matrixform and solve. In matrix form,
0.35 - 0 .1 5 - 0 .2 0 -1 1
- 0 .1 5 0.2 0 -1 3 (3.22)
- 0 .2 0 0.45 1 25
0 -1 1 0 hb 440
Because o f the extra variable, the equations become too large for hand calculation. Hence, we use
MATLAB as follows:
>>M = [0.35 - 0 .1 5 - 0 .2 0 ;-0 .1 5 0.2 0 - 1 ;
- 0 .2 0 0.45 1 ; 0 - 1 1 0];
»b = [-11 3 25 440]';
»x = M\b
X =
-9.0000e+ 01
-3 .1 0 0 0 e + 0 2
1.3000e+02
-5.1500 e+ 01
Hence,
= - 90 V, = - 3 1 0 V, = 130 V, = -5 1 .5 A
Chapter 3 “ Nodal and Loop Analyses 12 5
In a conventional nodal analysis, all unknowns are node voltages. Here we have the additional
unknown current, . Because o f this additional unknown current, the method is called a mod
ified nodal analysis.
Also, in this example, node d was taken as the reference node. However, one could just as easily
take node b as the reference node, in which case, the voltage source would not have been floating.
A home problem investigates this choice o f reference node.
The next example investigates a circuit having floating independent and dependent voltage
sources. By convention, the reference node o f this circuit, figure 3.10, and all subsequent circuits,
will be the bottom node o f the circuit unless stated otherwise.
E X A M P L E 3.6.
The circuit o f Figure 3.10 contains a floating independent and a floating dependent voltage
source. Find the node voltages V^, Vy, V^, and the unknown currents and Then find the
power delivered by the 30 V source and the dependent source.
500 1
FIG U RE 3.10. Resistive circuit containing a floating dependent voltage source and a floating
independent voltage when node d is chosen as the reference node.
So l u t io n .
Step 1. Sum currents leaving node a. Here,
126 Chapter 3 • Nodal and Loop Analyses
Equivalently,
(3.23)
100 “ 100
100 500
Equivalently,
(3.24)
Step 4. Write an equation relating the terminal voltages o f the independent voltage source. Here,
K ,-n = 3 0 (3.26)
Step 5. Write an equation relating the terminal voltages o f the dependent voltage source. Here,
40
K ,- K = 4 0 ., = —
Equivalently,
0.6V;, + 0 .4 y ^ - V ^ = 0 (3.27)
Step 6. Write equations 3.23 through 3 .2 7 in matrix form an d solve in MATLAB. Combining
the above equations into a matrix produces
0 .03 -0 .0 1 0 1 0 ■2.2'
- 0.01 0.012 0 0 -1 V'i 0
0 0 0.0 0 1 2 5 -1 1 = 0
________
0 -1 0 0 ^ac 30
0 .6 0.4 -1 0 0 Icb, 0
Again, this matrix equation is too large for hand computation. Hence in MATLAB,
»M = [0.03 - 0 .0 1 0 1 0;
- 0.01 0 .012 0 0 - 1;
0 0 0.00125 - 1 1;
0 - 1 1 0 0;
0.6 0 . 4 - 1 0 0];
Chapter 3 • Nodal and Loop Analyses 127
l.OOOOe+02
5.0000e+01
8.0000e+01
-3 .0 0 0 0 e -0 1
-4 .0 0 0 0 e -0 1
Hence,
100 ■
50
= 80
^ac - 0 .3
Icb_ - 0 .4
Step 4. Compute the power delivered by the 3 0 Vsource. The power delivered by the 30-V source is
^^./ = - 3 0 / ,^ = 3 0 x 0.4 = 1 2 W
Step 5. Compute the power delivered by the dependent source. The power delivered by the depend
ent source is
Pdel = = - 4 0 X
^ ( - 0 . 3 ) = 0.12(V^ - V^) = 6 W
100
Exercises. 1. For Example 3.6, compute the voltages and the power absorbed by the 800
Q resistor.
A N SW ERS in random order: 8 watts, 20 V, - 8 0 V
2. Suppose the two independent voltage source values in Example 3.6 are doubled. W hat are the
new node voltages? W hat are the new branch currents?
A N SW ERS: Node voltages are doubled and branch currents are doubled.
3. Suppose all resistances in the circuit o f Figure 3.10 are doubled and the value o f the parameter
on the dependent source is also doubled. W hat are the new branch currents?
AN SW ER: All branch currents are cut in half.
The above example increases the number o f unknowns beyond the node voltages to include the
two currents through the floating voltage sources. However, we could have included additional
currents to the set o f equations making the dimension even higher. W ith a tool like MATLAB,
this poses no difficulty. However, it does make hand computation a challenge. For example, we
128 Chapter 3 • Nodal and Loop Analyses
could have included = l^y as an additional variable with a corresponding increase in the num
ber o f equations. By adding addirional unknowns we would simplify the writing o f the individual
node equations but increase the dimension o f the matrix equation. Specifically, the node equation
at “a” becomes
_ t i I IV
50 50
0.02 0 0 1 0 2.2
0 0.002 0 0 -1 0
0 0 0.00 1 2 5 -1 1 0 0
0 -1 1 0 0 0 ^ac 30
0 -1 0 0 -4 0 Ic h 0
-1 0 0 0 -100 0
This completes our discussion o f the standard nodal equation method o f circuit analysis. T he next
section takes up a discussion o f an alternative analysis method entided loop analysis.
5. LOOP ANALYSIS
Loop analysis is a second general analysis technique for computing the voltages and currents in a
circuit. Mesh analysis is a special type o f loop analysis for planar circuits, i.e., circuits that can
be drawn on a plane without branch crossings. For planar circuits, loops can be chosen as mesh
es, as illustrated in Figure 3.2, or as in 3.11 below. Associated with each loop is a loop current.
Loop currents circulate around closed paths (loops) in the circuit. Similarly, for planar circuits,
the term mesh current is used traditionally for loop current. By KVL, the sum o f the voltages
across each branch in a loop is zero. By expressing each o f these branch voltages in terms o f the
designated loop currents, one can write an equation in the loop currents for each designated loop
in the circuit. For branches that are often common to two or more designated loops, the branch
current equals the net flow o f the loop currents incident on the branch. Writing an equation for
each loop produces a set o f equations called loop equations. If sufficient independent loops are
defined, one can solve the loop equations for the loop currents. Once the loop currents are known,
we can easily compute the branch currents and then the branch voltages in the circuit. Then we
can compute any other quantities o f interest, such as power absorbed, power delivered, voltage
gain, etc.
So l u t io n .
Step 1. Write a KVL equation based on loop 1 by summing voltages around this loop. Summing
the voltages around loop 1 using Ohm’s law and the defined loop currents produces
Here, observe that the 4 Q resistor is incident on two loops; the net current flowing from top to
bottom, i.e., with respect to the direction o f loop 1, is /j - l 2 - The idea is analogous to a pair o f
distinct water pipes that share a common length. The common length is analogous to the 4 -Q
resistor. The flow rate in each pipe is analogous to the currents /] and I 2 , which in fact, are rates
at which charge flows past a cross sectional area o f the conductor. It follows that the net flow
through the common length o f pipe with respect to the direction o f loop 1 is the difference in the
net flow rates o f pipes 1 and 2, respectively. This is precisely the meaning o f /j - /2 . A similar
explanation can be made for the 1 -Q resistor common to loops 1 and 3 for which the net flow
rate with respect to the direction o f loop 1 is 7j - ly
Step 2. Write a KVL equation based on loop 2 by summing the voltages around this loop. Applying
Ohm’s law and KVL to loop 2 produces
Notice that with respect to the direction of loop 2, the net flow rate through the 4 Q resistor is h'
Step 3. Finally, write a KVL equation based on bop 3. Stmiming the voltages around loop 3 yields
Step 4 . Write eqtiations 3.25b, 3.26, an d 3 .2 7 in matrix form an d solve. Writing the above three
loop equations in matrix form yields
'6 -4 -r W ■20'
-4 8 -2 h - 0 (3.28)
-1 -2 4 h 20
Solving Equation 3.28 by the matrix inverse method (by a numerical algorithm or by Cramer’s
rule) yields the loop currents in amps as
-1
h' 6 -4 -r '2 0 ' 0 .35 0 .225 0 .2 ■20' 11 ■
h = -4 8 -2 0 = 0 .2 2 5 0 .2 8 7 5 0 .2 0 = 8.5
h -1 -2 4 20 0 .2 0 .2 0 .4 20 12
Step 5. Compute the power consumed by the 2 Q resistor. Knowledge o f the loop currents makes
it possible to compute all voltages and currents in the circuit. For our purpose, the voltage
3. Suppose the source values are doubled. W hat are the new values o f the loop currents?
AN SW ER: loop currents are doubled
4. Suppose the resistance values are multiplied by 4. W hat are the new loop currents? W hat are
the new node voltages?
AN SW ERS: Loop currents are 0.25 times their original values, and node voltages are unchanged.
Observe that there are no dependent current or voltage sources in the circuit. Similar to the nodal
analysis case, whenever dependent sources are absent and the equations are written in the natural
order, the loop (or mesh) equations are symmetric, as illustrated by the coefficient matrix o f equa
tion 3.28 where, for example, the 1 -2 and 2 -1 entries coincide, as do the 1 -3 and 3 -1 entries,
etc. Also, the value o f all entries can be computed by inspection. The 1-1 entry o f the matrix is
the sum o f the resistances in loop 1; the 2 - 2 entry is the sum o f the resistances in loop 2, etc. In
general, the i- i entry is the sum o f the resistances in loop i. T he 1 -2 entry o f the matrix is 'L{±R^
(the large sigma means summation), where each is a resistance common to both loops 1 and 2.
Use the + sign when both loop currents circulate through in the same direction, and use the -
Chapter 3 • Nodal and Loop Analyses 131
sign otherwise. Further, if independent current sources are absent, then the right-hand side o f the
loop equations can also be written by inspection. The i-th entry is simply the net voltage o f the
sources in the i—th loop that tends to deliver a current in the direction o f the loop current.
Exercises. 1. Use the inspection rules described above to write two mesh equations for the circuit
o f Figure 3.12, when both mesh currents are assigned clockwise direction.
FIG U RE 3.12.
2. Use the inspection rules described above to write two mesh equations for the circuit o f Figure
3.12, when the left mesh current is clockwise and the right mesh current is counterclockwise.
3. Use the inspection rules described above to determine the right-hand side o f the mesh equation
for the circuit o f Figure 3.13.
AN SW ERS: 8, 0, 10
A simplifying reduction to the set o f loop equations occurs if an independent current source coin
cides with a single loop current. The analysis becomes simpler because that loop current is no
longer an unknown; rather it is equal to the value o f the source current if their directions coin
cide, or to the negative value if their directions are opposing. Because the associated loop current
is known, there are fewer loop equations to write and solve. One would apply KVL to such a loop
only if it were necessary to compute the voltage across the independent current source, which
might be necessary for determining the power delivered by the source. Th e entire situation is anal
ogous to an independent voltage source tied between a node and the reference in nodal analysis.
The following example illustrates the details o f this discussion.
132 Chapter 3 • Nodal and Loop Analyses
FIG U RE 3.14. A resistive circuit containing an independent current source on the perimeter
o f loop 3 forcing /^ = 8 A.
So l u t io n .
Step 1. So/ve fo r by inspection. Because is the only loop current circulating through the
branch containing the independent 8 A current source, /j = 8 A. This phenomena is similar to the
fact that in nodal analysis, the node voltage o f a grounded voltage source is fixed at the voltage
source value.
Step 2. Write a KVL equation fo r loop 1 by summing voltages around this loop. Summing the
voltages around loop 1 using Ohm’s law and the designated loop currents produces
28 = / j + 4 ( /j - /2) + 12 + ( /j - 8) = 6 /, - 4 /2 + 4
Hence,
6 /, - 4 /2 = 24 (3.29)
Step 3. Write a KVL equation fo r loop 2 by summing the voltages around this loop. Applying KVL
and Ohm’s law to loop 2 produces
Equivalently,
- 4 /1 - 8 /2 = 16 (3.30)
Chapter 3 • Nodal and Loop Analyses 133
Step 4 . Write above loop equations in matrix form and solve. The matrix form o f equations 3.29
and 3.30 is
-1
II" ■6 -4 24 ' 1 '8 ' 24'
4 's'
I2 -4 8 16 “ 32 4 6 16 6
Step 6. Compute power delivered by 8 A source. Observe that the 8 A current source is labeled
according to the passive sign convention, in which case,
Pdel = - = - (8 8) = - 64 watts
Exercise. In the circuit o f Figure 3.15, nvo o f the three mesh currents coincide with independent
source currents. By writing and solving just one mesh equation, find /j.
A N SW ER: 3 A
Not only do independent current sources constrain loop currents, but dependent currents sources
do also. This situation is illustrated in Example 3.9.
E X A M PLE 3 .9 .
This example illustrates the writing o f loop equations for a simplified small signal equivalent cir
cuit, Figure 3.16, o f a two-stage amplifier that contains a current-controlled current source
(C C C S) and a current-controlled voltage source (CC VS). This process extends the techniques o f
Examples 3.7 and 3.8 to find some important characteristics o f the amplifier. Specifically, find
(a) the input resistance seen by the source, i.e., = v-Jij^,
134 Chapter 3 * Nodal and Loop Analyses
+ V
FIG U RE 3.16. Small signal equivalent circuit for a two-stage amplifier. Signals in amplifiers are
usually time dependent, so we adopt the lowercase notation for voltages and currents.
So l u t io n .
The circuit o f Figure 3.16 contains three loop or mesh currents. The direction o f the loops is a
user-chosen preference. For convenience, we have chosen mesh current z'2 to be consistent with
the direction o f the arrow in the dependent current source. Because this dependent current source
lies on the perimeter o f the circuit, it constrains the value o f /'2> i-e-) ?2 ^ P ^b- the control
ling current, , z'2 = . ih ~ ^ h' relationship implies that the mesh current o f loop 2
depends directly on the mesh current o f loop 1. This observation allows us to skip constructing a
mesh equation for loop 2. Only equations for loops 1 and 3 are needed, thereby reducing the
number o f simultaneous equations from three (because there are three loops) to two.
Step 1. Apply KVL to loop/mesh 1. Here, by KVL and the observation that 12 = P zp
Step 2. Apply KVL to loop/mesh 3. In this case, observe that z^ = — (z^ + z'g) = - (|3 z'j + Zj). By BCVL,
K + ph + '3 = 0 (3.32)
Chapter 3 * Nodal and Loop Analyses 135
Step 3. Write equations 3.31 and 3.32 in matrix form an d solve. The matrix form o f these equa
tions IS
/?^ + (P + l ) ^ , 0 h ^in
0 (3.33)
'3
V. )
Because the solution is desired in terms o f the hteral variables, we solve equation 3.33 using
Cramer’s rule, which utilizes determinants. In this task, first define
Using the notation A for the determinant, Cramers rule provides the solution for i^ according to
the formula
^in 0
det
0 R^ + R ^+r„ (3.34)
‘I =
Step 4. Compute in terms o f v-^ and then the voltage gain vjv^^. As per the circuit o f Figure
3.11 and equations 3.34 and 3.35,
o
P/?
I— d
---- r= y —r
V.
in
A ^
Step 4 . Compute v. To compute the voltage across the dependent current source, apply KVL to
mesh 2 to obtain
n = R^ (z'2 + /'i) + R, («2 + ^3) = [P h + h
136
Chapter 3 * Nodal and Loop Analyses
Exercise. Find a simplified loop equation for in the circuit o f Figure 3.17.
FIG U RE 3.17.
c' '1
To see the importance o f the calculations o f the amplifier circuit o f Example 3.9, suppose two
amplifiers are available for use with a non-ideal voltage source. The non-ideal voltage source is
modeled by an ideal one-volt source in series with a 100 Q source resistance. Suppose amplifier 1
has a voltage gain, vjv^^ = 1 0 and = 100 k£2. Suppose amplifier 2 has a voltage gain o f 100
and R -^ 2 = 5 £3. If amplifier 1 is attached to the non-ideal source, then by voltage division, =
100,000/(100,000 + 100) = 0.999 V, whereas in the case o f amplifier 2, v -^ 2 = 5/(100 + 5) =
0 .0 4 7 6 V. In the first case, the gain from to is 10, yielding = 9.99 V. In the second case,
the same gain is 100, yielding v^ 2 = V. One concludes that amplifier 1 is better suited to this
particular application, although it has a lower voltage gain than amplifier 2. Hence, Example 3.9
illustrates the need to know both the voltage gain and the input resistance to determine the out
put voltage in practical applications. Further, using the literal solution to the example allows us to
apply the formulas to different sets o f parameter values without repeating the complete analysis.
In the previous two examples, there were current sources on the perimeter o f the circuit. Such cur
rent sources were incident to only one loop. It often happens that independent and dependent
current sources can be common to two or more loops. When this happens, a situation analogous
to floating voltage sources in nodal analysis occurs. To handle such cases, many texts define some
thing called a supermesh and write a special loop equation for this supermesh. Supermeshes often
confuse the beginner. There is an easier way.
Example 3.10 below illustrates how to write “loop” equations when current sources are common
to two or more loops. In such cases, we introduce auxiliary voltage variables across current
sources common to two or more loops. The resulting set o f simultaneous equations will contain
not only the loop currents as unknowns, but also the auxiliary voltages as unknowns. Because the
resulting set o f equations contains both loop currents and additional (auxiliary) voltage variables,
the equations are called modified loop equations. T he process o f writing modified loop equa
tions is extremely systematic and straightforward. Further, it allows us to avoid explaining the very
confusing concept o f a supermesh. On the other hand, the presence o f auxiliary voltage variables
increases the number o f “unknowns,” i.e., the number o f simultaneous equations increases.
Because o f the availability o f software packages such as MATLAB, M ATH EM ATICA, and
MAPLE, this increased dimension is not a hindrance.
Chapter 3 * Nodal and Loop Analyses 137
E X A M P L E 3 .1 0 .
Consider the circuit o f Figure 3-18 in which = 28 V and 7^2 = 0 .06 A. Note that the inde
pendent current source is common to loops 1 and 3 and a voltage-controlled current source is
common to loops 1 and 2. Find values for the loop currents 7p Ij, l y and the power delivered by
each independent source.
200Q
So l u t io n .
To begin the solution, we introduce two auxiliary voltage variables Vj and Vj associated with the
current sources common to two (or multiple) loops. The purpose o f these variables is to facilitate
the application o f KVL for constructing the loop equations. This will require that we obtain three
KVL equations, one for each loop, and two constraint equations, one for each current source.
28 = 2007] - V] - V2
(3.36)
Step 2. Apply KVL and Ohm’s law to loop 2. Again applying KVL and Ohm’s law to loop 2, we
obtain 100 7j + 200 (Jr^ —7^) + V2 = 0. After grouping like terms.
3 0 0 7 , - 20 0 /3 + V2 = 0 (3.37)
Step 3. Apply KVL to loop 3. Applying KVL to loop 3 yields 150 7^ + Vj + 200 (/j - Tj) = 0.
Equivalendy,
- 200/2 + 350/3 + V| = 0
Step 4. Write a constraint equation determined by the independent current source. Here, loops 1
and 3 are incident on the independent current source so that
0 .0 6 = / , -/ 3 (3.39)
138 Chapter 3 * Nodal and Loop Analyses
Step 5. Write a constraint equation determined by the dependent current source. In a straightfor
ward manner, we have
/i + 3/2 - 4 /3 - 0 (3.40)
Step 6. Write equations 3 .3 6 to 3.40 in matrix form an d solve. The matrix form o f these equa-
tions is
0 0 -1 -r w ■ 28
300 -2 0 0 0 1 h 0
-2 0 0 350 1 0 h = 0 (3.41)
0 -1 0 0 ''I 0 .0 6
3 -4 0 0 V2 0
Solving equation 3.41 by the matrix inverse method or by an available software package yields the
solution (currents in A and voltages in V) given by equation 3 .42 below:
w 200 0 0 -1 -r -1 ■ 28 ■ ■0.1 ■
h 0 300 -2 0 0 0 1 0 0 .02
= (3.42)
0 -2 0 0 350 1 0 0 = 0 .0 4
Li j .
1 0 -1 0 0 0 .06 -To
I 3 -4 0 0 0 2
,''2.
»LoopIplus= M\b
Looplplus =
l.OOOOe-01
2,0000e-02
4,0000e-02
-l.OOOOe+01
2.0000e+00
Chapter 3 • Nodal and Loop Analyses 139
Step 7. Compute the powers delivered by the independent sources. First, the power delivered by the
independent voltage source is
Py-source = 28 /j = 2.8 watts
Pl-source = V^ = - 0 .6 watts
This last value indicates that the independent current source actually absorbs power from the cir
cuit.
Exercise. For the circuit o f Figure 3.19, write the modified loop equations having two unknowns
/j and V, following the procedure described in Example 3.10. Solve the equations and find the
power absorbed by the 2 -Q resistor.
4A
AN SW ER: 18 watts
One final point before closing our discussion o f loop analysis. Loops can be chosen in different
ways. Cleverly choosing loops can sometimes simplify the solution o f the associated equations. For
example, by choosing a loop that passes through a current source so that no other loop is com
mon to the source, the loop current is automatically specified by that current source.
6. SUM M ARY
This chapter introduced the technique o f nodal analysis. Nodal analysis is a technique for writing a set
o f equations whose solution yields all node voltages in a circuit. With knowledge o f all the node volt
ages and all the element values, one can compute all branch voltages and currents. As mentioned, when
ever there are no dependent sources present, the coefficient matrix o f the node equations is always sym
metric. Hence, whenever dependent sources are absent, it is possible to write the nodal equation coef
ficient matrix by inspection. Further, if independent voltage sources are absent, then the right-hand side
o f the matrix form o f the nodal equations can also be written by inspection: the entry is simply the sum
o f the independent source currents injected into the node at which KCL is applied. When VCCSs are
present, the steps for writing nodal equations are the same as illustrated in Example 3.3. Generally, in
such cases, the resultant coefficient matrix is not symmetric.
140 Chapter 3 • Nodal and Loop Analyses
W hen floating dependent or independent voltage sources are present with respect to a given ref
erence node, we introduce new current variable through these floating sources as unknowns. The
node equations then incorporate these additional unknown currents, as was illustrated in
Examples 3.5 and 3.6. This method increases the number o f equations but simplifies the con
struction o f the individual equations. W ith a tool like MATLAB to compute solutions, there is no
difficulty, although hand computation may become more difficult. This concept is the basis o f the
modified nodal analysis method used in circuit simulation programs like SPICE.
Loop/mesh analysis, an approach dual to nodal analysis, was introduced in Section 5. Mesh analy
sis is a special case o f loop analysis for planar circuits when the loops are chosen to be the obvious
meshes, similar in geometry to a fish net. In loop analysis, one sums the voltages around a loop or
mesh to zero. Each o f the branch voltages in the loop is expressed as a product o f resistances and
(fictitious) loop currents that circulate through the branch resistance, as illustrated in Figures 3.10,
3.14, and 3.16. The branch current o f the circuit are equal to the net flow o f the loop currents
incident on a particular branch, meaning that each branch current is expressible as a sum o f loop
currents. The desired set o f loop equations is produced by summing the voltages around each loop,
expressing these voltages either as source values or as resistances times loop currents. One solves
the loop equations for the loop currents. Once the loop currents are known, we can then compute
the individual branch currents and then the branch voltages, and thus any other pertinent current,
voltage, or power. Whenever there are no dependent sources present, the coefficient matrix o f the
loop equations is always symmetric. Whenever dependent sources are absent, it is possible to eas
ily write the loop matrix by inspection.
As the size o f an arbitrary circuit grows larger, there are two good reasons for choosing the nodal
method over the loop method; (i) the number o f nodal equations is usually smaller than the num
ber o f loop equations, and (ii) the formulation o f nodal equations for computer solution is easier
than methods based on loop equations. Writing nodal equations is particularly easy if the circuit
contains only resistances, independent current sources, and VCCSs — for short, an R—I—
g„ net
work. For an network, one simply applies KVL to every node (except the reference node)
and obtains a set o f node equations directly. For floating independent or dependent voltage
sources, the task is more complex. Examples 3.5 and 3.6 illustrate cases where, besides the node
voltages, additional unknown auxiliary currents are added. By adding additional auxiliary variables
to the formulation o f the nodal equations, we described the concept behind the modified nodal
analysis (MNA) method. The MNA method retains the simplicity o f the nodal method while
removing its limitations and is the most commonly used method in present-day computer-aided
circuit analysis programs.
Connected circuit: every pair o f nodes in the circuit is joined by some set o f branches.
Cram er’s rule: a method for solving a linear matrix equation for the unknowns, one by one,
through the use o f determinants; the method has serious numerical problems when
implemented on a computer, but is often convenient for small, 2 x 2 or 3 x 3, hand cal
culations.
Floating source: neither node o f the source is connected to the reference node.
Chapter 3 ®Nodal and Loop Analyses 141
Gaussian surface: a closed curve or a closed surface surrounding two or more nodes.
Linear matrix equation: an equation of the form Ax =b, where A i s z n x n matrix, x is an n-vec-
tor of unknowns, and b is an n-vector of constants.
Loop (closed path): a contiguous sequence of branches that begins and ends on the same node
and touches no node more than once.
Loop analysis: an organized method of circuit analysis for computing loop currents in a circuit.
Knowledge of the loop currents allows one to compute the individual element currents
and, consequently, the element voltages.
Loop current: a (fictitious) current circulating around a closed path in a circuit.
Matrix inverse: the inverse, if it exists, of an n n matrix yl, denoted b y ^ “ ^ satisfies the equation
A A~^ = A~^ ^ = I,, where / is the w x « identity matrix; the solution of the linear matrix
equation, ^ is given hy x = A~^b,
Mesh: After drawing a planar graph without branch crossing, the boundary of any region with
finite area is called a mesh. Intuitively, meshes resemble the openings of a fish net.
Mesh analysis: the special case of loop analysis for planar circuits in which the loops are chosen
to be the meshes.
Mesh current: a fictitious current circulating around a mesh in a planar circuit.
Modified nodal analysis: a modification of the basic nodal analysis method in which the
unknowns are the usual nodal voltages plus some naturally occurring auxiliary currents.
Nodal analysis: an organized method of circuit analysis built around KCL for computing all node
voltages of a circuit.
Node voltage: the voltage drop from a given node to a reference node.
Symmetric matrix: a matrix whose transpose is itself I f ^ = is a « x « matrix whose i- j entry
is then A is symmetric if a - = a-.
142 Chapter 3 • Nodal and Loop Analyses
=
PROBLEMS 3. For the circuit o f Figure P3.3, suppose
1.2 A. Write a single node equation in the volt
age V and solve.
SIN G LE N ODE PROBLEM S
1. For the circuit o f Figure P3.1, write a single
node equation in , Gj > <^3 > ^ 2-
For a fixed K > 0 , R-^ = R , R2 = 2R , = 2R.
Compute Kj in terms o f R and V^j Figure P3.3
4 Ksi- A N SW ER: - 6 V
Figure P3.2
Chapter 3 “ Nodal and Loop Analyses 143
0.4, G2 = 2, G3 = 3, and = 5. Use nodal out. Then find the linear equation
analysis to find and V^. Then compute the relating to temperature. Plot Vq
power delivered by the independent sources - y c as a W c tio n of temperature, i.e.,
and the power absorbed by G2 . over the range - 2 5 ° C to +25°C . Over
what range of temperatures about 0
degrees would the sensor be reason
ably accurate? W hy and why not?
144 Chapter 3 • Nodal and Loop Analyses
Figure P 3 .1 1
ion. ©
Reference node
Figure P 3 .1 0
Figure P 3 .13
Chapter 3 “ Nodal and Loop Analyses 145
14. Consider the circuit of Figure P 3.14 17. Use nodal analysis to find the voltages
(a) Write two node equations in terms of , and in the circuit of Figure P 3.17.
the literal variables in Figure P 3.14 Suppose = 20 Q, i?2 = 10 Q, = 4 Q, =
and put in matrix form. 0.1 S, = 10 Q, and I^= 6 A. Note that in
(c) Solve the node equations for the volt solving this problem, you are to generate three
ages and Vq when = 0.1 A, 7^2 = (nodal) equations in which the unknowns are
0-2 A, = 7 mS; = 2 mS, = you could eliminate the equation
500 Q ; i?2 = 333.33 £2; and /?3 = 1 fl. for but this problem is to illustrate that
(d) Determine Kq. such elimination is not necessary. Finally, deter
(e) Determine the power delivered by mine the equivalent resistance seen by the inde
each source. (Be careful of sign.) pendent current source.
. 9.iVa
'4 "■
Figure P3.14
Figure P3.17
-------- --------------
■ 0 .0 0 8 0 .0 1 9 - 0 .0 0 5 ' ■ 0 ■ 0 ■
-0 .0 0 1 0 .0 0 5 5 - 0 .0 0 2 = 0 = 0
-0 .0 0 5 - 0 .0 0 2 0 .1 0 7 O .n/ G A.
25. Consider the circuit o f Figure P3.25 in 27. T he modified nodal equations for the cir
which l/j = 2 50 V, = 5 0 V, = 50 Q , cuit o f Figure P 3.27 are
: 20 Q, = 5 0 Q, = 40 Q , and R^ = 10 Q.
■0.004 -0 .0 0 1 -0 .0 0 2 0 •T 4 ' O'
(a) Identify the floating voltage source and
add a current label through the source. -0 .0 0 1 0.001 0 -1 yB
(b) W rite modified nodal equations, -0 .0 0 2 0 0.004 1 0
which include both node voltages and 2 -1 -1 0 ^CB 0
unknown currents through any float
ing voltage sources. Compute all four resistor values and (3. Hint:
(c) Solve the equations for the node volt- Find all the conductances first and then convert
ages Vg and Vq and the current to resistances.
through the 50 V source. C H EC K :
5 0 V a n d V^= 100 V.
(d) Find the power consumed by R^ .
(e) Determine the power delivered by
each o f the sources.
Rc
Figure P3.27
Figure P 3.32
Figure P3.33
Reference node
AN SW ERS T O (C) IN RA N D O M O R D E R :
40, -2 5 ,3 8 .7 5 ;/ ? = 12 Q .
Figure P3.31
Chapter 3 • Nodal and Loop Analyses 149
Figure P3.37
C H EC K : /j = 100 mA.
(b) Now suppose /^j = 4 0 0 mA and 1 ^ 2 =
100 mA and the loop equation for /j
C H EC K S: R^^ = 2 50 Q, and = 8 watts.
written in the standard way directly
yields 2000/|= 60. Find R-^ and R2 if
3 5 . In the circuit o f Figure P 3.35, = 4 00 £2,
R^ = 600 fil. Note: If the equations are
= 5 0 Q, and = 50 Q, and jj. = 0.5. If
not written in the standard way, the
= 50 V, find , the power delivered by the
solution is not unique. For example,
independent and dependent voltage sources,
multiplying both sides o f the above
and the equivalent resistance, R^^ , seen by the
equation by 0.5 yields a different
independent source.
answer in which R-^ = 140 £2 , as
opposed to the correct answer o f R-^ =
400 Q.
Figure P3.38
39. Consider the circuit o f Figure P3.39. with internal resistances R ■20 Q and R2 = 80
(a) Suppose = 2 50 Q, = 5 0 0 Q, Q (faulty connection) respectively connected in
= 100 V, and /S = 0.5. Use loop analy parallel to supply power to a load o f 7?^ = 80 Q.
sis to find /j and R^^ . Compute the power absorbed by the load R^
(b) Compute the power dehvered by each and the power delivered by each independent
source and absorbed by each resistor. source. W hich battery supplies more current to
Verify conservation o f power. R^ and hence more power to the load? How
(c) Compute R^^ as a function o f R^, much power is wasted by the internal resistanc
and 13 . Suppose = 250 Q and = es o f the battery?
5 00 Q , plot R as a function o f ,0
s 13 s 2 .
Pi,
0 '
Figure 3.42
Find the value of each resistance and each (b) the current in the locomotive motor and
source in the circuit. the power absorbed by the locomotive
(c) repeat parts (a) and (b) when the loco
motive is 1/3 distant from either station
CHECKS: V^2 = ^0 V, = 40 Q.
■40 -8 0 - 1 0 ‘ \h' ■n ■
-3 0 130 -5 0 = -V2
h
-1 0 -5 0 70 [*3. 0
Figure P3.45
Find *
46. Reconsider Problem 3.45. Let = 590 V
and R^ = 1.296 Q. This time, suppose there are
two locomotives on the track. One is 1/3 dis
tant from the East side station, and the other is
1/3 distant from the West side station.
(a) Determine the resistance R in Figure
P 3.46.
(b) Using the indicated currents, write a
set of three mesh equations and solve
for , ^2 ’ H•
(c) Determine the two motor currents.
45. Figure P 3.45a shows an electric locomotive (d) Determine the power delivered by
propelled by a dc motor. The locomotive pulls each of the 660 V sources.
a train of 12 cars. The motor behaves
---------------------- -------- ------------- 1--------- ------------------
like a 590 V battery in series with a R __ I ^ R A
1.296 Q resistor. Suppose the train is
midway between stations, West side 660 V
and East side, where 66 0 V dc T l= 660VVf©
sources provide electricity. The T
resistance of the rails affects the cur
rent received by the locomotive. The R R
equivalent circuit diagram is given by Figure Figure P3.46
P3.45b, where R 0.15 Q. Using m e ^ analy-
sis find 4 7 . Reconsider the Problem 3.5 and the circuit
(a) the currents and I 2 of Figure P3.5. Draw two loop currents and
152 Chapter 3 ® Nodal and Loop Analyses
solve for these currents. Then compute the 50. Consider the circuit o f Figure P 3.50 in
node voltages > and . which = 40 V and V^ 2 = 20 V. Write a set o f n\
three loop equations by inspection. Refer to
48. In Figure P 3.48, let = 9 kQ, = 18 kQ, Example 3.7 and the discussion following the o
= 6 kQ, = 9 kQ, = 3000 Q, and = example. Solve the loop equations using matrix
20 mA. methods via your calculator or an appropriate
(a) Write two mesh equations in and /2. software program. Compute the voltage v. Note
Put in matrix form and solve. that /| and I 2 should have values identical with
O
(b) Specify the voltages 1^^, , those in example 3.7. Finally, find the power
^BO ^CD’ K iD' delivered by each o f the sources.
(c) Find the power delivered by each o f
the sources. IQ
n
Figure P 3 .4 9 Figure P 3.51
Chapter 3 • Nodal and Loop Analyses 153
(a) If the loop equation matrix is o f the A, 7?, = 200 £2, 7?2 = 400 Q, = 100
R3 ???? -1 0
C H EC K : /j = 0.4 A and = 50 V.
Ri R3
R.
-VESA
's ■<------
—N/N^
Rs
Figure P3.58
Amplification o f voice allows announcers at sports events to convey their comments on the play-
by-play action to the crowd. At concerts, high-powered amplifiers project a singer’s voice and the
instrumental music into a crowded auditorium. Electronic amplifiers make this possible. One o f
the simplest and most common amplifiers is the operational amplifier, the subject o f this chapter.
The word “operational,” though, suggests a purpose beyond simple amplification. Often one must
sum signals to produce a new signal, or take the difference o f two signals. Sometimes one must
decide whether one dc signal is larger than another. The operational amplifier is operational pre
cisely because it can be configured to do these things and many other tasks, as we will see later in
the text.
SECTIO N HEADIN GS
1. Introduction
2. The Idealized Operational Amplifier: Definition and Circuit Analysis
3. The Design of General Summing Amplifiers
4. Saturation and the Active Region of the Op Amp
5. Summary
Terms and Concepts
Problems
156 Chapter 4 • T h e Operational Amplifier
1. IN TRO D U CTIO N
Chapters 1 and 2 defined and discussed independent and dependent voltage and current sources.
Chapter 3 investigated the nodal and loop analysis o f resistive circuits containing such sources.
Ofi:en, dependent sources supply energy and power to a circuit, making them so-called active ele
ments. O n the other hand, resistors are passive elements because they only absorb energy. Circuit
models o f real amplifiers (see Examples 3.3 and 3.4 with associated Figures 3 .7 and 3.8, respec
tively) contain controlled sources that underlie their analysis and performance evaluation. Indeed,
the VCVS is the core component o f the operational amplifier (op amp), the main focus o f this
chapter. Thus, the op amp is an active circuit element whose analysis is done with the techniques
o f Chapters 1 through 3.
A real op amp is a semiconductor device consisting o f nearly two dozen transistors and a dozen
resistors sealed in a package from which a small number o f terminals protrude, as shown in Figure
4.1(a). Despite its apparent internal complexity, advances in integrated circuit manufacturing
technology have made the op amp only slightly more expensive than a single discrete transistor.
Its simplicity, utility, reliability, and low cost have made the op amp an essential basic building
block in communication, control, and the instrumentation circuits that can be found in all under
graduate EE laboratories.
Top View
Balance 1 [
TO-5
Dual-in-line
Inverting
input ^
Noninverting ^ [
3 6 Output
input
m
E- 4 [ H 5 Balance
(a)
Inverting
E+«- input
— V- Output
Reference
node
E- Noninverting
input
(c)
(d)
FIG U RE 4.1 (a) Typical op amp packages; (b) typical terminal arrangement of an op amp package;
(c) dual power supply notation; (d) essential terminals for circuit analysis.
Figure 4.1 (b) shows a typical arrangement o f terminals for a dual-in-line op amp package. Th e ter
minal markings and the symbol shown in Figure 4.1(b) do not appear on the actual device, but
Chapter 4 • T h e Operational AmpUfier 157
are included here for reference. In Figure 4.1(b), the terminal labeled “N C ” (no connection) is not
used. The E+ and E- terminals (Figure 4.1(b)) are connected to a dual power supply, illustrated in
Figure 4.1(c), where typically ranges between 3 V and 15 V, depending on the application;
adequate voltage is required for proper operation. The three terminals in Figure 4.1(b) marked
“inverting input,” “non-inverting input,” and “output” interact with a surrounding circuit, and
correspond to V, and Vq in Figure 4.1(d). The two terminals labeled balance or ojf-set have
importance only when the op am is part o f a larger circuit: resistors o f appropriate values are con
nected to these terminals to make sure the output voltage is zero when the input voltage is zero.
This “balancing process” is best discussed in a laboratory session.
This chapter sketches the basic properties o f the op amp: just enough to understand some o f the
interesting applications. The ideal op amp model and the saturation model are described. Using
these models and the principles o f analysis covered in Chapters 1 through 3, we then analyze the
behavior o f some widely used op amp configurations. These application examples hint at the
importance o f the op amp and furnish motivation for the study o f electronic circuits.
Several o f the examples include a SPIC E simulanon o f the circuit being analyzed. SPIC E is a
sophisticated circuit simulation program. Behind the user-interface, SPIC E uses complex models
o f the real operational amplifier. Our purpose in using SPIC E simulation is to verify or test the
theoretical analysis set forth in the examples. W hat we show is that the simplified theoretical
analysis provides a very good approximation to the actual circuit behavior represented in the
SPIC E simulation results. Industrial circuit designers often use SPIC E to visualize the expected
behavior o f very complex circuits. Later chapters cover some o f the more complex op amp appli
cations.
This section analyzes resistive circuits containing an operational amplifier. Figure 4 .2 explicitly
shows an op amp embedded in a surrounding resistive circuit.
FIG U RE 4.2
One possibility for analyzing op amp circuits is to represent the op amp by one o f the simplified
models shown in Figure 4.3 that do not account for saturation effects. The first model o f Figure
4.3(a) consists o f an input resistor, an output resistor, and a V CV S with finite gain A.
O f practical import is the idealization o f this model (Figure 4.3(a)) to the one o f Figure 4.3(b) by
158 Chapter 4 • T h e Operational Amplifier
(1) letting R-^ become infinite, setting up an open circuit condition at the input terminals; (2) let
ting become zero, making the output voltage o f the op amp equal to that o f the V C V S; and
(3) letting the gain A approach infinity. These conditions are idealizations because (1) with R-^
infinite, there is no loading to a circuit attached to the input; (2) with R^^^ = 0, the full output
voltage appears across any circuit connected to the output; and {5) A ^ leads to a simplifica
tion o f the associated analysis. These conditions, stated below as equation 4.1, define the so-called
ideal operational amplifier;
Rest of Circuit
Rout +
+ V
A(v - V
(a) —
FIG U RE 4.3
To see how this idealization simplifies op amp circuit analysis, consider an equivalent set o f con-
ditions for the ideal op amp, called the virtual short circuit model:
(4.2a)
(4.2b)
v^ = v_ (4.2c)
From Figure 4.3(b), the conditions that = 0 and i_ = Q follow directly from the open circuit con
dition at the input terminals. The condition that = v_ (hence, the term “virtual short circuit”)
will be discussed later, but occurs because A ^ co, forcing {v^ —v j 0.
Chapter 4 • T h e Operational AmpUfier 159
The recommended way to analyze circuits containing op amps is to replace any ideal op amp by the
model o f Figure 4.3{b), the virtual short circuit model o f equation 4.2. Th e following examples illus
trate the use o f the virtual short circuit model.
EX A M P L E 4 .1 . This example investigates the inverting am plifier o f Figure 4.4, which is used in
a wide range o f commercial circuits. The objective is to compute in terms o f Rjs and
+
V
Rf
FIG U RE 4 .4 Inverting amplifier, assuming an ideal op amp in which V„„, = ---- —V:„ .
So l u t io n
Step 1. Compute and v_. Since the + terminal is grounded, = 0. From the virtual short prop
erty o f the ideal op amp, v_= v^ = Q.
Step 2. Compute i^-^. Since v_ = 0, the voltage across R-^ is v-^. From Ohm’s law,
v,„
R
Step 3. Compute iy. Again, since v_ = 0, the voltage across Rf is From Ohm’s law.
if = -
R
Step 4. Relate the currents i-^ and ip and substitute the results o f Steps 2 and 3. From KCL, i^^ —i_
+ ir= 0. From the properties o f the ideal op amp, i_ = 0, in which case, ir= -i-^. This imphes that
R*
V
Hence, the voltage gain relationship o f the inverting op am circuit is
Rf
V ou t= --^ ^ in (4.3)
Rin
Equation 4.3 shows that the input and output voltages are always o f opposite polarity, hence the
name inverting am plifier. One also observes that by choosing proper values for Rjr and R^ a volt
age gain o f any magnitude is possible, in theory. In practice, other factors limit the range o f obtain
able gains.
160 Chapter 4 • T h e Operational Amplifier
lOOkO
25kQ
-s/ W '
50 mV
lookn
A N SW ER F O R BO T H : = - 200 mV.
A few remarks are in order. Op amp configurations in which one o f the input terminals is ground
ed, as is the non-inverting terminal in Figure 4.4, are said to operate in the single-ended mode. The
input terminal can be grounded directly or through a resistor, as in Exercise 2 above. Also, since
v_ = or v_-v^ = 0, the terminals are virtually short circuited even though there is no hard-wired
direct connection between them. This condition is called a virtual short circuit. Further, if one
o f the terminals is grounded, then the other terminal is said to be virtually grounded, as is the
case in Figures 4.4, 4.5, and 4.6. Specifically, in Figure 4.4, there is a virtual ground at the invert
ing input terminal.
The next example continues the investigation o f the ideal inverting amplifier for the two-input,
single-output op amp circuit o f Figure 4.7. The solution again makes use o f the virtual ground
and virtual short circuit properties o f the ideal op amp.
Chapter 4 • T h e Operational Amplifier 161
E X A M P L E 4 .2 . For the circuit o f Figure 4.7 , our objective is to compute in terms o f R^, ^2>
and the two input voltages and v^2 -
Rf
+
V
FIG U RE 4.7 Inverting (ideal) amplifier with rwo inputs for which
Rf Rj
v„„,=-^v,|V ,2.
Step 1. Compute ij and Since v_ = v^ = 0, the voltage across is and the voltage across Rj
is v^2 - From Ohm’s law,
R^
Step 2. Compute ij-. Again, since v_ = 0, the voltage across Rj- is and from Ohm’s law.
'/ = ■R
/
Step 3. Relate the currents /j, ^2 and then substitute the results o f Steps 1 and 2. From KCL,
h H —i_ + ijr = ^- From the properties o f the ideal op amp, i_ = 0, in which case, ijr = -(/j +
This implies that
Hence,
Rf Rj
(4.4)
162 Chapter 4 • T h e Operational Amplifier
Exercise. In Figure 4.7, suppose Rjr= 100 IcD. Find and ^2 so that
AN SW ER: = 25 kD and = 50 kD.
''out
V.
So l u t io n
Step 1. Compute and v_. Since the + terminal is connected to the input voltage source, = v-^.
From the virtual short property o f the ideal op amp, v_ = v^ = Vj^.
Step 2. Compute zj . Since v_ = the voltage across the resistor is Observe that the current,
Zj, has reference direction different from the passive sign convention. Hence, from Ohm’s law,
h=-
Step 3. Compute ip Again, since v_ = v-^, the voltage across R^-is From O hm s law,
Vout - Vin
lf =
Step 4 . Relate the currents z'j and ip and substitute the results o f Steps 1 and 2. From KCL, z'j - z_ +
Zyr= 0. From the ideal op amp property o f equation 4.2, z_ = 0, forcing z^= -Zj. This implies that
Rf /?,
Chapter 4 • T h e Operational AmpHfier 163
Rf\
1+ ^ (4.5)
R i)
/ R \
From equation 4.5, the voltage gain is greater than 1, i.e., i 1 + - i - i > 1, and and always
I }
have the same polarity; the circuit is naturally called a non-inverting amplifier.
Exercise. For the non-inverting amplifier o f Figure 4.8, find and so that the gain is 2, and
when v-^ = 5 V, the power absorbed by R^ is 5 mW.
AN SW ER: Rf = 5\<Q. and /?,^ = 5 kQ
i,n
EXA M PLE 4 .4 . This example analyzes the ideal general di^Ference amplifier circuit o f Figure
4.9. We show that
Kf
^’s 2 - - r ^ s \
'’" " ' I R\
In a basic difference amplifier, the output is the difference o f two input voltages. For the gener
al difference amplifier o f this example, the output is a difference o f the scaled input voltages,
= for appropriate positive and a^.
So l u t io n . From the ideal op amp property o f equations 4.2, v_ = and no current enters the
inverting and non-inverting op amp input terminals.
Step 1. Write a node equation at the non-inverting input terminal o f the op amp. Summing the cur
rents leaving the + node o f the op amp yields
G 2 { v + - v , 2 ) + GgV+=Q
164 Chapter 4 • T h e Operational Amplifier
(4.6a)
Step 2. Write a node equation at the inverting input terminal o f the op amp. Recall v_ = v^. T he sum
o f the currents leaving the - node satisfies
G i ( v+ - v, ] ) + G ^ ( v+ - v„„,) = 0
Thus,
Cl
^out = + 1+ (4.6b)
Step 3. Combine Steps 1 and 2. Substituting equation 4.6a into 4.6b yields
(4.7a)
G t + Go
or, in terms o f resistances,
^s2 (4.7b)
[R g + R2 j
Equations 4 .7 have the desired form: appropriate positive constants <?2
a ^, which can be obtained by proper choices o f the resistors.
Two special cases o f Example 4 .4 are o f practical importance. First, if and R^ = /?2> then
equation 4.7b reduces to the classical difference amplifier equation.
out = K {v , 2 -V si)
with K = I and for an arbitrary K > 0 , Rj- = KR^ and R^ = KR2 fits the bi
bill.
2. Using the circuit o f Figure 4.9, design a difference amplifier so that = 4(t^^2 “ ^ji) ^^d the
feedback resistance R^ = 20 kQ.
A N SW ER: R, = R , ^ 5 kQ and R^ = 20 kQ
3. In Exercise 1, suppose R^ and /?yare scaled by a positive constant A'j, i.e., R^^^ = K^R^y and R2
and R are scaled by a positive constant K^. Determine the new input-output relationship.
o
A N SW ER: w ith K the sam e as in Exercise 1
Chapter 4 • T h e Operational AmpUfier 165
The point o f Exercise 3 is that the group , Rj] can be independently scaled by and the group
{Rj, R^ independently scaled by K 2 without affecting the gain o f equation 4.7b.
E X A M P L E 4 .5 . This example analyzes a special case o f the non-inverting amplifier called the
buffer or isolation amplifier, shown in Figure 4.10, where = v-^. W hen connected between
two circuits, the buffer amplifier prevents one circuit from having a loading effect on the other.
So l u t io n . From the connection shown in Figure 4 .1 0 , v-^ = and = v_. From the proper
ties o f the ideal op amp, = v_, in which case = v-^.
Exercise. Compute the power delivered by the source in Figure 4 .1 0 and the power delivered to
the load R^. ^
AN SW ER: 0 and
R,
The circuit o f Figure 4.10 is called an isolation or buffer amplifier, because no current is drawn
from the source However, the op amp does supply current (and power) directly to the load by
maintaining under the condition that not exceed the manufacturer’s maximum out
put current rating. Since Vg„f{t) = the circuit is also called a voltage follower.
Figure 4.11 shows a SPIC E simulation that verifies the behavior arrived at in Example 4.5. Here
a dc voltage sweep, Q <v-^< 12 V, was input to a highly accurate SP IC E model o f a Burr Brown
741 connected to ±10 V power supply. Observe in Figure 4.11 that the output follows the input
up to the 10-volt value, after which, the output remains at 10 V despite increased input values.
This non-ideal phenomenon, called saturation, is due to the power supply voltage level and is dis
cussed in Section 4.
166 Chapter 4 • T h e Operational Amplifier
Exercise. Find for the circuit o f Figure 4.12, the power supplied by the source and the
power supplied to the 12 kQ load.
AN SW ERS: =v,.0.
1 2 x 10-’
Often data acquisition equipment and active fdters require multi-input single-output amplifiers
having a more general summing characteristic, such as
where the constants ay>0 and |3>0. The inverting and non-inverting amplifier configurations
(Examples 4.1, 4.2, and 4.3), as well as the difference amplifier configuration o f Example 4.4, are
special cases o f equation 4.8. W ith a little cleverness, it is possible to design by inspection an op
amp circuit whose input-output characteristic is precisely equation 4.8. The op amp circuit o f
Figure 4.13 having the four inputs V^2 > ^hv ^bl accomplishes this. The circuit looks ordi
nary except for the presence o f one additional conductance, AG, incident on the inverting termi
nal o f the op amp. T he dashed lines in Figure 4.13 are present because this conductance may or
may not be needed. Computation o f the values o f AG and are explained in design Step 2,
below.
b2 0 — s / \ / V
G
The first two design steps constitute a preliminary or prototype design, meaning that the feedback
resistor is normalized to 1 Q , or equivalently, 1 S. After completing the prototype design, an engi
neer would scale the resistances to more practical values without changing the gain characteristics.
T he scaling procedure is explained in Step 3.
Design Step 2. Prototype design continued: Computation ofG^ andhr tS.G so that the total conduc
tance incident at the inverting terminal o f the op amp equals the total conductance incident at the non
inverting terminal.
To achieve this equality, recall that in design Step 1, Cy= 1 S, S, and S. Define
a numerical quantity
^ = (1 + a i + ) - (^1 + )
Design Step 3. Scaling to achieve practical element values. Multiply all the resistances (divide all
conductances) incident at the inverting input terminal o f the op amp by a constant Similarly,
multiply all resistances (divide all conductances) incident at the non-inverting terminal o f the op
amp by It is permissible to choose but this is not necessary.
>^o»/ = - 7 V a l - 3 v ^ 2 + 2 v i , + 4 v^2
(4.9)
So l u t io n
Step 1. Prototype design. Using Figure 4.13, choose Gy- = 1 S, G^j = 7 S, G^2 = 3 S, G^, = 2 S, and
G^2 = 4 S .
Step 3. Scaling. To have practical element values, let us choose = 10^. This scaling leads
to a design with resistances Rjr= 100 kQ, 7?^, = 14.28 kQ, = 3 3.33 kQ, = 50 kQ, R^^ =
25 kD and R„ = 20 kQ.
Chapter 4 • T h e Operational Amplifier 169
14.28 kQ
V3, o------ lOOkO
25 kQ
(a) (b)
FIGURE 4.14. (a) Prototype design of equation 4.9; (b) final design
after scaling with = lO^.
So l u t io n
Step 1. Prototype design. Again, using Figure 4.13, choose 1S,G „=2S,G ,2 = 4S,G^,=7
S, and G^2 = 5 S.
Step 3. Scaling. To have practical element values, let us again choose = 10^. This scal
ing leads to a design with resistances Rjr = 100 kQ, R^j = 50 k fl, R ^2 = 25 kQ, = 14.28 kQ,
R^j = 20 kQ, - 100 kQ, AR = 16.67 kQ. The final design is set forth in Figure 4.15(b).
16.67kn
50 kQ
V3, o------v X / X . - lOOkQ
25 kQ
'^a2 O------- s/\v^v^
14.28 kQ
'^b2 O--- lOOkQ
20 kQ
(a) (b)
FIGURE 4.15 (a) Prototype design of equation 4.10; (b) final design after scaling with = 10^.
170 Chapter 4 • T h e Operational Amplifier
Exercise. 1. Obtain an alternative design for Example 4 .7 such that = 0, implying the saving
o f one resistor.
AN SW ER: In prototype design, AG = 5 mho.
10 kQ
At this point, the reader may wonder how this simple procedure is derived. The derivation o f this
procedure is beyond the scope o f the Hght edition^ Th e interested reader is directed to the 2nd
edition o f this text.
6K U
(a)
FIG U RE 4 .17
Chapter 4 • T h e Operational Amplifier 171
In the previous sections, we assumed the op am functioned ideally: = v_ and = i_=Q. For the
inverting amplifier o f Example 4.1, this led to the very simple gain formula,
_
'^out
Rh,
Thus, as the input voltage increases, the output voltage increases proportionately. For real circuits,
this proportional relationship holds only when < V^^^for some value o f that is associat
ed with the power supply voltage. Intuitively speaking, an op amp cannot generate an output volt
age beyond that o f its power supply voltage, typically less than or equal to 15 V. W hen the V^^-
limit is reached, further increases in the magnitude o f v-^ produce no change in the value o f
This behavior is called saturation.
saturation occurs are . If = 15 V and A= 10^, the critical threshold voltages are ±0.15
A
mV; if A is infinite, as in Figure 4.19(b), then saturation occurs when \v^ > 0.
FIG U RE 4.18 Practical op amp model with a nonlinear controlled voltage source.
The linear r e l a t i o n s h i p , - v^=A{v^ - v J , holds for segment 1 in Figure 4.19(a), which is said
to be the linear region or active region o f the op amp, denoted by
172 Chapter 4 • T h e Operational Amplifier
Typical values o f finite A range from 10'^ to 10*^. The active region is the ordinary region o f oper
ation. In the active region, the op amp provides a very high (open loop) voltage gain A, the slope
o f segment I. The phrase “open loop” gain means that there is no connection through a wire, a
resistor, or some other device back to the input terminals.
Models o f the three operating regions o f the op amp are summarized in Table 4.1.
A f(v^-v)
V
\
Positive
Saturation
/
Active / d
Region = V -V
Negative
Saturation
\ -V
(b)
FIG U RE 4.19 A piecewise linear (three-segment) curve for the op amp that specifies the active
and positive/negative saturation regions of operation: (a) finite gain A, and (b) (ideal) infinite gain A.
f(Vd)
Vcl =
A
Active and
*^sar
Vsa,
Positive A
II
saturation and
sat
<-
Negative
III
saturation and
Chapter 4 • T h e Operational Amphfier 173
The use o f a three-segment curve in Figure 4 .1 9 is different from the techniques o f earlier chap
ters. The operating point, determines the proper segment to be used for analysis. If the
input is small, one reasonably assumes the operation is in the active region, segment I. However,
when the input magnitude is large, one must “guess and check” to determine the appropriate oper
ating region. For example, should the guess be incorrect, then the model for one o f the other
regions must be used and the analysis repeated until a valid solution (and operating region) is
obtained. The following example illustrates the approach.
E X A M P L E 4 .8 . The purpose o f this example is to illustrate that an op amp may operate in any
o f three regions and also to illustrate that the determination o f the region o f operation using the
“guess and check” method. Recall the inverting am plifier o f Figure 4.5. Suppose is infinite,
= 50 kQ, = 10 kQ and = 15 V. Find and for the following three cases: (a) = 0.5
V; (b) = 4 V; and (c) = - 5 V. Finally, verify the theoretical analysis using SPICE.
So l u t io n
(a) Assume the op amp operates in its active region. From equation 4.3 in Example 4.1, the out
put voltage is
vout
„ = - ^ v , m, = - ^JQx 0 . 5 = - 2 . 5 V
Since |-2.5| < = 15 V the op amp operates in its active region; the answers = - 2 .5 V and
v^ = 0 are valid.
^ 50
(b) W ith v-^ = 4 V, assuming operadon in the active region, = -----= -------------- x 4 = - 2 0 V.
Rl 10
However, since |-20| > = 15 V, the op amp does not operate in its active region. Therefore,
^out ~ ^ invalid, but does suggest operation in the negative saturation region. The negative
saturation model o f Table 4.1 yields the circuit o f Figure 4 .20 in which = - 1 5 V.
By writing and solving a single node equation at the inverting input terminal designated by the
minus sign in Figure 4.20, we obtain Vj = - 0 .8 3 V.
R' 50
(c) With v-^ = - 5 V, assuming operation in the active region, ~ ^ = 25 V
This result suggests that the op amp is really operating in the positive saturation region. Using the
positive saturation model o f Table 4.1, Figure 4.21 shows the proper circuit configuration with
174 Chapter 4 • T h e Operational Amplifier
^out ~ 15 V. As in the previous case, by writing and solving a single node equation at the invert
ing terminal designated by the minus sign in Figure 4 .2 1 , = 1.667 V. In this case and in case
(b) above, 0, as we were not in the active region o f operation, and it was necessary to change
the guessed region o f operation to obtain valid results.
A SPIC E simulation was used to validate the theoretical analysis^. A D C sweep, - 4 < < 4 V,
is an adequate input to demonstrate the saturation effects. In the SPIC E simulation, an accurate
model for a 741 op amp manufactured by Burr Brown was used. T he resulting dc transfer curve
is shown below in Figure 4.22.
Lin/Decarlo E xI-D C Transfer-4 Vin
Output voltage
(V) -4.000 - 1.000 + 0 .0006+000 + 1.000 +3.000 +4.000
V(IVM)
From this curve, one can see that the op amp saturates for input voltages v-^ such that > 3,
and the op amp operates in its linear region whenever \v-^ < 3. As hoped, the simplified three-seg
ment model in Table 4.1 yields very good results in all regions o f operation relative to the realis
tic SPIC E simulation.
O ne can conclude from the above example that for the purpose o f faithfully amplifying an input
signal, the input should not be so large as to drive the op amp into saturation. Driving an op amp
into saturation distorts the output signal relative to the input. O n the other hand, for some spe
cial applications, such as the com parator, saturation is precisely the property to be utilized. Figure
Chapter 4 • T h e Operational AmpHfier 17 5
4.23 shows two comparator circuits. A com parator circuit compares tlie input voltage v-^ with a
reference voltage Vygjr{or some multiple o f Only two different output voltages are produced,
one for and the other for v-^ <
EXA M PLE 4 .9 . For the com parator circuits shown in Figure 4.23, each op amp has infinite gain
and a saturation voltage = 1 5 V"^:
(a) Find the v -^ relationship for the comparator o f Figure 4.23(a).
(b) Repeat part (a) for Figure 4.23(b).
Note that in both circuits, there is no connection between the output and inverting input
terminals, a departure from all the previous circuit configurations. Because o f this, for
almost all voltages, ^ v_, and =
Vin O------s / W -
2 0 kQ
''re f O --------
= -2 0 V 80 kn
1 .
1 .
(a) (b)
FIG U RE 4 .23 Two comparator circuits that are used to determine when an input
voltage is above or below a reference voltage.
S o l u t io n
(a) For > - 5 V, the voltage v^= v_ = -5 - v-^ < 0. Referring to Figure 4.19(b),
^out ^ ~^sat = - 1 5 V. Similarly, for v-^ < - 5 V, the voltage v^ = v^ - v_ = -5 - v-^ > 0, and hence
^out = ^sat = 15 V.
(b) By the fact that no current flows into the input terminals o f Figure 4.23(b), using nodal analy
sis, we have that
V - V
20x10^ 80x10^
m which case,
- 4 + 0.8v,„
For > - 5 V, the voltage v^ = - v_ = 0 - {-A + Here, referring again to the satu
ration curve o f Figure 4.19(b), = - 1 5 V. Similarly, when v-^ < 5 V, the voltage
v^ = v ^ -v _ = Q -{-A + 0.8V.J > 0; hence, = 15 V.
176 Chapter 4 • T h e Operational Amplifier
To verify this analysis, the circuit o f Figure 4.23(b) was simulated in B2Spice using a Burr Brown
741 op amp model. T he results o f the simulation are given in Figure 4.24. The theoretical analy
sis based on the simplified models o f Table 4 .1 shows a very good match with the more realistic
SPIC E simulation results.
example4.8-DC Transfer-6
Output voltage
(V ) -1 0 0 ° +2.000 +3.000
V(IVM)
Exercise. For the circuit o f Figure 4.25, suppose = 12 V, find the range o f v^2 for which the op
amp is in positive saturation. Then find the range o f v^ 2 fof negative saturation.
AN SW ER: when < -^ V, and =- when
V , C3------- v N / \ - -
75 kO
-o
''s2 o-----s /s y \/- +
25 kQ
1 .
FIG U RE 4.25
Chapter 4 • T h e Operational Amphfier 177
5. SUM M ARY
This chapter has introduced the operational amplifier and a number o f practical circuits that uti
lize this new device. These circuits include the inverting and non-inverting amplifiers, the buffer
amplifier, the difference amplifier, and the general summing amplifier. W ith regard to the gener
al summing amplifier, a simple design algorithm is described and exemplified. The analysis o f
these circuits builds on the definition o f an ideal op amp, meaning that, when properly config
ured, no current enters the input terminals and the voltage across the input terminals is zero; these
properties are referred to as the virtual short circuit model o f the op amp, i.e., the ideal op amp
has infinite input resistance, zero output resistance, and an infinite internal gain, A. (See equations
4.1 and 4.2.) Practically speaking, the gain A, is not infinite, but ranges between lO'^ and 10*^.
After exploring properties o f the ideal op amp, we discussed the phenomena o f output voltage sat
uration. By introducing output saturation, the ideal model o f the op amp gives way to a more real
istic one, characterized by three regions o f operation, each having its own “ideal” model, as set
forth in Table 4.1. In practical design and applications, output saturation is either to be avoided
or utilized to some advantage, as in the case o f the comparator circuit studied in Example 4.8. For
a faithful amplification o f an input signal, saturation is to be avoided.
Active element: A circuit element that requires an outside power supply for proper operation and
has the capability o f delivering net power to a circuit such as is the case for an op amp or
negative resistance.
Buffer: A circuit designed to prevent the loading effect in a multistage amplifier. It isolates two
successive amplifier stages. Characteristics o f an ideal buffer are infinite input impedance,
zero output impedance, and constant voltage gain.
Com parator: an op amp circuit that compares the input voltage with a reference voltage
(or some multiple o f only two different output voltages are produced, one for <
v-^, and the other for v-^ < v-^.
Difference amplifier: given two inputs, and a difference amplifier produces the output
^out ^ appropriate constant k, often taken as 1.
General summing amplifier: an op amp circuit having the input-output relationship =
- + ^ n ^ a r) + + - + ^ m ^ bn ) constant a - and p ..
p o s itiy c
Ideal op amp: An operational amplifier with infinite input resistance and infinite open-loop gain.
Inverting amplifier: An operational amplifier connected to provide a negative voltage gain at dc.
Linear active region: In the op amp output vs. input transfer characteristic, the region where the
curve is essentially a straight line through the origin is called the linear active region.
Non-inverting amplifier: An operational amplifier connected to provide a positive voltage gain
at dc.
Open-loop gain: The ratio o f the output voltage (loaded, but without any feedback connection)
to the voltage across the two input terminals o f an op amp. The slope, p, o f the straight
line in the active region o f an op amp is the open loop gain under no load condition.
When a load is present, the open loop gain is reduced to + R^, where R^ is
the output resistance o f the op amp.
178 Chapter 4 ®The Operational Amplifier
Operational amplifier (abbreviated op amp): A multi-stage amplifier with very high voltage
gain (exceeding 10"^) used as a single circuit element.
Passive elements: a circuit element that cannot deliver net power to a circuit such as a resistor.
Saturation regions: In the op amp output vs. input transfer characteristic, the region where the
curve is essentially a horizontal line is called the saturation region. There are two such
regions: one for positive input voltage, and the other for negative input voltage.
SPICE: Acronym for Simulation Program with Integrated Circuit Emphasis. It is a very sophisti
cated software tool for simulating electronic circuit behavior.
Virtual ground: When an ideal op amp has one of its input terminals grounded, and is operat
ing in the active region, then the other input terminal is also held at the ground poten-
tial because of the virtual short effect (see below). Such a condition is called a virtual ^
ground (in contrast to a physical ground).
Virtual short circuit: When an ideal op amp is operating in the active region, the voltage across
the two input terminals is zero, even though the two terminals are not hard-wired togeth
er. Such a condition is called a virtual short circuit (in contrast to a physical short circuit).
Voltage follower: A voltage-controlled voltage source with gain equal to 1, often utilized to sep
arate stages of amplification in a multi-stage amplifier device.
^The circuit proposed in this section is a modification of one proposed in W. J. Kerwin, L. P. Huesman, and R.
W. Newcomb, “State-Variable Synthesis for Insensitive Integrated Circuit Transfer Functions,” IEEE Jr.
of Solid State Circuits, Vol. SC-2, pp. 87-92, Sept. 1967. The modification consists of an additional
resistor, which greatly simplifies the design calculations and was published by P. M. Lin as “Simple /—
Design Procedure for a General Summer,” Electron. Eng., vol. 57, no. 708, pp. 37-38, Dec. 1985. '
^ See Linear Circuit Analysis by DeCarlo and Lin, 2nd edition, New York: Oxford University Press, 2002.
^ Any of the SPICE or PSPICE software programs available by a variety of vendors will suffice to obtain the indi
cated curve.
^ An op amp and a comparator as seen in a parts catalog are essentially the same, except that the comparator device
has a modified output stage that makes it compatible with digital circuits.
o
Chapter 4 • T h e Operational Amplifier 179
Problems
ANALYSIS USING IDEAL OP
AM P M O D EL
1. Consider the inverting amplifier circuit o f
Figure P 4 .1, in which = 4 V.
(a) If = 2 kD, find i ?2 that the power Figure 4.3
dehvered to = 100 is 4 W.
(b) Now suppose T(’2 = 12 kQ. Find so 4. In the circuits o f Figure P4.4, a source is rep
that the power delivered to Rj^ = 2 Id l resented by an ideal voltage source, v-J^t) = 4 V,
is 450 mW. Then find the power con
in series with R^ = 10 Q resistor. The loadin
sumed in i?, and Rj-
both cases is = 40 Q.
(a) W ith the load connected directly to
the source, as shown in Figure P4.4(a),
find the load voltage, the load current,
the source current, and the power
delivered to the load.
(b) As in Figure 4.4(b), a buffer amplifier
Figure P 4.1 separates the source and the load.
Check: (b) 1500 <R^< 2000, and = 10 m W Again, find the load voltage, the load
current, the source current, and the
2. Consider the non-inverting circuit o f Figure power delivered by the op amp to the
P 4.2, in which v- = 4 V. load.
(a) If R^ = 2 kQ, find 7?2 so that the
power delivered to = 100 Q is 4 W.
(b) Now suppose i ?2 = 13 kO. Find R-^ so +
that the power delivered to = 2 kH Rs
V,
' l >r
is 450 mW. Then find the power con-
■6
sumed in R-^ and i?2-
(a)
(b)
Figure P4.4
SCRAMBLED ANSWERS: (a) 0.256, 3.2, 0.08
computed in part (a) because the 80- A N SW ER: (a) R, = 5 kQ; (b) 6 .76 m W
£2-240 Q resistor combination is iso
lated from the 320-Q. resistor.
7. (a) For the op amp circuit o f Figure
(c) For the circuit o f Figiire P4.5(c), again
P4.7(a), find as a function
compute v-^ and in terms o f v^. Your
and the R^.
answers should be the same as those in
(b) Repeat part (a) for the circuit o f Figure
part (b). The buffering op amp circuit
P4.7(b).
again isolates the 80-£l-240 Q. resistor
(c) If for Figure 4.7(b), 12 kD, =
comination from the 320 Q resistor.
3 kn , = 4 kQ, Rj = 1 k£2, = 1.5
80 n 80 n V, and v^ 2 = 2 V, find the power deliv
-o
+ ered to the load R^= 100 Q.
V r + ^ <,v, 240 AN SW ER: (c) 0.04 watts
320 n<
(a)
son 80 n
-o
+
240 n
320 n<
(b)
80 0
son -v \ ^ - -o
+
240 n
320 n
Figure P4.5
AN SW ER: (a)-V / 0 5V
3
6. In the circuit below, R^= 10 kQ.
(a) Find R^ and R^ so that = -2t^^j -
">^a- 8. (a) For the circuit o f Figure P 4.8, suppose
(b) Given correct answers to part (a), sup out
= 6 kQ, and find R^ so that
pose a 1 kO resistor is attached as a load.
Find the power delivered to the load if = 20. If = 0.6, find the power deliv
= 200 mV and = “ 600 mV. ered to the 8 -Q load.
(b) Now suppose R^^'^ kQ, and find R^
out
so that = 20
out
their common value so that = 20
F igu re P 4 .6
Chapter 4 • T h e Operational Amphfier 181
Figure P4.12
12R
R, 2R 5R
-V S/V
r e -O
3R _ d +
(a)
Figure P 4 .15
C H EC K : (a)
12R
5R
2R 16. For the circuit o f Figure P 4.16, find in
rO -o
+ terms o f V, v^2
2.5R
(b)
Figure P4.13
AN SW ERS; (a) 6R-, (b) 3R
4R
2R 1.5R
0.5 R
-o-
+
0.75R
0 - . ^
Figure P4.14
C H EC K : i = 0.9 mW, = 0-20667 Figure 4.17
CH ECK: = 8R
watts
15. For the circuit o f Figure P4.15, the input volt 18. For the circuit o f Figure P 4.18, find R^ and
age = 5 V, and the input voltage i?2> and Rj in terms o f R so that = 8f^j +
(a) If = 8 R and Rj - Ry find 10v^2~2^s3- Hint: Consider Problem 17 first.
(b) If R^ = 8R and R^ = ARy find
(c) If /?2 = find in terms o f R so
that = lOV.
Chapter 4 • T h e Operational Amphfier 183
Time in s
(0
AN SW ERS: = AR, R^ = R, = O m
Figure P4.19
(a) 4R
1.5R
-O
2R
Figure P4.21
AN SW ER: (a) 15 V; (b) = 3.75 V
(b) For the circuit o f Figure P4.22(b), (b) For the circuit o f Figure P4.24(b),
compute the power delivered by compute the power delivered by
the source, and power delivered to the the source, and power delivered to the
load in terms o f v-^^. load in terms o f
(c) Discuss the differences in your solu (c) Discuss the differences in your solu
tions to (a) and (b). Specifically, dis tions to (a) and (b). Specifically, dis
cuss the effect o f using a voltage fol cuss the effect o f using a voltage fol
lower to isolate portions o f the circuit. lower to isolate portions o f the circuit.
(a) (a)
(b)
Figure P4.24
AN SW ERS: (a) = 0.003i^y„; (b) Pl = 27. Two non-ideal voltage sources are each rep
3.7037 x l O 'V . resented by a connection o f a (grounded) inde
pendent voltage source and a series resistor.
26. Figure P 4.26 contains three circuits that Denote the parameters o f each connection by
explore loading and the elimination o f loading {Vs^, ^st)' Design an op amp cir
effects using either a dependent source or an cuit such that the output voltage with respect
equivalent buffering op amp circuit. to ground is
(a) For the circuit o f Figure P4.26(a),
compute I'j and Observe that the
8-Q -24 Q. resistor combination loads
down the 3 2 -Q resistor. for all values o f and R^2 - tie greater
(b) For the circuit o f Figure P4.26(b), than or equal to 100 kS2 so that only small
compute and Notice that is amounts o f current are drawn from the buffer
different from the answer computed amplifiers. Note that the general difference
in part (a) because the 8-Q -24 Q resis amplifier circuit o f the chapter will not work
tor combination is isolated from the here because o f the presence o f the resistances
3 2 -Q resistor. and R^2 - To achieve such a design, it is nec
(c) For the circuit o f Figure P4.26(c), essary to isolate the (practical) sources from the
again compute Pj and Your difference amplifier inputs using buffer ampli
answers should be the same as those in fiers, as shown in Figure P 4.27. Explore your
part (b). The buffering op amp circuit design for various values o f and R^2 using
again isolates the 8-Q -24 O, resistor SPICE. Do the SPIC E simulations verify that
combination from the 32 Q resistor. the output is independent o f the values o f
Figure P4.27
-o
8Q + 28. Following Example 4.9, for the comparator
80
circuit o f Figure P 4.28, suppose the op amp has
24 n infinite gain and a saturation voltage =15
32 0 <
V. Find the versus relationship and plot
as a function o f v-^. Verify your analysis using
(c) SPICE. Assume that the op amp is a type 741
whose model should be available within your
Figure P4.26 SPIC E program.
AN SW ERS: (a) 0.6665 V^, 0.5 V^; (b) and (c)
0.8 V^, 0.6
186 Chapter 4 • T h e Operational Amplifier
29. (a) Find the versus relationship for Vin > , then and when
the comparator circuit o f Figure
P4.29. Specifically, show that when ^in<-^^ref , then =
R,
Vin <
Rf Rf
V l-
R, Ro
Figure P4.29
(b) Suppose 25 kO. Find , i?2>
AN SW ERS: (b) = 15 V if < 2 V, and
so that is the negative o f the
average o f , V2 , and Vy
34. Using the topology o f Figure 4.13, design 38. Using the topology o f Figure 4.13, design
an op amp circuit to have the input-output an op amp circuit to have the input-output
relationship relationship
- 3^.2 + +H i +H i
Two different designs are to be produced for
Two different designs are to be produced for comparison and selection:
comparison and selection: 1: . = 50 kQ.
(a) Design 1: Rjr= 100 kQ. ^ iqq
(b) Design 2: Rj^= 50 kQ. Specify all final ^
values in terms o f Q.. 39. Generalizing the topology o f Figure 4.13,
C H ECK : = 25 kO, and R^2fiil ^ 33.33 k£2 design an op amp circuit to have the input-out-
put relationship
35. Using the topology o f Figure 4.13, design
an op amp circuit to have the input-output ^out ~^a\ ~ '^'^al ~
^out ~ ~^^a\ ~ '^^al ‘^^h\ ^'^bl 40. Generalizing the topology o f Figure 4.13,
design an op amp circuit to have the input-out-
Two different designs are to be produced for put relationship
comparison and selection:
out
= -4v^j + 2z;^j +
(a) Design 1: Rr= 100 k£l. , ^ , n / r^
(b) Design 2: R^= 50 k tl. Specify all final ^/=
values in terms o f Q.
VARIABLE GAIN AM PLIFIERS
36. Using the topology o f Figure 4.13, design 4 1x h e circuit o f Figure P4.41 is a modifica-
an op amp circuit to have the input-output non o f the basic non-inverting amplifier. In the
relationship modification, a potentiometer R^ is connected
between the output terminal and Rq, with the
^out ^ ~ '^'^al ^ ^'^bl sliding contact between points A and B, as
shown. Show that as the sliding contact o f the
Your design must have Rj-= 10 kX2 in the final potentiometer is moved between positions A
circuit, and all other resistors should be within and B, the range o f voltage gain achievable is
the range 2 kQ to 20 kQ. P
^out
37. Using the topology o f Figure 4.13, design
an op amp circuit to have the input-output
relationship
■o
10 kn
In the mid-nineteenth century, before the introduction o f the alternating current (ac), electricity
was available mainly as direct current (dc). This time period saw the evolution o f basic laws for
the analysis o f electrical circuits composed o f dc voltage sources and resistors: Ohm’s law, KVL,
and KCL. Application o f these laws to the analysis o f circuits led to the development o f the mesh
and nodal techniques requiring the solution o f simultaneous equations. Before the computer age,
manual solution o f a (large) set o f equations was very difficult. To circumvent this difficulty,
researchers developed a number o f network theorems that (i) simplified the aforementioned man
ual analysis, (ii) reduced the need for repeated solution o f the same set o f equations, and (iii) pro
vided insight into the behavior o f circuits. These network theorems remain useful even in the pres
ent day o f high-powered computing.
SECTIO N HEADIN GS
1. Introduction
2. Linearity
3. Linearity Revisited: Superposition and Proportionality
4. Source Transformations
5. Equivalent Networks
6. Summary
192 Chapter 5 * Linearity, Superposition, and Source Transformation
1. IN TRO D U CTIO N
Chapter 3 covered nodal and loop/mesh analyses. Node voltage or loop current calculation pro
ceeds by constructing a set o f simultaneous node or loop equations and solving them by hand, by
MATLAB, or with some equivalent software package. Few o f us will attempt a paper-and-pencil
solution o f four equations in four unknowns. Yet, MATLAB, Mathematica, or some other com
putational software program, can easily and reliably crunch numbers, relieving us o f tedious hand
calculations. Nevertheless, manual analysis in some form remains important for a deeper under
standing or insight into a circuit’s behavior, as well as a way to check the validity o f a program
output.
Experience teaches us that manual analysis is ordinarily practical only for small circuits.
Fortunately, the network theorems studied in this chapter and the next can often reduce seemingly
complex circuits to simpler ones amenable to manual analysis. They also provide shortcuts for
computing outputs and allow us to obtain deeper insights into a circuit’s behavior.
This chapter talks about linearity and superposition, which are motivated by the following ques
tions: What is the effect on the circuit output (voltage or current) o f a single independent voltage
source, say acting alone. “Acting alone” means that the independent source, Vj^, has a nonzero
value, while all other independent sources are set to zero. A deactivated voltage source acts as short
circuit (see Chapter 2), and a deactivated current source acts as an open circuit (again, see Chapter
2). Is there a shortcut to computing the response if Vj^ is doubled in value?
To answer the above questions and others, our discussion begins with the important property o f
linearity. Linearity relates the values o f independent sources to a circuit output with a very com
pact equation. This equation defines the effect o f any independent source on a circuit output.
After studying linearity, we discuss two special consequences called superposition and propor
tionality. Each o f these concepts helps reduce manual computation o f responses, and each provides
insight into circuit behavior. Next, we state the source transformation theorem and show how this
method can reduce a complex circuit to a more simple form. Finally, we set forth the notion o f an
equivalent two-terminal network and then outline a proof o f the source transformation theorem.
Chapter 5 * Linearity, Superposition, and Source Transformation 193
2. LIN EARITY
This section investigates the circuit property o f linearity, which we introduce with a motivating
example.
E X A M P L E 5 .1 . For the circuit o f Figure 5.1, find the outputs and Vg in terms o f the source
current /^j, and the source voltage V^2 - will derive the relationships Vg = 40/^j + ^ 1^2 and
A 3 jgQ .sz
I.
+ 60 0
V„ 120Q
So l u t io n
Step 1. Find the voltage Vg. A node equation at the top o f the current source is
Vb , V B - V s l ^ r
120 60
= 4 0 7 ,1 + - V ,2
Here, Vg appears as a constant times /^j, plus another constant times K^2>^ so-called linear com
bination.
Step 2. Find the current I From Step 1, we know Vg. The current satisfies
40 / , i + - V , 2 - V ,2
3 180 •'
Similar to Step 1, the output current is a constant times /^j plus another constant times V^2>^
linear combination.
194 Chapter 5 ' Linearity, Superposition, and Source Transformation
Exercise. 1. In Example 5.1, suppose the 60 Q resistor is changed to 120 Q. Find the outputs
and Vg in terms o f the sources, /^j and V^2 -
AN SW ER: Vg = 60/ ,, + 0 .5 V ,2 and = 0 .5 / ,, -
In the above example and exercises, the desired output voltage or current was a so-called linear
combination o f the independent source values. This is, in fact, a quite general phenomena, as indi
cated by the linearity theorem below.
LIN EA R ITY TH EO R EM
For all practical linear resistive circuits, as per Figure 5.3, any output voltage, or any cur
rent, ig, can be related linearly to the independent source values, as in the following equa
tions:
V
Linear Circuit
containing
no
independent
sources.
V
+ ■
FIG U RE 5.3. A linear circuit driven by n independent voltage sources and m independent current
sources with outputs of and
A rigorous proof o f the linearity theorem entails solving a set o f modified nodal or loop equations
using matrix algebra and is beyond the scope o f this text.
EXAM PLE 5.2. For the circuit o f Figure 5.4, our objective in this example is to express lin
ear combination o f /^j, Iq, and as per equation 5.1a. In doing this, we review nodal analysis.
g .v .
)v .
FIG U RE 5.4
So l u t io n
Step 1. Write nodal equation at A. For node A,
(5 .2 )
196 Chapter 5 • Linearity, Superposition, and Source Transformation
or equivalently,
= (5.3)
Step 3. Write equations 5.2 and 5.3 in matrix form. In matrix form, the nodal equations are
G ^ -g^ 0 ■ ■ ■ /.I •
8m G 2+ G 3 Vou t, J s l + G ^ V ,, (5.4)
Step 4 . Solve equation 5.4. Solving equation 5.4 for and yields
-1
-V a ■ 'G l- g m 0 - h i
Vou t, 8m G 2 + G 3
1 G2 + G 3 0 ‘ s\
{G :-gJ(G 2+ G ,) -8 m G, - I fs2+G3V,3
It follows that
8m Gi-g. jGl-8m)G3
Vout-- hl+- -V.s3
(G i-g ,)(G 2 + G 3 ) (G ,-g „,)(G 2+ G3) (G i-g ,„)(G 2+ G 3) (5 3 )
2. Suppose the dependent current source in Figure 5.4 is changed from^^V^ Sm^ouf Compute
the new expression for if G j = 1 S, G2 = 2 S, G3 = 3 S, and = 5 S.
AN SW ER: V ;„,= 0.17,2+ 0 .31/3
3. Suppose the dependent current source in Figure 5.4 is changed from to + ^0 ^)-
Compute the new expression for if G j = 1 S, G2 = 2 S, G3 = 3 S, and = 5 S.
E X A M P L E 5 .3 . A linear resistive circuit has two inputs and z,2 with output as shown in
Figure 5.5. Rows 1 and 2 o f Table 5.1 list the results o f two sets o f measurements taken in a lab
oratory. The measurements are taken in a practical way by first setting the value o f the current
ch a p te r 5 " Linearity, Superposition, and Source Transformation 197
source to zero, i.e., i^ 2 = 0 exciting with a dc power supply set to 5 V; then the voltage
source is removed and replaced by a short circuit using a jumper cable, i.e., = 0, and the cur
rent source is excited by a power supply producing a constant current o f = 0-2 A.
(a) Derive the linear relationship + 50/^2 using the data in Table 5.1.
(b) Find when = 10 V and ^'.2 = 0.5 A, i.e., complete the third row o f Table 5.1.
5 0 4
0 0.2 10
10 0.5
So l u t io n
From the linearity equation 5.1(a),
^out = +hhl
for appropriate ttj and ^2 - From the data in rows 1 and 2 o f Table 5.1,
So from row 3 o f Table 5.1, if i',] = 10 V and z'^2 = 0.5 A, we have that
0 .8 X 10 + 5 0 X 0 .5 = 33 V.
198 Chapter 5 • Linearity, Superposition, and Source Transformation
2. Suppose the data in row 1, column 3, o f Table 5.1 is changed to 10 V. Find y^^^when = 50
V and = 0.4 A.
AN SW ER: 120 V
Comparing the development o f equation 5.7 in Example 5.3 with equation 5.1 suggests that the
coefficients Qj and P 2 can be defined as ratios:
''out
«1 = and 132= —
I,-,v2
Example 5.3 and these equations suggest the algorithm for finding the coefficients in equation 5.1
by setting all inputs to zero except the input associated with the desired coefficient. This approach
is sometimes impractical. It is not always possible to set an independent source voltage or current
source to zero: imagine turning off a generator for downtown Manhattan to obtain a coefficient.
T he following example illustrates an alternate approach.
E X A M P L E 5 .4 . Consider Figure 5.6, which has two inputs and output Table 5.2
lists measurement data taken in a laboratory. Row 1 ofTable 5.2 lists the nominal operating con
ditions o f the circuit. Rows 2 and 3 illustrate measurements in which one source has its value only
slightly changed (although the change may be arbitrary) while keeping the other source value the
same. From the linearity theorem, we know + ^2^s2- Compute and P j, and then
find to complete row 4 ofTable 5.2.
5 0.25 -1
5+0.1 0.25 - 1 .0 3
5 0.25+ 0.05 - 0 .9
15 0.5 ????
So l u t io n
From rows 1 and 2 o f Table 5.2,
0.1 = p 2 x 0 .0 5 ^ Pa = 2
Equation 5.1 for the given data has the linear form
(5.9)
2,„, = - 0 . 3 x 1 5 + 2 x 0 . 5 = - 3 .5 V
200 Chapter 3 * Linearity, Superposition, and Source Transformation
20 100 15
22 100 15.9
20 110 15.6
28 80 ???
AN SW ER: 17.4 A
As a final comment on linearity, we note that by simply using the data o f rows 1 and 2 o f Table
5.2, one can solve for the coefficients by solving simultaneous equations. Specifically, using the
data o f rows 1 and 2 ofTable 5.2, we have the following matrix equation
■5 0 .2 5 ' ■ «i’ ■ -1 ■
5.1 0.25 -1 .0 3
.^1.
Exercise. Find the unknown entry in Table 5.4 after finding a , and (3j in the equation
+ Pl^.2-
10 100 15
20 100 20
30 150 ???
Here, j denotes an output, whether it be current or voltage, and each denotes a source input,
whether it be voltage or current. A special consequence o f the linearity principle is the superpo
sition property. Equation 5.10 says that the total response jy is the sum o f the responses '"aju".
Each “a-u” is the response o f the circuit to u- acting alone, i.e., when all other independent
sources are set to zero. Although implied by linearity, this property is so important that we single
it out.
EXA M PLE 5 .5 . A linear resistive circuit has two inputs and with output as shown in
Figure 5.7, where = 2 Q,, Rj = 2.5 £2, and R^ = 10 Q. Find by the principle o f superposi
tion. Then, compute the power absorbed by the 10 resistor. We show that =
0.5 V^l + 0.4V^2’ where is the contribution o f the source acting alone for k = 1, 2.
So l u t io n
Step 1. Fini/ the contribution to due only to V^j. Denote this contribution by W ith V^2
= 0, the equivalent circuit is shown in Figure 5.8(a). Here, the 2.5 Q and 10 Q, resistors are in par
allel, yielding an equivalent resistance o f 2 = 2.5 x 10/12.5 f i. By voltage division,
V -
2 +2
202 Chapter 5 • Linearity, Superposition, and Source Transformation
Vo^ut
2.5 + -
3
(youtT
Pr3 = = 0 . 1(0.5 + 0.4V,2 f = 0 . 1(o .2 5 v /i + 0 .2 V ,iy ,2 + 0 . 16V / 2 )
Note that the total power, Pj^^, is not the sum o f the powers due to each source acting alone
because o f the presence o f the cross product term. Hence, in general, superposition does not apply
to the calculation o f power.
For dc circuit analysis, the principle o f superposition does NOT apply to power calculations.
Exercise. Reconsider Figure 5.7 in which = 2 Q , 7?2 = 4 Q, and R^ = A Q.. Find by the
principle o f superposition.
AN SW ER; 0.5K^, + 0.25 2
Chapter 5 • Linearity, Superposition, and Source Transformation 203
The next example adds a controlled source to the circuit o f Figure 5.7 and repeats the superposi
tion analysis.
q V.A
So l u t io n
Step 1. Compute the contribution due only to Setting v^ 2 ~ ^ leads to the circuit o f Figure 5.10,
where we note that
0 - 5 ( - L - " . l ) + (0-2 + 0 . 1 + 0 .2 ) .i „ ,= 0
Therefore,
d r = 0 - 5 ^ .i
Step 2. Compute the contributions due only to Setting = 0 in Figure 5.9 leads to the circuit
o f Figure 5.11, where this time, = ^out~
204 Chapter 5 • Linearity, Superposition, and Source Transformation
Step 3. Using superposition, add up the contributions due to each independent source acting
alone.
^out = + 0-4 ^',2 (5.11)
The above examples used voltage division and superposition to compute an output voltage due to
two independent voltage sources.
E X A M PLE 5 .7 . This example illustrates the principle o f superposition for the three-input op
amp circuit o f Figure 5. 12 . Show that + 2 . 5 K^2 +
So l u t io n
Step 1. Find the contribution to due only to Denote this output by W ith V^ 2 = ^ 3
= 0, the circuit o f Figure 5.12 reduces to that o f Figure 5.13(a). The properties o f an ideal op amp
ensure that i^ = 0, making = -0.5i?z^ = 0. Thus, v_ = v^=Q implies
V’i r = - — K . = - 4 V ,
2R
(a)
FIG U RE 5.13
Step 2. Find the contribution to due only to V^2 - W ith = 0, the equivalent circuit is
shown in Figure 5.13(b) where we denote the output as From op amp properties and volt
age division,
^ -V ,2 = 0 .5 V ,2
R+R
Hence, from Example 4.3,
Step 3. Find the contribution to due only to T he equivalent circuit in this case is the same
as that o f Figure 5.13(b) with V^ 2 replaced by Therefore, the output due to source acting
alone is
The above examples have generated the linearity formula, equation 5.1, using superposition, i.e., the
response o f a circuit is the sum o f the responses due to each source acting alone. The technique is equiv
alent to that described in Example 5.3. However, superposition alone is not equivalent to linearity.
Linearity is equivalent to the properties o f superposirion AND proportionality, which is now stated.
However, for dc analysis, the proportionality property does N O T apply for power calculations.
If . K, and . V„ . 0, then V ^ . - 4 (« V „ ) . « - 4 K „ ) -
Exercises. For certain nonlinear circuits, the principle o f superposition may be satisfied, but pro
portionality not satisfied, or vice versa. This exercise explores these distinctions.
1. If a circuit has input-output relationship sV ^^^ow that the principle o f
superposition is satisfied, but proportionality is not satisfied.
2. If a circuit has input-output relationship = a,Wjj + ^^ow that the
principle o f proportionality is satisfied, but superposition is not satisfied.
A very interesting and significant application o f the proportionality property occurs in the analy
sis o f a resistive ladder network. A resistive ladder netw ork is one having the patterned structure
shown in Figure 5.14, where each box represents a resistor.
-H
v.Q V
neous equations, as follows: assume = 1 V. We can sequentially compute currents and voltages
in a backwards fashion to obtain the required source value to yield = \ V. Suppose we call this
voltage Define K = to be the proportionality constant, where is the actual source
E X A M P L E 5 .8 . Find all the voltages V^-, i = 1, ..., 6 in the resistive ladder network o f Figure 5.15.
So l u t io n
Assume Vj = 1 V. Repeatedly apply Ohm’s law, KCL, and KVL as follows: (Q, V and A are used
throughout):
(Ohm’s law)
/2 = / i = 0 .1
(KCL)
V3 = + V2 = 2 (KVL)
Vg = \/3 + y4 = 5 (KVL)
y, = y5 + y6 = io (KVL)
We conclude that if Vj = 1 V, the source voltage must be = 10 V. But the actual source voltage is
50
50 V. Define K = = — = 5 . By the proportionality property, if = 50 V, then - I^x 1 =
5 V Similarly, - 5 V, = W V, = 15 V, and V5 = 25 V.
208 Chapter 5 • Linearity, Superposition, and Source Transformation
In the solution given above, we have separated the expressions into calculation blocks to empha
size the repetitive pattern. For example, the expressions in block #3 are simply obtained from
block #2 by increasing all subscripts by 2. When the ladder network has more elements, the
sequence o f expressions contains more blocks, each o f which entails two additions and two mul
tiplications. This method then allows us to straightforwardly solve ladder networks o f any size
without writing or solving simultaneous equations.
Exercise. In Example 5.7, change all resistances to 2 Q and find V^. Would it make any differ
ence in the voltage Kj if all the resistors were changed to R ohms?
50
AN SW ER: Vi = — = 3.85 V, and no difference.
' 13
The words “source transformation” refer to the conversion o f a voltage source in series with an R-
ohm resistor to a current source in parallel with an R-ohm resistor, and/or vice versa. This section
explains the details o f such transformations and how they can simplify analysis. But first we must
recall from Chapter 2 that voltage sources in series add together (such as batteries added to a flash
light) and that current sources in parallel combine into an equivalent single current source. This
is illustrated for multiple voltage sources in series in Figure 5.16. Similarly, Figure 5.17 shows how
multiple current sources combine into a single source.
4.5 V
FIG U RE 5.16 (a) Three voltage sources in series; (b) equivalent single voltage source.
s,eq
V =RL
A justification for the source transformation theorem will be given in the next section. Practically
speaking, it can save significant computational effort. For example, in the circuit o f Figure 5.19 in
Example 5.9 below, a solution approach using mesh analysis requires writing and solving three
simultaneous equations. Nodal analysis at A and B requires writing and solving two simultaneous
equations. Applying the source transformation theorem is a third avenue that avoids all simultane
ous equations.
5 kO 6kQ
50 V
So l u t io n
Step 1. Substitute all series V^- R combinations by their parallel - R equivalents, where in each
case, /^ = — . Applying the source transformation theorem four times results in Figure 5.20.
R
10 mA 10 mA
FIG U RE 5.20 Circuit equivalent to that o f Figure 5.19 by source transformation theorem.
Step 2. Combine the parallel resistances and the parallel current sources.
To the left o f point A are two independent current sources and two resistors, all in parallel.
Similarly, to the right o f B are two current sources and two resistors in parallel.
Combining current sources and resistors to the left o f A results in a single current source o f 5 mA
directed upward and an equivalent resistance o f 4 kD. To the right o f B, the current sources can
cel each other out, and the equivalent resistance is 2 kXl. T he resulting simplified circuit is shown
in Figure 5.21.
5 mA 10 mA
Step 3. Apply the source transformation theorem a second time to each o f the - R pairs. These par
allel I ; - R pairs become series V^- R pairs, as illustrated in Figure 5.22.
4kO
---
A 4kQ B
20 V 20 V
20 + 20
^AB - = 4 mA
4+4+2
Thus, = 4 0 0 0 X (0 .0 0 4 )2 ^ <54
Exercises. 1. For the circuit o f Figure 5.23(a), /^j = 50 mA and = 500 Q. Convert the parallel
R combination to a series V^- R combination, where = ? and = ? in Figure 5.23(b).
AN SW ER: 25 V, 500 Q
A f^series
-o -
Circuit V. r + Circuit
B B
-O ■O
(a) (b)
FIG U RE 5.23
2. For the circuit o f Figure 5.24(a), /^j = 50 mA and R^ = 500 Q , while I ^ 2 = 1 5 0 mA and R^ =
300 Q . Convert the two parallel 1^—R combination to a single series V^—R combination, where
= ? and = ? in Figure 5.24(b).
AN SW ER: - 2 0 V, 800 Q
A A
ho- -o -
Circuit V Circuit
B B
-O -O -
(a) (b)
FIG U RE 5.24
3. Consider the circuit in Figure 5.25(a). Using a source transformation and resistance combina
tions, determine the values o f and Rp^^^ in Figure 5.25(b).
212 Chapter 5 * Linearity, Superposition, and Source Transformation
100Q 100Q
FIGURE 5.25
AN SW ER: 150 LX 50 niA
The source transformation theorem above is based on the notion o f equivalent networks, as is the
material o f the next chapter. So we now explore a precise understanding o f equivalent 2-tenninal net
works. Figure 5.26 illustrates four 2-terminal networks, all enclosed by dashed boxes, labeled N j,
N 2, Ng, and N^. Their characteristic is that there are only two accessible nodes for connection to other
circuits. Note however that any controlling voltage or current must be contained within the dashed-
line box. Such dashed-line boxes are often omitted to avoid cluttering in circuit diagrams.
FIGURE. 5.26 Examples of 2-terminal networks, i.e., networks in which only two terminals are
available for connection to other networks.
Observe that networks N j and N 2 in Figure 5.26(a) and (b) have the same terminal characteris
tics: at the terminals o f N j, the v - i characteristic is
v = 2i+ 10
Chapter 5 • Linearity, Superposition, and Source Transformation 213
i= - - 5 v = 2( + 10
2
The two equations are identical. We then say that a pair o f 2-terminal networks are equivalent if
they have the same terminal characteristics. Therefore, and N 2 are equivalent.
Now, observe that networks and are also equivalent to A^j and A^2- To see this, note that
for N y
V v-1 5
6/ = 3 v - 3 0 v = 2i + 10
And for N^, first observe that i = \0i^ from KCL, in which case = O.lz; further, from KVL,
10 + 20/'a ^ v = 2 i + 10
as was to be shown.
Because equivalent 2-terminal networks have the same terminal v —i characteristic, if one network
is interchanged with its equivalent, all currents and voltages outside the box remain the same as
illustrated in Figure 5.27; i.e., all voltages and currents in the “rest o f the circuit” are the same as
before.
A
■o
N +
Rest of
N.1 V
Circuit
J
-O -
B
FIG U RE 5.27 The networks denoted N^, i = 1,2, are equivalent when the v-i values at the terminals
are identical; logically then, all voltages and currents inside the “Rest o f Circuit” remain the same.
These examples allow us to justify the source transformation theorem as follows. Both 2-terminal
networks in Figure 5.18 have the same v - i relationship: v = Ri + and
v = Ri + Rlj, = Ri + .
Therefore, the two networks o f Figure 5.18 are equivalent, and the source transformation is a valid
analysis technique.
214 Chapter 5 * Linearity, Superposition, and Source Transformation
6. SUM M ARY
This chapter covers the notions o f Unearity, superposition, proportionaHty, and source transfor
mations. Linearity states that for any hnear resistive circuit, any output voltage or current, denot
ed as y, is related to the independent sources by the formula j = + ... + where through
are the voltage and current values o f the independent sources, and through are appropri
ate constants. O nce values for the are known, one can compute the output for any (new) set o f
input values without having to resolve the circuit equations, a tremendous savings in time and
effort. A special consequence o f linearity is the widely used principle o f superposition.
Superposition means that in any linear resistive circuit containing more than one independent
source, any output (voltage or current) can be calculated by adding together the contributions due
to each independent source acting alone with the remaining independent source values set to zero.
Practically speaking, this is the customary path to computing the coefficients, a^, in the linearity
formula.
Proportionality, another consequence o f linearity, means that if a single input is scaled by a constant,
with the other inputs set to zero, then the output is scaled by the same constant. This property led
to a clever technique for analyzing ladder networks without writing simultaneous equations.
Using the notion o f an equivalent 2-terminal network, the chapter set forth the theorems on
source transformations for source-resistor combinations: a 2-terminal network consisting o f a
series connection o f an independent voltage source and a nonzero finite resistance R is equiv
alent to a 2-terminal network consisting o f an independent current source, = V^IR, in parallel
with R, as illustrated in Figure 5.18. These transformations, applied multiple times to a circuit,
often simplify the analysis o f a circuit.
Chapter 5 • Linearity, Superposition, and Source Transformation 215
2-term inal network: an interconnection o f circuit elements inside a box having only 2 accessible
terminals for connection to other networks. T he concept is extendible to n-terminal net
works.
Equivalent 2-term inal networks: two 2-terminal networks having the same terminal voltage-cur
rent relationship. I f two 2-terminal networks and N 2 are equivalent, then one can be
substituted for the other without affecting the voltages and currents in any attached net
work.
Linearity property: let the responses due to inputs and Uj, each acting alone, be y-^ and y 2 -
W hen the scaled inputs and . U2 are applied simultaneously, the response is =
0 2
+ ®2^2- Linearity implies both superposition and proportionality, and vice versa.
Linear resistive element: a 2-terminal circuit element whose terminal voltage and current rela
tionships is described by Ohm’s law.
Linear resistive circuit/network: a network consisting o f linear resistive elements, independent
voltage and current sources, op amps, and controlled sources.
Proportionality property: when an input to a linear resistive network is acting alone, multiply
ing the input by a constant, K, implies that the response is multiplied by K.
Source transformation: a 2-terminal network consisting o f an independent voltage source in
series with a resistance is equivalent to another 2-terminal network consisting o f an inde
pendent current source in parallel with a resistance o f the same value.
Superposition property: when a number o f inputs are applied to a linear resistive network simul
taneously, the response is the sum o f the responses due to each input acting alone.
216 Chapter 5 • Linearity, Superposition, and Source Transformation
Figure P5.1
A N SW ER: (b) 8 cos(lO^) - 8 V
Figure P5.2
G,
— <+ i;
I .Q
>R, © '
C H EC K : (a) ????? I/, + 16/^2’ 39 m W > 10. A linear resistive circuit has two independent
sources, as shown in Figure P5.10. If = 0
with v^2 ^t) = 10cos(2 ?) V, then = 20 cos(2 i)
8. Consider the circuit o f Figure P5.8. in which V. On the other hand, if = 10cos(2^) mA
Ry = 18 Q. =9 ^3 = 18 Ra = 36 Q, with v^^{) = 0, then = 2 cos(2z-) V. Find the
and R^= 18 D. linear relationship between and the inputs,
(a) Find the linear relationship between and Now compute when
out
it) and the four independent sources. = 20cos(2/) mA and v^2 (^) = 20 V.
(b) (Challenge) If each o f the resistances
is doubled, what is the new linear rela
tionship. (Reason your way to the
answer without having to resolve the
circuit. Hints: Investigate the effect o f
changing the resistance in Ohm’s law Linear
resistive circuit
for fixed current. Investigate the effect with
o f equal changes in all resistances on a dependent sources
Figure P5.10
13. Consider the linear network o f Figure (a) Compute the coeiTicients o f a linear
P 5.13, which contains, at most, resistors and relationship among the output and
linear dependent sources. Measurement data is three inputs.
given in Table P5.13. (b) If 7,1 = - 1 A, 1/2 = 4 0 V, and I/3 = 10
(a) Find the linear relationship = V, find the power absorbed by R^.
AN SW ER: 16 W
(b) Find the power consumed by the 10
resistor when = 20 V and = 500 mA
Resistive Circuit
witli
Dependent
Sources
+ V. . -
Figure P 5.I3
I',, (volts) z'^2 (amperes) Km 16. Again consider Figure P5.15. Suppose the
5 0.4 -1 data measurements are given in Table P5.16.
10 1 2 Table P5.16
14. Reconsider Figure P5.13 Two separate dc /j, (mA) ^.2 (V) (V)
measurements are taken. In the first experi
ment, = 7 V and hi - 3 A, yielding =1 Case 1 30 2 -1 11.5
A. In the second experiment, = 9 V and z'^2
= 1 A, yielding = 3 A. Case 2 40 2 -1 13
(a) Find the coefficients o f the linear rela
Case 3 30 2.2 -1 11.6
tionship + p 2i,2 -
(b) Given the equation found in part (a), Case 4 30 2 - 0 .9 11.9
compute when = 15 V and z^2
= 5 A. Case 5 40 8 10
AN SW ER: (b) 90 watts
(mA) 1^.2 (V) ^^.3 (V) 17. Again consider Figure P 5.15. Suppose the
1 data measurements are given in Table P5.17.
Case 1 50 -2 5 -1 3
Case 2 0 3 5 2
Case 3 0 2 4 0
Chapter 5 • Linearity, Superposition, and Source Transformation 219
Case 1 30 2 -1 11.5 1 6
Case 2 -2 0 4 2 27 2 10
Case 3 -1 0 -3 1 -1 4 5 ?
(b) Repeat part (a), but express your answers (a) Using superposition, find due to
in terms o f the literals R-,i= 1, 2, 3. acting alone, and then find due to
(c) Find and and the v^2 acting alone. What is ^
power delivered to when = 100 (b) Redo part (a) using the literals G/ = —
V, = 2 A, and = 4 A. Ri
(d) Repeat part (c) when is doubled, i^2 1 5 2 7
is tripled, and is halved. AN SW ER: (a) = -''.v i- '' m » = - ‘',v2
AN SW ER: (c) 100 V, 500 watts
Figure P5.22
Figure P5.20
23. For the circuit o f Figure P5.23, suppose R^ =
21. In the circuit shown in Figure P5.21, R-^ = 20 Cl, /?2 = 50 Q, = 100 Q and = 0.02 S.
180 Q, R^ = 360 a , T?3 = 90 Q, an R^ = 720 Q. (a) Using superposition, find due to
(a) Find the coefficients o f the linear rela acting alone, and then find
tionship + due to v^ 2 acting alone. W hat is
fi by superposition.
0 -2 1 A N SW ER: = 0.5j',| + 0.9i^,,
(b) Repeat part (a), but express your
answers in terms o f the literals G- = (b) Redo part (a) using the literals G,- = —
l lR- ,i = 1 , 2 , 3 , 4.
(c) Find and the power absorbed by
R^ when = 100 V and v^ 2 = 50 V.
AN SW ER: = 60 V; = 5 watts
225 n i8on
■^900 0
0 20V 0.1 A
28. (a)
Figure P5.27
For the circuit shown in Figure P 5.28,
If /j = 1 A, find /j by writing a
MATLAB program.
Figure P5.24
(b) If it is known that = 200 mA, find /j.
25. For the circuit shown in Figure P5.25, = (c) Find the equivalent resistance seen by
160 V. Find Then find the intermediate the current source.
node voltages. Hint: Assume = 1V
and use proportionality, as per Example
5.11.
C H E C K : Answer is an integer.
+ Figure P5.28
26. For the circuit shown in Figure P5.26, = 29. Consider the circuit in Figure P5.29.
64 mA. Find Hint: Assume = 1 A and (a) Find the contribution to due only
then use proportionality. to
(b) Find the contribution to due only
1n 1n 1n 1n
III “ '^^2-
o>^i (c) Find by superposition.
2n
© 1n
2n 2n r2n
2n
Figure 5.26
Figure P5.30
31. (a) For the circuit of Figure P 5.31, find Figure P5.32
V^ut^ the voltage due to each source
acting alone in terms of the literal values, 33. (a) For the circuit in Figure P 5.33, find
Figure P 5 .3 4
n
Chapter 5 • Linearity, Superposition, and Source Transformation 223
35. Consider the circuit in Figure P5.35. 37. Consider the circuit in Figure P 5.37 in
(a) Find the contribution to K^^^due only which = Q.25R^= R2 = Ry
(a) Find the contribution to due only
(b) Find the contribution to due only to
to 1/2. (b) Find the contribution to due only
(c) Find by superposition. to
(c) Find by superposition.
(d) Find the power delivered to Rj^ when
is acting alone, i.e., ®
then find the power delivered to R^
when V^2 acting alone when Rj =
R and R^ = R.
(e) Find the total power delivered to R^
when R 2 = 4 R and R^ = R.
Figure P5.37
200 n 80 40 0
0 son 20 n 40 n
Figure P5.41
Figure P5.39 AN SW ER: 13.5 and 9.1125
AN SW ER: 6 V source in series with 12 Q resis
tor 42. Use source transformations on the circuit o f
Figure P5.42, to compute the value o f need
40. Consider the circuit o f Figure P 5.40 in ed to deliver a current o f I = 0.25 A.
which /,i = 10 mA, 1^2 = 20 V, and I/3 = 80 V.
(a) Use a series o f source transformations
to find a single voltage source in series
with a resistance that is in series with
the 9.6 kQ resistor.
(b) Then find the power absorbed by the
9.6 kQ resistor.
Figure P5.42
A N SW ER: 28 V
Figure P5.45
AN SW ER: 2.8 V, - 0 .4 V
C H A P T E R
In the early days o f electricity, engineers wanted to know how much voltage or current could be
delivered to a load, such as a set o f street lamps, through a complex transmission network. Before
the days o f computer-aided circuit simulation, simplification o f complex circuits allowed engi
neers to analyze these very complex circuits manually. In 1883, a French telegraph engineer, M.
L. Thevenin, first stated that a complex (passive) network could be replaced by an equivalent cir
cuit consisting o f an independent voltage source in series with a resistor. Although stated only for
passive networks, the idea o f a Thevenin equivalent evolved to include active networks. Its wide
spread use has simplified the homework o f students for many years now and probably will con
tinue to do so for many years to come.
A more recent but quite similar idea is the Norton equivalent circuit consisting o f an independ
ent current source in parallel with a resistance. At the time o f E. L. Norton (a scientist with Bell
Laboratories), the invention o f vacuum tubes made independent current sources a realistic possi
bility. Many electronic circuits were modeled with independent and dependent current sources.
The appearance o f Norton’s equivalent circuit was a natural outcome o f advances in technology.
CHAPTER OBJECTIVES
1. Define and construct the Thevenin and Norton equivalent circuits for passive networks.
2. Define and construct the Thevenin and Norton equivalent circuits for active networks
containing dependent sources or op amps.
3. Illustrate several different techniques for constructing the Thevenin and Norton equiva
lent circuits.
4. Investigate maximum power transfer to a load using Thevenin or Norton equivalents.
228 Chapter 6 • Thevenin, N orton, and M axim um Power Transfer Theorem s
SECTIO N H EADIN GS
1. Introduction
2. Thevenin and N orton Equivalent Circuits for Linear Passive Networks
3. A General Approach to Finding Thevenin and N orton Equivalents
4. Thevenin and N orton Equivalent Circuits for Active Networks
5. Thevenin and N orton Equivalent Circuits for Op Amp Circuits
6 . Thevenin and N orton Equivalent Circuits from Measured Data
7. M aximum Power Transfer Theorem
8 . Summary
9. Terms and Concepts
10. Problems
1. IN TRO D U CTIO N
Practicing electrical engineers often want to know the power absorbed by one particular load. The
load may be a large machine in a factory or a lighting network in the electrical engineering build
ing. Simple resistances often represent such loads. Usually the load varies over time in which dif
ferent resistances are used at different times to represent the load. W hat is the effect o f this load
variation on the absorbed power and on the current drawn by the load? To simplify analysis, the
rest o f the linear network (exclusive o f the load) is replaced by a simple equivalent circuit consist
ing o f just one resistance and one independent source.
For our purposes, a (resistive) load is a two-terminal network defined in Chapter 1, meaning that
the current entering one o f the terminals equals the current leaving the other. More generally, a
two-terminal network is any circuit for which there are only two terminals available for connec
tion to other networks. (See Figure 6.1.) T he important question for our work in this chapter is:
How does one characterize a two-terminal networks As is shown in Figure 6.1(a), there is a voltage
v{t) across the terminals and a current i(t) entering one terminal and leaving the other. T he rela
tionship between the voltage v{t) and the current i{t) characterizes the two-terminal network. For
example, if v{t) = Ri{t), we would recognize the terminal network as an equivalent resistance R.
Or, if v{t) = Ri{t) + Vq, we might recognized this equation as that o f a resistance in series with a
voltage source. In fact, this equation could be represented as graph, e.g. Figure 6.1(b).
This leads to our next question: When are two 2-terminal networks equivalent'^ As developed in
Chapter 5, two 2 -terminal networks are said to be equivalent when their terminal v-i characteris
tics are the same. O f particular interest for this chapter is an equivalent network consisting o f a
voltage source in series with a resistance, called the Thevenin equivalent network, and a current
source in parallel with a resistance, called a Norton equivalent network. Figure 6.1c shows a
Thevenin equivalent for a linear resistive circuit.
Chapter 6 • Thevenin, N orton, and M axim um Power Transfer Theorems 229
N- i(t)
2-terminal —o
Linear +
V(t)
Resistive
Network
FIG U RE 6.1. (a) a 2-terminal linear network with terminal voltage v{t) and current i{t)\
(b) graphical representation of the equation v{t) = + Vq,
(c) Thevenin equivalent network having the same terminal v{t) and i{t) relationship as (b).
This chapter investigates the replacement o f a network N by its Thevenin equivalent or its Norton
equivalent. The first section describes the Thevenin and Norton equivalent theorems for passive net
works, those containing only independent sources and resistors. Following that, we generalize the
statements to include active networks. However, because op amps have peculiar properties, Thevenin
and Norton equivalents o f circuits with op amps are explored exclusively in Section 4. Following
this, in Section 5, we describe how to obtain a Thevenin or Norton equivalent from measured data
without having to know anything about the internal circuit structure. This is particularly useful
when one has equipment such as a power supply but no schematic diagram o f the internal circuit
ry. Unfortunately, not all linear devices have a well-defined Thevenin or Norton equivalent. The
homework exercises illustrate a few cases. Section 6 explores the problem o f maximum power trans
fer to a load in the context o f the Thevenin equivalent circuit, which ends the chapter.
Our first objective is to develop and illustrate the celebrated Thevenin theorem for passive net
works. Then we will state and illustrate Norton’s theorem, dual to Thevenin’s theorem.
To develop Thevenin’s theorem, consider Figure 6.2(a) consisting o f two 2-terminal networks, N
and Nj^ , joined at A and B. Only resistors and independent sources make up N, while con
230 Chapter 6 • Thevenin, N orton, and M axim um Power Transfer Theorem s
tains arbitrary even nonlinear elements. Suppose undergoes various changes as part o f an
experiment, while N, complicated in its own right, remains unchanged. To simplify repeated cal
culations, N is replaced by its Thevenin equivalent, as illustrated in Figure 6.2(b). The more sim
ple Thevenin equivalent consists o f a single voltage source, in series with a single resistance.
Rth-
f
Resistances ------ o------
+
and Arbitrary
independent Networi< r-\
Sources
V J B V J
(a)
i, = 0
N- -A N-
Resistances -o Reslstances -o
-I-
and with independent
R..
independent Sources
Sources -o Deactivated
B
(c) (d)
In the above theorem, “for almost all” means there are exceptions. For example, an independent
current source does not have a Thevenin equivalent. More generally, any two-terminal network
characterized by i{t) = constant does not have a Thevenin equivalent. This leads us to suggest that
there ought to be an equivalent current source formulation o f an equivalent network. From
Chapter 5, the source transformation theorem tells us that the Thevenin equivalent o f Figure
6.2(b) when 0 is equivalent to a current source in parallel with R^j^, as in Figure 6.3(b).
Figure 6.3 leads us to a formal statement o f the so-called Norton theorem.
■N-
Resistances
and
Independent
Sources
(c)
FIG U RE 6.3 (a) Arbitrary 2-terminal linear network o f resistors and independent sources;
(b) Norton equivalent circuit; (c) circuit for computing with computed, as per Figure 6.2(d).
A single voltage source does not have a Norton equivalent, and— as mentioned— a single current
source does not have a Thevenin equivalent. Both Thevenin and Norton equivalents exist for a 2-
terminal linear circuit when R^^ 0 and is finite. W hen both the Thevenin and Norton equiva
lents exist for the same network, the source transformation theorem and Ohm’s law imply that
thhc'kt)
(6.1a)
232 Chapter 6 • Thevenin, Norton, and M axim um Power Transfer Theorem s
r,
th - ■ (6.1b)
This formula turns out to be useful in calculating for a variety o f circuits, especially op amp
circuits.
E X A M P L E 6 .1 . For the circuit o f Figure 6.4, using literals, find the open circuit voltage, the
short circuit current, and the Thevenin equivalent resistance, Then, if R^ = 50 Q, =
200 Q , = 100 V, and i^ 2 = 2 A, construct the Thevenin and Norton equivalent circuits.
A A
-O -O
o
40 n
40 0
160V 4A
-o -O
(b) (c)
FIG U RE 6.4. (a) Resistive 2-terminal network; (b) Thevenin equivalent; (c) Norton equivalent.
S o l u t io n
Step 1. Find Using superposition, we have by voltage division and Ohm’s law,
/?2 ^1^2
~ R\+R2 ^1 + ^2
Step 2. Find i^^. As per Figure 6.4, with terminals A and B shorted ^together, all the current from
flows through the short circuit. From superposition, isc = 's2'^----- ■ Substituting numbers
into this formula yields
Step 3. Find Replacing by a short circuit and z'^2 t>y an open circuit implies that
^ ^ 40 q
R, +R2
Step 4 . Determine the Thevenin and Norton equivalent circuits. The Thevenin equivalent circuit fol
lows from Steps 1 and 3 and is illustrated in Figure 6.4(b). The Norton equivalent circuit follows
from Steps 2 and 3 and is illustrated in Figure 6.4(c). We also note that
R ,, = V =155 = 40 Q
he 4
as expected.
It is important to note that for many circuits, especially when the deactivated circuit is a series-
parallel connection o f resistances, one can obtain the Thevenin equivalent by a series o f source
transformations.
Among the three quantities, R^j^, v^^, and if two have been calculated, then the remaining one
follows easily from Equation 6.1. In some cases, the choice o f which two to find first either
increases or decreases the amount o f calculation. The following exercises illustrate this point.
Exercises. 1. For the circuit o f Figure 6.5, R^^= 200 Q, R2 = 50 O., R^= 10 Q, = 100 V, and v^2
= 50 V. Find R^f^, i^^ and v^^ in this order.
A N SW ERS: 8 Q, 1.5 A, 12 V
2. For the circuit o f Figure 6.5 with the same values as in Exercise 1, find v^^, i^^, and this order.
AN SW ER: Same as in 1, but v^^^. is harder to find.
3. For the circuit o f Figure 6.5, find the Thevenin equivalent circuit using a series o f source trans
formations.
The next example illustrates the computation o f the Thevenin and Norton equivalent circuits
using loop analysis.
234 Chapter 6 • Thevenin, Norton, and M axim um Power Transfer Theorem s
E X A M PLE 6 .2 . Find the Thevenin and Norton equivalent circuits seen at the terminals A-B for
the circuit depicted in Figure 6.6, where = 100 V and = 3.2 A. We show that = 4 0 0 Q,
= 200 V, and = 0.5 A.
500 Q 1500
A
FIGURE 6.6 Two-source circuit for Example 6.2 with loop currents shown; = 100 V and i^ 2 = 3.2 A.
So l u t io n
Step 1. Compute To compute R^^^, we set all source values to zero. Each voltage source becomes
a short, and each current source becomes an open. This leads to the circuit o f Figure 6.7. Here, we
have a 500 Q in series with 100 Q, yielding 600 Q. Since this 600 Q resistance is in parallel with
400 £2, the resulting equivalent resistance is 240 Q. Hence, = (150 + 240 + 10) = 400 Q..
500 0 1500
FIG U RE 6.7 The circuit of Figure 6.6 with all independent sources deactivated.
Step 2. Compute an expression for Because we are computing the short across the terminals
A-B is N O T present. Hence, = 0 and no current flows through the 150 Q resistor. This means
its voltage drop is zero. (One ofi:en says that the 150 Q resistor is dangling.) Thus, from KVL we
have
Step 3. Compute iy The only unknown in Equation 6.2 is /j, since ~ 3-2 A. Hence, around
loop 1,
^s\ =
Step 4 . Construct the Thevenin and Norton equivalent ciraiits. Equation 6.3 with = 400 Q yields
the Thevenin equivalent o f Figure 6.8(a). Further, from the source transformation theorem,
(6.4)
400 Q A
Step 5. Compute i^^ directly so as to verify the above calailation. This step is merely given to illus
trate the direct calculation o f i^^ and is unnecessary at this point to the solution o f the problem.
Referring again to Figure 6.6 and assuming that the short acro.ss A-B is present, then /2 = i^^.
Hence, around loop 1,
Thus,
-1
' i\ ■ 1000 -4 0 0 ' ■420‘ ■0.62'
A
j.sc -4 0 0 560 32 0.5
Vo,.
Consequently, / = 0.5 A as was found earlier using the easier method o f =
th
Exercises. 1. Suppose all source values in the circuit o f Figure 6.6 are doubled. What is the new
v j Does change?
A N S W E R : /•, =-)()() V. no
2. Suppose all resistances in the circuit o f Figure 6.6 are multiplied by 4 and the independent cur
rent source is changed to 0.6 A. Find and Hint: For in equation 6.3, the value “5 0 ”
is in ohms, so if the resistances are multiplied by four, what is the new value?
AN SW ER: r = 160 V. A',
//» 4 x 400 1600 Q. and /SC
. = 0.1 A
3. A 400 £L resistor is connected in series with terminal A o f the circuit o f Figure 6.6. Find the
V “ ‘I V
ANSWLR: im V, A',,, =»K) ti. ,nd - «.2S A
4. A 400 Q resistor is connected across terminals A and B o f the circuit o f Figure 6.6. Find the
V “" ‘I V
ANSV('-UR: O.SA, /(•, titJi - 2 0 0 U . a n d r oC- lOl) V
In the above two examples, deactivation o f all independent sources led to a series-parallel network.
Calculation o f w^as then straightfor\vard. In fact, we can state a corollary to Thevenin and
Nortons theorems.
Consider Figure 6.10(a) where we have a network N connected to the remainder o f a larger cir
cuit. Our goal is to replace the net\vork N by its Thevenin equivalent, as shown in Figure 6.10(b).
Chapter 6 •The\’enin, Norton, and Maximum Power Transfer Theorems 23'
The terminal v-i characteristics o f the network N and its Thevenin equivalent must be the same.
Consider that the v-i characteristic at A-B o f the Thevenin equivalent o f N is
(6.5)
while the Norton equivalent o f N as per Figure 6.10(c) has the v-i relationship
•
'a
- 1 - ^sc - (6 .6 )
^th
These relationships tell us that if we have a linear net\vork and assume there is a voltage across
its terminals and a current /^j entering the network, as shown in Figure 6.10(a), then obtaining an
equation o f the form
(6.7)
or o f the form
Vj n — (6.8)
allows us to match the coefficients o f equations 6.7 and 6.5 to determine and or to march
the coefficients o f equations 6.8 and 6.6 to determine ^th = — - and This sometimes proves
^th
an easier approach for non-simple circuits, as the next two examples illustrate.
Remaining
Network
^ A
------o-
Linear
A8
Network
- o .......+ —
B
(a)
Remaining Remaining
Network
(c)
E X A M PLE 6 .3 . This example revisits Example 6.2 using the new approach. Again, we find the
Thevenin and Norton equivalent circuits seen at the terminals A-B for the circuit depicted in
Figure 6. I I , where = 100 V and = 3.2 A. Our goal is to find the v-i characteristic at the ter
minals A-B.
FIGURE 6.11 Two-source circuit for Example 6.2 with loop currents
shown; = 100 V and ip = 3.2 A.
So l u t io n
Step 1, Consider i^ loop. Around the loop for i^, we have
Thus,
42 0 - 400/^
= (6 . 10)
1000
4 2 0 - 400/4
V.
AUo = 560/.A + 4 0 0 ----------------^
jQQQ + 32 = 400/.A + 200 V (6.11)
Step 4. Match coefficients. Matching coefficients o f equations 6.11 and 6.5 implies that
EXAM PLE 6.4. For the circuit o f Figure 6. ] 2, find theThevenin equivalent o f the 2-terminaJ Network
N defined by the dashed line box. We show that = 9.6 V, = 4.4 Q., and = 2.1818 A.
So l u t io n
Our objective is to compute the relationship o f the form o f equation 6.7 using Nodal analysis and
then match coefficients with equation 6.5 to obtain R^j^ and Assume /^, = 2 A and ~^
Step 1. Write nodal equations. For writing the equations o f this circuit, the reader might first
review Example 3.2. Alternately, using the inspection method, the matrix nodal equations are
- 0 .4 -0 .1 0.5 ^'ab js l + U .
Step 2. Solve equation 6.12 fo r using Crammer's rule. First, we note that
0.8 - 0 .2 - 0 .4 '
det(A/) = det - 0.2 0.6 - 0.1 = 0.:
- 0 .4 -0 .1 0.5
■0.8 - 0 .2 h\
det - 0 .2 0 .6 0
- 0 .4 -0 .1 ‘s2 + U
^’ab - V
d et(M )
Equation 6.13 shows that is calculated finding four determinants numerically using MATLAB
or equivalent.
Step 3. Match coefficients o f equations 6.13 and 6.5. Matching coefficients o f equation 6.13 with
equation 6.5, we obtain
V, = 4.4 a and 2.1818 A
Exercises. 1. If the independent current sources in the circuit o f Figure 6.11 arc set to zero, find
the Thevenin equivalent circuit.
AN.SW'j-^R: The I hcvcnin ci|uiv.ilcnt «.onsist.s oi a single rc.sisn)r. = 4.4 12.
3. A 4.4 d resistor is connected in series with terminal A o f the circuit o f Figure 6.12. Find the
V V
•ANSW'l-.R: . •).(, \’. = 8.8 12. and =1 A
4. A 4.4 Q resistor is connected across terminals A and B o f the circuit o f Figure 6.12. Find the
At this point, we end our development in this section with an example that shows how to com
pute a Thevenin equivalent from measured, e.g., in a laboratory setting where there is a power sup
ply with an adjustable voltage.
EX A M PLE 6 .5 . Consider Figure 6.13, which show^s the Thevenin equivalent o f an unknow'n net-
w’ork N attached to a variable voltage, power supply, which also shows the current delivered
to the unknown network N, i.e., Two measurements o f the unknown network N are taken, and
the data is displayed in Table 6.1. Find the Thevenin equivalent o f N.
N
r
R..
Variable
Voltage
oc
Power
Supply
TABLE 6.1
- 10 24
20 40
S o l u t io n
Substituting the measured data in Table 6.1 into equation 6.5 yields
24 = 0.01«,;,+ V
40 = 0 . 0 2 « ,^ * V
Solving produces
Thus, one can use the technique o f Example 6.5 to determine Thevenin equivalent circuits in the
laboratory.
Constructing Thevenin and Norton equivalents for active networks, those containing dependent
sources and op amps, presents us with some unique challenges. Except with one extra condition,
Thevenin and Norton’s theorems and their corollary are valid for active networks. Because active
networks contain dependent sources, the extra condition is that all controlling voltages or currents
be within the 2-terminal network whose Thevenin/Norton equivalent are being sought.
242 Chapter 6 *Thevenin, Norton, and Maximum Power Transfer Theorems
N- /^Neq
Resistances, R..
+
independent
and V
dependent
sources
N- N-
-o—
Resistances,
Independent independent
sources and
deactivated dependent
sources
(d) (e)
As in the previous section, a corollary to Thevenin and Nortons theorems is that if the nerwork
N has no internal independent sources, then the Thevenin and Norton equivalent circuit consists
o f a single resistance /?^yr. However, in contrast to passive networks, R^f^ can be negative. As a first
example illustrating the above theorems, we consider an active nervvork containing no internal
independent sources.
Chapter 6 • Thevenin, Norton, and Maximum Power Transfer Theorems
EXA M PLE 6.6 . Find the Thevenin equivalent circuit for the 2-terminal network (marked by
dashed line box) in Figure 6.15(a) using the method o f Section 3. (The dependent source acts as
a voltage amplifier.)
B
O-
(b)
FIGURE 6.15. (a) circuit with terminal voltage i/^^and input current /^;
Step 3. Match coefficients with equation 6.6. Matching coefficients implies that = ((7j + (p +
1)G^) in which case,
^2
1 R^R
j A2 '(A^ + 1) (6.14)
- R^
{ 1 J_ \ R2+(M + ^)R\ d 4.
We recognize equation 6.14 as the parallel combination o f the resistance and _.^2—
To illustrate a typical calculation, suppose p = 199, /?j = 100 and Rj = 4 kf2.
Then
Exercises. 1. For the above example, suppose p = 99, R^ = 500 Q, and /?2 = 1 P*rid R^jy
A N SW ER: 10 Q
244 Chapter 6 • Thevenin, Norton, and Maximum Power Transfer Theorems
2. For the circuit o f Figure 6.16, find the Thevenin equivalent resistance by obtaining in terms
of V
'AB-
^ih -
('jth
E X A M PLE 6 .7 . Find the Thevenin and Norton equivalent circuits seen at the terminals A-B in
Figure 6.17 when = 50 mA. Our computations will proceed using loop analysis to find the ter
minal v-i characteristic A-B.
5001
S o l u t io n
Step 1. Write a set o f loop equations for the circuit o f Figure 6.17.
which simplifies to
- 5 = 2000/, - 1500/;^
Step 2. Write the loop equations in matrix form and solve. Writing the loop equations in matrix
form yields
■ 2000 - 1 5 0 0 ' 'i\ ■
■-50'
-1 0 0 0 1100
50 2
------- 1* AR------ (6.15)
700 700
or equivalently,
Step 3. Match coejficietits o f equation 6.15 with equation 6.6 or equation 6.16 with equation 6.5 to
obtain 50 1 1
= — = — A, G,i, = — = — S. R„, = 350 Q, and v,,,. = 25 V
700 14 700 350
Exercises. I. In Example 6.7, if /j= 5 n-L\, find R^f^, and Hint: Use proportionalir)^
I
A. = S. = 350 a a n d i„ ^ = 2.5 V
140 350
2. In Example 6.7, if /^= 5 mA and the 100 Q resistor is replaced by a short circuit, find
and Hint: We have removed the dangling resistor in this case.
3. Find the Norton equivalent at the terminals A-B o f the circuit o f P'igure 6.18 when mA.
AN SW ER: = 200 LI and = 0.1 25 A
8 0 0 I.
200 0
--- O
<—
I
<
^ 800 0 800 Q
----- (
B
FIG U R H 6 .1 8 Modification o f the circuit o f Figure 6 .1 7 .
2-i6 Chapter 6 • Thevenin, Norton, and Maximum Power Transfer Theorems
Op am circuits arc active circuits. However, because the op amp is a device with special proper
ties, such as the virtual short circuit in the ideal case and such as output saturation in the non
ideal case, their discussion warrants special consideration. Our discussion begins with a Thevenin
equivalent o f a non-inverting amplifier with a dangling resistor at the output terminal.
EXA M PLE 6 ,8 . Find the Thevenin equivalent seen at the terminals A-B for the op amp circuit
o f Figure 6.19.
S o lution
Step 1. Find
2 0 -h 5
' ’Cfi -
Step 3. Match coefficients with equation 6.5. Matching coefficients we obsen'e that
R.i. = 16 Q. V’ r = , and =
16
Our next example illustrates how to construct a negative resistance using an ideal op amp.
Chapter 6 *Thevenin, Norton, and Maximum Power Transfer Theorems
EXA M PLE 6 .9 . Find the Thevenin equivalent seen at the terminals A-B for the (ideal) op amp
circuit o f Figure 6.20.
lO k O
S o lution
By V-division and the properties o f the op amp,
VcB -
5 ,
. ^’AB - Vc _
10x10^ lOxlO-"' ■■
Exercise. For the circuit o f Figure 6.21, find the Thevenin equivalent circuit at A-B.
Our third example constructs a Thevenin equivalent of the standard inverting op amp configura
tion with a terminal resistance. However, we will consider both the ideal and non-ideal cases.
EXA M PLE 6 .1 0 . Find the Thevenin equivalent seen at the terminals A-B for the op amp circuit
o f figure 6.22 when
(a) w'hen the op amp is assumed ideal, and
(b) when the op amp has a saturation voltage, = 15 V.
S o lution
Step 1. Fiuci the Theveniti equivalent seen at the terminals A-B assuming an ideal op amp. I'h e prop
erties of an ideal op amp imply that
On the other hand, with set to zero, = 0. In flict, = 0 for all possible currents, inject
ed into node A. Hence, x -5V^ implies = 0 and = -5K^. The
Thevenin equivalent seen at the terminals A-B consists only o f a voltage source o f value =“5
for the ideal op amp case.
Step 2. Find the Thevenin equivaloit seen at the termitials A-B assuming an op amp ivith output sat
uration. When the non-ideal op amp operates in its linear region, the Thevenin equivalent by Step
I is a voltage source having value v^^^ = - 5 V^. When, |-5 V} > = 15 V , or equivalently, when
I > 3 V, then the op amp saturates at ± 15 V. Specifically, when K > 3 V, then v^^ = - \5V and
when K, < - 3 V, then v^^ = 15 V. The Thevenin equivalent for an op amp with output saturation
is summarized in Figure 6.23, where v^^. takes on three separate values depending on the region o f
operation o f the amplifier.
1 5 V fo rV ^ < -3 V
{ - 5 V f o r |V j< 3 V
1 5 V fo rv / > 3 V
FIGURE 6.23 Thevenin equivalent at output terminals of an inverting amplifier (Figure 6.22)
with non-ideal op amp.
Chapter 6 • Thcvcnin, Norton, and Maximum Power Transfer Theorems 249
This cnd-s our investigation of I hevenin equivalents o f op amp circuits. There arc many more
interesting examples that are beyond the scope o f this text.
Figure 6.24 shows the Thevenin equivalent o f a network N connected to a variable load desig
nated Rj. 1'he load voltage, Vj, the load current, ij, and the power, delivered to the load arc
all functions o f Rj. The main objective o f this section is to show that for fixed R^j^, maximum
power is transferred to the load when R^ = R^f^. We illustrate this assertion with an example that
shows the power delivered to Rj as a function o f Rj. Throughout this section, it is assumed that
all resistances are non-nesative.
FIG URE 6.24 Thcvcnin equivalent o f network N connectcd to a variable load, Rj.
E XA M PLE 6 .1 1 . For the circuit o f Figure 6.24, suppose that R^j^ = 20 Q and = 20 V, Plot the
power delivered to the load R^ as a function o f
SO L U T IO N
The power delivered to the load Rj is
L (6.17)
I’L = — ^ (2 0 )'
To obtain the plot w'e use the following MA'FLAB code, resulting in the plot o f Figure 6.25.
RL in Ohms
FIG URE 6.25 Plot of power delivered to the load in Figure 6.24 as a function of Rj.
This experimentally observed fact, that maximum power transfer occurs when R^ = plays an
important role when matching speaker “resistances” to the output “resistance” o f a stereo ampli
fier or when trying to get as much power as possible out o f an antenna and into a receiver.
=«,/,
rL.max . „
A verification o f the maximum power transfer theorem proceeds using differential calculus. From
equation 6.17, the power absorbed by the load is
Chapter 6 • Thevcnin, Norton, and Maximum Power Transfer Theorems 2S1
R
Pl =
L____ V-
V? o c
[^L + ^th)‘
Following the standard procedure o f calculus for determining a maximum/minimum, we com
pute the derivative o f w i t h respect to Rj, set to zero, and solve for R^.
dPL ^ d Rr
-9 .
Rl v I c
2 oc
cJRl (IRl
= y2
/ \3
from which R^ = R^^^ and R^ = ^ are the only possible solutions. But, if R^ = oc, then = 0.
Hence, because equation 6.17 is positive for /?^ > 0, /?^ = R^j^ produces maximum power,/>^, deliv
ered to the load.
Rfll ^n/-
P L ,m a x =
(6.18)
E X A M PLE 6 .1 2 . Consider the circuit o f Figure 6.26a. Find (i) the value o f 7?^ for maximum
power transfer and (ii) the corresponding
■.......N ................................
< V -/ 300 0
600 O
Thevenin
equivalent
(b)
FIGURE 6.26 (a) A network N connected to a load /?/-, (b) Thevenin equivalent of N connected to
2^2 Chapter 6 • Thevenin, Norton, and Maximum Power Transfer Theorems
SO L U T IO N
Step 1. To compute ■urrent
the independent voltage source becomes a short and the independent cur,
source becomes an open. Finding the equivalent: resistance seen at the terminals produces = 200
Hence, maximum power is transferred when = 200
Step 2. may be computed by any o f the methods discussed throi4ghoi4t this chapter. For example, by
repeated source transformations, the network N reduces to its Thevenin equivalent shown in Figure
6.26(b) with 16 V. In fact, this approach would have found R^j^ and at the same time.
vi (1 6 )“
PU = 320 mW
4R,th 800
Exercise. Suppose the 400 Q resistor in Figure 6.26(a) is changed to 100 Q. Find R^j^, and pj
A N S W l'R S : 24 V. ISO 12, 0.96 wat t s
EXA M PLE 6 .1 3 . This example shows that the Thevenin equivalent cannot be used to calculate
power consumption within the network N it represents. For this demonstration, consider the net
work N given in Figure 6.27(a) with its Thevenin equivalent given in Figure 6.27(b). Compute
the power loss within the actual network N and within its Thevenin equivalent. We show that
these are different.
•N.
-O-
1n
1A
20
3V
--Thevenin equivalent --
(a) (b)
S O L U T IO N
W ithin the network N, the power loss is
PN, actual = 2 x 2 “ + 2 x 1 - = 10 w;
watts
Pm - 1 X 1- - 1
Chapter 6 • Thevenin, Norton, and iMaximum Power Transfer Theorems 2^3
This means that theTIicvenin equivalent is not, in general, representative o f pov/er relationships
within the network, i.e., the losses that are dissipated as heat, for example.
When a network N is a voltage source in series with a resistance R^, and hence is its own Thevenin
equivalent, one may ask about maximum power transfer when is variable and the load Rf is
fixed, assuming is also fixed. The following example is an experiment for investigating this sit
uation.
E X A M P L E 6 .1 3 . For the circuit o f Figure 6.28, suppose /?^ = 20 Q and = 20 V. Plot the power
delivered to the load as a function o f R^ along with the power loss, in R^.
FIGURE 6.28 A network N in which R^ can be adjusted with and /?^ fixed.
SO L U T IO N
The power delivered to the load R^ is
R, 0 20 X 400
•vr =
(R l + Rs ) (20+ / ?,)
To obtain the plots, we use the following MATLAB code, resulting in the plot o f Figure 6.29.
FIGURE 6.29 Plot of power delivered to load as a function of for circuit o f Figure 6.28.
According to Figure 6.29, the maximum power o f 20 warts is delivered when = 0. Observe that
if R^< R^ (the usual case), then mi n i m i z i n g m a x i m i z e s However, if R^ > Rj, minimizing
Ploss niaximizep^.
The proof for the maximum power transfer theorem given earlier considers Rj as the independ
ent variable and sets to zero, standard practice in calculus. There is, however, an alternate
dRi
approach whose derivation is simpler mathematically, but is more meaningful for applications in
the sense that the load can be a general 2-terminal linear network, N^, instead o f a single resistor.
For this alternate derivation, refer to Figure 6.30. We ask the question. What v-i characteristic
should the load network have so that maximum power is transferredfrom N to N J
FIGURE 6.30 The Thcvcnin equivalent of a network N connected to a loading ncrvvork N^.
Chapter 6 • Thevcnin, Norton, and Maximum Power Transfer Theorems
To find the value o f v, we note that the power transferred from N lo Nj^ is
p = y i= y l^ - V
R,th
To find the value o f v that maximizes we differentiate with respect to v and set the result to zero:
lIPL ^oc - 0
dv R,i,
V = 0.5v„ (6.19a)
at which value
(6.19b)
ih th
which are the conditions on i/and i for maximum power transfer to If consists o f a single
resistor it has the v-i characteristic o f equations 6.19. Then from O hm s law,
^ L - ~ - ^ih
I
At V= 0.5v , the corresponding maximum power is
PLmax = V X /= - ( 6 .20 )
4/?,th
EX A M PLE 6 .1 4 . In the circuit o f Figure 6 .31, and R^f^ = 2 0.. Find the value o f that
maximizes power transfer to the network N^.
S o lution
According to equation 6.19(a), maximum power transfer occurs when v = = 1 V and from
V= \ = 0 .2 / + = 0.\ + Vj =o Vf = 0 .9 V
2S 6 Chapter 6 • Thcvcniii, Norton, and Maximutn Power Transfer'I’hcorcnis
Exercises. 1 If the 0.2 £2 resistor is changed 4 Q, find the value o f that maximizes power trans
fer to the network N^.
ANSWHR; -1 V
2. If the 0.2 Q resistor is changed 2 Q, find the value o f Vj that maximizes power transfer to the
network N^.
AN SW ER: 0 V
3. IF the 0.2 Q. resistor is variable and Vj = 0.5 V, find the new' value of the 0.2 resistor that
maximizes power transfer to the Nj.
AN SW ER: 1 Q
8. SUM M ARY
This chapter has set forth a powerful strateg)' for analyzing complex networks by replacing por-
tions o f the nerwork by their simpler Thevenin and Norton equivalents. The Thevenin and
Norton theorems assure us that almost any 2-terminal linear nerwork, no matter the number o f
internal elements, is equivalent to a simple nerwork consisting of an independent source either in
series with or in parallel with a resistance. O f course, an independent current source does not have
a Thevenin equivalent, and an independent voltage source does not have a Norton equivalent. '
\4ore generally, there are some circuits that have one but not the other. Further, some circuits have
neither.
rhe chapter has illustrated various techniques for constructing the Thevenin and Norton equiv- ^
alents. For passive networks, the ordinary approach is to find first by deactivating all internal
independent sources. If the resultant circuit is series-parallel, then can be found by combin- ^
ing series and parallel resistances as learned in Chapter 2. If the resultant nerwork is not series-
parallel, then one should use the main technique set forth in this chapter, which is to find the v-
i characteristic o f the terminals. This technique is valid for all circuit t\'pes.
With the ideas o f a Thevenin and Norton equivalent circuit, we then investigated the problem of
transferring power to a load. When R^f^ is fixed, maximum power is transferred when R^ is adjust
ed to be R^j^. If is adjustable and R^ is fixed, then maximum power is transferred when R^j^ =
0. It is important to imderstand that a practical dc voltage source (such as a battery in an auto
mobile) is designed to provide nearly constant output voltage for the intended load current.
Accordingly, it has a rather small source resistance R^. Any attempt to transfer the maximum
power from such a source continuously will overload the source and may cause damage to its
internal structure. For example, in a lead acid battery, the plates may warp or the solution bub
ble. Hence, maximum power transfer is not o f critical importance for power transmission net
works, whereas for communication networks, maximum power transfer is important.
Chapter 6 •Thcvcnin. Norton, anil iMaximum Power Transfer Theorems 25'
2-term inaI network: an interconnection o f circuit elements inside a box having only 2 accessible
terminals for connection to other nervvorks.
D eactivating an independent current source: replacing the source by an open circuit.
D eactivating an independent voltage source: replacing the source by a short circuit.
Equivalent n-term inal networks: t\vo n-terminal networks having the same terminal voltage-
current relationships. Alternately, two n-terminal networks N j and N-, are equivalent
when substituting one for the other in every possible network N; the voltages and cur
rents in N are unaffected,
the current through a short circuit placed across the output terminals of a 2-terminal network.
M aximum Power Theorem : let an adjustable load resistor be connected to the Thevenin
equivalent o f a 2-terminal linear network. Maximum power is absorbed by the resistor
when Rj = R^i^.
N orton’s equivalent circuit: any 2-terminal net\vork consisting o f independent sources and lin
ear resistive elements is equivalent to an independent current source in parallel with a
resistance.
R(h (Thevenin’s equivalent resistance): the resistance that appears in the 7'hevenin equivalent
circuit o f a 2-terminal linear network. It is also the equivalent resistance of the 2-termi-
nal net\vork w'hen all internal independent sources are deactivated.
Thevenin’s equivalent circuit: any 2-terminal network consisting o f independent sources and lin
ear resistive elements is equivalent to an independent voltage source in series with a resist
ance.
V : the open circuit voltage o f a 2-terminal network N when no load is connected.
' For a generalization o f this condition to the case where the controlling voltage or current is out
side o f N, see the article by Peter Aronheim entitled “Frequenc)' Domain Methods” in The Circuits
and Filters Handbook, BocaRaton, FI.: C R C Press, 1995, pp. 682-691.
258 Chapter 6 • Thevenin, Norton, and Maximum Power Transfer Theorems
R.
Figure P 6.1
5kQ
-o-r
Figure P6.2
-O -v
'■ 6
- 6 ^«(l) «
Figure P 6 .6
Figure P6.3 CH ECK: = 2 mA
ANSWF.R: /?;/; =2 kil, isr = 20 mA, // . = 40 V
Chapter 6 • Thcvcnin, Norton, and Maximum Power Transfer Theorems 2S9
7. Find the Thevcnin equivalent seen at A-B o f (c) Find the value o f Rj for maximum
the circuit o f Figure P6.7, where R, = 18 kH, R-, power transfer and the resultant power
= 9 kD, /?3 = 3 kD. /?4 = 6 kD. = 3.6 kQ. delivered to the load.
= 32 k n . = 48 V, and s2 = 8 mA. Hint: Use (d) If the value o f is doubled, what is the
the result o f Problem 6 to find the Thevenin power delivered to the load under the
equivalent of the network between C and D. condition o f maximum power transfer?
R,
'■ 6 8R
6R
Figure P6.9
C H ECK S: (a) 900 Q, 30 V; (c) 250 mW, (d) 1
Figure P6.8
2kn
TH EVEN IN /N O R TO N FOR
A CTIVE CIRCU ITS
13.(a) Find the value o f so that the Figure P6.15
Thevenin equivalent resistance o f the
circuit shown in Figure P 6 .13 is 5 16.(a) Find a so that R^^^ = 5 kQ for the cir
(b) Repeat part (a) for the case when = cuit of Figure P6.16.
- 2 5 0 Q. C H EC K : (b) 6.25 mS (b) Repeat part (a) so that = -1 kl^ .
200 n 1 kn
2k O
»800 n
Figure P 6 .17
ANSWFR: - 6 0 0 /„ = 0
Chapter 6 • Thevenin, Norton, and Maximum Power Transfer Theorems 261
18, Use loop analysis to compute the Thevenin 2 1 .(a) Find the Thevenin and Norton equiva
equivalent for the circuit shown in Figure lent circuits for the network shown in
P6.18, What is the Norton equivalent? Figure P6.21, assuming that k = 0,025 S
and = 20 V,
(b) For what value o f k is the open circuit
100 0 voltage zero. For this value o f k, deter
-----1----- -
mine R^fj.
loon '300 0
' 0.01 0>
800 n
Figure P 6,18
ANSW I-R; = 0, = 250 12
'■©
50 0
bv
20. Find the Thevenin equivalent o f the circuit 23. Consider the circuit shown in Figure
in Figure P6.20 where = 0.2 A. P6.23.
(a) Find the Thevenin equivalent.
(b) If a load resistor R^ is connected across
400 Q terminals A-B, determine R^^ for maxi
—►
mum power transfer and determine the
1 kO
'• 0 '400 0
maximum power delivered to R^.
200i
(c) If a resistor /?, were added in series with
terminal A o f figure P6.23, what is the
Thevenin equivalent resistance o f the
Figure P6.20
AN SW FR: r^. = 60 \‘. A’,;, = SOO LI augmented network.
2(>: Chapter 6 •Thevenin, Norton, and Maximum Power Transfer Theorems
'■ 6
Figure P6.23
27. Find the Thevenin and Norton equivalents of
2 4 .(a) Find the Thevenin equivalent for the the circuit o f Figure P6.27 when = 20 mA.
network shown in Figure P6.24. AA
(b) If the values of each source are cut in
half, what is the new
ANSWER: (a) 1.6 kLl - 260 V: (b) /• = - 130 V
400 n
2000i
io o v (^ ^
0 0.1 A
ANSWI-:R: = 12.S V, = 650 Q
Figure P6.24
Figure P6.25
C H EC K : 0.09 A
Figure P6.28
26. Consider the circuit o f Figure P6.26, where CH ECK: = 8 mA
= 32 V, /?, = 80 Q, /?2 = 240 Q, = 60 Q,
and = 2.
OP AM P PROBLEM S
(a) Replace the circuit to the left o f nodes A
and B with its Thevenin equivalent.
2 9 .(a) Find the Thevenin equivalent seen at the
(h) Given your answer to (a), assume that Rf
terminals A-B for the op amp circuit o f
= 150 n , and find and the power con
Figure P6.29.
sumed by Rj.
C H EC K : R^,^ = iQ Q.
Chapter 6 • Thevenin, Norton, and Maximum Power Transfer Theorems 263
Figure P6.29
C H EC K : /?,/,=/?3
30. (a) Find the Thevenin and Norton equiva Figure P6.32
lents seen at the terminals A-B for the op amp AN SW ER: ( b ) = A*,, =0
circuit o f Figure P6.30.
3 3 .(a) Find the Thevenin and Norton
Equivalent circuits o f the op amp con
figuration o f Figure P6.33 seen at A-B.
(b) Determine the value o f a load resistor R^
connected across the terminals A-B for
maximum power transfer. If = 4 V
and ryp = 5 V, determine the maximum
power transferred to this R^.
C H E C K : (b) = 0.9 watts
1 ( /?! + /?2 \
20 kQ 50 kO
rA «i ‘I
3 1.(a) Find the Thevenin equivalent to the
right o f the terminals A-B for the (ideal)
op amp circuit o f Figure P6.31.
(b) If the practical source indicated in the
figure is attached to A-B, find the cur-
rent in terms o f
15 kn
N
Linear resistive
network with
Power
dependent sources
Supply
and fixed
independent
sources
Figure P6.36
Tabic P6.36
3 5 .(a) Find the Thevenin equivalent seen at the
/^l (mA)
terminals A-C for the op amp circuit of
Figure P6.35 when the op amp has out 1 6
put saturation, = 12 V.
4 12
(b) Find the Thevenin equivalent seen to the
left o f the terminals B-C and the maxi C H E C K : ( b )P „ ,,,= 2 m W
mum power that will be absorbed by the
24 k n resistor for all variations in V.. 37. Repeat Problem 36 with the data given in
■Rible P6.37.
Table P6.37
r j (mA)
10 54
40 66
Figure P6.35
C H E C K : (b) .5625 W
144
C H EC K : /W = = 6 mW
2800 38. The data listed in Table P6.38 was taken for
the network N o f Figure P6.38.
THEVEN IN AND NORTON
(a) Fill in the values for the third column o f
EQUIVALENTS FROM Table P6.38 and find the Thevenin and
M EASURED DATA Norton equivalents o f the linear resistive
36. In a laboratory, the data set forth in rows I nervvork N.
and 2 o f Table P6.36 were taken. (b) To what resistance should be changed
(a) Compute the Thevenin and Norton to achieve maximum power transfer?
equivalents o f N. What is P..
(b) After the power supply is removed, what
resistance, Rj, should be connected Table P6.38
across A-B for maximum power transfer?
(niA )
What i s />„„„?
2 4
10 10 j>
Chapter 6 •lhc\xnin, Norton, and Maximum Power Transfer Theorems 26S
Table P6.40
R, (M Q) (pA)
2 0.4 >
10 1 >
N
Linear resistive 42. 'Fhis problem is the second of rwo problems
network with
dependent sources that outline a laboratory measurement procedure
and fixed for finding the The\'enin equivalent of a linear
independent
sources resistive 2-terminal net\vork. For this problem,
consider the new configuration o f Figure P6.42 in
which the circuit under test contains independent
Figure P6.40 sources and has a non-zero 'Fhe experimental
=4 V
apparatus includes a resistance decade box, denot
ed R, and a dc voltmeter with internal resistance
266 Chapter 6 • Thcvenin, Norton, and Maximum Power Transfer Theorems
Figure P6.42
(c) If R = 500 Q, V^it) = 50 sintdOOO t) mV, dated voltage and the power delivered to
and 100 mV, find i{t). Hint: Use a the load.
suitable Thevenin equivalent for N.
(d) If R = 50 v^{t) = 200 sintdOOO t) mV,
and = 500 mV, find /(/).
R
scale)
'■ 6
Figure P6.50
(a)
Figure P6.47 51 (a) For the circuits o f Figure P6.51, find the
ANSWl-.R: (a) = 0. = 2 kl2; (h) = load resistance R^ needed for maximum
0.195V, R,/, = 5 o 'h ; (c) 0.04 + 0.02 sin (1000 power transfer, the associated voltage
/) mA: (d) .^.05 + 2 sindOOO t) mA and the power delivered to the load.
(b) If the load resistance is constrained as 5
k n < R^ < 10 kD, repeat part (a).
M AXIM UM POW ER TRANSFER (c) If the load resistance is constrained as 15
48. For the circuit o f Figure P6.48, /?, = 160 Q., k n < R^ < 20 k n , repeat part (a).
R-, = 480 and = 80 V. Find the value o f
for maximum power transfer and
e
12 kO 2 mA
8kn
6 24 V
6kfi
Figure P6.48
C H EC K : 7.5 watts
Figure P 6.51
49. For the circuit o f Figure P6.49, R^ = 900 Q,
R j= 180 /?3 = 50 Q, = 60 mA, and v^ 2 = 52. Consider the circuit o f Figure P6.52.
21 V. Find the value o f for maximum power (a) Find the value o f for maximum
transfer and power transfer to the three-resistor load.
(b) Find the power delivered to each load
resistor, i.e., to R^, R J2 , and /?^/3.
40 V
Figure P6.49
lon
50. For the circuits o f Figure P6.50, = 10 V
and v^ 2 - ^5 V. Find the load resistance R^^ Figure P 6 .5 2
needed for maximum power transfer, the asso-
268 Chapter 6 • I'hevcnin, Norton, and Maximum Power transfer Theorems
(a) (b)
Figure P6.56
R.
30 Q + V 15Q I
(b 30 V .s v Q
Figure P6.57
(b)
58. The i-v relationship o f certain type o f LED
Figure P6.54 (light emitting diode) in its operating range o f
C H EC K : (a) 10 watts
1-7 V -3 V is represented by a 2 V voltage-source
in series with a 50 H resistance. The load con
55. Suppose the polarity o f the 15-V-source in sists o f a network o f n such diodes connected in
Problem 54 is reversed. Repeat Problem 54 and parallel. The source network is represented by a
determine which configuration transfers more 5 V voltage-source in series with a 50 ^2 resist
power to the load. ance. Assume that the power delivered to each
diode is totally converted into light. Determine
56. T he linear resistive circuit o f Figure how' many LEDs should be connected in paral
P6.56(a) is found experimentally to have the lel for maximum brightness. W hat is power dis
voltage-current relationship plotted in Figure sipated by each diode?
P6.56(b). Find the maximum power that can
be absorbed by placing a load resistor across ter ANSW ER: M - 5, = 2S mW
minals a-b?
C H A P
Every non-portable personal computer contains a power supply that converts the sinusoidal volt
age o f the ordinary household outlet to a regulated dc voltage. “Regulated” means that the output
voltage stays within very tight limits o f its nominal value (e.g., 12 ± 0.1 V) over a wide range o f
power requirements. Engineers design power supply circuits with regulators that produce voltages
with a small oscillation because to generate a truly dc voltage is practically impossible.
RECTIFICATION REGULATION
O
O
SMOOTHING
This process o f converting ac to dc has three stages: First, the ac waveform is rectified into its
absolute value. Then a smoothing operation takes place that reduces the variation in the voltage
to a reasonable but still unacceptable level. This first level o f smoothing is nccessar)’ becausc the
voltage regulator is a precision subcircuit that requires a fairly constant voltage for its proper oper
270 Chapter 7 • Inductors and Capacitors
ation. The partially smoothed waveform is fed into a voltage regulator, which limits the voltage
oscillation between critical levels even when the load drawn by any connected device (e.g., your
computer) varies in the course o f its operation.
As mentioned, the rectified sine wave is smoothed before entering the voltage regulator. A crude
smoothing can be accomplished with a capacitor, a device studied in this chapter. Intuitively,
capacitors resist voltage changes and are designed to steady the voltage at a constant level. In this
chapter, we will study the capacitor and investigate a simplified smoothing operation for a power
supply.
CH APTER O BJECTIVES
1. Define the notion o f inductance and introduce the inductor, whose terminal voltage is
proportional to the time derivative o f the current through it.
2. Investigate the ability o f an inductor to store energy and the computation o f the equiva
lent inductance o f series-parallel connections.
3. Define the notion o f capacitance and introduce the capacitor, whose current is propor
tional to the time derivative o f its terminal voltage.
4. Investigate the ability o f a capacitor to store energy and the computation o f the equiva
lent capacitance o f series-parallel connections.
5. Define and illustrate the principle o f conservation o f charge.
CHAPTER O U TLIN E
1. Introduction
2. The Inductor
3. The Capacitor
4. Series and Parallel Inductors and Capacitors
5. Smoothing Property o f a Capacitor in a Power Supply
6. Summary
7. Terms and Concepts
8. Problems
1. IN TRO D U CTIO N
This chapter introduces two new circuit elements, the linear inductor and the linear capacitor,
hereafter referred to as an inductor and a capacitor. The inductor, shown in Figure 7.3, is a device
whose voltage is proportional to the time rate o f change o f its current with a constant o f propor
tionality I , called the inductance o f the device, i.e.
as set forth in equation 7.1. The unit o f the inductance Z., is the henry, denoted by H.
Macroscopically, inductance measures the magnitude o f the voltage induced by a change in the
current through the inductor.
Chapter 7 • Inductors and Capacitors
The capacitor, shown in Figure 7.15, is a device whose current is proportional to the time rate o f
change o f its voltage, i.e.,
dvcO)
ic(0= C
(it
as set forth in equation 7.5. Here, the constant o f proportionality, C, is the capacitance o f the
device with unit farad, denoted by F. Capacitance measures the devices ability to produce a cur
rent from changes in the voltage across it.
By adding the inductor and the capacitor to the previously studied devices (the resistor, inde
pendent and dependent sources, etc.), one discovers an entire panorama o f possible circuit
responses, to be explored in the next four chapters. Together, these devices allow one to design
radios, transmitters, televisions, stereos, tape decks, and other electronic equipment. In this chap
ter, our goal is to understand the basic operation o f inductors and capacitors.
2. TH E IN DU CTO R
SomePhysics
In Figure 7.1, a changing current flowing from point A to point B through an ideal conductor
induces a voltage between points A and B according to Faradays law. Joseph Henr}' inde
pendently observed the same phenomenon at about 1831. The induced voltage, was found
The following experiment illustrates the idea. Suppose the conductor in Figure 7.1 is 6 feet o f #22
copper with resistance 16.5 Q/1,000 ft. The 6-foot length has a resistance o f about 0.1 U. Using
a current generator, we apply a pair of ramp currents (shown in Figure 7.2a) to the conductor, as
per Figure 7.2b. The measured responses are shown in Figure 7.2c and, as expected, satisfy Ohms
law.
Chapter 7 * Inductors and Capacitors
(a)
(b)
(c)
(d)
(e)
FIGURK 7.2 (a) Ramp currcnt inputs to iincoilcd and coilcd wire, (b) Six feet of #22 wire attached
to a current generator, (c) Voltage responses to ramp current inputs of uncoiled wire.
(d) Six feet o f #22 wire coiled into 45 turns 1” long and 1” in diameter.
(e) Voltage responses to ramp current inputs o f coiled wire.
Chapter 7 • Inductors and Capacitors 273
Now suppose the wire is coiled into a qrlinder 1” in diameter and 1” long, as in Figure 7.2d. Apply
the same ramp currents o f Figure 7.2a to the coiled wire. This time, the measured responses are as
shown in Figure 7.2e. These responses have the same shape as those of Figure 7.2c, except for the
offsets o f 30 mV and 60 mV, respectively. These offiet voltages are proportional to the derivatives
of the input currents, i.e., Offset -
for k = 1 , 2 , where L is the proportionality constant, called the inductance o f the coil. Since the
derivative o f i s lO'^ A/sec, and the derivative o f ini^t) is 2 x lO'^ A/sec, the inductance L
of
the coil can be computed as
^ Offset 0.03 0.06 , .
3x10 = —j:— = — - hennes
^ 10^ 2x10^*
dt
As mentioned earlier, the heniy, equal to 1 volt-sec/amp and abbreviated H, is the unit of induc
tance. Also, from the above experiment, one concludes that the inductance o f a cylindrical coil o f
wire is much greater than the inductance of a straight piece of wire, which in the above experi
ment was not measurable by our apparatus.
The physics of the preceding interaction is governed by Maxwell’s equations, which describe the
interaction between electric and magnetic fields. A time-varying current flow through a wire cre-
ates a time-varying magnetic field around the wire. The magnetic field in turn sets up a time-vary-
ing electric field, i.e., an electric potential or voltage. One can verify the presence o f this magnet-
ic field by bringing a compass close to a wire carrying a current. The magnetic field surrounding
the wire will cause the compass needle to deflect. Physically speaking, a changing current causes a
change in the storage o f energy in the magnetic field surrounding the conductor. The energy trans
ferred to the magnetic field requires work and, hence, power. Because power is the product of volt
age and current, it follows that there is an induced voltage between the ends of the conductor.
W hat is even more interesting is that if a second wire is immersed in the changing magnetic field
of the first wire, a voltage will be induced between the ends of the second wire. A proper (math-
ematical) explanation of this phenomenon is left to a fields course. For our purposes, three fects
are important: (1) energy storage occurs, (2) the induced voltage is proportional to the derivative
o f the current, and (3) the constant of proportionality is called the inductance of the coil and is
denoted by L.
As mentioned, a straight wire has a very small inductance, whereas a cylindrical coil o f the same
length o f wire has a much greater inductance. This inductance can be increased many times over,
possibly several thousand times, simply by putting an iron bar in the center o f a cylindrical coil.
Alas, the calculation o f inductance is the proper subject of more advanced texts, e.g., on field the
ory or transmission line theory. Nevertheless, there are empirical formulas for estimating the
inductance of a single-layer air-core coil as described in the homework exercises.
274 Chapter 7 • Inductors and Capacitors
(7.1)
dt
i,(t)
h/Y Y V
+ V jt) -
0.5H
S o lution
From equation 7.1, direct differentiation o f the inductor current /^(/) leads to
2. In Example 7.1, suppose ii{t) = (1 - V for / > 0 and 0 otherwise. Find t> 0 .
AN SW ER: ///(/) = lOOe--”'*'V.
The differential equation 7.1 has a dual integral relationship. Safely supposing that at / = the
inductor had not yet been manufactured, one can take = 0, in which case
Chapter 7 • Inductors and Capacitors 275
(7.2)
L-fh
The time represents an initial time tiiat is o f interest or significance, e.g., the rime when a switch
is thrown or a source excitation is activated. The quantity
specifies the initial current flowing through the inductor at ^q. This quantity, sums up the
entire past history o f the voltage excitation across the inductor. Because o f this, the inductor is said
to have memory.
EXA M PLE 7 .2
For the circuit o f Figure 7.5a, determine /^(O) and ij{t) for / > 0 when Vj{t) = V as plotted in
Figure 7.5b.
\(t)
ijt)
,(t, L = 0.5H
(a) (b)
FICJURE 7.5 (a) Simple inductor driven by a voltage source, (b) Source waveform Vj{t).
(c) Resulting inductor current
1~ G Chapter 7 • Inductors and Capacitors
S olution
A direct application o f equation 7.2 leads to
I T, 1
= z j l ’. i / o " " ' ' ’’ = z + i ’
It follows that
2. Repeat Example 7.2 with L = — H and with v,{t) = cos(27ir) V for r > -0 .2 5 sec and zero
I . 4jt
otherwise.
ANSWl'.R: //(()) = 2 A, /y(/) = 2 + .sinUni) A for r > 0.
E X A M PLE 7 .3
Consider the circuit o f Figure 7.6a with voltage excitation v^{t) shown in Figure 7.6b. Find the
inductor current /^(r) for f > 0, assuming that /^(O) = 0.
i,(t)
,(t ) Q L = 0.5H
(a)
FIGURE 7.6 (a) Voltage source driving inductor, (h) Square wave excitation !ȣ(/).
S o lution
It is necessary to apply equation 7.2 to each interval, [0, 1], [1, 2], ... , [;;, n + 1], .... For this we
need to first specify the initial conditions for each interval.
Step 3. Compute the iuitial condition for the interval [n,u + 1] for u even. Again from equation 7.2,
with t = n and n even, we liave
Step 4. Compute the initial condition for the interval [n,n + 1] for n odd. From equation 7.2, with
t = n and n odd, we have, utilizing steps 1 and 3,
/I-1 /j «
f
I j j
+“ +j f y i(r)d T = — J v7 (tV /t = 2
ri-1 n-l
since n - 1 is even.
Step 5. Compute ij^t) over [n,n + 1] ivith n even. If n is even, then the value o f the inductor cur
rent over the interval [;/,;/ + 1] is
Observe that i^it) = 2 t - In A is the equation of a straight line having slope +2 and^-intercept -In .
Step 6. Compute i^(t) over [n,n + 1] with n odd. If n is odd, then for the inter\'al [n,n + 1], the
inductor current is
Here, i^{t) = 2 + 2n - 2t is the equation o f a straight line, with slope - 2 and )'-intercept 2+2«.
Step 7. Piece segfnentsfivm steps 5 and 6 together. Thus the segments computed in steps 5 and 6 inter
cept the /-axis at the same points. Figure 7.7 sketches the resulting triangular response for /> 0.
FIGURE 7.7 Triangular shape o f inductor current for the square wave voltage excitation of
Figure 7.6b applied to the circuit of Figure 7.6a.
27H Chapter 7 * Inductors and Capacitors
Exercises. (All time is in seconds.) 1. Again consider the circuit o f Figure 7.6a. Compute iyr (?) for
(i) 0 < r < 1, (ii) 1 < t < 3 , and (iii) 3 <t For the waveform o f Figure 7.8a, assuming i^(0) = 0.
2. Again consider the circuit o f Figure 7.6a. Compute i^(^) for (i) 0 < r < 1, (ii) 1 < ^ < 3, (iii) 3 <
( < 4, and (iv) 4 < /, for the waveform o f Figure 7.8b, assuming i^(0) = 0.
It is important to recognize that the square wave voltage input o f Figure 7.6b is discontinuous but
the current waveform o f Figure 7.7 is continuous. Integration (computation o f “area”) is a smooth
ing operation: it smoothes simple discontinuities. This means that the inductor current is a con
tinuous function o f t, even for discontinuous inductor voltages, provided that the voltages are
bounded. A voltage or current is bounded if the absolute value o f the excitation remains smaller
than some fixed finite constant for all time. Thus, equation 7.2 leads to the continuity property
o f the inductor: if the voltage Vf{t) across an inductor is bounded over the time interval /] < t <
tj, then the current through the inductor is continuous for < t < tj. In particular, if
then /^(^o”) = ^ The notation and “+” on /q is used to dis
tinguish the moments immediately before and after /q- For example, in Figure 7.9, t = 2 shows a
discontinuity o f The value o f *^^(2“) is 1 and the value o f is - 1 . The value Vj{2*) can
be seen as the limiting value o f z^^(r) when approaching r -» 2 from the right, whereas Vf{2~) can
be seen as the limiting value o f v^{t) when approaching t 2 from the left.
Chapter 7 • Inductors and Capacitors 279
\ {2 )
/
c
(U
L-
u
T3
C
rtJ
01
CT>
TO
*->
o
>
u
■O
c
Time (seconds)
PotverandEnergy
Rccall that the instantaneous power absorbed by a devicc is the product o f the voltage across and
the current through the device assuming the passive sign convention. For an inductor,
Plit)=\'L{t)ilU) =
dt w atts.
Since energy (absorbed or delivered) is the integral o f the instantaneous power over a given time
interval, it follows that the net energy stored' over the inter\'al [/q, /■,] in the magnetic
field around the inductor is
f'o V (h /
(7.3)
= /£(/,)-//^(/q ) joules.
for L in henries and in amps. From equation 7.3, whenever the current waveform is bounded,
the net energy stored in the inductor over the interval [/q, rj] depends only on the value o f the
inductor current at times r, and /q, i.e., on //(^j) and //(/q)’ respectively. This means that the
stored energy is independent o f the particular current waveform between and
If the current waveform is periodic, i.e., if ij{t) = + T) for some constant 7'> 0, then over any
time interval o f length T, the net stored energ)' in the inductor is zero because = /^(/-q + 7)
280 Cliapter 7 • Iiuluctors and Capacitors
forces equation 7.3 ro zero. To further illustrate this propert)', consider Figure 7.10a, which shows
a 0.1 H inductor driven by a periodic current /^(^) = sin(27tr) V. This current signal has a funda
mental period T = I, i.e., the smallest 7 'over which the signal repeats itselh From equation 7.3,
However, we can interpret this result in terms o f the waveform o f pi{t). First note that the volt
age across the inductor in Figure 7.10a is Vjit) = 0.27Tcos(27if) V. Hence, the instantaneous power
is pf{t) = = 0.2jtcos(2Tt^)sin(27tr) watts, as plotted in Figure 7.10b. Observe the shaded
regions o f Figure 7.10b in which the area under the power curve has equal parts ot positive and
negative area. This means that all the energ)' stored by the inductor over the part o f the cycle o f
positive power is delivered back to the circuit over the portion of the cycle when the power is neg
ative. Fhis is true for all periodic signals over any period. Because no energ)' is dissipated, and
because energy is only stored and returned to the circuit, the (ideal) inductor is said to be a loss
less device.
IlW
v jt)
sln(27T) A 0.1 H = 0.2n cos(2n) V
(a)
P lW
(b)
FIGUllE 7.10 (a) Inductor excited by periodic current, (b) Plot of the power absorbed by the inductor.
= (7 .4 )
Chapter 7 • Inductors and Capacitors 281
for all t. Equation 7.4 can be viewed as a special case o f equation 7.3 in which r,) = -oo and /^(-oo)
= 0. Thus, equation 7.4 can be interpreted as the change in stored energ)' in the inductor over the
inten'al (^x>, t].
E XA M PLE 7 .4
Find the instantaneous energy stored in each inductor o f the circuit o f Figure 7.11 a for the source
waveform given in Figure 7.1 lb. In Figure 7.1 lb , note that ij^t) = 0 for r < 0.
FIGURE 7.11 (a) Series inductors excitcd by a source current, (b) Graph o f the source current.
S olution
From KCL, i^{t) = for all t. Since i^{t) = 2r A for 0 < r < 1 and i^{t) = 2 A for /> 1, equa
tion 7.3 or 7.4 immediately yields the instantaneous stored energies (in J) as plotted in Figure 7.12:
r 0^/<l 4r 0:sr<l
1 1s
> t > t
(a) (b)
FIGURE 7.12 (a) Encrg)' (in J) stored in inductor Z.,. (b) Energy (in J) stored in inductor Ly
282 Chapter 7 • Inductors and Capacitors
Exercises. 1. For the circuit o f Figure 7.1 la, find analytic expressions for the instantaneous stored
energ)' for the current excitation in Figure 7.13a for r > 0.
i.(t)
L^ = 1H
FIGURE 7.14 Parallel inductive circuit with switch in which v^{t) = cos(t) V for /> 0 and 0 otherwise.
Chapter 7 • Inductors and Capacitors 283
S olution
Step 1. Since no voltage is applied to either inductor for / < 0, /,(0) = 0. Further, no voltage
appears across the second inductor until r > 1. Hence, /^(l) = 0.
Step 3. At ^ = 1, the switch closes. T he r\vo inductors are then in parallel, and the source voltage
appears across each. Hence, by equation 7.2,
From the KCL, the input current ij^t) = /,(/) + ijit) = 2sin(/) - sin(I) A for / > 1.
Step 4. Compute the energy stored in the inductors over the interval [0, t]. From equation 7.3, it fol
lows that for 0 < r < 1, t) - 0.5 sin^(r) joules, whereas ->((), t) = 0.
Step 5. Compute the energy stored in the inductors over the interval [0, f] for 1 < t. Again from equa
tion 7.3, for 1 < t, t) = 0.5 sin^(^) joules and t) = 0.5[sin~(r) - 2 sin (l) sin(/) +
sin^(l)] joules.
Exercise. Repeat the calculations o f Example 7.5 for = 2 sin(r) V for / > 0 and 0 otherwise.
A N SW ERS: For 0 < / < 1. U'} ,(0, f) = [2 - 2 cos(r)]“ J, whereas W) ,(0, t) = 0; for 1 < U'^^,(0.
/) = [2 - 2 cos(/)]- J and U'} ,(0, ;) = [ 1.0806 - 2 cos(/)]- J.
28-» Chapter 7 • Inductors and Capacitors
3. THE CAPACITO R
DefinitiojisandProperties
D EFIN ITIO N OF T H E C A PA CITO R
Like the inductor, the capacitor, denoted by Figure 7.15a, is an energy storage device.
Physically, one can think o f a capacitor as two metal plates separated by some insulating mate
rial (called a dielectric) such as air, as illustrated in Figure 7.15b. Placing a voltage across the
plates o f the capacitor will cause positive charge to accumulate on the top plate and an equal
amount o f negative charge on the bottom plate. This generates an electric field between the
plates that stores energy. Hence, for a capacitor.
(7.5)
d\ clt
where q{() is the accumulated charge on the top plate, which is proportional to the voltage
V({t) across the plates; thus q{i) = Cv^t), with proportionality constant C denoting capaci
tance and having the unit o f Farad (F). One Farad equals 1 amp-sec/volt. The capacitance C
is a measure o f the capacitor’s potential to store energ)' in an electric field.
ic(t) >r ^ + -h
-i- + 4-
, 4- + + -I--1- +
V ,(t) A- + + + + + +.
(a) (b)
FIG URE 7.15 (a) The symbol for the capacitor with conventional voltage and current direc
tions. (b) Illustration o f electric field between plates of a parallel-plate capacitor.
Modern-day capacitors take on all sorts o f shapes and sizes and materials. In keeping with cradi-
tio n , the parallel-platc concept remains the customar)' perspective. Calculating the capacitance o f
t^vo arbitrarily shaped conducting surfaces separated by a dielectric is, in general, ver)- difficult
Fortunately, the ordinary capacitor o f a practical circuit is o f the parallel-plate variety, with the
plates separated by a thin dielectric. The two plates are often rolled into a tubula, ruu„, .,„>1
complete structure is sealed.
ic(t)
v Jt )
2mF
(a)
Time in milli-seconds
(b)
FIG URE 7.16 (a) A 2 mF capacitor connectcd to a voltage source,
(b) Plots of capacitor voltage and current waveforms.
S o lution
A direct application o f equation 7.5 yields
Exercises. 1. In Figure 7.16, suppose =e V for r > 0. Compute for r > 0. Sketch
(preferably in MATLAB) i^^t) for 0 < r < 0.5 sec.
ANS\V1-:R: - 0 . 0 5 , A.
2. Repeat Exercise 1 with = e~-^^ cos( 100/) V for f > 0 but plot over die time interval [0, 0.15 sec].
A.\’S\V1-:K: -.- -^ q 0 .0 5 COS. KJOr) r 0.2 sin(lOOr)] A.
The differential relationship o f equation 7.5 has the equivalent integral form
(7.6)
C •'^0
where is in volts, /^r) is in amps, and C is in farads, and where we have taken =0
because the capacitor was not manufactured at t = -oo. The time /q represents an initial time o f
interest or significance, e.g., the time when the capacitor is first used in a circuit. The quantit)'
286 Chapter 7 • Inductors and Capacitors
specifies the initial voltage across the capacitor at ^q. This initial voltage, sums up the entire
past history o f the current excitation into the capacitor. Because o f this, the capacitor, like the
inductor, is said to have memory.
EXA M PLE 7 .7
Suppose a current source with sawtooth waveform shown in Figure 7.17h, drives a relaxed
0.5 F capacitor (zero initial voltage) as in the circuit o f Figure 7.17a. Compute and plot the volt
age across the capacitor.
i.(t) A/v^(t)V
(a)
(b)
S o l u t io n
The input waveform is periodic in that it repeats itself every 2 sec. Therefore, the solution will pro
ceed on a segment-by-segment basis.
Step 1. Consider the interval 0 < f < 2. For this interval ij^t) = {It - 2) A. With = 0, it fol
lows from equation 7.6 that
Step 2. Consider the interval 2 < t <A. Observe that at / = 2, = 0; hence, the capacitor volt
age over the interval 2 < /“< 4 is simply a right-shifted version o f the voltage over the first inter
val. Right-shifting is achieved by replacing t with t - 2 . In other words,
Step 3. Consider the general interval 2k < t < 2{k + 1). For interval 2k < t < 2{k + 1),
Lastly, obser\^e that the voltage across the capacitor, as illustrated in Figure 7.17b, is continuous
despite the discontinuity o f the capacitor current. Again, this follows because the capacitor volt
age is the integral (a smoothing operation) o f the capacitor current supplied by the source.
Chapter 7 * Inductors and Capacitors 287
Exercise. Consider the capacitor circuit o f Figure 7.18. Suppose the current source is i^{t) = e~‘ A
for /■> 0 and = 1 V. Compute the capacitor voltage the resistor voltage and the
voltage vj^t) across the current source for r > 0.
-H V„(t) -
+ 20
I (t) = e-'u(t) 0 M O ; v,(t)
. 0.5 F
FIGURE 7.18 Scries RC circuit driven by a currcnt source for accompanying exercise.
AN SW ERS: v^^t) = 3 - for t> 0, = lc~' for t > 0. and, by K\'l., v^{t) = 3 V for t > 0.
It is important to emphasize that the sawtooth current input depicted in Figure 7.17b is a dis
continuous function, but the associated voltage waveform is continuous because integration
(equation 7.6) is a smoothing operation. This means that the capacitor voltage is a continuous
function o f t even for discontinuous capacitor currents, provided they are bounded. This obser
vation leads to the continuity property o f the capacitor: if the current i(^t) through a capacitor
is bounded over the time interv'al < ^ < ^2> then the voltage across the capacitor is continuous
for fj < r < tj. In particular, for bounded currents, if fj < < tj, then = V(- (tQ"^), even when
At the macroscopic level, there appear to be some exceptions to the continuit}' propert}' o f the
capacitor voltage, e.g., when two charged capacitors or one charged and one uncharged capacitor
are instantaneously connected in parallel. In such cases, KVL takes precedence and will force an
“instantaneous” equality in the capacitor voltages, subject to the principle o f conservation o f
charge, to be discussed shortly. Another example is w'hen capacitors and some independent volt
age sources form a loop. When any o f the voltage sources has an instantaneous jump, so will the
other capacitor voltages. Upon closer examination, however, we see that there is really no excep
tion to the stated continuity rule: it can be shown that in all o f the cases where the capacitor volt
age jumps instantaneously, an “impulse” current flows in the circuit. Physically, an impulse cur
rent is one that is ver)' large (infinite from an ideal viewpoint) and o f very short duration. The cur
rent is not bounded, and consequently, the capacitor voltage may jump instantaneously. This
jump does not violate the rule, which presumes that the currcnt is bounded.
Relatio7ishipofChargetoCapacitorVoltageandCurrent
We have defined the capacitance o f a two-terminal device strictly from its terminal voltage-current
relationship— the differential equation 7.5 and the integral equation 7.6, which is now repeated:
v 'c ( 0 = V c(fo) + ^ f ^
C •'M)
Physically speaking, the integral o f i(^t) over [/q, t\ represents the amount o f charge passing
through the top wire in Figure 7.19 over [rQ, r].
288 Chapter 7 • Inductors and Capacitors
ic(t)
/ ''+ + +q + + /
/+ + + ^
/ ++++++
A- + + + + + +y
-q
FIGURE 7.19 Capacitor cxcitcd by a currciu.
Bccausc o f the insulating' material (the dielectric), this charge cannot pass through to the other
plate. Instead, a charge o f +q{t) is stored on the top plate, as shown in Figure 7.19. By KCL, if
i(it) flows into the top plate, then must flow into the bottom plate. This causes a charge
of to be deposited on the bottom plate. The positive and negative charges on these two
plates, separated by the dielectric, produce a voltage drop from the top plate to the bottom
plate. For a linear capacitor, the only t)'pe studied in this text, the value o f V(^t) is proportional
to the charge The proportionalit)' constant is the capacitance o f the device. Specifically,
where q{t) is in coulombs, Cis in farads, and t^i^) is in volts. Thus, equation 7.6 has the following phys
ical interpretation: the first term, is the capacitor voltage at /q; the integral in the second term,
ic(T )d r.
represents the additional charge transferred to the capacitor during the interval [r,j, /]. Dividing
this integral by Cgives the additional voltage attained by the capacitor during [^q, ^]. Therefore,
the sum o f these rwo terms, i.e., equation 7.6, is the voltage o f the capacitor at r. Since q(/) =
it follows direcdy that
= (7.8)
(It cl!
ThePj'hicipleofConservationofCharge
It is important in terms of modern trends in circuit applications to further investigate the rela
tionship o f charge to capacitor voltages and currents. The principle o f conservation o f charge
requires that the total charge tramferred into a junction {or out o f a junctiori) be zero.~ This is a direct
consequence of KCL. To exemplify, consider the junction o f four capacitors shown in Figure 7.20.
Chapter 7 • Inductors and Capacitors 289
v,(t)
+ V ,(t) - i3(t)
i,(t) - V3(t) +
v,(t)
+
l4(t)
FIGURE 7.20 Junction of four capacitors.
By KCL
/,(/) + ijit) + i^{t) + i^{t) = 0
Since the integral o f current with respect to time is charge, the integral o f this equation over
( - 00, t] is
where qj^{t) is the charge transferred to capacitor k. By equation 7.6, at ever}' instant of time,
which defines the relationship between transported charge, capacitance, and the voltage across the
capacitor. Hence, from equations 7.9 and 7.10, at every instant of time,
C^v^it) + C2 V2 U) + + C^v^{t) = 0
This simple equation relates voltages, capacitances, and charge transport. The following example
provides an application o f these ideas.
EXA M PLE 7 .8
This example shows that under idealized conditions, capacitor voltages can change instanta
neously. Consider the circuit o f Figure 7.21 , in which £^q(0~) = 1 V and = 0 V. Find
and V(^{t) for f > 0.
290 Chapter 7 • Inductors and Capacitors
t=0
C l = 1F
Cl C2
C2 = 1F
S o lution
At t = 0“ , the charge stored on C, is C, ^ ^ (0 “ ) and that o f C j is For t > 0, KVL
requires that Therefore, after the switch is closed at r = 0, some charge must be
transferred between the capacitors to equalize the voltages. According to the principle o f conser
vation o f charge, the total charge before and after the transfer is the same. Thus, conservation o f
charge requires that
C, = 0
C, v - c , ( 0 ^ ) - l l + C 2 [ i ’c i ( 0 ^ ) - 0 = 0
Exercises. 1. In Example 7.8, make C, = 0.75 F and C2 = 0.25 F, and compute ^/^(O'^).
AN SW ER: /Y-,(0") = 0.75 V.
2. In Example 7.8, sufipose ^’q ( 0 " ) =10 V and Vqj,{Q~) = - 8 V. Also let C, = 0.75 F and C-, = 0.25
F. Compute
ANSW ER: /.v.,(0^) = 5.5 V.
Example 7.8 is illustrative o f a charge transport that is germane to switched capacitor circuits,
which are o f fundamental importance in the industrial world.
EnergyStorageinaCapacitor
As with all devices, the energy stored or utilized in a capacitor is the integral o f the power absorbed
by the capacitor. The net energ}' entering the capacitor over the interval [/q, /J is
ch’ciT )] Vcih)
= c r '' ( dT = (7.11)
(It Oo)
= -C
1
for C in farads, in volts, and energy in joules (J). From equation 7. 11, the change in energy
stored in the capacitor over the inter\'al [rQ, r j depends only on the values o f the capacitor volt
ages at times /q and i.e., on and v^t^. This means that the change in stored energy is
independent o f the particular voltage v/aveform between and r,. If the voltage waveform is peri
odic, i.e., if V(\t) = + T) for some r > 0, then over any time interval [t, / + 7], the change in
the stored energy in the capacitor is zero because + 7) = forces equation 7.11
to zero. Analogous to the inductor, for all periodic voltages, the capacitor stores energy and then
returns it to the circuit and is thus called a lossless device.
As with the inductor, it is convenient to define the instantaneous stored energy in a capacitor as
which is really the integral o f power over the interval (—x , /], assuming that all voltages and cur
rents are zero at r = - x .
E X A M PLE 7 .9
Consider the circuit o f Figure 7.22, in which = 0. It is known that
for f > 0, the source current is i^{t) and the voltage across the capacitor is ) - 4/? \ - e
V for ^ > 0. Compute (i) the energ)', in joules, stored in the capacitor for
/ > 0, (ii) and (iii)
S o lution
(i) Since = 0, from equation 7.11 (or 7.12),
VV^^(0,/) = lc v J ( / ) = 8 C / ? \-e RC
= C ^ ^ ^ = 4— A
<li RC
2 ‘)2 Chapter 7 * Inductors and Capacitors
\ -e A
Thus
1- e- + 4^ = 4 A
ijt )
E X A M PLE 7 .1 0 — ►
/ \
For the circuit o f Figure 7.23a, it is known that the voltage across
the capacitor is = 20sin(2f + rr/6) V for r > 0. Compute and +
plot the instantaneous power absorbed by the capacitor and the
energy stored by the capacitor during the time interval [0, f]. N
5mF
k -
T im e t in seconds
(b)
F IG U R E 7.23 (a) Capacitor with known voltage v^^t) connectcd to a network N. (b) Plot of power,
and the net cncrg)', t), stored over the interval lO, ^].
Chapter 7 * Inductors and Capacitors
S o lution
Step 1. Compute From equation 7.5, for / > 0
,-^(,) = C ^ ^ = 0 .2 co s 2t + -
dt 6}
Step 2. Computep(\t). By direct multiplication and a standard trig identit}',
/ k\ / k\ {A
P c(t)= V(-(/)/(-(/)= 20 sin 2 / + - X 0. 2 cos 2/ + - = 2 sin 4t + - watts
l 6j 6j \ 3!
Plots o f pf4f) i^nd V\^^0, t) arc given in Figure 7.23b. Notice that WT^O, /) can be negative,
because W^—oo, 0) = 0.25 joules, meaning that at r = 0, there is an initial stored energy that can
be returned to the circuit at a later time . Figure 7.23b substantiates this.
Sei'iesInductors
Just as resistors in series combine to form an equivalent resistance, inductors in series combine to
form an equivalent inductance. As it turns out, series inductances combine in the same way as
series resistances.
E X A M PLE 7 .1 1 .
Compute the equivalent inductance o f the series connection o f three inductors illustrated in
Figure 7.24. Then find the voltages as a fraction o f the applied voltage
Leq
o
+
V ..
+
Leq
Vl2 Leq
+
o-
+
V .,
^eq
Q- o -
(a) (b)
FICIJRH (a) Scries connection of three inductors, (b) Equivalent inductance.
29-4 Chapter 7 • Inductors and Capacitors
S o lution
First we must answer the question o f what it means to be an equivalent inductance. Earlier, we
defined the inductor in terms o f its terminal voltage-current relationship. Two 2-terminal induc
tor circuits have the same inductance if each circuit has the same terminal voltage-current rela
tionship as defined in equation 7.1.
Step 1. The voltage labeled appears across the series connection, and, by KCL, the current
il flows through each o f the inductors, i.e., = ‘l y equivalent inductance,
is defined by the relationship
= (7.13)
dt
O ur goal is to express in terms o f Z j, L-,, and Ly
Step 2. Find in terms ofi^^^. To obtain such an expression, observe that, by KVL,
dt
it follow's that
^'U’q - (^1 + ^2 + ^ 3 )
dt
Hence, the series inductors o f figure 7.24a can be replaced by a single inductor with inductance
= -^-1 + ^2 +
_ Lj
dt (L| 4- Zy-> -l- L-^) ^
Exercises. 1. If, in Example 7. 11, Z,, = 2 mH, = 5 mH, and = 1 mH, find .
AN SW ER: = 8 mH. “
2. Find in terms o f
3
ANSW ER: '■/.:=
S
Chapter 7 • Inductors and Capacitors 295
Extension o f the formulas in the above example to n inductors is fairly clear, and we state the
results without rigorous proof: the formula for series inductances is
(7.14a)
^'IJ = (7.14b)
L\ + Lo + ... + Lfj
InductorsinParallel
The same basic question as with inductors in series arises with a parallel connection o f inductors:
what is the equivalent inductance? Rather than derive the general formula, let us consider the case
o f three inductors in parallel, as illustrated in Figure 7.25a.
E X A M PLE 7 .1 2
For this example our goal is to show that the equivalent inductance o f the circuit o f Figure 7.25a
is given by the reciprocal o f the sum-of-reciprocals formula,
- ~\ i T
— +— +• (7.15)
U Ly
Leq Leq
O -
L2, L3 +
Leq ^eq
L, L,
o - o-
(a) (b)
FIG URE 7.25 (a) Parallel connection of three inductors, (b) Equivalent inductance.
So l u t io n
Once again, equation 7.13 defines the relationship for the equivalent inductance:
^U-q ^eq
The goal is to construct in terms o f Z ,, and in a way that satisfies equation 7.13. This
will produce equation 7.15.
Step 1. Write KCL for the parallel connection shown in Figure 7.25a. Here, by KCL,
‘U q = 'Z.1 + + 'L 3
2% Chapter 7 • Inductors and Capacitors
_ (Hl\ ^ I
ill dt dt dt
di,
Step 2. Find — Ul in terms o f a n d L^,. From equation 7.1, For each inductor
dt
dt Li
Substituting into the result o f step 1 and noting that ~ ^i\ ~ ^L1 ~ ^L5
di, ( \ 1 n
— +— +— Li-q
dt Ln L'S L] L t^)
This has the form o f equation 7.13, which implies equation 7.15, i.e.,
Le, = 1 1 1
— + ---- + —
Z/j L~i L,'^
Thus
(/)
Exercises. 1. If, in Flxample 7.12, Z.j = 2.5 mH, Z., = 5 mFl, and = 1 mH, find
ANSWq-R: = 0.625 mH.
The above arguments easily generalize. Suppose there are u inductors, /,,, Z-,, ... , Z.,^, connected in
parallel. Then the equivalent inductance is given by the reciprocal o f the sum-of-reciprocals formula.
(7.16a)
U
~i-------- [ ii.jn (7.16b)
Exercise. For two inductors Zj and L-, in parallel, show chat the equivalent inductancc satisfies the
formula
(7.17)
Series-ParallelCombinations
This subsection examines series-parallel connections o f inductors. This allows us to use the for
mulas developed above in an iterative way.
EXA M PLE 7 .1 3
Find the equivalent inductance, o f the circuit o f Figure 7.26.
S o l u t io n
Step 1. In the circuit o f Figure 7.26, several inductors are enclosed by an ellipse. Let denote
the equivalent inductance o f this combination. Observe that the series inductance o f the 5/6 H
and 0.5 H inductors equals 4/3 FI. This inductance is in parallel with a 1 H and a 4 Fi induc
tance. Hence,
I
4 , =
1 I 3= 1 H
—!----- !----
1 4 4
Step 2. The equivalent circuit at this point is given by Figure 7.27. This figure consists o f a series
combination o f a 1.5 FI and a 0.5 FI inductor connectcd in parallel with a 6 H inductor. It fol
lows that
29S Chapter 7 • Inductors and Capacitors
= = -= 1 .5 H
1 I 4
-------------+ -
0 .5 + 1.5 6
Exercise. In Example 7.13, suppose the 5/6 H and 0.5 H inductors are both changed to 0.4 H
inductors. Find L o f the circuit.
AN SW ER: 1.443 H.
CapacitorsinSeines
Capacitors in series have capacitances that combine according to the same formula for combining
resistances or inductances in parallel. Similarly, capacitances in parallel combine in the same way
that resistances or inductances in series combine. This means that the equivalent capacitance o f a
parallel combination o f capacitors is the sum o f the individual capacitances, and the equivalent
capacitance o f a series combination o f capacitances satisfies the reciprocal o f the sum-of-recipro-
cals rule. These ideas are illustrated in the examples to follow.
EXAM PLE 7 .1 4
Compute the equivalent capacitance, o f the series connection o f capacitors in Figure 7.28a.
Chapter 7 • Inductors and Capacitors 2‘)9
o-
Cl
'C2
o-
+
C3
eq
Q- o
(a) (b)
FIGURI^ 7.28 (a) Series combination of three capacitors, (b) Equivalent capacitance,
So l u t io n
The equivalent capacitance denoted in Figure 7.28b is defined implicitly by the current-voltage
terminal conditions according to equation 7.5, i.e.,
; - r ^
^ dt
Our goal is to express this same terminal v-i relationship in terms o f the capacitances, C j, C 2, and
Cy After this we set forth a formula for voltage division.
Step 1. Set forth the i-v relationship fo r each capacitor. For each capacitor, k = 1, 2, 3,
‘Ck - Q
dt
dt C,
Differentiating this expression with respect to time and using the result o f step 1 yields
\
iQ -
dt dt
H--------h --
[C l Cl C^j
It follows that
Q ,=
C, G C3
Thus
C, C2 C3
2. Find in terms of
.ANSWER: /V;: = 0 .1 6 /y ;
Generalizing the result o f Example 7.14, we may say that capacitors in series satisfy the reciprocal
o f the sum-of-reciprocals rule. Thus, for n capacitors C j, C2, ... , C^, connected in series, the
equivalent capacitance is
1
C =■ - 1 -------------- r (7.18a)
— + — + ... + —
C, G C„
Exercise. Show that if two capacitors Cj and C-y are connected in series, then
c (7.19)
c ,+ c .
CapacitorsinParallel
If rvvo capacitors are connected in parallel as in Figure 7.29a, there results an equivalent capaci
tance = Cj + C2 and a simple current division formula to be derived.
I. 'c
a o
+ ‘Cl 'C2 +
eq
a- a-
(a) (b)
FIGURE 7.29 (a) Parallel combination of two capacitors, (b) Equivalent capacitance,
Since the voltage appears across each capacitor, and since /^= /q + by KCL it follows that
Hence,
^eq - ^2
One surmises from the above example that, in general, capacitors in parallel have capacitances that
add. And, indeed, this is the case: if there are n capacitors C j, C2, ... , in parallel, the equiva
lent capacitance is
C = C, + C2 + ... + C„ (7.20a)
Ck . Cf, .
Ceq Q + Q
and that for n capacitors in parallel.
Q ( 7 .2 0 b )
'CA- - 'c
C] + C j +••• + C„
302 Chapter 7 • Inductors and Capacitors
Series-ParallelCombinations
Wc round out our discussion o f capacitance by considering a simple series-parallel interconnection.
EX A M PLE 7 .1 5
Consider the circuit o f Figure 7.30. Compute the equivalent capacitance,
0.45 mF
0.6 mF
S o l u t io n
Step 1. Combine series capacitances. Observe that the rwo series capacitances o f 0.5 mF and 0.5
mF combine to make a 0.25 mF capacitance.
Step 2. Combine parallel capacitances. First, as a result o f step 1, the three capacitances, 0.3 mF,
0.25 mF, and 0.45 mF, add to an equivalent capacitance o f 1 mF. Further, the two parallel capac
itances, 0.3 mF and 0.6 mF, at the bottom o f the circuit, add to make a 0.9 mF capacitance. The
new equivalent circuit is shown in Figure 7.31.
a- II mF
1.125 mF
a- 0.9 mF
FIGURE 7.31 Circuit equivalent to that o f Figure 7.30.
Exercise. Suppose the two 0.5 mF capacitors in Figure 7.30 are changed to 2.5 mF capacitors.
Find the new
A N SW ER: 0.4 mF.
Chapter 7 * Inductors and Capacitors 303
As mentioned in the chapter opener, a power supply converts a sinusoidal input voltage to an
almost constant dc output voltage. Sucii devices are present in televisions, transistor radios, stere
os, computers, and a host o f other household electronic gadgets. Producing a truly constant dc
voltage from a sinusoidal source is virtually impossible, so engineers design special circuits called
voltage regulators that generate a voltage with only a small variation between set limits for a given
range o f variation in load. The voltage regulator is a precision device whose input must be fairly
smooth for proper operation. A capacitor can provide a crude, inexpensive sm oothing function
that is often sufficient for the task. This section explores the design o f a capacitive smoothing cir
cuit. In practice, such a circuit is used only for low-power applications.
i.(t)
FIGURE 7.32 Simple power supply with capacitive smoothing for low- power applications.
Consider, for example, the circuit shown in Figure 7.32. The four (ideal) diodes are arranged in a
configuration called a fidl-wave bridge rectifier circuit. An ideal diode allows current to pass only
in the direction o f the arrow. The diode configuration ensures that i^{t) remains positive, regard
less o f the sign o f the source current. Specifically, the diodes ensure that /j(f) = Using the
integral relationship (equation 7.6) o f the capacitor voltage and current, it follows that
Because o f the difference |/j(r)| - i^it) inside the integrand o f the integral, i^{t) tends to increase
the capacitor voltage, whereas i^^{t) tends to decrease the capacitor voltage. Further, because the
diodes are assumed ideal, it follows that
To see this, suppose the opposite were true; i.e., suppose One o f the diodes would
then have a positive voltage across it in the direction o f the arrow. The diode is said to be forivarcl
biased. But this is impossible, because an ideal diode behaves like a short circuit when forward
biased. The consequence is that V(4,t) will be 12 V whenever |?>'^(^)| is 12 V. This occurs every 1/120
o f a second. Thus, the rectifier output will recharge the capacitor every 1/120 o f a second. Between
charging times, the current, i(){t), will tend to discharge the capacitor and diminish its voltage.
304 Chapter 7 • Inductors aiui Capacitors
The design problem for the capacitive smoothing circuit is to select a value for C that guarantees
that v^t) is sufficiently smooth to ensure proper operation of the voltage regulator. Here, “suffi
ciently smooth” means that the maximum and minimum voltages differ by less than a prescribed
amount. To be specific, suppose that i>(\t) must remain between 8 V and 12 V. Recall that i^{t)
tends to increase the capacitor voltage, while tends to decrease it. The design requires select
ing a value for C to ensure that i^{t) can keep up with so that the capacitor voltage remains
fairly constant. The value for /(,(^) is obtained from the specification sheet o f the voltage regulator.
Suppose this value is a constant 1 A. It remains to select C so as to ensure that V({t) remains above
8 V between charging times. From equation 7.21, it is necessary that
CJio
Now we need consider only values for t between 0 and 1/120, because the capacitor will recharge
and the process will repeat itself every 1/120 o f a second. Thus, because i^{t) will only increase the
capacitor voltage, to ensure that v^{t) remains above 8 V, it is sufficient to require that
1 A X 120 sec
= 2.083 mF
4 V
A 2,100 |.iF capacitor satisfies this requirement. A method for computing the capacitor voltage
waveform is described in Chapter 22 o f o f 2"^ edition. However, using SPIC E or one o f the other
available circuit simulation programs, one can generate a plot o f the time-varying capacitor volt
age produced by this circuit, as shown in Figure 7.33. In the figure, it is seen that the capacitor
voltage varies between 12 and 9.02 V, which is smaller than the allowed variation o f (12 - 8) V.
Two factors contribute to this conservative design: (1) we used C = 2,100 uF instead o f the cal
culated value, C = 2,083 uF, and (2) the increase in the capacitor voltage due the charging current
is is not included in the calculation.
volts
The preceding brief introduction made several simplifying assumptions to clarify the basic use o f
a capacitor as a smoothing or filtering device. Practical power supply design is a challenging field.
A complete design would need to consider many other issues, some o f which are the nonzero
resistance o f the source, the non-ideal nature o f the diodes, the current-handling abilit)' o f the
components, protection o f the components from high-voltage transients, and heat-sinking o f the
components.
6. SUM M ARY
This chapter has introduced the notions o f a capacitor and an inductor, each o f which is a lossless
energ}' storage device whose voltage and current satisfy a differential equation. The inductor has a
voltage proportional to the derivative o f the current through it; the constant o f proportionalit}^ is
the inductance L. T he capacitor has a current proportional to the derivative o f the voltage across
it; the constant o f proportionalit}' is the capacitance C. It is interesting to observe that the roles of
voltage and current in the capacitor are the reverse o f their roles in the inductor. Because o f this
reversal, the capacitor and the inductor are said to be dual devices.
That the (ideal) inductor and the (ideal) capacitor are lossless energy storage devices means that
they can store energ)- and deliver it back to the circuit, but they can never dissipate energ)^ as does
a resistor. The inductor stores energy in a surrounding magnetic field, whereas the capacitor stores
energy in an electric field between its conducting surfaces. Unlike energ)' in a resistor, the energy
stored in an inductor over an interval [r^, fj] is dependent only on the inductance L and the val
ues o f the inductor current //(/^()) and //(/^j). Likewise, the energ)' stored in a capacitor over an
interval [r^, /,] is dependent only on the capacitance C and the values o f the capacitor voltage
and
Both the inductor and the capacitor have memor)'. The inductor has memory because at a partic
ular time Tq, the inductor current depends on the past histor}' o f the voltage across the inductor.
The capacitor has a voltage at, say, time that depends on the past current excitation to the capac
itor. The concept o f memory stems from the fact that the inductor current is proportional to the
integral o f the voltage across the inductor and the capacitor voltage is proportional to the integral
o f the current through the capacitor. This integral relationship gives rise to the important proper
ties o f the continuity of the inductor current and the continuit)' o f the capacitor voltage under
bounded excitations.
rhe dual capacitor and inductor relationships are set forth in Table 7.1.
lABLE 7.1. Summary of the Dual Relationships ot the Capacitor and Inductor
ic « )
d\'c ) dilit)
icit) = C V[{t) = L
dt dt
Bounded voltage or current; voltage or current signal whose absolute value remains below some
fixed finite constant for all time.
Capacitance o f a pair o f conductors: a propert)' o f conductors separated by a dielectric that per
mits the storage o f electrically separated charge when a potential difference exists
between the conductors. Capacitance is measured in stored charge per unit o f potential
difference between the conductors.
Capacitor (linear): a two-terminal device whose current is proportional to the time derivative o f
the voltage across it.
C oil: another name for an inductor.
Conservation-of-charge principle: principle that the total charge transferred into a junction (or
out of a junction) is /.ero.
C ontinuity property o f the capacitor: property such that if the current i(\t) through a capaci
tor is bounded over the time interval < t < t-,, then the voltage across the capacitor is
continuous for /, < r < tj. In particular, if fj < < tj, then t^(^tQ~) = when
C ontinuity property o f the inductor: propert}' such that if the voltage across an inductor
is bounded over the time interval r, < r < then the current through the inductor is
continuous for r, < r < In particular, if r, < then /^(/‘o~) = when
Coulom b: quantit)' o f charge that, in 1 second, passes through any cross section o f a conductor
maintaining a constant 1 A current flow.
Dielectric: an insulating material often used between two conducting surfaces to form a capacitor.
Farad: a me;Lsure o f capacitance in which a charge o f 1 coulomb produces a 1 V potential difference.
Faradays law o f induction: law' asserting that, for a coil of wire sufficiently distant from any mag
netic material, such as iron, the voltage induced across the coil by a time-varying current
is proportional to the time derivative o f the current; the constant of proportionality,
Chapter 7 • Inductors and Capacitors 307
denoted Z, is die inductance o f the coil. Faradays law is usually stated in terms of flux
and flux linkages, which are discussed in physics texts.
H enry: the unit o f inductance; equal to 1 V-sec/amp.
Inductance: property of a conductor and its local environment (a coil with an air core or iron
core) that relates the time derivative o f a current through the conductor to an induced
voltage across the ends o f the conductor.
Inductor (linear): a two-terminal device whose voltage is proportional to the time derivative of
the current through it.
Instantaneous powen p(t) = */(/)/(<), in watts when v{t) is in volts and i{t) in amps.
Lossless device: device in which energy can only be stored and retrieved and never dissipated.
Lossy device: a device, such as a resistor (with positive R), that dissipates energy as some form o f
heat or as work.
Maxwell’s equations: a set o f mathematical equations governing the properties o f electric and
magnetic Beids and their interaction.
M emory: property o f a device whose voltage or current at a particular time depends on the past
operational history o f the device; e.g., the current through an inductor at time /q depends
on the history o f the voltage excitation across the inductor for t< /q.
Unbounded voltage or current: a voltage or current whose value approaches infinity as it nears
some instant o f time, possibly r = oo.
Voltage r^ u la to r: circuit that produces a voltage having only a small variation between set lim
its for a given range o f load variation from a fairly smooth input signal.
^The word “stored” emphasizes that the energy in the inductor is not dissipated as heat and can be recovered by
the circuit, whereas the word “absorbed” is used to mean that the energy cannot be returned to the circuit. In a
resistor, energy absorbed is dissipated as heat.
^ More generally, conservation of charge says that the total charge transferred into a Gaussian sur&ce (or out of a
Gaussian surface) is zero.
31)8 Chapter 7 • Induccors and Capacitors
-2 -
L(t)
© t(s) (a)
(a)
-2 + , 0.8 mH
Figure P7.2
(b) „(t)
i.JA )
(b)
Figure P7.5
6. (a) Find and plot for 0 < r < 6 sec the induc
Figure P 7.3 tor current /^(f) for the circuit o f Figure
Chapter 7 • Inductors and Capacitors 3 0 ')
P7.6a driven by the voltage waveform of (a) the voltage waveform sketched in Figure
Figure P7 .6 b. P7.8b, and
(b) Find and plot the instantaneous stored (b) the voltage waveform sketched in Figure
energy. P7.8c.
(c) Find and plot the stored energy
as a function o f time for 5 > ^ > 1.
(d) Find and plot t/j (/)
i,(t)
vJt)
(a)
(b)
Figure P7.6
Figure P7.8
+ v,(t) -
% iQ °
ic(t)
Figure P7.12
Figure P7.9
13. In Figure P7.13a, the capacitors C, = 4 mF
and C j = 12 mF are driven by the voltage
10. The circuit o f Figure P 7 .10 has two induc
specified in Figure P7.13b. Plot /q(/),
tors, Z., = 20 mH and Lj = 50 mH, in parallel.
and iA i).
The input is v^{t) = 200cos(5007rr) mV for t >
0 and zero otherwise. The switch between the
two inductors moves down at r = 4 ms.
Compute the currents and for 0 <
r < 4 ms and 4 ms < /. Also find the energy
stored in each inductor as a function o f t for the
same time intervals.
(a)
Figure P 7.10
O '
O '
Figure P7.16
w
(b) 17.(a) Consider the circuit sketched in Figure
0 P 7.17 in which Cj = 20 pF and C2 = 0.1
Figure P7.15
mF. Suppose v^{t) = 5sin(2000f) V for t
0 > 0 and suppose = 10 V. Find
16. For the circuit in Figure P7.16a, C = 0.25
for r > 0. Is the output voltage
mF. Compute and plot the waveforms of the
0 independent of the initial voltage on Q ?
voltage, given as sketched in Figures
Why?
P7.16B and c.
0
(b) W hat is the instantaneous power deliv
ered by the dependent source?
w
(c) Find the energy stored in Cj over the
interval [0 , t].
0
i jt )
(a)
0
0 Figure P 7 .1 7
312 Chapter 7 • Inductors and Capacitors
(c) Compute and plot the energ)' stored in f = 0 . Compute V({t) for 0 < r < 2 sec and 2 sec
< t when C, = 4 0 0 mF.
C,.
t = 2s
Figure P7.22
rent waveform = 200fr“ ’^' mA for ^ > 0 27. For the circuit o f Figure P7.27,
and 0 otherwise. (a) Compute as a function o f i^{t) and
(a) Compute and sketch I'lit), and the capacitances Cj and C j.
(b) Now find in terms o f i^{t) and the
(b) Compute and sketch the energy stored circuit parameter values.
in the inductor for r > 0.
(c) Compute and sketch the energy stored
L. +
in the capacitor for t> 0.
Figure P7.27
SERIES-PA RALLEL IN D U C TO R S
29. In the circuit o f Figure P7.29, all inductors
are initially relaxed at /^= 0 and /.j = 6 mH, L-,
i.(t)
Figure P 7 .2 6
314 Chapter 7 • Inductors and Capacitors
/ Y Y V
L. -200 --
Figure P7.32
(!)
33. In Figure P7.33, Z-j = 5 mH, L-, = 20 mH,
= 20 mH, = 80 mH, and ij^t) =
lOsin(lOOOr) mA for r> 0 and 0 otherwise.
Figure P7.30 (a) With the switch in position C, find the
C H EC K : 12 mH, - 1 .4 4 sin(lOOOr) V, 90 equivalent inductance, Vj^, and
cos(lOOOr) mA. (b) Repeat part (a) with the switch in posi
tion D.
31. For the circuit o f Figure P7.31, Z., = 260
mH, = 26 mH, L-^ = 39 mH, and =
10^"^“ tiiA.
(a) Find and i[^2 -
(b) Compute and
(c) Com pute the instantaneous energy
stored in as a function o f t.
32. Repeat Problem 31 for the waveform of CH ECK: L = 20 mH; /.„(1) = /.„(3) = 0.8 A
Figure P7.32. while /y,/2) = =0
Chapter 7 * Inductors and Capacitors 31
ijt )
____rvY V
v X L. L,
i„(t)
1-6 L,
f r r \ ____ T Y Y \
(a)
16'
0 1 2 3 4
16-
L. L,
(b)
rOA___ TYYV
Figure P7.34 (b)
Figure P7.36
SCRA M BLED ANSWERS: 0.1, 0.08, 0.6 (in 11)
38. Find L for each o f the circuits in Figure AN.SWER: C//I for
. all / values.
P7.38. 40. Like Problem 39, this is a conceptual prob
lem and requires no calculations for the answer.
/ Y Y Y
4 mH lOmH 1 mH I Consider circuits 1 and 2 o f Figure P7.40. All
inductors are 1 H except the one labeled L. We
36 mH 3mH' 5 mH' wish to determine the relationship bet\veen
Bo- and L^^-, in the presence o f the finite posi
(a) tive inductor o f L henries between points a and
b. W hich of the following statements is true?
/YYV
7mH
(h) < /v,2-
/ Y Y V / Y Y V
2.4 mH 1.2 mH
(d) There is no general relationship between
0.6 mH
^eq\ relationship depends
on the value o f L
(b) Explain your reasoning.
Figure P7.38
ANSWHR: (a) 13 m il: (b) 2 in 11
<!>
Vw/' -C,-
- ^ 1
C. c,
Figure P7.43
CHECKS: 6 pF, 66 ^F
4 2 .(a) Find the indicated equivalent capaci
+ If
c.
if if
c,
if
tance for the circuit o f Figure P7.42a c, c
assuming Cj = 48 pF, Cj = 16 |jF, =
20 )jF, Q = 80 pF, and C5 = 8 pF.
(b) Repeat for Figure P7.42b assuming Cj = Figure P7.44
3 pF, C2 = 6 pF, C3 = 3.6 pF, Q = 6 pF,
C5 = 4.5 pF, Q = 48 pF, = 48 pF, Cg 4 5 . Three 12 pF capacitors are available for
= 24 pF, Cg = 24 pF. interconnection. List all equivalent capaci
tances obtainable over all possible interconnec
tions of these capacitors.
O ’
o-
c C -L C, 4 6 . This is a conceptual problem and requires
c. no calculations for the answer. Consider cir
cuits 1 and 2 o f Figure P7.46. All capacitors are
(a)
1 F except the one that is labeled C. We wish to
determine the relationship between S ' and
^eql presence o f the finite positive C F
capacitor between points a and b. Which of the
following statements is true?
(b) ^ ^eq2 '
Figure P7.42 (W
^eq\ - ^eql-
43. Find for the circuit of Figure P 7.43, (a) (d) There is no general relationship between
when the switch is open, and (b) when the and Any relationship depends
switch is closed, assuming that Cj = C4 = 12 pF, on the value of C.
C2 = C5 = 40 pF, C3 = Cg = 2 0 pF. Explain your reasoning.
(c) Repeat parts (a) and (b) for Cj = 12 pF, Cj
= 40 pF, C3 = 20 pF, C4 = 4 0 pF, C5 = 20 pF,
Cg = 100 pF.
318 Chapter 7 • Inductors and Capacitors
c, V,
v jt)
©
Figure P7.48
C H EC K ; = 6(1 - V
Circuit 2
Figure P7.46
ANSWMR: C ;„ < v jt)
©
47. In Figures P7.47a and b, the charge on Cp
and C3 is Q = 72 X C. hi Figure
P7.47a, the voltages on C ,, C , and are 2 V, Figure P7.49
3 V, and 4 V, respectively; and in Figure 7.47b
Vq = A \ while the charges on Cp C , and 50. In the circuit o f Figure P7.50, suppose C,
are Qj = 48 x 10“^ C, Q 2 = ^0 ^ ^0“^ C, and = 5 mF, Cj = 20 mF, = 4 mF, Q = 80 mF,
Q 3 = 72 X C, respectively. = 10 0 e~^^ V for f > 0 and 0 otherwise.
(a) Find for the circuit o f Figure P7.47a. (a) Find and for r > 0 .
(b) Find for the circuit o f Figure P7.47b. (b) Compute the energy stored in the C-, over
the interval [0 , t].
/ v,„(t)
(a) (b)
Figure P7.50
Figure P7.47
ANSWHK: (a) = 8 mF 51 . In the circuit o f Figure P7.51, suppose Cj
= 4 mF, C-) = 80 mF, C3 = 20 mF, and =
48. In the circuit o f Figure P7.48, C| = 6 mF, 1OO^*”^^ mA for t > 0 and 0 otherwise.
C2 = 12 mF, C3 = 36 mF, v J t ) = 20(1 - (a) Find /q(^) and for t > i).
V for t> 0 and 0 otherwise. If all capacitor volt- (c) Compute the energy stored in C-, over
the interval [0 , t].
Chapter 7 • Inductors and Capacitors 31')
v^(/) = - WYiLit)
at
Figure P7.52
6 0.5H L(t)
(a)
C H EC K : v,„,,(/) = :^ v ,(/ )
Exponential Approximation
Ideally, the sawtooth voltage increases linearly with time until reaching a threshold where it imme
diately drops to zero, which reinitiates the process. The threshold voltage corresponds to a fixed
unit o f time. The linear voltage increase then acts as an electronic second hand, ticking o ff the
322 Chapter 8 • First Order RL and RC Circuits
smaller units o f time. In practice, the linear increase in voltage is approximated by the “linear” part
o f an exponential response o f an RC circuit. ^X1^en the voltage across the capacitor reaches a cer
tain threshold, an electronic switch changes the equivalent circuit seen by the capacitor, allowing
the capacitor to discharge ver)^ quickly, i.e., the capacitor voltage drops to zero almost instanta
neously. Once the voltage nears zero, the electronic switch reinstates the earlier circuit structure,
causing the capacitor to charge up again. The process repeats itself indefinitely.
CHAPTER O BJECTIVES
SECTIO N HEADIN GS
1. Introduction
2. Some Mathematical Preliminaries
3. Source-Free or 2^ro-Input Response
4. D C or Step Response o f First-Order Circuits
5. Superposition and Linearity
6 Response Classifications
7. Further Points o f Analysis and Theory
8 . First-Order RC Op Amp Circuits
9. Summary
10. Terms and Concepts
11. Problems
1. IN TRO D U CTIO N
Our study prior to Chapter 7 focused exclusively on resistive circuits. Recall that all nodal equa
tions and loop equations for resistive circuits lead to (algebraic) matrix equations whose solution
yields node voltages and loop currents, respectively. Chapter 7 then introduced the capacitor and
the inductor. Interconnections o f sources, resistors, capacitors, and inductors lead to new and fas
cinating circuit behaviors. How? Inductors and capacitors have differential or integral voltage-cur
rent relationships. Interconnecting resistors and capacitors or resistors and inductors leads to cir
cuits that must satisfy both algebraic (KVL, KCL, and Ohm’s law) and differential or integral rela
tionships for L and C values. When only one inductor or one capacitor is present along with resis
tors and sources, these relationships lead to first-order RL and RC circuits. When the sources are
Chapter 8 • First Order RL and RC Circuits
dc, such circuits have vohages and currents o f the form A + Be~^ for constants A, B, and X. The
main purpose o f this chapter is to develop techniques for computing the exponential responses o f
first-order RC and RL circuits driven by dc sources. A simple example serves to explain some of
these points.
In the series RC circuit o f Figure 8.1, suppose an initial voltage is present on the capacitor,
where 0~ designates the instant immediately before zero. Often vve distinguish among 0“, 0, and
O'*' when switching occurs or when discontinuities o f excitation functions occur at r = 0.
R
+
v(t)
© v,(t)
v,(/) = +Vc(/)
at
Dividing through by RC yields the constant-coefficient first-order linear diflferential equation
( 8 .2 )
dt RC RC
subject to the initial condition Vf^Qr). This equation says that the derivative o f the capacitor volt
age plus MRC times the capacitor voltage equals MRC times the source voltage. The equation
enforces constraints on the capacitor voltage, its derivative, and the source voltage, and is differ
ent from the algebraic node or loop equations studied earlier. The terminology first-order differ
ential equation applies because only the first derivative appears. Equation 8.2 is linear because it
comes from a linear circuit. Our goal is to find capacitor voltage waveforms that satisfy the con
straints imposed by the differential equation 8.2.
Exercise. For the circuit o f Figure 8.1, show that the capacitor current i({t) satisfies a differential
equation o f the form
cti.it) 1 . 1 dv,{t)
dt RC R dt
32 4 Chapter 8 • First Order RL and RC Circuits
Our scope in this ciiaptcr is limited to circuits containing one inductor or one capacitor— equiv
alently, first-order RL or RC circuits. W ithin this category we further constrain our investigation
to circuits with no sources but nonzero initial conditions, circuits driven by constant (dc) sources,
circuits driven by piecewise constant sources, and circuits containing switches. First-order circuits
driven by arbitrary source excitations are covered in later chapters using the Laplace transform
method.
Ver)' often our interest is in source excitations such as v^{t) = 2e~^’ V for /> 0 and 0 otherwise. To
conveniently represent such time-restricted waveforms, we define a signal called the unit step
function, denoted by u{t), as
1
«(/) =
0 / < ()
The unit step function is a universally used function and will appear many times in the remain
der o f this text. MATLAB code for specifying the step function is
function f = ustep(t)
t = t + le-12;
f = (sign(t)+l)*0.5;
With the unit step so defined, v^{t) - 2e~"'u(t) V, and both relations are plotted in Figure 8.2.
Further, if v<^t) = l e for t>t^ and 0 for r < /q, then v^{t) = lOf’ r^) would be the prop
er representation because the shifted unit step function, //(/- /q), means
Plots o f v^{t) = 2e~^‘u{t - ^q) and u{t - t^) are given in Figure 8.3 for = 0.5.
Exercise. Plot //(—/) and «(/q — t). Hint: For what values o f t are the functions zero and for what
values are they 1?
dxU)
(8.3b)
dt
valid for t> /q, where a-(/q) = is the initial condition on the differential equations 8.3. T he term
J{t) denotes a forcing function. Usually, y(r) is a linear function o f the input excitations to the cir
cuit.
326 Chapter 8 • First Order RI. and RC Circuits
The parameter X, denotes a riaturalfrequency o f the circuit. Natural frequencies are natural modes
o f oscillation such as, for example, in the ringing o f a bell. For physical objects natural frequen
cies are called natural modes o f vibration. All physical objects have a vibrational motion even
though it may be imperceptible. Knowledge o f these modes is important for the safety and relia-
bilit)' o f large buildings and bridges. For example, the Tacoma Narrows Bridge had natural modes
o f vibration that the wind excited. Undulations in the wind intensit}' resonated with the natural
vibrations of the bridge, causing a swaying motion to increase without bound until the bridge col
lapsed. In circuits, the natural modes o f oscillation are reflected in the shapes o f the voltage and
current w'aveforms the circuit produces. A more thorough and mathematical discussion o f the
notion o f natural frequency will take place in the next chapter, when we study second-order {RLQ
circuits.
Let us return to the goal of finding a solution to the differential equations 8.3. The solution to
equations 8.3 (a derivation will appear shortly) for t >tQ has the form
This means that the expression on the right-hand side o f the equal sign (1) satisfies the differen
tial equations 8.3 [its derivative equals K times itself plus/r)], and (2) it satisfies the correct ini
tial condition, xit^) = x^^. A simple example illustrates this point.
E XA M PLE 8.1.
Compute and verify the solution o f equation 8.3a using equation 8.4.
SO L U T IO N
Suppose in equation 8.3a., J{t) = u{t - 1), a shifted unit step function, X = - 1 , = 1, and at(1) =
10, in which case
cit
,v(/) = \)(k= + I
Chapter 8 • First Order Rl. and RC Circuits 32'
To verify that [9^ + 1] does indeed satisfy the differential equation, observe that for ? > 1,
dx{t) _ d
= -9 e + I = -,v(/) + 1
dt ~~dt
Further, at ^ = 1, [9f’ 1] = 10, which is the mandatory initial condition. Thus, x{t) = 9^
+ 1 is a valid solution for r > 1.
Example 8.1 spells out the application o f the solution (equation 8.4) to the differential equation
8.3a. It also verifies that the computed solution satisfies the differential equation and the proper
initial condition. Although not shown, equation 8.4 also satisfies equation 8.3b. A formal deriva
tion o f the solution o f equation 8.4 requires the use o f the integrating factor method, the sub
ject o f a differential equations course. Briefly, the first step o f this method entails multiplying both
sides o f equation 8.3a or 8.3b by a so-called integrating factor e~^. For equation 8.3b, this results
in
dx(t) _>j f/,\
e — ------ h e x i l ) = e f{ t ) (8.5)
dt
By the product rule for differentiation, the sum on the left equals
d_
e""x{t)
dt
( 8 .6)
dt
f -dr e dr = e
J'o
(8.7)
JI q
Bringing the term e'^‘Ox{tQ) to the right-hand side o f equation 8.7 and multiplying through by
results in the solution to the differential equation 8.3a or 8.3b, given by equation 8.4. This
completes the derivation o f the very powerful formula o f equation 8.4.
There are four points to remember about the preceding discussion: (1) circuits have behaviors
modeled by differential equations such as equations 8.3; (2) the solution to a first-order differen
tial equation is a waveform (also called a signal or response) satisfying equation 8.4; (3) the for
mula o f equation 8.4 works for all continuous and piecewise continuous time functionsy(/); and
(4) a solution to a differential equation means that the waveform satisfies the given differential
equation with the proper initial condition.
328 Chapter 8 • Hirst Order RL and RC Circuits
Exercise. Show that the hinction .v(/) = (1 — 0. ^or r > 0, is a solution to the difFerential equation
^ = -.v(/) + //(/) with initial condition a-(0) = 0 by showing that x{t) satisfies the difFerential
dt
equation and has the proper initial condition at r= 0.
Figure 8.4 depicts the most basic (undriven or source-free) RL or RC circuit: a parallel connection of
a resistor with an inductor or a capacitor without a source. In these circuits, one assumes the pres
ence of an initial inductor current or initial capacitor voltage. The complication introduced by a volt
age or current source is taken up later. Once the source-free or zero-input behavior is understood,
one can understand more easily the responses resulting from constant source excitations.
(a) (b)
FIGURK 8.4
Our first goal is to derive differential equation models for the RL and RC circuits o f Figures 8.4a
and 8.4b, respectively. We do this in parallel.
(1) At the top node o f Figure 8.4a, KCL implies (1) Similarly, for Figure 8.4b, KVL implies
Vi it) L d ii U) .d\'c{t)
n it ) — R icit) ——RC-
dt
(3) Making the obvious substitution and (3) Making the obvious substitution and
multiplying by R/L yields the differential dividing by RC yields the differential
equation model equation model
^ = hc{t) = - - m (8.9)
dt X
i.e., the derivative of x(t) is a constant, X = -1/x, times itself. Applying equation 8.4 to equation
8.9 implies that both equations 8 .8a and 8 .8 b have solutions given by
where x is a special constant called the time constant o f the circuit. Equation 8.10 means that the
responses for (q o f the undriven JiL and /?C circuits are, respectively, given by
W L
where the time constant o f the RL circuit is T = — and the time constant o f the RC circuit is x
= RC. ^
The time constant of the circuit is the time it takes for the source-free circuit response to drop to
e~^ = 0.368 o f its initial value. Roughly speaking, the response value must drop to a little over one-
third of its initial value. This is a good rule of thumb for approximate calculations involving decay-
ing exponentials.
The mathematics that underlie the solution to the differential equation 8.9 given in equation 8.10
is nothing more than elementary calculus. To see this, consider the exponential solution form
(8. 12)
where K 'ls an arbitrary constant. The fiinaion has the property that its derivative is----- e~' ^
This is precisely what equation 8.9 requires. Therefore equation 8.12 satisfies the differ- ^
ential equation 8.9 and is said to be a solution. To completely specify x{t) it only remains to iden
tify the proper value of K from the initial condition. Evaluating x(r) at ^ yields
Mt„) = /Cf-Vr
in which case
Substituting this value o f i n t o equation 8.12 produces the solution given in equation 8.10,
which is adapted to specific RL and R C circuits in equations 8 . 11. Figure 8.5 plots equation 8.12
for arbitrary K and x > 0 . This plot proves instructive for understanding how the response decays
as a function o f the time constant.
330 Chapter 8 • First O rder RL and RC Circuits
Time
FIGURI: 8.5 Plot o f equation 8.12. For f = x, one time constant, decays to 0.368
of its maximum value.
In summary, the circuits o f Figure 8.4 motivate the development o f the rudimentary machinery
for constructing solutions to undriven RL and RC circuits. For more general circuits, those con
taining multiple resistors and dependent sources, it is necessary to use the Thevcnin equivalent
resistance seen by the inductor or capacitor in placc o f the R in equation 8.11. Figure 8.6 illus
trates this idea.
f \ f \
Linear i,(t)
>f Linear
Resistive Resistive
Circuit Circuit +
L -N V jt)
No p C
No
Sources Sources
i,(t)
v,(t)
These facts imply that the general formulas for computing the responses o f undriven RL and RC
circuits have the structures
(8.13)
VrU) = e
The difference between equations 8.11 and 8.13 is that in equations 8.13 R^f^ is the Thevenin
equivalent resistance seen by the inductor or capacitor.
t = 0.4sec
S
+
20Q
v,(t)
5Q 8H
S o l u t io n
Step 1. With switch S open, compute the response for 0 < /< 0.4 sec. From the continuity property
o f the inductor current, /^(O'*’) = ^^(0“) = 10 A. Using equation 8.13,
'1 ( 0 = c A
Step 2. With switch S closed, compute the response for t > 0.4 sec. For this time interval the Thevenin
equivalent resistance seen by the inductor is = 20||5 = 4 Q, i.e., the equivalent resistance o f a par
allel 20 Q and 5 ^ combination. According to equation 8.13, the response for t>tQ = 0.4 sec is
,- ,( 0 = / t
Step 3. Write the complete response as a single expression using step fitnctions:
Step 4. Plot the complete response. To plot this using MATLAB, we use the following m-file along
with the code given earlier for the unit step function:
»t = 0:0.005:1.4;
»iL= 10*exp(-2.5*t) .* (ustep(t).* ustep(t - 0.4)) + 3.679*exp(-0.5*(t-0.4)) .* ustep(t - 0.4);
»plot(t,iL)
»grid
Using this code, Figure 8.8 illustrates the complete response, showing the two different time con
stants. The 0.4 sec time constant has a much faster rate o f decay than the lengthy 2 sec time con
stant.
a
E
<
Time (seconds)
vi^t) =
.IL l
For t> 0.4, however, the circuit structure changes and 4 in which case i'^(0.4'^) = 4 x 3 .679
= 14.716 V. Thus,
yi(f) = e (0 .4 "') = - 1 4 . 7 1 V
Chapter 8 • First Order RL and RC Circuits
Step 6. Compute the energy dissipated in the 5 O. resistor over the interval [0.4, oo). The power
absorbed by the 5 resistor for 0.4 < / is
v’i ( 0
= 43.31
5 5
The energ)' dissipated over [0.4, oo) is given by
Exercises. I. Plot v^it) using the above m-file, ustep, and the appropriate .code.
2. Repeat the calculations o f Example 8.2 with the 8 H inductor changed to 8 mH and a switch
closing time o f 0.4 ms.
A N SW ER: ij{t) = l{)r-“^'^"'//(f)//(0.4 x lO"-^ - r) + - 0.4 x Ur-^) A
E X A M PLE 8.3
Find V(^t) for r > 0 for the circuit o f Figure 8.9 given that y^^^O) = 9 V.
So l u t io n
Because there is a switch that changes position at r = 1 sec, there are two time intervals to consider.
Step 1. Compute the response forO < t < 1. Over this time interv\il, the equivalent circuit is a par
allel /?C circuit, as shown in Figure 8.10a.
334 Chapter 8 • First Order RL and RC Circuits
+
0.1 F 0.1 F
80 V jt) 3Q V jt)
(a) (b)
FIG URE 8.10 Equivalent circuits for Figure 8.9: (a) 0 :s r < 1 and (b) 1 s t.
1
------- 1
v c(t)= e Vc(O^) =9^’- ’ “5^V
Step 2. Compute the response for r > 1. Figure 8.10b depicts the pertinent equivalent circuit.
Observe that = 2.58 V and =3 Again by equation 8.11, for r > = 1,
r-I
-St-to)
V c (0 = f v .c (^ ) = 2 .5 8 e V
Step 3. Use step functions to specify the complete response. By using the shifted unit step function,
the two expressions obtained previously can be combined into a single expression:
t-\
V cir)=9e~^-'^'[u{t)-u{f-\)] + 2.5Se - \) V
Step 4. Obtain a plot o f the response. Using MATLAB and code similar to that used in Example
8.2, the plot in Figure 8 . 1 1 w'as obtained. Here the part o f the response with the 0.3 sec time con
stant shows a greater rate o f decay than the longer 0.8 sec time constant.
Time (seconds)
For all o f these examples x > 0 and the response is a decaying exponential. Intuitively, the response
decays because the resistor dissipates as heat the energy initially stored in the inductor or capaci
tor. One o f the homework exercises will ask the student to show that the total energ)' dissipated
in the resistor from to oo equals the decrease in energ)^ initially stored in the inductor or capac
itor at ^Q. When controlled sources are present, may be negative, in which case x < 0. Here the
negative resistance supplies energ)' to the circuit and the source-free response will grow exponen
tially. This is illustrated in the next example.
E X A M PLE 8 .4
Find Vf^t) for the circuit o f Figure 8.12, assuming that^^^^ = 0.75 S and i^(;(0“) = 10 V.
-o- -o-
b b
So l u t io n
It is straightforward to show that theThevenin equivalent seen by the capacitor is a negative resist
ance, - -2 Q, as shown in Figure 8.12b. Again, by equation 8.11,
v cit) = e vc{0^)=\0e^'u{t) V
Because o f the negative resistance, this response grows exponentially, as shown in Figure 8.13. A
circuit having a response that increases without bound is said to be unstable. Practically speaking,
an unstable circuit will destroy itself or exhibit a nonlinear phenomenon that clamps the voltage
at a finite value, as in the case o f saturation in an op amp.
336 Chapter 8 • First Order RL and RC Circuits
Tim e (seconds)
FIGURE 8.13 Plot of unbounded voltage response due to presence of negative resistance.
Circuits with such responses arc said to be unstable.
2. in Example 8.3, = 0.125 S. Find the equivalent resistance seen by rhc capacitor and
f > 0.
ANSWI-.l^S: 8 £2. V
3. Show that in general, for / > the form o f the capacitor current is similar to the voltage form.
Hint: Apply the capacitor v-i relationship to equation 8.11.
The circuits o f the previous section had no source excitations. This section takes up the calcula
tion o f voltage and current responses when constant-voltage or constant-current sources are pres
ent. It is instructive to start with the basic series RL and RC circuits as shown in Figure 8.14.
Chapter 8 • First Order RL and RC Circuits
O i^Ct)
Linear Linear
Resistive Resistive
Circuit Circuit
with with
Constant Constant
Sources Sources
O
(a) (b)
\{t)
FIGURE 8.14 (a) Driven first-order RL circuit, (b) Driven first-order RC circuit, (c) Thevenin
equivalent representation oF (a), (d) Thevenin equivalent representation of (b).
Given these basic circuit representations and initial conditions at Tq, what is the structure o f a dif
ferential equation mode! that governs their voltage and current behavior for t>tQ' The first objec
tive is to derive the “differential equation” models characterizing each circuits voltage and current
responses. It is convenient to use ij{t) as the desired response for constructing the differential
equation for the series RL circuit (Figure 8 .l4 c ), whereas for the series RC circuit (Figure 8 .l4 d ),
is the more convenient variable.
(1) The circuit mode! for the inductor is (i) The circuit mode! for the capacitor is
dilit) dv(^(t)
vl O )= L ic(t)= C
dt dt
^’oc -
ic (t) -
R,ill
(iii) Substituting for ;^(^) leads to the (iii) Substituting for /^r) leads to the differentia!
differential equation model equation model
d iijt) ^ R. dvc(r) _
+ (8-> 5a) -^ ^ 'C (0 + - — v o c (8 .1 5 b )
dt dt
338 Chapter 8 • First Order RL and RC Circuits
Exercise. Construct differential equation models for the parallel RL and RC circuits o f Figure
8 . 15 . Note that these circuits are Norton equivalents o f those in Figure 8.14a and Figure 8.14b.
Again choose /^(r) as the response for the RL circuit and v^^t) as the response for the RC circuit.
(constant) (constant)
A simple application o f basic circuit principles has led to the two differential equation models of
equations 8.15. The next important question is: What do these t%vo differential equation models
tell us about the behavior o f each circuit? Equivalently, how do we find a solution to the equa
tions? Observe that both differential equations 8.15 have the same structure:'
dx(t) 1
(8.16a)
dt X
where the time constant T = for RL circuits and x = R^j^C for RC circuits, and F= v J L for RL
circuits and F = vJiR^i^j for the RC case. This equation is valid for t> Equation 8.4, rewritten
here with/^y) = F, presents the general formula for solving the differential equation 8.16a:
t ( I-In
(t-tn
x( 0= f t + U 4 )- ft
” = -r (8.17)
which is valid for t > /q. After some interpretation, this formula will serve as a basis for comput
ing the response to RL and RC circuits driven by constant sources. A homework exercise will ask
for a different and direct derivation of this formula.
At this point it is helpful to interpret the quantity i r in equation 8.17. For RL circuits, when x(r)
= /*£(/), equation 8.15a implies that v J L , x = UR^f^ and hence F l = vJR^f^ = For RC cir
cuits when ;c(/) = V({t), equation 8.15b implies that F = vJR ^ C , x = Rf^,C, and hence Fz =
This interpretation is valid for both positive and negative values of x. If x > 0, then
t-tQ y
isc = for RL case
jr(oo) = lim x(t) = lim Fx + \ x {t^ )-F x y T = Ft = Rih (8.18)
/—*00 t~ * 0 0
for R C case
This means that for the RL case, /^(oo) = = vJR^f^ and for the RC case, V({<x>) = O f course,
is computed by replacing the inductor with a short circuit, and is computed by replacing
the capacitor with an open circuit. See Chapter 6 for details. Mathematically, any constant, such
as x(/) = constant, that satisfies a differential equation is called an equilibrium state of that dif
ferential equation. Since the constant x(/) = F i satisfies the differential equation 8.16, / r is an
o
equilibrium state of the differential equation 8.16.
Whenever X > 0, equation 8.18 implies that the formula (equation 8.17) for the solution o f equa
tion 8.1 6 given constant or dc excitation becomes
. _ ^-'o
a: ( 0 = j :(«>) + U ( ^ ^ ) - J c( oo) e (8.19a)
and when x(^) = /£(/),
(t-to )
iL (0 = / z .(“ ) + k ( 4 ) - ' L ( “ ) (8.19b)
and when x(/) = V({t)y t-t
o
K/.C
vc(0 = vc(00) + Vc(^o )-^ c (°°)
(8.19c)
Note that x > 0 is true whenever R^j^ > 0, C > 0, and Z, > 0, i.e., the circuit is said to be passive.
This allows us to state a nice physical interpretation of equation 8.19a:
elapsed time
w x { t )= [ F in a l value^ + i^Initial v a lu e ]-[F in a l value\)e tt^e constant
's . ;
340 Chapter 8 • First Order RL and RC Circuits
Elapsed Time
I'iG U RE 8.16 Graphical interpretation o f equation 8.19a for the case xico) > x{tQ).
Exercise. Redo the curve o f Figure 8.16 for the case x{<x)) <
The initial value computed from initial conditions and possibly the value o f the source
excitation, or it can be computed from past excitations up to Several examples will now illus
trate the use o f equation 8.19.
R = 5Q
v^(t) 1^(1-)=1 A
S o l u t io n
Step 1. Determine the circuit’s differential equation model. Since the circuit o f Figure 8.17 is a driv
en series RL circuit, equation 8 .1 5a implies that the differential equation model o f the circuit valid
for r > 1 is
cl'iAt) R 1 1 10
Step 2. Determine the form o f the response. Since /^(1“ ) = equation 8.19b implies that
/^(/)=/^(oo)+^/^(r) ^
Here the presence o f u{t - 1) emphasizes that the response is valid only for / > 1.
Step 3. Compute i^i^X)) and set forth the fin al expression for /^(r). Since x = 0.4 > 0, we replace the
inductor in the circuit o f Figure 8.17 with a short circuit to compute i^^. = //(oo) = 2 A. It follows
that
[2 + (1 - - 1) = (2 - - 1) A
Step 4 . Plot i^it). One cannot presume that the response is zero for r < 1. Hence, using MATLAB
or the equivalent, one can construct the graph o f i^{t) for / > 1 as given in Figure 8.18.
c
0;
3u
O
tj
3
■o
c.w
Time (sec)
Step 5. Compute v^it). Given the expression for the inductor current in step 3, it follows that for
t> 1,
/-I
cliLit)
Vi{t) = L- u{t -\^ ) = 5 e * ^ / (r -l'* ')V
dt
342 Chapter 8 • First Order RL and RC Circuits
Exercises. 1. Verify that in Example 8.5 v^it) can be obtained without differentiation by =
V -V iW -
2. In Example 8.5, suppose R is changed to 4 Q. Find i^it) at r = 2 sec.
ANSW ER: 1.8647 A
Note that we have used the differential equation 8.16 (or equations 8.15) to obtain the solution
form o f equation 8.19. However, when using equations 8.19, it is not necessary to reconstruct the
differential equation of the RL or RC circuit. Specifically, we need only compute xit^), x{cc), and
the time constant x = LIR^i^ or
The method described for computing final values can also be used to find the initial values o f
and i^ at f = if dc excitations have been applied to the circuit for a long time before t = ?q. The
next example illustrates this technique and extends the preceding discussion.
R = 6 kn
v,(t)
So l u t io n
Step 1. Compute initial capacitor voltage. For r < 0, the 12 V excitation has been applied for a long
time. Therefore, at r = 0“ , the capacitor has reached its final value and looks like an open circuit
to the source. Hence the entire source voltage o f 12 V appears across the capacitor at r = 0~, i.e.,
^(40~) = = 12 V by the continuity property o f the capacitor voltage.
Step 2. Use equation 8.19c to obtain v^it) fb r Q < t< 10 sec. Equation 8 .1 9c requires only that we
know V(iQ*) (step 1), x, and Vf^co). For 0 < t< 10 sec, r = R^C= 3 sec. It is important to realize
here that for 0 < r < 10 sec the circuit behaves as if the switch were not present. Hence, the com
putation o f v^(co) proceeds as if no switching would take place at r = 10 sec. Here v^^ = v^^oo) =
24 V. Hence, for 0 < ^ < 10, equation 8.19c implies
Chapter 8 • First Order RL and RC Circuits .vi3
Step 3. Compute the initial condition fo r the interval 10 < i.e., t'^^lO'^). Plugging into equation
8.20 and using the continuity property o f the capacitor voltage yields
= 2 4 - 12f>-io/3 = 2 3 .57 V
Step 4. Find V(^t) for / > 10. For r > 10, the resistive part o f the circuit can be replaced by its
Thevenin equivalent, which yields Figure 8.20.
Rtn = 2 k O
V(t)
FIGURE 8.20 Circuit equivalent to that of Figure 8.19 for / > 10.
Here, equation 8.19c applies again. The value for y^^co), however, is now 8 V and the new time
constant is = 1 sec. Hence, for ; > 10,
/-lo
v c (/ )= V c (“ ) + [v c (IO *)-V (-(= c )]e = 8 + ( 2 3 . 5 7 - 8 ) t '- * '- '" ’ = 8 + I5 .5 7 e ‘ * ' '" ” V
Step 5. Set forth the complete response using step functions. Using step functions, the response V(\()
for f > 0 is
v ^ {t)= 2 A -\ le + S + \5.51e^'
Step 6. Plot V(it). Plotting i^(^t) yields the graph of Figure 8.21.
3-m
Chapter 8 • First Order RL and RC Circuits
o
OJ
O)
Q.
u
Time (seconds)
Exercise. Suppose the switch in Example 8.6 opens again at r = 20 sec. Find v^it) at r = 25 sec.
ANSW ER: 20.98 \’
EXA M PLE 8 .7 The circuit o f Figure 8.22a has a capacitor voltage given by the cur\'e in Figure
8.22b. We note that ^f^O.l) = 7.057 V. Find, Kq, y^^O), the time constant r = RC, the exact value
of j/(j(Q.25), and the value o f C i f /? = 100^2.
------ O-
v.(t)= V „u (t)
v,(t)
-o-
(a)
Chapter 8 • First Order lU. and RC Circuits 345
01
IB
I
Q.
fO
U
T im e (s)
(b)
FIGURE 8.22 (a) Scries /?Ccircuit, (b) Capacitor voltage,
So lu t io n . A simple inspection o f the graph indicates that V(^{Q) = 2 V. One recalls that
/
) - \’(-(0C) e ^
Hencc as / -> oo, v^{t) -> v^co) = 10 V. Since the capacitor looks like an open circuit at / = oo, Vq
= = 10 V. From the given problem data,
y ^ 0 .1 ) = 7 .0 5 7 = 10
Simplifying yields
- 0.1
T = = 0.
/ 1 0 - 7 .0 5 7
8
When switching occurs frequently, or the excitation changes its constant level frequently, then
hand analysis, as in lixample 8.6, becomes very tedious. For such problems a SPIC E simulation
(or the equivalent) proves useful and saves time. The next example u.ses SPIC E to compute the
waveform o f a simple RC circuit whose input excitation is a square wave. Like the previous exam
ple, the solution is broken down into time intervals such that during each time interval inputs are
constant. Because no switching occurs, the time constants for all time intervals are the same. In
applying equation 8.19c, the quantities that vary from one time interval to the next are the initial
values and final values.
346 Chapter 8 • First Order RL and RC Circuits
(a) Plot for 0 < r < 60 ms, using SPIC E or equivalent software.
(b) Find the initial value and the final value in equation 8.19 when t is very large, for exam
ple, at the beginning and end o f the interval 1 < t< 1.01 sec. Plot the v^t) wave for this
interval using MATLAB or the equivalent.
V (t) (V)
■> t (msec)
10 20 30
(b)
FIG U RE 8.23 (a) Series RC circuit excited by the 50 Hz square wave o f (b).
So lu t io n
Part (a)
Doing a SPICE or equivalent simulation gives rise to the response curve shown in Figure 8.24, over
which the square wave input is superimposed. Observe that the response V(\t) has an approximate tri
angular shape. What is happening is that from zero to 10 msec, the circuit sees a step and hence the
capacitor voltage rises toward one volt. At 10 n:isec, the square goes to zero for 10 msec. The capacitor
then discharges its stored energ)' through the resistor, causing a decrease in its voltage value. The decrease
does not go to zero, however. So when the square wave again is at 1 volt the capacitor voltage begins to
rise again and achieves a slighdy higher value at f = 30 msec compared to f = 10 msec. In fact, one notices
in Figure 8.24 that the peak and minimum values are increasing slighdy as time increases. Eventually the
peak and minimum values will reach their respective fixed viilues, c;illed steady-state values. To find these
values, a simulation program could require a very lengdiy simulation interv'al, which often proves
impraaical. The steady-state values can be computed analytically as in part (b).
10 20 30 40 50 t(m se c)
FIG URE 8.24 Response of circuit of Figure 8.23a calculated using SPICE. For reference, the input
square wave excitation is superimposed on the plot.
Chapter 8 • First Order Rl. and RC Circuits _____________________________________________________
Part (b)
Let Tq = mT, where 7'= 20 msec is the period o f the square wave and m is some large integer. Then,
0.57-
+ 0 .5 7 ) = V(itQ)e^ (8.22b)
= 1 + [i/c(ro)-
= 0-2689 V
^ " 1+e
It follows that
r/J/o + 0 . 5 7 ) = i^(3<ro)e> = 0 .7 3 1 1 V
An examination o f the response in Figure 8.24 shows that the minimum and peak values are
approaching the steady-state values o f 0.26 8 9 V and 0.7311 V, respectively.
Exercise. Based on the response in Figure 8.24, roughly sketch the capacitor current, At
what time instants is the capacitor current discontinuous?
Superposition, a special case o f linearity, helps simplify the analysis o f resistive circuits, as discussed
in Chapter 5. Recall that linear resistive circuits are interconnections o f resistors and sources, both
dependent and independent. Does superposition still apply when capacitors and inductors are
added to the circuit? The answer is yes, provided one properly accounts for initial conditions.
In order to justify the use of superposition for RC and RL circuits, consider that resistors satisfy
O hm s law’, a linear algebraic equation. Capacitors satisfy the differential relationship
; - r ------
ir
^ dt
r\
348 Chapter 8 • First Order RL and RC Circuits
which is also a linear equation. To see linearity in this i-v relationship, suppose voltages and
V(^ individually excite a relaxed capacitor producing the respective currents
: ,■ _^ dvc2
‘a - c , >c2 - c —
Let be the current induced by a voltage equal to the sum o f and i.e.,
‘C3 = (^-^(^Cl+^C2)
. _ dvf
^
ic=C- + C — — = ic i + t 2
dt dt
By the same arguments, the current due to the input excitation t/Q - ^\^c\ ^2^Cl ^C3 “ ^l^Cl
+
On the other hand, suppose two separate currents / q and individually excite a relaxed capac
itor C Each produces a voltage given by the integral relationship
By the distributive property of integrals, the combined effect of the input, + <?2^C2’ would
be a voltage,
vc3 (0 = ^ f_ Ja \ ic \ (T ) + a2ic2('^)] d r
= ai
\ 7 S ' - J c 2 W ‘‘^
Arguments analogous to the preceding imply that a relaxed inductor satisfies a linear relationship,
and thus superposition is valid, whether the inductor is excited by currents or by voltages.
The interconnection of linear capacitors and linear inductors with linear resistors and sources sat
isfying KVL and KCL produces linear circuits because KVL and KCL are linear algebraic con
straints on the linear element equations. Hence, the property of linearity is maintained, and as a
consequence superposition holds for the interconnected circuit.
To cap off this discussion we must account for the presence of initial conditions on the capacitors
and induaors o f the circuit. For first-order RC and RL circuits, this need is clearly indicated by
the first term o f equation 8.4. For a general linear circuit, one can view each initial condition as
being set up by an input that shuts off the moment the initial condition is established. Hence the r-
Chapter 8 • First Order RL and RC Circuits
efFect o f the initial condition can be viewed as the effect o f some input that turns o ff at the time
the initial condition is specified. This means that when using superpositioti on a circuit, one first
looks at the effect o f each independent source on a circuit having no initial conditions. Then one
sets all independent sources to zero and computes the response due to each initial condition with
all other initial conditions set to zero. The sum o f all the responses to each o f the independent
sources plus the individual initial condition responses yields the complete circuit response, by the
principle o f superposition. A rigorous justification o f this principle is given in a later chapter using
the Laplace transform method.
E XA M PLE 8.9
The linear circuit o f Figure 8.25 has two source excitations applied at r = 0, as indicated by the
presence o f the step functions. The initial condition on the inductor current is = -1 A.
Compute the response /^(r) for r > 0 using superposition.
So lu t io n
Because the circuit is linear, having a linear differential equation, superposition is but one o f sev
eral methods for obtaining the solution. An alternative approach is to find the Thevenin equiva
lent circuit seen by the inductor. As we will see, the superposition approach sometimes has an
advantage over the Thevenin approach.
Superposition must be carefully applied, however. First one computes the response due only to the
initial condition with the sources set to zero. Second, one computes the response due to Vj with
all initial conditions and all other sources set to zero. Third, one computes the response due to /j
with all initial conditions and all other sources set to zero. Finally, one adds these three responses
together to obtain the complete circuit response.
Step 1. Compute the part o f the circuit response due only to the initial condition, with all independ
ent sources set to zero. With both sources set to zero, there results the equivalent circuit given by
Figure 8.26. The Thevenin equivalent resistance is = 4 Q, resulting from the parallel combi
nation o f and R->. Figure 8.26 depicts the equivalent undriven RL circuit having response
352 Oliapccr 8 • First Order RI. and RC Circuits
An approach based on the Thevcnin equivalent circuit seen by the inductor would allow one to
quickly compute the complete response, but not in a way that identifies the contributions due to
each o f the individual sources. Answ^ers to the preceding three questions would have required
repeated solutions to the circuit equations. However, if one keeps the source values in literal form,
then the Thevenin equivalent approach would be as efficient.
Having gained some understanding of the form o f the behavior of RL and RC circuits, it is instruc
tive to classify the responses into categories. The zero-input response o f a circuit is the response
to the initial conditions when all the inputs are set to zero. The zero-state response o f a circuit is
the response to a specified input signal or set of input signals given that the initial conditions are
all set to zero. By linearit)', the sum o f the zero-input and zero-state responses is the com plete
response o f the circuit. This categorization is the convention in advanced linear systems and lin
ear control texts.
I'Vequently circuits texts include two other notions o f response, the natural response and the
forced response. However, decomposition o f the complete response into the sum of a natural and
a forced response applies only when the input excitation is (1) dc, (2) real exponential, (3) sinu
soidal, or (4) exponentially modulated sinusoidal. Further, the exponent of the input excitation,
for example, a'm j{t) = must be different from that appearing in the zero-input response.
Under these conditions it is possible to define the natural and forced responses as follows: (1) the
natural response is that portion of the complete response that has the same exponents as the zero-
input response, and (2) the forced response is that portion of the complete response that has the
same “exponent” as the input excitation provided the input excitation has exponents different
from that of the zero-input response.
This decomposition is important for rwo reasons. First, it agrees with the classical method o f solv
ing linear ordinary differential equations with constant coefficients where the natural response cor
responds to the com plem entary function and the forced response correspontls to the particular
integral. Students fresh from a course in linear differential equations will feel quite at home with
these concepts. The second reason is that the forced response is easily calculated for dc inputs. For
general systems this type o f decomposition is not used.
In deriving equations 8.17 and 8 .1 9a, the quantin' x{t) was thought o f as a capacitor voltage or an
inductor current. It turns out that any voltage or current in an RC or RL first-order linear circuit
with constant input has the form
{I- to)
x{l) = X^,+ ~ r~ (8.23)
For T negative or positive equation 8.23 is identical to equation 8.17 with Pi = X^. Further, T is
the circuit time constant and X is that voltage or current of interest computed under the condi-
Chapter 8 • First Order Rl. ami RC Circuits 3 S3
tion that the inductor is replaced by a short circuit for the RL case or the capacitor is replaced by
an open circuit for the RC case.
How do we justify the form o f equation 8.23 for all variables? We invoke the linearity theorem o f
Chapter 5 and the source substitution theorem o f Chapter 6 o f 2"^ edition. Suppose in a first-
order RC circuit we have found V(\t). By the source substitution theorem, the capacitor can then
be replaced by a voltage source whose voltage is the computed V(it). This new circuit consists o f
constant independent sources, one independent source o f value and resistors and depend
ent sources. By linearity, any voltage or current in the circuit has the form
x(r) =
Note that equation 8.23 requires that the initial value be evaluated at ^ instead o f r = This
is because only the inductor currents and the capacitor voltages are guaranteed to be continuous
from one instant to the next for constant input excitations. The capacitor current and the induc
tor voltage as well as other circuit voltages and currents may not behave continuously.
E XA M PLE 8 .1 0
This example illustrates the application o f equation 8.23. For the circuit o f Figure 8.29, =
-18//(-r) + V. Find i- it) for f > 0.
6kQ 2kQ
v jt )
3kQ
0.5 mP
SO L U T IO N
Step 1. Compute /y„(0^). To obtain /y„(0*), we first compute V(\Q~) = Since for r < 0, - 1 8
V has excited the circuit for a long time, the capacitor looks like an open circuit. By voltage divi
sion,
j = —( - 1 8 ) = - 6 V
9 - ( - 1 .5 )
= 1.75 mA
6x10-^
Step 2. Find the circuit time constant and the equilibrium value o f ijj,t). From Figure 8.29, the
equivalent resistance seen by the capacitor is = 4 kl^. Hence, the time constant is t = 2 sec.
Further, since t><d,X^ = ij„{'^), which is computed when the capacitor is replaced by an open cir
cuit. In this case.
= = 1 mA
9 x 1 0 -'
Step 3. Apply equation 8.23. Using equation 8.23, we have, for r > 0,
-0.5/
= 1 + 0.75^' mA
Exercise. In Flxample 8.10, find /^O ), /(^O"^), and i^^t) for r > 0 using equation 8.23 directly.
AN SW ERS: 0, 2.25 mA. 2.25^’-** "^ mA
Note that in Example 8 .1 0 we used instead o f /,„(0~) to obtain the correct answer. Some
straightforward arithmetic shows that /y„(0“) = - 2 mA. Since /y„(0'^) = 1.75 niA, the input current
is discontinuous at f = 0, unlike the capacitor voltage, which is continuous at f = 0. This empha
sizes the need to use x{t^) in equation 8 .23.
Chapter 8 • First Order RL and RC Circuits 3S5
In several o f the examples o f sections 3 and 4, the circuits contain switches that operate at pre
scribed time instants. In some electronic circuits, the switch is a semiconductor device whose
on/off state is determined by the value o f a controlling voltage somewhere else in the circuit. If the
controlling voltage is below a certain level, the electronic switch is off; if the voltage moves above
a fixed level, the electronic switch is on. The time it takes for a controlling voltage to rise (or fall)
from one level to another is very important because timing is as critical in electronic circuits as is
scheduling for large organizations. For first-order linear networks with constant excitations, cal
culation o f the time for a voltage or current to rise (or fall) from one level to another is straight
forward because all waveforms are exponential functions, as per equation 8.19. The situation is
illustrated in Figure 8.31.
c
0)
k_
I—
D
u
<U
cn
(Z
4->
o
>
FIGURE 8.31 First-order response showing a rise from the voltage or current level A'j to the voltage
or current level A'2, for which the elapsed time is h - ty
In equation 8.19a, let .v(rj) = and .v(^^) = X , be the two levels o f interest. A straightforward
manipulation o f equation 8.19a leads to the elapsed tim e formula for first-order circuits.
E X A M PLE 8.11
This example uses the elapsed time formula o f equation 8,24 for the circuit o f Figure 8.32. The
switch in this circuit is used to produce r^vo different “final” capacitor voltages. When the switch
is open, the final capacitor voltage is 12 V. When the switch closes, the final capacitor voltage, by
V-division, changes to 4 V. Thus the switch causes the capacitor to charge and discharge repeat
edly. For our purposes we show that the choice o f resistances produces an approximate triangular
waveform.
For the purposes o f this example, suppose the switch in Figure 8.32 is controlled electronically so
that it closes when rises to 9 V and opens when V(^ falls to 5 V. Find and plot v^^t) for several
SNvitchings.
Chapter 8 • First Order RL and RC Circuits
FIGURE 8.32 Switched driven RC circuit used to generate an approximate triangular waveform.
So lu t io n
Suppose the switch first closes at f = subsequently opening at ^ = tj^ and closing again at t^tc,
and so on. For 0 < r < the time constant T = 3 sec, V(\Q) = 0, and = 12 V. From equa
tion 8 .1 9c,
t
VcU) = 12 1 - e 3 V
- 0 = 3//? = 3 X 1.386 = 4 .1 5 9 s
19-12^
Now, for t < tj^, there is a new' time constant x= 1 sec, = 9, and = 4. .^gain using
equation 8.19c,
//9-4\
- t „ =/n - /n{5) = 1.61 s
\5-4/
Finally, for the time interval f , x= 3 sec, = 5, and t'^co) = 12. Using equation 8 .1 9c,
JJz Itl
v c(/) = 1 2 - 7 e ^
From the preceding solurion = 4.16 sec, = 4.1 6 + 1.61 = 5.77 sec, and t^ = 5.77 + 2.54 = 8.31
sec. If we proceed to calculate the waveform for t > the waveform begins to repeat itself, as is
evident Irom Figure 8.33. Practically speaking, the first c)'cie o f a periodic, approximately trian
gular waveform occurs in the time interv^al [t^, r j , and the period is = 8.31 - 4.1 6 = 4.15
sec. Note that the triangular waveform has a frequency
I 1 1
/ = = 0.241 Hz
period 2.5 4 + 1.61 4. 15
The waveform in figure 8.33 is approximately triangular. This is due the fact that two time con
stants, 1 and 3 s, have the same order o f magnitude. If we select the resistances so that the charg
ing time constant is much larger than the discharging time constant, then the capacitance voltage
waveform will look more like a sawtooth waveform. Sawtooth waveforms arc used to drive the
horizontal sweep of the electronic beam in an oscilloscope or a T V picture rube.
RC op amp circuits have some singular characteristics that set them apart from standard passive
RC and RL t)^pes o f circuits. Specifically, because ol the nature o f the operational amplifier, the
time constant o f the circuit will often depend only on some o f the resistances. We present four
important examples to illustrate the behavior o f RC op amp circuits.
3^H Chapter 8 • First Order RL and RC Circuits
EXA M PLE 8 .1 2
Compute the response for the ideal op amp circuit o f Figure 8.34.
1.
FIGURE 8.34 Differentiating op amp circuit.
So lu t io n
Observe that by the virtual short-circuit propert)' o f the ideal op amp, as set forth in
Chapter 4. Also, Hence, from these equalities and the definition o f a
capacitor,
= « ,,( ,) = -« / c (0 = - R C ^ =- R C ^ (8.25)
Since the output is a negative constant (user chosen) times the derivative o f the input, the circuit
is called a differentiator.
Exercise. Suppose y/„(^) = cos(250^). Find R for the circuit o f Figure 8.34 so that = sin(250/)
V a n d C = 1 pF.
AN SW FR: 4 kH
EXA M PLE 8 .1 3
Compute the response for the ideal op amp circuit o f Figure 8.35 assuming V(^0~) =
= 0.
1.
So lu t io n
Observe rhat = v-^{t)IR by the virtual short circuit property o f the ideal op amp. Also, i^^t)
= Hence, from these equalities and the integral i>-i relationship o f a capacitor,
(8.26)
Since the output is a negative constant (user chosen) times the integral o f the input, the circuit is
called an integrator.
Exercise. Suppose = cos(250r) V. Find R for the circuit o f Figure 8.35 so that =
sin(250r) V and C = 1 pF.
A N SW ER: 4 kD
EXA M PLE 8 .1 4
This example considers the so-called leaky integrator circuit o f Figure 8.36, which contains an
ideal op amp. The input for all time is v^{t) = V. Rj represents the leakage resistance o f the
capacitor. Given C and Rj, the resistance /?| is chosen to achieve a dc gain o f 10. The objective is
to compute the response assuming = 0 and compare it to a pure integrator having a
gain o f 1. This problem is reconsidered in Chapter 13.
+ v jt )
1
FIG U RE 8.36 Leaky integrator op amp circuit in which v^{t) = -5u{t) V.
So lu t io n
Because there is only one capacitor, the circuit o f Figure 8.36 is a first-order linear circuit. Because
the inverting terminal o f the op amp is at virtual ground, and the capacitor sees
an equivalent resistance R^i^ = /?2- Hence x = /?,C = 10 sec > 0, Equation 8.19 implies that
360 Chapter 8 • First Order RL and RC Circuits
(8.27)
Because the voltage source v^{t) = 0 for r < 0, ) = -V(i^ ) = 0. For f > 0, v^{t) = - 5 V. Since
the source voltage is constant, the capacitor looks like an open circuit at r = co having final value
= 50 + (0 = 5 0 (l ii{t) V
A plot o f the op amp output voltage appears in Figure 8.37 along with that o f an ideal integrator.
O ne observes that the more realistic leaky integrator circuit approximates an ideal integrator only
for 0 < r < 0 .1 5 t before the error induced by the feedback resistor R-^ becomes noticeable. Such
integrators need to be reinitialized periodically by resetting the capacitor voltage to zero.
OJ
Ol
ra
>
4-'
D
CL
4-1
D
o
Q.
E
<
Q.
O
Time (s)
So h r we have assumed an ideal op amp. In practice, the output voltage will saturate at a level
determined by the power supply voltage and the specs o f the particular amplifier used. Further,
practical op amps have complex models. To evaluate the preceding analysis, Figure 8.38 shows a
SPIC E simulation using the standard 741 op amp.
Chapter 8 * First Order RL and RC Circuits 361
Observe rhar the response approximaces the ideal up to about 0.1 5t = 1.5 sec, which corroborates
our analysis using the ideal op amp. Note, however, that the simulation accounts for saturation
present in practical op amps but absent from the ideal.
4kO
Capacitance
(a) (b)
F IG U R E 8.39 (a) Incorrectly connectcd Inverting amplifier, (b) Circuit model, including a stray
capacitancc C = 1 pF and a finite gain A = lO"^.
362 Chapter 8 • First Order RL and RC Circuits
So l u t io n
C'hapter 4 op amp models contain only resistors and controlled sources. One way to explain the
situation described in this example is to postulate a small stray capacitance, C = 1 pF, across the
input terminals. In fact, this is a more accurate circuit model and is shown in Figure 8.38b. This
means that the response will be o f the form o f equation 8.23.
The first quantit)' to compute is the circuit time constant t = C. The equivalent resistance
looking to the right of C, is obtained from the circuit o f Figure 8.40.
FIG URE 8.40 Circuit for computing note the artificial 1 V excitation.
' 4000
Hence,
R = 1 = Q
/, 1 -1 0 ^
Observe that in the actual circuit, is in parallel with 1 k il. Hence, the Thevenin equivalent
resistance seen by C is
The negative time constant spells instabilit)'. The complete response may be written directly with
the use o f equation 8.23, where x{t) = To use equation 8.23, we need a^O"^) = and
= lO-le V
Chapter 8 • First Order Rl. and RC Circuits 363
To compute the equilibrium output voltage we open-circuit the capacitor and compute
Kg = with C open-circuited
/ -2 5 \ j
= 1 X ------ — 10'’ = -2 5 .0 6 3 V
V 1000-2.5/
0.339xl0‘^r (8.28)
= -2 5 .0 6 3 + (lO'^E + 2 5 .0 6 3 )e
For any small positive initial capacitor voltage E, equation 8.28 implies that the output would
increase exponentially, had the op amp been ideal. Because this particular real op amp saturates at
15 V, the output more or less instantaneously saturates at 15 V, the phenomenon observed by the
student. Had the initial capacitance voltage been sufficiently negative,
-2 5 .0 6 3
£ < ------- j ---- V,
10-^
9. SUMMARY
This chapter has explored the behavior o f first-order RL and RC circuits (1) without sources for
given ICs, (2) for constant excitations (dc), (3) for piccewise constant excitations, and (4) with
switching under constant excitations. In general, first-order RL and RC circuits have only one
capacitor or one inductor present, although there are special conditions when more than one
inductor or capacitor can be present. Our discussion has presumed only one capacitor or one
inductor is present in the circuit.
Using a first-order constant-coefficient linear differential equation model o f the circuit, the chap
ter sets forth rwo t}'pes o f exponential responses, the source-free response and the response when
constant independent sources are present. The source-free responses for the RL and RC circuits
have the exponential forms
When independent sources are present in the circuit, the response o f a first-order RC or RL cir
cuit has the general form
I-10
x{t) = ,v(3c) + [,y(/o ) - lie-
36-i Chapter 8 • First Order RL and RC Circuits
provided the time constant t > 0. When the time constant t < 0, then it is necessary to modify
the interpretation as discussed in section 7, with equation 8.23 identifying the form o f any volt
age or current in the circuit:
The time constants of- a circuit can be changed by switching within the circuit. By changing time
constants in a circuit, one can generate different t)'pes o f waveforms such as the triangular wave
form o f Figure 8.32. As mentioned at the beginning o f the chapter, wave shaping is an important
application o f circuit design. When inductors, resistors, and capacitors are present in the same cir
cuit, many other wave shapes can be generated. RLC circuits arc the topic o f the next chapter and
allow even greater freedom in waveform construction.
As a final application o f the concepts o f this chapter, we looked at the leaky integrator op amp cir
cuit. Integrators are present in a host o f signal processing and control applications. Unfortunately,
ideal integrators do not exist in practice. The leaky integrator circuit o f Figure 8.35 provides a rea
sonable model o f an ideal integrator.
Natural iirequency of a circuit: natural mode of “oscillation” of the circuit. For a first-order cir-
cuit having a response proportional to it is the coefficient X in the exponent.
Natural response: that portion of the complete response that has the same exponents as the zero-
input response.
Passive RLC circuit: circuit consisting of resistors, inductors, and capacitors that can only store
and/or dissipate energy.
Sawtooth waveform: triangular waveform resembling the teeth on a saw blade and typically used
to drive the horizontal sweep of the electronic beam in an oscilloscope or a T V picture
tube.
Source-free response: response of a circuit in which sources are either absent or set to zero.
Step response: response, for ^> 0, of a relaxed single-input circuit to a unit step, i.e., a constant
excitation of unit amplitude.
Stray capacitance: small capacitance always present between a conductor and ground. It usually
can be ignored, but as Example 8.14 shows, it can critically affect the response of a cir
cuit.
Superposition: in linear RC and RL circuits, the complete response is the superposition of the
relaxed circuit responses due to each source with all other sources set to zero, plus the
responses to each initial condition when all other initial conditions are set to zero and all
independent sources are set to zero.
Time constant: in a source-free first-order circuit, the time it takes for the circuit response to drop
to e~^ = 0.368 of its initial value. Roughly speaking, the response value must drop to a
litde over one-third of its initial value or rise to within one-third of its final value. For RL
circuits x = LIR^f^ and for RC circuits x =R^/jC.
Unit step (unction: function denoted «(/) whose value is 1 for f > 0 and 0 for ^< 0.
Unstable response: response whose magnitude increases without bound as t increases. The time
constant for first-order circuits is negative for an unstable response.
Zero-input response: response in which all sources are set to zero.
Zero-state response: response to a specified input signal or set of input signals given that the ini-
tial conditions are all set to zero.
w
^ It happens that all variables in a first-order RL or R C circuit satisfy a differential equation o f the same form. The
interpretation o f the solution is somewhat different fi-om what follows. A detailed explanation o f the general solu
tion is presented in section 7.
366 Chapter 8 • First Order RL and RC Circuits
Problems i,(t)
Figure P8.1
R.
7. In the circuit o f Figure P8.7, suppose /?j = 5
kQ, /?2 = 20 k n , C = 50 pF H, = 20 V,
and the switch opens at r = 0.4 sec.
R, 0.25 mF
(a) Find v^^t) for f > 0 . Plot your answer in
MATLAB for 0 < f < 4 sec.
(b) Find and /^(/), t > 0. Plot your Figure P8.9
answer in MATLAB for 0 < r < 4 sec. C H E C K : (c) a > - 3
t=0.4s
10. In the circuit o f Figure P8.10, /?, = 100 Q,
= 20 P = 2 0 0 ,1 = 0.5 H, and /'^(O) = 250
i„{t)
+ mA. The switch opens at / = 0.03 sec. Find the
,Vc(t) Thevenin equivalent resistance seen by the
inductor before the switch opens, and then
compute /^(r) and for r > 0.
Figure P8.7
ijt)
mA.
1
(b) Given your answer to parr (a), compute
and plot /^(r) for 0 < r < 5x, where I is Figure P8.10
the circuit time constant. C H EC K : = -2 5 n
ijt )
1 kf) 11. Consider the circuit o f Figure P 8 .ll in
which /?j = 25 ^2, “ 50 Q, and L = 2.5 H.
80 mH Suppose /^(O) = 2 A.
(a) With CX = 0.1, compute the Thevenin
Figure P8.8 equivalent resistance seen by the induc
AN SW ER: (a) 600 Q, 25 mA tor; then compute /^(f) for f > 0. Plot in
M A T I^ B for 0 < r < 5 t, where T is the
9. Consider the circuit of Figure P8.9, in which time constant o f the circuit.
= 100 Q and /?2 = 50 Q. Let = 500 mV. (b) W ith a = 0.1, compute
(a) Let a = 7. Compute the equivalent (c) Repeat part (a) for a = 0.02. Determine
resistance seen by the capacitor. Find the time, say r,, when the inductor has
i/^r) for f > 0. Plot for 0 < ? < 5 t lost 99% o f its initial stored energy.
368 Chapter 8 • First Order RL and R C Circuits
- + t=4RC
i,(t) A —
R. R 1=0
.4R \4R
R,
Figure P 8 .1 1
Figure P8.14
.4R 4R
Figure P8.15
Figure P8.12
RESPONSE OF DRIVEN
CIRCUITS
18. Consider the RC circuit o f Figure P8.18 in
which R = 10 k n and C = 0.4 niF.
Figure P8.13
(a) If = 0 and V, find
U(^t). Plot your answer for 0 < r < 4x,
14. In Figure PS. 14 the voltage source =
where x is the circuit time constant.
has been applied for a long time before the
(b) II /^(^-(O) = 10 V and = 0, find
switch opens at r = 0. Find and t»^r) for
Plot your answer for 0 < / < 4x, where x
r > 0 in terms o f V^, R, and C, where is in V,
is the circuit time constant.
R in Q, and C in F. Sketch V(^t) for 0 < r < 3x,
Now making use o f linearit)' and its
where x is the circuit time constant for / > 0.
associated properties, compute the indi
cated responses without any further cir
cuit analysis.
(c) If V(^Q) = 10 V and = 2i)ti(t) V,
find V(^t).
Chapter 8 • First Order RL and RC Circuits 369
(d) If «c<0) = - 2 0 V and i/,.„(») = -I0 « (» ) V, 20. In Figure P 8.20, = 50 Q , 7?2 = 200 Q, C
find V(it). = 2.5 mF, and the voltage excitation is given by
(e) If i/^O) = 10 V and v-J^t) = 20«(r) V. where = -1 0 V
find i({t) without differentiating your and V^2 = 20 V.
answer to part (c). Plot your answer for (a) Find ^(;^0'*^) and V(it) for / > 0.
0 < r < 4 t , where x is the circuit time (b) Sketch v^^t) for 0 < f < 5x, where x is the
constant. circuit time constant for r > 0.
(c) Identify the zero-input response (f > 0)
yoj and the zero-state response (r > 0) for the
answer computed in part (a).
'.w O
(d) Now compute for f > 0 assuming
the switch opens at r = 0.2 5 sec. Plot
Figure P8.18 your result ForO < f < 0.5 sec.
I'
'Vw>
370 Chapter 8 • First Order RL and RC Circuits
n
+
'.( s O ;vc(t)
r\
(b)
Figure P8.23
(a)
i,(t) r>
V„u(t)
d) Figure P8.24
Figure P8.25
20 msec
=0
2kO 1.6 kn
+
'. w ( D
8kO
Vc(t).
S jjF
o 20 V
+
vjt) .Vc(t)
Figure P8.29
Figure P8.31
30. Consider the circuit o f Figure P 8.30 in
which /?j = 300j^, /?2 = 800 Cl, R^ - 600 Q, L 3 2 . For the circuit of Figure P 8.32, -V^ = - 1 0
= 4 H, - -24u{-t) + 24«(/) V, and = V, Kj = 20 V, R^ = 6 0 0 Q, /?2 = 2 0 0 Cl, and C
24u{t) V. = 12.5 hF-
(a) Compute the response ii{t) for ? > 0. (a) Find y^O"^).
Plot for 0 < ^ < 4x where x is the circuit (b) Using the initial condition computed in
time constant. part (a), find Plot the result for 0
(b) Find the inductor voltage v^{t) for r > 0 < t< 160 msec.
372 Chapter 8 • First Order RL and RC Circuits
>
20V
o
v„,(t)
5 t(Msec)
2.5
(b)
--10V
-----------
Figure P8.32 (a) Pulse driving (a)
RC circuit of part (b). IMegO
- v,(t) +
400 5kO
(b) C = 0.1 F
6 IV V, lOlv.
Figure P 8 .3 3
Figure P 8 .3 6
Chapter 8 • First Order RI. and RC Circuits 373
Figure P8.37 A . KR
Figure P8.41
Figure P 8 .4 0
374 Chapter 8 * First Order RL and RC Circuits
(a) (b)
Figure P8.43 ‘r
Figure P8.45
44. In the circuit of Figure P8.44, suppose /?j =
50 Q, /?2 = 200 a , q = 0.06 F, Cj = 0.3 F, 46. Repeat Problem 45 for the circuit of Figure
yQ(0“) = 15 V, and i'qCO") = 5 V. Let = P8.45 when q = 1 F, Cj = 0.25 F, i/^iCO-) = 3
40tt(r) V. Use the equivalence set forth in V, and t/f^(0~) = 8 V.
Problem 42 to compute for r > 0.
47. In the circuit of Figure P8.47, =
1 lOtt(-f) + 220u(t) mV, Zj = 110 mH, ^2 = 11
C. - ± V,
mH, and /? = 10 ft. Compute and plot the
waveforms for inU) and /^(/). Hint: Adapt the
C, - t
results of Problem 43 to ^e case of two paral
lel inductors with initial currents.
Figure P 8 .4 4
r>
Chapter 8 • First Order RL and RC Circuits 375
49. Consider the circuit of Figure P 8.49 in 51. The solution to the basic RL or RC differ
which Cj = 1 F, Cj = 4 F, v-^ = 10 V, and R = 2 ential equation in this chapter, equation 8.3,
a. builds on the integral solution o f equation 8.4,
(a) Compute v^Qr) and Vj^{Q*). Hint: How which is valid for arbitrary y(/). This powerful
does the charge distribute over the two formula will be studied in a course on differen
capacitors at ^ = 0^? tial equation theory. Wheny(r) = F, a constant,
(b) Compute Vjfi) for r > 0. it is possible to develop an alternative deriva
tion o f equation 8 .1 7 using no more than some
basic knowledge of calculus. Since the solution
to the source-free case is the exponential, it is
reasonable to expect (or to try) a solution for
the constant input case o f the form
'> ^ '
( 1)
Figure P8.50 x Uq ) - F t
(2)
(a) Suppose vj^i) = K^u{t) V. Show that for t (c) Finally, substitute K2 = Fx and equation
> 0, the voltage v{i) satisfies the first- 2 into equation 1.
W
order differential equation
dv
= -K iv + K 2
It
3 "6 Chapter 8 • First Order RL and RC Circuits
A slope
= 1m U
/ • (not to scale)
/ : negative resistance
region
slope i
Figure P8.52
V = 0.5 mU :
60
N fc i
----------- ►
90
v(V)
(b) (0
Remark: All answers to parts (a) to (f) should be 5 5 . The circuit o f Figure P8.55 is a transistor
in terms o f /?,. and R^. photo tim er used for timing the light in pho
(a) Draw a simplified equivalent circuit tographic enlarger and printing boxes. Briefly,
with three circuit elements for 0 < ? < T j, the circuit operates as follows. When the relay
indicating all device values. contact closes, the lamp is lit. When the contact
(b) Construct an expression for 0 < ^< opens, the lamp is turned off. The relay has a
4 0 0 0 Q dc resistance and a negligible induc
(c) Draw a simplified equivalent circuit tance. The pickup current is 2 mA, and the
with two circuit elements for T-^<t< dropout current is 0.5 mA; i.e., the contact
indicating all device values. closes when the relay current increases from
(d) Construct an expression for /^(/), T^^<t zero to 2 mA, and it opens when the current
< Tj. You will need a value or expression drops below 0.5 mA.
for After obtaining your expres
sion, for simplicity, let denote To use the timer, switch S2 is closed first.
Switch Sj is normally in the B position. When
(e) Draw a simplified equivalent circuit it is thrown momentarily to position A, the bat
with two circuit elements for T2 < t, tery charges the 1000 pF electrolytic capacitor
indicating all device values. C to 1.5 V. When Sj is then thrown back to
(f) Construct an expression for Tj < t. position B at /^= 0, the capacitor discharges and
You will need an expression for produces a current iy, which, after amplifica
After obtaining your expression, for sim tion by the transistor, actuates the relay and
plicity, let ^Tl- denote /^(jT2 )• turns on the lamp. At some later instant the
Remark: For the remaining parts, all amplified current drops below a point for the
answers are to be given in terms of relay to open and the lamp is turned off.
K kf tori’ ^1> ^ 2> “ d
This is to prevent the substitution of Compute if the 10 kH potentiometer is set at
possibly incorrect answers from prior the middle of its full range (i.e., only 5 kQ is
parts for used in the circuit).
(g) For each o f the four devices
^ston> ^store down an expres
sion for the power absorbed at time t.
Call the results
and PRstore’
(h) For time t give an expression for the
energy Wj(J) stored in the inductor if
W i( r = 0 ) = 0.
^soljr -N
-------------- ^
Figure P 8 .5 4
378 Chapter 8 • First Order RL and RC Circuits
1.5V
lO kn E
Potentionmeter crude transistor
circuit model
Figure P8.55
56. The circuit o f Figure P 8.56 suggests a way of generating a sustained sinusoidal oscillation. All
op amps are assumed to be ideal. Capacitors, C = 0.1 fiF are uncharged at f = 0. The first two op
amps are differentiators and the last is an inverting amplifier.
(a) With switch S at position A and v^{t) = sin(lOOOr) V, find vj^t), and for r > 0.
(b) If at a later instant switch S is quickly moved to position B, what would you expect
to be?
Figure P8.56
C H A P T E R
food food
(a) (b)
food
0
Reversal of the polarity o f the applied voltage at a low frequency can be easily achieved with the
circuit elements studied in earlier chapters: the resistor, the capacitor, and the inductor. However,
the friction-induced heat production is inefficient at low frequency, l b produce a useful amount
o f heat for cooking purposes, ver)' high frequencies must be used. The t}'pical frequency used in
a microwave oven is 2.45 gigahertz, i.e., the water molecules reverse their orientations 2 x 2.45 x
10^ times per second. At such a high frequency, capacitors and inductors are quite different in
their behavior from their conventional forms. For example, the LC circuit becomes a “resonant
cavit}” and the connecting wire becomes a “waveguide.” These microwave components will be
studied in a future field theory course. T he theory studied in this chapter will enable us to under
stand the low-frequenc}' version o f the phenomenon, i.e., how a connected inductor and capaci
tor can produce oscillator)' voltage and current waveforms.
CHAPTER OBJECTIVES
1. Investigate the voltage-current interactions that occur when an ideal inductor is con
nected to an ideal capacitor with initial stored charge.
2. Use a second-order differential equation for modeling the series RLC and parallel RLC
circuits.
3. Learn to solve a second-order differential equation circuit model by first finding the nat
ural frequencies o f the circuit, then looking up the general solution form, and finally
determining the associated arbitrary constants.
Chapter 9 • Sccond Order Linear Circuits 3<S I
SECTION HEADINGS
1. Introduction
2. Discharging a Capacitor through an Inductor
3. Source-Free Second-Order Linear Networks
4. Second-Order Linear Networks with Constant Inputs
5. Oscillator Application
6. Summary
7. Terms and Concepts
8. Problems
1. IN TRO DUCTION
The previous chapter developed techniques for computing the responses of first-order linear net
works, either without sources or with dc (constant) sources, having first-order linear differential
equation models. Recall that the source-free response contains only real exponential terms.
This chapter focuses on second-order linear networks having second-order linear differential equa
tion models. Usually, but not always, a second-order net\vork contains t^vo energ)^ storage ele
ments, either {L, Q , (C Q , or {L, L). Second-order circuits have a wide variety o f response wave
forms: exponentials sinusoids (/l,cos(coy) + y4-,sin(coy)), exponentially damped
sinusoids, and exponentially growing sinusoids, among others. Tables 9.1 and 9.2 catalogue the
various response types. W ith no sources or with constant-value sources, some straightforward
extensions o f the solution methods o f Chapter 8 are sufficient to compute the various responses.
The behavior of second-order circuits is a microcosm o f the behavior o f higher-order circuits and
systems. Many higher-order systems can be broken down into cascades o f second-order systems or
sums o f second-order systems. This suggests that our exploration o f second-order circuits can
build a core knowledge base for understanding the behavior o f higher-order, more complex phys
ical systems.
Many introductory texts discuss only parallel and series RLC circuits, stating separate formulas for
the responses o f each. Our approach seeks a unified treatment. To this end, we formulate a basic
second-order differential equation circuit model. The associated solution techniques become
applicable to any second-order linear nersvork and, for that matter, to second-order mechanical
systems.
An oscillator circuit (section 5) motivates our study o f second-order linear networks. The chapter
contains several other practical examples illustrative o f the wide variety o f second-order circuit
382 Chapter 9 • Second Order Linear Circuits
applications. Some advanced applications pertinent to higher-level courses include low-pass, high-
pass, and bandpass filtering (covered later in the text); dc motor analysis; position control; and
many others. Most important, the concepts presented in this chapter are common to a host o f
engineering problems and disciplines. Hence, the techniques and concepts described here will
prove useful time and time again.
Chapter 8 showed that the voltage o f an initially charged capacitor in parallel with a resistor
decreases exponentially to zero: the capacitor discharges its stored energy through the resistor.
When an inductor replaces the resistor, as in Figure 9.1a, very different voltage waveforms emerge
for V(-{t) and In order to construct these new waveforms, we first develop a differential equa
tion model o f the LC circuit.
EX A M PLE 9 .1 . The goal o f this example is to develop a differential equation model o f the cir
cuit in Figure 9.1b. In Figure 9.1a, with the switch S in position A, the voltage source charges the
capacitor to volts. At / = 0, the switch moves to position B, resulting in the new circuit o f
Figure 9.1 b, valid for f > 0.
A B
+ +
(a) (b)
FICiURF 9.1 (a) A voltage source charges a capacitor, (b) An LC second-order linear network in
which the cnerg)’ stored in the capacitor in part (a) is passed back and forth to the inductor.
S o l u t io n .
Step 1. Write down the terminal i-v relationship for the capacitor and itiductor; then apply KCL and
KVL, respectively. Using the i-v relationships for L and C (see Chapter 7) in conjunction with KCL
and KVL, it follow's that
ih C _ >c _ ±L_
dt c c (9.1a)
V J V
C definition KCL
and
di
(9 .1 b )
V
L definition KVL
Chapter 9 • Second Order Linear Circuits 383
Step 2. Obtain a dijferential equation in the capacitor voltage, V(j^t). For this, first differentiate
equation 9.1a to obtain
1 dii^
dt C dt
Substituting equation 9.1b into this equation yields
^ = - — VC (9.2)
dt LC ^
Equation 9.2 is a second-order linear differential equation circuit model o f Figure 9.1b in terms o f
the unknown capacitor voltage, v^t). Equation 9.2 stipulates that the second-order derivative j
o f the unknown function, v^it), must equal the function itself multiplied by a negative constant,-----
LC
Step 3. Obtain a differential equation in the current, ij{t). An alternative circuit model in i^ is obtained
by first differentiating equation 9.1b and substituting equation 9.1a into the result to produce
Equation 9.3 has precisely the same form as 9.2: the second-order derivative o f the unknown
function, ij{t) , equals the function itself multiplied by a negative constant, -----. This similarity
suggests a similarity o f solutions, which we shall pursue further.
Exercise. Fill in the details o f the derivation o f equation 9.3 from 9.1.
Our next goal is to construct the waveforms V(^t) and ij{t), which are the solutions o f the differ
ential equations 9.2 and 9.3. Although differential equations are not usually part o f the common
background o f students in a beginning course on circuits, the solutions o f equations 9.2 and 9.3
do not demand this background. Some elementary knowledge o f differential calculus is sufficient.
Specifically, recall the differential properties o f the sine and cosine functions:
d d
— sin(co/+ 0 ) = (o co s(o )/-t-0 ) and — c o s ( o ) / + 0 ) = - a ) s i n ( c o / + 0 )
dt dt
Differentiating a second time yields
d2 ^y2
— 7 sin((or-f-0) = - 0) “ sin(co/+ 0 ) and — 7 cos(o)/+ 0 ) = - c a “ cos(co/+ 0 )
dt- dt-
In both cases, the second derivative equals the function itself multiplied by a negative constant.
This is precisely the propert}' required by equations 9.2 and 9.3. Thus one reasonably assumes that
the solutions o f equations 9.2 and 9.3 have the general forms
V(^t) = K cos(ojr + 0) (9 .4 a )
and
38^ Chapter 9 • Sccoiul Order Linear Circuits
7'hese forms are general because the cosine function can be replaced by the sine function with a
proper change in the phase angle. Specifically, we note that A'sin((i)/ + (j)) = A'cos(co/ + (j) - 0.5ti)
= K cos((or + 0) with 0 = ([) - ().5ti. Computing values for (o, A', and 0 specifies the solutions to
equations 9.2 and 9.3.
E X A M PLE 9.2. Find A"and B for the capacitor voltage in equation 9.4a.
So lu t io n .
Step 1. Differentiate equation 9.4a to obtain ^
dvrit)
— — = -A c o sin (O )/+ 0 ) /'o S'!
(It
Step 2. Dijferentiate n second time. Differentiating equation 9.5 (the second derivative o f 9.4)
yields
(0 ^ = -----
' o1r CO = — j= = fo 7) ^
LC fiC ^ ^
Equation 9.7 specifies co, the angular frequency o f oscillation, in rad/sec, o f the capacitor voltage.
Step 4. Compute K an d f) in equation 9.4a. These two constants depend on the initial conditions
as follows: when the switch is at position A, the capacitor is charged up to I^q volts and the induc
tor current is zero; immediately after the switch moves to position B, i.e., at r = 0+, the continu- ^
it>' properry o f the capacitor voltage ensures that y^^O*) = Vj, and^the continuity property o f the
inductor current ensures that //(O^) = 0. The initial value, ^ , ’ is now calculated from ^
equation 9.1a as
^/r(--(0 ) i( ^ ( 0 )
------------ = ----------- = ------------- = ()
dt C C
= K c o m = v;, (9.8a) ^
and
</v'r(0"’ )
— ^ = -A o )s m (e ) = 0 (9 g y
From equation 9.8b, 0 = 0. Consequently from 9.8a, K = V^^. Hence the capacitor voltage, i.e.,
the solution o f the second-order differential equation 9.2, is
As per equation 9.1a, one can obtain ij{t) directly by differentiating equation 9.9 and multiply
ing by —C. However, one could aLso solve equation 9.3 by repeating the above steps to arrive at
the same answer.
Exercise. Assuming that /^(r) = K coslcor + 6), solve for (o, K, and 0 in terms o f the initial condi
tions, and show that
Several very interesting and significant facts about this parallel ZCcircuit and the solution method
are apparent:
(1) For the source-free LC circuit o f Figure 9.1, the voltage and current responses are sinusoidal
waveforms with an angular frequency equal to —_L _ . Since the amplitude o f sinusoidal
oscillations remains constant (i.e., does not VZc
damp out), the circuit is said to be undamped.
(2) rhe frequenq^ (o, depends on the values o f L and C only, while the amplitude K and the
phase angle 0 depend on L, C, and the initial values o f the capacitor voltage and inductor
current.
(3) Although the instantaneous energ}" stored in the capacitor, \V^{t), and the instantaneous
energy stored in the inductor, both vary with time, their sum is constant. (This is
investigated in a homework exercise.) Physically there is a continuous exchange o f the ener-
g)' stored in the magnetic field o f the inductor and that stored in the electric field of the
capacitor, with no net energy loss. This is analogous to a frictionless hanging mass-spring
system: because o f the absence o f friction, the up-and-down motion of the mass never
stops; in such a mechanical system there is a continuous interchange between potential and
kinetic energy.
Figure 9.1 shows what is, in theory, the simplest circuit that generates sinusoidal waveforms. Such
an electronic circuit is an (idealized) oscillator circuit. Oscillator circuits play an important role
in many communication and instrumentation systems.
Unlike their ideal counterparts, real capacitors and inductors have resistances. A better under
standing o f a realistic oscillator entails the analysis o f an RLC circuit. This section investigates
source-free RLC circuits having two energ)' storage elements. Our investigation begins with the
development o f the differential equation models o f the series and parallel /^//.'circuits. Both mod
els are special cases o f an undriven general second-order linear diflerential equation. Hence we will
discuss the solution o f a general second-order linear differential equation and adapt the solution
to the series and parallel RLC circuits. We will also illustrate the theory with a second-order cir
cuit that is not a parallel or series RLC.
386 Chapter 9 • Second Order Linear Circuits
The first goal o f this section is to develop differential equation models for series and parallel RLC
circuits as detailed in the following example.
EXA M PLE 9 .3 . For the series and parallel RLC circuits shown in Figure 9.2, develop two sec
ond-order differential equation models (one in and one in for each circuit.
+ +
V
''c = V''l = V
''r
(b)
FIGURE 9.2 (a) Series RLC circuit, (b) Parallel RLC circuit. Passive sign convention is assumed as usual.
S o l u t io n
Step 1. Apply KVL to the series RLC. Step 1. Apply KCL to the parallel RLC.
d ir
Rin -f" L ------- 1- = 0 — -H/, + C — ^ = 0
^ dt ^ R ^ dt
Chapter 9 • Second Order Linear Circuits 38‘
R dvr llL +— = 0
d t‘ L dt LC dr RC dt LC
Each circuit has two second-order differential equation models, one each for and as the
unknown quantity.
Exercise. Show that and Vj^ satisfy second-order differential equations similar to those
derived in Example 9.3.
The final differential equations o f Example 9.3 force the current or the voltage to satisfy cer
tain differential constraints. All four (differential) equations have the general form
d^x dx
— ~ + h ----- h (.'.V = 0 (9 . 10)
dt- dt
for appropriate constants b and c, where .v is either ij or Equation 9.10 stipulates that the sec
ond derivative o f the function x{t) plus h times the first derivative o f x{t) plus c times x(t) itself
adds to zero at all times, t. Unlike the example o f section 2, where a sinusoidal solution was easi
ly predicted, the present differential equation requires a more careful mathematical analysis. Recall
from elementary calculus that the derivative o f an exponential is an exponential. Thus the first
and second derivatives o f an exponential are proportional to the original exponential. This sug
gests postulating a solution o f the form at(^) = Ke^‘ where we make no a priori assumptions about
s. If it is truly a solution, it must satisfy equation 9.10. Under what conditions will x{t) = Ke^‘ sat
isfy equation 9.10?
E X A M PLE 9.4. Determine conditions under which the postulated solution x{t) = Ke^‘ satisfies
equation 9.10.
S o l u t io n .
rfV '
K + bK — + cK e” = Ke" (,v^ + hs + c) = 0 (9.11)
dr <h ^ I
Step 2. Interpret equation 9.11. For nontrivial solutions, K\s nonzero. The function e^‘ is always
different from zero. Hence the quadratic in s on the right side o f equation 9.11 must be zero. This
necessarily constrains $ to be a root of
r + bs + c =0 (9 .1 2 )
3<S8 Chapter 9 * Sccond Order Linear Circuits
Step 3. Solve equation 9.12. From rhe quadratic formula, rhe roots o f equation 9.12 are
-b ± - 4c
.V|, .v. = (9.13)
C O N C LU SIO N : x(/) = Ke^' satisfies equation 9.10 provided s tatces on values given by equation
9.13. Equation 9.10 does not constrain K\ however, the initial conditions will.
Equations such as 9.12 whose solution is given by equation 9.13 are a common characteristic o f
second-order networks. Hence, equation 9.12 is called the characteristic equation o f the second-
order linear circuit. The associated roots, equation 9.13, are called the natural frequencies o f the
circuit. These are the “natural” or intrinsic frequencies o f the circuit response and are akin to the
natural frequencies o f oscillations o f a pendulum (for small swings) or o f a bouncing ball.
From elementary algebra, a quadratic equation (the above characteristic equation) can have dis
tinct roots or equal roots. Distinct roots can be real or complex. Thus and Sj can be two dis
tinct real roots, two distinct conjugate complex roots, or two repeated (equal) roots, depending
on whether the discriminant, Ip- - 4r, is greater than, less than, or equal to zero. This trifold
grouping separates the solution o f equation 9.10 into three categories, listed below as cases 1, 2,
and 3:
Case 1. Real and distinct roots, i.e., b~ -A c> 0. If the roots are real and distinct, then for arbitrar}'
constants and K2 , both
and
xif) = X2it)=K2e-'^-' ^
satisfy the second-order linear differential equation 9.10, i.e., are solutions to the differential equa-
tion. Since equation 9.10 is a differential equation, by superposition the sum x(r) = x,(^) +
X2 (t) is also a solution, a fact easily verified by direct substitution. Therefore, whenever s-y, the
most general form o f the solution to equation 9.10 is
The constants and K-, depend on the initial conditions o f the differential equation, which
depend on the initial capacitor voltages and inductor currents. For example, if a.*(0'^) and a''(0'*')
are known, then from equation 9.14,
.v (0^ )= A y '*'-h
and
If and $2 are negative, the response given by equation 9 .1 4 decays to zero for large f and the cir
cuit is said to be overdamped.
Case 2. The roots, and $2 , o f the characteristic equation are distinct hut complex, i.e., iP- -A c <Q.
Since Jj ^ general form of the solution to equation 9.1 0 is again given by equation
9.14, i.e.,
= +K2e"‘^^
y l^ c-b ^
where a = bH and co^ = ------ ^ S i n c e s^ and S2 are conjugates, so are d^i^and e^2^in equation 9.14.
For x(t) to be real, the constants /Cj and A'j in equation 9.1 4 must also be complex conjugates,
i.e., Using Euler’s formula,
the two terms in equation 9.14 combine to yield a real time function:
Thus the solution to equation 9.1 4 with and S2 complex is given by the (damped) sinusoidal
response
where ^4 = +^ = 2 Re[^j] = + A^2 and 5 - y Af] = - 2 ImATJ = /A j - yATj are real con
stants and where Re[ ] denotes the real part and Im[ ] denotes the imaginary part. The solution
expressed in equation 9.16 is completed by specifying ^4 and B. As before, A and B depend on the
initial conditions, x(0‘^) and xXO"*^) as follows:
= - o A + (o^B
and
Making use of a standard trigonometric identity, the general solution of equation 9.16 has the
equivalent form
and the quadrant of 0 is determined by the signs o f - B and A. In MATLAB, one uses the command
“atan2(-BvA)” to obtain the angle in the proper quadrant. Note that the response waveforms have
oscillations with angular frequency O)^ These oscillations are bounded by the envelope ±Ke~^*. If
Re[jj] = - a < 0, the amplitude of the oscillations decays to zero and the response is said to be under-
damped. If Re[jj] = - a > 0, the amplitude of the oscillations grows to infinity.
C ased. The roots are real and eqml, i.e.,lP' - 4 c = 0 . When the two roots of the characteristic equa
tion are equal, equation 9.1 4 does not represent the general solution form because if the
two terms collapse into a single term. However, the general solution for is
x(t) = {K^+K2t)e^^^
x{0-) =
and
i(0+) =
Substituting the value o f into x'(O^) yields a simple calculation for Kj.
If^l = i2 is negative, the response decays to zero and is said to be critically damped. “Critically
damped” defines the boundary between overdamped and underdamped. This means that with a
slight change in circuit parameters, the response would almost always change to either over
damped or underdamped.
d~x dx
— ^ + h — + cx - 0
dr dt
having characteristic equation + r = (j - 5 j)(j - = 0, where
- h ± yjb" —4 c
^2= -----------
x {t)= + Kie^-
where
40^) = and .v (0") = s^K^ + s^K^,
Figure 9.3 displays the various response forms described above for the case where Re[xj] and Re[j-,]
are negative or zero. Because o f their similarity, it is not possible to distinguish between the over
damped and the critically damped responses by merely looking at the waveforms. Both types o f
response may have at most one zero-crossing.
392 Chapter 9 • Second Order I.incar Circuits
> t t
FIGURE 9.3 (-'tncric waveforms corresponding to the four cases o f damping: (a) undamped (sinu
soidal) response, (b) undcrdamped (exponentially decaying oscillatory) response, (c) overdamped
(exponentially decaying) response, and (d) critically damped (exponentially decaying) response.
The terms “undamped,” “underdamped,” “overdamped,” and “critically damped” stem from an
intuitive notion o f “damping.” The sourcc-frce response o f an undamped second-order linear sys
tem, whether electrical or mechanical, has an oscillatory response (waveform) o f constant ampli
tude. Damping, due to system elements that consume cnerg\', means a monotonic decrease in the
amplitude o f oscillation. In electrical circuits, resistances produce the damping effect. In mechan
ical systems, friction causes damping. When the amount o f damping is just enough to prevent
oscillation, the system is critically damped. Less damping corresponds to the underdamped case,
where oscillation is present but eventually dies out. A greater amount o f damping corresponds to
the overdamped case, where the waveform is non-oscillatory, and a ver\’ small perturbation o f any
circuit parameter will not cause oscillations to occur.
In summary, once the roots o f the characteristic equation are found and the expression for the
general solution selected from the above cases or Table 9.1, it remains to find the constants A', and
Chapter 9 • Sccond Order Linear Circuits 3^)3
K j (or A and 5 ) from the initial conditions on the circuit. In the above development, and Kj
(or A and B) are given in terms of a:(0^) and x'(O^). Since jf(/) represents either a capacitor voltage
or an inductor current, its value at f = 0"^ is usually given, or can be determined from the past his
tory o f the circuit. (See Example 9.5.) The value o f x'(O^), on the other hand, is often unknown
and must be calculated. If >:(/) = then the capacitor v-i relationship implies that
^ xX0*) = v'c{0*) = ^ £ ^ .
To find or y^(O^), we construct an auxiliary resistive circuit valid at r = 0^. Since the initial
values, and /^(O^), are known, we replace (each) capacitor in the original circuit by an inde
pendent voltage source o f value and (each) inductor in the original circuit by an inde
pendent current source o f value /£(0'*^). Here the current /(;^0^) retains its original direction and
the voltage ^^(0^) retains its original polarity. After the replacements, the (new) circuit is resistive.
Values for and y^(O^) follow by applying any of the standard methods of resistive circuit
analysis learned earlier. This allows us to specify x(0^) and x'(O^) in terms of the initial conditions
on the circuit. Two equations in the two unknowns A'j and K2 (or A and B) result. Example 9.5
and, in particular. Figure 9.4c illustrate this procedure.
Before any additional circuit examples, let us summarize the solution procedure.
The following example illustrates these calculations for the three cases described in Table 9.1.
W
39-'i Chapter 9 • Second Order Linear Circuits
E XA M PLE 9 .5 . In the circuit o f Figure 9.4a, the 1 pF capacitor is assumed to be ideal, and the
inductor is modeled by a 10 mH ideal inductor in series with a 20 Q resistor to account for the
resistance o f the coiled wire. Suppose the switch S in Figure 9.4a has been in position A for a long
time. The capacitor becomes charged to 10 V. Then the switch moves to position B at f = 0. Find
and plot V(\t) for r > 0 for the following three cases; (1) = 405 (2) /?2 = 0, (3) 180 ti.
Each o f these cases produces a different response type.
lO O :L=10mH R=200
+
10V Practical Inductor
C=1 mF
(a)
i,(0*) = 0
L=10mH
+
C=1 mF
(b)
FKJURH 9.4 (a) Discharge of a capacitor through a practical inductor in series with a resistance /?,.
(b) Kquivalent circuit for t > 0. (c) Equivalent circuit at / = 0^ for calculating in which the
inductor has been replaced by an independent current source of value //(O^) and the capacitor by an
independent voltage source of value .
SoL u rioN
From the problem statements, = 10 V. When the switch moves to position B, =
- 10 V by continuit)' o f the capacitor voltage; the circuit now becomes a series RLC, for t
> 0, with / ?= / ?!+ R-y as shown in Figure 9.4b. The first step in the calculation o f the circuit
response is to find a second-order differential equation in the unknown From Example 9.3,
for the series RLC,
R dvf'
- *- + -------^ + \’r = 0 (9.19)
dt~ L dt LC ^
o R 1
s~ H— .y H------ = .v“ + ( 2 0 + /?-,)10^y-r-1 0 ^ = 0
L LC ( 9 .2 0 )
With this framework, we can separately investigate each o f the three cases.
Chapter 9 • Second Order Linear Circuits
2 = - 2 1 ,2 5 0 ± 18,750 = -2 5 0 0 , - 4 0 ,0 0 0 sec’ ^
v^t) = 10.667^-2’500^-0.667e’-^0’®®°'V
Exercise. You may verify' the above answer with the Student Edition o f MATLAB (version 4.0 or
later) by typing the command: y = dsolve(‘D 2y+42500*D y + leS^y = 0,y(0) = 10, Dy(0) = O’).
Case 2: /?, = 0 or /? = 20 Q
Step 1. If /?2 = 0> then from equation 9.20, the characteristic equation is + 2,000^ + 10^ = 0.
Since tr - Ac = -3 9 6 ,0 0 0 ,0 0 0 < 0, the roots are complex. From the quadratic root formula,
Step 2. Fiiid A and B. Ir remains to determine A and B in equation 9.23. From equation 9.23
and its derivative,
i / ^ 0 ^ ) = 1 0 = /I (9 .2 4 a )
and
= -oA + ( .) / = -1000/1 + 9950/y (9.24b)
A = 10 and = — = 1.005
CO,/
Exercise. You may verify the above answer with the Student Edition oF MATLAB (version 4.0 or
later) by typing the command: y = dsolve(‘D 2y+2000*D y + le8*y = 0,y(0) = 10, Dy(0) = O’).
Step 2. Find /Tj and Kj. From equation 9.25, its derivative, and the known initial conditions
from cases 1 and 2,
v^iO^) = 10 = A", (9.26a)
and
^yO ") = + K, = -10^ A', + A'2 = 0 (9.26b)
The waveforms o f V(fJ) for the three cases (underdamped, critically damped, and overdamped)
are plotted in Figure 9.5.
Chapter 9 • Second Order Linear Circuits 39'
FIGURU 9.5 Waveforms of vj,t) in Example 9.5 for three different degrees of damping. Critical
damping represents the boundary between the overdamped condition and the oscillatory behavior of
underdamping.
Exercise. Verify the answer calculated in Example 9.5 using the Student Edition o f MATLAB and
the “desolve” command.
On a practical note, commercially available resistors come in standard values each with an associ
ated tolerance. Tolerances vary from ±1% (precision resistor) to as much as ±20% . Further,
because o f heating action over a long period o f time, resistance values change. Given the above
example, in which the type o f response depends on the resistance, one can imagine the care need
ed in the design o f such circuits: without consideration o f precision and long-term heating effects,
a desired critically damped response could easily become oscillator)'.
Not all second-order circuits arc RLC. Some are only RC but with two capacitors and some are
RL with two inductors. Passive RC or RL circuits cannot have an oscillatory response. The proof
o f this assertion can be found in texts on passive network synthesis. However, with controlled
sources a second-order RC or RL circuit can have an oscillatory response that is not characeristic
o f a first-order circuit, but o f a second- or higher-order circuit. The example below illustrates the
analysis o f a second-order RC circuit containing controlled sources that has an oscillator)^
response.
398 Chapter 9 • Second Order Linear Circuits
EX A M PLE 9 .6 . This example illustrates the analysis o f the second-order RC circuit shown in
Figure 9.6. The objective is to find and for / > 0 given the initial conditions V(^{0)
= 10 V and = 0.
1 kn
I
'Cl
mF
r
FIGURE 9.6 Sccond-order RC circuit with controlled sources that has an oscillatory response.
S o l u t io n
Step 1. Write a dijferential equation in V(^^{t). From the properties o f a capacitor and KCL at the
left node,
di
d\
■^^ — 10 — 10^ v^2 ~ 1 0 'v’f'i (9.27a)
dt
_ 1()5 , Q3 (9.27b)
dt~ dt dt
10“6 — C l _ _ -0 .
dt
or equivalently,
= -1 0 ^ Vc, (9.28)
dt
*2 dt
After rearranging terms, we obtain a second-order differential equation in
(9.29)
'2
dt^ dt
Step 2. Determine the characteristic equation, its roots, and the form o f the response. The differential
equation 9.29 has characteristic equation
Chapter 9 • Second Order Linear Circuits
+ 1 0 ’° = 0
= - 5 0 0 + 799,998.75 = - a ± y W
From Table 9.1, complex roots imply an underdamped response o f the form
Step 3. Find A and B. \x. t = 0, z^qCO) = \() = A. Also, from equation 9.27a and the initial con
ditions,
Differentiating equation 9.30, evaluating at r = 0, and equating the result with equation 9.31
produces
dl "
Step 4. Set forth the final form o f V(-^{t). The final form o f the response is
Step 5. Plot the response. A plot o f the (underdamped) response is given in Figure 9.7.
Chapter 9 • Second Order Linear Circuits
Exercise. Construct a parallel RLC circuit to have the same second-order difFerential equation
model as 9.29. Note that there is no unique solution.
C H EC K : /?C= 10-^ and ZC = \0~^^
It is important to observe here that the design o f Example 9.6 achieves a second-order RLC
response without the use of an inductor, which is important for integrated circuit technolog)'.
The preceding section studied source-free second-order linear networks. When independent
sources are present, such as in the circuit o f Example 9.7 below, the network (differential) equa
tions are similar to the source-free case except for an additional term that accounts for the effect
o f the input:
Since the expressions o f Table 9.1 satisfy the homogeneous diflferential equation 9.10, the gen
eral solution to equation 9.32 follows by adding a constant to each o f the solution forms given
in Table 9.1. Specifically, the general solution o f the driven differential equation
tl~x dx
- + h — ^ cx = F (9 . 33 )
dr dt
where-v^j(f) is the solution to the homogeneous equation 9.10 (equivalently, equation 9.33 with F
= 0). Recall that the form o f x j t ) is determined by the roots of the characteristic equation r + bs
+ r = (i - ^])(j - ^2) = 0, given by the quadratic formula
h yjh~ - 4 c
i |-) = — ± --------------
9 9
Chapter 9 * Scconcl Order Linear Circuits *0 1
To verify that the structure o f equation 9.34 is a solution to 9.33 and to compute the value o^Xp sub
stitute the structure given by equation 9.34 into 9.33. Since satisfies the homogeneous equation
9.10, it contributes zero to the left-hand side. What remains is cXp= F. Therefore, = — which
is independent o f the roots o f the characteristic equation. However, if Re[j|] and Re[^2] < 0, then
tends to zero for large t. Hence x{t) tends to Xp for large t. Consequently Xp is termed the
final value o f the response.
Because o f the trifold structure o f as summarized in Tible 9.1, the solution form o f equation
9.34 once again breaks dow'n into three distinct cases. We summarize this trifold structure for the
constant-input case in Table 9.2.
-z- + h — + cx = F
dr dt
h yjb^ - 4 c
-7 = ---- ± ---------------
“ 2 2
x {t)= + Xp
F
with Xp = —. Further,
c
F
where again Xp = —, with
c
x{0^) = A + X p x(0^) = - g A +
K
V )
Case 3. The roots are real and equal, i.e., s, = s, and Ir -A c= Q . The solution
form is
.v(/) = (/ r ,+ ^ 2 0 ^ '
where Xp = —, and
c
x{Q^) = K^+Xp and .v W "-) = .v ,^ i-h AT2
402 Chapter 9 * Second Order Linear Circuits
The interpretation Xp = F/c. is a mathematical one. When the differential equation describes a lin
ear circuit with constant inputs, there is a physical interpretation o f Xp and a circuit theoretic
method for computing its value, even without writing the differential equation. Since = Xp
= a constant” satisfies the differential equation 9.33, it is also a constant solution to the circuit.
Hence Xp\s either a constant capacitor voltage or a constant inductor current. If a capacitor volt
age is constant, its current is zero; this is interpreted as an open circuit. Similarly, if an inductor
current is constant, its voltage is zero; this is interpreted as a short circuit. Therefore, Xp is an
appropriate (capacitor) voltage or (inductor) current obtained when the capacitor (or capacitors)
are open-circuited and the inductor (or inductors) arc short-circuited. The value o f Xp can be
obtained by analyzing the resistive network resulting when all capacitors are open-circuited and
all inductors are short-circuited. Recall that if Re[^j] and Re[^2] < 0, then x{t) tends to the constant
value Xp Physically speaking, then, equals either or /^(o))when Re[;,] and Re[^2] < 0.
Once the proper general solution structure is ascertained from Table 9.2 and the constant Xp is
found, the parameters and (or A and B) are computed by the same methods used in the
source-free case. The following example illustrates the procedure for a parallel RLC circuit.
E XA M PLE 9.7 . A step current input, = u{t) A, excites the parallel /^ZCcircuit o f Figure 9.8,
whose initial conditions satisfy /)(0) = 0 and y^^O) = 0. This simply means that the current source
turns on with a value of 1 amp at r = 0 and maintains this constant current excitation for all time.
The objective is to find the inductor current, for / > 0, for three values o f R: (i) R = 500 Q,
(ii) R = 25 n , and (iii) R = 20 Q..
So l u t io n
Because the circuit is a parallel RLC, the characteristic equation is
elicit)
0= = -0/4 + (HjB
dt
1 = 0^
^ \
From physical circuit From derivative o f
expression for ij^t)
fi = — = - 5 .0 0 6 3 x 1 0 " “
(0,y
Hence for / > 0,
/•^(r) = ^.-10.000^ [ cos(1.9975 x IQ5f) + 5.0063 x 10"2 sin(1.9975 x lO^r)] + 1
= 1 .0 0 13^>-*^’‘^^®'cos( 1.9975 x 105^+ 2.866'’) + 1 A
Case 2. For R = 25 O., the characteristic equation 9.35 reduces to + 4 x 10^5 + 4 x lO'® = 0.
From the quadratic formula, the roots are
j, 2 = - 2 . 0 x 105
-2xio-\
+ 1A
Case 3 . For R = 20 Q, the characteristic equation 9.35 is + 5 x 1O^j + 4 x 10^® = 0. From the
quadratic formula, the roots are
L L dt t=(f
1
Equivalently, -K^ ~ Solving these two equations yields K\ = ---- and K-, = —.
Therefore the actual response for r > 0 is ^ ^
Figure 9.9 displays a graph o f the response for each o f the three cases.
Exercises. 1. Show that for t > 0 , the differential equation for the circuit o f Example 9.7 with R
= 500 Q is 2
^ ^ + 2 x l 0 ‘' ^ + 4xl0'»,',.<n = 4 x l 0 JO
'
clr dt
2. Use MATLAB s “dsolve” command to verify the solution obtained for case 1 in Example 9.7.
Chapter 9 • Second Order Linear Circuits
In a linear circuit or system, the response to a step input often indicates the quality o f the system
performance. The problem o f measuring a batter}' voltage using a voltmeter is illustrative o f this
indicator. Here the battery dc voltage is the input and the output is the meter pointer position.
Connecting the meter probes to the batter)'^ terminals amounts to applying a step input to the
voltmeter circuit that drives a second-order mechanical system consisting o f a spring and mass
with friction. Naturally, one would like the pointer to settle on the proper voltage reading quick
ly. If the mechanical system is underdamped, then the pointer oscillates (undesirably) for a short
time before resting at its final position. On the other hand, if the mechanical system is over
damped, the pointer will not oscillate but may take a long time to reach its final resting point.
This also is undesirable. A near critically damped response is the most desirable one: the pointer
will come to rest at the proper voltage as quickly as possible without being oscillatory, and small
changes in the mechanical system will not make it oscillatory.
In the next example, we reverse the process o f analysis and ask what the original circuit parame
ters are given a plot o f the response that might have been taken in a laboratory.
E X A M PLE 9 .8 . Consider the circuit o f Figure 9.10, which shows the response, of a
(relaxed) series RLC circuit to the voltage input i/y^(r) = 10«(r) V. In laboratory, you have meas
ured the capacitor voltage values (approximately). If the response has the form v^it) = Ke~^^
cos(ojy + 0) + vYp find a , 0, K, and the values o f R and C i f it is known that L = 0.5
H. Your lab instructor has told you that to^and a are integers.
m \ ___________,
(a)
406 Chapter 9 • Second Order Linear Circuits
(b)
TABLE 9.3
S olution
Step 1. FindXp. By inspection, the curve is settling out at X p - 10 V.
Chapter 9 • Second Order Linear Circuits 40
Step 2. Find Now obscr\'e that the first two crossings o f v^^t) = 10 occur at r = 0 .3 1 6 sec,
0.839 sec (Table 9.3). This means that a full k radians is traversed by tlie cosine over [0.316,
0.839] , which is a half c>de or half period. So the period o f the cosine is 7 = 2(0 .8 3 9 - 0.316)
= 1.046 sec, making 2 ti
CO^/ = = 6.007 = 6 rad/sec.
Step 3. Find O . From Table 9.3, we know that two successive “peaks” occur at = 0.523 sec
and t-y = 1.5708 sec. This means that for ^ = 1 ,2 ,
COS(Ci),y/l + 0 ) C0.s((0^/r2 + 0 )
Thus
cos((0 ,//| +Q) V c i t j ) - ^ ^,-a(/2-/,)
cos(co^y/'> + 0 ) Xf,- (9.37)
Equation 9.37 simplifies because two adjacent positive peaks must be 2 tu radians apart, i.e.,
(co/2 + ^) = which means cos(to^^^ + H) = cos((i)y, + 0). It follows that
Solving leads to a = 2.
0 = 0.316 0 . 3 1 6 + B)
Thus 6 X 0.316+ 0 must equal 0.5?! or 1.5tt radians. We also know that since =0 =K
cos(0) + 10, we must have K cos(0) = - 1 0 . Since A"> 0 (always by convention), the value o f
cos(0) must be negative. This means 0.57t < 0 < 1.5Tt. So therefore, it must be that at the first cross
ing
D
o f 10 V
6 X 0 .3 1 6 + 0 = — or 0 = 2.1864 rad.
Therefore K = — ^— = 10.553.
cos(0)
Step 5. Find R and C. We know that the characteristic equation o f the series RLC circuit must be
^f? 1 ^ 02
s~ H—- . V.V +H-------= 2Rss +H—— = (5' + 2)*' + 6 " = s~ + 4.V + 40
------- = .v“ + IR
L LC C
Therefore
/? = 2 n and C = 0 .0 5 F
-H)K Chapter 9 • Sccontl Order Linear Circuits
In the previous examples one obsen'e that the characteristic equations are independent o f the
source values. I'his is a general property of linear circuits with constant parameters. Hence when
constructing the characteristic equation we may without loss of generality set independent source
values to zero; i.e., independent voltage sources become short circuits and independent current
sources become open circuits. With this operation, some circuits that appear to be non-series/par-
allel, become series/parallel. This allows us to easily compute the characteristic equation and then
use Table 9.2 and physical reasoning to obtain the solution without having to construct the dif
ferential equation explicitly. The following example illustrates this procedure for a pseudo-paral-
lel/.series RLC. The example will also illustrate the computation o f initial conditions due to past
excitations and the computation o f the complete response w'hen the input changes its dc level.
EXA M PLE 9.9 . The circuit of Figure 9. l i b is driven by the input o f Figure 9.11 a, i.e., vj^t)
-60//(-r) + G^u{t) + 60//(/‘ - 1) V. Our goal is to find the response for / > 0.
FIG URE 9.11 (a) Input cxcitation whose dc level changes at r = 0 and t = 1 second, (b) A pseudo-
parallel RLC circuit; i.e.,when the voltage source is replaced by a short, the circuit reduces to a paral
lel RLC whose characteristic equation is p- + — s+— = 0.
RC LC
So l u t io n
Step 1. Analysis at 0“. Here the circuit has been excited by a constant - 6 0 V level for a long
time. Therefore at ^ = 0", the capacitor looks like an open circuit and the inductor a short cir
cuit. Because the inductor looks like a short, the entire - 6 0 V appears across the 6 i l resistor,
making = - 60 V and /^(O") = - 60/6 = — 10 A.
Step 2. Analysis at 0^. By the continuity o f the capacitor voltage and the inductor current, the
equivalent circuit at 0"^ is given in Figure 9.12.
Chapter 9 • Second Order Linear Circuits 409
= -6 0 V
^ (0 *)= i,(0 )
= -10A
FICJURE 9.12 Equivalent circuit for analysis at 0^; the capacitor is replaced by a voltage source
o f value and the inductor by a current source o f value /^(O*) = //;(0“).
From the circuit diagram o f Figure 9.12, v^{0*) = 60 - ( - 60) = 120 V, = - 60/6 = - 1 0 A, and
iff-) = if^{0*)l5 = 40 A. It follows that
Step 3. Find the characteristic equation and the form ofthe response using Table 9.2. To find the char
acteristic equation, we set the independent voltage source to zero. I'he resulting circuit is a paral
lel RLC with characteristic equation
-) 1 I T -)
.v“ + ----- .V+ ------= .v“ + 4.V + 4 = (.V + 2 )“ = 0
RC LC
where R = 2 LI is the parallel combination o f 6 Q and 3 i i . The characteristic roots are s^i = - 2 ,
which correspond to a critically damped response o f the form (Table 9.2)
Step 4. Find constants in the response form for 0 < /■< 1. The input is constant for 0 < ^ < 1, but
changes its value to 120 V at r = 1 sec. However, the circuit does not know the input is going to
change, and so its response behaves as if the input were to remain at 60 V for all time: the circuit
cannot anticipate the future, and thus its response over 0 < r < 1 behaves as if no further switch
ing were going to occur. If no further switching were to occur and if the input remained at 60 V,
then in Figure 9.10b for large t the capacitor is an open circuit and the inductor is a short circuit;
hence Xp= 60 V. Under these same conditions we find and K-y. To find A^j, observe that from
step 2, = - 6 0 V. Evaluating the response form o f step 3 yields + Xp. Equating
these rwo expressions produces - 6 0 = ^'(;(0■^) = + X^ which implies thar = - 1 2 0 . To calcu
late Kj, observe that from step 2, /(-(O^) = 40 A. Since C = 0.125, it follows that
il{t) = (-2 0 + + 10 A
Step 5. Analysis at t = 1“. Although the circuit does not know the input will change at r = 1 sec,
we do and we must prepare for the analysis for f > 1. To do this we must evaluate the initial con
ditions at / = 1“ and then use the continuity o f the capacitor voltage and inductor current to
obtain the initial conditions at / = At r = 1“, using step 4 we have V(^\~) = = 54.59 V
and/^d") = /^ (r) = 10 A.
Step 6. Analysis att= I This step mimics step 2 for r = P . The capacitor is replaced again by an inde
pendent voltage source and the itiductor by an independent current source as shown in Figure 9.13.
Here v^{V) = 1 2 0 - 54.59 = 65.41 Vand =-in\ + + //(O'*') = ^ 120 - 54.59 ^
= - 9 .0 9 8 + 21.8 + 10 = 22.71 A. ’ I 6 3 >
Step 7. Computation o f the responsefor t> 1. Because the characteristic equation is independent o f
the input excitation, the form o f the response is almost the same as in step 3, except for the
replacement o f t by (/ - 1); this substitution follows by the time invariance (constant parameter
values) o f the circuit. Thus, for r > 1,
Since the source excitation for r > 1 is 120 volts, by inspection o f Figure 9.1 lb f^^oo) = 120
V. To find /f], 54.59 = ^'c^l"^) = ^p+ A',. This implies = -6 5 .4 1 . Finally, to find K-, consider
that
= 1 8 1 .7 = .ViA:, + K2 = 1 3 0 .8 + K 2
dt
/=r
Time in seconds
FIGURE 9.14 Complete analytical response of the capacitor voltage for 0 < r < 3 sec.
Exercise. Fill in the details for the computation o f i^it) = (-2 0 + 20/)^^ + 10 A for 0 < r < I and
then compute /y(/) for 2.5 > ^ > 1. Also, compute /^(/) for t> 2.5.
Despite the idea illustrated in Example 9.9, many second-order RLC circuits are not reducible to
series or parallel RLC circuits when the independent sources are set to zero. Furthermore, when a
dependent sourcc is present, the circuit is generally not reducible to a series or parallel RLC. In
such cases one ordinarily uses a systematic methodolog)' to compute the circuit’s differential equa
tion and, subsequently, the characteristic equation. This systematic procedure is described in more
advanced texts and in the second edition o f this text. Nevertheless, for some situations one can use
the earlier method s integro-differential equations, which must be differentiated again to eliminate
the integral. This is illustrated in the example o f the next section.
•il2 Chapter 9 • Sccond Order Linear Circuits
5. OSCILLATOR APPLICATION
An imporrant difference between first-order and second-order linear networks is the possibilit)' o f
oscillatory responses in the latter. In some applications sinusoidal oscillations are intended
responses, while in other applications oscillations arc undesirable. This section presents an exam
ple o f a Wien bridge oscillator circuit.
The goal is to build a circuit that generates a pure sinusoidal voltage waveform at a specified fre
quency. In theory, as per section 2 of this chapter, this is achievable by discharging a capacitor
through an inductor. In practice, both capacitor and inductor have losses. Losses cause the oscil
lation amplitude to decay eventually to zero. For sustained sinusoidal oscillations, some “active”
element such as a controlled source or op amp must replenish the lost energy. Note that these
active elements require a dc power supply for their operation. Ultimately the dc power supply
replenishes the power losses due to various resistances in the circuit.
E XA M PLE 9 .1 0 . Figure 9.15 shows a Wien bridge oscillator constructed with an op amp as the
active clement. Find the condition on the circuit parameters R^, and C for sustained sinusoidal
oscillation, and the frequency' o f oscillation.
r R.
(b)
S o l u t io n
From the principles described in Chapter 4, the non-inverting amplifier enclosed in the dashed box of
Figure 9.15a is equivalent to a voltage-controlled voltage source with a gain equal to {2Rj- + RJ}IRr= 3.
(See Chapter 4.) Replacing the dashed lx)x with this equivalent yields the simplified circuit o f Figure
9.15b. Using the simplified circuit, the first task is to derive die differential equation model of the circuit.
(9 .4 0 a )
Chapter 9 • Second Order Linear Circuits 413
(9.40b)
dt /?,C 7?1 dt
, V’2
^c\ ~ ^ ----- '---- (9.41a)
dt /?,
(9.41b)
dt dt^ l<2
Step 3. Substitute equations 9.41 into equation 9.40b. Substituting as indicated yields
- ± ^ =0 (9.42a)
dt~ /?2 dt R^C ^ dt Rj y /?, dt
d Vf 1 dv-> 1 dv-y
- + ----------- ^ + ---------- ^ +
dt~ Ro_C dt /?,C dt R^R^C- R^C dt
which simplifies to
d-V2 dvo Vo
= 0 (9.42b)
dt~ RjC R\C}
Step 4 . Compute the characteristic equation and determine the conditions for sustained oscillations.
The resulting characteristic equation is
1 1
s~ + hs + c = s~ + = 0 (9.43)
\ R ,C R ,C )
For sustained sinusoidal oscillations to occur, the roots must be purely imaginary. Thus the coef
ficient o f s must be zero, i.e.,
b= - ^ =0 (9.44)
R2C /?,c /e|/?2C
Thus the condition for sustained sinusoidal oscillations reduces to /?, = R-,.
Step 6. Find the frequency o f oscillation. Under the condition R^ = Rj, the roots o f the character
istic equation are
/?,C
(Oo = (9 .4 5 )
/?,C
A14 Chapter 9 • Second Order Linear Circuits
An examination o f equation 9.44 shows that if > R^, then ^ > 0 and the unforced response is
an exponentially decreasing sinusoid. On the other hand, if R^ < Rj, then b <0, and the unforced
response is an exponentially growing sinusoid. For the oscillations to start, the value of/?| should
be designed to be slightly smaller than Rj- Then the value for b in equation 9.58 will be negative,
producing an exponentially growing sinusoidal response. If all circuit parameters are truly con
stant, the amplitude o f oscillation would theoretically grow to infinit)'. In real oscillator circuits,
such growth is limited to a finite amplitude by saturation effects or nonlinearities that clamp the
response when the voltage swing grows large. The resulting waveform then only approximates a
pure sine wave. The analysis o f this nonlinear effect is beyond the scope o f this book. However,
the next example illustrates the growing oscillation when < /?2 and also shows the effect o f sat
uration to produce an approximate sinusoidal oscillation.
EXA M PLE 9 .1 1 . The circuit o f Figure 9.16a is a B2 Spice schematic for the Wein bridge oscilla
tor o f Figure 9.15. The op amp is a 741 with = 15 V. Suppose that (0) = 10 m V and
= 0. Observe that = 10 = 9.5 kQ. According to the analysis o f Example 9.10, the out
put voltage labeled IVout should be a growing sinusoid. The output response o f Figure 9.16b
shows this growth and the saturation effects induced by the op amp. The waveform is not a pure
sinusoid due to these saturation effects. Also note that the frequency o f oscillation is approximately
16 Hz, which is consistent with equation 9.45, i.e..
= 16.3 Hz
2k 2 k ^R^R2 C
C2 R1
(a)
Chapter 9 • Second Order Linear Circuits 15
(b)
FIGURE 9.16 (a) Schematic diagram of Wein bridge oscillator, (b) Voltage response showing grow
ing oscillation clamped at ±15 V due to saturation effects of op amp.
6. SUMMARY
'I'his chapter has explored the differential equation modeling and response computation of^ sec
ond-order linear circuits having either no input or constant input excitation. Such second-order
circuits contain at least two dynamic elements, either an LC, CC, or LL combination. Second-
order circuits may also contain active elements such as op amps. In contrast to first-order circuits,
second-order linear circuits allow for the possibilit)^ of damped and undamped sinusoidal oscilla
tions.
Analysis o f second-order linear circuits has two phases. Pha.se 1 entails the formulation o f the sec
ond-order differential equation circuit model. For simple I C parallel RLC, or series RLC, the cir
cuit model can be found by inspection.
Phase 2 o f the development centers on the solution o f the second-order differential equation
model o f the circuit. I'he first step here is to compute the (quadratic) characteristic equation and
then solve for the two roots. The roots o f the characteristic equation determine the t}'pe o f
response. The three t)'pes o f roots for a quadratic— real distinct, real identical, and complex—
specify the three response types of overdamped, critically damped, and underdamped, respective
ly. These three types o f responses characterize all second-order linear differential equation models,
be they o f electrical circuits, mechanical systems, or electro-mechanical systems.
Since sinusoidal waveforms are germane to many electrical systems, this chapter presented an
oscillator circuit that generates a sinusoidal waveform. O f the many types of oscillator circuits, we
chose one containing an RC circuit built around an op amp, avoiding the use o f an inductor.
Chapter 9 • Second Order Linear Circuits 41
' T h e notations and KO arc used interchangeably in the literature to den ote the first derivative ol v{t).
418 Chapter 9 • Sccond Order Linear Circuits
d^\>C
where all parameters are real, but not necessar
dr LC
ily positive. Prove that x(/) = 0 at some t = T <
COif and only if and K j have opposite signs.
By superposition, then, v^t) = A cos(ojr) + B
sin((or) V. Show that for a given A and B, there
6. (a) Consider case 2 o f Example 9.5. The
exist K and 0 such that Vf\t) = A cos(tof) + B
circuit is underdamped and the
sin(cof) = K cos(tor + 0). One concludes that
response is given by
the two solution forms are equivalent.
v^t) = [10 cos(9950^) + 1.005
sin(9950/)] = 10.05^’" ‘®‘^®' cos(9950f
2. Find the expressions o f the instantaneous
+ 5.7°)
energy stored in C and L for the circuit o f
How many cycles o f “ringing” occur in
Figure 9.1b. Show that the sum o f and
the voltage waveform before the peak
W^{t) is constant.
value drops from its largest value o f
10.05 to 10.05/f = 0.3 6 8 X 10.05?
3. By direct substitution, show that
(b) Suppose the characteristic polynomial
is written as + 2as + + to^ with
x{t) = (A'l + K^j)e-^‘
response form x{t) =
cos(w y + 0). Prove that for the under
satisfies the differential equation
damped case, the circuit will ring for
N = iy ijiln o) cycles before the ampli
x"{t) + 2}^'{t) + h~x{t) = 0
tude decreases to Me o i its initial
value.
where and AT, are arbitrar)^ constants,
x{t) =
Chapter 9 • Second Order Linear Circuits 419
Figure P9.7
1 ]
10. In the circuit o f Figure P9.10, suppose
.V l c ', Vi„{t) = 10 V, /? = 10 n , C = 0.4 mF, Z = 0.25
and H, and the switch opens at /^= 0.
(a) Compute , //(0~), and
f 1
yc(t) = Vo 1 - COS
■Jlc\ (b) Compute the energ)^ stored in the
inductor and the capacitor at / = 0.
(c) Using only energy considerations,
8. The circuit in Figure P9.8 is a dual o f the compute the maximum value o f
previous problem. f > 0.
(d) Find the analytical expression for
and verify the maximum value
of computed in part (c).
i,(t)
v„(t) (a) (t>0)
Figure P 9.11
100 0
© 25 mV (b) If v^O) = 2 V and /^(O) = 2.5 A, find
v^t) for t> 0 . Use MATLAB or equiv
alent to plot the V(^t) waveform and
Figure P9.14 verify that there is no zero crossing.
0
ijt)|
Figure P 9.18
'n^
Figure P9.19
422 Chapter 9 • Sccond Order Linear Circuits
i jt )
/ Y W
Figure P9.21
27. Almost 75% of fliilures in circuits, i.e., situa DRIVEN SERIES AND
tions where a circuit dramatioilly fails to perform
PARALLEL /?/.C CIRCUITS
as designed, are due to opens and shorts o f indi
28. (Initial condition calculation) For the cir
vidual circuit elements. Heating, c\'cling a circuit
on and off, etc., cause degradation in the circuit cuit shown in Figure P9.28, find //(O'*'),
parameters, resistances, capacitances, inductances, ;^ 0 ^ ), and in two steps:
etc. that often precipitates the short or open situa Sff/> I. Find V(^0~) and /^(O") by open-circuit
tion. For example, the material inside a resistor ing C and short-circuiting L.
might become brittle over a period of time and Step 2. Construct a resistive circuit valid at t =
finally crumble, leaving a break in the circuit. On O'*’ and from this find and
the other hand, the material might congeal or
become dense, decreasing the resistance. In the
problems below you are to determine the length of
time it takes for a circuit to move from an over
damped behavior to an imderdamped behavior
due to changes in the resistor characteristic as a
fiinction of time.
(a) For the parallel RLC circuit in Figure
P9.27a, suppose R = Rq + exp(/ - 5) H
where f > 0 constitutes time in years.
Determine the time f’ for which the cir 29. Consider the circuit o f Figure P9.29 in
cuit changes its behavior from over which = -10w(-/) + 20«(f) V, 7? = 20 Q,
damped to underdamped. C = 0.1 niF, and L = 0.25 H.
(b) For the series RLC circuit of Figure (a) Compute /^(0“), and
P9.27b, the resistor satisfies R = +
exp(r - 5)] n , where again t is time in (b) Compute the energ)'^ stored in the
years. Here it is presumed that the circuit inductor and the capacitor at r = 0.
is pan o f a larger piece o f electronic appa (c) Find the analytical expressions for the
ratus, such as a TV, which is used exten zero-input, zero-state, and complete
sively over a period o f years. The tiine t'
responses for Identif}' the tran
then is not connected with the response
sient and steady-state responses. Plot
time o f the circuit. Determine the time
V(^t) using MATLAB over [0, 40
f ' for which the circuit changes its behav
ior from overdamped to underdamped. msec].
(d) Find the analytical expressions for the
zero-input, zero-state, and complete
R„ = 0.8 Q R= responses for i/{t) . Plot ij{t) using
MATLAB over [0, 40 msec].’
i,(t)
ijO L= 1H
R„ = i s n
C = 1/36
i,(t)
33. Repeat Problem 32 for /? =40 Q and v-JJ)
fY Y \ k
= -0.5tt(-/) + 2«(/) V. Plots in M ATLAB L
should be for 0 < f < 800 msec. + n
,Vc(t)
(t)
34. Repeat Problem 32 for /? = 50 Q and and
v.^{t) =-0.5«(-/) - 2u{t) V. Plots in M ATLAB
should be for 0 < ? < I sec.
o
Figure P9.36
35. For the circuit of Figure P9.35, =
-0.5«M + 2u{t) A. 37. For the circuit of Figure P9.37, R^= 5 0.^
(a) I f /? = 2 Q, Z = 1 H, and C = 8 mF, = 2 0 a , C = 2.5 mF, L = 0.25 H, and v j t ) =
find and plot the zero-input, zero- 20«(^) - 20u(t-7) V, where T = 0.25 sec. o
state, and complete responses of v^{t) (a) Find the zero-input, zero-state, and
and ij\t) for / > 0. Identify the tran complete responses of v^t) for f > 0.
sient and steady-state parts of the Plot the complete response for 0 < r <
complete response. 0.25 sec. n
(b) Repeat part (a) for R= 22.5 Q.
Chapter 9 • Sccond Order Linear Circuits ■li')
(b) Find the zero-input, zero-state, and 42. Repeat Problem 40 for = 50 and Rj =
complete responses o f for t > 0. 200 n .
Plot the complete response for 0 < r <
0.25 sec. 4 3 . Consider the RLC circuit in Figure P 9.43
where /?^, = 60 Q, R^2 = Q, I = 4 H, and C
= 5m F
(a) = 100«(-f) mA and r^2(^) =
2 0u (-t) V. Find the response, for
t> 0. Plot for 0 < f < 1 sec.
(b) /y,(r) = 100u(-t) + 500u(t) mA and
1/^2 ^) = 20u(-t) V. Find the response,
for ^ > 0. Plot for 0 < / < 1 sec.
38. Repeat Problem 37 for = 50 Q, /?j = 200
Q, C = 0.05 mF, L = 0.5 H , and v j t ) =
- 5 0 « (- r ) + 50k(?) - 50«(r - 7) V, where T =
0.08 sec. However, only plot the complete
I V ,(t)
responses for 0 < f < 200 msec.
48. Repeat Problem 47, except find r> 0. PSEUDO SERIES AND
PARALLEL /?/.C CIRCUITS
49. riic current source with /^|(^) = 5//(-/‘) mA
and the voltage source = 10 V in Figure
52. Consider the circuit o f Figure P9.52 in which
P9.49 drive the circuit in which /? = 1 k^2, C =
/?! = 4 a = 4 a /. = 5/12 H, C = 25 mF
0.5 //F, and Z. = 0.184 H.
(a) Find the roots o f the characteristic
(a) Find t'(-(0‘"), /^(O"^), /^(0‘^), and
equation. C H EC K : - 8 , - 1 2
.,(0 ^ ).
(b) If v.^,{t) = -2Qu{-t) V, find 1/(^0-),
(b) Compute for r > 0. Plot for 0 < r
/j^(0"), ^i(0*), and for t >
< 5 msec.
0. Hint: After finding the initial volt
(c) Compute for / > 0. Plot for 0 < f
ages and currents, draw the equivalent
< 5 msec.
circuit valid for t > 0. Check your
(d) Compute for r > 0. Plot for 0 < r
answer for using the “dsolve”
< 5 msec.
command in MATLAB. Plot for
(e) Compute for r > 0 . Plot for 0 <t
0 < r < 2 sec using MATLAB or the
< 5 msec.
equivalent. Be sure you properly label
your plot.
(0 If = - 2 0 « (- f) + 20u(r) V, find
for / > 0.
(d) If y,./r) = -2 0 u (-t) + 20u(r) -2 0 u (t- 1)
V, find and plot v^^t) for 0 < r < 2 sec.
Figure P9.49
(d) Find die characteristic equation and 60. Reconsider the circuit of Figure P9.57.
natural frequencies o f the circuit. Suppose /?j = 80 Q, /?2 = 4 0 ^ = 2 H, and
(e) Determine the general form o f the C = 0.625 mF with = -1 5 0 tt(-/) + 150«(/)
response, ij{t)y valid for 0 < ? < 1. mA.
(f) Determine all coefficients in the gen (a) Find V(iO~) and V(^0*).
eral form o f the response. (b) Find ii(0~) and /^(O'*’).
(g) Determine the form o f the response, (c) Find t//^(0+) and /’c(0+).
for f > 1. Find the response. (d) Find the characteristic equation and
(h) Plot the response for / > 0 using natural frequencies o f the circuit.
MATLAB or the equivalent. (e) Find the response, t> 0.
(f) Find the response, t> 0.
56. Repeat Problem 55, except find Vf4,t).
6 1 . Consider the circuit of Figure P9.61
57. Consider the circuit o f Figure P 9.57 in Suppose /?j = 80 Q, R2 = 40 Q., L = 2 H, and
which = 2 £2, /?2 = 2 Q, Z = 0 .4 H, and C = C = 0.6 2 5 mF with = 300u(f) mA, Vq =
0.1 E 50 V, and /?3 = 20 Q.
(a) Find the roots of the characteristic (a) Find V(iO~) and y^^O'*’).
equation. (b) Find i^(0-) and z^(0+).
(b) If = -2u{-t) A. find t;c<0-), (c) Find v^{0*) and /’c (0 ‘*‘).
^> (d) Find the characteristic equation and
0, Hint: After finding the initial volt natural frequencies o f the circuit for t
ages and currents, draw the equivalent >0.
circuit valid for ^ > 0. Check your (e) Find the response, v^t), t> 0.
answer for ij{t) using the “dsolve” (f) Find the response, ijXt), t> 0.
command in MATLAB. Plot for
0 < ^ < 1 sec using MATLAB or equiv
alent. Be sure you properly label your
plot.
(c) If = -2u{-t) + 2«(?) A, find
for ^ > 0.
(d) If = -2u{-t) + lu{t) - l u { t - 1) A,
find and plot for ^ > 0.
compute its roots. Given tiie roots, tion with ij{t) as the unknown. Observe that
write down the general form o f the the derivative o f is present on the right-
response hand side.
(e) Solve for the unknown coefficients in
the response form o f part (d) and write ANSW'F.R: /"(/) + i\t) + i{t) = ^
down the exact expression for v^^t)
valid for ^ > 0.
40 fi
40 0
Figure P9.62
65. Consider the circuit shown in Figure
P9.65.
GENERAL SECOND-ORDER (a) Write a second-order differential equa
CIRCUITS tion with the unknown. Find the
63. In the operational amplifier circuit o f roots o f the characteristic equation.
Figure P9.63 is a second-order circuit. Suppose (b) If V, find V(^t) for r > 0.
= R^ = 50 k n .
(a )‘ Determine the values o f C, and Cj
that produce a characteristic equation
having natural frequencies at - 5 and
- 10 .
(b) Adjust the value o f /?j so that for a
step function input voltage, the value
o f the output voltage for large t is for
all practical purposes is 5. 66. Consider the circuit shown in Figure
(c) Compute when v^{t) = 2u{t) P9.66.
and all capacitor voltages are zero at t (a) Write a second-order differential equa
= 0. tion with the unknown. Find the
roots o f the characteristic equation.
Then find V(^t) when = u{t) A.
(b) Repeat part (a), for when is the
unknown.
67. The second-order circuit shown in Figure nearly 1 V in about 5 nsec, it behaves approxi
P 9.67 contains two capacitors. mately as an open circuit. The 0.1 pF capacitor
(a) Find the second-order differential is then charged up with a time constant of
equation with V(^ as the unknown. about 0.1 msec. As the larger capacitor is
Give the values o f s. charged up, the output across the smaller one
(b) If vci (0) = 2 V, and ^ ^ (0 ) = 4 V, find decreases toward zero.
V(^{t) for t> 0.
0.5 0
0.5 n
2F 2F
0.5 0
Figure P9.67
Figure P9.69
-R .
V,F
120
Figure P9.70
o
430 Chapter 9 • Second Order Linear Circuits
The control o f high-performance jet engines requires highly accurate pressure measurements, with
errors less than one-tenth o f 1% o f a full-range measurement, over a wide range o f temperatures,
- 6 5 ° to 200° F. The pressure range may be as low as 20 psia or as high as 650 psia. In jet (turbine)
engine applications, knowing pressure and temperature allows one to compute the mass (volume)
air flow, a critical aspect o f an engines performance. A pressure sensor is also a critical component
in the regulation o f aircraft cabin pressure. Such a sensor is depicted here along with a functional
block diagram o f its operation. A diaphragm consisting o f t\\'o fused quartz plates separated by a
vaciumi has a capacitance that changes as a function o f pressure and temperature. This quart/
capacitive diaphragm is an element in a bridge circuit. It is this bridge circuit, in conjunction with
detailed knowledge o f the characteristics o f a pair o f quartz capacitors over the required operating
range o f pressure and temperature, that enables accurate pressure measurements.
Because o f the small capacitances, on the order o f picofarads, associated with the quartz
(diaphragm) capacitor, the bridge circuit is driven by an ac source and is called an ac bridge.
Driving the bridge by an ac source moves its analysis outside the realm of the dc and step response
techniques studied in earlier chapters. New methods o f analysis, such as phasor analysis, are nec
essary. Phasor methods, the primary focus o f this chapter, allow us to analrze capacitive and induc
tive circuits excited by sinusoidal (ac) inputs. In particular, phasor techniques permit us to anal)'ze
an ac bridge circuit. Although the analysis ot the pressure sensor shown here is beyond the scope
o f this text, the chapter will end with a simplified pressure sensor circuit based on the one shown.
CHAPTER OBJECTIVES
1. Review and elaborate on the basic arithmetic and essential properties o f complex num
bers pertinent to sinusoidal steady-state analysis o f circuits.
2. Develop two complementary techniques for computing the response o f simple RL, RC,
and RLC circuits excited by sinusoidal inputs and modeled by differential equations.
3. Define the notion o f a (complex) phasor for representing sinusoidal currents and voltages
in a circuit.
4. Using the notion ol phasor, introduce the notions of impedance, admittance, and a gen
eralized Ohm’s law lor two-terminal circuit elements having phasor currents and voltages.
5. Utilizing the methods o f nodal and loop analysis and the nerwork theorems o f Chapters
5 and 6, analyze passive and op amp circuits by the phasor method.
6. Introduce the notion o f frequency response for linear circuits, i.e., investigate the behav
ior o f a circuit driven by a sinusoid as its frequency ranges over a given band.
SECTION HEADINGS
1. Introduction
2. Brief Review o f Complex Numbers
3. Naive Technique for Computing the Sinusoidal Steady State
4. Complex Exponential Forcing Functions in Sinusoidal Steady-State Computation
5. Phasor Representations of Sinusoidal Signals
6. Elementary Impedance Concepts: Phasor Relationships for Resistors, Inductors,
and Capacitors
7. Phasor Impedance and Admittance
8. Steady-State Circuit Analysis Using Phasors
9. Introduction to the Notion o f Frequency Response
10. Nodal Analysis o f a Pressure-Sensing Device
11. Summary
12. Terms and Concepts
13. Problems
Chapter 10 • vSinusoicial Steady State Analysis by Phaser Methods -133
1. INTRO DUCTION
Perhaps you have experienced the bouncing motion of a car with broken shock absorbers or
watched the (mechanical) oscillations o f a sw'inging pendulum. These motions reflect the sinu
soidal and damped sinusoidal oscillations in circuits with conjugate poles o f the characteris
tic equation, as detailed in Chapter 9. In this chapter we allow sources with sinusoidal forcing
functions (such as such as cos(ov + 0) or perhaps sin(d)/ + f))), which almost always result in
sinusoidal responses regardless o f the root locations o f the characteristic equation. A sinusoidal
voltage source models the voltage from the ubiquitous wall outlet.
If one hooks up an oscilloscope to measure a voltage in a linear circuit driven by sources with sinu
soidal values, the voltage may not look sinusoidal at first. However, if the circuit is stable, after a
sufficiently long period o f time the screen o f the scope will trace our a sinusoidal waveform. (Here
“stable” means that any zero-input response consists o f decaying exponentials or exponentially
decreasing sinusoids.) The eventual sinusoidal behavior is not immediately apparent because at
startup, stable circuits exhibit a transient response. “Transient” means that the circuit response is
transitioning— for example, from an initial voltage or current value to another constant value.
Flickering lights during a thunderstorm illustrate the phenomenon o f transient behavior: light
ning may have struck a transmission line or pole, causing the power system to waver briefly from
its nominal behavior.
Because sinusoidal excitations and sinusoidal responses are so common, their study falls under the
heading o f sinusoidal steady-state (SSS) analysis. Here “sinusoidal” means that source excita
tions have the form cos((or + 0) or K sin((0/ + 0). For consistency with traditional approaches,
we take cos(tor + 0) as the general input excitation, as shown in Figure 10.1, because sin(to/
+ 0) = cos(tof + 0 - 7t/2). Steady state mean that all transient behavior o f the stable circuit has
died out, i.e., decayed to zero. Observe that every sinusoidal waveform is periodic with angular
argument (to/ + 0). In terms o f angle, each cycle o f the waveform traverses 2 n radians. In terms of
time, each cycle covers a time interval o f T = 27r/o) seconds, called the period o f the waveform.
The number o f cycles contained in 1 second is called the frequency o f the sinusoidal waveform
and is denoted by / T h e unit for/is the herrz (Hz), meaning “cycles per second.” The quantity
OJ, which specifies the variation o f the angular argument (tor + 0) in 1 second, is called the angu
lar frequency o f the w^aveform. The unit o f (O is radians per second (rad/sec). From these defini
tio n s,/ = 1/7'= ti)/27i and o) = 2 k /
Stable circuits driven by sinusoidal excitations produce sinusoidal voltages and currents, as illus
trated in Figure 10.1. The output excitation in Figure 10.1 has the general form cos((0/ + (j))
to distinguish it from the input excitation, cos((0/ + 0). Because o f linearity, the circuit can
change only the magnitude o f the input sinusoid is changed to K^^) and the phase angle o f the
input sinusoid (0 is changed to (j)) while ensuring that the angular frequenq^ to remains the same.
For nonlinear circuits, to can and usually does change.
434 Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
CO
FlGURl'l 10.1 Graphical illustration ot steady-state sinusoidal linear circuit behavior. (K^and
could just as well be and or any combination thereof.) Note that 0 and (}) are often different and
that (0 is the same for both input and output excitations.
In Figure 10.1 the steady-state (voltage) response is cos(a)r + ({)). Alternatively this could have
been a current response, cos(a)r + (j)). Such waveforms have the equivalent structure A cos(wr)
+ 8 sin(oj/), deducible from trigonometric identities,
= A cos(a)f) + B sin(tor) ( 10 . 1)
where A = cos(({)) and B = — sin((})). Conversely, by summing the squares oi'A and B, one
obtains
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods 4 3 “^
n ------ ^ (10.2a)
V„, = + B-
By taking the inverse tangent o f the ratio o f - B and A, one obtains
(10.2b)
In using equation 10.2b it is important to adjust the resulting angle for the proper quadrant o f
the complex plane. Equations 10.1 and 10.2 turn out to be useful in developing a conceptually
simple, although naive, technique for computing the steady-state response using a differential
equation model o f the circuit, as explained in section 3.
Sinusoidal steady-state (SSS) analysis o f circuits draws its importance from several areas. The
analysis o f power systems normally occurs in the steady state where voltages and currents are sinu
soidal. Music is a rhythmic blend o f different notes. Mathematically, a musical (voltage) signal can
be decomposed into a sum o f sinusoidal voltages o f different frequencies. The analysis o f a sound
system typically builds around the steady-state behavior o f the microphone, the amplifier, and the
loudspeakers driven by sinusoidal excitations whose frequency varies from around 40 Hz to 20
kHz. Indeed, almost any form o f speech or music transmission requires an understanding o f
steady-state circuit behavior. There are many other areas o f applicability.
This chapter will introduce three techniques for computing the SSS response. The first two, some
what naive, approaches map out a natural motivation and path to the third, ver)' powerful tech
nique o f phasor analysis. Phasor analysis builds on the arithmetic o f complex numbers and the
basic circuit principles studied thus far. To set the stage for phasor analysis, section 2 reviews the
necessary basics o f complex number arithmetic. O f course, the student is assumed to have stud
ied complex numbers in high school and in prerequisite calculus courses.
Let = a + jb be an arbitrary complex number, where ^ . The real number a is the real
part o f 2 j, denoted hy a = Re[z,]. The real number b is the im aginary part o f z^, denoted by b
= Im[ 2 j]. It is simple to verify that
« = Re ^1 ^ i L ± i L
2
and
_
b = Im -1
2y
In Figure 10.2, the number Zj can be thought o f as a vector o f length p = >/«“ + b" = |zj|, which
makes an angle 0 = tan"*(^/<z) with the horizontal in the counterclockwise direction. (In comput
ing tzn~^{bln) it is important to adjust the angle (principal part) to be in the proper quadrant o f
the complex plane.) Hence Zj = /? + jb is completely specified by its magnitude p and angle 0 i.e.,
is the famous Euler identity. The Euler identity can be demonstrated by writing the Taylor series
for and recognizing it as the sum of the Taylor series for cos(0) added to j times the Taylor
series for sin(0). Note that the symbol Z. has two meanings, depending on the context o f its use:
(1) L z means angle o f the complex number z, and (2)pZ. 0 means the complex number whose
magnitude is p and whose angle is 0. The properties o f the exponential immediately imply that
(10.4)
/T r ; /4S'’
AN SW ERS:
2. Let z = 6 where ti/6 has units o f radians; for example, n rad equals 180°. Find the real
and imaginary parts o f z.
ANSW ER; 2 = S. 1962 + /3
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods 43*
3. Show by direct computation that since eJ^\ = cos(0j) + j sin(0j) and eJ^i = cos( 02) + j sin(f)2),
then = cos(0,+ 0-,) + j sin(0,+ (),) =
With these simple definitions, the product oF t^vo complex numbers z-^ = a +jb = and z-y
can be found using rectangular coordinates as
~ P 1P 2 ^2^
Equivalently,
z ,z 2 = 5 ^ - 53.130 X 1 0 ^ 3 6 .8 7 ° = 50 Z ( - 5 3 .1 3 ° + 3 6 .8 7 °)
= +36.87«)
Similarly, in rectangular coordinates the arithmetic for the division o f two complex numbers is
Cl _ a + jb _ {a + jb ){c - jd ) _ {a + jb ){c - jd )
Z2 c + jd ic + j d ) { c - j d ) c -+ d ~
{ac + bd) + j ( b c - a d )
~ 2 ^2
c- + d
-i38 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
O f particular concern in this chapter are equations involving mixed representations o f complex
numbers. For example, suppose an unknown complex number z = satisfies the equation
{a + jb) = f + j d
c + jd
a + jh
Since Vis the magnitude o f the complex number on the right-hand side o f the equal sign, it fol
lows that
+ _ V c-+ t/ ~
+ V «^+/r
Here we have used the fact that a complex number that is the ratio o f two other complex num
bers has a magnitude equal to the ratio o f the magnitudes. To determine the angle 0, one uses the
property that 0 equals the angle o f the complex number in the numerator minus the angle o f the
complex number in the denominator,
-1 -I
0 = Z (c + jd ) - Z {a + jb )= tan'
tan — -ta n —
S o l u t io n
Factoring Ve-^ out to the left and dividing by (-8 + j6) yields
- 8 + j6
Sometimes a function v{t) is a complex number for each t, such as if{t) = and v{t) will
satisfy some specific algebraic or differential equation. W hen this is the case, it is possible to use
the differential equation to find values for V and (j). The next two examples illustrate this strategy.
v{t)
satisfies the differential equation
^ + 2 - + 2v = 10e-'<“ ” “ °'
dr dl
Find the values o f A and (}) if O) is known to be 2 rad/sec.
S o l u t io n
Since the function v{t) must satisfy the differential equation, the first step is to substitute into the
differential equation. Substituting ‘1’^ into the differential equation and taking appropriate
derivatives yields
2 - ( 0 " + y2 (0 -2 + y4
440 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
The techniques of circuit analysis in this chapter will often require complex number arithmetic.
The voltages and currents o f practical interest are always real. T he complex arithmetic is a short
cut to computing “real” voltages and currents. The real quantities are obtained by taking the real
part o f the complex number or complex function. The various manipulations depend on some
general properties related to the real part o f complex numbers.
This property has a particularly nice application to summing trigonometric waveforms. Let v^{t)
= cos(oj/ + 55°) and r-y(t) = 10 sin(d)/ - 3 0 °) = 10 cosCd)/ - 120°). Note that a - 9 0 ° shift con
verts the sine to a cosine. Hence,
= Relt’-''^^^ + lOf*"/’ ’ *' )] by equation 10.4 and then factoring to the left
This sequence o f manipulations shows that the magnitude and phase o f two cosines at the same
frequency O) can be represented by distinct complex numbers. One can then add the complex
numbers and determine the magnitude and angle o f a third cosine equal to the sum o f the origi
nal two cosines. This presents a shortcut for adding two cosines together.
R e [a , 2 , + = a jR e [ 2 j] + a,Re[z-,]
which is a linearity property for complex numbers with multiplication by real scalars. 'I'he next
property, which underpins the techniques of this chapter, defines how differentiation can be inter
changed with the operation Re[-].
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
= Re - ( / t e - '" ' ) = Re
dt
Our fourth property tells us the conditions for the equality o f two complex-valued time functions.
Property 10.4. For all possibly comple.x numbers A and B, Re[/lf’-^^'^T = for all t if and
only A = B.
Taken together, the preceding properties imply a fifth, very important propert)'. Here note that a
complex exponential is sometimes referred to as a complex sinusoid.
Property 10.5. I he sum of any number o f (1) complex exponentials, say AjeJ^'^‘, or (2) derivatives
o f any order o f complex exponentials o f the same frequency co, or (3) indefinite integrals of any
order o f a complex exponential o f the same frequency O), is a complex exponential o f the same fre
quency (1).
This property is another foundation stone on which the phasor analysis o f this chapter builds.
Table 10.1 summari?,es the properties o f complex numbers.
Differentiation propert)' = Re
dt L J dt^ }
Property 10.5 o f the previous section suggests a technique for computing the SSS response o f a
circuit. The technique uses a differential equation model of an RL, RQ or RLC circuit as devel
oped in Chapters 8 and 9. In contrast to the dc sources in those chapters, suppose the source exci
tations have the form cos(cor+0). In addition we assume that the zero-input response consists
o f (eventually) decaying exponentials or (eventually) exponentially decaying sinusoids to ensure
that there is a valid sinusoidal steady state. Thus the form o f a first-order circuit differential equa
tion model with a sinusoidal excitation is
^/■v(/)
+ ax{t) = K^cos(co/+ 0 ) (10.5a)
cit
dt (10.5b)
df
Property 10.5 guarantees that the sum o f any number o f cosines or derivatives o f any order o f
cosines o f the same frequenc)' OJ is a cosine o f the same frequency O). Hence, the circuit response
x{t) in equations 10.5 has a steady-state cosine form o f frequency O). Further, the scaled sum ofx(f)
and its derivatives on the left-hand side o f each differential equation 10.5 must equal V^cos(o)/+0),
the input excitation. This also implies that the steady-state circuit response, x{t), is a cosine o f the
same frequency as the input, but not necessarily the same magnitude or phase. We conclude that
•’^■(^) = cos(co/‘+(j)) = A cos(tor) + B sin(to/). The SSS response is then specified upon finding A
and B. The following example illustrates this calculation.
E X A M PLE 10.4. Let the source excitation to the circuit o f Figure 10.3 be = /^cos(o)^).
Compute the SSS response /jr(r).
______________ i^lt)
i,» ©
So lu tio n
Step 1. Determine the differential equation model o f the circuit. From KCL applied to the top node
o f the circuit, i jt ) = i^(t) + //_(/)• Sincc the resistor and inductor voltages coincide, the t/-i rela
tionship o f the inductor implies that the inductor current satisfies the difTerential equation
( 10.6)
Step 2. Determine the form o f the response. Since the input is a cosine wave, the SSS response will
have the sinusoidal form
ij{t) = A cos(cor) + B sin(oj/) (10.7)
Step 3. Substitute the form o f the response {equation 10.7) into the differential equation 10.6.
Inserting equation 10.7 into the differential equation 10.6 and evaluating the derivatives yields
RA RB
= - 0) /\sin(coO + to cos((or) + — cos(co/) + — sin(co/)
I-t L
Step 4. Group like terms and solve for A and B. Grouping like terms leads to
r R R ] r R ]
Bco H— A ----- 1^ cos(o)r) + — B -A ii) sin(coO = 0 (10.8)
L L ', .L
To determine the coefficients A and B, we evaluate equation 10.8 at two distinct time instants.
Since equation 10.8 must hold at every instant o f time, it must hold at / = 0; i.e., at /= 0,
=0
or, equivalendy,
(10.9a)
-COA + - 5 = 0
L ( 1 0 .9 b )
444 Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
/?-/, „ CO/?L/
= — ------f — , B=
R~ + L^(0-' /?- + L“co“
. , , O^RLL .
Il (0 = — ------ ^ cos(o3 0 + — ^ SI n(co /)
R~ + L~0)~ R~ +
In the more common alternative form o f i^{t) = cos((or + ({)) as per equation 10.2a,
where
= - tan *
(10.10b)
This example has illustrated a procedure for finding the SSS response o f a circuit. Step 1 is to sub
stitute an assumed sinusoidal response form, such asy4cos(cor) + Bs\n{a)t), having unspecified con
stants A and B, into the differential equation and evaluate all derivatives. Step 2 is to group like
terms, and step 3 is to compute the constants A and B. After finding A and B, one computes the
magnitude, and phase (j) o f the cosine cos(oj^ + (|)) via equations 10.2.
The next section offers an alternative approach. Using complex excitation signals o f the form
y^eJ^Mt +0)^ computes and (j) by a more direct route.
Complex exponential forcing functions are simply complex exponential input excitations o f the
form or ^;(o)/ + 0) PfQji-j properties o f complex numbers in section 2, we can
replace the input excitation cos{wt + 0) and the assumed circuit response cos(cor + cj)) = /I
cos(to^) + j5sin(co^) with their complex counterparts and respectively, with
out any penalt)'. To recover the actual real-valued responses, we simply take the real parts o f the
complex quantities. Again this is justified by properties 10.1 through 10.5. This process o f sub
stitution and subsequent taking o f real parts actually simplifies the calculations developed in sec
tion 3, because of the simple differential and multiplicative properties o f the exponential function.
The following example illustrates a more efficient calculation o f the steady-state response using
complex exponentials.
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods ■vn
EX A M P L E 10.5. For the series RC circuir o f Figure 10.4 let v^it) = K^cos(cof). Compute the
steady-state response
Step 1. Construct the dijferential equation o f the circuit. Writing a loop equation and substituting
for i(^t) yields
at ( 10. 11)
Step 2. Substitute complex forms o f the input and response into the differential equation. If v^{t) were
to be equal to the complex exponential then the response would be ^ However,
if vj^t) = = V}cos(cor) (as is the case), then V(^t) = from the properties
o f complex numbers. Hence, for the moment, let us set vj^t) = and agree that =
appropriate real parts.
Substituting the complex expressions into the circuit differential equation 10.11 yields
After canceling the terms, factoring V„,e^^ out to the left, and dividing through by {jixiRC +
1), we obtain
1 + jOdRC ( 10 . 12)
Step 3. Determine the magtiitude and the angle (}). Equating magnitudes on both sides o f equa
tion 10.12 yields
(10.13a)
Vl + ( o V c '
Step 4. Determine the steady-state response. Using equations 10.13 the desired response is comput
ed by taking real parts:
4 i6 Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
\'cU) = Re (10.14)
In deriving the relationship 10.12 from the ciifterential equation 10.11, we utilized a complex exponen
tial ftinction as the circuit input. A complex exponentid input is not a signal that cm be generated in the
laborator)'. Nc-vertheless, it is often used in advanced circuit theor)' to simplift' the derivation of many
important results, as was done in the preceding example. If one does not mind a more lengthy derivation,
then the s;imc result (equations 10.12 through 10.14) am be obtained without the flaitious complex
exjxjnentid excitation. For example, let the voltage source in Figure 10.4 represent a reiil signal source
/ec— ( rc v...e + Re = Re
Making use of properties 10.2 and 10.3, move the position o f the operator Re[] outside the first
term to obtain
Re + Re = Re
(10.15)
^ + <l>) =
This is precisely the equation following equation 10.11 that leads to equations 10.12, 10.13, and
finally 10.14.
As we can see, the use o f complex exponentials does indeed lead to a more direct calculation of
the SSS response. However, this method and the method o f section 3 require a difFerential equa-
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods 44 /
tion model o f the circuit. For circuits with multiple sources, dependent sources, and many inter
connections o f circuit elements, finding the differential equation model is ohen a nontrivial task.
In the next section we eliminate the need to find a differential equation model o f the circuit by
introducing the phasor concept.
Recall is shorthand iov = /lcos(({)) +yy4sin((})). If the frequency co is known, then the
complex number AL^ completely determines the complex exponential In turn, if to is
known, then AL^ completely specifies A cos(ojr +({)) = This means that the com
plex number A/L^ can represent a sinu.soidal function A cos(ojf + (})), whenever (O is known.
Complex number representations that denote sinusoidal signals at a fixed frequency are called
phasors. A phasor vo\iz^c or current will be denoted by a boldface capital letter. A typical voltage
phasor is V = and a ty'pical current phasor is I = Por example, the current i{t) =
25cos((Of + 4 5 °) has the phasor representation I = 25Z-45°. The voltage v{t) = - 1 5 sin(tor + 3 0 °)
= 15 cos(w^ + 120°) has the phasor representation V = 15/-120°.
As all voltages and currents satisfy KVL and KCL, respectively, one might expect phasor voltages
and currents to do likewise. This is not patendy clear. The following simple example demonstrates
why this is true for KCL.
^4(^) = - ijit) +
Using trigonometric identities or property^ 10.1 to combine terms on the right-hand side leads to
i^it) = 1 0 c o s (tO f- 6 0 °)
4 <S Chapter 10 • Sinusoidal Steady State Analysis by Pliasor Methods
10Z-60O = = I, - I 2 + I3
Thus the phasors (which have both a real and an imaginar)' part) satisfy KCL. KCL is satisfied
because i^{t) = /j(^) - i-,{t) + /^(r) implies
It is the properties o f complex numbers and the fact that an equation is true for all t that guaran
tee that phasors satisfy KCL. Although not general, the argument is sufficient for our present ped
agogical purpose. A similar argument implies that phasor voltages satisf}' KVL, as illustrated by
the following example.
EXA M PLE 10.6. Determine the voltage across the resistor in the circuit of Figure 10.6 using the
phasor concept.
+ . -------------- •+
S o l u t io n
Firsr note that 19.68 sinltDr+l 52.8”) = 19.68 cos(co/+152.8“—90") = 19.68 cos(cor+62.8°). From KVL,
Vf^it) = v^{t) - Vjit) + v^{t)
Exercise. In Figure 10.6, suppose yj(r) = 10 cos(cor) V, v-^ (/) = 10 co s((o r- 0.5tc) V, and
V3(/)= I 0 V 2 cos(cOf - 0.25ti ) . Find the phasorV^and then
Given that phasor voltages and currents satisfy K\^L and KCL, respectively, it is possible to devel
op phasor O hm s law-like relationships for resistors, capacitors, and inductors operating in the
SSS. This would allow us to do SSS circuit analysis with techniques similar to resistive dc analy
sis. The next section takes up this thread by introducing the notion of (phasor) impedance.
Ohm’s law-like relationships do exist for resistors, capacitors, and inductors operating in the SSS.
The constraint, operating in the SSS, suggests that any Ohm s law-like relationship should be
dependent on the sinusoidal frequency.
The first objective o f this section is to derive three Ohm’s law-like relationships, one each for the
resistor, the capacitor, and the inductor. The relationships each take the form V = Z(/to)I, where V
is a phasor voltage, I is a pha.sor current, and Z(/to) is called the impedance o f the device: Zy^(;tij) for
a resistor, for a capacitor, and Z/(/co) for an inductor. The fact that the phasor voltage V is a
function Z(/co) times a phasor current I indicates a clear kinship with Ohm’s law for resistors. Indeed
the unit o f impedance is the ohm because it is the ratio of phasor voltage to phasor current. The
impedance Z(/cij) explicitly shows that the relationship is potentially frequency dependent.
450 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
The derivation o f these elementary impedance concepts will build on the assumption that all volt
ages and currents are complex sinusoids o f the same frequency represented by complex phasors.
This is permissible because real sinusoids can be recovered from complex sinusoids simply by tak
ing the real part. To this end consider the resistive circuit o f Figure 10.7a.
V „(t) =
i,(t) =
j 0j(u)t + 6)
(a) (b)
FIGURE 10.7 (a) Resistive circuit driven by complcx current,
(b) Equivalent phasor representation of the circuit in (a).
where and Z^(yco) = R\s the impedance o f the resistor defined by equa
tion 10.17. Ideally the resistor impedance is independent o f frequency. Thus = Rlj^ Z.0. If
= cos(u)r + 0) = Kq[ I ^ then Vj^it) = Rlj^ cos(co/ + 0) = This phasor
relationship restates Ohm’s law for complex excitations. The distinctiveness o f phasors comes with
their application to inductors and capacitors.
Now consider the inductor circuits o f Figure 10.8. Assume the circuit o f Figure 10.8a is in the
steady state.
i jt ) =
A L
-I-
Remainder Remalnder
V = jcoL 1^
of circuit of circuit
(a) (b)
FIGURE 10.8 (a) Inductor having complex exponential voltage and current, (b)
Phasor relationship of (a).
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods /is:
The complex current and voltage associated with the inductor are, respectively, /^(f) =
and v^it) = Substituting these expressions into the defining equation for an inductor
yields
j = jcoz.
( 10. 18)
in which case the inductor Impedance is derived as Z^(/co) = p iL . The inductor impedance clear
ly depends on the value ol the radian frequency CO. Specifically, if U) = 0, then the impedance o f
the inductor is 0, i.e., in SSS the inductor looks like a short circuit to dc excitations. If co = oo, the
impedance is infinite, i.e., in the steady state the inductor looks like an open circuit to signals o f
very high frequency.
Equation 10.18 exhibits a frequency-dependent Ohm’s law relationship for the inductor. From the
properties o f the product o f two complex numbers, the polar form o f the voltage phasor is
\ l = (yco£)I^ = (03/./,) ^ ( 0 + 9 0 °)
Hence if
/^(/) = ^ = /, cos(tor + 0) A
then
p^(t) = Re[/wZ/^f’>(‘'^' ^ 0)] = RelcoL/^e^^^" " ‘-^0“)] = coL/^ cos(tof + 0 + 9 0 °) V
From this relationship one sees that the voltage phase leads the current phase bv 9 0 ” . Equivalently,
one can say that the current lags the voltage by 9 0°. This leading and lagging takes on a more con
crete meaning when one views phasors as vectors in the complex plane, as per Figure 10.9, which
shows that the voltage phasor o f the inductor always leads the current phasor by 9 0 ” .
452 Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
The capacitor has a similar impedance relationship, derived as follows. Assume the circuit o f
Figure 10.1 Oa is in the steady state.
C'
►
+
Remainder Remainder
of circuit of circuit
C -
V J L J
(a) (b)
FIGURF. 10.10 (a) Capacitor having complex exponential voltage and current.
(b) Phasor relationship o f (a).
The complex current and voltage associated with the capacitor are, respectively, i^^t) =
and V(^t) = Substituting these expressions into the defining equation for a capacitor
yields
Vce 7(o)r+<!>)■
I^=yo)C V ^
or, equivalently.
- — — i c - ^ cO ^ ^ )Ic (10.19)
ycoC
Equation 10.19 defines the capacitor impedance as Z(^j(a) = l/(/a)Q. if co = 0, the impedance
o f the capacitor is infinite in magnitude. This means that in SSS the capacitor looks like an open
circuit to dc signals. On the other hand, if OJ = co, then the capacitor has zero impedance and looks
like a short circuit to large frequencies.
Chapter 10 * Sinusoidal Steady State Analysis by Phaser Methods 4 S3
Vr = (10.20)
70) C I c =-
(oC
t: - 90°)
Equation 10.20 has a vector interpretation in the complex plane, as shown in Figure 10.11
Imaginary axis
F ' l C l J R l i 10, 1 1 Diagram of capacitor voltage and current phasors where the voltage phasor lags the
current phasor by 90°.
The diagram o f Figure 1 0 .1 1 indicates that the capacitor voltage lags the capacitor current phasor
by 9 0 ° or that the capacitor current leads the capacitor voltage by 9 0 °, which is the opposite o f
the case for the inductor.
^ ycoL ^
and that
V/
i r ( t ) = — ^COS(Cl)/ + 0 - 90°)
(OL
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
I^ = ya)C V ^
and that
i(^t) = + ^ + 90”)
iiCt) ijt )
V ,(t) = V ,(t) =
\J 0 j((i)t + 0) \J g j((» t + 0)
(a) (b)
I-ICURE 10.12 (a) inductor driven by vohagc source, (b) Capacitor driven by voltage source.
Recall that resistance has a reciprocal counterpart, conductance. Likewise, impedance has a recip
rocal counterpart, admittance. Admittance has units o f siemens, S, as does conductance, fh e
admittance, denoted by l^yto), associated with an impedance, Z(/co), is defined by the inverse rela
tionship
K(./co) =
ZOCO) ( 10.21)
provided Z{j(M) is not equal to zero ever)^vhere. What this means is that the phasor i-v relation
ship o f a resistor, capacitor, and inductor satisfies an equation o f the form I = K(y(o)V. Hence, the
admittances o f the resistor, inductor, and capacitor are respectively given by
The impedance and admittance relationships o f the resistor, capacitor, and inductor are summa
rized in table 10.2.
Impedance Admittance
2/?(y(0) - R K^(;o)) = i
yV-(./co) = ycoC
j(oC
Z Lijoi) = j(i)L
JY Y V j(i)L
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
In the next section the notion o f impedance is applied to an arbitrar)' two-terminal network. This
generalization will allow us to consider the impedance and admittance o f interconnections o f
capacitors, inductors, resistors, and dependent sources.
For the resistor, the inductor, and the capacitor, the impedance equals the ratio o f the respective
phasor voltage to the phasor current. Analogously, the impedance o f any t\vo-terminal circuit, as
illustrated in Figure 10.13, is the ratio o f the phasor voltage to the phasor current, i.e..
Z „ S j^ ) = ^ = R + jX (10.23a)
1 in
O- Two
+
Terminal
Circuit
Z (j(o)orY,„(j(o)
Because impedance is the ratio o f phasor voltage to phasor current, its unit is the ohm. Inverting
the relationship o f equation 10.23a defines the adm ittance o f a two-terminal device as the ratio
o f phasor current to phasor voltage, i.e..
I;
(10.23b)
As an example, the impedance o f an inductor is jwL and its admittance is \/(J(.oL). Historically,
impedance and admittance were first defined as per equation 10.23. However, with the wide
spread use and utility o f the Laplace transform (Chapter 12) in the past several decades, imped
‘0 6 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
ance and admittance have become understood as much broader and more useful concepts than
the steady-state presumptions o f equation 10.23, as set forth in Chapter 13.
In general, admittances and impedances are rational functions with real coefficients o f the com
plex variable Ju). At each d) the impedance and the admittance are generally complex numbers.
Since a complex number has a real part and an imaginary part, we can further classify the real and
imaginary parts o f an impedance or an admittance. For an impedance Z(yw) the expression
lm[Z(/co)] = X is called the reactance o f the two-terminal element, while Re[Z(/(o)] = R refers to
its resistance. 1‘urther, for an admittance Vijto), Im[K(/(o)j = B is called the susceptance of the
two-terminal device whereas Re[)1[yoj)] = G is referred to as the con d u aan ce. These definitions
are summari/,cd in Table 10.3.
Impedance Admittance
V/„
Z{J(}^) = — = R + jX Y( jo) ) = ^ = G + jB
I in V/«
Using equations 10.23, one can compute the equivalent impedance o f two devices in
series, as in Figure 10.14a. Flere = ^] + % Ohm’s law for impedances, V , = Z|(/to)I, and
V 2 = Zol/w) I ,. But I, = I 2 = I/„- Mence,
I.e.,
This simple derivation has another consequence: given Zy^j(yco) = Z,(/to) + Z 2 (/co) and the fact that
i = 1, 2, a simple substitution yields the voltage division formula,
Z/O )
V;.,
' Z,(./(D) + Z2(7(0)
(10.25)
Kquations 10.24 and 10.25 are consistent with our early development o f series and parallel resist
ance.
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods 4V
V. Y,(jco) V. V,
o-
YJjco)
(b)
FIGURE 10.14 (a) Two impcclanccs in series, (b) Two admittances in parallel.
Exercises. 1. Duplicate the derivation ol" equation 10.23 for three impedances in .series.
2. Derive a formula for voltage division when there are three impedances in series.
v,„ v,„ V, V,
since
^1 (./“ ) = “ f Y2 ijc o ) = ^
V| Vo
we conclude that
(10.26)
Exercises. 1. Duplicate the derivation o f equation 10.26 for three admittances in parallel, i.e.,
show that )^y„(/co) = Kj(/o)) +
2. Show that the equivalent impedance o f two devices, Z,(/co) and Z t(/co), in parallel is given by
(10.27)
3. Show that the equivalent admittance o f two devices, Kjlyco) and Y-,{ji.o), in series is given by
458 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
>U/co)>'2(yw)
■ K|aco) + K ,0 (o )
(10.28)
C 1C 2
4. Show that the admittance o f two capacitors, Cj and C^, in series is y’CO
C\ + C-)
5. Show that the impedance o f two inductors, Z.j and Z.^, in parallel is /O)
Lt + Lo
Now the derivation o f equation 10.26 leads to a current division formula as follows. Since
= Kj(/co) + and since for /= 1 ,2 , one immediately obtains the current divi
sion formula,
n-(7(0)
I; = I, (10.29)
' K ,(7co)+ r2(y(o)
Since devices represented by impedances or admittances must satisfy KVL and KCL in terms of their
phasor voltages and currents, and since each device so represented satisfies a generalized Ohms law, i.e.,
V = Z(;o))I or I =
it follows that impedances can be 7nauipiilated in the same manner as resistances, and admittances in the
same manner as conductances. The voltage division formula o f equation 10.25 and the current divi
sion formula o f equation 10.27 illustrate this fact. Example 10.7 further clarifies these statements.
3. Compute the equivalent inductance for Figure 10.15a and the equivalent capacitance for the
circuit ol Figure 10.15b.
ANSWI-RS: , Cj + Ct+C^
L| L. L,
4. Find ^ in terms o f for each circuit in Figure 10.15.
I
O- I
+
(a) (b)
FIGURE 10 . 15 . (a) Set of three parallel inductors, (b) Set of three parallel capacitors.
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods 4S9
EXA M PLE 10.7. For the circuit o f Figure 10.16, compute the input impedance ^/„(/Co) when co
= 500 rad/sec.
S o l u t io n
As shown in Figure 10.16, Z^-^^(/500) can be seen as the sum o f three impedances, + ^2 + Zy
Our approach is to first calculate Zy for each /.
Step 1. Compute Z y Since this is an I C series combination,
1
A = j 5 0 0 x 0 .0 0 5 - = - j 2 .5 Q
5 00 X 0 .0 0 0 4 )
Step 2. Compute Z-) = MYj. From the propert)' that parallel admittances add and series imped
ances add.
1
Y2 = y'500 X 0.0 0 0 2 -h :— = JO. 1 -f 0 .0 4 - jO .02 = 0 .0 4 y0.08
10 + ( 1 0 -H7 IO)
H en ce,Z 2= 1/^2 = 5 - ; 1 0 a
Hence, Z 3 = 5 + J5
Calculations performed in this example are most easily done with an advanced calculator or in
M ATLAB. For example, in M ATLAB the command for computing is “Z3 =
l/(sqrt(0.02)*exp(i*pi/4) - j/(50()*0.01)).”
EXA M PLE 10.8. Compute the input impedance Zy^^(/a)) o f the ideal op amp circuit o f Figure
10.17.
I,
V,
S o l u t io n
The trick to solving this problem entails full use o f the ideal op amp properties discussed in
Chapter 4.
Step 1. From the properties o f an ideal op amp, from KVL, and from Ohms law,
V 2 - V „ , - M 3 = V,„ (10.30)
This follows because the voltage across the input terminals of each ideal op amp is zero and no
current enters the + or - terminal o f each ideal op amp. This implies that
Step 2. Using the phasor voltage division formula o f equation 10.25, it follows that
R
=
T " '
R+ -
j(oC
or, equivalently,
1
v,= 1+ -
(10.31)
Jc^RC)
Here, o f course, because o f the idealized properties of the op amp, the voltage appears across
the resistor R in the leftmost op amp.
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
Step 3. Writing a node equation at the inverting terminal o f the rightmost op amp yields
2V,„ = Vi + V , (10.32)
Step 4. Substituting equations 10.30 and 10.31 into equation 10.32 yields
V.
=\. + — + -/?!■
jCdRC
Equivalently,
Z i„ u ^ )= ^ = m ~ c
I//J (10.33)
Equation 10.33 suggests that the op amp circuit o f Figure 10.17 can replace a grounded inductor
whose impedance is jii)L with proper choice of R and C, i.e., L = R^C. In integrated circuit tech
nology it is not possible to build a wire-wound inductor. Instead, inductors are “simulated” by cir
cuits such as that o f Figure 10.17.
The next section continues to develop our skill with and deepen our understanding o f the phasor
technique by computing the steady-state responses o f various circuits.
This section presents a series o f examples that illustrate various aspects of’ the phasor technique.
Our purpose is not only to demonstrate how to compute the SSS, but also to illustrate the pha
sor counterparts o f Thevenin equivalents, nodal analysis, and mesh analysis. Our first example
reconsiders the parallel RL circuit o f Example 10.4, together with the series RC circuit o f Example
10.6. We will demonstrate the superiority o f the phasor technique over the methods presented in
sections 3 and 4.
i6: Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
E X A M PLE 10.9. Compute the steady-state voltage V(^t) for the circuit o f Figure 10.18 when i^{t)
= cos(lOOr) A.
R=10Q
21,(t)
i ( t ) 0
R = 100 C=1mF
L = 0.1H
S o l u t io n
R 1 1 1
Z -4 5 " A
(10.34)
R
Step 2. Use equation 10.34 and voltage division on the RC part o f the cirniit to compute \ q Using
voltage division and equation 10.34, the capacitor voltage phasor is
ycoC
j(oC
Step 3. Determine V(^t). Converting the phasor o f equation 10.35 to its corresponding time
function yields
v^t) = co s(1 0 0 ^ - 9 0 °) = sin(100/) V
The next example illustrates voltage division with phasors as well as the basic impedance relationships.
SO LU TIO N
Zr,.U2) = - 4 - = 2 v /2 Z 4 5 " = 2 + j2
The next example illustrates the computation o f a Thevenin equivalent circuit with the aid o f
nodal analysis. Because impedances may be manipulated in the same manner as resistances and
admittances in the same manner as conductances, the Thevenin theorem, the source transforma
tion theorem (Chapter 5), and node and mesh analysis (Chapter 3) carr)' over directly.
46^ Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
S o l u t io n
Find theThevenin equivalent circuit, and then using theThevenin equivalent, find Vj{t).
Step 1. Establish nodal equation. A nodal equation at the left node o f Figure 10.20 in terms o f pha-
sors is given by
Step 2. Determine the relationship between and The relationship between and as
determined by the dependent source is
V z - V , , = 0 .2 5 [ ;2 V J
Equivalently,
V , = (1 - 10.37)
h = m -j)y o c
Step 4 . Compute the Thevenin equivalent impedance Consider the circuit o f Figure 10.21,
which is the phasor version o f Figure 10.20 with the output terminals short-circuited. Hence, the
short-circuit current phasor is
1= A
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods 4(n
Z,/,0'4) = ^ = ( 0 . 4 - y O . S ) Q
.VC
Step 5. Interpret to generate the Thevenin equivalent circuit. To physically interpret theThevenin
equivalent impedance, consider that
Thus, = 0.4 Q and C = 0.3125 F. Hence, the desired Thevenin equivalent circuit (valid at O)
= 4 rad/sec) has the form sketched in Figure 10.22.
0 .3 1 25F
-O-
Step 6. Compute by voltage division. Using voltage division on the circuit o f Figure 10.22,
1.2
= = (0.6 + y 0 .3 )(0 .8 9 4 Z 6 3 .4 3 °)
l.2 + ( 0 .4 - y 0 . 8 )
= 0 . 6 Z - 36.87° V
Converting the load voltage phasor to its corresponding time-domain sinusoid yields
y^^(t) = 0 . 6 cos(4/^ - 3 6 . 8 7 ° ) V
166 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
EX A M PLE 10.12. Determine the phasor voltage and the corresponding time function vj,t)
for the circuit o f Figure 10.23 if co = 100 rad/sec.
j60Q
I'lG U R E 10.23 Phasor domain circuit for Example 10.12. Ail clement values indicate phasor
impedances at 100 rad/sec.
S o l u t io n
To solve this problem, it is convenient to execute a source transformation on the independent cur
rent source and to combine the impedances o f the parallel combination o f the capacitor and
inductor on the right-hand side o f the circuit. After executing these rwo manipulations, one
obtains the new circuit of Figure 10.24.
FIGURE 10.24 Phasor domain equivalent circuit to that of Figure 10.23. All element values indi
cate phasor impedances at 100 rad/sec. I denotes a phasor loop current.
The frequency response o f a circuit is the graph o f the ratio o f the phasor output to the phasor
input as a function o f frequency, i.e., as the frequency varies over some specified range. Since the
phasor input and the phasor output are complex numbers, the frequency response consists o f tw^o
plots: (1) a graph o f the magnitude o f the phasor ratio and (2) a graph o f the angle o f the phasor
ratio. Such graphs indicate the magnitude change and the angle change imposed on a sinusoidal
input to produce a steady-state output sinusoid. In steady state, the magnitude o f the output sinu
soid is the product o f the magnitude o f the input sinusoid and the magnitude o f the frequency
response at the frequency o f the input. Similarly, the phase o f the output sinusoid in steady state
is the sum o f the input phase and the frequency response phase at the input frequency. This prop
erty takes on greater importance once one learns that arbitrary input signals can be decomposed
into infinite sums o f sinusoids o f different frequencies, i.e., each signal has a frequency content.
This notion is made precise in a signals and systems course, where one studies Fourier series and
Fourier transforms. The frequency response o f a circuit describes the circuit behavior at each fre
quency component o f the input signal. This permits one to isolate, enhance, or reject certain fre
quency components o f an input signal and thereby isolate, enhance, or reject certain kinds of
information.
-o- -O
+
0.01 F
10
-o
FIG URE 10.25 RC circuit passing high-frequency content of an input signal.
S o l u t io n
Using voltage division, the ratio o f the output phasor voltage to the input phasor voltage N^
is given by
___
out __________________yO.Ola)
1
= m p ))
Vi„ " i + ----- ! _ “ l + iO .O lc o
/).01(0
The two universally important frequencies are O) = 0 and co= oo. At these frequencies, H{jO) =
0Z .90° and //(;“ ) = 1^-0°. Asymptotically then, the magnitude |//(/ca)| 1 as CD ^ oo and
468 Chapter 10 * Sinusoidal Steady State Analysis by Phaser Methods
Frequency (rads/sec)
(a)
Frequency (rads/sec)
(b)
FIG URE 10.26 (a) Magnitude plot of frequenc)' response for Example 10.13.
(b) Phase plot of frequency response.
Do these frequency responses make sense? They should. Going back to the circuit, observe that at
to = 0, the capacitor impedance is infinite. Physically, then, in steady state, the capacitor looks like
an open circuit for dc, i.e., at zero frequenc)'. The magnitude plot bears this out. For frequencies
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
close CO zero, the capacitor approximates an open circuit and, hence, the magnitude remains small.
O n the other hand, for large frequencies, the capacitor has a very small impedance. This means
that most o f the source voltage appears across the output resistor. The gain then approximates 1,
as indicated by the magnitude plot. The frequency response o f the circuit is such that the high-
frequency content o f the input signal is passed while the low-frequency content o f the input sig
nal is attenuated. Such circuits are commonly called high-pass circuits.
EXA M PLE 10.14. Investigate the frequency response o f the parallel /^Z,C circuit o f Figure 10.27.
L=0.04H
R=10
C=0.25F
S o l u t io n
LC RC
R jOiL . CO
w
J y'4co
C
J_ _ ^ 2 ^ ._ o L I00-co-+y4co
LC RC
Clearly, Hence the ratio o f the output phasor to the input phasor is simply
Zy^^(yoj). Once again, co = 0 and co = oo are the first two frequencies to look at. Here Zy^,(0) =
0Z .90° and -2^,„(oo) = 0Z.-90^’. Also at co = 10, the impedance is real, i.e., 0) = I . These three
points provide a rough idea o f the magnitude and phase response. Two more points are necessary
for a real .sense o f the frequency response. At what frequency or frequencies does the magnitude
drop to 0.707 o f its maximum value or when does the phase angle equal ±45°? This will occur
when 1100 - co^| = |4co|. This is a quadratic equation. Flowever, because o f the absolute values,
there are rvvo implicit quadratics, co" - 4co - 1 0 0 = 0 and co^ + 4co - 100 = 0. Solving using the
quadratic formula yields co = ±8.2, ±12.2. Since the magnitude plot is symmetric with respect to
the vertical axis (co = 0 axis), we consider only the positive values o f co. This information provides
a good idea o f the magnitude and phase plots. A computer program was used to generate the fre-
470 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
quenc)' response plots in Figure 10.28a (magnitude) and Figure 10.28b (phase). The magnitude
plot shows that frequencies satisfying 8,2 < O) < 12.2 are passed with little attenuation. Frequencies
outside this region are attenuated significantly. Such a characteristic is said to be o f the band-pass
type, and the corresponding circuit is a band-pass circuit.
Frequency (rads/sec)
(a)
Frequency (rads/sec)
(b)
F IG U R t 10.28 (a) Magnitude plot o f frequency response for band-pass circuit o f Figure 10.27. (b)
Phase plot of frequency response.
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods •n
EX A M PLE 10.1 5 . As a final example vve consider the so-called band-reject circuit o f Figure
10.29. A band-reject circuit is the opposite o f a band-pass circuit. A band-reject circuit has a band
o f frequencies that are significantly attenuated while it passes with little to no attenuation those
frequencies outside the band. In this example our goal is to compute the magnitude and phase o f
the frequenc)' response o f the band-reject circuit o f Figure 10.29.
R =10
S o l u t io n
-(0 ‘
' t)Ul LC 1 0 0 - CD"
_ L _ c o 2 + /o)-5 1 0 0 - 0 ) 2 +^25co
LC ■ L
es 0 and CO. Also at (o = 10“, //(/co) = 0 Z .-9 0 ” while at to = 10+, Hijo)) = 0 Z .-2 7 0 ° = 0Z.90". For
this example, to find the frequencies where |//(/‘to)| drops to l/ V I o f its maximum value o f 1, it
is necessary to equate the magnitudes o f the real and imaginary parts o f the denominator. This
produces two quadratics whose positive roots are to = 3 .5 0 7 8 and OJ = 28 .5 0 7 8 . At these frequen
cies the angles o f //(/w) are —45® and 45°, respectively. Fhe computer-generated plots o f Figures
10.30a and 10.30b are, o f course, consistent with these quickly computed values.
472 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
Frequency (rads/sec)
(a)
Frequency (rads/sec)
(b)
F I G U R t 10.30 (a) Magnitude plot of frcquenc)^ response for band-reject circuit of Figure 10.29. (b)
Phase plot of frequency response.
As wc can see, a wealth o f different kinds of frequency response are obtainable by different inter
connections o f resistors, inductors, and capacitors. Historically, phasor techniques were the essen
tial tool for the analysis and design of such circuits. Nowadays, engineers ordinarily use either
M A T L A B or SPICE to obtain frequency response plots. Two examples follow where we use MAT-
LA B , SPICE, or both to obtain the frequenc)’ response.
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods 4 '’3
E XA M PLE 10.16. Compute the frequency response o f the circuit o f Figure 10.31 using MAT-
LAB and SPICE.
R=10Q
S oluti on
This circuit was originally analp^ed in Example 10.9. You might want to refer to that example
before proceeding.
SPICE Part. A SPIC E simulation produces the result shown in Figure 10.32.
FIGURE 10.32 SPICE plot of capacitor voltage for the circuit of Figure 10.31.
MATLAB Part. Although the analysis appears in Example 10.9, we can use MATLAB to more
easily obtain the frequency response. First define Zj(/co) = ytoA and Z-,{p)) = l/y'coC Then from
current division,
= -7 “ " r 7 — 1 ,0 )
R+ Z|(yco)
2-. (/CO) I
•21^( 700)
R+ Z 2 (y co ) R K2( 7(0) +
474 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
Assuming a frequency range o f 0 < co < 1000 rad/sec, the following MATLAB code will result in
a suitable magnitude frequency response plot, as shown in Figure 10.33.
FIGURE 10.33 Magnitude plot of frequenc)’^response o f capacitor voltage in the circuit of Figure
10.31. The response is of the low-pass t)'pe.
Now suppose the inductor in the circuit o f Figure 10.31 is replaced by a capacitor C, = 1 mF with
the controlling current changed to /q (0- The frequency response is easily computed with a sin
gle change to the MATLAB code, namely, “Z , = 1. ./(j*w *0.001).” The resulting plot shows a
band-pass characteristic, as illustrated in Figure 10.34.
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods 47S
F IG U R E 10.34 Magnitude plot of frequency response of capacitor voltage in circuit of Figure 10.31
when inductor is replaced by a 1 mF capacitor. The response is of the band-pass type.
E X A M PLE 10.17. In Chapter 9 we investigated the Wien bridge op amp oscillator circuit,
redrawn in B2 Spice in Figure 10.35.
4" 6 Chapcer 10 • Sinusoidal Steady State Analysis by Phaser Methods
Two difTerences are notable: ( !) there is a current sourcc present across the combination,
and (2) is now 10 kQ, as opposed to 9.5 kH in Example 9.14, forcing /?, = /?2- This means
that the characteristic equation for the circuit is
1
■V*' + / ? i ' + ( • = .y~ +
{RlC R,C)
f o = - ^ = -----!— = 15.92 Hz
" 2n 2 k R,C
for any initial condition on C ,. In fact one might recall that R^ < Rj causes a growing oscillation
that is limited by the saturation effects o f the op amp.
The current source, set at 1 A, is present in Figure 10.35 so that we can obtain the frequency
response cur\'e shown in Figure 10.36. In Figure 10.36 observ'e that the magnitude response peaks
at/q, as expected from the theoretical analysis. In an actual circuit, the current source would not
be present. Nevertheless, a sustained sinusoidal oscillation will occur because o f the presence o f
noise. W ithout going into the analysis, noise contains an infinite number o f frequency compo
nents, each o f which has a minute magnitude. In particular, noise contains frequency components
around / q that drive the circuit into oscillation. This is precisely what the peak in the frequency
response means: a very small (noise) voltage on Cj will cause a very large-magnitude sinusoid out
put voltage at /q. However, the presence o f nonlinearities such as saturation keep the magnitude
at an acceptable level.
The bridge circuit presented in Figure 10.37, or some variation of this bridge circuit, has been and
continues to be a widely used approach to accurate measurement technology. In this section we
will analyze the ac bridge circuit o f Figure 10.37 as a pressure measurement device. The capaci
tance C t is a diaphragm capacitor consisting o f a hollow cylinder capped on either side by fused
quartz wafers. Bersveen the wafers is a vacuum. The capacitance o f the diaphragm changes with
temperature and pressure. For our analysis we will assume that the temperature is constant and
that the pressure is constant for a time period greater than five times the longest time constant o f
the circuit. This will allow the voltages and currents in the circuit to reach steady state and thus
allow us to use phasor analysis to compute their values.
R =100 0
15cos(20,000nt)V
F1GUR1-' 10.37 Bridge circuit diagram of pressure-sensing device. The capacitance changes as a
function of pressure, which causes the voltage to changc as a function of pressure. This is
registered on the attached voltage meter, which has a 1 MQ internal impedance.
As a rule o f thumb, the capacitance C-, » Q.llAKAId. This means that the capacitance is inverse
ly proportional to the distance d between the plates and proportional to the area A o f the plates
and to the dielectric constant K o f the material between the plates. Increasing the pressure on the
diaphragm decreases the distance rf'between the wafers, increasing the capacitance. Conversely, a
decrease in pressure will increase the distance between the wafers, thereby decreasing the capaci
tance. As the capacitance changes, the magnitude o f the ac voltage appearing across the voltage
meter will vary accordingly. Hence, two relationships are necessary: (1) the relationship berween
the capacitance C , and the magnitude o f the voltage - V e and (2) the relationship berween
the pressure applied to the diaphragm and the associated capacitance. Our first task will be to
specify the relationship between the pressure applied to the diaphragm and the resulting capaci
tance. Following this, we will use nodal phasor analysis to determine the magnitude of
and finally, the relationship between pressure and the magnitude V ^ - V e
Pressure is measured in various units. Millimeters o f mercury (mm Hg) is a common standard; 1
mm Hg = 1 torr, and 760 torr = 1 atmosphere (atm), where 1 atm is the pressure o f the earth’s
atmosphere at sea level, which supports 76 0 mm o f mercur\' in a special measuring tube. Suppose
4 ■’8 Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
it has been found experimentally that the capacitance C j (in pF) varies as a function o f pressure
according to the formuhi
C 2 ( A P ) = Q ) + ^ lo g 10
^0
(10.39)
(16i)+A P^
= 2 6 .5 + 68 log 10
760
QJ
(T3
Q.
U
Change in Pressure
Our next task is to develop the relationship between the capacitance of^ the bridge circuit and the
magnitude o f the phasor voltage In our analysis, G, = (/?,)“ ', Gj = = (^ 3)"^
and = 10~^ S is the conductance o f the meter , According to Figure 10.37, Cj = 20 pF. We
will let Cj range as 0 < C 2 < 40 pF. Finally, co = 2h x 10"^ rad/sec. The following phasor analysis
will be done symbolically so as not to obscure the methodology.
Summing the phasor currents leaving node A leads to the phasor voltage relationship
- G 2 V ^ .( G 2 .G 3 .G J V ^ - G ,,V c =0
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods 479
- > C ,V ^ .;( o ( C , . q t G JV c- =0
G] + G 2 + -G i -/■(oC, V / ■ l5 G ,'
G 2 + G 3 + G ,„ = 0
-G j ~^m (10.40)
-yco C , - G ,„ G „i + ^ ( C | + C 2 ) Vc. 0
The matrix on the left is said to be a nodal admittance matrix. Its entries can be real or complex,
as indicated. It is nor advisable to solve such a set o f equations by hand over the range o f possible
C2 values. However, using MATLAB one can solve this matrix equation over the range 0 pF < C,
< 40 pF to produce the plot o f Figure 10.39.
CO
QJ
01
nj
*->
<U
cn
-o
o
cu
■o
D
'E
cn
fO
10 15 20 25 30 35 40
C, in pF
FIGURE 10.39 Plot of the magnitude o f the phasor voltage V ^ as a function of capacitance.
O f course, one could measure the voltage appearing across the meter, from Figure 10.39 deter
mine the associated value o f Cj, refer to Figure 10.38 for AP, and then determine P = 760 + AP.
This is a long route. To complete our analysis, then, we need to develop the relationship between
pressure and bridge voltage. As we have the relationship between C j and APand the relationship
between C2 and | |, it is a matter o f using equation 10.39 to derive the value o f in
equation 10.40. This is best done with a simple MATLAB routine, which yields the plot given in
Figure 10.40.
480 Cliaptcr 10 • Sinusoidal Steady State Analysis by Phasor Methods
>
m
>
'o
01
T3
D
'c
ro
300 400 500 600 700 800 900 1000 1100 1200
Pressure in mm Hg
I'lG U RE 10.40 Relationship herween magnitude of bridge output voltage and pressure applied to
diaphragm capacitor C-,.
An actual pressure sensor would, oFcourse, be more complex. For example, there would probably
be a difterential amplifier such as the one shown in Figure 10.41 across the terminals oFthe bridge
circuit, and this would probably drive a peak (ac) detector to determine the maximum value oF
the ac signal appearing at the output oF the dlFFerential amplifier. Fiu ther, the peak value would
probably be read by a digital voltmeter. Nevertheless, our analysis illustrates the basic principles
involved in such a measurement. OF course, one could just as easily use loop analysis to solve the
problem. This is leFt as an exercise in the problems.
kR.
-o V,
Exercise. Prove that = k{v-, - for the differential amplifier o f Figure 10.41.
The two primar)' goals o f this chapter were (1) to de\'elop the phasor technique for the analysis of circuits
having a sinusoidal steady state and (2) to illustrate how this technique leads to die idea of a circuit fre-
quenc)^ response, which characterizes the circuits behavior in response to the frequency content of an input
excitation. In the development, sinusoids were first represented ;is the real part of a complex sinusoid. As
a motivation for the delineation of die phasor method, we showed how the complex sinusoids could be
urilized to compute the sinusoidal steady-state response using difierential equation circuit models. We then
pointed out that a complex (voltage or current) sinusoid is specified by a complex number or pha.sor rep
resenting its magnitude aind phase. After introducing the notions of impedance and admittance for the
capacitor, the inductor, the resistor, and a general two-terminal circuit element, we showed how the pha
sor voltage and phasor current for each such element satisf}" a frequency-dependent Ohms law. I'his
iillowed us to adapt the ;inalysis techniques and network dieorems of Chapters 1 through 6 to the steady-
state analj'sis of circuits excited by sinusoidal inputs. For example, diere are voltage division formulas, cur
rent division formulas, source transformations, and The\'enin and Norton theorems all valid for phasor
representations. This permits us to effectively analv/e circuits diat have a src*ady-state response.
The phasor technique opens a door to seeing how circuits behave in response to sinusoids. Given that
input excitations are composed o f different frequenc)' sinusoids, such as a music signal, phasor analy
sis shows why a circuit will behave differently toward the different frequencies present in the input sig
nal. This fact prompts the notion o f a circuits frequency response, which is defined as the ratio o f the
phasor output to the phasor input excitation as a function o f (u in the single-input, single-output case.
The frequency response consists of two plots. The magnitude plot shows the gain magnitude o f the
circuits response to sinusoids o f different frequencies, and the phase plot shows the phase shift the cir
cuit introduces to sinusoids o f different frequencies. The notion o f frequency response will be gener
alized in Chapter 14 afrer we introduce the notion of the Laplace transform.
Admittance: o f a two-terminal device, the ratio o f the phasor current into the device to the
1;in
phasor voltage across the device, y< ( /CO) =
in
Band-pass circuit: circuit in which frequencies within a specified band are passed while frequen
cies outside the band are attenuated.
Band-reject circuit: circuit in which one band o f frequencies is significantly attenuated while
those frequencies outside the band are passedes with little to no attenuation.
Com plex exponential forcing function: function o f the form v{t) = , where V = and
^ = a +yco are complex numbers. A special case (a = 0), f(t) = is used through
out the chapter as a shortcut for sinusoidal steady-state analysis.
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
^ ^
Conductance: real part of a possibly complex admittance.
Current division: in a parallel connection of admittances driven by a current source, the current ^
through a particular branch is proportional to the ratio of the admittance of the branch
to the total parallel admittance.
Euler identitjr: = cos(0) + j sin(0).
Frequency: in a sinusoidal function A cos((j)t + d) or B sin((Of + 0), the quantity co is the angular ^
frequency in radians per second (rad/sec). Equivalendy, A cos(o)f + 6) = A cos(2K/t + 0),
where/is the frequency in hertz (Hz or cycles per second). Note that O) = 2n f.
Frequency response: (of a circuit) graph of the ratio of the phasor output to the phasor input as
a function of frequency. It consists of two parts: (1) a graph of the magnitude of the pha
sor ratio and (2) a graph of the angle of the phasor ratio.
High-pass circuit: circuit with a frequency response such that the high-frequency content of the
input signal is passed while the low-frequency content of the input signal is attenuated.
Imaginary part: the imaginary part of a complex number z = a + Jh for real numbers a and b,
denoted by Im[z], is b.
Impedance: ordinarily complex frequency-dependent Ohms law-like relationship of a two-ter
minal device, defined as Z(/a)) = V/I, where V is the phasor voltage across the device and
I is the phasor current through the device. For the resistor, = /?; for the capacitor,
= l/(/a)Q; and for the inductor, = ycoZ.
Magnitude (modulus): the magnitude of a complex number z = a + jb , denoted by |z|, is
Phason complex number representation denoting sinusoidal signals at a fixed frequency. Bold&ce
capital letters denote phasor voltages or currents; a typical voltage phasor is V =
and a typical current phasor is I =
Polar coordinates: representation of a complex number z as p?-^, where p > 0 is the magnimde of z and
0 is the angle z makes with respea to the positive horizontal (real) axis of the complex plane.
Reactance: imaginary part of an impedance.
Real part: real part of a complex number z = a + jb for real numbers a and b, denoted by Re[«], is a.
Rectangular coordinates: representation of a complex number z as coordinates in the complex
plane, i.e., zs a + jb for real numbers a and b.
Resistance: real part of a possibly complex impedance.
Sinusoidal steady-state response: response of a circuit to a sinusoidal excitation after all transient
behavior has died out. This definition presumes that the zero-input response of the cir
cuit contains only terms that have an exponential decay.
Stable circuit: circuit such that any zero-input response consists of decaying exponentials or expo
nentially decaying sinusoids.
Susceptance: imaginary part of an admittance.
Voltage division: in a series conneaion of impedances driven by a voltage source, the voltage
appearing across any one of the impedances is proportional to the ratio of the particular
impedance to the total impedance of the connection.
Zero-input response: response of the circuit when all source excitations are set to zero.
’ In the literature, both z and z* are used to denote the conjugate o f a complex number z. However, in matrix
arithmetic, Z* usually means the conjugate transpose o f the matrix Z. We will sometimes interchange the usage.
In MATLAB, * means multiplication and conj(Z) means conjugated. So there is some ambiguity in the usage.
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods 483
Problems
SO LUTIO N OF DIFFEREN TIAL O '
EQUATIONS W ITH CO M PLEX
EXPO N EN TIALS Figure P I0.3
A N SW ER S: 0 .8 sin (2500r - 3 6 .8 6 ‘>). 80
1. Construct the difFerential equation model o f s in (2 5 0 0 r-3 6 .8 6 » )
the series RL circuit o f Figure PI 0.1 in which L
= 0.25 H and R= 100 Q. Then use the method 4. Construct the differential equation model of
o f section 4 to compute the steady-state the parallel /?Z.Ccircuit o f Figure PI 0.4 for R =
response when v’,„ (/) = 20V 2 cos(400/) V. 100 C = 1 pF, and Z. = 40 mH. Then use
the method o f section 4 to find the steady-state
response when / (f) = 20 cos(2500f) mA.
vJt) R < v .J t )
Figure P I0.1
AN SW TR: 20 cos(400/ - jr/4) V Figure P I0.4
AN SW ER: 1.6 cos(2500; + 36.8~») V
2. Find the differential equation model o f the
series RC circuit o f Figure P I0.2 in terms o f
and V(it) assuming that C = 5 pF and R KCL AND KVL W ITH
= 800 Q. Write as a function o f
PHASORS
and V(^t). Then use the method o f section 4 to
5. Find the phasor current I and /(/) for each
determine the steady-state response when
circuit o f Figure Pi 0.5 when (O = IOOti rad/sec.
iV„(/) = 20>/2sin(250/) V.
+ v,(t) -
(2+j4)A
+
v„(t) R
V „(t)
j20
(26+j12)V
0 20
(a)
Figure P i0.6
e
C H E C K : /,(/) = 25 cos(100r + 0.9273) A j20
(26+ jl2)V L
20
7. Suppose that in Figure P I 0.7, v^{t) = 4
cos(o)f) V and ''2^^ ^~ 4-s/2 cos(o)/ —0.25ti ) V.
Find V[{t) = A"cos(ior+ (J)). (b)
+ v^(t) -
e
j20
v,(t) Vj(t)
(26 + jl2)V
0 20
Figure P I0.7 (0
ANSW ER; vjU) = 4 cos((.»/ - ‘)0") \'
Figure P I0.9
8. Use KVL to determine the phasor voltage
in the circuit o f Figure PI 0.8. 10. For the circuit o f Figure P i 0.10, use KCL
and KVL to find the phasor voltage V^. and the
4jV 8jV
phasor current I^,. If the frequency co = 2 0 0 0 ;:
rad/sec, find the associated voltage and current
time functions.
- V -
V _y (2+ j
(2+j4)A
(4 + ]4)V
I / Ti (4 + j6) A 1
Figure P i0.12
fi.(t)
v,(t)
o
Figure P I0.13 Figure P I0.16
AN SW ER: (b) 1.2Z-9()'\ 1.2 sin(200r) A
14. Find the phasor currents and
and then determine i j p ) for the circuit o f 17. In the circuit o f Figure P 10.17, /? = 6 f i, Z.
Figure P I 0.14 in which /? = 1 k li, = 0.5 H, = 80 mH, C = 0.5 mF, Vp{t) = 8 cos(200r) V,
C = 1 pF, and = 20 cos(1000r + 60°) V. and I, = 0.5Z 90® A. Find the source voltage,
which operates at the same frequency o f
200 rad/sec.
■m Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
V jt)
Z Jj(o )
Figure P I0.17
ANSWHR: S.831 1 cos(200r +30.96<’) V Figure P 10.20 Parallel Z.C circuit.
ANSWl-.RS: (a) -/1.25 12, (b) 1.6 ml-
18. The circuit o f Figure PI 0.18 operates in the
sinusoidal steady state at a frequency o f cOq = 21. Consider the circuit o f Figure P I0.21.
2000 rad/sec, /?, = 7?^ = 10 Q, V-„ = 50 V, and (a) Find the impedance at (o = 100 rad/sec.
= 2Z. - 53.13® A. Compute the phasor volt (b) What happens to the impedance as to
age across Rj and then find the impedance gets large?
Now construct a simple series circuit (c) Ifv/„(/) = IOV 2 cosdOOOV, find /.(/).
that represents this impedance at cOq.
ijt )
1 mF
I
v jt)
6 lOO
0.1 H
Figure P I0.18
ANSWHR: Z = 2.5 +ylO Q Figure P I0.21
A N SW FR: (a) S - p U
19. (a) Find the steady-state response o f the
circuit o f Problem 3 using the phasor 22. For the circuit o f Figure PI 0.22, suppose R
method. Discuss the relative advan = 100 Q, ^ = 0.5 H, C = 5 ^iF.
tages o f the phasor method. (a) If = 0.1 cos(500/) A, find v^it).
(b) Find the steady-state response o f the (b) Find (O ^ 0 in rad/sec so that the input
circuit o f Problem 14 using the phasor admittance is real.
method. Discuss the relative advan
tages o f the phasor method.
+
V ,( t)
SERIES-PARALLEL
IM PED A N CE AND
A D M ITTA N CE
CA LCU LA TIO N S
20 . Consider the circuit o f Figure P I 0.20. Figure P i0.22
(a) If C = 0.01 F, findZ.„(/100).
(b) If Zy,^(/100) = 25j O., find the appro 23. Consider the circuit o f Figure P I0.23 in
priate value o f C. which /?j = 20 /?2 = 10 Z, = 20 mH.
(c) Using the impedance o f part (b), if (a) I f C = 0.3 mF, find K.„(;-500).
ij„U) = 100 cos(100r + 45®) mA, find (b) Find the value o f C that makes the input
v^t). admittance real at OJ = 500 rad/sec.
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods 487
(c) If C = 0.3 mF and /.„(f) = 100 zero. Determine the minimum value
cos(500r) niA, find /^](^) and /^-(r) o flK „(/ 0 ))l.
using current division.
/Y Y V
L
(a)
Figure P I0.23
(b)
A N SW ERS: (a) 0.1 + yO.l S, (b) 0.1 mF
Figure P I0.25
24. Consider the circuit o f Figure P I 0.24 in
which C = 1 F. At C0= 2 rad/sec, = 4 + jl 26. For a particular two-terminal device,
n. = 0.002 + y0.002 at O) = 500 rad/sec.
(a) Find the appropriate values o f R and Construct a parallel RC circuit having this
L admittance at (0 = 500 rad/sec. If the circuit is
(b) For 0 < Z < 0.2 H, specify the range o f excited by a currcnt source with ij^t) = 10
possible reactance values for cos(500f) niA, find the voltage appearing across
(c) If = 10 cos(2/) V, find Vj^i). the current source.
A N SW ER : C = 4 uF, A* = 500 Q ..
5
\'^(/) = —;= cos 500/ - 4 5 " ) V
R V2
27. For a particular two-terminal device,
'jt ) Q
Z .„(;1000) = 2000 + ;2 0 0 0 Q.. Construct a
series RL circuit having this impedance at CO =
Z,„(jw) 1000 rad/sec. If the circuit is excited by a volt
age source with v^{t) = 10 cos(1000/) V, find the
Figure P I0.24 current through the resistor.
AN SW ERS: 4 Q, 0.2 H, 0 to x AN SW ER: Z. = 2 H, /^ = 2 kQ.,
5
/(/) = - = cos 1000/ - 4 5 " ) mA
25. Consider the circuits o f Figure P I 0.25 in
which R = 5 O., L = 32 mH, and C = 5 ^F.
(a) Consider Figure P 10.25a. Find 28. The circuit o f Figure P i 0.28 operates in the
Zy^^(/(0) as a function o f (0. Then com sinusoidal steady state at the frequency' (0 =
pute the frequency at which Zy^^(/CO) is 5000 rad/sec with R =4 Q., L = 0.4 mH, and C
purely real, i.e., the reactance is zero. = 0.1 mF. Find Z-^^{j5000) and ^/„(y5000).
Determine the minimum value o f Construct a simple series circuit that is equiva
lent to this circuit at 0) = 5000 rad/sec. Finallv,
(b) For the circuit o f Figure P 10.25b, find construct a simple parallel circuit that is equiv
Ky^;(/0)) as a function o f (0. Then com alent to this circuit at 0) = 5000 rad/sec. In both
pute the frequency at which is cases specily the element values.
purely real, i.e., the susceptance is
488 Chapter 10 * Sinusoidal Stoatly State Analysis by Phasor Mcthoils
Figure IM0.28
C H EC K : Z.„(/‘5000) =1 + ;2 ^ -H f
R
SERIES/PARALLEL
^,(t)
IMPEDANCES WITH V/l
DIVISION
30. Consider the circuit o f Figure P I0.30 in
Figure P I0.32
which /? = 20 Q, Z. = 4 H, and
( 0 = 10>/2 cos(5/) niA. ANSW ER: (a) r(^(/) = ^ c o s -4 5 ‘
v2 RC
(a) Find the input impedance ■^,„(/w) and
33. Consider the circuit o f Figure P I 0.33 in
the input admittance
which y? = 8 Z, = 8 mFi, and C = 0.125 mF.
(b) At OJ = 5 rad/sec determine the steady-
(a) Determine the values o f the phasors
state current ;’^(r).
I^, and V(^ when = 2 A and (d =
i,(t) 1000 rad/sec. Specify the correspon
ding time functions.
© (b) Repeat part (a) for o) = 500 rad/sec.
Figure P I0.30
y20co 0.25
ANSW'F'.KS: (a) . 0.05 - /■
5+ /(I) (I)
31. (a) For the circuit o f Figure P I0.31, find 34. In Figure P I0.33, suppose R = 500 Q, I. =
the ratio in terms of/^, C, and 0 .5 mH, C = 0 .1 2 5 mF, and
CO. Express the answer in polar form. i:,Ar) = loV2co.s(co/+ 60^) A.
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods •hS9
(a) Compute the values o f the phasors I^ Using the phasor method, find when
I^, and when = 2 A and OJ = v j t ) = 50 cos(4000^) mV.
4000 rad/sec. Specify the correspon
ding time functions. He 100v„,(t)
(b) Repeat part (a) for (O = 8000 rad/sec.
v„.(t)
''" 6
and C = 5 pF. Suppose = 120 cos(400r)
mA. Find and
Figure P I0.37
AN SW TR: 0.625 mF or 0.2083 mF
2Q
> j2 Q -j2 0
NETWORK THEOREMS IN
CONJUNCTION WITH V/l
Figure P I0.44
DIVISION. AN SW ERS; t'fit) = 2 cos(lOOOr) V, i/(/) = cos
(You should consider applying one or more (1000/-O.S3T) A
network theorems to simplify the solution to
the problems in this section.)
45. In the circuit of Figure Pi 0.45, R=20Q., L = 20
mH, C= 100 pF, and v^.(0 = 2 0V 2 cos(IOOO/)V.
42. In the circuit o f Figure P 10.42, R^ = 60
Compute the value of in steady state.
/?2 = 40 n , and C = 0.1 mF. Find the phasor Ij
and the corresponding steady-state current /j(/) /Y Y V
L
when /^.(/) = 5>/2cos(100/) mA. This prob
lem can be solved by direct current division or
by source transformation and impedance con
6
cepts. Which method is easier?
Figure P i0.45
A N SW ER: 20 co s(1 0 ()0 r- 135°) V
i(t)
0 46. Consider the linear circuit o f Figure
PI 0.46, which operates at 50 Hz and for which
V, = + 1,1^2-
(a) Find the values o f a and b.
Figure P i0.42 (b) If v^^{t) = 10 cos(lOOTCf) V and =
200 sin(lOO)/^) mA, find v^it).
43. In the circuit o f Figure P I 0.43, CO = 400
rad/sec, = Rj = 2 L = 5 mH, and C =
625 |.iF. Find I^^and the corresponding in
steady state. vp)= 12 cos(400^) V j5on
■ -©
J200Q
+ V. -
Figure P i0.46
AN SW ERS (in random order);
Fiaure P I 0.43 - 0 .8 . ./40 Q. - 1 6 cos(IO().-t O V
Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods 491
47. The linear circuit o f Figure P I 0.47 is such can do this with some straightforward
that in the steady state, if = 10 cos(2007U/) reasoning without writing any equa
A with v^2 ^t) = 0, then \t) = 20 cos(2007ir + tions.
45®) V. On the other hand if i^^{t) = 0 with
/;2 (/) = 10 cos(2007if + 45°) V, then v^{t) = 5
cos(200)r + 900) V
(a) Find a linear relationship between ,
v^2 >and .
(b) If"/^,(/) = 5 cos(2007tr - 45«) A and
= 20 cos(2007rr) V, then in the
steady state find v^{t).
(c) Find Zj and Z - ,. Develop simple cir Figure P I0.48 AC Wheatstone bridge circuit.
cuit realizations o f these impedances
valid at OJ = 20071 rad/sec.
THEVENIN AND NORTON
EQUIVALENTS
49. Find theThevenin equivalent for the circuit
o f Figure PI 0.49 when R = AQ. L = 20 niH, C
= 1.25 mF, = 2Z.45® A and (O = 200 rad/sec.
Figure P I0.47 Be sure to express the open-circuit voltage as a
ANSWHRS: (b) r,(/) = 18.46 cos(200nr - time function.
22.5") V; (c) Z, = 0.763 + y2.6()5 H Z, = 4 Q OA
OA
Figure P I0.54
-OA
e R,
Figure P I0.55
■OB
GENERAL SSS ANALYSIS
Figure P I0.52 Two coupled R C circuits.
(NODE OR LOOP ANALYSIS)
56. (a) Find the phasors in the cir
53. For the circuit o f Figure P I 0.53 R = 2.5
cuit o f Figure Pi 0.56 when
k n . Find the Thevenin equivalent when =
= 2 0 V 2 Z 4 5 " V, R = AQ., L = A
10 cos(4000r) mA.
mH, and oj = 1000 rad/sec. Specify
the corresponding time functions.
(b) Determine the value o f oj for which
the magnitude o f the output voltage
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods 493
'“6 -OA
2R
2R
;v,(t)
Figure P i0.56
C H E C K S: 5 A, 2o’ v -O B
Y.(j«)
57. In the circuit o f Figure P I0.57, = y'30 Q,
= - ; 4 0 Q, V^, = 28 V, V^2 = ^ Figure P I0.59
sinusoidal sources have been operating for a
long time, and Z = 50 - j 4 0 Q. Find V^. 60. Consider the circuit o f Figure P i 0.60 in
which v(,t) = 20 cos(lOOO^) V, = 40 Q., Rj =
20 Q.,L = 20 m H .an d C = 7 5 \i¥.
(a) Write and solve a nodal equation at
the top node for Then write the
6 ''“ corresponding time-domain expres
sion for v^t).
(b) Calculate and then write the corre
Figure P I0.57
sponding time-domain expression for
the inductor current
58. For the network o f Figure P i 0.58, a = 20
(c) Find the Flievenin equivalent circuit at
a z^ = y io a z^^ = - y i o a ^ = lo a and
the source frequenc)' relative to terminals
= 20 V. Find the phasor current I^..
A and B. Draw the Thevenin equivalent
Z, circuit showing the Thevenin impedance
as a series circuit of two elements.
Figure P I0.58
AN SW FR: O.r. + 0.2/
Figure P I0.61
C H EC K : = ;1 0 0 0 Q, = 25 + ;2 5 V, '■©
= 25 + y25 mA
ijt)
66. Inside the black box o f Figure P I0 .6 6 a
R there is a two-element circuit composed o f a
v,(t)
6 resistor o f 10 ^2, capacitors, inductors, or some
combination o f these elements. A variable-fre-
quency voltage vp) = lOcos(O)r) V is applied to
Figure P i0.63
the box and the voltage v(t) = cos(cor + 0) is
observed. A plot o f the magnitude o f v{t) with
FREQUENCY RESPONSE
respect to CO is given in Figure P 10.66b.
64. Compute the magnitude and phase func
(a) Draw the circuit contained inside the
tions o f the frequency response o f the circuit o f
box. (There are two solutions.)
Figure P I 0.64 in which L = 4 mH and C = 0.25
(b) Specify the element values.
mF. Plot your response in MATLAB (0 < CO <
5000 rad/sec). Before sketching the responses,
Chapter 10 • Sinusoidal Steady State Analysis by Phaser Methods
v,(t) =
Black +
10cos(cot)V v(t)
Box
(a)
(b)
67. Compute the frequency response o f the cir write a set o f mesh equations that describe the
cuit o f Figure PI 0.67, where R= 100^2, L = 10 circuit. Solve the equations for 1 pF < € 2 ^
mH, C = 0.1 mF, and is the output. Use pF using MATLAB or some other, equivalent
MATLAB or its equivalent to generate the software program. Plot the magnitude o f V ^ -
magnitude and phase (in degrees) plots. V ^as a function o f Cj- Now construct a plot o f
OP AMP CIRCUITS
70. (a) Compute when = sin(200r)
mV for the circuit of Figure P10.70a.
Figure P I0.67
(b) For the circuit o f Figure PI 0.70b, find
C so that when = cos(400r) mV,
68. The box labeled V{joi) in Figure P I0 .6 8 a
= sin(400r) mV.
contains a single resistor, a single capacitor, and
(c) Find the phasor transfer function,
a single inductor. Let v^{f) be the input excita
//(yco), and plot the magnitude o f the
tion and ip ) the circuit response. The magni
frequency response (using iMATL^B
tude frequency response is given by Figure
or the equivalent) as a function o f 03 =
P I0 .6 8 b . Draw^ the circuit inside the box and
Inf, where/is in Hz and to in rad/sec.
assign component values if it is known that L =
40 mH.
496 Chapter 10 • Sinusoidal Steady State Analysis by Phasor Methods
r>
20 kO 100 kfi 80 kO
(a) (b)
Figure P I0.70 Op amp differentiation circuits.
(a)
Compute when = n
71. (a)
sin(400/) V for the circuit of Figure
P10.71a. n
(b) For the circuit of Figure P I 0.7 lb, find
C such that when =sin(500/) V, o
= 5 cos(500/) V. This represents
an integration of the input with gain.
(c) Find the phasor transfer function,
and plot the magnitude of the
frequency response (using M ATLAB (b) n
O ' ^
function as a function o f COusing M ATLAB or frequency increases to infinity? What
w the equivalent, assuming /? = 10 k fl and C = happens as the frequency decreases to
0.01 mE zeroi^
VO
o Figure P I0.75
Figure P I0.77 Op amp circuit having a band
o 76. For the operational amplifier circuit of pass type of response.
Figure P I0.76, /?j = 5 kQ, C, = 0.02 |iF, /?2 =
5 kfi, and C2 = 0.08 pF. Note that is an intermediary variable useful
o
(a) Write two node equations and solve to in the nodal analysis of the circuit.
find a relationship between the output
o
phasor and the input phasor
at the frequency/= 1000 Hz. Note
o that the voltage from the minus termi
nal of the op amp to ground is
o which equals the voltage from the plus
terminal to ground, assuming the op
o amp is ideal.
(b) Repeat the calculation at/= 100 Hz
and/= 3000 Hz. What happens as the
o
w
C H A P T E R
The AM or FM receiver that is often part o f a home stereo system receives signals from radio sta
tions through an attached antenna. The intensity o f these signals or radio waves depends on the
power radiated into the atmosphere by the broadcasting station, the distance between the receiv
ing and transmitting antennas, and the design o f the receiving antenna. The intensity or magni
tude o f the signals picked up by the receiving antenna is very small. The power available from the
antenna and deliverable to the receiver is typically in the microwatt range. Again, this is ver)^ small.
Hence, it is important to have maximum power transfer from the antenna to the receiver input so
that the music signals received can be properly amplified and enjoyed. Since the signals in the
antenna are sinusoidal at very high frequencies, the antenna is represented by a phasor Thevenin
equivalent circuit as is the input circuit o f the receiver. Hence we must understand maximum
power transfer in the context o f sinusoidal steady-state analysis to describe and analyze this prob
lem. An example at the end o f the chapter illustrates some impedance matching techniques to
achieve maximum power transfer from an antenna to a receiver.
CHAPTER OBJECTIVES
SECTION HEADINGS
1. Introduction
2. Instantaneous and Average Powers
3. Effective Value o f a Signal and Average Power
4. Com plex Power and Its Com ponents: Average, Reactive, and Apparent Powers
5. Conservation o f Com plex Power in the Sinusoidal Steady State
6. Power Factor and Power Factor C orrection
7. M aximum Power Transfer in the Sinusoidal Steady State
8. Summar}^
9. Terms and Concepts
10. Problems
1. INTRODUCTION
Chapter 1 defined the concept o f power. The following chapters were primarily devoted to the cal
culation o f voltages and currents. This does not mean that the consideration of power is o f sec
ondary importance. The very opposite is true. A homeowner pays for the energ}' used, not for volt
age and current. The integral o f power over, say, a 30-day period determines the household ener
gy consumed in a month. Hidden in the homeowners cost is an adjustment to cover the power
losses incurred in transmitting energy from the generating station to the home. Thus power con
siderations have a significant impact on everyday life.
A second reason for understanding ac power usage is safety. Each appliance, and its cord that plugs
into the wall outlet, has a maximum safe power-handling capacit)'. Misunderstanding such infor
mation and/or misusing an appliance can lead to equipment breakdown, fire, or some other life-
threatening accident.
Even for electronic equipment in which power consumption is low, such as laptops and handheld
PDAs, power consumption and, thus, battery life are important design factors. Power drainage direct
ly determines the PDAs operating time before the battery needs recharging. In fact, optimizing power
management in laptops and hybrid electric vehicles is an important research area in todays world.
In this chapter we will investigate different notions o f power in ac circuits and discuss their sig
nificance and application. The term “ac circuits” has a narrow meaning here. It refers to linear cir
cuits having all sinusoidal sources at the same frequency and consideration o f responses only in
steady state. The basic analysis tool is the phasor method o f Chapter 10.
Figure 11.1 shows an arbitrary two-terminal circuit element isolated from a larger circuit. With
the voltage (in V) and current (in A) having indicated reference directions, the instantaneous
power (in watts) absorbed by the element is given by equation 11.1:
p{t) = v{t)i{t) (11.1)
Chapter 11 • Sinusoidal State State Power Calculations 501
Evaluating battery life or the length o f operation o f your cell phone involves consideration o f a
quantit)' called average power, P (or for emphasis), defined as the average value o f the instan
taneous power over an interval [T*!, T^. The idea is based on the average value o f a function, say
j{t) , which is defined as
T-,
1
Le =
T
h - TM r,J
Using this idea we define the average power consumed by a rwo-terminal element as shown in
Figure 11.1 over the interval [T j, Tj] as
T,
1
Pave(TiJ2) = piOclt
r .-r , ( 11. 2)
When the signal is periodic with period T, we speak o f the average power consumed by an ele
ment over the period T as
T T
(11.3)
It is not necessary that T b e the fundamental period; the evaluation o f the integral is the same for
any integer multiple o f the fundamental period.
EXA M PLE 11.1. Compute the average power absorbed by the resistor R connected to an inde
pendent voltage source as shown in Figure 1 1.2b with the excitation shown in Figure 1 1.2a.
v.(t) f +
So lu t io n
Step 1. Compute the instantaneous power for 0 < f < 7q. Here
IV J
P {t) = 0 < / < 0 .5 ^
o .5 r „ < / < 7 ;
Step 2. Compute Using equation 11.3 and observing that the Fundamental period is Tq, we
have
0.57n
/
‘ >ave rj. p {t)d t = —
TnR 6R
Exercises. 1. Suppose the sawtooth in Figure 11.2a does not drop to zero at r = 0.5 T'q , but rather
continues to increase until reaching ^ = T’q when it drops to zero and repeats. Find the average
power consumed by R.
AN SW ER:
3/^
. K?
2. Show that the average power absorbed by an R Q resistor in parallel with a Vq V dc source is—
over any time interval [T'j, 7'^]. ^
O f particular importance is the average power consumed by devices in the SSS assuming all exci
tations are at the same frequency, to. Consequently, all voltages and currents are sinusoids at the
sa7nefrequency. To compute the average power absorbed by a circuit element as depicted in Figure
11.1 (assuming a linear circuit), suppose v{t) = cos(u)t + 9^^ and i(t) = cos(a)r + 0^) . The
associated instantaneous power is
(11.4)
= cosO , - e,-) + cos( 2 (o/+ e,, + e,.)
Equation 11.4 follows from the trigonometric identit)' cos(x) cos(y) = 0.5 c o s (x - y) + 0.5 cos(a' +
y). Observe that the instantaneous power o f equation 11.4 consists o f a constant term plus anoth
er component varying with time at tivice the input frequency. Figure 11.3 shows typical plots o f
/>(/), v{t), and i{t).
Chapter 11 • Sinusoidal State State Power (lalculations 503
FIGURE 11.3 Plots o f tit) = 10 cos(377f) A, v{t) = 2 cos(377/ + 45°) V, and p{t).
Using equation 11.3 witii T = 23t/ol), and observing that the integral o f a sinusoid over any peri
od is zero, we obtain the following formula for average power in SSS:
T
*Pnvp = —
_ co s(0 , - 0 ; )dt + ^ j c o s ( 2 c o r + 0 , + 0,-)d! = c o s(0 , - 0 , ) (11 -5)
If the two-terminal element is a resistance R, then v{t) = Ri{t) and 0^^ - 0y = 0 . It follows from
equation 11.5 that for a resistor
Before closing this section, we need to investigate the question o f superposition o f average powder.
Is there a principle o f superposition o f average power? If so, when is it valid? When is it not valid?
The following example provides the answers.
504 Chapter 11 • Sinusoidal State State Power Calculations
EXA M PLE 11.2. Consider the circuit o f Figure 11.4, which consists o f a series connection o f two
(sinusoidal) voltage sourccs in parallel with a 1 Q resistor. For this investigation v^{t) =
cos(ojjr + 0|) (having fundamental period /'] = 2:t/tO|) and v-,(f) = cos(t02^+ B2) (having fun
damental period T-, = 271/(0,). For simplicit)^ we assume that v^{t) and v-,{t) have a common peri
od o f 7 'seconds, i.e., there exist Integers ni and ;/ such that T = uT^ = niT^.
v,(t) Q
1n < v_(t)
So lu t io n
Compute the average power consumed by the 1 Q resistor. First observe that the power consumed
by the 1 Q resistor with source 1 acting alone, i.e., v-j{t) = 0, is
1 1
0 0
Also note that the power consumed by the 1 H resistor with source 2 acting alone, i.e., u^{t) = 0,
is
I T , r ^
0 0
W ith both sources active, linearit)^ (or KVL) implies that = v^{t) + . By equation 11.3,
1 1
Pave = - \ V R {t)i,i{t)d t = - \ (^ ’ 1 (/) + v'2 (r ))“ dr
0 0
T
^ove.l ^a\r,2 ' ^ Vj (/)V2(/)f/f
0
When the integral term in this last equation is zero, then , indicating that
superposition o f average power holds. When this integral term is nonzero, superposition o f aver
age power does not hold. The next question is, under what circumstances is the integral zero and
nonzero.^ There are three cases to consider. Case 1 is when cuj co^ , which will result in a zero
value o f the integral. In this case, the integral consists o f rwo sinusoids integrated over a common
period T. The integral of a sinusoid over any period is zero. Thus, the integral is zero and super
position o f power holds when cOj ^ O)-,.
Case 2 is when C0 | = co-, but with (Bj - O2) = ±knl2, k an odd integer. In this case, the integral is
again 0 . This follows because the first term o f the integrand is a sinusoid whose integral is zero
over the period T. The second term o f the integrand is a constant, cos(0j - (),) = cos{±kKl2) = 0 ,
also resulting in a zero integral. Hence for case 2, superposition o f power holds.
Finally, we have case 3, for which tOj = co-, but with (6 j - 62 ) ^ ±kKll, k an odd integer; here
superposition ot power does not hold. The second term of the integrand is a constant, cos(0j - 62 )
0 , resulting in a nonzero integral over the period T. So P^^^^ j + For case 3, it is
desirable to use the phasor method o f Chapter 10 to compute the desired voltage and then use
equation 11.5 to compute average power.
Exercises. 1. In Example 11.2, suppose t/j(r) = 3cos(107rr) V and V2 {t) = 4 cos(15n:r + 0.25ti) V.
Compute T, a common period for the two sinusoids, and then compute the average power con
sumed by the 1 Q resistor.
C H EC K ; T = 0.4 sec w-ill work, and = 12.5 watts
Equation 11.6 resembles equation 1.18b for the dc power absorbed by a resistor connected to a
dc source. However, in equation 11.6 the factor 1/2 is present. With the introduction o f a new
concept called the efifective value o f a periodic waveform, the formulas for the average power
absorbed by a resistor can be made the same for dc, sinusoidal, or any other periodic input wave
forms.
S06 Chapter 11 • Sinusoidal State State Power Calculations
From section 2, a resistor o f R ohms excited by a periodic voltage or current absorbs an average
power, . T he effective value o f any periodic current, i{t), denoted by is a positive con
stant such that a dc current o f value /^-exciting the resistor causes the same amount o f average
power to be absorbed, i.e., The same holds for a resistor excited by a periodic volt
age v{t). Mathematically,
(11.7a)
or
(11.7b)
R
Hence, the mathematical definition o f the effective value of a periodic current i{t) is
/m+7
U>ff.R (11.8a)
T
h\+T
ef/M (11.8b)
In + T
F e jf- (11.8c)
Observe that the expressions under the radical sign in equations 11.8 constitute the average value
ot the square o f the signal. Hence, the expressions give rise to the alternative name for the effec
tive value, the root-m ean-square (abbreviated rms) value oiJ{t), since the square rooto'i the
mean value o f the square j{t) over one period.
Chapter 11 • Sinusoidal State State Power Calculations SO'
Exercises. 1. Show that the average power absorbed by an R Q resistor carrying a periodic current
5
ANSWlvR: = 1 2 .5 watts, and l^,g- = = 2.5\/2 A
7?
EXA M PLE 11.3. Compute the efFective value o f the periodic voltage waveform sketched in
Figure 11.5.
So lu t io n
= T 4J 4J 4J 3
Therefore, = 2.3094 V,
Exercise. Repeat the calculation o f Example 11.3 for the case where the values on the vertical axis
o f Figure 11.5 are doubled.
A N SW l'R : V
For a sinusoidal signal/r) = cos(tor + 0) , the effective value can be calculated using the identit}'
Since by assumption to 7^ 0, the average value o f the cosine term is zero. The average value o f the
first (constant) term is itself Hence, by equation 11.8c,
F~
I itt Eml (11.9)
72
Thus, for a sinusoidal waveform, the effective or rms value is always 0 .707 times the maximum
value or, equivalently, the ma.ximum value divided by >/2— a basic fact well worth remembering.
The ac voltage and current ratings o f all electrical equipment, as given on the identification plate,
are rms values unless explicitly stated otherwise. For example, the household ac voltage is 110 V,
with a maximum voltage o f 1 lOx-s/^ = 1 5 6 V. A typical appliance such as a coffee maker will
have a 110 V rating, ac, at say, 900 watts. The effective values o f a few other periodic waveforms
are listed in Figure 11.6, with their derivations assigned as exercises.
Feff
«=F r
dc
sinusoidal
triangular
square
F.»= F .
Exercises. 1. Derive the formula for riie cfFectivc value o f a triangular waveform shown in Figure 11.6.
AN SW FR:
We recurn now to the case of single-frequency SSS analysis. The average power as per equation 11.5
absorbed by an arbitrary rwo-terminal element mav now be rewritten in terms o f effective values:
where 0^ = 0,^ - 0 y , V= 0.707 , and / = 0.7071^^ . The angle 0^ is the angle o f the impedance
Z (/co) o f the two-terminal element and is also interpreted as the angle by which the voltage pha-
sor leads the current phasor.
EXA M PLE 11.4. Figure 11.7 shows two t)'pes o f household loads connected in parallel to a 110
V, 60 Hz source, = 1 1 0 ^ 2 cos(120:tr) V. Lainp 1 and lamp 2 have effective hot resistances
o f 202 and 121 respectively. The impedance o f the fluorescent light is Zjfjwi) = 60 + j70
(a) Find the average power consumed by each light.
(b) Find the average power delivered by the source.
Lamp 1 Lamp 2
FIGURK 1 1.7 An example of load current calculation.
So lu t io n
(a) For lamp 1, Z|(/to) = 202Z.0® Cl. Hence, Ij = Vy^Zj = 0.5446Z .0° A. From equation 11.11,
Similarly, for lamp 2, Z^C/co) = 121Z.0° Hence, 1, = ^ iJZ j = 0.909 IZ.O^’ A. From equation
11. 11,
Finally, for the fluorescent light, = 56 + y'66 = 86.56Z.49.7® O.. Hence, ^ jJZ jj =
1.2 7 Z .-4 9 .7 ° A. From equation 11.11,
(b) For this part we first compute and then apply equation 11.11 to compute the average power
delivered by the source. Here by KCL,
= I, + I 2 + I 3 = 0 .5 4 4 6 Z 0 " + 0 .9 0 9 1 ^ 0 ° + 1 .2 7 ^ -4 9 .7 °
= 2.2759 -y 0 .9 6 9 0 = 2 .4 7 3 6 -^ -2 3 .0 6 ° A
Observe that the sum o f the individual average powers is 250.3 watts, which equals the power
delivered by the source within the accurac)' o f our calculations, where we have rounded our
answers.
Recall the notion of a phasor. When all source excitations are sinusoidal at the same frequency,
voltages and currents in the SSS can be represented by phasors. Our question here is, can the pha
sor method aid the computation o f power consumption in a circuit? The answer is yes. However,
the formulation will bring out several other concepts o f power associated with the sinusoidal
steady state.
In dc power calculations, the average power consumed by a two-terminal device is the product o f
the voltage and current, assuming the passive sign convention. In SSS, the complex power
absorbed by a two-terminal device, as shown in Figure 11.8, is a complex number defined by the
formula
(1 1 .1 2 )
eff
o - Two
+
Terminal
Veff
Device
o-
FIGURE 11.8 Two-terminal device with phasor voltage and current consistent
witli passive sign convention.
The first useful result o f this definition is that = Re S = Re . To see this result,
i(t) = y/l cos((or + 6,-) , which is represented by the phasor The average power consumed by
a rwo-terminal device excited by this voltage-current pair (Figure 11.8) is given by equation 11.10 as
S = \ //C /f = j = h r
in which case
R e[S]= R e = Re
The curious reader may ask why a conjugate o f the current is used in the definition o f complex
power. Suppose one did not have the conjugate o f the current. Then
Re = VeffI(,ffCOs{Q^,+Qj)^I\,y^, i.e., the resulting product would have no physical
meaning. Now because S is a complex number, it has an imaginary part, a magnitude, and an
angle. T he imaginary part o f S defines a quantity called the reactive power absorbed by the two-
terminal device in Figure 1.18; i.e., reactivepoioer \s defined as
The unit o f reactive power, Q, is VAR, which stands for volt-amp-reactive. It follows immediately
that
where P = . Also, the magnitude o f S is defined as the apparent power absorbed by the two-
terminal device o f Figure 1.18, i.e., ________
\^\ = Kff^eff = \P ^ JQ\ = 4 p ~ ^Q~ = apparent poNver
^12 Chapter 11 • Sinusoidal State State Power Calculations
The Linir o f apparenr power is VA, short For volt-arnp. The iinerrclationship o f these different pow
ers is illustrated by the right triangle diagram in Figure 11.9, which is often helpful in solving
problems. Observe that the apparent power is always greater than or equal to the average poiuer,
with equality applying to the case o f a purely resistive load.
FIGURE 11.9 Relationships among complex, average, reactive, and apparent powers.
I'h e distinction among these various powers is best understood by computing the powers for some
basic circuit elements. For simplicit)', except when needed or for emphasis, from this point on we
will often drop the subscripts <^’and rfw'as given in equations 1 1 .1 3 -1 1 .1 6 .
EXA M PLE 11.5. This example explores the computation o f the various powers for a simple
inductor. Given that 1^0 ) = v2/sin(CO/) in the circuit of Figure 11.10, compute V^, S^,
Ql, the instantaneous absorbed power />/(/), and the instantaneous stored energ)' Wjit) in terms
o f L, CO, /^, and Vj. After this show that
(i) max Ql
Wiit) max Ql
(ii)
(0
Remainder
of Circuit
FIG U R E. 1 1 . 1 0 Isolation o f an inductor for investigating the concept o f com plex power.
Chapter 11 • Sinusoidal State State Power Calculations “SI 3
So lu t io n
= -jl^,
By inspection, and noting that wc again presume effective values, = /oZ.1^ = cdZ./^ = V^.
= V^I jVJi^ = I\ + yQ^. This implies that = 0 and (X = V Ji- Further, the instantaneous
absorbed power is Pi ( f ) = v = COLy/lIi^cosUor)x J l l i sin(o)/) = V^// sin(2(0/), which
Since the bracketed quantity varies between 0 and 2, (f) \Qi\ , as was to be shown.
0)
EXA M PLE 11.6. This example, like the previous one, investigates the concept o f reactive power,
but in the case o f a capacitor. The calculations will all be dual to those o f Example 11.5. Hence,
given that ^ ^ ( 0 = >/2V(-sin((0/) in the circuit o f Figure 11.11, compute S^^ Qq the
instantaneous absorbed power and the instantaneous stored energ)^ 'r» terms o f C, (O ,
(i) P c (0 Qc and
(ii) lV c (0
Qc
(0
Remainder
of Circuit
FIGURE 11.11 Isolation of a capacitor for investigating the conccpt of complex power.
So lu t io n
In Examples 11.5 and 11.6, one observes that the inductor absorbs reactive power while the capac
itor absorbs negative reactive power or, equivalently, delivers reactive power. This follows from the
definition o f complex power (equation 11.13, i.e., S = T he structure o f equation 11.13
derives from the convention that whenever the phasor current lags the phasor voltage (as with the
inductor), the device is considered to absorb reactive power, whereas if the current phasor leads
the voltage phasor (as with the capacitor), the devicc is considered to deliver reactive power.
Indeed, the overwhelming majority o f loads (toasters, ovens, hair dryers, motors, transformers,
TV s, etc.) have lagging currents.
EXA M PLE 11.7. This example investigates the computation o f the various powers defined above
for an /?Ccircuit. Here, consider the circuit o f Figure 11.12, where v,„(/) = 100 V 2 c o s (200071:/)
V. Find the complex, average, reactive, and apparent powers absorbed by the load.
o 100 Q
lO kO
16nF i
S o lu tio n
S te p 1 . Compute
Z,„ry20007Cj = 100
-9
j -I- 720007T X 16x10
10
Chapter 11 • Sinusoidal State State Power Calculations SI^
I,v, = ^ = 10 + j9 .8 5 mA
^in
Step 3. Cornpute the complex power absorbed by the load. By equation 11.12,
Before doing a more complex example, we will discuss the particulars o f the principle o f conser
vation o f power in the sinusoidal steady state.
The validity o f this principle follows from KVL and KCL. This principle leads to the particular
fact that complex power is conserved in ac circuits operating in the SSS.
Note however, that the conservation principle does not hold for apparent power, i.e., for the mag
nitude o f the complex power. The following example illustrates a basic use o f the conservation law.
S I6 Chapter 11 • Sinusoidal State State Power Calculations
EXA M PLE 1 1.8. This example illustrates the application o f the principle o f conservation o f com
plex power in determining power delivered by a source and the input current to a circuit. We also
show that conser\'ation o f apparent power does not hold. Consider the circuit o f Figure 11.13.
Find the power delivered by the source and the phasor input current, given that S j = 360 +
;1 6 0 VA, S 2 = 360 - ; 1 2 0 VA, S 3 = 420 + y540 VA, S 4 = 130 + >80 VA, S 5 = 40 - ; 1 0 0 VA.
/
/
lOOV
\
V- \
S.
\ A
FIGURE 11.13 Bridge circuit where S- represents the complex power absorbed by the element.
S o lu tio n
By the principle o f conservation of power in ac circuits,
= S , + S , + S 3 + S 4 + S 5 = 1310 + j5 6 0 VA
This means that the circuit absorbs 1310 w'atts o f average power; the reactive power is 560 VAR,
and the apparent power is 1425 VA. Notice that the large component o f reactive power makes the
apparent (consumed) power larger than the actual consumed power, P .
Exercise. Repeat the above example calculations for S j = 300 + y'400 VA, S 2 = 300 - y 4 0 0 VA, S 3
= 600 + j\ 000 VA, S^ = 60 + y’80 VA, S^ = 120 -> 1 6 0 VA. What are the average and reactive pow
ers delivered by the source?
AN SW FR: S^ = 1380 + /920 VA :uid = 13.8 + / ).2 A. 1380 w ;itts and 020 VAR
The next example illustrates the computation o f various powers through basic definitions and
application o f the principle o f conservation o f power.
EXAM PLE 11.9. Consider the circuit of Figure 11. 14, which depicts a motor connected to a commer
cial pow'er source. The motor absorbs 50 kW of average power and 37.5 kVAR o f reactive power, and has
a terminal voltage = 230 V. Find IIj, the complex power delivered by the source, S^, and IV^I.
Chapter 11 • Sinusoidal State State Power Calculations
R,line = 0 .5 0
FICfURE 11.14 Motor absorbing 50 kW and 37.5 kVAR at a terminal voltage o f 230 V; the value of
^line exaggerated for pedagogical purposes; electrical code requires that the size of the connecting
wire be large enough that the voltage drop is only a small percentage of the source voltage.
S o l u t io n
Step 1. FiJid the apparent power, |S^^J, absorbed by the motor. Since
0-5 X 271.72 = 36 .9 2 kW
Step 4. Compute the complex power delivered by the source. From conservation o f power,
s. s.
V. =- *
—
= 3 4 8.4 V
h I.
In the above example we choose large to illustrate the calculations. In practice a line loss of
36.92 kW for a 50 kW motor operation would not be permitted.
In a resistor, average power is dissipated as heat. In a motor, most ol the average consumed power
is converted to mechanical power, say, to run a fan or a pump, with a much smaller portion dis
sipated as heat due to winding resistance and friction. The ratio o f the average power to the appar
ent power is called the power factor, denoted by pf, i.e..
The right-hand portion o f equation 11.25 follows directly from equation 11.13. Equation 11.25
specifies the power factor as cos(6^, — Oy) , i.e., the cosine o f the difference between the angles o f
the voltage phasor V and the current phasor I. Clearly, 0 < pf < 1 . The angle
Since cos(x) = cos(-a:), the sign o f (0^^- 0^) is lost when only the pf is given. In order to carry the
relative phase angle information along, the common terminology is p f lagging ov p f leading. A l/ig-
ging power factor occurs when the current phasor lags the voltage phasor, i.e., 0 < (0^, - 0y) < 180® .
A leaditjg power factor occurs when the current phasor leads the voltage phasor, i.e., 0 < (0y - 0^j <
180® . Practically all t)^pes o f electrical apparatus have lagging power factors. Some typical power
factor values are listed in Table 11.1.
T ype o f L oad P o w e r F a c t o r ( L a g g in g )
Fluorescent lighting 0 .5 -0 .9 5
To illustrate the idea o f leading and lagging pf, consider the circuits o f Figure 11.16. Suppose the
circuits operate at a frequency o f 400 Flz or co = 2 5 13.3 rad/sec. For the circuit o f Figure 1 1.6a,
I = (1 -jO.995) 10“^ V = 1.41 10“^/.—44.85° V. Hence, the current phasor lags the voltage pha
sor, i.e., ( 0 J ,- 0y) = 44.8 5 ° and the pf is cos(44.85°) = 0.709 lagging. On the other hand, for the
circuit o f Figure 11.16b, I = (1 + ;2 .5 ) 10-^ V = 2.7 10“3 Z 6 8 .3 ° V. Hence, the current phasor
leads the voltage phasor by 68.3", i.e., (0 — 0^^ = 6 8 .3 ° and the pf is cos(68.3°) = 0.688 leading.
o- O-
i0 .4 H IH F
1 kO 1 kQ
O-
(a) (b)
A load with a required average power demand, operating at a fixed voltage with a low pf, say 0.6,
has a relatively high reactive power component. This results in a relatively high apparent power.
Chapter 11 • Sinusoidal State State Power Calculations SI 9
Since the operating voltage is fixed, the line current needed to drive the load is higher than if the
load operated at a higher pf, say 0.95. Relatively speaking, a higher pf has a lower reactive power
component with correspondingly lower apparent power. Figure 11.9 helps to visualize the rela
tionships. For fixed line voltage, lower apparent power (higher pO means lower line current and
hence lower power loss in the connecting transmission line. In todays world o f energy conserva
tion, it is important to be energy efficient. The following example illustrates how improved pf on
a load can reduce line losses and thus decrease cost o f operation.
EXA M PLE 11.10. This example reconsiders Example 11.9, involving a motor connected to a
commercial power source as illustrated in Figure 11.17. The solution process will emphasize the
basic definition o f pf and the use o f voltage and current phasors. Suppose the motor absorbs 50
kW (about 67 hp) o f average power at a pf o f 0.8 lagging. The terminal voltage, is 230 V. The
frequency o f operation is 60 Hz or co = 120 j1 . For the first part o f the example the capacitor in
Figure 11.17 is not connected to the motor. In part (c), the capacitor is connected to the motor
to improve the p f This will reduce the magnitude o f the current supplied by the source and hence
reduce the line losses.
(a) Find the complex power delivered to the motor.
(b) Find I., V^, and the power delivered by the source, which might represent the power
delivered by the local electric company.
(c) Correct the power factor o f the combined motor-capacitor load to 0.95 lagging by choos
ing a proper value for C.
(d) Compute the new power delivered by the source to the combined motor-capacitor load.
R,line =0.5 0
FIG U RE 11.17 Motor absorbing 50 kW and 37.5 VAR at a terminal voltage of 230 V. Again, the
value of is exaggerated for pedagogical purposes; clectrical code requires that the size of the con
necting wire be large enough that the voltage drop is only a small percentage of the source voltage.
S o l u t io n
(a) Find the complex power delivered to the motor.
Step 1. Use the p f o f 0.8 lagging and the given average power to fin d the apparent power. From the
definition o f pf,
50
!s„, = 6 2 .5 k V A = \/„,/,„ ( 1 1 .2 7 )
0. 8
S20 Chapter 1 1 • Sinusoidal State State Power Calculations
Step 2. Compute . Ligging means that current phase lags behind voltage phase, i.e., 0 ^ -
= 0 - 0^ > 0 . Consider the diagram in Figure 11.18, which shows that the current phasor I^ la g s
the voltage phasor, i.e., the current phasor makes an angle o f - 3 6 .8 7 “ = cos” ' (0.8) from the volt
age phasor. Hence,
= 3 6 .8 7 “ > 0
S m= V IDP m
s,,, = | S ,„ k S m = = 6 2 .2 Z 3 6 .8 7 ° kVA
= 5 0 + ;3 7 .5 k V A = /’„„^t>Q
Step 1. FindX.. From equation 11.27 and the fact that for this part.
62.5 X 10'^
= 2 7 1 .7 4 A
V„, 230
Norc that it takes 86.918 kW to run a 50 kW motor. The difference is the loss in the power line.
If we have a way o f reducing the magnitude o f this line loss will be reduced. In fact, we do,
and this strateg)^ is the goal o f the next part o f the example.
(c) Correct the power factor o f the combined motor-capacitor load to 0.95 biggifig- Since motors are
inductive, a properly chosen capacitor can improve the pf to 0.95 lagging. The new motor con
figuration is that o f Figure 11.17, with the capacitor connected across the motor. The proper value
o f C must be found.
Step 1. What does a p f o f 0.95 lagging require in terms ofcomplex power absorbed by the motor-capac-
itor combination'^
Recall that
= (50 + > 37.5) kVA
Step 2. Find a capacitor value to reduce the reactive power. For this step consult Figure 11.19.
In Figure 11.19, one observes that is the same for both the new and old complex powers since
that is what the motor requires for its operation. The reactive powers are different. The new com
plex power with the 0.95 lagging pf has a smaller reactive power component. The capacitor must
be chosen to reduce the old reactive power to this new level. Hence,
Therefore,
It follows that
C = ^ = ^ H lZ iii4 = ,o 5 7 x lO - ’ F
co|V,„ 12071(230)“
(d) Compute the new power delivered by the source.
Step 1. Compute the complex power, denoted absorbed by the motor-capacitor load. The com- .— ^
plex power absorbed by the load is the sum o f the complex power consumed by the motor and
the reactive power of the capacitor, as illustrated in Figure 11.9, i.e.,
Step 2. Compute the new denoted Since S ”^“'is the complex power o f the combined motor-
capacitor load,
♦
V ;^ “’ = 0 . 5 l f + V,„ = (3 3 8 .5 - y 3 5 .7 2 ) V
Step 4. Compute the new complex power delivered by the source. By definition,
Hence, the new average power delivered by the source is 76 kW with pf correction as opposed to
86.9 kW without pf correction. With this pf correction, there is a reduction o f 86.9 - 76 = 10.9
kW o f power loss in the line connecting the source to the load.
Example 11.10 illustrates how adding a parallel capacitor can improve the pf o f a load. The main
motivation for improving the pf was to reduce the power loss in However, even if is
negligible, another strong reason exists for improving the load pf. Example 11.11 illustrates how
an improved power factor allows a single generator to run more motors. Example 11.11 will fully
utilize the principle of conservation o f complex power and the two consequences o f equation
11.25.
From equation 11.25 and the fact that S = /^ + jQ, we can express pf directly in terms o f P an d Q
as follows:
Chapter 11 • Sinusoidal State State Power Calculations 523
with a lagging pf for Q > 0 and a leading pf for Q < 0. Solving for Q from equation 11.29, we
obtain
e= ± P .H r - i (11.30)
P f‘
E X A M P L E 11 .1 1 . An industrial plant has a 100 kVA, 230 V generator that supplies power to one
large motor and several identical smaller motors. The resistance o f the connecting line is assumed
negligible in the approximate analysis below'. The large motor, labeled t^-^pe A, draws 50 kW at a
pf o f 0.8 lagging. Each smaller motor, o f type B, draws 5 kW at a p f o f 0.7 lagging. The configu
ration is illustrated in Figure 11.20.
Generator
230 V ^
60 Hz
lOOkVA
FIGURH I 1.20 A generator supplying power to one large motor and several smaller motors.
(a) Can the generator safely supply power to one large motor and three small motors? What
are the generator current (magnitude) and the power factor o f the combined loads?
(b) Compute the number o f small motors (besides the one large motor) that can be run
simultaneously without exceeding the generators rating.
(c) If the power factor for all motors, large or small, is corrected to 0.9 lagging by connect
ing appropriate parallel capacitors (as done in Example 11.10), how many small motors
(besides the one large motor) can be run simultaneously without exceeding the genera
tors rating?
(d) Compute the capacitances required in part (c) for the large and the small motors.
S o lu tio n
(a) Compute the reactive power for each motor type. Using equation 11.30, the reactive power for
each type o f motor is given as
1
- 1 =50 -1 = 37.5 kVA
0.8^
and
By the principle ot conservation o f power, the complex power (in kVA) supplied by the generator
is
V = ^ 3(7^^ + ;Q ^ ) = (50 . 15) + ;(3 7 .5 -h 15.303)
= 65 + ;5 2 .8 = 8 3 .7 4 ^ 3 9 . kVA
By inspection, the apparent power is 83.74 kVA, which is below the generator capacity o f 100
kVA, meaning that the generator can safely operate the large motor and three smaller motors.
The magnitude o f the generator current is 83,740/230 = 364 A. From equation 11.29, the pf o f
the combined loads is
pf = . ^ = 0 .7 7 6 2
P- + \/65“ + 5 2 . 8 “
(b) Compute the number o f stnall motors {besides the one large motor) that can be run simultaneous
ly. When one large type A motor and n smaller t)^pe B motors are connected in parallel, the com
plex power delivered by the generator is
Replacing the inequalit)' sign in equation 11.31 by an equalit)^ results in the quadratic equation
51.020/r + 882.5750/7 - 6 ,0 9 3 .8 = 0
The resulting zeros are //, = 5.288 and //2 = - 22.58. The largest positive integer satisfying the
inequality 11.31 is // = 5. Thus, at most, five small motors can be run simultaneously with the
large motor without exceeding the generators capacit)'
(c) I f all power factors are corrected to 0.9 lagging, fin d the number o f small motors {besides the one
large motor) that can be run simultaneously. We essentially repeat the calculations o f part (b) with
the new given power factor o f 0.9 lagging:
30.8641;/2 + 6 l7 .2 8 2 9 n - 6 ,9 1 3 .6 = 0
The roots o f this quadratic are = 8 and «2 = —28. The largest positive integer that satisfies the
above inequalit}' is n = 8. Thus, eight small motors, as opposed to five in the earlier case, can be
run simultaneously with the large motor without exceeding the generators capacity.
(d) For the large motor, the capacitor must absorb a negative reactive power equal to Q^.
Equivalently, the capacitor must supply a reactive power equal to
Equating equations 10.32 and 10.33, we have In x 60C^|2302 = 13284. Solving produces =
66 6 .1 6 X 10-<^ E
Equating these two quantities and solving for C^, w'e obtain Cg = 134.35 x 10“^ F. We note that
in the power industry, such capacitors are usually specified only by their kVAR rating, with no
mention o f their actual capacitive value in F.
In the above example the generator capacit}^ was given in terms o f VA, the unit o f apparent power.
The example points out the importance o f reducing reactive power to more fully utilize the power
526 Chapter 11 • Sinusoidal State State Power Calculations
capacity o f the generator. Use o f VA for generator, motor, and transformer capacit)' arises out o f
safety considerations. Most ac machiner)' operates at a specified voltage depending on the insula
tion strength. The size o f the wire and other heat transfer factors determine the maximum allow
able current o f a machine or transformer. Also, the cost and physical size o f most ac equipment
are more closely aligned to the VA rating than to other measures. Hence, the VA rating better
reflects the safe operating capacity o f ac equipment.
Another motivation for improving the power factor is economical. A power company charges a
consumer only for the actual electrical energy used. A meter measures this energy usage in units
o f kWh (kilowatt-hour). As mentioned earlier, most clectrical loads have lagging currents. As
shown in Examples 11.10 and 11.11, for a given required average power, a higher pf means lower
transmission line losses. Also, loads that operate at low pf force power companies to pursue high
er kVA ratings o f the generator equipment. Thus utilities companies encourage consumers to
operate their equipment and appliances at high pfs. Since power companies can supply more
power with the same equipment if the pf is high, they adjust their rates so that energy costs are
less with a high pf and are greater with a low pf.
(11.33a)
and
(11.33b)
p ^ <e ff (11.33c)
max
Chapter 11 • Sinusoidal State State Power Calculations >2.
To derive the conditions o f the maximum power transfer theorem, observe that the current pha-
sor, I, is
1= (11.34)
{R, + RO + j(X , + XO
(11.35)
Here Pave a hmction o f two real variables and Xj . To find the conditions for maximum set
the partial derivadves ------ and ------ to zero and solve for and X^. Differendating equation 11.35
V f [ ( R , + R i f + {X, + X 0 - - 2R l (R, + R , , )
= 0 (1 1 .36a)
dRi -)12
{R^ + Rl)~ + +
dP v "-[-1R l {X, + Xl )
= 0 (11.36b)
ax,
which is equation 11.33b. Substituting this result into the numerator of equation 11.36a yields
v3 “ ^
(R s + Rl T
R ,-R , ^
which produces equation 11.33a. (Note that this is the condition for maximum power transfer in
purely resistive circuits.) Substituting these results into equation 11.35 produces equation 11.33c,
Ir
4/?, _
The theorem can be established less formally as follows. W ith any existing connected to the
source, if the total reactance +X^) is not zero, we can always increase the magnitude of the
current, and hence the power delivered to the load, by “tuning out” the reactance, i.e., by adjust
ing to be -X ^. This implies condition 11.33b. Under such a condition, the circuit becomes
resistive, and the maximum power transfer theorem of Chapter 6 may be applied to obtain equa-
tions 11.33a and c. The maximum power obtainable with a passive load, given by equation
11.33c, is called the available power of the fixed source.
The conditions for maximum power transfer, as given by equation 11.33, are valid when both
and Xj^ are adjustable. If X^ is fixed and only is adjustable, then the condition for maximum
power transfer is
= + (11.37) n
which is obtained by solving equation 11.36a for fixed X^ and X^ .
r\
If the source is a general two-terminal linear network, then its Thevenin equivalent must be found
before application of the maximum power transfer theorem. If the source is represented by a '
Norton equivalent circuit, we can use a source transformation to obtain the Thevenin form and
then apply equations 11.33.
As pointed out in Chapter 6, maximum power transfer is not the objective in electric power sys-
tems, as the sources usually have very low impedances. On the other hand, it is a very important
factor to be considered in the design of many communication circuits, as illustrated in the fol-
lowing example.
o
Chapter 11 • Sinusoidal State State Power Calculations S29
EXA M PLE 11 .1 2 . The radio receiver shown in Figure 11.22a is connccted ro an antenna. The
antenna intercepts the electromagnetic waves from a broadcast station operating at 1 MHz. For
circuit analysis purposes, the antenna is represented by the Thevenin equivalent circuit shown in
Figure 11.22b.
(a) Find the input impedance o f the receiver if maximum power is to be transferred
from the antenna to the receiver.
(b) Under the condition o f part (a), find the magnitude of the voltage across the receiver ter
minals, and the average power delivered to the receiver.
S o l u t io n
(a) From the maximum power transfer theorem, the answers are = 1\ i l and = 1070.
(b) Since the reactances in the circuit have been “tuned out,” the input current to the receiv
er is simply 14.6/(21 + 21) = 0.348 mA. The input impedance has a magnitude
7. + 1070- = 10 7 0 .2 n
Therefore the magnitude o f the voltage across the receiver terminals is 0.348 x 1070.2 = 3 7 2.4
mV (rms). The power transferred from the antenna to the receiver is 0.348^ x 21 = 2.54 uW.
In the preceding discussions of maximum power transfer, we have assumed that the load is
adjustable. In practice the load is often fixed, as for example, in the case o f a loudspeaker having
a4 voice coil. In such cases, one designs coupling networks consisting o f lossless passive com
ponents. These coupling networks transform the fixed load impedance into one whose conjugate
matches the fixed source impedance. This permits maximum power transfer to the load. The fol
lowing example illustrates the principle. A design procedure for some simple coupling networks
will be discussed in the second volume o f this text.
530 Chapter 11 • Sinusoidal State State Power Calculations
S o l u t io n
(a) From equation 11.33c, as used in Example 11.12, the available power from the source is
2.54 j.iW. If all o f the power is delivered to R^ , then the voltage, must be
(b) The input impedance o f the coupling network with load must be the conjugate o f
the source impedance. Specifically
^in = yw ^ + ------------- r~
j(dC +
R,
Substituting the values OJ = 10^’, L = 400.9 x 10“^, C = 109.8 x 10 and /? = 100 x 10^ into the
above expression yields
= 21 + ;1 0 7 0 Q
which is indeed the conjugate o f the source impedance. Since is conjugate-matched to the
source impedance, the maximum power o f 2.54 |aW is transferred to the coupling network. Since
the coupling network consists o f L and C, neither o f which consumes average power, the 2.54 f,iW
power must be transferred out o f the coupling network and into the load resistance. The voltage
across the load resistor, Vj, is given by
= 4 ^ = V 2 .5 4 X 10“^ X 100000 = 0 .5 0 4 V
This verifies that the coupling network o f Figure 11.23 enables the largest voltage to appear across
the load resistor.
Chapter 11 • Sinusoidal State State Power Calculations 531
8. SUMMARY
Fundamental to the material in this chapter is the definition o f the effective value (rms value) o f
a periodic voltage or current waveform. For a sine wave, the effective value is the maximum value
divided by -v/2 • For a general periodic voltage or current, the effective value is the value o f a dc
waveform that will produce the same amount o f heat as the periodic waveform when applied to
the same resistance. Using the definition o f the effective value o f a waveform, formulas for the
average power absorbed by a linear two-terminal network in ac steady state were set forth and
derived. Recall that for a two-termlnal element with sinusoidal voltage v{t) = ^/2 V'^ cos(co/- 0y) an
current i{t) = 0/), the absorbed average power \s P = ^^^^cos(Oj; - G/), assuming
the passive sign convention. Next w'e presented the definition o f complex power and its compo
nent parts, which include its real part or average power, its imaginary part or reactive power, and
its magnitude or apparent pow’er. Various examples illustrating the calculation o f these powers
were given. Again, for a two-terminal element w'ith sinusoidal voltage v{t) and current i{t) as
above, the reactive power absorbed is defined to be Q = ^ ^ ^ s i n ( 0 y - 0/) VAR (volt-ampere-
reactive). After introducing these different types o f power, we proved the principle o f conserva
tion o f complex power, which implies the consen^ation o f real power and the conservation o f reac
tive power. This was followed by the definition o f power factor, pf, the ratio o f average power to
apparent power, which takes on values between 0 and 1. The need for improving a low power fac
tor and a method for achieving an improved power factor were illustrated with two examples.
The maximum power transfer theorem, first studied in Chapter 6 for the resistive nervvork case,
was taken up again in this chapter for the sinusoidal steady-state case. Here, maximum power
transfer to the load requires that the load impedance be the conjugate o f theThevenin impedance
seen by the load. As pointed out earlier, the theorem has no application in electrical power sys
tems. However, for communication circuits the maximum power transfer theorem is o f extreme
importance. The power that can be extracted from the antenna o f a radio receiver is usually in the
microwatt range, a very small value. It is therefore necessary to get as much power as possible from
the antenna system. Example 11.13 illustrates this principle.
Apparent power: the apparent power absorbed by a two-terminal element is assuming the
use o f a passive sign convention . The unit is VA (volt-ampere).
Average power: the average value o f the instantaneous power. For a two-terminal element with
sinusoidal voltage u{t) = yfz V^^cos{o)t + Qi,) and current i{t) = ^ V^jj-cos{o)t + 0/), the
absorbed average power is P = ^ ^ ^ ^ c o s (0 y - 0/), assuming a passive sign convention.
Com plex power: for a two-terminal element absorbing average power P and reactive power Q,
the complex power is defined to be S = P + jQ. The unit o f measuremcnr is VA (volt-
ampere). The magnitude of S is the apparent power.
Conservation o f powers: for any network, the sum o f the instantaneous powers absorbed by all
elements is zero. For any linear network in sinusoidal steady state, the sum o f the aver
age powers, reactive powers, or complex powers absorbed by all elements Is zero. This
property is a consequence of KCL and KVL.
532 Chapter 11 * Sinusoidal State State Power Calculations
Efifiecdve value (nns value): for a sine wave, the effective value is the maximum value divided by
V 2 •For a general periodic voltage or current, the effective value is the value of a dc wave
form that will produce the same amount of heat as the periodic waveform when applied
to the same resistance.
Instantaneous power: the power associated with a circuit element as a function of time. The
instantaneous power absorbed by a two-terminal element is p{t) = y(/)/(/), assuming that
a passive sign convention is used.
Maximum power transfer theorem: if a variable load 7.^^ = -^j^i is connected to a fixed source
Vy having a source impedance =7?^+/A^, then the largest average power is transferred
to the load when is the complex conjugate of Z j , i.e., = R^ and AT^ =-X^.
Power factor: the ratio of average power to apparent power. The pf value lies between 0 and 1. '
For a passive load, the power factor is said to be lagging when 90° >0j, - 6,->0, and lead
ing when 90° > 0^- 0^, > 0.
Real powen in ac circuits, real power means average power. It is the real pan of the complex
power.
Reactive power: for a two-terminal element with sinusoidal voltage v{t) = -y/2 V^cosisat + 0^)
and current i{t) = ^ /^ co s(co f + 0^, the reactive power absorbed, denoted by Q, is
defined to be Q = ^ ^ ^ sin (0 y - 0/), assuming that a passive sign convention is used.
The unit of measurement is VAR (volt-ampere-reactive).
o
Chapter 11 • Sinusoidal State State Power Calculations ^33
Problems 20
AV(t)(V)
INSTANTANEOUS AND
AVERAGE POWERS t(sec)
(a)
Figure Pi 1.3 (a) Rectangular waveform, (b)
Triangular waveform.
EFFECTIVE VALUE OF
NONSINUSOIDAL SIGNALS
(b)
i(t)0 60 O < = 30 Q
(c)
Figure PI 1.5
12 15
R
(a)
Figure PI 1.8
ANSWl-.R: 25 watts
t(sec)
-4 -■
(b)
Chapter 11 • Sinusoidal State State Power Calculations
9. In the circuit shown in Figure PI 1.9, i-{t) = (b) Compute the average power delivered
5 cos(30/) A, /? = 5 Q, and C = 5 niF. by each source.
(a) Find v^it). (c) Repeat parts (a) and (b) for R = 50 Q,
(b) Using equation 11.10, find the instan Zi = p O Q, and a = 49.
taneous and average power absorbed
by the load.
/Y Y V
— ►
Load
' ■6 aV
Figure 1’ 11.1 I
ANSW'FRS: (b) in random order: - 9 0 . 100
watts; (c) 4. - 3 .9 watts
Figure P l l .9
AN SW ER: (b) 40 warts 12. Consider the circuit o f Figure PI 1.12,
where = 120 R = 32 W, Z^ = ;2 0 0 Q.
10. In the circuit o f Figure PI 1,10, = Z q = y’80 ri, and ^ = 4.
5 0 z i-9 0 ° /? = 6 Q, = ;1 2 Q, and = (a) Find the average power (in watts)
- ;4 Q. absorbed by the resistor.
(a) Find the phasor current and deter (b) Find the average power delivered by
mine its magnitude. each source.
(b) Using equation 11.10, find the aver
age power (in watts) delivered by the /Y Y V
z. -
source.
(c) O nly R absorbs average power.
' •6
Therefore, once III is
known, the bl
average power consumed by R is =
y?| Check your answer to part (b) Figure PI 1.12
using this formula. ANSWF.RS: (b) in random order: S7.6, 230.4
(d) Repeat parts (a) and (b) when 7? = 30 watts
Q, Z^ = y'50 Q, and Z q = -ylO Q..
13. A coil is modeled by a series connection o f
/Y Y V L and R. When connected to a 110 V 60 Flz
R
source, the coil absorbs 300 watts o f average
pow'er. II a 10 Q resistor is connected in series
with the coil and the combination is connected
to a 220 V 60 Hz source, the coil also absorbs
300 watts o f average power. Find L and R.
Figure Pi 1.10
AN SW FR: R = 0.9901 LI and L = 16.6 mH.
A N SW FR; (b) ISO wans
v.(t)
Figure P 11.17
C H EC K : Complex power is 7+y’6 VA
Figure P 11.14
18. In the circuit shown in Figure Pi 1.18, to =
ANSW ER: /. = i2.9 niH, R - 14.SS3 Q 64 rad/sec, = 120/ i60" = 20 £2, /?,
= 4 12, /., = 0.375 H, and L , = (^-125 H. Find
15. Repeat Problem 14 for = 4 £2, = the complex and average powers absorbed by
400 watts, and nns with all other values the the load.
same.
C H EC K : L = 7.958 mH, R = 17.794 £2 Load
—TY"YA--- Q
R. L. +
COMPLEX POWER
CALCULATIONS
Figure Pi 1.19
C H EC K : complex power delivered by source h
Z,
is 10 + 260; VA.
v .Q
CONSERVATION OF POWER
Figure PI 1.20
C H EC K S: (a) Complex power delivered by
v,(t)
source: 48 + y'36 kVA; (b) |lj = 26.087 A
24. The circuit shown in Figure P I 1,24 is in 26. The circuit in Figure P I 1.26 operates at CO
the sinusoidal steady state. Suppose that = 500 rad/sec and = 100Z0° V ^ . The com
absorbs 3000 W of average power at pf = plex power drawn by the load without the
0.7905 lagging, Ri-^^ = 0.1 £2, and = 120 capacitor attached is S^; = 100Z30® = (86,6 +
V rms’ y‘50) VA. This constitutes a pf of 0.866 lagging.
(a) Find the average power absorbed in (a) Find the values of R and L
the transmission line resistance (in (b) Find the value of C in fiF that pro
W). duces a pf of 0.95 lagging.
(b) Find =A cos(1207t^ + 0) V and Load
the complex power delivered by the 0 -"0 -
source.
!'4<
Transmission
line resistance
— 'N/S/'------ O-
R .,„ + I L
o - -o-
FigurePll.26
CHECK; 3 pF < C< 5 pF
Figure P I 1.24
CH ECK: (a) 100 watts
27. The circuit shown in Figure P I 1.27 is operat
ing in the SSS with v^(0 = 120>/2cos(1207cO
25. As shown in Figure P I 1.25b, a capacitor is
V. Device 1 absorbs 360 W with pf =0.9. Device
put in parallel with a motor using average
2 absorbs 1440 W with a pf of 0.866 lading.
power = 40 kW operating at a power fec-
Find the value of the capacitor C such that the
tor of 0.7 laggmg to boost it to a power factor
magnitude of the source current equals 16 A ^ .
of 0.9 lading. The voltage across the parallel
What is the pf of the two-device-plus-capacitor
motor-capacitor combination is 230Z0° V ^ .
combination?
The power relationships are shown in Figure
P I 1,25a. If the frequency of operation is CO =
120tc, compute the proper value of the capaci
tance, C (in mF). Device Device
1 2
Figure P I 1.27
CH ECK: 0.108 mF
(c) Determine the annual savings from By what amount in kVAR must
installing the capacitor bank if a the reactive power be reduced to
demand charge (in addition to the produce a pf o f 0.94 lagging?
charge for the kilowatt-hours used by (iii) Compute the needed capacitor
the induction motors) is applied at current
S20.00 per kVA per month. (iv) Compute Z^j(£)) as the ratio o f
AN SW ERS: (a) 0.1771 niF; (b) Minimum the capacitor phasor voltage to
kVA rating: 3.8457 kVA; (c) S608.14 the capacitor phasor current, at
the indicated frequency.
29. Consider a source that drives an electric (v) Compute the proper value o f C in
motor that consumes an average power o f 94 mF.
kW (about 125 hp) at a pf o f 0.65 lagging, as (i) Compute the new I"^ .
show'n in Figure PI 1.29, where = 0.07 Cl. (j) Compute V 'f":
(k) Compute the complex power deliv
ered by the source and the new effi
— o- ciency.
Later
Addition of
M o to r) — Capacitor
for p.f. MAXIMUM POWER
Correction
TRANSFER
Figure P11.29
30. Consider the circuit shown in Figure
The phasor voltage across the motor is ^eff~ PI 1.30 in which = 100 Q, R-> = 25 C=
2 3 0 Z 0 V. The sinusoidal frequency is 60 Hz. 1 mF.
(a) Find the apparent power delivered to (a) Find the value o f the load impedance
the motor in kVA. that will absorb maximum power
(b) Find the complex power, delivered at CO = 100 rad/sec.
to the motor. (b) Given the conditions o f part (a) and
(c) Determine the reactive power in VAR = 100 find the average power
delivered to the motor. absorbed by the load.
(d) Compute I
f u—
(e) Compute V^. R.
(f) Compute the complex power deliv
v .Q
ered by the source.
(g) Determine the efficiency o f the con o—
figuration, i.e., the ratio o f average Figure PI 1.30
power delivered to the motor to the C H EC K : 20 + ;1 0 Q
average power delivered by the source
as a percentage. 31. In the circuit o f Figure PI 1.31,
(h) Add a capacitor across the motor to v^(r) = 100V 2cos(1000/) V, /?, = 80 Q, R, =
improve the power factor to 0.98 lag 20 Q, Z. = 5 mH. Find the value o f the resist
ging. Then ance R^ (in Q) and the capacitance C (in mF)
(i) Compute Compare with such that maximum average power is absorbed
sou
by the load.
(ii) Recall that the role of the capaci
tor is to reduce the reactive power.
540 Chapter 11 • Sinusoidal State State Power Calculations
Load
/ Y Y V
R L
zrc,
Figure P 11.38
'" ‘" Q A N SW FRS: (a) Z^ = 100 f ylOOO Q: (b) 99 pi-
and 0.2 niH
/ Y Y V
37. The circuit o f Figure PI 1.37 operates in the
6Q
sinusoidal steady state with
v .(t )Q
v^(O = 5 0 V 2 co s(2 0 0 0 r) V. -jisn
(a) Choose R and C such that the maximum
average power is absorbed in the load
resistor R^^ = 5 when Z. = 0.1 mH. Figure P I 1.39
What is this maximum average power?
542 Chapter 11 • Sinusoidal State State Power Calculations
WC,ave _ Qc
2(0
HISTORICAL NOTE
The Laplace transform converts a time function into a new function o f a complex variable via an
integration process. The name Laplace transform comes from the name o f a French mathemati
cian, Pierre Simon Laplace (1 7 4 9 -1 8 2 7 ). Pierre Laplace adapted the idea from Joseph Louis
Lagrange ( 1 7 3 6 -1 8 1 3 ), who in turn had borrowed the notion from Leonhard Euler
(1 7 0 7 -1 7 8 3 ). These early mathematicians set the stage for converting complicated diff-erential
equation models o f physical processes into simpler algebraic equations. The Laplace transform
technique allows engineers to analyze circuits and to calculate responses quickly and efficiently.
In turn engineers became better able to design circuits for radio communication and the tele
phone, not to mention other, earlier electronic conveniences. This chapter introduces the notion
o f the Laplace transform, a mathematical tool that is ubiquitous in its application to an army o f
engineering problems.
CHAPTER OBJECTIVES
1. Explain and illustrate the benefits o f using the Laplace transform tool for solving circuits.
2. Develop a basic understanding o f the Laplace transform tool and its mathematical prop
erties.
3. Develop some skill in applying the Laplace transform to differential equations and cir
cuits modeled by differential equations.
SECTION HEADINGS
1. Introduction
2 Review and Summary o f Deficiencies o f “Second-Order” Time Domain Methods
3. Overview o f Laplace Transform Analysis
4. Basic Signals
5. The One-Sided Laplace Transform
6. The Inverse Laplace Transform
7. More Transform Properties and Examples
Chapter 12 • Laplacc Transform Analysis 1; Basics
1. INTRODUCTION
I'his chaprer introduces a powerful mathematical tool for circuit analysis and design named the
Laplace transform. Later, more advanced courses will describe the design aspects. Use o f the
Laplace transform is commonplace in engineering, especially electrical engineering. A student
might ask why such a potent tool is necessary for the analysis o f basic circuits, especially since
many texts use an alternative technique called complex frequency analysis. Complex frequency
analysis does not permit general transient analysis; rather, it restricts source excitations to sinu
soids, exponentials, damped sinusoids, and dc signals. This class of signals is small and does not
begin to encompass the broad range o f excitations necessary for general circuit analysis and the
related area o f signal processing. The Laplace transform framework, on the other hand, permits
both steady-state and transient analysis of circuits in a single setting. Additionally, it affords gen
eral, rigorous definitions of impedance, transfer fimctio}!, and various response classifications perti
nent to more advanced courses on system analysis and signal processing. Introducing the Laplace
transform early allows students an entire semester to practice using the tool and learn about its
many advantages.
Section 2 describes some of the difficulties associated with the methods o f circuit analysis intro
duced in earlier chapters when applied to circuits o f order 3 or higher. Following this, we present
an overview o f Laplace transform analysis in section 3, define important basic signals in section 4,
and introduce the formal definition o f the one-sided Laplace transform in section 5. The inverse
Liplace transform and important properties o f the transform process arc introduced in sections 6
and 7, with numerous illustrative examples. Section 8 applies the technique to circuits modeled
by differential equations. Such models were developed in Chapters 8 and 9.
Recall that the output or response o f a circuit depends on the independent source excitations, on
the initial capacitor voltages, and on the initial inductor currents. Calculation o f the output often
begins with the writing o f an algebraic or a differential equation model o f the circuit for the out
put variable in terms o f the source excitations or inputs and element values. For first- and second-
order circuits with simple .source excitations, such as dc or purely sinusoidal, the solution o f the
differential equation circuit model has a known general form containing arbitrary constants. See,
for example, Tables 9.1 and 9.2. The arbitrary constants depend on the initial conditions and the
magnitude o f the dc excitation or on the magnitude and phase o f the sinusoidal excitation.
Specifically, the steps in finding the response o f a second-order circuit to a constant input are as
follows;
Chapter 12 • Laplace Traiisforni Analysis 1: Basics
or if A, = X-),
Step 4. Compute the constant D by shorting itiductors, open-circuiting capacitors, and analyzing the
restdting resistive circuit.
Step 5. Compute the constants A and B using the initial conditions on the circuit.
For circuits beyond second order, the approach in the above algorithm tends to break down.
Example 12.1 demonstrates how the approach breaks down with a simple third-order circuit.
As mentioned earlier, the foregoing technique, although quite useful for simple circuits, has seri
ous drawbacks for circuits with more than two capacitors or inductors. This is because higher-
order derivatives o f circuit output variables generally have little or no physical meaning. Such
derivatives are complicated linear combinations o f initial capacitor voltages and initial inductor
currents. The following example illuminates the difficulties.
E X A M P L E 12.1. Figure 12.1 shows three circuits coupled through the use o f dependent volt
age sources. The goal o f this example is to construct a differential equation model, determine the
solution form in terms o f arbitrary constants, and demonstrate the difficulties w'ith the simple
recipe o f the above algorithm by attempting to relate the arbitrary constants to the initial condi
tions.
1Q 1Q 1O
-I-
FIG URE 12.1 A cascade of three RC circuits coupled by means of dependent voltage sources. The
differential equation model of the circuit is third order.
S o l u t io n
Step 1. Construct the differential equation o f the circuit. For this task, first write a dilTerential equa
tion relating to Then write one relating to and finally, write one relating to
Some straightfor^vard algebra leads to the following three differential equations:
Chapter 12 • Laplacc Transform Analysis 1: Basics
^ + v c i ( 0 = 0.5iv„ ( 1 2 .!) ^
at
0 . 2 5 ^ . v „ „ , ( O = 0 ,5 v „ ^ , ^ 3 ^
Successively substituring equation 12.1 into equation 12.2 and equation 12,3 into the result pro
duces the input-output differential equation model, ^
^’ont -7 ^~^out , 1 A , o
dt^ dr d, (12-4)
Step 2. Compute the characteristic equation and its roots. The characteristic equation for differen
tial equation 12.4 is
^ + 7s^ + \As + ^ = {s - a) {s - b) {s - d) = {) ^
Step 3. Determine the form o f the solution. If v-^j^t) = then the complete solution has the
form
Step 4 . Compute A, B, D, and E in equation 12.5. Using the rule o f thumb mentioned earlier, a
simple calculation yields E = 0.125 Calculation o f A, B, and D specifies the solution in equa
tion 12.5. Applying the recipe described earlier, we take derivatives o f equation 12.5. set t = 0, and
relate them to the circuit initial conditions:
^out (0) = ^ + dD
'^out (0) =
Again, one dot over a variable means a first-order time derivative, and two dots denotes a second-order
time derivative. A, B, and D are computed by solving this set o f equations. The difficulty is in spec-
ifying and First, is simply the initial capacitor voltage on the third capaci
tor. However, v^^^^it) is proportional to the current through it, which depends on all the initial
capacitor voltages. Further, what is the physical interpretation o f And how do v^^^{0)
and v^^fiO) relate to the initial capacitor voltages? The relationship is complex and lacks any mean
ingful physical interpretation. Finally, even for this simple example, computation and solution o f
the differential equation 12.4 proves tedious.
Chapter 12 • Liplacc Transform Analysis I: Basics
One o f the advantages o f Laplace transform analysis is that it does not destroy the physical mean
ing o f the circuit variables in the analysis process. Chapter 13 addresses how the Laplace transform
approach explicitly accounts for initial capacitor voltages and initial inductor currents.
Laplace transform analysis is a technique that transforms the time domain analysis o f a circuit, sys
tem, or differential equation to the so-called frequency domain. In the frequency domain, solu
tion o f the equations is generally much easier. Hence, obtaining the output responses o f a circuit
to known inputs proceeds more smoothly.
To apply the technique, one takes the Laplace transform o f the time-dependent input signal or
signals to produce new signals dependent only on a new' frequency variable s. In an intuitive sense,
and as precisely derived later, one also takes the Laplace transform o f the circuit. Assuming zero
initial conditions, one multiplies these two transforms together to produce the Laplace transform
o f the output signal. Taking the inverse Laplace transform o f the output signal by means o f known
algebraic and table look-up formulas yields the desired response o f the circuit. The effect o f initial
conditions is easily incorporated.
Figure 12.2 is a pictorial rendition o f the method. As just mentioned, one transforms the input
signal, transforms the circuit to obtain an equivalent circuit in the Liplace transform world, and
computes the Laplace transform o f the output by “multiplying” the two transforms together.
Inverting this (output) transform with the aid o f a lookup table or MATLAB produces the desired
output signal.
T '' r
In a mathematical context, one executes the same type o f procedure on a difFerential equation
model o f a circuit and, indeed, difFerential equations in general. Figure 12.3 illustrates the idea.
FIG URE 12.3 Diagram showing flow of Laplace transform analysis for solution
o f differential equations.
The benefit o f this t)'pe o f analysis lies in its numerous uses. Some o f these uses include steady-
state and transient analysis o f circuits driven by complicated as well as the usual basic signals, a
straightforward lookup table approach for computing solutions, and explicit incorporation of
capacitor and inductor initial conditions in the analysis. The forthcoming sections will flesh out
these applications.
4. BASIC SIGNALS
Several basic signals are fundamental to circuit analysis, as well as to future courses in systems
analysis. Perhaps the most common signal is the unit step function.
1, r>0
«(r) = ( 12.6)
0, r<0
defined in Chapter 8. The bold line in Figure 12.4, resembling a step on a staircase, represents the
graph o f u{i).
u(t)
FKiURE 12.4 Graph of the unit step function. It often represents a constant voltage or current le\'el.
The unit step function has many practical uses, including the mathematical representation o f dc
voltage levels. Any t)'pe o f sustained, constant physical phenomenon, such as constant pressure,
constant heat, or the constant thrust o f a jet engine, has a step-like behavior. In the case o f jet
engine thrust, a pilot sends a command signal through the control panel to the engine requesting
a given amount of thrust. The step function models this command signal.
Chapter 12 * Ijp lacc Transform Analysis 1: Basics
The shifted step, shown in Figure 12.5, models a rime-delayed unit step signal.
u(t-T)
A
1 --
Shifted steps, u (t- 7), often represent voltages that turn on after a prescribed time period T. The
flipped step fioiction, u{T - t), o f Figure 12.6 depicts yet another variation on the unit step. Here
the step takes on the value o f unity for time t ^ T. Often it provides an idealized model o f signals
that have excited the circuit for a long time and turn o ff at time T. The key to knowing the val
ues o f these various step functions is to test whether the argument is non-negative or negative.
Whenever the argument is non-negative, the value is 1; when the argument is negative, the value
IS zero.
u(T-t)
•1 ■
FIGURE 12.6 Graph of flipped and shifted unit step. This function is often used to model signals
that have been on for a long time and turn ofTat time ’/'.
1, 0</<2
(0 / ( 0 =
0, otherwise
I, -3</<6
0. otherwise
1, / < -l
ANSVC^R.S: in random t)rder: //(-I - r) + u {t - 1), wu! - u{i - 2), ii{ t + 3) - i d t - {>)
The pube function, p-^i), o f Figure 12.7 is the product o f a step and a flipped step or, equivalent
ly, the difference o f a step and a shifted step. Specifically, a pulse o f height A and width 7" is
Pr(t)
> t
T
A signal sharing a close kinship with the unit step is the ramp function r{t) depicted in Figure
12. 8.
r(t)
/•(0=
J —oo
ii(T)dx =ln{t) ( 12.8)
w h e re t is s i m p l y a d u m m y v a r i a b l e o f i n t e g r a t i o n .
f,(t) y t)
S o l u t io n
For the signal/j (/), observe that the signal begins with a ramp with a slope o f 2. Thus we have
f\ U) = 2r(f) + ? . At ; = 7, the signal/j(^) levels off. Since the 2r{t) part o f the signal continues to
increase, the increase must be canceled by another ramp o f slope 2, but shifted to the right by T
units. Thus,/j(r) = 2r{t) - 2r(t - T).
The signal j 2(^) replicates/j(^) up to 3T . After 37', the signal drops to zero. Hence we must sub
tract a shifted step o f height 2 7 ’ from/j(/‘), Thus fjit) =/j(f) - 2Tu{t - 3 7 ) = 2r{t) - 2r{t - T) -
2T u {t-5 T ).
Exercises. 1. Figure 12.10 depicts a sawtooth waveform denoted hy J{t). Sequences o f sawtooth
waveforms are used as timing signals in televisions and other electronic devices.
f(t)
A
2. For/r) o f Figure 12.10, p lo t/ l - t) and represent the ftmction in terms o f steps and ramps.
AN SW ER: /(I - /) = ;-(l - r) - >i-t) - u{-t)
EXA M PLE 12.3. Express Figure 12.11 in terms o f steps and ramps.
f(t)
> t
S o l u t io n
Observe that the signaly(f) begins with a ramp with a slope o f 1 at r = -T . Thus /(f) = r{t ■¥ 7) +
?. The signal falls o ff with a linearly decreasing ramp for 0 < r < T. If we subtract r{t) the signal
would become flat for r > 0. Thus we subtract 2r{t) to obtain the linear decrease. Hence, y(r) =
^52 Chapter 12 • Liplacc Transform Analysis 1: Basics
r{t + 7) - 2>it) + ?. For r > T, this signal, r{t + T) - 2Kr), continues to linearly decrease. Hence
for the signal to be zero for t > T, we cancel this decrease with an additional ramp. T h u s/ r) =
r{t + 7) - lr{t) + r { t - T).
Newtonian physics provides a good motivation for defining the ramp signal. Applying a constant
force to an object causes a constant acceleration having the functional form Ku{t). 'J'he integral o f
acceleration is velocity, which has the form a ramp function.
A very common and conceptually useful signal is the (Dirac) delta function, or unit impulse
Function, implicitly defined by its relationship to the unit step as
h{cj)dq
(12.9)
s/ . , V I- u{t)~ ii{t - h)
0 {{) = — //(/)= hm-----------------
(h //->{) h ( 12. 10)
Strictly speaking, the derivative o f u{t) does not exist at t = 0, due to the discontinuit)' at that
point. Without delving into the mathematics, one typically interprets equations 12.9 and 12,10
as follows: define a set o f continuous differentiable functions as illustrated in Figure 12.12a.
The derivatives, 5^(/) = -y//^(/) >of these functions are depicted in Figure 12.12b.
5A(t)
-►t
(a) (b)
Clearly, 6^(/-) has a well-def'med area o f 1, has height I/A, and is zero outside the interval 0 s t
s A. In addition, u[t) = lim //^(r), and lim 6^ = b{t) as A -♦ 0. Hence, although the definition
o f equation 12.10 is not mathematically rigorous, one can interpret the delta function as the limit
o f a set o f well-behaved functions. In fiict, the delta function can be viewed as the limit o f a vari-
et)' o f different sets o f functions. A problem at the end o f the chapter explores this phenomenon.
Despite the preceding mathematics, the delta function is not a function at all, but a distribution,^
and its rigorous definition (in terms o f so-called testing functions) is left to more advanced math
ematics courses. Nevertheless, we shall still refer to it as the delta or impulse function. The stan
dard graphical illustration o f the delta function appears in Figure 12.13, which shows a pulse of
Chapter 12 • Laplace Transform Analysis I: Basics
infinite iieight, zero width, and a well-defined area o f unit}', as identified by the “ 1” next to the
spike. Visualization of the delta function by means o f the spike in the figure will aid our under
standing, explanations, and calculations that follow.
FIGURE 12.13 Standard graphical illustration o f a unit impulse lunction having a wcll-dcfincd area
of 1. The function typically represents an energ)' transfer, large force, or large impact over a very
short time duration, as might occur when a bat hits a baseball.
where 0“ is infinitesimally to the left and 0"^ infinitesimally to the right o f zero. I f the area is dif
ferent from unity, a number Kalongside the spike will designate the area; i.e., the spike will be a sig
nal Kb{t).
One motivation for defining the delta function is its abilit)' to “ideally” represent phenomena in
nature involving relative immediate energ\' transfer (i.e., the elapsed time over which energy trans
fer takes place is very small compared to the macroscopic behavior o f the physical process). An
exploding shell inside a gun chamber causing a bullet to change its given initial velocity from zero
to some nonzero value “instantaneously” is an example. Another is a barter who hits a pitched ball,
“instantaneously” transferring the energy o f the s\vung bat to the ball. Also, the delta function pro
vides a mathematical setting for representing the sampling of a continuous signal. Suppose, for
example, that a continuous signal v{t) is to be sampled at discrete time instants t^ r,* ••• •
v{t) is to be physically measured at these time instants. The mathematical representation o f this
measuring process is given by the sifting property of the delta function.
In other words, the value o f the integral is the non-impulsive part o f the (continuous) integrand,
replaced by that value o f r which makes the argument o f the impulse zero, in this case r = tj.
Verify'ing equation 12.11 depends on an application o f the definition given in equation 12.9.
Specifically, if v{t) is continuous at f = t-, then
by equation 12.9, where are infinitesimally to the right and left o f f,.
Exercises. 1. Compute the derivatives o f the signals in Figures 12.9a, 12.9b, and 12.11.
A N SW E R S;/,(/) = 2|//(/)-/ / (/ -T ) ] . / .(z) = 2[//(/) - / / ( / - 7 ')| - 27’6 (/ - 3 7 ').
Intuitively, a transform is like a prism that breaks white light apart into its colored spectral com
ponents. T he one-sided or tmilaternl Laplace transform is an integral mapping, somewhat like a
prism, between time-dependent signals y(r) and functions F{s) that are dependent on a complex
variable s, called complex frequency.
LAPLACE TRANSFORM
Mathematically, the one-sided Laplace transform J{t) is
( 12. 12)
As the equation makes plain, the Laplace transform integrates out time to obtain a new func
tion, F{s), displaying the frequenc)' content o f the original time function/r). In the vernacu
lar, F{s) is the frequency domain counterpart of/ r). Analysis using Laplace transforms is often
called frequency domain analysis.
Exercises. 1. Find the Laplace transform o f a scaled Dirac delta function, Kb{t).
ANSWF.R; K
Answer: It is called unilateral because the lower limit o f integration is 0“ as opposed to -oo. If the
lower limit of integration were -oo, equation 12.12 would be called the two-sided Laplace trans
form , which is not covered in this text.
Answer: Our future circuit analysis must account for the effect o f “instantaneous energy transfer”
and, hence, impulses at / = 0. The use o f 0"^ would exclude such direct analysis, since the Laplace
transform of the impulse function would be zero. Using ^ = 0 is simply ambiguous.
Answer: Because the lower limit o f integration in equation 12.12 is 0“, the Laplace transform does
not distinguish between functions that are different for f < 0 but equal for ^ > 0 (e.g., «(/) and
u{t + 1) would have the same unilateral Laplace transform). However, since ^ = 0 designates the
universal starting time o f a circuit or system, the class o f signals dealt with will usually be zero for
t < 0 and thus will have a unique (one-sided) Laplace transform. Conversely, each Laplace trans-
form F{s) will determine a unique time fiinaion J{t) with the property that f^t) = 0 for ^ s 0.
Because o f this dual uniqueness, the one-sided Laplace transform is said to be bi-unique for sig-
nals/^) yfirh J{t) = 0 for ^ < 0.
Question 4 : Does every signalj{t) such th a tfj) = 0 fo r f < 0 have a Laplace transform?
Answer: No. For example, the function f i ) = ^ «(/) does not have a Laplace transform because
the integral o f equation 12.12 does not exist for this function. To see why, one must study the
Laplace transform integral closely, i.e..
Observe that e~j^* = cosicot) -jsm{(Ot) is a complex sinusoid. As f approaches infinity, the real and
imaginary parts o f the integrand in equation 12.13 must blow up, due to the term. Hence,
the area underneath the curve e^~^ grows to infinity, and the integral does not exist for any value
of a.
W h enever//) is piecewise continuous, a sufficient condition for the existence o f the Laplace trans
form is that
I f.
\f{t)\<kx€^^ (12.14)
for some constants and This bound restricts the growth of a function; i.e., the fimction can
not rise more rapidly than an exponential. Such a fiinaion is said to be exponentially bounded. The
Chapter 12 * Uplacc Transform Analysis 1: Basics
condition, however, is not necessary for existence. Specifically, the transform exists whenever the
integral exists, even if the function/f) is unbounded. W ithout belaboring the mathematical rigor
underlying the Laplace transform, we will presume throughout the book that the functions we are
dealing with are Laplace transformable.
Question 5: Why does the existence o f the Laplace transform integral depend on the value o f a , men
tioned in the answer to question 4?
Answer: If the condition in equation 12.14 is satisfied, then there is a range o f a s (recall that s =
a + yoo) over which the Laplace transform integral is convergent. This is explained in the follow
ing example.
EXA M PLE 12.4. Find the Laplace transform o f the unit step. By equation 12.12,
£{u(i)\ = U(s)= \Z
O'
(12.15)
1
a + o + 7 (0 .V
provided that o > 0. Notice that if a > 0, then -* 0 as t This keeps the area under
neath the curve finite. For a < 0, the Laplace transform integral will not exist for the unit step.
The smallest number Oq such that for all a > Oq the Laplace transform integral exists is called the
abscissa o f (absolute) convergence. In the case o f the unit step, the integral exists for all a > 0;
hence, Oq = 0 is the abscissa o f convergence. The region a > 0 is said to be the region o f conver
gence (RO C) of the Laplace transform o f the unit step. Figure 12.14 illustrates the R O C for the
unit step.
j (o-axis
-f->a-axis
FIGURE 12.14 Region of convergence, a > 0, o f the Laplace transform of the unit step function
(i.e., the Laplace transform integral will exist for all a > 0).
Question 6: Is the unilateral Laplace transform valid only in its region o f convergence?
Answer: the answer is no. There is a method in the theory o f complex variables called “ana
lytic continuation” which, although beyond the scope o f this text, permits us to uniquely and ana-
l)aically extend the transform to the entire complex plane.- Analytically means smoothly and also
that the extension is valid ever)^vhere except at the poles (to be discussed later) o f the transform.
Thus, the region o f convergence goes unmentioned in the standard mathematical tables o f one
sided Laplace transform pairs.
Chapter 12 • Liplace Transform Analysis I: Basics
EXA M PLE 12 . 5 . Find F{s) iorJ{t) = Ke u{t), where K and <{ are scalars.
So lu t io n
poo / K
Ke~^“ e~^‘dt = K
s + ci (12.16)
The integral exists if Re[j + ^] > 0. If^7 is real, then the R O C is a > -a. As mentioned in the answer
to question 6, by analytic continuation, F{s) = M{s + a) is valid and analytic in the entire com
plex plane, except at the point s = -a . The point s ■=-a\s a pole o f the rational function M{s + a)
because as s approaches -a , the value o f the function becomes infinitely large.
The preceding discussion and examples set up the mathematical framework o f the Laplace trans
form method. Our eventual focus rests on its application to circuit theor)\ which builds on two
fundamental laws: Kirchhoffs voltage law (KVT) and KirchhofT’s current law (KCL). KVL
requires that the voltage drops around any closed loop sum to zero, and KCL requires that the
sum o f all the currents entering a node be zero. For the Laplace transform technique to be useful,
it must distribute over such sums o f voltages and currents. Fortunately, it does.
l« l/ | (0 + «
E X A M PLE 12.6. Find F{s) wheny(r) = K^ti{t) + for real scalars A', , and a.
So lu t io n
The Laplace transform o f u{t) is \h by equation 12.15 and that of is ]/{s + a) by equa
tion 12.16. By the linearity property (equation 12.17),
F(.v) = ^ +- ^ ,
s s +a
with region o f convergence (a > 0} H {a > - a }, where H denotes intersection. By analytic con
tinuation, the transform is valid in the entire complex plane except at the poles, s = 0 and s = -a.
(Henceforth we will not mention the RO C in our calculations.)
Chapter 12 • Laplacc Transform Analysis 1: Basics
The transform integral o f equation 12.12 has various properties. These properties provide short
cuts in the transform computation o f complicated as well as simple signals. For example, the
Laplace transform o f a right shift o f the s i g n a l a l w a y s has the form e~^^F{s), T > 0. Shifts are
important for two reasons:
1. Many signals can be expressed as the sum o f simple signals and shifts o f simple signals.
2. Excitations o f circuits are often delayed from t = 0.
Hence, provisions for shifts must be built into analysis techniques.
£ [ f { , - T l u i i - T ) ] = r ‘'rF{s) (12.18)
Verification o f this property comes from a direct calculation o f the Laplace transform for the shift
ed function, i.e.,
Let cj = t -T ,W k WJ q = dt. Noting tliat the lower limit o f integration becomes 0“ with respect to q,
Observe that if T < 0, the property fails to make sense, since J { t - T )ii{t- 7) would then shift left.
Since the transform ignores information to the left o f 0“ one cannot, strictly speaking, recover
J{t) from the resulting transform.
Exercises. I. Find L\J{t - T)] w hen/f) is (i) Ad(t), (ii) Au(r), and (iii) Ae~'’'u(t).
2. pyU) = /iu(t) - Aii{t - T).
A N SW ERS: In random order: ----- .\----- ^ / s
s+ a s
Chapter 12 • Laplacc Transform Analysis 1: Basics
E X A M PLE 12.7. Using the tim e shift property, find F{s) for the signal sketched in Figure
12.15.
f(t)
1
1 2
-1
-2
S o l u t io n
3 5e~^ 2e~
Direct application o f linearity and the time shift propert)’ yields F {s) = —------------f- - —
s s s
Exercises. 1. Find the Laplace transform o f the pulse signal o f equation 12.7.
2. Find F{s) when/^) = A^u[t - T^) + A2 ^(t - T-^ + A^u{t - T^).
1 ~sl 1 — v7i .4 4 .—
One more property allows us to revisit the signals discussed in section 3 and take their Laplace
transforms. The new propert)^ is multiplication of//) by t. This always results in a Laplace trans
form that is the negative o f the derivative o f F{s).
Verification o f this property follows by a direct application o f the Laplace transform integral to
with the observation that te~^^ = ------ • In particular,
ds
S60 Chapter 12 • Laplace Transform Analysis 1: Basics
■oo d
-St dt =
j() n o
.d s ds J as
Table 12.1 lists this transform pair, as well as numerous other such pairs, without mention o f the
underlying region o f convergence. As mentioned earlier, we shall dispense with any mention of
the ROC, assuming that all functions are zero for r < 0.
EXA M PLE 12.8. Find the Laplace transform o f the ramp function, r(r) = tu(t).
S o l u t io n
S o l u t io n
The quickest way to obtain the answer is to apply equation 12.19. Specifically, since
^ r —nt , , T 1
s+a
1
L {s + a ) -1 ( 12.21)
ds s + a ds
An alternative, more tedious approach is to use integration by parts as follows:
oo
fOO •oo
F {s)= L te =: ydii = uv _
udv
0“ 0" O'
le •oo
+ ■dt
.V+ a
The RO C is a > -a , in which case the first term on the right-hand side is zero. Thus, in this RO C ,
evaluation of the integral term implies that
1
F {s)- dt =
•'0“ s+a (s + aY
nl
£\t''e-^^‘u(t)] = (1 2 .2 2 )
(s +a)71+ 1
Chapter 12 • Liplace Transform Analysis 1: Basics 561
AN SW ER: F{s) =
(\ + a r
EXA M PLE 12.10. Find F(s) for the signal depicted in Figure 12.16.
f(t)
S o l u t io n
From Example 12.2,/ r) = 2r(^) - 2r(t - T) - 2 T u { t - 3 7 ). Hence by linearity, the time shift prop
erty', and equations 12.15 and 12.20,
2 - 2 £ ” '^
F {s) = L \ 2rU )- 2r{t - T ) - 2Tu(t - 3 7 )
Exercises. 1. Note that the sawtooth o f Figure 12.17 is/ f) = t[u{t) - u {t- 1)]. Suppose = z/(r)
2. Use equation 12.22 to compute the Laplace transforms o f/ ;) = tr{t) for the ramp function r{t)
and forjit) = p-r{t).
2 6
AN SW TRS: — . —
.s' s
562 Chapter 12 • Laplace Transform Analysis I: Basics
EX A M PLE 12 .1 1 . The circuit o f Figure 12.18a has two source excitations, /j(/) and /2W> shown
in Figures 12.18b and c. Compute
V o Jt)
(a)
(c)
S o l u t io n
= 10/, U) + 10/2(r)
= 10/, W + lOAW
Step 2. Compute /,(j) and Some reflective thought yields /,(?) = 2ti{t) - 2r{t) + 2r{t - 1) A
and ijit) = \.5u{t) - \ 5 u {t- 1) A. From linearity, the time shift propertys and the previously com
puted transforms.
2 l + 2e
h{s) = - ~ and l 2 {s) =
S s'
20 l5^
your(s) = - - ^ + e~^
Chapter 12 * Laplace Transform Analysis I: Basics S63
Step 4. As an introduction to the next section, by inspection we can compute the time function
o f the output voltage:
= 35«(/) - 20r{t) - \5 u (t- 1) + 2 0 r{t- 1)
i\NSV('T,RS: in random order: — !----- 1-----!-----j------- !------- 1------- ?----- . — !------- y — 5— -— --------h
.v + c/ .v + /> (.v + f/)- (.v + /?)~ (.v + f/) (.v-c/) .v+ ^/ .V
2. Recall that cos(cor) = ------------------ . Show that the Laplace transform ofJ{t) = cos{cot)u{t) is
. r + 0)
3. Recall that sin(CO/) = ------------------ •Show that the Laplace transform o f/ r) = sin(ct)f)z^(f) is
(0
F{s) = - -------
, 4 4 4 . - “’
4. Find the time functions associated with Fi ( j) = — , F) (s) = ----------^ = ---------.
^ (s + 2)- -v+ 4
We end this section with Table 12.1, which lists a number o f Laplace transform pairs. Some o f
these will be developed later in the chapter and some in the homework exercises. We will refer to
this table in the next section when computing inverse transforms.
KV4 Chapter 12 • Liplace Transform Analysis I: Basics
Item Number
m t) K
Ku{t) or K KIs
m t) KlP-
+1
\I {s+ a)
]/{s+ dp-
(OqUP- + OJ^)
(Oo
10 e s in {(O Q t)u {t) s2 ~>
(i + a) + coq
is + a)
11
{s + Cl)~ + (Oq
2(0o^
12
(.v^ +toi5)“
17 te cosi(OQt)uit) is + a)~-(OQ
((5 + rt)^ +C0n)^
18
2(0o
[(5+C/)^ +(OoJ“
Cj ~ C\Ci C|5' + C 2
19 Cl cos(coor) + sin(o)o/) n it)
(.? + « )“ +0)^
A + jS A —j B
20 2^I a ^ + B ~ cos (Oq/- tan -1
' B^ A + jB ^ A -jB
21 2^A~ + B- te~^' cos cogr - tan *
.A)j [s + a + ;cOo)“ {s + a - ycoo)"
Chapter 12 • Laplace Transform Analysis I: Basics
For the Laplace transform tool to effectively anal)'ze circuits, one must be able to uniquely recon
struct time functions /(r) from their frequenc}'" domain partners F{s). Theoretically, this is attained
through the inverse l-aplace transform integral.
/ ( ,) = r V u ) , = ^ J ^ f ( . ) e V ,
over a particular path V in the complex plane. T he path F is typically taken to be the vertical
line Oj + jio where OJ ranges from -oo to +00 and Oj is any real number greater than Oq, the
abscissa o f absolute convergence.
This integral uniquely reconstructs the time structure o f F{s) to obia.\n J{t) in whichy(r) is zero
for ^ < 0. Conceptually, the process resembles the reverse action o f a prism, to produce white
light from its spectral components. An appreciation for the power o f this integral requires a
solid background in complex variables and would not aid our purpose, the analysis o f circuits.
In fact, the evaluation o f the integral is carried out using the famous residue theorem o f com
plex variables. Further discussion is beyond the scope of this text.
Just as the Laplace transform is linear, so, too, is the inverse Laplace transform, as its integral
structure suggests, i.e., Also, the unilateral transform pair
[fit), f(j)} is uftique, where by unique we mean the following: let F^{s) = L\f^{t)] and Fjis) =
Z[^(r)] coincide in any small open region o f the complex plane. Then F^{s) = Fjii) over their com
mon regions o f convergence, and/j(f) = f-y{t) for almost all r > 0, “Almost all” means except for a
small or thin set o f isolated points that are o f no engineering significance. Hence, there is a one-
to-one correspondence between time functionsy(f) for whichy(f) = 0 for f < 0 and their one-sided
Laplace transforms. Linearity and this uniqueness make the Liplace transform technique a pro
ductive tool for circuit analysis.
Virtually all the transforms o f interest to us have a rational function structure; i.e., F{s) is the ratio
o f two polynomials. Rational functions may be decomposed into sums o f simple rational func
tions. These simple rational functions are called partial fractions and their sums are known as par
tial fraction expansions. Two o f the more common “simple” terms in partial fraction expansions
have the form K b and K!{s + a). Such simple rational functions correspond to the transforms o f
steps, exponentials, and the like. Table 12.1 lists these known inverse transforms. With the table,
direct evaluation o f the line integral in equation 12.23 becomes unnecessary. Our goal is to
describe techniques to compute the simple rational functions in a partial fraction o f F{s). Once
these are found, the transform dictionar)- in Table 12.1, in conjunction with some well-known
properties o f the Laplace transform, will allow us quickly to compute the time function y(r).
Chapter 12 • Laplace Transform Analysis I: Basics
^ (^ )_ + --- + ^^l-y + ao _ + +
5" + +--- + biS + bo is - Pi )(.v - 7^2 - Pn )
where m s « and p^, ... , are the zeros o f the denominator polynomial, + ... +
+ l?Q, and are called the finite poles o f F{s). For the most part, rational functions are sufficient for
the study o f basic circuits. There are three cases o f partial fraction expansions to consider:
(1) the case o f distinct poles, i.e., p - p - for all i j;
(2) the case o f repeated poles, i.e., pj = pj for at least one i j\ and
(3) the case o f complex poles. Although case (3) is a subcategory o f case (1) or (2) or both,
its attributes warrant special recognition.
If F{s) is a proper rational function with distinct (equivalently, sitnple) poles />j, ... , then
where
K = lim F(s)
5-400 (12.25a)
The numbers Aj in equation 12.24 are called the residues o f the pole p- and can be computed
according to the formula
The rightmost equality o f equation 12.25b is valid only when the numerator factor {s - p ) has
been canceled with the factor [s - p ) in the denominator of F{s)\ othenvise, one will obtain zero
divided by zero which, in general, is undefined. As intimated earlier, this partial fraction expan
sion should enable a straightforward reconstruction o f/ r). Indeed, from Table 12.1, we immedi
ately conclude
jr(^+ a )
S o l u t io n
The solution proceeds by executing a partial fraction expansion (equation 12.24) on F(s) to pro
duce the Laplace transform o f two elementar>- signals, a step and an exponential. Specifically,
F{s) = ----------- = - - h
5(5 + a ) s s +a
Chapter 12 • Laplace Transform Analysis I: Basics S6 '
where Ah is the Laplace transform o f a weighted step, Au{t), and B!{s + a) is that o f a weighted
exponential, To find A, multiply both sides by^, cancel common numerator and denom
inator factors, and evaluate the result at j = 0, to produce A = Ma. Similarly, to find B, multiply
both sides by i cancel common numerator and denominator factors, and evaluate the result
at j = -ay to obtain B = -Ma. Recall that, by iinearit)', X “ ’ [aF{s)] = aL~^ Hence,
I ^Cl
Exercises. 1. Findy(f) when F{s) = --------- ^
AN SW ER: J{t) = 2u(t) -
'j
AN SW ER: / ( ,) = c 5 ( 0 + ^ K / ) - —
a a
2Q
V .. '> 80>v^,
S O L U T IO N
Step 1. Detemihie By voltage division, = 0.8t'y„(r), in which case
2 5 “ + 35 + 2
56H C'haptcr 12 • Liplace Transform Analysis 1: Basics
Step 2. Construct a partial fraction expansion ofV^^^^{s). Since the numerator and denominator are
both o f degree 2,
16.v" + 2 4 i-+ 1 6
= K + — + ------- (12.27)
.v(.v + 2) .v + 2
The value o f K in equation 12.27 is determined by the behavior o f F(s) at infinity (equation
12.25a), i.e..
^ 1 6 i“ + 24.v+ 16
K - lim
i —^OO i(A- + 2)
\
165 “ + 24.V + 16 Bs
Ks + /4 +
(.v + 2) .sT 2 .v=0
,v=()
Similar!)’,
16.v- + 2 4 a + 16
Ii = = -1 6
i= -2
v v ,„ ,( o = r '
r 8 161
16 + -------------- = r ' [ Li 6 Ji + r ‘
' 8 ' 1 16
- r ‘
.V .v+ 2. s ..v + 2.
= 166(r) + 8u{t) - ]6e--'u{t) V
ijt)| 1Q
40
-e
F{s) =
So lu t io n
.v(.v+l)
fijs)
F{s) =
is-a)^d(s)
S70 Chapter 12 • Liplace Transform Analysis 1: Basics
where the denominator factor - a)^ specifies a repeated root o f order k, d{s) is the remaining fac
tor in the denominator o f the rational function F{s), and n{s) is the numerator o f F{s). The struc
ture o f a partial fraction expansion with repeated roots is
Ak
FCv) = - ^ +
(s-af (12.28)
where A^, ... , are unknown constants associated with s - a, ... ,(s - a)^, respectively, and ^i^(s)
and <^(s) are whatever remains in the partial fraction expansion o f f(s). The formulas for comput
ing the y4y o f equation 12.28 are
n(s)
A k=(s-arF(s) (12.29a)
-*i-a f/(5)
n{s) (12.29b)
^A-l = y ( ( . v - « / 'F ( ^ ) )
[d(s)
and, in general,
1 njs)''
(s-afF(s)
/! ds ' (12.29c)
O f these expressions, only the first looks like the case with distinct roots; the others require deriv
atives o f (.f - a)^ F{s). Computation o f high-order derivatives borders on the tedious and is prone
to error. The above formulas, equation 12.29c in particular, are included for completeness.
Computer implementation circumvents these difficulties. An example that illustrates the use o f
the preceding formulas, as well as a usefiil trick, comes next.
EXA M PLE 12.15. The goal here is to illustrate the computation when
s +2
F(s) = —
.v“ (.v + l)“ .V .s- .v + 1 (.v+l)'^ (12.30)
The two easiest constants to find are A2 and Bj, as their calculation requires no differentiation.
From equation 12.29a,
-I 5-1-2
A, = s-Fis) =2
-5=0
i=0
and
j'-i-2
B 2 = ( s + \r F (s )
.v=-l _ s~ .9 = - !
Finding A^ and is more difficult, since formula 12.29b requires some differentiation.
According to equation 12.29b,
as
Chapter 12 • Laplacc Transform Analysis I: Basics 571
To implement this formula multiply both sides o f equation 12.30 by take the derivative o f the
resulting expression with respect to s, and evaluate at i = 0:
d ’ 5+2 _ d ^ B|i-
y4|5 + A-y H----------- h
ds (5+l)-_ ds S+\ (5+1)“
5=0 i= 0
Observe that, on the right-hand side, it is not necessary to differentiate the terms that contain A 2 ,
and ^ 2» since these terms disappear at / = 0, as the formula for requires. Consequently,
Similarly,
0 ” _ d 5 +2 ■| ^s + 2'
Bx=- ( 5 + l ) ^ F ( 5 ) = = 3
ds s= - 1 ds . 5“ . A--1 .5 * ' S' . ,v=-l
Note that since Aj, and B2 were known, a simple trick allows a more direct computation o f
Bji merely evaluate equation 12.29 at j = 1 (in fact, any value o f s, excluding the poles, will do),
to obtain
0.75 = - 3 + 2 + 0.25 + 0 .5 5 j
.v+ 2 -3 2 3
F(s) = — = — + ^ + ------ +
i “ (.v + l)- 5 “ s + \ (5 + 1 )2
The above result can also be found with the MATLAB command “residue” as follows. Let F{s) =
7i{s)ld{s). It follows that n{s) = s + 1 and d{s) = + r . In MATLAB,
»num = [1 2];
>>den = [ 1 2 1 0 0 ] ;
»[r, p, k] = residue(num, den)
The answers from MATLAB are:
r = -3 2 3 1 (the residues associated with the poles)
p = 0 0 -1 -1
and constant
k = 0.
2 5 ^ + 2 r S 3 ^ + 35 4-2
Exercises. \. Find the partial fraction expansion o f F{s) =
AN SW ERS (residues in random order): 2, 2. 2, - 1 . -1 .v^(.9+n^
3s^ + \0s + 9 A B C
F{s) = - + ------ +
+ 45* + 55 + 2 5+ 2 5+1
Chapter 12 • Laplacc Transform Analysis 1: Basics
Compute A, B, and C
ANSWHRS: In random order: 2. 1. 2
3. Use MATLAB to find a partial l^raction expansion o f F{s) = where n{s) = G{s + 2)^(^ - 2)“
(Iis)
and d{s) = s{s + 1)“(^ + 4)“. Hint: Use num = 6*poly([-2 -2 -2 2 2]) and den = poly([0 -1 -1 -4 -4]).
ANSWHR: [r,p,kj = residue(num,den) \'iclds
r = - 4 4 4 8 - 1 6 - 6 12
p = _4 .4 _| _ ] 0
, n .r -^8 -1 6 -6 12 ^ ^
1 his results m rlie r r h : r(.s) = ■ -{------------ - h----- H--------------- -----------h 6
_______________________________ •'>+ 4 (.v + 4 ) v+l (v + l)~ ________________________
The derivative formulas o f equations 12.29 are often difficult to evaluate for complicated ration
al functions, such as s
55^ + 955-^ + 692^^ + 2369.V- + 3715.9 + 2076
F{s) = ---------------------------------------------------- r--------------
(.v+l)(:? + 2)(.s- + 3)(.v+5r
A B _ C ^ D\ 1)2 D3
s+\ s +2 i + 3 .y + 5 (.v + 5 ) - (.^•+ 5 )'*'
For these functions, it is very efficient to find A, B, C, and directly. Then one evaluates F{s) at
two values o f j, e.g., j = 0 and s = 1, to obtain two equations in the unknowns D j and Dj.
Typically, solving the resulting two equations simultaneously is much easier than evaluating Z),
and D j directly by equations 12.29. Alternatively, one can use a software program such as MAT
LAB to compute the answers. In particular, in MATLAB:
n{s) n{s)
F(s) =
|(.v + a ) - + ( 0 -](/(. 5) (s + a + ju>)(s + a - j w ) d ( s ) (12.31)
Since the poles - a - j c o and - a + jco are distinct, the partial fraction expansion o f equation 12.24
is valid. Since the poles are complex conjugates o f each other, the residues o f each pole are com
plex conjugates. Therefore, it is possible to write the partial fraction expansion of f{s) as
r(s)- I
.v + « + yto 5 + a -y c o d{s) (12.32)
for appropriate polynomials and d{s). As per equation 12.25b, the first residue in equation
12.32 is
jco (12.33)
With A and B known, executing a little algebra on equation 12.32 to eliminate complex numbers
results in an expression more amenable to inversion, i.e..
with A and B specified in equation 12.33. W ith Cj and Cj given by equations 12.35, it is straight
forward to show that
C \S + C 2 ( C2 -C ^a \ to
F'ois) = =C
1, x2 2 to {s + « )"
{s + a)~ +oj^ (.9 + a) + 0) +03" (12.36)
(C2-Cici)
M O = e' C| cos(tor) + sin (to/) Hit)
[ to )
574 Chapter 12 • Laplace Transform Analysis 1: Basics
The following example illustrates the algebra for computing C, and Cj without using complex
arithmetic.
3 .r + 5 + 3 D A + jB A - j B D C^s + C^ , ^
F{s) = ------------^------- = -------- + ------ =;- + ------^ = ------ + - ^ ------- - (12.37)
(5 + 1 )(5 “ + 4 ) ^+1 s + j2 s-jl i +1 s +A
Step 1. Compute the coefficients D, C ,, and C2 in the partial fraction expansion o f equation 12.37.
First we find D by the usual techniques:
3 .r + 5 + 3
= 1
.v^ + 4
s=-\
Given that D = 1, to find C , we evaluate F{s) at j = 0, in which case 0.75 = 1 + O.2 5 C2, or Cj =
- 1 . With D = 1 and C2 = - 1 , we evaluate F{s) at j = 1 to obtain 0.7 = 0.5 + 0.2(C j - 1) or, equiv
alently, Cj = 2. Thus,
■V+ 1 s^ + 4 + 4
Step 2. Compute j{t). Using Table 12.1, items 8 and 9, to compute the inverse transform yields
Alternative Step 1. Compute A and B in equation 12.37 by hand or with MATLAB. In MATLAB,
Alternative Step 2. One must exercise caution here and note the difference between the MAT-
LAB output and the form o f the partial fraction expansion. From equation 12.39, w = +2, A = 1,
and B = - 0 .2 5 . Again using MATLAB to obtain the form needed in item 20 o f Table 12.1,
»K = 2*sqrt(A^2 + B^2)
K = 2.0616
»theta = atan2(B,A)* 180/pi
theta = -1 4 .0 3 6 2
Thus
Example 12.16 illustrates not only the computation o f an inverse transform having complex poles,
but also the computation o f Cj and C , without resorting to complex arithmetic, as was needed in
equation 12.32. The trick again was to evaluate F{s) at two distinct ^-values different from the
poles o f F{s). This yields two equations that can be solved for the unknowns Cj and C ,.
5 ^ 8 4
Exercises. 1. Find Kt) when F{s) = ------- z----------.
s(s-+4)
AN SW ER://) = [1 + 4 cos(2r) - 4 sin(2r)]//(/)
, 5s" - 2 ^ + 5
2. Find/(r) when F ( s ) - — ^ .
. v ( r + 2 5 + 5)
Another handy propert)' o f the L'lplace transform is the frequenc)' shift property, which permits
one to readily compute the transform o f functions multiplied by an exponential. With knowledge
o f the transforms o f u{t) , sin(o)/), and other functions, computation o f e~^‘u{t) and f’“"^sin(to/)«(^)
becomes quite easy.
L[ e~^'p)] = F{s * a) (1 2 .4 0 )
Chapter 12 • Laplacc'Iransform Analysis I: Basics
Xlc’- “7 ( ' ) l = = F (s + c)
where we have viewed the sum s + a m tlie integral as a new variable p, which leads to F{p) with
p replaced hy s + a.
EXA M PLE 12.17. Let//) = sin(wr)//(r). D efine^/) = e~‘" p ) = e~‘*‘ s\n{iot)u{t). Suppose it is
known that
Compute G{s).
So lu t io n
CO CO
G( j ) = £| )| = L\e-"\m \ = F (s + « ) =
5“ + (0 “
Exercise. Lcijit) = cos(cor)u(i) for which r ( s ) = ___ - ___ D efin e^r) = e J{t) = e cos{LOt)u{t).
Compute G{s). +C0“
.V+ a
AN’SW I-R:
i.s + a )- + (0 "
Another property o f particularly widespread applicability is the time differentiation formula. Its
utility resides not onl)' in obtaining shortcuts to transforms o f signals, but also in the solution o f
differential equations. Differential equations provide a ubiquitous setting for modeling a large
variety o f physical systems— mechanical, electrical, chemical, etc. In terms o f signal computation,
recall that the velocity of a particle is the derivative o f its position as a function o f time. The accel
eration is the derivative o f the velocity. After computing the Laplace transform o f the position as
a function o f time, one finds that a differentiation formula allows direct computation o f the trans
forms of the velocity and acceleration. Also, as discussed at the very beginning o f this chapter, cir
cuits have differential equation models. For example, weighted sums o f derivatives o f the response
of- the circuit are equated to weighted sums o f derivatives o f the input signal. Therefore, a differ
entiation formula is an essential ingredient in the analysis o f circuits.
y / (/ )
(It
4/(0V' ' d t = f U ) e
JO'
The following examples explore some clever uses o f the first-order time difTerentiation formula
E XA M PLE 12.18.^^
Recall that 6f/) = — u ( t ) . Using the sifting propert)', a direct calculation yields £ [(5(/)] = 1. Is
this consistent with the differentiation propert}^? Interpreting the delta function as the derivative
o f the step function and applying the differentiation formula yields
d I
£ 6 { l ) = £ - H i t ) = s £ u{r) - u { 0 ) = s
dt s) (12.42)
Exercises. I. The Liplace transform o f a signal/r) is F{s) = — . What is the Laplace trans
f o r m o f e - 2 '4 / W ?
dt'
2s + 4
AN SW I-R: -— — ^-----
(.Y+ 2 r + 4
2. £[sm{wt)6it)] = ?
AN SW ER: u
EXA M PLE 12 .1 9 . Suppose^r) = sin(wr)w(/) and we know (for example, from Table 12.1) that
F { s ) = £ sin((0/);/(/) Compute £ [cos(wr)//(r)] using the time differentiation formula.
.V" + (0
So lu t io n j
2. Now suppose it is known that £ cos(co/)/KM -. Use the result of Exercise 1 and the
-t
+ 0)' S~
differentiation propert)' to compute the Liplace transform of/r) = sin(oj/)«(/) noting that^^O ) = 0 .
1 _ 1 1 (0
ANSW ER; £ sin((0/);/(/)
(0 (!) (I) (0 V +(0' s' - (O'
578 Chapter 12 • Laplacc Transform AnaK'sis I: Basics
EXA M PLE 12.20. Lety(r) and its derivative iiave the shapes shown in Figure 12.22. Th e goal o f
this example is to explore the relationship between the Laplace transforms o f/ r) and f'{t) in light
o f the differentiation property.
FIGURE 12.22 A pulse and its derivative for Example 12.20. Observe how the derivative o f the
pulse leads to a pair o f delta functions.
£ l / ’(r)] = i : [ 5 ( 0 - 5 ( f - I ) ] = l - ^ - ^
From the differentiation formula, it must follow that L\f{t)] = sL\J{t)]. Thus,
£|/'(/)l = sL U (t)] = j i d - « - * ) = I -
demonstrating consistency.
As might be expected, the formula for the first derivative is a special case o f the more general dif
ferentiation rule:
L m = s"F(s) - )
cit” (12.43)
This rule proves useKil in the solution o f general «th-order difTerential equations. O f particular
use is the second-order formula:
(1 2 .4 4 )
Chapter 12 • Laplace Transform Analysis 1: Basics 579
The inverse o f differentiation is int^ration. The following property proves useful for quantities
related by integrals.
■
c Jo-/(9 )d? =— (12.45a)
and
F (s) (12.45b)
As with many o f the justifications o f the properties, integration by parts plays a key role. By direct
computation (using equation 12.16),
u= fiq ) d q and dv = e
Then
-too
rr
JO'
For the appropriate region of convergence, the first term to the right o f the equal sign reduces to
fO'
et f(q ) d q
f( q ) d q
JO*
Since the second term to the right of the equal sign is F{s)/s, as per equation 12.45a, the proper
ty is verified.
S80 Chapter 12 • Laplace Transform Analysis I: Basics
E XA M PLE 12 .2 1 . Find the Laplace transform o f the signal/f) sketched in Figure 12.23a using
the integration property.
f(t)
FIGUIIE 12.23 (a) A triangular signaly(/) for Example 12.21. (b) The derivative
S o l u t io n
Observe that the triangular waveform y{r) o f Figure 12.23a is the integral o f the square wave^^).
Since ^t) is easily represented in terms o f steps and shifted steps as
its Laplace transform follows from an application o f linearity and the time shift property:
x u (o i= -
g{q)dq
.S -
EXA M PLE 12.22. This example explores the voltage-current {v-i) relationship o f a capacitor in
the frequency domain by way o f the integration property. Recall the integral form o f the voltage-
current dynamics o f a capacitor:
1 r'
Taking the Laplace transform o f both sides and applying the integration property produces
J I 1 fO-
£[vcit)] = £ ic(T)dx = 7T^c(-^) + —
C-> Cs Cs
But this expression depends on the initial condition ), because
Chapter 12 * Laplacc Transform Analysis 1: Basics S8]
Therefore,
(12.46)
Cs s
Equation 12.46 says that the voltage V(^s) is the sum of rvvo terms: a term dependent on the fre
quency domain current I^^s) and a term that looks like a step voltage source and depends on the
constant initial condition V(^0~). The quantity Z^^s) = MCs looks like a generalized resistance—
“generalized” because it depends on the frequency variable s and a “resistance” because it satisfies
an Ohm’s law-like relationship, V^^s) = Z^^s) I({s). These analogies prompt a series-circuit inter
pretation o f equation 12.46 as depicted in Figure 12.24. An application o f this equivalent fre-
quenc)' domain circuit to general network analysis appears in the next chapter.
ic(t) \,{S)
O— >■ o — >-
4- +
£ [ •] Cs
V,(t) c ----- > VJs)
O 'f
O-------
FIG U RE 12.24 Equivalent circuit interpretarion of a capacitor in the frequency domain. This equiv
alent is arrived at by applying the integration propert)' of the Laplacc transform to the capacitor volt
age, seen as the integral of the capacitor current.
A second example interpreting the v-i characteristics o f the capacitor in the frequenc)' domain
ensues from the differentiation rule. Instead o f winding up with a series circuit, one obtains a par
allel circuit. The interpretation is thus said to be dual to the one just described.
E X A M PLE 12.23. This example has two goals: (i) Verify that equation 12.46 is consistent with
the differentiation formula interpretation o f the capacitor; (ii) Build a dual frequenc)' domain
interpretation o f the v-i characteristic o f a capacitor analogous to that o f Example 12.22.
Vc(-
Cs s
which, after some algebra, becomes
Notice that equation 12.47 is consistent with the application o f the derivative formula to i(^t) =
C[civ(Jdt]. This consistency offers some reassurance in the accuracy o f our development. The
interpretation o f equation 12.47, however, is quite different from that o f equation 12.46. In the
latter equation, the current /^j) equals the sum o f two currents, CsV^s) and -C t/^ 0“). This sug
582 Chapter 12 • Laplace Transform Analysis 1: Basics
gests a nodal interpretation, resulting in an equivalent circuit having two parallel branches. One
branch contains a capacitor with voltage V^s). The other, parallel branch contains a current
source with amperage Cv(\Qr). The current through the capacitive branch is where
“G ” now acts like a generalized conductance because it multiplies a voltage, similar to Ohm’s law.
“Q ” is generalized because it depends on s. Figure 12.25 presents the equivalent circuit o f the
capacitor in the frequency domain and is dual to the circuit o f Figure 12.24. Chapter 13 covers
in detail the role o f these equivalent circuits in analysis.
Ic(s)
The last elementar)' property o f the Laplace transform that we consider in this chapter is the time-
scaling property, also called the frequency-scaling property. Its importance is fundamental to net
work synthesis. Here, numerical problems, such as roundoff, prevent engineers from directly
designing a circuit to meet a given set o f specifications. Instead, the design engineer will normal
ize the specifications through a frequency-scaling technique. Once the normalized circuit is
designed, frequency-scaling techniques arc reapplied in an inverse fashion to obtain a circuit meet
ing the original specifications.
L[J\at)\ = - F
a \ci) (12.48)
Since the proof o f this property is straightforward, it is left as an exercise at the end o f the chap
ter.
E XA M PLE 12.2 4 . Figures 12.26a and b show impulse trains that model sampling in signal-pro
cessing applications. The impulse train o f Figure 12.26b is the time-scaled counterpart to that o f
Figure 12.26a.
Chapter 12 • Laplacc Transform Analysis 1: Basics ^83
f(t) f(2t)
A A
2 -- 2 --
i<
1 -- 1 - -
> t > t
1 1
FIG URE 12.26 (a) Unit impulse train, (b) Time-scaled imit impulse train. Unit impulse trains such
as these model sampling in signal-processing applications.
The time-scaled impulse train in Figure 12.26b increases the frequency at which the impulses
occur (twice as often as in the original signal). This is reflected in the Liplace transforms o f the
two signals:
'Z&U-k) = I (12.49)
k=0 A-=0 \-e'
By the time-scaling property,
Exercise. Verify, by direct calculation, that L\J{2t)\ is given by the right side o f equation 12.50.
Several more properties o f the Laplace transform are germane to our purpose. However, these
properties have a systems flavor and are postponed until Chapter 13. We close this section by pre
senting Table 12.2, which lists the Laplace transform properties and the associated transform pairs.
Chapter 12 • Laplacc Transform Analysis 1: Basics
Linearity
n cr r j s )
Multiplication by t“
ds"
d - f{ t )
Second-order differentiation = r F c v ) - 5 / ( ( r ) - / '\ ( D
dr
d''fU)
£
wth-order differentiation
dt’'
Fis) m dq
(i)X fUl)dq
Time integration
Fis)
(ii) £ .-fU l)d q
d'\-
+ a..V. +b
dl dt'
for constants and bj might model the behavior o f a linear circuit. We may use the following steps
to solve this differential equation for using the Laplace transform procedure:
1. Take the Laplace transform o f both sides o f the equation, using the appropriate deriva
tive formulas, equations 12.41 and 12.43.
2. Algebraically solve the resulting expression for
3. Compute a partial fraction expansion o f the expression for
4. Inverse-transform the partial fraction expansion to obtain the time function
If the equation is an integro-differential equation, i.e., a mixture o f both derivatives and integrals
of the input and output signals, then we simply apply the same algorithm, except we use the inte
gral formula where appropriate. Some examples ser\-e to illuminate the procedure.
EXA M PLE 12.25. Consider the pulse current excitation o f Figure 12.27a) to the RC circuit o f
Figure 12.27b. The goals o f this example are (i) to use and illustrate Laplace transform techniques
to solve a difTerential equation derived from a simple RC circuit and (11) to find the response volt
age r > 0, when V(^0~) = 1 V.
F IG U R E 12.27 Excitation currcnt (a) fora simple /?Ccircuit (b) for Example 12.25.
S o l u t io n
\ -e
X [/ (0 ] = 0 .5
Chapter 12 • Laplacc Transform Analysis I: Basics
Step 2. Find the circuit’s dijfereutial equation model that links the excitation current i{() to the response
voltage, V(it). Since ij^it) = 0.5v(\t) and = ^.‘b d v jd t, summing the currents into the top node
of the circuit yields
d\’c {t)
dt
dv(-
{f) + v c ( f ) = 2 i(t)
dt
Step 3. Take the Laplace transform o f both sides, apply the differentiation rule to the left side, and solve
for V(i{s). Applying the Laplace transform to both sides yields
2 v r(0 ” ) \-e ~ ^
Vcis) = ----- :/(.v) + - ^ ^ = -----------+
.v+1 i’ + l ^(i-i-1) .v + 1
1 1 _ 1
5(.v+ l) s (.v+1)
Thus, with the aid o f the shift propert)' and the transform pairs o f Table 12.1, we obtain
V /
Figure 12.28 presents the graph o f this response. Because o f the initial condition and the magni
tude o f the pulse input, the capacitor voltage is constant for 0 < r < 1 second. At r = 1 second, the
pulse magnitude drops to zero, making the circuit equivalent to a source-free RC circuit in which
the capacitor voltage decays to zero as shown in the figure.
V,(t)
EXA M PLE 12.26. The goal o f this example is to compute the response, denoted here by the
input current /,„(^), to the input voltage cxcitation given the scries RLC circuit o f
Figure 12.29. Suppose the initial conditions are /^(O-) = 1 A and V(^0~) = -2 V.
— TYYY
40 1H
FIGURE 12.29 Series RLC circuit for Example 12.26. Here the current //„(^) =
So lu t io n
Step 1. Compute the Laplace transform o f the input. From the tables or by inspection, X[6(r)] = 1.
Step 2. Compute the integro-differential equation o f the circuit o f Figure 12.29. The first task is to
sum the voltages around the loop to obtain
Substituting for each o f the element voltages using the mesh current, ij„{t), yields the desired inte-
gro-difFerential equation,
diin
(12.51)
Step 3. Take the Laplace transform o f both sides, substitutefor R, L, C, Vf^Qr), and and
solve for W ith the aid o f the differentiation and integration formulas, taking the Laplace
transform o f both sides o f equation 12.51 produces
I
•') ~ ~ ~
5 + 4 5 + 4 5 +2 5+ 2 ( 5 + 2 )-
Step 4 . F in d iinit). Taking the inverse Laplace transform yields the desired result:
i d t ) = (2 - 2,)e-^-‘u(t) A
588 Chapter 12 • Liplace Transform Analysis I: Basics
cli,C
dl C —oo
L.(s) r . ( ( ) “ )
ANS>X^R: .v/,-.(.v) + -^----- + -^-------- = 0 = > / ..(/ )= 10sui(/)//{/) A
^ .S .V
dxU)
+ 2v(/) = 45(/) and 2.v(/)- y{z)dz = 2ii(t)
dt
where x(0 ) = 2 , u{t) is the unit step function, and 8{t) is the Dirac delta function, then findATi).
AN SW ER: X{s) = -
•V
EXA M PLE 12.27. The final example o f this chapter looks at the leaky integrator circuit o f
Figure 12.31, which contains an ideal operational amplifier (op amp). represents the leakage
resistance o f the capacitor. Given C and /?, is chosen to achieve an overall gain constant, in
this case, 1. The objective is to compute the response assuming that t'(j(0~) = 0, and com
pare it with that o f a pure integrator having a gain constant o f —1.
Chapter 12 • Liplacc Transform Analysis 1: Basics SH‘)
R, = IMegO
V (t) = 5 u (t )
''ou.W
So lu t io n
First, note that since the op amp is ideal, -V(^t) = The goal, then, is to write a difFerential
equation that relates to z^^-and solve for using the Laplace transform method.
Step 1. Determine the dijferential equation. Since the op amp is ideal, it follows that ijr= -i^. From
O hm s law, i^ = vJR^^. On the other hand,
^ /?2 dt
dt I<2C ~ R^C
where, as indicated before, = -v^^t). Note that if /?,C= 1 and R2 is infinite, then the cir
cuit works as a simple integrator. The circuit is called a leaky integrator because /?-,C is large but
finite. Since /?, C = 1, one expects the gain constant to be 1 as well.
Step 2. Substitute values, take the Laplace transform o f both sides, a7jd solve for Vg,,f{s). Taking the
Laplace transform o f both sides, one obtains
'w ( 0 " ) + =- -
-5 0 50
+
5( 5 + 0 . 1 ) .V .9 + 0 .1
sinh(/ - 2 r , ) 6 ( f - r , )
(g) fj{t) = sm{2m - Ji)d{2t - 4) 7. Represent each o f the following signals using
(sums of) steps, ramps, shifts o f basic signals,
2. Find the Laplace transform o f each o f the etc. Then find the Laplace transform.
following signals. Use Tables 12.1 and 12.2 as
f,{t)
needed.
(a) f^{t) = lu{t) + u{t - 1) + u{t - 2) 2K -
-Au{t-A)
(b) f^{t) = 2 r { t ) - l r { t - l ) - r ( r - 3 ) +
At - 5)
(c) -> t(s)
(e) f,(t)
f,(t)
2K
K
> t
■> t
T 2T
(d)
i k f^ft)
’
-2 f,(t)
1 K
■> t
■> t 1 2
-T.
-K
(e)
f«(t) (d)
4K -
(f)
8. Represent each of the following signals using
2K
(sums of) steps, ramps, shifts o f basic signals,
etc. Then find the Laplace transform. -I---------- '----------1--------- 1— ► t(s)
1 2 3 4
(a)
(a)
(b)
5 94 Chapter 12 • Laplace Transform Analysis I: Basics
r>
-► t
(0 r^
i,(t)
(!)
Figure P I2,11
2
■>
1
(a) o
■> t Figure P I2.12
2
r>
> t
r^
Chapter 12 • Laplace Transform Analysis I: Basics
Figure P I2.18
22. (a) Considery(/) in Figure P i 2.22.
(i) Express as a sum o f appropri
19.(a) Using the formulas cosh(/7r) = +
ate step functions. Compute F{s)
and sinh(^/r) = - e~“^, find
{\) £ [cosh(/7r)] (ii) Compute £ \g {t ) \=£ — /(/)
Idt ,
(2) L [sinh(^r)]
using the derivative property.
(b) Using the formulas o f part (a) and the
multiplication-by-r property, compute (iii) Com putc£[//(0] = £ f(X )dx
(1 )X [/ T r c o s h (/ 7 r )]
using the integral property.
(2) £ [Kt sinHat)]
(b) Repeat part (a) for =J{t + 4).
(c) Again using the formulas o f part (a)
and the multiplication-by-r property, f(t)
compute
{\) £ [K r cosh(/7r)] 3 -
(2) £ [Kr s m h ia t)]
2 -
2.V + 4
20. Suppose/r) = 0 for r < 0 and F(.v) =
s +1 ■> t
Find the Laplace transforms o f the functions 1 2 3 4 5
below, identifying each o f the properties used Figure P I2.22
to compute the answer. Solutions obtained by
finding/f) are not permitted. 23. The Laplace transform is given as
(a) g^{t) = 5 J { t - T), T> 0 -(s-a)
\-e
(b) g,{t) = 2r^^t) F{s) =
s-a
(c) g^it) = 2e-‘^ % t - D , T > 0 (a) Find the Laplace transform o f e “%t).
(d) g^{t) = 5 i f i t - r i , T > 0 (b) Find the Laplace transform of tj{2t).
25. Supposey(/) = S(t) —d{t — T), T > 0. transform o f / r ) and then, using the relation
(a) Find L\J{2t)\ by direct calculation of ship, find the Laplace transform o f^ f).
the Laplace transform integral.
(b) Find L\J{2t)] by computing Hs) and
then using the scaling property.
->
Figure P I2.27
(a) F,Cv) =
s{s a){sb)
C H E C K : One residue is a.
^ - 7.v^ + 4.V + 2
( b ) F ^ s ) = ---------------------------- ^
. X s 2 / + 18i-^ + 4 6 5 ^ + 4 4 5 + 12
(c) F2,{s ) = --------------------r---------- z-----------
30. Let j{t) and ^t) be as sketched in Figure
(5 + l ) ^ 5 + 2 r
P I 2.30. Find G{s) in terms o f F{s). C H E C K : Two residues are at 2 and - 2 .
f(t)
Remark: Check answers using MATLAB.
20 -
Use the help command to make sure you
understand the terms used. For example
10 ■
for part (b),
■> n = [2 -7 4 2];
d = conv([l -1],[1 -4 4]):
[r,p,k] =residue(n,d)
Figure P12.30
{a + b ) s + l a b
31. Using partial fraction expansions and your (a) Fi(.v) =
(5 + fl)(5 + b )
knowledge o f the Laplace transform o f simple
Check: One residue is a.
signals, find j{t) when F{s) equals
( a + b + c ) s ‘' + ( b e + 2 a b + a c ) s + a b c
2 5 ^ + 1 3 r + 305 + 32 (b ) F2{ s ) =
(a ) s { s + a ) { s + b)
s{s^ + 6^ + 8)
C H E C K : One residue is a.
- s-6
(b) cs~ + { a + 2 a c ) s + il + c ) a ^ - a
Cv + 2 ) ( . v - - l )
(c) F3(.v) =
2.v'^ + 12.v“ +22.y + 8
(c)
( r +25 + 1)( j + 2) / +185^+ 9 8 5 -+ 2085+ 144
(d ) F4(s ) =
/ + 1 2 A - ^ - 2 4 r - 3 2 .y + 16 (5 + 2 ) “ (5 + 4 ) “
(d)
C H EC K : Two residues are at 2 and - 2 .
55^ + 1 4 4 5 + 2 0 4
(e) (e ) F ^ ( s ) =
(5+1) (5+ 2)“ + 6 4
'w '
35. F in d /r) when F{s) equals Vg„f{s) and then find for the
input current
25 + 16
(a)
+16 r /X “ 45 + 9
lin ( s ) = 2 0 -------------- 2— 2---------
245-72
(b)
5^ + 4 5 + 40 Check: One residue is at 20.
25^ + 885
(c)
(5 + 4)(5^ +64) +
0
20
2 5 ^ + 2 5 ^ -2 5 -6
(d) v.(t)
24 0 80
36. Find (i) the partial fraction expansion and
(ii) the inverse Laplace transform for each of
(a)
the following functions by hand. Show all
work.
C H EC K : Residue at j = - 2 is - 1 .5 .
25^ + 9 5 ^ + 1 6 5 + 1 1
(b) F2(5) =
(5 + i)(5 + 2 y
C H EC K : Two o f the residues are 2 and 1.
5^ + 4 5 ^ - 2 5 ^ - 9 5 - 3
(C) ^3 ( 5 ) = -------------=-------- T-----
( 5 - 1 ) 2 ( ^ + 2)2
CHECK: Two residues are at I and two are at - 1 . 38. Suppose F {s) it follows that
(5 + 2)^ + 9
4 5 ^ - 1 2 5 ^ + 325 + 16 fj) = cos{(ji)i)u{i) + Ar2^“'“sin(cflf)«(r). Find
(d) /=4(5)=r
(5 + ir+ 1 (5 + 2)2 + 16 tf, A^j, Kj, and CO. Now express J{t) =
cos{(Ot + 9)u{i) by finding K y a, (O and d.
37. Find the partial fraction expansion and the
inverse Laplace transform for each of the indi 39 . The Laplace transform of f j ) =
cated output voltages or currents. All answers [A^jf“'^'cos(fi)?) + sin(fi)r) + K^e~^^u{t) is
must be in terms o f real functions with real
coefficients or symbols. Show a ll work, - 5 2 5 + 228
F {s)
(a) For the circuit o f Figure P 12.37a, find (5 + 4 ) ( 5 + 0 ^ + 100
the partial fraction expansion o f
and then find for die Find a, b, K^, K y and 0).
input voltage
40. Consider the resistive circuit in Figure
P 12.40. Use Table 12.1 and the shift property
5(5 + 4 ) to find for each Vj^(s) below. Sketch
(b) For the circuit o f Figure P 12.37b, find t^ouM) by hand or with the help o f MATLAB.
the partial fraction expansion of
600 Chapter 12 • Liplacc Transform Analysis I: Basics
use superposition?
Figure P i2.40
(0 If ij{0~) = 100 mA and = 10//(r
- 7) V, where 7 = 1 0 msec, find i/it),
41. Use MATLAB to compute the partial frac
for t > 0.
tion expansions o f the rational functions listed
below. Then use Table 12.1 and MATLAB to R
obtain the associated time function. -O
i,(t) +
3.v'^ + 3 0 5 - + 86.9 + 6 4
(a) F,(5) =
.s'* + 8 i - +20.V+ 16 '■">6 v,(t)
- 4 6 .2 5 s - 6 9 2 . 8 125
(b) F2(s ) =
.s-'* + 14.5.V- + 169.5625s + 510.25
-2.s-^ + 23.V- - 68.V - 3265 Figure P i2.42
(c) F^{s) =
.V - S 3 .55 " + 134.V + 797.5
10.5.r'^ + 47.875.s- + 151.875 - 108.5938
(d) 74(‘' ) =
+ 6.5 + 36.5625.v- + 101,5625.v + 2 07.0312
- 1 .5.s-'^ - 25.75.v‘^ - 127.5.y'* - 2 9 1 .5.y~ - 330^ - 143.75
(e) F^{s) =
.s-^’ + 10.5.v-'^ + 50.v-^ + 141.V-'' + 250.V- + 262.5.V + 1 2 5
-O
i,(t) ijt)
\r
Figure P I 2.43 ;r v^(t)
vJt)
^ 4-
C=-,F
M t )
45. Consider the Z,C circuit ol^ Figure P i 2.45, -t-^/|.v(/) = oiyit)
dt
for w'hich /^(O") = 7q and = V^. Since and
there is no resistance present in the circuit,
there is no damping; hence, one expects a pure
ly sinusoidal response. Such circuits are called dt
lossless.
602 Chapter 12 • Laplacc Transform Analysis I: Basics
with initial conditions ;c(0~) = 1 andyO “ ) = 2. 51. The op amp in the circuit o f Figure Pi 2.51
Suppose , = 1 , = 1. ^3 = 1 . ^3 = 1. and/r) is assumed to be ideal. /?, = 20 R-, = 40 ki2,
= 2u{t). Find j/(^) and A-(r). and C = 10 |iF.
(a) Use nodal analysis to construct a first-
48. Reconsider Problem 47 with = 2, = 2, order differential equation describing
= 4, b^ = 3, and/r) = 2u(t). the input-output relationship o f the
voltages.
49. The inductor current i{t) in a second-order (b) If u jt ) = 2uU) V, and j/c(0) = -1 V,
RLC circuit satisfies the following integro-dif- and then =
ferential equation for / > 0. Sketch the response in
MATLAB.
(c) If v-it) = 2e~^-">‘tt{t) V and ^C^O) = 0,
v'c(() ) + 8 j^ _ //(X )f/r
find
(c) Assuming that vc(0 ) = ) = 0, (a) Three node equations in the time
^ = 0.8 Q, Z. = 1 H, C = 0.25 F, and domain have been given in equations
= 5(r), show that 12.1, 12.2, and 12.3. Take the Laplace
transform o f each o f these three node
v’c(/) = ~
equations, accounting for initial con
ditions.
(b) If = lOu(t) V and the initial
L=1H
— ► / Y Y Y
capacitor voltages are ^^(O) = 12 V,
ijt) + = 6 V, and t/^(0) = 3 V, find
;±
v„(t)
6 R=0.8Q I
v,(t)
(d) Now do a partial fraction expansion of
Vgut^^^ and determine for /> 0.
Figure P I 2.50
C H A P T E R
A FLUORESCENT LIGHT
APPLICATION
A starter circuit must quickly generate a sufficient quantity of free electrons and crcate a suffi
ciently high voltage to initiate the arc that vaporizes the mercury inside the tube. One t)'pe o f
starter circuit contains a special heat-sensitive switch in series with an inductor. We will model this
special switch by an ideal heat-sensitive (bimetal) switch in parallel with a capacitor. The concepts
developed in this chapter will allow us to analyze the operation o f such a starter circuit as set forth
in Example 13.1 1 .
CHAPTER OBJECTIVES
1. In terms o f the Laplace transform variable s, define the notion o f impedance, denoted
Z (j), and the notion o f admittance, denoted y\s). Impedances and admittances will sat
isfy a type o f O hm ’s law. These ideas are generalizations o f the phasor-based notions o f
impedance and admittance introduced in Chapter 10.
2. Learn the arithmetic o f impedances and admittances in the Laplace transform domain,
which is analogous to the arithmetic o f resistances and conductances in the time domain.
Chapter 13 • Laplacc Transform Analysis 11: Circuit Applications
3. Apply the new concepts o f impedance and admittance to redevelop the notions o f volt
age/current division, source transformations, linearity, and Thevenin and Norton equiv
alent circuits in the /-dom ain.
4. Define /-domain-equivalent circuits o f initialized capacitors and inductors for the pur
pose o f transient circuit analysis.
5. Introduce the notion of a transfer function.
6. Define rvvo special types of responses: the impulse and step responses.
7. Redevelop nodal and loop analyses in terms o f impedances and admittances.
8. Utilize the l-aplace transform technique, especially the /-domain-equivalent circuits o f
initialized capacitors and inductors, for the solution of switched /?ZC circuits.
9. Introduce the notion o f a switched capacitor circuit, which has an important place in
real-world filtering applications.
10. Set forth a technique for designing general summing integrator circuits.
SECTION HEADINGS
1. Introduction
2. Notions o f Impedance and Admittance
3. Manipulation o f Impedance and Admittance
4. Equivalent Circuits for Initialized Inductors and Capacitors
5. Notion o f Transfer Function
6. Impulse and Step Responses
7. Nodal and Loop Analysis in the j-Dom ain
8. Switching in RLC Circuits
9. Switched Capacitor Circuits and Conservation of Charge
10. The Design of General Summing Integrators
11. Summary
12. Terms and Concepts
13. Problems
1. INTRODUCTION
Chapter 12 cultivated the Laplace transform as a mathematical tool particularly useful for circuits
modeled by differential equations. This chapter adapts the Laplace transform tool to the peculiar
needs and attributes o f circuit analysis. W ith the Laplace transform methods described in this
chapter, the intermediate step o f constructing a circuit’s differential equation, as was done in
Chapter 12, can be eliminated.
Available for the analysis o f resistive circuits is a wide assortment of techniques: O hm s law, volt
age and cu rren t division, nodal and loop analysis, linearit)', etc. For the sinusoidal steady-state
analysis o f RLC circuits, phasors serve as a natural generalization o f the techniques o f resistive cir
cuit analysis. The Laplace transform tool permits us to extend the sinusoidal steady-state phasor
analysis methods to a much wider setting where transient and steady-state analysis are both pos
sible for a broad range o f input excitations not amenable to phasor analysis. Recall that transient
an;ilysis is not possible with phasors.
Chapter 13 * Laplacc Transform Analysis II: Circuit Applications
T he keys to this generalization are the i-domain notions o f impedance and its inverse, admit
tance. Instead o f defining impedance in terms ofyoj, as in phasor analysis, we will define it in
terms o f the Laplace transform variable s. This definition allows the evolution of a frequency- or
j-dependent O hm s law, j-dependent voltage and current division formulas, and ^-dependent
nodal and loop analysis; in short, all o f the basic circuit analysis techniques have analogous s-
dependent formulations. W hat is most important, however, is that with the ^-dependent formu
lation, it will be possible to define .^-dependent equivalents for circuits containing initialized
capacitors, inductors, and other linear circuit elements. These equivalent circuits make transient
analysis natural in the i-domain.
In the final section o f the chapter, we introduce the notion o f a switched capacitor circuit.
Switched capacitor circuits contain switches and capacitors, and possibly some op amps, but no
resistors or inductors. Present-day integrated circuit technolog)' allows us to build switches,
capacitors, and op amps on chips easily and inexpensively. This has fostered an important trend
in circuit design toward switched capacitor circuits. A thorough investigation o f switched capac
itor circuits is beyond the scope o f this text. Nevertheless, it is important to introduce the basic
ideas and thereby lay the foundation for more advanced courses on the topic.
at
to obtain
V^is) ^ Lsliis), (1 3 .1 )
assuming /^(0“ ) = 0. Here, the quantit)' Z^(s) = Ls multiplies an ^-domain current, /^(^), to yield
an j-domain voltage, Vjis), in a manner similar to O hm s law for resistor voltages and currents.
T he units o f Z^(j) = Ls are ohms. The quantit)' Ls depends on the frequency variable s and gen
eralizes the concept o f a fixed resistance, and it is universally called an im pedance. This complex-
frequency or ^-domain concept has no time-domain counterpart.
Although the inductor served to motivate ^-domain impedance, in general an impedance can be
defined for any two-terminal device whose input-output behavior is linear and whose parameters
do not change with time. A device whose characteristics or parameters do not change with time
is called time invariant.
606 Chapter 13 • Laplacc'Iransforni Analysis II: Circuit Applications
IMPEDANCE
T he impedance, denoted Z{s), o f a linear time-invariant rwo-terminal device, as illustrated in
Figure 1 3 .1 , relates the Laplace transform o f the current, /(s), to the Laplace transform o f
the voltage, V^j), assuming that all independent sources inside the device are set to zero and
that there is no internal stored energy at ^ = 0 “. Under these conditions.
V(s)
Z(s) =
I(s) (13.2b)
in units o f ohms.
l(s)
vis) Device
Z(s) orY(s)
Exercise. For an unknown linear circuit, Vj,j{s) = —------ and I- (s) = 2. Com pute
5“ + 4
4
ANSW'HR; —
.y" + 4
T he inverse o f resistance is conductance, and the inverse o f impcdance is admittance. For exam
ple, if we divide both sides o f equation 13.1 by Ls, we obtain
(1 3 .3 )
Ls
This suggests that \/Ls acts as a. generalized conductance universally called an adm ittance, which
is defined as follows.
Chapter 13 • Laplacc Transform Analysis II: Circuit Applications 607
ADMITTANCE
T he admittance, denoted y(^), o f a two-terminal linear time-invariant device, as illustrated
in Figure 1 3 .1 , relates the Laplace transform o f the voltage, V(j), across the device to the
Laplace transform o f the current, /{^), through the device, assuming that all internal inde
pendent sources are set to zero and there is no internal stored energy at f = 0~. Under these
conditions,
Ijs)
(13.4b )
V{s)
in units o f S.
From equations 13.2 and 13.4 , impedance and admittance satisfy the Inverse relationship
1
Y(s) =
Zis) (1 3 .5 )
16
Exercise. For an unknown linear circuit, = ---- arid A/l(-^)="
C om pute Ky,//). ' (.S- + 2) (>v + 2 )(i + 4 )
2
ANSWER: ^+ 4
As a first step in deepening our understanding o f these notions, we compute the impedances and
admittances o f the basic circuit elements shown in Figure 13.2.
i,(t) ic(t)
O- O— o — >■
-I- +
o-
From O hm ’s law, the resistor o f Figure 13.2a satisfies = Rij^it). Laplace-transforming both
sides yields the obvious, K^(j) = Rlj^is). From equations 13.1 and 13.2, the im pedance o f the
resistor is
Zf^is) = R
608 Chapter 13 • l-aplacc Transform Analysis II: Circuit Applications
(!v(^{ t )
/c ( 0 = C-
dt
I^s) = CsV^s)
Y(is) = a
1
Z eis) =
Cs
Repeating this process for the inductor o f Figure 13.2c , )= L — —— the im pedance and
ad m ittan ce o f the in d u cto r are
Z,,U) = U . YLU) = - j-
Ls
Exercises. 1. Given the integral form o f the v-i capacitor relationship, assume no initial stored
energy and take the Laplace transform o f both sides to derive the impedance o f the capacitor. This
provides an alternative, more basic means of deriving the impedance characterization.
2. Given the integral form o f the v-i inductor relationship, assume no initial stored energy and
take the Laplace transform o f both sides to derive rhe admittance o f the inductor.
Throughout the rest o f the text, whenever we refer to an impedance the unit o f Ohm is assumed,
and similarly, admittance is assumed to have the unit of siemens (S). The units for KW and I{s)
are usually not shown, although strictly speaking they are volt-second and ampere-second, respec
tively.
Chapter 13 • Laplace* Transform Analysis II: Circuit Applications 609
Recall that the Laplace transform is a linear operation with respect to sums of signals, possibly
multiplied by constants. KVL and KCL are conservation laws stating, respectively, that sums of
voltages around a loop must add to zero and sums o f all currents entering (or leaving) a node must
add to zero. Since the Laplace transform is linear, it distributes over these sums, so the sum of the
Liplace transforms o f the voltages around a loop must be zero and the sum o f the Laplace trans
forms o f all the currents entering a node must be zero. In other words, complex-frequency
domain voltages satisfy' KVL and complex-frequency domain currents satisfy' KCL. Because of
this, and because impedances and admittances generalize the notions o f resistance and conduc
tance, one intuitively expects their manipulation properties to be similar. In fact, this is the case.
1. Impedances, Z{s), can be manipulated just like resistances and, like resistances, have units
o f ohms.
2. Admittances, K(/), can be manipulated just like conductances and, like conductances,
have units o f S.
This manipulation rule suggests, for example, that admittances in parallel add. T he following
example verifies this property for the case o f two admittances in parallel.
EXA M PLE 13.1. C om pute the equivalent admittance, and impedance, o f three
general admittances, K,(/), V^{s), and Y:^{s) in parallel, as shown in Figure 1.3.3. Then develop the
current division formula.
Yji-s]
An(-v) (1.3.6)
K,(.v)+Ko(.v)+r,(.v)
Z Js)
S o lution
W c seek the relationship vvhich implicitly defines From the definition
o f the admittance o f a two-terminal device, /^(j) = for = 1 ,2 , 3. From KCL,
= AW ^3(5) = >3(5))
/;,(,C) = Y / : ( S ) V , „ ( S ) = y , ( s ) Z , „ ( s ) / , ; , ( s ) =
Y^{s )+Y 2( s )+Y2{ s )
Exercises. 1. Show that for t%vo impedances, Z^{s) and Zjis), in parallel, Z ,„(^) =
2. Show that the equivalent impedance o f two capacitors in parallel is 2 i ( ‘^) +
Z (5 )= ‘ ‘
C]^+C25 (C j+ C 2 )5
L^ + Lo
6. In the circuit o f Figure 13.3, suppose Kj(j) = MR, K,(j) = M{L$), and = Cs, a resistance,
an inductance, and a capacitance. Derive the relationship
1
fin (s)
L C s ^ + -s +
R
1 s
7. In the circuit o f Figure 13.3, suppose Kj(j) = 0 .5 , ^ ^ Find
25+1
the equivalent admittance, and find I^{s) in
terms o f I;„{s).
in' ■ ^
Ai\S\V1-:R: y j s ) = l and /^(.v) =
EXAMPLE 13.2. Compute the input impedance o f the parallel RLC circuit sketched in Figure 13.4.
o—
+
VJs)
Z Js) =
Y Js)
S o l u t io n
For parallel circuits, it is convenient to work with admittances, since parallel admittances add.
Thus, for the circuit o f Figure 13.4,
^ 1 1
r +— .v+'
RC LC
1 1
C , 2 ^ _ 1L , + 1 (1 3 .7 )
RC LC
Exercises. 1. C om pute the equivalent impedance o f a parallel connection o f three inductors hav
ing values 4 m H , 5 m H , and 2 0 m H .
A N SW ER : 2 x 1()--S-
2. Com pute the equivalent impedance o f a parallel connection o f six elements: rsvo resistors, o f 6
kQ and 3 kQ; two inductors, o f 3 mH and 6 m H ; and two capacitors, o f 0 .2 |.iF and 0 .0 5 |.iF.
A N SW E R ; 4 X 1()^V(r + 2 x 1 i)-^s + 2 x 1 0 * ’)
1 he dual o f the parallel circuit o f Figure 13.3 is a series connection o f three impedances as shown
in Figure 13.5. T he following example verifies that impedances in series add, and simultaneous
ly develops a voltage division formula.
Zj{s)
= V^n(.v)
Z,(.v) + Z2(.v) + Z3(i-) (1 3 .8 )
+ V,(s) - + V^(s)
S o l u t io n
(1 3 .1 0 )
T he voltage division formula o f equation 13.8 follows from a modified form o f equation 1 3 .1 0 ,
and equation 13.9, to yield
Z (s )
Z ,( 5 ) + Z 2 ( s ) ...+ Z„(5)
1 1
AN SW I-R;
C’|CS c V
Cj + C .
2. Show that the equivalent admittance o f rwo capacitors, Cj and C-,, in series is Y{s) = -------- — s.
C\ + Cl
3. Suppose Zj(s) = 10 Q, Z-,(s) = 2s, and Z^(s) = 6^ in Figure 13.5. Find V^-)W, and
.V
A N SW ER S: Z,,(s) = 10 + 8^. \ s ( .0 = . /-.(/) = '
■ 4.V + 5
2
4. Suppose Zj(s) = 10 Q, Z-,(s) = 2s, and Z t^(s ) =— in Figure 13.5. Find Z^-^^(s) and K^(j) in
terms o f V^Js). ^
-7 + 10.v + 2 1
A N SW ER S: Z,„(.v) = ---------------------, =- -------------- V;„(.v)
■V ,v‘' + 5 .v + !
O f course, there are series-parallel connections of circuit elements that combine the concepts illus
trated in Examples 13.1 through 13.3, as set forth next.
EXA M PLE 13.4. Com pute the input impedance Z-J,s) o f a series connection o f t\vo pairs o f par
allel elements, as shown in Figure 13.6, in which = 10 Ci, C = 0.1 ¥, = 5 O., and Z. = 1 H.
Then compute in terms o f If = u{t), find Vjit).
614 Chapter 13 • l^place Transform Analysis II: Circuit Applications
S o l u t io n
V,(s)
Here
1 10
Z ,U ) =
O.l+O.l.v .y + 1
and
Is
Z2{s) =
7+2
in which case
Finally, if = -,
s
(^■+1) (-^ + 0
^2(^') = -3 ---------------= ---------- ^
.9“ + 6 . 9 + 10 ( 5 + 3)" +
Exercise. Repeat Example 13.4 with the following changes: C = 0.01 F and /?, = 10 Q.
A N SW ER S: Z Js ) = 10 VS(.v) =
\+10
Another basic and useful circuit analysis technique is the source transform ation property,
exhibited now in terms o f impedances and admittances. The first case we will examine is the volt-
age-to-current source transformation, illustrated in Figure 13.8,
(a) (b)
FIG U R E 1 3 .8 Illustration of the transformation of a voltage source in series with Z^[s), as shown in
part (a), to an equivalent current with a current source in parallel with Z^(s), as shown in part (b).
Often, voltage-to-current source transformations provide an altered circuit topology that is more
convenient for hand or calculator analysis. Mathematically, the goal is to change the structure o f
a voltage source in series with an impedance to a current source in parallel with an admittance
while keeping both Vjis) and / 2 W fixed. To justify this, one starts with Figure 13. 8a, in which
voltage division implies
V,(.v) = ^ ^ V^^is)=Z.is)l2is)
Z ,(5 )+ Z 2 (5 )
H en ce, ZAs) ^ 0 ,
Z2(.V)Z, is ) 1 (V iu (s)\ (1 3 .1 2 )
V ;(.v) =
Z,(.v)+Z2(.v) I 2 ,(5 )J ■r,(5)+K2(.v) U i„(.v )j
616 Chapter 13 • U p lace Transform Analysis 11: Circuit Applications
where Yjis) = [Zj{s)] ^ This equation identifies the parallel structure of Figure 1 3 ./b ; i.e., Figure
13. 8b is a circuit equivalent o f equation 13.8.
(a) (b)
FIG U R E 13.9 Illustration of (a) current source to (b) equivalent voltage source transformation.
Clearly, the manipulation of impedances and admittances parallels that o f resistances and con
ductances, as suggested earlier. Indeed, for a rigorous statement o f the soiuce transformation tech
nique developed above, refer to the source transformation theorem in Chapter 5 and replace R
by Z{s), by and by Ijp). Indeed, all such values in Chapters 5 and 6 have i-domain
counterparts.
This section ends with a demonstration of finding a Thevenin equivalent in the /-dom ain.
VJ s ) V Js ) v„(s)
(b)
FIG U R E 1 3.10
SOLUTIO N
From the material in Chapter 6, our new concepts o f admittance and impedance, and Figure
13.10b,
(13.13a)
o r equivalently,
o s-B b )
Chapter 13 * L^placcTransform Analysis II: Circuit Applications 61
= I {a + 1).vC + C ] ( 5 ) - (« + 1).vCV/^(.v)
1
(13.15)
(a + \)sC + G ' (W +D.9C + G
2-2t;\ .. 2/ / , (. s)
2 + 2.V +
FIG U RE 13.11
3. For the circuit o f Figure 13.12, use source transformations to find I^p) and Y^jj^s) for the indi
cated terminals. .
A N SW ER S: I. is) = 0 . 2 .a ' (>■) and = 0.2.^ + - + 0.4
0.2 F
(-
V Js ) 2.5 Q
1 H
F IG U R E 1 3 .1 2
618 Chapter 13 * Laplace Transform Analysis 11: Circuit Applications
T he notions o f impedance, admittance, and transfer function do not account for the presence of
initial capacitor voltages and initial inductor currents. Hoiv can one incorporate initial conditions
into various analysis schemes? For an answer vve look at the transform o f an initialized capacitor
and inductor and interpret the resulting equation as an equivalent circuit in the complex-fre-
quency domain. For the capacitor and the inductor, rvvo equivalent circuits result for each; a series
circuit containing a relaxed (no initial condition) capacitor/inductor in series with a source, and
a parallel circuit with a relaxed capacitor/inductor in parallel with a source. Example 1 2 .2 3 pre
viewed this notion.
c ^ = /c (0
(it
Taking the Laplace transform and allowing for a nonzero initial condition yields
Cs - Cv(J,Q~) = I (is) (1 3 .1 6 )
T he left side of equation 1 3 .1 6 is the difference o f two currents, one given by the product o f the
capacitor admittance and the capacitor voltage {CsVf^s)) and the other by Cy^^O"). Thus the cir
cuit interpretation o f equation 1 3 .1 6 consists o f a relaxed capacitor in parallel with a current
source, as illustrated in Figure 13.13. In the time domain the current source o f Figure 1 3 .1 3 cor
responds to an impulse that would immediately set up the required initial condition.
1^(5)
^ .....................................................
V,(s)
Cs
FIGllRK 13.13 Parallel form of an equivalent circuit for an initialized capacitor. Here, the capacitor
within the dotted box is relaxed while the current source Cv(4S)~) accounts for the initial condition.
Cs s (1 3 .1 7 )
Example 12.22 previewed this equation by taking the transform o f the integral relationship o f the
capacitor. We observe that the right-hand side o f equation 1 3 .1 7 is the sum o f two voltages, one
o f which is the product o f the capacitor impedance and the capacitor, current and the other
z/^(0“)A'. Thus, the interpretation is a series circuit, as sketched in Figure 13.14.
Chapter 13 • Laplace Transform Analysis II: Circuit Applications 619
V Js)
FIG URE 1 3.14 The series form of an equivalent circuit for an initialized capacitor. Here the capacitor
in the dotted box is relaxed, and the voltage source accounts for the effect of the initial condition.
Initialized inductors have similar j-domain equivalent circuits analogous to those o f the capaci
tor. W ith the voltage and current directions satisfying the passive sign convention, the differen
tial inductor current-voltage relationship is
Again, this equation consists o f a sum o f voltages, Lsl^is) and - Z /^ ( 0 ). Thus equation 1 3.18 can
be interpreted as a series circuit, as depicted in Figure 1 3 .1 5 .
FIG U R E 13.15 Series form of equivalent circuit for an initialized inductor. Here the inductor with
in the dotted box is relaxed; notice the polarity orientation of the voltage source.
(1 3 .1 9 )
Ls s
T he right side o f Equation 13.19 is a sum o f currents that determines a parallel equivalent circuit,
as sketched in Figure 13 .1 6 .
620 Chapter 13 * Laplace Transform Analysis II: Circuit Applications
Two examples illustrate the use o f these four equivalent circuits for initialized capacitors and
inductors.
+ vjt) -
--------- (-
S o l u t io n
In this example, it is convenient to replace the capacitor by its (series) 5-domain voltage source
equivalent circuit, because the capacitor is in series with the input voltage source. On the other
hand, it is convenient to replace the inductor by its (parallel) /-dom ain current source equivalent
circuit, because the desired output is the inductor voltage. This results in a three-source or multi
input circuit. O nce the equivalent circuits are in place, one can apply superposition to obtain the
answer, although there arc many other ways to solve the problem.
Step 1. Using the voltage source ynodelfor the capacitor and the current source modelfor the induc
tor, draw the equivalent s-domain circuit. Using the equivalent circuits o f Figures 1 3 .1 4 and 1 3 .1 6 ,
we obtain the circuit of Figure 13.18. Here we note that
+ > —O
1 .5 0 +
V,(s)
0.5s
FIG U R E 1 3 .1 8 j-domain equivalent accounting for initial conditions of the circuit of Figure 13.17.
I.5 + - + 0.55 S- + 3S + 2
s
1 -.y
V l(s) = — X —=
.5 + - + 0.55 ^ s^ + 3s + 2
s
Step 4 . Find the contribution to Vj{s) from Z,;'^(0 ) = 1. Using O hm ’s law in the 5-domain,
0.5^ 1.5 + -
s/ 2 -3 s - 2
V t{s) = - X — = --------------
1.5 + - + 0.55 ^ ^“ + 3^ + 2
5
Step 5. Su?n the three contributions and take the inverse transform.
-2 2 2
V i(s) = v l u ) + v l { s ) + vl(.s) = -
- + 35 + 2 -V+ 2 .V+ 1
in which case
Vj{t) = 2e ^hi{t) - le ^u{t) V
Exercise. Find Ij{s) and ij{t) for the circuit o f Figure 1 3 .1 7 using the equivalent circuits o f Figures
1 3 .1 4 and 1 3 .1 5 . Hint: Write one loop equation.
/Y Y \ -o
1.5 Q +
v Jt) v,(t)
1F
S o l u t io n
In this example, it is convenient to replace the inductor by its (series) coniplex-frequency domain
voltage source equivalent circuit, because the inductor is in series with the input voltage source.
On the other hand, it is convenient to replace the capacitor by its (parallel) complex-frequency
domain current source equivalent circuit, because the desired output is the capacitor voltage. This
results in a three-source, or multi-input, circuit. O nce the equivalent circuits are in place, one can
combine the voltage sources and write a single node equation to find V(As).
Step 1. Using the voltage source model for the inductor and the current source ynodelfor the capaci
tor, draw the equivalent complex-frequency domain circuit. Using the voltage source equivalent for
the initialized inductor and the current source equivalent for the capacitor produces the circuit o f
Figure 13.20a. Combining the voltage sources and the series impedance into single terms results
in the circuit shown in Figure 13.20b.
0.5 i^(O-) = 1
/ Y Y V _ Q
1.50 0.5 s
CvJO-) = l
(a)
FIG U R E 1 3.20 (a) Complex-frequenc)’ domain equivalent accounting for initial conditions o f the
circuit of Figure 13.19. (b) Circuit equivalent to part (a) with voltage sources combined.
Chapter 13 * Laplacc Transform Analysis II: Circuit Applications 623
Step 2. Write a single node equation for V(^s). Summing the currents leaving the top node o f
yields
1
V c ( s ) ---------------- l + 5V ’c ( 5 ) = 0
1.5 + 0.5.y s
Grouping terms produces
1
+ 5 V c is )= -: ----- r+ 1
U . 5 + 0.5^ 5 (0 .55'+ 1.5 )
Solving for V(\s) leads to
s~ + 5s + 2
Vcis) =
5 ( 5 + 0 ( 5 + 2)
Step 3 . Exemte a partialfraction expansion on V(^s), and take the inverse transform to obtain V(^t).
Using the result o f step 2,
5^ + 55 + 2 1 2 -2
^(7(5) = -------------------- = — I----------- f-
s{s + 1) ( 5 + 2 ) S .V + 1 5 + 2
v^t) = [1 + le-'-2e-^~^u{i)V
Exercises. 1. In Example 13.7, change the resistance from 1.5 H to 2 .2 5 Find V(\t) for r > 0.
A N SW ER : v^p) = [1 » 0.57l4i>-'^-^'- 0 .5 7 l4 f -^ q « (/) V
2. Find I^{s) and i^it) for the circuit o f Example 13.7, using the equivalent circuits o f Figures
13.14 and 13.15. Hint: W rite one loop equation.
TV
A N SW ER : 1, ( 5 ) = -------- ----------- . i,{t) =- le -‘ii{t) + Ae--'u{t)
( 5 + I ) ( 5 + 2) ^
EXA M PLE 1 3 . 8 . T he chapter opened with a discussion o f the operation o f a fluorescent light
with classical starter, com m on in residential usage. For a fluorescent light to begin operating, there
must be a sufficient supply o f free electrons in the tube and a sufficiently high voltage between the
electrodes to allow arcing to occur. During arcing, mercury particles in the tube vaporize and give
off ultraviolet light. The ultraviolet light excites a coating o f phosphorus on the inside o f the tube
that emits light in the visible range.
For a simplified analysis, assume that all resistances are negligible and refer to Figure 1 3 .2 1 . The
source VjJ^t) is 120 V, 6 0 Hz, i.e., ordinary house voltage, which is too low to cause arcing inside
the fluorescent tube. Prior to arcing the gas inside the fluorescent tube acts like a very large resist
ance betvN'een the rwo electrodes. W hen the switch is turned on, the starter, a neon bulb with a
bimetallic switch inside, lights up and heats the bimetallic strip. This causes the metal to curl and
624 Chapter 13 • Laplacc Transform Analysis II: Circuit Applications
close the contacc. The bulb then looks like a short circuit, and a large current, limited by the
inductive ballast, flows through the heating electrodes o f the fluorescent tube, making them bet
ter able to emit electrons. During this time the neon bulb is shorted out and the bimetallic strip
cools and opens the circuit after a few seconds. At this point in time, which we will call ^ = 0 , the
inductor has an initial current Because o f the Z.Ccombination, a very high voltage will then
appear across the electrodes of the lamp, resulting in ignition or arcing. After the lamp ignites, the
voltage between the electrodes becomes “small” and is insufficient to relight the neon starter lamp.
Hence, the ac current flows between the two electrodes inside the fluorescent tube. The ballast
again serves to limit the current.
Direction
Heating of curl when
FIGIIRK 13.21 Wiring diagram of simple fluorescent light circuit, including an inductive ballast, a
capacitor, and a starter within which is a neon bulb containing a bimctallic switch.
Suppose L = 0 .8 H , C = 1 nF, and = 0.1 A. For r > 0, we find the com ponent o f due
to the initial inductor current, i.e., the zero-input response. The other com ponent, the zero-state
response, is not as important for ignition purposes. O ur strategy will be to use the ^-domain equiv
alent circuit for L, as illustrated in Figure 13 .2 2 .
0 .8 s I,(S)
Voltage Li,(0)
due to = 0.08
intial
inductor
current • u- u •*
High resistance
'.................................................... prior to
arcing
FIGUKI! 13.22 Equivalent complex-frequenc)' domain circuit immediately
prior to arcing and normal lamp operation in fluorescent lighting.
Chapter 13 • Laplacc Transform Analysis II: Circuit Applications 625
Since we are assuming that all resistances are negligible and that the internal resistance (between
electrodes) o f the fluorescent lamp prior to arcing approximates infinity, voltage division in terms
o f impedances yields
— + Ls ^ r + — J l .25x10^ r + 1 .2 5 x 1 0
Cs LC
J l . 2 5 x lo '
2 ,8 2 8 ^ o
. r + 1.25 X 10^
= - 2 ,8 2 8 sin (3 5 ,3 5 5 /) V
which is sufficiently high to induce arcing and cause the fluorescent lamp to operate.
See the homework exercises for an extension o f this analysis to the case where the ballast model
includes a resistance o f 100 Q.
Besides impedances and admittances, other quantities such as voltage gains and current gains are
critically im portant in amplifiers and other circuits. T he term transfer fu nction is a catchall
phrase for the different ratios that might be o f interest in circuit analysis. Impedances and admit
tances are special cases o f the transfer function concept.
TRANSFER FUNCTION
Suppose a circuit has only one active independent source and only one designated response
signal. Suppose fiirther that there is no internal stored energy at f = 0~. T he transfer func
tion o f such a circuit or system is
Thus if the input and the response is^(f), then y(^) = which is a handy for
mula for computing responses. Notice that if the input is the delta function, then F{s) = 1
and Y{s) = H{s). This means that the transfer function is the Laplace transform o f the so-
called im pulse response o f the circuit, i.e., the response due to an impulse applied at the
circuit input source when there are no initial conditions present. The idea is easily extend
ed to multiple inputs and multiple outputs to form a transfer function matrix. This exten
sion, however, is beyond the scope o f this text.
626 Chapter 13 • Laplacc Transform Analysis II: Circuit Applications
A transfer function, as defined by equation 1 3 .2 0 , has broad applicability to electrical and elec-
tro-mechanical systems. For example, the designated output may be a torque while the input
might be voltage. However, in the context o f circuits, a transfer function is often called a netiuork
function. T he literature distinguishes four special cases: (i) driving point impedance, where the
input is a current source and the output is the voltage across the current source; (ii) driving point
admittance, where the input is a voltage source and the output is the current leaving the voltage
source; (iii) transfer impedance, where the input is a current source and the voltage is across a des
ignated pair o f terminals; and (iv) transfer admittance, where the input is a voltage source and the
output is the current through another branch in the circuit. In cases (i) and (iii), the voltage polar
ity must be consistent with the conventional labeling o f sources as set forth in Chapter 2. In gen
eral, however, we will adopt the ordinary language o f transfer function.
EXA M PLE 1 3 . 9 . T he circuit o f Figure 1 3 .2 3 has elements with zero initial conditions at f =
0 “ Find
^out
V :Js)
S o l u t io n
There are many ways to solve this problem. O ur approach is to execute a source transformation
on the R-L impedance in series with the voltage source. After the source transformation, we use
current division to obtain the necessary transfer function.
Step 1. Execute a source transformation to obtain three parallel branches as per Figure 13.24.
Chapter 13 • Laplacc Transform Analysis II: Circuit Applications 627
,(S)
V Js ) r A
sT T
Y,
©
FIG U R E 1 3 .2 5 Parallel admittance form of Figure 13.24.
Step 2 . Use current divisiofi. Since the output current, ^ current through one o f three
parallel branches, the current division formula (equation 13.9) applies, producing
>3(^) M l
y,(A-)+r2(.s') + >3(-^'V 5 + 1
Hence,
Y^is)
H(s) = (1 3 .2 1 )
Vinis) U W + i 2 ( > ^ ) + W / s+ 1
Step 3 . Compute K, (^), Y2 {s), and Y^^is). Because impedances in series add, and admittance is the
inverse o f impedance (equation 13 .7 ), some straightforward algebra yields
2.5s
K,(.v) = ----- K2 (.v) = ------------. = ----- ^3(5) =
S+\ 5+1 0.4 , 2^1
0 .4 5 +
5
Step 4 . Substitute into equation 13.21 to obtain H{s):
2.5 s
5^ +1 1 2.5s
H(s) =
1 2.55
+ - —I— X---- \ 5 + 1/ “ + 1 + 5 ( 5 “ + 1) + .5 5 ( 5 + 1)
5
5 2
5+1 5+1 s +l
2.5s 2.5s
(5 + 1 ) ( 5 - + 2.5s + 1) (s + 1)(^ + 0 . 5 ) ( 5 + 2 )
(i2« Chapter 13 • Laplace Transform Analysis II: Circuit Applications
Exercise. For H{s) as computed in Example 13.9, find the so-called impulse response h{t) =
- -0 .5 / -2 /
A N S W FR : //{/) = --C u(t
3
V,„(s)
S o l u t io n
Since no current enters the inputs o f an ideal op amp, I - p ) = - I p ) . Further, the voltage at the
negative op amp terminal is driven to virtual ground; hence, V-p) = Z -p )I-p ), and Vg^,f{s) =
Z p ) I p ) . Combining these relationships with I - p ) = - I p ) yields
(1 3 .2 2 )
Vi„{s) Zj„{s) Yj{s)
Equation 1 3 .2 2 is a verv' handy formula for computing the transfer functions and responses o f
many op amp circuits.
3. Find the value o f C for which the transfer function o f the op amp circuit in Figure 1 3 .2 7 is
0.2 5 Q
R,
S o l u t io n
Step 1. From the given data, compute the actual transferfunction o f the circuit. By definition o f the
transfer function,
2
L[resp(mse\ Voui{s) (s + l)“ 2
H{s) = ( 1 3 .2 3 )
£\input\ Vi„(s) 1 5+1
.y+1
(i3 0 Chapter 13 • Laplacc Transform Analysis II: Circuit Applications
Step 2 . Using Figure 13.28, fin d the transferfiinction o f the circuit in terms o f R^, Rj, and C. Here,
obsen^e that Figure 1 3 .2 8 has the same topolog)^ as Figure 13 .2 6 , where
Cs + -
Step 3 . Match coejfcients in equatiotis 13.23 and 13.24a to obtain the desired values ofR^, Rj, and
C. Equating the coefficients yields
_1_
Cv+—
Rl
H (s)= - (1 3 .2 4 b )
CRj
other solutions are also possible. For example, for any > 0, ^2new ~ ^new
= represents a valid (theoretical) solution. In Chapter 14 we encounter a concept called
magnitude scaling. is called a magnitude scale factor, which leaves this transfer function
unchanged but produces more realistic values for the circuit elements.
Exercises. 1. In equation 13.24b, it is required that C = 10 uE Find appropriate values o f /?j and R-^.
AN SW ER: R^ = 50 kLl R, = 100 kLl
t
A N SW ER : _JL H,C
tl(l)
Chapter 13 • Laplacc Transform Analysis II: Circuit Applications 631
Suppose a circuit or system lias a transfer function representation H{s), with ^-domain input
denoted by F{s) and j-domain output given by y(y) in which case Y{s) = H W F{s). Assuming that
all initial conditions arc zero, if /(f) = ^{t), then the resulting^(r) is the system im pulse response.
Some simple calculations verify that the transform o f the impulse response is the transfer function,
1.e.,
Hence, the impulse response o f the circuit/system, denoted /;(/“), is the inverse transform o f the
transfer function
Exercises. l.T h e transfer function o f a certain linear network is H{s) = (s + 5)l[{s + 1)(j + 2)]. Find
the impulse response o f the network.
A N SW E R : [lc^‘ -
2. If the impulse response o f a circuit is a pulse ^(f) = u{t) - u {t- T), T > 0, compute the transfer
function.
A N SW ER : (I - e^'^)/s
3. Suppose t/{t) = 2b{r- 1) - 3 6 (/‘- 3) is the input to a relaxed (zero initial conditions) circuit hav
ing an impulse response h{t) = 2u(t) - 2u{t—5). Find the output ^(f).
ANSXXHER: y(r) = 2hU - 1) - 3A(r - 3)
Why is the impulse response important? hs we will see, it is because every linear circuit having con
stant parameter values for its elements can be represented in the time domain by its impulse
response. This is shown in Chapter 15, where we define a mathematical operation called convolu
tion and show that the convolution o f the input function with the impulse response function
yields the zero-state circuit response. In addition to this significant theoretical result, the impulse
response is im portant for identification o f linear circuits or systems having unknown constant
parameters. Sometimes a transfer function is unavailable or a circuit diagram is lost. In such a
predicament, measuring the impulse response on an oscilloscope as the derivative o f the step
response is quite practical.
What is the step response o f a circuit? T he step response is merely the zero-state response o f the cir
cuit to a step function. Observe that if the input f^t) to the circuit is « (/), then F{s) = 1/^ and K(j)
= H[s) (1/^). By the integration propert}’ o f the Laplace transform, it follows that the step response
032 Chapter 13 • Laplacc Transform Analysis II: Circuit Applications
is the integral o f the impulse response. Conversely, the derivative o f the step response is the
impulse response. In lab, many scopes can display the derivative of a trace and hence can display
the derivative o f the step response, which is the impulse response. Alternatively, a homework prob
lem will suggest a means o f directly generating an approximate impulse response.
Exercises. 1. If the transfer function o f a circuit is H{s) = 1/j, what arc the impulse and step
responses?
2. If the Laplace transform o f the step response o f a circuit is given by Y{s) = I/fi'U + I)], what is
the impulse response?
3. If the step response o f a circuit \sy{t) = [1 - 0.5^’“ “^- cos(2r)]//(r), what is the impulse
response?
A N SW ER S: in random order: H/). u(t), cus(2/ + 2 6 .5 7 ‘')//(r).
y(t)
h(t) A
3
A
2 - 2 --
1 1 -
H-----
1 2 1 2 3
(a)
(b)
FIG URE 13.29 (a) Impulse response of hypothetical circuit, (b) Response to 6(r) + b {t- 1).
S o lu t io n
Since X [6(r) + 6(/- 1)] = I + e~\ the response, is simply the sum o f /}{[) and h (t- \)u {t- 1).
Doing the addition graphically yields the waveform o f Figure 13.25b.
S o l u t io n
T he relationship between/( /) and b{t) identifies the strateg)' o f the solution. If the step function
is the integral o f the delta function and the ramp the integral o f the step, then the delta function
equals the second derivative o f the ramp. Hence, b{t) = 0 .1 2 5 /" (r ). By the linearity o f the circuit,
the impulse response h{t) = 0.1 2 5 7 "(^ ), and some straightforward calculations produce
But the right-hand term is zero. (Why?) Hence, y"{t) = [Sf* and
hit) = [e-^-
To see the utility o f this approach, try the alternative method o f computing F(s), V(s), and H(s) =
V(s)/f(s). The algebra is straightforward, but tedious and prone to error.
As a final example, we compute a circuits step response and verify that its derivative is the impulse
response.
/m
R = 40 L=1H +
v . » Q
C = 0 .2 F
-o
S o l u t io n
Cs 1 1 5
(1 3 .2 6 )
/^ + L .V + — s~ -\ --s+ — f.v + 2 ) “ + l
Cs L LC
His) 5 1 -.V - 4
= - +—
.y (.y + 2 ) “ + l l (.v + 2)^ + l
.V (5 + 2 ) - + 1 {s + 2)-+\
= 5e sin{t)u{t)
Thus
5
£
Cv + 2 ) - + 1
in which case
h{{) = 5^’" “^sin(r)«(f)
as expected.
This section develops ^-domain formulations of node and loop analysis. Nodal analysis o f circuits
builds around KCL, whereas mesh/loop analysis utilizes KVL. In Chapter 3 and, indeed, in most
beginning courses on circuits, loop and nodal analysis are taught first in the context o f resistanc
es and conductances and then (in Chapter 10 here) in the phasor context. Recall that KCL
requires that the sum o f the currents leaving any circuit node be zero. Further, KVL requires that
the voltages around any loop of a circuit sum to zero. By linearity, the Laplace transform o f a sum
is the sum o f the individual Laplace transforms. Hence, a KVL equation and a KCL equation have
an j-domain formulation where elements are characterized by impedances and/or admittances.
For loop analysis, one writes a KVL equation for each loop in terms o f the transformed loop cur
rents and element impedances. The set o f all such equations, then, characterizes the circuit’s loop
currents, which determine the ciu rents through each o f the elements. Knowledge o f the loop cur
rents and the element impedances permits the computation o f any o f the element voltages.
In nodal analysis, one writes a KCL equation at each node in terms o f the Laplace transform o f
the node voltages with respect to a reference, the transform of the independent excitations, and
the element admittances. The set o f all such equations characterizes the node voltages o f the cir
cuit in the ^-domain. Solving the set o f circuit node equations yields the set o f transformed node
voltages. Knowledge o f these permits the computation o f any o f the element voltages. W ith
knowledge of the element admittances, one may com pute all o f the element currents. Since nodal
analysis has a more extensive application than loop analysis, our focus will be on nodal analysis.
E X A M P L E 1 3 .1 5 . Figure 13.31 shows an ideal operational amplifier circuit called the Sallen and
Key normalized low-pass Butterworth filter. (See Chapter 19 for a full discussion o f filters.) A nor
malized low-pass filter passes frequencies below 1 rad/sec and attenuates higher frequencies. As we
will see later in the text, the 1-rad/sec frequenc)' “cu toff” can be changed to any desired value by
frequency-scaling the parameter values o f the circuit. (See Chapter 14 for a discussion o f frequen-
Chapter 13 • Laplace Transform Analysis II: Circuit Applications 63 S
cy scaling.) The goal here is to utilize the techniques o f nodal analysis to compute the (normal
ized) transfer function o f this circuit.
S o l u t io n
T he solution proceeds in several steps that utilize nodal analysis techniques in conjunction with
the properties o f an ideal op amp. Recall that for an ideal op amp, the voltage across the input ter
minals is zero and the current into any o f the input terminals is also zero. Finally, note that one
does not write a node equation at the output, which appears across a dependent voltage source
whose value depends on other voltages in the circuit.
Step 1. Find Vy. Because the voltage across the input terminals o f an ideal operational amplifier
is zero.
Step 2. 'Write a node equation at the node identified by the node voltage V^. Summing the currents
leaving the node yields
Step 3. Write a node equation at the node identified by the node voltage Vy. By inspection, the desired
node equation is
5+1 (1 3 .2 9 )
Step 4 . Write the foregoittg tivo node equations in matrix form. In matrix form, equations 1 3 .2 8 and
13.29 combine to give
- (V 2 . + I)'
■ K; ■ '^in (1 3 .3 0 )
-1 -|=5+ 1 0
IV 2 J
6.U> Chapter 13 • Laplacc Transform Analysis II: Circuit Applications
Step 5. Solve equation 13.30 for in terms ofV-^^ using Cramer's rule. From Cram ers rule,
'{ ■ J lS + l ) V;„
del
-1 0
N otice that for small values o f ^ = yto (i.e., low frequencies), the magnitude o f H{s) approximates
1, and for large values o f s = JiO (i.e., high frequencies, where |/b)| » 1), the magnitude o f H(s) is
small. Since = Myo)) ^^^(yco), such a circuit blocks high-frequency input excitations and
passes low-frequency input excitations. As mentioned at the beginning o f the example, the circuit
passes low frequencies and attenuates high frequencies.
T he preceding example used matrix notation, com m on to much o f advanced circuit analysis. In
one sense, matrix notation is a shorthand way of writing n simultaneous equations: the n variables
are written only once. More generally, matrix notation and the associated matrix arithmetic allow
engineers to handle and solve large numbers o f equations in numerically efficient ways. Further,
the theory o f matrices allows one to develop insights into large circuits that would otherwise
remain hidden. Hence, many o f the examples that follow will utilize the elementary properties o f
matrix arithmetic.
T he next example uses nodal analysis to compute the response to an initialized circuit. The exam
ple combines the equivalent circuits for initialized capacitors and inductors with the technique o f
nodal analysis.
E X A M P L E 1 3 .1 6 . In the circuit o f Figure 13.3 2, suppose /y^^(r) = 6(r), /^(O") = 1 A, and v^Qi~)
= 1 V. Find the voltages V(^t) and v^it).
Step 1. Draw the s-domain ecjuivaletit circuit with an eye toward nodal analysis. Inserting the equiv
alent current source models for the initialized capacitor and inductor in Figure 1 3 .3 0 , one obtains
the 5-domain equivalent circuit shown in Figure 13 .3 3 .
V Js) 1o VJs)
,Cv,(0 ) ii(O-)
O
_
= 1
1o 1F
1H
Step 2 . Write two node equations and put in matrix fonn. At the node labeled V(is) K C L implies
that
.9+1 1
-V c(^ ) + — = —
5 S
.v + 2 -1 2 1
, .9+1 _i
s
Step 3 . Solve the matiix equation o f step 2 for the desired voltages. Using C ram ers rule, computing
the inverse, or simultaneously solving the equations gives
2{s+\)-\
[5 +1 1 1 ■2 ■ (s+Vr + \
Vcis)- 2
s (1 3 .3 1 )
_i
5“+ 2 . 9 + 2 1 5+2 s C v + D -3
(.V+ 1r +1
Step 4 , Take the inverse Laplace tratisfonn to obtain time domain voltages. Breaking up equation
13.31 into its components yields
638 Chapter 13 • L iplacc Transform Analysis II: Circuit Applications
2(^ + 1)________ 1
Vcis) =
(5 + 1)- + ! (5+ 1)^ + 1
in which case
V(^t) = e ^[2 cos(t) - sin(/)]/^(r)
Also,
(.^ + 1) 3
V^(s) =
(A-+1)^ + 1 (.V+1)^ + 1
leading to
=e ^[cos(^) - 3sin(f)]«(r)
H G U R E 13.34 Plots of the capacitor and inductor voltages for Example 13.16.
Dual to nodal analysis is loop analysis. In loop analysis, one defines loop currents and writes KVL
equations in terms o f these loop currents. The following example illustrates the method o f loop
analysis for computing the input impedance o f a bridged-T network.
Chapter 13 • Laplace Transform Analysis 11: Circuit Applications 639
E X A M P L E 1 3 .1 7 . Use loop analysis to com pute the input Impedance o f the bridged-T network
illustrated in Figure 13 .3 5 .
/YY\ 1 H
20 2Q
20
0.25 F
Z Js)
S o l u t io n
Step 1. Smn the voltages around loop 1. Dropping the specific j-dependence for convenience, one
obtains
2(/, - / 3 ) + - ( / , - / 2 ) = 2 — - - / 2 - 2 /3 =
s s s
- ( / 2 - / | ) + 2( / , - / 3 ) + 2 / 2 = - - / , + 4 — / 2 - 2 /3 = 0
.V “ s s
Step 4 . Put the three loop equations in matrix form, and solve. In matrix form, the three loop equa
tions are
,£ ± 2 .1
/l1
.1 4— -2 0
s s
-2 -2 5+ 4 h 0
(viO Chapter 13 * I.aplacc Transform Analysis II: Circuit Applications
Using Cram ers rule to solve this equation for in terms o f = / ] W yields
.v“ + 4.V + 4
7 =2 Q
T + 25 0 -2 s 1
\h] -Vs:
4
0 2.V + - + 2 _2 -1 h 0
s
h 0
-2 i' _2 25 + 4 0
V 0
-I 1 -4 0
M GURE 13.36
Switches control lighting systems, furnaces, car ignitions, traffic lights, and numerous other
devices. Switching also takes place inside electronic circuits, as in switched power supplies and
switched capacitor filters. All have a functional element called a switch that affects and, indeed,
shapes the behavior o f the circuit. T he switch inside an electronic circuit is a special device that
we will model simply as an ideal on /off switch. This section investigates the behavior of switching
Chapter 13 • Laplace Transform Analysis II: Circuit Applications 64
in simple RLC circuits, as preparation for the understanding o f switching in more elaborate and
sophisticated electronic circuits. O ur immediate task is to apply the Laplace transform method to
com pute the responses of switched RLC circuits. The following example motivates a general pro
cedure.
+
R
v,(t)
v jt)
S o l u t io n
Because o f the switching at r = 1 sec, the first step is to determine V(-{t) over the time interval 0 s
f < 1, This allows us, in turn, to find which will serve as the initial condition over the time
interval 1 s r < 2. This then produces (2“), the initial condition for the interval 2<. t <A, etc.
Step 1. Compute the response for 0 ^ t < 1. Over the interval 0 :s r < 1 , the circuit o f Figure 13.37
is the simple RC circuit o f Figure 13.38.
20 0 +
V,(s)
0.2 5 F
Step 2 . Coynpute the response over 1 <. t < 2. After the switch moves ft-om position A to position B,
the source is decoupled from the right half o f the circuit; the response then depends only on the
initial condition at r = 1 “ , i.e., = 1 .7 2 2 V. T he goal is to compute v^^t) over the interval 1
^ f< 2 , or equivalently, over the inten'al 0 s r' < 1 where t' = t - \. The ^-domain equivalent cir
cuit that models the behavior o f the time domain circuit o f Figure 1 3 .3 7 over \ ^ t < 2 has the
form illustrated in Figure 13 .3 9 . We note that the value on the current source is Cvq (1“) =
0 .4 3 0 5 .
FIG U R E 1 3 . 3 9 Equivalent circuit of Figure 13.37 for the time interval 1 ^ t <2.
4 1 72^
Vc(.s) = X 0 .4 3 0 5 3 = — ^
s +1 .V + 1
For 0 s ^' < 1 , taking the inverse transform yields
= \.722e-^'ti{t)
or equivalently, for 1 ^ r< 2 ,
V(^t) = \.722e-^^-^\i{t- 1)
W e emphasize that this last equation is valid only for 1 :£ r < 2 . We note that (2') = 1 .7 2 2 e ‘ ^ =
0 .6 3 3 5 .
Step 3. Compute v(it) fo r 2 ^ t <4. For this interv'al, the capacitor is initialized at /' = t - 2 = 0.
Using the parallel current-source equivalent circuit, we note that (r') q = (2') =
0 .1584, as shown in Figure 13.40. We further note that
8 .1 8 7 3
5 + 0.1
Chapter 13 • Laplacc Transform Analysis II; Circuit Applications 643
4
____s 0 .1 5 8 4 0 . 2 x 8 .1 8 7 3 0 .6 3 3 5
+•
0.25.S + 0 .0 5 U + 0 .2 )(5 + 0 .1 ) 5 + 0 .2
20 + -
.V
Step 4 . Compute the response over 4 ^ t. After the switch moves from position A to position B,
again, the source is again decoupled from the right half o f the circuit. According to Chapter 8, we
can write the answer by inspection: the solution is simply v^^t') = 2.8548^ “^ n{t’). Equivalently,
f o r / > 4 , y J r ) = 2 .8 5 4 8 f - ( '- ^ \
Step 5. Combine results into a single expression and plot. I'h e combined expressions for V({i) arc
Os/ <l
2.85486^
FIGURK 13.41 Plot of capacitor voltage for switched circuit of Figure 13.37.
T he extension o f this method to more than three switching times is straightfoi^vard. Although the
preceding example uses an RLC circuit, the same strategy or algorithm is applicable to the calcu
lation o f switching transients in any linear dynamic circuit. T he following is a summary o f the
general procedure.
(2) Determine the proper form o f the input excitation(s) (if there are any), in
terms of t ’.
(3) I'incl the response by the Laplace transform method. Note that the time vari
able is t' . Then obtain the solution in t by substituting { t - t'^ for t .
(4) If / = w, stop. Otherwise, eN-aluate the capacitor voltages and inductor currents at
Note: In some situations, the first switching occurs at r = 0, but the capacitor voltages and induc
tor currents at r = 0 “ are not given. Instead, the problem specifies that dc and sinusoidal sources
have excited the circuit for a long time. If the network is passive— i.e., if it consists o f inductors,
capacitors, or lossy elcm ents-then the circuit will have rcached a steady state at t = 0 “ . The proce
dure then is first to find the steady-state solution and then to evaluate the capacitor voltages and
inductor currents at / = 0 “ . It is instructive to review the dc and sinusoidal steady-state (phasor)
analysis methods studied in a first course. Recall that under certain stabilit)' conditions (to be stud
ied in Chapter 15) on the network:
1. For dc steady-state analysis, open-circuit all capacitances and short-circuit all inductors
to find the steady-state voltages and currents.
2. For sinusoidal steady-state analysis, use the phasor method to find the steady-state
responses.
In addition to its many uses already described, the Laplace transform method is applicable to a
special class o f circuits called switched capacitor (abbreviated SC) networks. These circuits contain
only capacitors, switches, independent voltage sources, and possibly some operational amplifiers.
No resistors or inductors are present. One can dispense with resistors because it is possible to
approximate the effect o f a resistor with rwo switches and a capacitor. Similarly, inductors can be
approximated by circuits containing only switches, capacitors, and operational amplifiers. These
facts, coupled with the easy and relatively inexpensive fabrication ol switches, capacitors, and
operational amplifiers in M OS (metal-oxide semiconductor) technolog}^ have made switched
capacitor filters an attractive alternative to classical filters. Given this scenario, the purpose of this
section is to lay a foundation (i.e., introduce the principles) upon which switched capacitor cir
cuit design builds. More advanced courses delve into the actual analysis and design o f real-world
switched capacitor circuits.
Besides the Laplace transform approach, an alternative method for analyzing switched capacitor
networks builds on the principle o f conserv'ation o f charge.
Principle o f conservation o f charge: T he total charge transferred into a junction (or our o f
a junction) o f a circuit at any time is zero.
646 Chapter 13 • Laplace Transform Analysis II: Circuit Applications
This principle is a direct consequence o f KirchhofFs current law. For example, in Figure 1 3 .4 2 ,
KCL implies that
I,
Since charge is the integral o f current over a time interval, integrating both sides o f equation 1 3 .3 2
from -00 to t yields
f li,(r) + i 2 ( 0 + h ( 0 + U U m = 0
— X
or equivalently.
^i(^) + ^2^^) + ^3(^) + ^4(r) = 0
which is just another expression o f the principle o f conservation o f charge. A simple switched
capacitor circuit will now serve as a test bed for comparing the merits o f the foregoing analysis
techniques.
EXA M PLE 1 3 .1 9 . Consider the circuit shown in Figure 13.43a. The switch S is closed at / = 0.
Just before the closing o f S, the initial conditions are known to be = 1 V and = 0.
C om pute the voltages and for ^ > 0.
i(t)
“• ^ ----- -----
+ + +
Vc, - Vc, ^ C2
IF IF r IF IF
---- • -----
(a) t < 0 (b) t > 0
Chapter 13 • Ijp la c e Transform Analysis II: Circuit Applications 64"
l(s)
V,C2
(C )
S O L U T IO N
M ethod 1. Using the parallel current-source model o f a capacitor, we obtain the ^-domain equiv
alent circuit shown in Figure 1 3.43c. By inspection,
0 .5
s s s+ s s
M ethod 2 . Using the series voltage-source model o f a capacitor, we have the ^-domain equivalent
circuit shown in Figure 1 3.43d . Again, by inspection.
1 1
i + i ■' ■' i+ i
S S
M ethod 3 . Conservation-of-charge approach. For r > 0 , the network is shown in Figure 13.43b .
Clearly, After S is closed, some charge is transferred from C, to C2 . However,
according to the principle o f conservation o f charge, the total charge transferred out o f the junc
tion must be zero. Hence,
and
Solving these two equations for the two unknowns, and results in
Since there is no external input applied, the voltages remain constant once the equilibrium con
dition has been reached. Therefore,
(vi.S Clhapccr 13 * I.aplacc Transform Analysis il; Circuit Applications
for r > 0. For the specific capacitance values given in Figure 1 3.43a, we obtain =
0 .5 V for r > 0.
Computationally, the Laplace transform method is more straightforward. On the other hand, the
conservation-of-charge method is more basic and often provides better insight into what happens
to the charges stored in various capacitors. It is particularly useful for the purpose o f checking
answers obtained by other methods: the answers are correct when the conser\'ation-of-charge con
dition is met at every node.
lO V
S o l u t io n
We first construct the.f-domain equivalent circuit using admittances. The result is shown in Figure 13.45.
4 15
4s 3s
+ A IQ
fO 2s 12
s
5 10
4s + 3.V + I s Vt. = 4 + 1 5 + 12
from which wc derive = 9/s and = 9 V for r > 0. Obviously, Kj = 5 V and K2 =10 V. To
verify that = 9 V is indeed the correct solution, we check for conservation o f charge. From t=
0~ to A= O'*’, tiic voltage of the 4 F capacitor jumps from 1 V to 4 V (note that 4 = 9 - 5 ) , indi
cating that 4 X (4 - 1) = 12 coulombs o f charge have been transferred to this capacitor. The volt
age o f the 2 F capacitor jumps from 6 V to 9 V, indicating that 2 x (9 - 6) = 6 coulombs o f charge
have been transferred here. Finally, the voltage o f the 3 F capacitor changes from 5 V to - 1 V (note
that - 1 = 9 - 1 0 ) , indicating that 3 x (-1 — 5) = - 1 8 coulombs of charge have been transferred to
this capacitor. As a check for conservation o f charge, we have 12 + 6 + ( - 1 8 ) = 0, and the solution
is assured to be correct.
T he preceding examples considered idealized circuits, i.e., no resistances were present. In zny prac
tical circuit, the connecting wires have some resistance. W hat is our interest in the analysis of an
idealized circuit? T he analysis o f an idealized circuit is much more straightforward than that of a
realistic circuit yet provides relatively accurate answers. As a case in point, reconsider Example
1 3 .2 0 . Suppose we insert a 0.1 Q resistance in series with ever)' capacitor. The resulting transform
analysis would produce a rational function with a cubic denominator polynomial (a third-order
network) whose factorization would require the use of a root-finding program. In sharp contrast,
the idealized circuit o f Example 13.20 was anaK'zed by writing a single first-order node equation,
making a partial fraction expansion unnecessary.
Idealizations o f circuit models sometimes lead to phenomena that defy intuitive explanations. An
interesting case is given by Example 13.20. Before S is closed, the energ)' stored in the electric field
is 0 .5 X 1^ + 0 .5 x 0 “ = 0.5 joule. After S is closed, the stored energ)' becomes 0 .5 x (0 .5 )“ + 0.5 x
(0 .5 )“ = 0 .2 5 joule. Apparently, 0 .2 5 joule o f energy has been lost. Since there is no resistance in the
circuit to dissipate the energ)', what accounts for the lost energ}'? Is energ)' not conser\'ed?
Note that for idealized SC circuits, as long as the independent voltage sources arc piecewise con
stant, all capacitor voltages are piecewise constant and all currents in the circuit are impulses. These
650 Chapter 13 * Laplace Transform Analysis II: Circuit Applications
properties remain valid for more general idealized SC circuits that allow the inclusion o f VCVSs,
C C C Ss, and ideal op amps. T he reason is that the parameters characterizing these components are
dimensionless and hence do not result in a time constant. All voltage changes are instantaneous.
On the other hand, if the circuit contains resistances, we have a lossy switched capacitor circuit,
whose voltages arc no longer piecewise constant. The transient analysis o f a lossy SC circuit
requires the usual Laplace transform analysis.
One reason for our interest in SC circuits is that a SC combination can be used to approximate a
resistor. As a result, any RC-op amp circuit used for signal processing can be approximated by an
S C -o p amp circuit. A study o f the general theor)' o f such S C -o p amp circuits is beyond the level of
this book. We shall merely illustrate the approximation property with a simple integrator circuit.
E X A M P L E 13.21. Consider the RC-op amp integrator circuit shown in Figure 13.46.
R = 5kQ C,= 1 mF
Using the result o f Example 13.5 and the integration property o f the Laplace transform, it is
straightforward to show that
If the input is a constant voltage v{ = £ = 5 V, and if = 0, then the output waveform is a ramp
function vj,t) = - l ,0 0 0 ///( f ) (as long as the output has not reached the saturation level), as shown
in Figure 13.47.
t (msec)
- 10-
1. At r = 0, S is at position a.
2. At t = T, S is moved to position b.
3. At r = 27", S is moved to position a.
4. At / = 37", S is moved to position b, etc.
T he output waveform may be determined ver)' easily by the principle o f conser\'ation o f charge as
follows: for 0 :s / < T, y. = £,(]= CE, qj-= 0, and = 0. At / = f , switch S is moved to position
b. Because the op amp is assumed to be ideal, the voltage across the input terminals is zero, and
so is v^. Thus, C cannot store any charge. T he charge CE previously stored on C must be trans
ferred out o f C. Since the op amp is ideal, the input impedance is infinity and the input current
is zero. Therefore, none o f the charge can flow into the op amp. Instead, the charge must be trans
ferred to the capacitor This leads to CE and Vgi'T') = - C £/C ^
At r = 27', switch S is moved back to position a, causing C to be charged to E volts again. Since
the charge “trapped” on Cj-, the output voltage remains unchanged until S is moved to posi
tion b again. At that time, another CE coulombs o f charge are transferred to Cp and is incre
mented by -C EIC p Subsequent switching is similar.
To make the output waveforms o f the circuits o f Figures 1 3 .4 6 and 13.48 approximately the same,
the average o f the charges transferred to (y m u s t be the same in both cases. For Figure 1 3 .4 6 , the
current flowing into C^is at a constant value o f E/R. Therefore, every I T sec, the charge trans
ferred to C^is equal to 2TEIR. On the other hand, for Figure 1 3 .4 8 , the charge transferred to Cj-
every I T sec is CE. Equating these rwo quantities, we have CE = 2TEIR, or RC = IT . Thus, there
is no unique combination o f C and T that will produce the approximate effect o f a resistance. A
smaller T (i.e., a higher operating frequency o f the switch) in Figure 13.48 produces a staircase
output waveform that closely “hugs” the ramp output o f Figure 1 3 .4 7 . For the purpose o f com
parison, the output waveform corresponding to 7 = 1 msec and C = 0 .4 pF is shown in Figure
1 3 .4 9 , together with the ramp output from the RC-op amp integrator. It is worthwhile to note
652 Chapter 13 • Laplace Transform Analysis II: Circuit Applications
that each o f the circuits o f Figures 1 3 .4 6 and 1 3.48 drains the same average amount o f charge
from the voltage source and puts the same average amount of charge on the capacitor Cj; The only
difference is that in the former the process is coutinuous, whereas in the latter the process occurs
in quantized steps.
2 3 4 5 6 7 8 9 19
-I— I— I— I— I— I— I— I— I— I-
t (msec)
Output of
switched capacitor
circuit
Exercise. Plot the output waveform of the SC circuit o f Figure 1 3 .4 8 for 0 s ^ 2 msec if 7'= 0.1
msec and C = 0 .0 4 pF. Also, plot the ramp output on the same graph for comparison.
As recently as two decades ago, switched capacitor circuits were considered impractical. No longer!
Advances in semiconductor technology have made the fabrication o f high-speed electronic switch
es and op amps as cheap as resistors. Furthermore, large numbers of switches, capacitors, and op
amps can be fabricated on a single chip. Consequently, switched capacitor circuits hold an impor
tant place in future signal-processing applications. Although we cannot delve into the practical
aspects o f the design of such circuits, we have at least outlined the basic principles needed for their
approximate or exact analysis.
In the design o f active filters and in the design of data acquisition equipment, general summing
integrators play an important role. This section takes up the design o f such op amp circuits from
the transfer function perspective. However, to simplify the presentation, we consider only multi
input transfer functions o f the form
(1 3 .3 3 )
*oiit ( '0 -
s s s s
Chapter 13 • Laplace Transform Analysis II: Circuit Applications 653
where the a- and are positive gain constants and .... are ^-domain inputs. This
expression represents the ^-domain equivalent o f the following time domain equation assuming
zero initial conditions at ; = 0 “ on the variables:
The first two design steps constitute a preliminary or protorvpe design, meaning that the two
capacitors are normalized to 1 F. After completing the prototype design, an engineer would scale
the capacitances and resistances to more practical values without changing the gain characteristics.
T he scaling procedure is explained in step 3.
Design Step 2. Prototype design [continued): Computation o f or AG^ such that the total admit
tance incident at the inverting terminal o f the op amp equals the total admittance incident at the non
inverting terminal. To achieve this equality, recall that in design step I > - «1 S, - ” 2 ^b\
= h, S. G ,, = S. Define a numerical quantity 6 = + a-^) - [b^ + b.^)
6S4 Chapter 13 • Laplacc Transform Analysis II; C ircuit Applications
D esign Step 3 . Scaling to achieve practical element values. Multiply all the admittances incident at
the inverting input terminal o f the op amp by a constant K^. Similarly, multiply all admittances
incident at the non-inverting terminal o f the op amp by /Q,. It is possible to choose = /Q,, but
this is not necessary.
(1 3 .3 4 )
S o l u t io n
Step 1. Prototype design. Using Figure 1 3 .5 0 , choose C = 1 F. W ith all admittances in S, set
Step 3 . Scaling. To have practical element values, let us choose K^ = K y- 10“^. This scaling leads
to a design with C = 10 j_iF and resistances = 1 4 .2 8 kQ, R^2 = 3 3 .3 3 kQ, /?^, = 50 kQ, R^2 =
25 kQ, and A/?^ = 25 kQ.
75
Chapter 13 • Laplace Transform Analysis II: Circuit Applications 6SS
1 4 .2 8 kQ
lO p F
a2 '
3 3 .3 3 kO
+
5 0 kO
b2
o—
2 5 kO 2 5 kO- I O mF
(b)
FIGURE 13.51 (^) Prototype design of equation 13.34. (b) Final design after scaling with /T = AT^ = 10~5.
s s s s
is given by the circuit of Figure 13.52.
2S 6S
M GURE 13.52
Referring to Figure 13.50, from the properties of an ideal op amp, V^= V_ and no current enters
the inverting and non-inverting op amp input terminals.
Derivation Step 1. Write a node equation at the non-inverting input terminal of the op amp.
Summing the currents leaving the + node of the op amp yields
W b2
(13.35)
Derivation Step 2. Write a node equation at the inverting input terminal of the op amp. Recall that
V_ = Vy Summing the currents leaving the inverting node yields
- V^2) - . AG^V^ = 0
Thus,
1
/ _ ^ a l T, G a2 y I I ^ a\ + ^ a l + "^ 7 )
If we choose and AG^ to make AG^ + ^a\ + ^al = + ^b\ + ^bV ‘-e- if the total admit
tance incident on the inverting terminal is made equal to the total admittance incident on the
non-inverting terminal, then equation 13.37 simplifies to
I sC sC (13.38)
If we let C = 1 F (to be scaled later to a practical value), equation 13.38 further simplifies to
Equation 13.39 shows that the circuit of Figure 13.50 is a general summing integrating circuit
whose gains are proportional to the admittances G^- and/or % The sign o f each gain depends on
whether the corresponding input is connected to the inverting or non-inverting terminal o f the
op amp. This completes the derivation of the input-output characteristic of the op amp circuit and
is the basis for the prototype design. The extension to more than four inputs is straightforward.
Chapter 13 * Laplacc Transform Analysis 11: Circuit Applications 6^
It remains to justify the scaling o f step 3 in the design procedure. In step 3 , it is stated that all
admittances incident at the inverting input terminal o f the op amp can be multiplied by a (scal
ing) constant K^. Similarly, all admittances incident at the non-inverting terminal may be multi
plied by a (scaling) constant /Q,. It is possible to choose = /Q,, but, again, this is not necessar}^
The verification that these multiplications will not change the input-output characteristic follows
directly from equation 13 .3 7 . After inserting the scale factors, one immediately sees that they can
cel and have no effect on the overall gain.
11. SUMMARY
This chapter has presented the basic principles and techniques o f circuit analysis in the j-domain.
Impedance, admittance, Thevenin equivalents, superposition, linearity, voltage division, current
division, source transformations, and nodal and loop analysis have ^-domain forms that allow the
analysis o f complex circuits excited by a variety o f waveform types. Indeed, the simple starter cir
cuit o f a fluorescent light points to the usefulness o f Laplace transform methods in the analysis
and design o f simple, everyday electrical conveniences. Various op amp applications were also pre
sented. As subsequent chapters will illustrate, complex circuits and advanced analysis methods
build on these basic principles and techniques.
T he chapter also introduced the notion o f switched capacitor circuits. Integrated circuit technol
og)' has made such circuits easy and inexpensive to produce. Applications include speech process
ing and other t\'pes o f signal processing. Although a full-scale analysis o f such circuits is beyond
the scope o f this text, the basic principles o f their operation arc presented as a foundation for more
advanced analysis tools.
Finally the chapter introduced a general method for the design o f multi-input integrators having
both positive and negative gains. Such circuits can be used for implementing active filters and for
implementing controllers in practical situations.
65<S Chapter 13 * Lnplacc Transform Analysis II: Circuit Applications
Admittance: the ratio o f the Laplace transform o f the input current to the Laplace transform o f
the input voltage with the two-terminal network initially relaxed.
Admittance o f capacitor: the quantity Cs.
Admittance o f inductor: the quantity MLs.
Admittance o f resistor: the quantity MR.
Current division: a formula for determining how currents distribute through a set o f parallel
admittances.
Impedance: the ratio o f the Laplace transform o f the input voltage to the Laplace transform o f
the input current with the two-terminal network initially relaxed.
Impedance o f capacitor: the quantity MCs.
Impedance o f inductor: the quantit}' Ls.
Impedance o f resistor; the quantity R.
Impulse response: the response o f a circuit having a single input excitation o f a unit impulse;
equal to the inverse Laplace transform o f the transfer function.
Integrator circuit: usually an op amp circuit with transfer function K/s.
Source transformation property: in the ^-domain, voltage sources in series with an impedance
are equivalent to the same impedance in parallel with a current source whose value equals
the transform voltage divided by the series impedance.
Steady-state analysis: analysis o f circuit behavior resulting after excitations have been on for a
long time; often refers to finding the sinusoidal or constant parts o f the response when
the circuit is excited by sinusoids or dc.
Switched capacitor circuit: a circuit containing switches, capacitors, independent voltage
sources, and possibly op amps, but no resistors or inductors.
Time-invariant device: a device whose characteristics do not change with time.
Transfer function; the ratio o f the Laplace transform o f the output quantity to the Laplace trans
form o f the input quantity with the network initially relaxed.
Transient analysis: analysis o f circuit behavior for a period o f time immediately after independ
ent sources have been turned on.
Voltage division: a formula for determining how voltages distribute around a series connection
o f impedances.
Chapter 13 • Laplace Transform Analysis II: Circuit Applications 6S9
Problems R
IMPEDANCE, ADMITTANCE, 6
VOLTAGE DIVISION,
CURRENT DIVISION, SOURCE ZJs)
(a)
TRANSFORMATIONS,
/Y Y V ^ I
THEVENIN AND NORTON L
EQUIVALENTS
1. Consider the circuits in Figure P I 3 .1 .
= (5(r), then
5 + 40
2.5 0
0.2 F
(a) C om pute Y-^^{s) and Z-^^{s).
■6 1H
(b) If = 10u{t) V, find for t>Q.
(c) If v jt ) = 20e-^^^u{t) V, find i j t ) for
r > 0.
(a) (d) If v jt ) = 20e~^K(t) V, find ijt) for
0.25 H 0.
J T Y V _ ijt)
0.25 F
v.(t) I Z(s)
20
Figure P I3.3
(b,
Figure P I3.1
4. For each o f the circuits in Figure PI 3 .4 , find
C H E C K S : Figure P I 3 . la:
expressions for
+5
y,„(5) = 0 . 4 --------- ^ F i g u r e P I 3 . lb: (a) Z Js) and YJs)
( .y + l ) -+ 2 " and (as indicated)
J" + 1 6
(0
r\
Figure P I3.4
r\
5. In the circuit o f Figure P13.5, suppose
) = 0. Find . Z.„W. and i^^t)
when (a) = 5{t) and (b) = n\
50(1 - mA.
Z.(s)
Y Js)
o
Figure P I3.8
JL R. i
R +
V^(t)
(a)
20- 5-10
C H EC K : Z;„(5) = - 1 0 0 o
10- 5 + 10
(b) Z,(s)
Figure P I3.1 0
— — —
z (S)
11. Find theThevenin equivalent impedance o f Z,(s)
each circuit in Figure P 1 3 .1 1 . Flint: Label the
terminals and an input current to the top
terminal as determine or
Figure P I3 .13
-o — /Y Y V
ANSWF.R: (h) 2 + 2.V + i,(t) R, L +
YJs)
Z,(s)
Figure P i 3 .1 4
Z,(s)
CH ECK: K,„ = 1 0 —
.y“ -h 200.V + ? ? ? ?
(a)
662 Chapter 13 * Liplace Transform Analysis II: Circuit Applications
(h) 5^ + 16 5
n > ,(^ )= •+
45 '0 .2 5 5 ^ + 16
Figure P I3 . 15
TRANSFER FUNCTIONS „ , ,
16. Find an RC circuit for each o f the follow
V ( A")
19. Find the transfer function, H{s) = -----,
ing: impulse response, and step response
(a) K J j) = 0 .b + 0 .1 for each o f the following differential equation
1 models o f a physical process.
(b) Z:,As) = - (a) <’o,a(0+ P\Voui(0= Kvi„U)
’ 0 .2 5 + 0 .0 2 5
(b) iv„,,(/) + 20v,,„ , ( / ) + 100vv,,„(r) = 2 0 0 v v „ (0 + 10v,„(O
(c) Z ,„ ( 5 - ) = 1 0 +
0 .2 5 5 + 0 .2 ('= ) v „ , „ ( / ) + 1 0 v „ „ ,(/ ) + 2 5 ..„ ,„ (/ ) = 2 2 5 0 V ,. , , ( 0 + l O v ^
2.V + 8
(a) Z M =
2 0 . Find the transfer function for the general
s+ 2
4s~ + 36.? + 6 4 second-order differential equation
IUU.V
2 6. (a) Compute the transfer flinctions o f the
.s- + 400
op-amp circuit in Figures P i3 .2 6 a and b.
0
(b) Given your answers to part (a), com
pute the step responses.
- 0 . 1 / „ „ , ( 0 " ) + 4() + 40^^^^ = 1 0 l’„
.V .V (c) Assuming R= 1 kH and C = 0 .5 niF,
2 2 , An integro-difFerential equation for an compute the zero-state response to the
active circuit driven by a current source with C input 2u{t) - 4u (t- 2) V.
= 0 .5 F , Z = 1 H, ^ = 2 n i s (d) For the circuit o f Figure P I3 .2 6 b ,
assuming R= 1 kD and C = 0 .5 mF,
compute the output when the input is
-X
as given in Figure P i3 .2 6 c .
(a) Find the transfer function ^/(.y) = .
(b) Find the impulse response. %
(c) Find the step response. He
c 2R
V „(t) +
2 3. Suppose two signals and y{t) are
related by the equations
(a)
--- r ; — + ^’out ) + 2y(/) = 4 — ^—
dt~ at
He-
and He
(c) Repeat part (b) assuming the initial 29. (a) Find the transfer fimaion of the (ideal)
capacitor voltage is = -4 V. op amp circuit of Figure PI 3.29 in terms
of /?,, /?2, Cp and Cj. Make the leading
- v,(t) +
coefficient of s in the denominator 1.
5+ 2
(b) If it is desired to obtain H(s) = - 4 -------
^^ s+4
with € 2 = 100 jaF, find /?j, Rj, and Cj.
j +2
(c) Given H(s) = - 4 -------find the zero-
^^ s+4
state response to t',„(^)= sin(2r)«(f) V.
v,(t)
->t
(b)
Figure P I 3.2 7
Figure P I3.30
Figure P I3.28
CHECK: -1
Gi + G t H(s) =
s + —^------ ^ C 2R 1
CHECK: H(s) = -------- f — s+ s+
/ ? ,C j R 2C 2
s+
Chapter 13 • Liplacc rransform Analysis II: Circuit Applications 665
3 1. (a) C om pute the transfer functions o f 3 3 . C onstruct the transfer Kinction o f the d r-
each circuit in Figure P I 3.31 in terms cuit o f Figure P 1 3 .3 3 , assuming that all op
of C, R and K. amps are ideal, as follows:
(b) Suppose C = 1 i.i¥, K = 3, and R = R^ . .
1
= 1 k n . Find the zero-state response to (a) Let G| = — , C t = — , G t, =
R R.
2 sin(500r)//(r) for each circuit
in Figure P 1 3 .3 1 .
(c) Repeat part (h) for i’j„{t)=
2 sin(l 500/)//(r). (b) Com pute the ratio
V'i(.v)
IJs)
+ 2,(s)
Z.(s) Y,(s) VJs) Y,(s)
9 jY ❖ < p
Z,(s) V,(s)
Figure P i3.34
Figure P13.36
35. (a) For the circuit in Figure P13.35, find 37 Consider the circuit o f Figure P13.37. Find
die transfer fiinctions die transfer fiinctions
w i “ ■
1
(b) If Z 2 = — ,Z^ = j + 3, Y^=s, Y^=
3. and r ^ = 20 Q, compute the VJ„(j)
i(^)
impulse and step responses. as follows:
(c) Given the values of part (b), compute
(a) Compute the input impedance
the response to the input
_ V^is)
K^u(t) -Kjuit — 7) V using linearity Zin(s) =
and no further calculations, assuming
T>0. (b) Given the results from part (a), use
voltage division to compute
Vln(s)
in terms of Do not substitute
your answer to part (a) into the
obtained expression.
Figure P I3.35 (c) Compute
Figure P I3.37 r\
Chapter 13 * Laplace Transform Analysis II: Circuit Applications 667
K 1^2
.K Vout = det(MM)/det(M)
CHECK: Z ^ = Z i +
% You have now computed the symbolic
n + ^L
% expression for Vout in terms o f lin so you
38. Consider the circuit o f Figure P13.38.
% can now identify the transfer function as a
% symbolic ftinction o f the variable s.
% Now lets do some numerical work.
C l= 1; C2 = 2; G l = 1; R2 = 1; G3 = 2; a = 0 .5 ;
v„Jt)
% Re-enter M above
M = [Cl'*‘s+Gl 0 1; fill in the rest of your co ^ -
G is mhos; R is ohms
cient fnatrix]
% check: the roots o f the determinant should
Figure P I3.38 % be -1 and - 5 .
dt = det(M)
(a) Draw the equivalent frequency factor(dt)
domain circuit assuming t'QCO") ^ 0 % Re-enter MM above
V and ^^2(0") ^ 0 V. % Compute the actual transfer function
(b) Define a current Ij^{s) from left to Vout = det(MM)/det(M)
right through Then write three % Identify the transfer function
(modified) nodal equations for the cir
cuit o f part (a) similar to the text (d) Find the impulse response h(t) of the
example, but accounting for the initial circuit as follows:
conditions. Put equations in matrix % Rewrite the above expression for Vout with
form (unknowns on the left side and % out lin.
knowns, input, and initial conditions syms s t H h
on the right). You should have a 3 by H = -(1/2V1)/????????
3 matrix of literals. The unknowns are h = ilaplace(H)
two (node) voltages and one current,
(e) Now use the MATLAB command
(c) Determine the transfer ftinction H{s) “ilaplace” to compute the step
= y^out^s)/IjJs) of the circuit in MAT- response, again with zero initial condi
LAB using the following code: tions.
X..,(s)
E(s) C = 0.1 F, i^{0~) = 2 A, and v^iO~) = 2 V. Find
4 0 . G,(s) -> Y(s)
V(is) and v^it) using the et]uivalent models for
initialized L and C in the j-domain.
ijt)
Figure P I3.3 9 +
L v,(t)
ch ec k .
l-t-G](5) d{s) + n{s)
4 0. In the so-called feedback configuration of
Figure P I3.41
Figure P i 3 .4 0 , E{s) is the Laplace transform of
the error between the reference signal x^^t) and
4 2 . For the circuit o f Figure P 1 3 .4 2 , /^(0“) = 2
the response i.e., e{t) = - y{t). =
A. Draw an equivalent circuit for the inductor
nj,s)ldjj) is the plant o f the system (described
that accounts for the initial condition, write a
in the previous problem) and F{s) = )tj(s)lci^s) is
single node equation in and then find
a feedback controller to be designed.
(a) Find the transfer
function H{s) = E{s)IX^^,J(s). 50
for f > 0.
(b) I f v^0~) = 10 V and = 200
cos(2f)«(/) mA, find V(^s) and V(4J) Figure P I3.46
for ^ > 0.
47. Consider the circuit of Figure P I3.47, in
which R^=l k£2, R^ = S kH, and C = 50 ^F.
(a) Find the transfer fiinction H{s).
(b) Compute the step and impulse
responses.
(c) Compute the response to v-^(t) =
Figure P I3.44 10r'2-5^cos(25^)«W V.
(d) Find the response to V(iO~) = 20 V.
45. Consider the circuit of Figure P13.45, in (e) Using the principle of superposition,
which R = 200 £2. find the response to the input of pan (c)
(a) Compute the transfer function in with the initial condition o f part (d).
terms o f L. (0 From the principle o f linearity, what
(b) If the response to the input i-JJ) = would the response be to the input
Intuit) A is ijij) = {2.5t - 0.025 + = 20£-“^^-^^cos(12.5/)«(/) V and
O.O25e"^®°0«(^) A, assuming /^(O") = y j^ 0-) = 1 0 V ?
Vw> 0, find the values of L and /q.
(c) Assuming ij{0~) = 1 A and i-JJ) = + 1
2.5 sin(100f)«W A, find
v Jt)
lV (t) I 1
w
Figure P I3.47
w u « (t)
20
C H EC K : H{s) =
2 5 + 25
Figure P I3.45
w
670 Chapter 13 • Laplace Transform Analysis II: Circuit Applications
48. Consider the circuit o f Figure P I3.48, in (c) Find the response i({t) when v-JJ) =
which C = 2.5 mF, = 200 Q, = 50 £2, /?£ 0, V({Q) = 10 V, and /^(O) = 0.
= 5 n , Z = 0 .2 H , and^„ = 0 .5 S .
(a) Find the transfer functions
//2 (5 ) = - ^ ^ , a n d
^ VcisV
Figure P I3.49
W
(b) Given N(s), find the zero-state
50. In the circuit o f Figure P I 3.50, suppose L
response; i.e., assume /^(0“) = 0 and
= 0.04 H. 1
= 0 when = 25(1 - C = - F
V. 9
(c) Find the zero-input response if R^ = 0.5 £2, = 5 £2, C2 = 0.05 F, and a = 4 .
/•^(0) = 6 A a n d t;c(0 ) = -1 0 V . Compute Vciis), assuming diat
(d) Find the complete response (combine v jt ) = 36u{t) V. /^(O-) = 0. v c ^ m = - 1 8 V,
parts (b) and (c)) and identify the and v„ufiO~) = 0 V. (Hint; You must construct
transient and steady-state parts o f it. the equivalent circuit in the j-domain, account
(Note that the transient pan is the ing for initial conditions. Consider a source
part that is not constant or not period transformation on v-J^t), and then draw the
ic. Usually the transient part converges equivalent circuit in the j-domain so that you
to zero.) can combine sources in the front half o f the cir
cuit.)
^ g.Vc(s) 1 1 L
' '1 Y r 1
Figure PI3.50
Figure P I3.48
51. Consider the circuit of Figure P13.51.
49. Consider the circuit of Figure P 13.49. Note Suppose /?j = 2 £2, /?2 = 6 £2, i?3 = 3 £2,
that the computation o f transfer functions pre C = 0.125 F I = 1.6 H.
sumes no initial internal stored energy.
(a) Use a source transformation and the Suppose v-JJ) = yj(/) + 1/2W = -1 0 « (-f) +
current divider formula to show that 10 u (t) V (plot this input function so that you
the transfer function between V-JJ) know what it looks like).
(the input) and I(\s) (the output) is (a) Compute V(iO~) and /^(O").
(b) Compute the “zero-input response”
1
for ^> 0; i.e., assume you have the ini
1 1
■s + tial conditions computed in step (a)
RC LC
and that the voltage v-^(t) = 0 for r > 0.
(b) If Z = 0.2 H, C = 25 mF, R = 4/3 Q, First draw the equivalent j-domain cir
and v-^{t) = \Qe~^^^u{t) V, find /^j) cuit accounting for the initial condi
and i({t). tions.
Chapter 13 • Laplace Transform Analysis II: Circuit Applications 671
(c) Compute the transfer function H{s) of Plot the resulting time function in
the circuit valid for ^ > 0. MATLAB for 0 < r < 1 sec. The partial
(d) Compute the “zero-state response,” fraction expansion is most easily com
i.e., the response to the circuit o f the puted using MATLABs “residue”
input v-JJ) = V2 {t) = 10«(f) assuming command. {Note: Practicing hand cal
all initial conditions are zero. culation is important for the exams.)
(e) Compute the complete response. Now forget about the partial fraction
(0 Identify the transient and steady-state expansion and instead use MATLAB s
parts of the complete response. command “ilaplace” to compute the
time function You should define
H, Vc, s, t, lin, and vc as symbols
using “syms.” The program should be
something like the following:
R = ? ; L = 1 ; C = ?;
syms Yin Zin H Vc s t Is vc
lin = ?
% H will be the transfer function defined in
52. Consider the circuit o f Figure P13.52. terms o f s, a symbol, and R, L, C, and K.
Suppose Z = 1 H, = 15 C = 0.01 F, and % MATLAB will fill in the numbers.
^2 = 5 Q. Suppose is the desired output, H = ????
(a) Find the input admittance Vc = H*Iin
vc = ilaplace(Vc)
n\ ^
Chapter 13 • Laplacc Transform Analysis II: Circuit Applications 673
(c) Now suppose i-p) = 16[u{t) - u {t- T)] (d) If v jt) = l[u{t) - u { t- 0.5)] V, y^(0") =
mA, where 7'= 1 0 msec. Find and plot 5 \/ and /yr(0~) = 0, find
0.3 H 0.1 H
ijt) 0 1 5 0
90 n lO Q
Figure P I3 .56
Figure P I3.58
C H E C K : (a) {s + 100)(5 + 3 0 0 )/[(j + 200)(^ +
4 0 0 )] 59. in the circuit o f Figure P 1 3 .5 9 , /?j = 2 Q,
C, =1 F, /?, = 1.75 Q, and C , = 2 /7 F.
57. Consider the bridged-'F network in Figure (a) Show that the transfer function o f the
P I 3 .5 7 . Let C = 0 .2 5 F. Assuming no initial _ ^C2
circuit in literal form is H (.v) =
conditions, find /j(s ), ^ (s ), and V({s) when
Vj„{t) = 10(1 - e--^ii{t) V. You might use
R\C,
C ram er’s rule and the symbolic toolbox in
M A FLA B to solve the problem. Now find .v +
"JO - U 2C2 ^ ^ «|C|/?2C ,
(b) If = 15«(/) V and the capacitors
are initially relaxed, find V(^[t) for t>
0. Plot the input and the response on
the same graph using M ATLAB or the
equivalent.
(c) if = 0, ^^^(0“ ) = 0 , and i^Q(0~)
Z Js) = 15 V, find for t> 0.
(d) If = 0 , t^<^'i(0“ ) = 0, and /^ ^ (0“ )
Figure P I 3 .5 7 = 15 V, find V(^{t) for r > 0.
(e) If = 15//(/) V, i^Q(0“ ) = 15 V, and
58. Consider the circuit o f Figure P I 3 .5 8 , in ^(^(0“ ) = 15 V, find Vf^{t) for r > 0.
which /?^ = 2 k n , C = 2 0 }d¥, = 20 Q, =
80 n , and L =5 H.
(a) Com pute the transfer function
/V,(.v) =
v;,,(.v) 0
(b) Com pute the transfer function
, Figure P i3.59
Vf(.v)
(c) Com pute the transfer function
A N SW ER S: (b) - ~e^^uu) V; (c,-
y. [^-0.25;^
60. In the circuit o f Figure PI 3 .6 0 , /?, = 0.5 (c) Now suppose ) = 3 V, i^{Q) ) = 0
=1 H, Rj = 1.75 a and Z , = 7 /8 H. A, = ?>u{t) V, and = 5u{t) A.
(a) Show that the transfer function o f the R ecom pute the nodal equations
circuit in literal form is n {s) = ^ = accounting for the initial capacitor
voltage. Then find Vjft).
v,(t) V .(t)
L jG’2 ^2 m v
1 0.5 n -> 0 .1 H
— - + —+ >v+
^C|L| ^2^1/ ( t ) ' . ,(t)
0.8 F 1Q
/Y Y \
—► 6 3 . In the circuit shown in Figure P I 3 .6 3 , i\„U)
= 12 V for r > 0 , i/^ ,(0 ) = 6 V, and z^^,(0) = 2 V.
'• 6 (a) C onstruct the equivalent circuit in the
Laplace transform domain, account
ing for an initializ-ed capacitor.
Figure P i3 .60 (b) Write a nodal equation for the circuit
constructed in part (a).
A N SW FR S: (b) (30 - A: (c) Find l/^ W .
(c) A; (d) + (d) Find for t> 0 .
7e^‘\u{t) .A; (e) sum o f parts(b). (c). and (d).
tance voltages are t/c,(0) = 0, t/Q(0) = 6 V, equations for the circuit of part (a)
*;C3(0) = 2 V . only in terms of the variables V({s)y
(a) Construct the Laplace transform ^2^’ Initisl
domain equivalent circuit. conditions. Simplijy each equation.
(b) Find Vf^{s) in terms o f V^J^s) and the (c) Put equations in matrix form.
initial conditions. (d) Assuming /jr(0“) = 0, (0~) = 5 V,
(c) For the given initial conditions, if W = 5<5(/) A, and (^) = \06{t) V,
v-^{t) = \2u{t) V, find for r > 0. use Cramers rule to find the current
Hint: After obtaining Vq W, use the “residue” (f) and then i^ {t).
conmiand in MATLAB to obtain the partial (e) Now suppose that ii(0~) = 0,
flection expansion. You might also investigate = OV. = OA. and «„(») = 10»(() V.
the use o f the command “ilaplace.” Find 1^2 W-
1 0 I.
65. In the circuit of Figure P I3.65, = 0.25 67. Consider the circuit of Figure PI 3.67. This
S, C2 = 0.25 F, and /? = 1 Q. Write a set o f mod problem is to be solved using (modified) nodal
ified loop equations and solve for the three loop analysis.
currents and V, assuming V(^0~) = 4 V and (a) Draw the equivalent frequency
v jt ) = \2u{t) V. domain circuit assuming /^(0“) = 0
and «'ci(0“) = 2 V.
(b) Write three nodal equations for the
circuit o f part (a) only in terms o f the
voltages Ifis), and the
initial conditions. Simplify your equa
tions.
(c) Put equations in atrix form.
(d) Using Cramers rule, find the transfer
fimction _
H (s)^
I;in(s)
o f the circuit.
(e) Find the impulse response h{t) of the
66. Consider the circuit of Figure PI 3.66. This circuit.
problem is to be solved using (modified) nodal (0 Find the response of the circuit to
analysis. i.J^t) = -8u(t) A, assuming the initial
(a) Draw the best equivalent frequency conditions are zero.
'O domain circuit for nodal analysis (g) Find the response due only to the ini
accounting for the as-yet-unknown tial condition on the capacitor. A sim
W initial conditions ifiO') and i/q(0 ). ple observation leads to the answer
(b) Following the procedure explained in direcdy.
the text, write three (modified) nodal
W
6“6 Chapter 13 • Laplace Transform Analysis 11: Circuit Applications
L -
+ V .(t) -
Figure 1M3.67
v jt)l
A N SW ERS: H {s)=
4,s(.v+ 1)
and /-(/) =
0 .2 5 1 )//(/)dcrivcd from Figure P i3.69
A B
Finally {t) = 2{t - \ + e and
V \ /V
Cv^O-)hit) = (e-‘ - \)u{t). R, +
-
c,
R a W . b
7 3 . In the circuit o f Figure P i 3 .7 3 ,
+
V „(t) = iX
Rs = Ri,= 1 .4 4 2 7
lo g ,( 2 )
/?, over [0, oo). Assume R^ = 2 Q, R^ = 8 Q., R^ (b) Find and for 0 < r < 1 sec.
Chapter 13 • Laplace Transform Analysis II: Circuit Applications 677
Figure P I3.75
32 V
c, c 76. In the circuit in Figure P I3.76, the switch
has been in position A for a long time. The
Figure P I3.73 switch moves to position B at r = 0, back to A
at ? = 2 sec, and finally back to B at r = 5 sec,
74. In die circuit of Figure P13.74, Vq = 10 V, where it remains. Suppose = 2.5 =
= 4 Q, = 3i2, /?2 = 24Q, Zj = 3 H, and 1.5625 H, R2 = 2.5 Q, ^2 = 1 H, C = 0.04 F,
^2 = 6 H. Suppose switch S j has been closed and = 10 V for all t.
for a long time while S2 has been open. Ax. t =
A B
0 S2 is closed, after which Sj is opened. At ^ =
1 sec, S j is closed and S2 is then opened.
(a) Find /j(0") and vjt)l
:\{t)
(b) Find z‘2(/) and for 0 < f < 1 sec.
(c) Find Zj(l“) and
Find and /2(1'^). Figure P I3.76
(d)
(e) Find and for 1 < /sec.
(a) Find Vf^s) and for 0 < / < 2 sec.
(b) Draw the equivalent frequency
domain circuit that is valid for / > 2
sec.
(c) Find an expression for with the
switch in position B and then deter
mine for 2 < f < 5 sec.
(d) Compute V(jJ) for / > 5 sec. Hint:
This can be done without any further
Figure P I3.74
computation.
for 0 < r < 10 sec. (e) Set up the Cram ers rule equations for
(c) Com pute the initial conditions at / = solving the equations o f part (e) for
10 sec, i.e., find ^ ^ (1 0 ") and
y ^ ( 1 0 “). Also assume that at r = 10
sec, the switch moves back to position
A. Draw the equivalent circuit in the s- SW ITCH ED CAPACITOR
domain valid for r > 10 or r' > 0. NETW ORKS
Com pute for r > 1 0 sec. 7 9 . Consider the circuit o f Figure PI 3 .7 9 .
(a) If v^{t) = 25 V and the 100 m F capac
itor is uncharged at r = 0 , find V(^t) for
2 sec < t.
(b) Now suppose vp) = 25 V and =
10 V. Find v^it) for 2 sec < t.
t=2 sec
Figure P I3.7 7
C H E C K : (d) v^Jt) = [ 1,846^(-2-5') -
150 mF 100 mF
'• " '6 v,(t)
8 1. For the circuit in Figure P I 3.81 ^^(0') = 0. itor voltages are zero at f = 0 . Com pute v\(t)
At time t=0' both switches arc flipped to posi and //2 (^) at / = O'", r = 2~ , and t = 2* .
tions A. At r = 1 sec, both switches move back
to their original positions, and then back to
positions A at r = 2 sec.
(a) Determine the vakie o f yj(O^).
(b) Determine the value of
lO F
85. In the switched capacitor circuit o f Figure
20 5(t)
5F 5F 5F P I 3 .8 5 , all capacitor voltages are zero at r = 0~.
20u(t)V
Switches S arc moved to positions A at r = 0 s
Figure P i 3.81 and to positions B at /■= I sec, and then back to
CH ECK : =3 V A at / = 2 sec. C om pute and plot v^{t) and V2 {t)
for 0 > /“> 3 sec.
82. Repeat Problem 81 for the circuit o f Figure ^ + V, - 3 mF
A 2m F
P 1 3 .8 2 .
20 V 2m F
1 mF
> r i4 F 6 mF
4F
40 8(t) Figure P I3.85
(!) 2F 2F 4F
10u(t)V
lO F f + Figure P I3.8 6
lO F
20u(t)V
(b) Repear part (a) for ^ = 0 .5 . increasing bur less rhan 8 0 V, and is
(c) Repeat parr (a) for k = 2. moved to position B when v^^ reaches
80 V. (2) S remains at position B when
kC
is decreasing bur is greater rhan 5 V,
and is moved to position A when
reaches 5 V. Find and make a
rough sketch of the waveform for one
cycle o f operation.
C H E C K : charging time s 4.?? msec;
discharging time s 3? jisec
M ISCELLAN EO U S
8 8. Use the material on op amp integrator
design to achieve the following input-output
characteristics. In your final design, capacitors
should be 100 nF.
0 .5 0
(a) V,u,b) = - — \<l\ + - V ;b\
S S Figure P i3 .8 9 Switching circuit lor gen
erating a sawtooth waveform.
(b) = I',h\
S S S
9 0 . (Fluorescent light) Reconsider the
0 .7 5 1 fluorescent light starter circuit
(C ) V , (•0 = - — K .- — + — hi
s s s described in the text (Example 13.8).
^ In this problem, suppose the ballast is
b\ realistically modeled as an ideal
S S
inductor in series with a 100 Q resistor,
as shown in Figure P i 3 .9 0 . Using SPIC E or
89. (Sawtooth waveform generation) This prob some other circuit simulation program, com
lem uses some simple switching techniques in an pute the starting voltage, due to the ini
RC circuit to generate an approximate sawtooth tial condition on the inductor, as depicted in
waveform. Sawroorhs are common to a number Figure P i 3 .9 0 . Estimate the starting voltage
o f devices, such as televisions and test equipment. from a plot o f the response over one half-peri
Consider the circuit of Figure P i 3.8 9 . od. Com pute the difference between the peak
(a) Assume the circuit is initially at rest. voltages with and without the 100 Q resistor
Beginning at time / = 0, the switch, S, is present. Is the lossless circuit o f Example 13.8 a
alternately closed to the left position. A, good approximation o f the starter response?
for 1 msec and then to the right posi
tion, B, for 50 |isec. Find z-'Q(r) for 0 s r
^ 1.05 msec (one q^cle of operation),
and sketch the waveform.
CH ECK : 19 V
(b) T he circuit in the shaded box is a
crude model o f a neon lamp (costing
less than a dollar) and operates as fol
lows: (1) S is at position A when is
u Chapter 13 • Laplace Transform Analysis II: Circuit Applications 681
1000
0.001 pF
Li,(0)5(t) =0.088(t)
Figure P I3.90 Model of fluorescent light starter circuit that includes a ballast resistance of 100 Q.
Note that in the time domain, the effect of the initial inductor current appears as an impulse in this
model.
U
C H A P T E R
Electric motors turn fans, run air conditioners, pull trains, rotate antennas, and help us in a wide
variety o f ways by efficiently converting electrical energy into mechanical energy. These electro
mechanical devices are o f two general types: ac and dc. D c motors are ordinarily used in electric-
powered transit cars, often with rwo motors per axle to propel the car. A typical rating o f such a
m otor is 140 hp, 3 1 0 V, 2 ,5 0 0 rpm. Another type o f dc m otor is a high-performance dc servo
motor, found in computer disk drives and microprocessor-controlled machinery. These motors are
very useful in applications where starts and stops must be made quickly and accurately. As an
application, we will represent a dc m otor by an equivalent circuit-like model and analyze its oper
ation using the transfer function method.
O f the several kinds o f dc motors, the t)'pe most pertinent to the analysis techniques o f this chap
ter is a permanent magnet type typically found in low-horsepower applications. They are reliable
and efficient. Further, for a permanent magnet dc motor, a plot o f the torque produced on the
rotating shaft o f the motor versus the input current to the motor is almost a linear curve. Hence,
the m otor has a linear circuit-like model that most nearly describes its performance over a large
range o f operating conditions. Since the output o f the m otor is a mechanical quantity, e.g., angu
lar velocit)' o f the m otor shaft:, the transfer function is the only viable modeling tool available to
us at this stage o f our development. Section 9 presents the circuit model o f the m otor and devel
ops the transfer function analysis o f its operation. Several problems at the end o f the chapter
extend the analysis. In more advanced courses, time domain analysis is developed and used.
CHAPTER O BJECTIVES
1. Characterize the transfer function of a circuit in terms of its poles, zeros, and gain constant.
2. Use knowledge o f the pole locations o f a transfer ftjnction to categorize generic kinds of
responses (steps, ramps, sinusoids, exponentials, etc.) that are due to different kinds o f
terms in the partial fraction expansion o f the transfer function.
3. Identify, categorize, define, and illustrate various classes o f circuit responses, including
fi.S-4 Chapter 14 • Laplacc Transform Analysis III: Transfer Function Applications
the zero-srate (zero initial conditions) response, the zero-input response, the step
response, the impulse response, the transient and steady-state responses, and the natural
and forced responses.
4. Define the notion o f the frequency response o f a circuit, explore its meaning in terms of
the transfer function, and introduce the concept o f a Bode plot, which is an asymptotic
graph o f a circuit’s frequency response.
5. Introduce the notions o f frequency and magnitude scaling of circuits.
6. Illustrate the applicabilit)- o f the transfer function concept to a circuit model o f a dc
motor.
SECTIO N HEADIN GS
1. Introduction
2. Poles, Zeros, and the 5-Plane
3. Classification o f Responses
4. Computation o f the Sinusoidal Steady-State Response for Stable Networks and
Systems
5. Frequency Response
6. Frequency Scaling and Magnitude Scaling
7. Initial- and Final-Value Theorems
8. Bode Plots
9. Transfer Function Analysis o f a DC Motor
10. Summary
11. Terms and Concepts
12. Problems
1. IN TRO D U CTIO N
O ur experience o f using Laplace transforms to calculate responses makes clear that the pole-zero
structure o f the transfer function sets up generic kinds o f circuit behaviors: constants, ramps,
exponentials, sinusoidals, exponentially modulated sinusoids, etc. Such knowledge leads to a qual
itative understanding o f the circuits response. For example, if a pole is in the right half o f the com
plex plane, then we know that the response will grow with increasing t. Identifying this kind of
behavior allows us to define the notion o f stabilit)' o f a circuit or system. Generally, the pole-zero
locations allow us to categorize and com pute various special t)'pes o f responses, including tran
sient, steady-state, natural, forced, step, and impulse responses. Coupling the transfer function
with the presence o f initial conditions in the circuit permits us to define two further types of
responses fundamental to both this text and advanced courses in circuits, systems, and control: the
zero-input response (due only to the initial conditions o f the circuit or system) and the zero-state
response (due only to the input excitation, assuming that all initial conditions are zero).
These time domain notions are balanced by the concept o f the frequenc)' response of the circuit
or system. Briefly, the frequenc)' response is the evaluation o f the transfer function, H{s), for s=
Jii). Since H{Jw) has a magnitude and phase, the frequenc)' response breaks down into a magni-
Chapter 14 * Liplacc Transform Analysis 111: Transfer Function Applications 6S"S
rude response and a phase response. A technique for obtaining asymptotic (straight-line) approx
imations (called Bode plots) is also outlined in this chapter.
As a final introductor)^ remark, unless stated otherwise, all circuits in this chapter are linear and
have constant parameter values. Such circuits are said to be linear and time invariant. Also, for
convenience in this chapter, the symbol Z. will be used to denote either o f two things: (i) the angle
o f a complex number, L{a + jb) = arg(/z + jb) = z^n~^{bla), with due regard to quadrant, or (ii) Z.({)
= The context will determine the actual usage o f the symbol.
In all our circuits, impedances Z{s), admittances }\s), and transfer functions H{s) are rational
functions of^, i.e., they are ratios o f a numerator polynomial ;/(j), divided by a denominator poly
nomial, r/{s). Mathematically,
where s = pj is a finite pole o f H{s), s = zj is a finite zero o f H{s), and A" is the gain constant of the
transfer function. We assume that common factors o f n{s) and d{s) have been canceled. A finite pole
satisfies M /’,) = which is shorthand for //(s) -♦ x as j “♦ />y, and a finite zero satisfies = 0. i-e.,
the transfer function takes on the value zero at each z-. If pj = pj, i j, the pole is a repeated pole. A
pole repeated twice is second order; one repeated three times is third order, etc. The terminology is the
same for zeros. Also, transfer functions sometimes have infinite poles or infinite zeros. If m < n and s
-* 00 , then H{s) 0, suggesting the term “zero at infinit}'.” In such cases, H{s) has a zero of ordern
- m at infinit)'. If, on the other hand, n < m, H{s) has a pole of order rn - n at infinit)'.
O ut o f all this terminolog)' comes one striking fact: transfer functions, impedances, and admit
tances are characterized by their finite poles, their finite zeros, and their gain constant.
S o l u t io n
From equation 14.1 and the given locations o f its poles and zeros, the transfer function must have
the form
3 l(l + 2 ) ( l - h 4 ) 15
This implies that K = 2 .5 . Equation 14.2 specifies the transfer function with K = 2 .5 .
686 Chapter 14 • Laplace Transform Analysis 111; Transfer Function Applications
Exercises. 1. Suppose a transfer function has a zero at j = 1 and a pole at x = - 1 and that as ^
0 0 , H{s) -* H{s).
- 3 . Find
A N SW E R : His) = -Ms - 1 )!{s + 1)
2. A transfer function H{s) has poles at j = - 1 and - 2 and zeros at j = - 3 and - 5 . It is further
known that H{0) = 15. Find //(oo).
A N SW E R : 2
Because the essential information about transfer functions resides with the poles and zeros, a plot
of these locations in the 5-plane, called a pole-zero plot, proves informative.
S o l u t io n
T he transfer function given in equation 14.2 has the pole-zero plot shown in Figure 14.1,
JW
>k
■j
-o--------- ►'G
-4 -3 -2 -1
...J
FIG URE 14.1 Pole-zero plot of //(j) given by equation 14.2, where the poles arc flagged by and
the zeros by “o.” Since j = a + ya), the real axis is labeled o and the imaginar)' axisyw.
(5 + 1)(.V4-3)
(1 4 .3 )
l(5 + l)“ -H l](5 + 2 )
X ...J
FIG U R E 14.2 Pole-zero plot of H{s) given by equation 14.3. Again, the a-axis represents the real
part of the pole or zero and the /o -axis represents j times the imaginary part of the pole or zero.
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications 68’
Plots such as those in Figures 14.1 and 14.2 com m unicate much about the nature o f the imped
ance, admittance, or transfer function o f a circuit. For example, an RC input impedance,
satisfies the following properties: (i) all o f its poles and zeros are on the non-positive C7-axis o f the
complex plane; (ii) all o f its poles are simple (o f multiplicit)' 1) with real positive residues, i.e., the
coefficients in a partial fraction expansion are real and positive; (iii) does not have a pole at
j = CO; and (iv) poles and zeros alternate along the a-axis. Proofs o f these assertions can be found
in texts on network synthesis.
Exercise. C om pute the input impedance o f the circuit in Figure 14.3, and show the pole-zero plot
if / ? ,= / ? - ,= 1 Q, Cj = 0 .2 5 F, and C2 = 0 .5 F. Are the poles on the non-positive a-axis? Do the
poles and zeros alternate? Is there a pole at s= oc?
J_
A N SW E R : c.
A’.C ,
More commonly, pole-zero locations provide important qualitative information about the response of
the circuit. Pole locations determine the inherent, natural behavior of the circuit, and the poles are
commonly called tmttiralfrequencies. Howe\'er, the complete set o f natural frequencies o f the circuit
may be larger than the set o f poles o f the transfer function. 'Fhis is because there might have been a
pole-zero cancellation in constructing the transfer function. The canceled pole would amount to a nat
ural frequency o f the circuit that is not present in the poles o f the transfer ftmction.
T he terms in a partial fraction expansion o f the response establish the types o f behavior present in
the response. Each term has only one o f several possible forms. Four ver)' com m on terms are Kh,
KjJ, K/(^ - />y), and K/{s - pj real in each case. Figure 14.4 sketches each o f the associat
with
ed responses. In Figure l4 .4 a , the term Kh leads to a dc response and KlJ to a polynomial response
proportional to . In Figure l4 .4 b , the term K l{s- p-) leads to an exponential response that is
increasing if pj > 0 and decreasing if pj < 0. Finally, in Figure l4 .4 c , if pj < 0, the response curve
has a hump.
These qualitative behaviors suggest that one important application o f the transfer function is
determining the “stability” o f the response; i.e., under what conditions will the circuit response
remain finite for all time?
688 Chapter 14 • Laplacc Transform Analysis III: Transfer Function Applications
(a) — —
s
(b)
■
V- Pi
(c)
(S - Pi)
FIGURE 14.4 Response t)'pcs common to partial fraction expansion terms, (a) The term Kh leads to a dc
response and KJ^ to a polynomial response proportional to (b) The term KJ{s — leads to an expo
nential response diat is increasing i f > 0 and decreiising if pj < 0. (c) If pj < 0, the curve has a hump.
Chapter 14 • Laplacc'Iransform Analysis III: Transfer Function Applications 68^)
In addition to the preceding response t)'pes, there is the sinusoidal response associated with terms
o f the form , „
/4.V+ B
-----------^------- T
f.v + a r +
FIG URE 14.5 Various sinusoidal responses, (a) Exponentially decaying, - a < 0.
(b) Pure sinusoidal, 0 = 0. (c) Exponentially increasing, - a > 0.
690 Chapter 14 • l^placeTransform Analysis III; Transfer Function Applications
Referring again to Figure 14.5 , the real part o f a pole, i.e., - a , specifies the decay rate o f the
response. Often, the word damping is used. I f -O < 0, then the response is damped and the oscil
lations die out. T he farther -O is to the left o f the imaginary axis, the greater the damping. If a =
0, there is no damping and the response is a sustained oscillation. I f - a > 0, the response is nega
tively damped, i.e., the response is unstable And increases without bound.
O ne concludes that pole locations specify the type o f time domain behavior o f a circuit or system.
A very important type o f circuit behavior characterized by the pole locations is stability.
Despite the need for stability, some circuits utilize an unstable-like response for a finite duration.
Circuits that exhibit both stable and unstable-like responses are studied in electronics courses.
A transfer function with first-order poles on the imaginary axis is sometimes called metastable.
Such a classification has no practical or physically meaningful significance, since the ubiquitous
presence o f noise would excite the mode and cause instabilit}' o f the circuit. Moreover, in power
systems engineering, i.e., the study o f the generation and delivery o f electricit)' to homes and
industry, transfer function poles that are in the left half-plane, but close to the imaginary axis, are
highly undesirable. Such poles cause wide fluctuations in power levels. The situation is analogous
to the way a car without shock absorbers would bounce. Much work has been done on how to
move the poles that are close to the imaginary axis farther to the left. Moving these poles to the
left increases the damping in the system and maintains more stable power levels. Summarizing,
the requirement that the transfer function have no poles on the imaginar)’ axis is both theoreti
cally and physically meaningful.
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications 691
Exercises. 1. If H{s) = find a bounded input that will make the response
unbounded.
2. If H{s) = = l/(^^ + 1), find a bounded input that will make the response unbound
ed. Use M ATLAB or some other program to plot the response for 0 ^ r s 10 s.
= 4kO R^ = 4 k n
(a) (b)
FIGURE 14.6 (a) Correct wiring of op amp circuit, (b) Accidental, improper wiring of op amp circuit.
S o l u t io n
Assume that the op amp is modeled as a voltage-controlled voltage source with a finite gain o f 10"^
and that there is a very small stray capacitance o f 1 pF across the input terminals. Figure 14.7 illus
trates the equivalent circuit model for each o f the circuits in Figure 14.6.
= 4kO R^ = 4kQ
P art 1: Analysis o f the correctly wired op amp. W riting a node equation at to compute the trans
fer function o f Figure l4 .7 a yields
1 1
.V + - - - - - - - + • v.,-— v,- — v = o
Y ’ ’ CR| ’ ’ C « , j CRi '' CR,
Substituting x 10“^ for Ky produces the stable transfer function
\
10 13 (1 4 .4 )
His) = 12
Vi(s) R^C 1 l+ I O ' ^ + 2 .501 X 10
.y + ------ +
R^C RjC I
P art 2 : Analysis o f the incorrectly wired op amp. To com pute the transfer function o f the circuit o f
Figure l4 .7 b , we write a node equation at Vj.
C i V , , + ^ ( V ^ - V '. ) + -^ (V r f -V '„ ) = 0
Kj /<2
which produces the transfer function
10" 1013
H{s) = 12
(1 4 .5 )
Viis) R^C 1 1 -u r . y - 2 .4 9 9 x 1 0
5 + -------- h
T he transfer function o f equation 14.5 has a right half-plane pole, in contrast to that o f equation
14.4. This implies that the incorrectly wired circuit is unstable, which explains the students con
cern over the surprising performance o f the op amp.
A brief interpretation o f the zeros o f a transfer function ends this section. This is best done in
terms o f a simple example. Suppose
( .v + l )- + l
His) =
V;., (5 + l)U + 2)(^ + 3)
Let v- it) = e ‘ sin(/)/^(f) V, so that
1
Assuming that the system is initially “relaxed,” i.e., all initial conditions are zero, we obtain
1
V^,As) = H is )V :Js ) =
(.9 + l)(5 + 2 )C y + 3)
Chapter 14 • Laplace Transform Analysis 111: Transfer Function Applications 693
for appropriate constants A^-. Observe that the response dies out very quickly and does not have
any term similar to the input signal. This follows because tiie input signal has transform poles, s
= -1 ± J, that coincide with the zeros o f the transfer function. One can think o f the pole locations
in the transform o f the input signal as identifying frequencies that are present in the input. Hence,
the effect o f these input signal frequencies (poles) is canceled out by the transfer function zeros,
eliminating them from the circuit response.
3. CLASSIFICATIO N OF RESPONSES
In addition to the various response behaviors discussed in section 2, there are other general
response classifications. Three fundamentally important general response classifications germane
to all o f circuit and system theory are the zero-input response, the zero-state response, and the
complete response.
Zero-state response;The response o f a circuit to a specified input signal, given that the ini
tial conditions are all set to zero. Figure 14.8 illustrates this idea.
Output
Relaxed Circuit I
Input F(s) > V(s) = H(s)F(s)
H(s)
Zero-State Response
FIGURE 14.8 Relaxed circuit having transfer function H(s) and zero-state response.
Complete response: The response o f a circuit/system to both a given set o f initial condi
tions and a given input signal. For linear circuits, the complete response equals the sum o f
the zero-input and zero-state responses.
Recall that a circuit is linear if, for any tv.'o inputs,yj(f) and/jU), whose zero-state responses are
and_y2 W> respectively, the response to the new input [AT|/j(r) + K^ 2 ^t)] is +
where /T, and A'-, are arbitrary scalars. The circuits studied in this book are linear unless
otherwise stated.
T he decomposition o f the complete response into the sum o f the zero-input and zero-state
responses is important for three reasons:
1. It is defined for arbitrary input signals.
6 9 ‘» Chapter 14 • Laplace Transform Analysis 111: Transfer Function Applications
2. T he zero-state response is given by £~^[H{s)F{s)], l^or tiie arbitrary i-domain input F{s).
3. It illustrates a proper application o f the principle o f superposition for linear dynamic net
works having initial conditions.
T he following example illustrates point 3.
= 6u(t)
20 + 2 0
Va(t) v jt)
Vo(t); ■
1F
= 3u(t) = 6u(t)
P art 1: Response o f linear resistive network o f Figure 14.9. For the resistive network o f figure 14.9,
the contribution to vj^t) due to vj^t) with = 0 is v^J^t) = l//(f), and the contribution due to
VjJ^t) with vj,t) = 0 is = lu{t). By superposition.
For this type o f circuit there is no initial condition and the complete response consists o f only the
zero-state response, which decomposes into the superposition o f each source acting alone.
P art 2 : Response o f linear RC network o f Figure 14.10. Now consider the dynamic network o f
Figure 14 .1 0 . Suppose the capacitor has an initial voltage o f 2 V at f = 0, i.e., ^^(0) = 2 V.
Step 1. W ith applied, y^(0) = 2 V, and Vf^set to zero, the response is
v j t ) = {0.5e-‘ ^ \.5)ii{t)W
Step 2 . W ith the input Vy{t) applied, y^(0) = 2 V, and set to zero, the resulting response is
= {-e-^ + 5)u{t) V
Chapter 14 • Laplace Transform Analysis 111: Transfer Function Applications 695
T he last answer is wrong because the response due to the initial condition has been added in twice.
Hence, the complete zero-state response, by superposition, is the sum o f and Further,
the zero-input response is le~Ui{i). Hence, the complete response is
= (4-5 - 2.5rO /K r) V
It is important to note that the transfer function is defined only for circuits whose input-output
behavior is linear. In terms o f the zero-state response, if a circuit has a linear input-output behav
ior characterized by a transfer function H{s), then H{s)[K^Fj^{s) + K2 Fi{s)] = K^H{s)F^{s) +
K2 H{s)p2 {s) = ATj KjU) + KjY^is), where Yf,s) = H{s)Ff^s) is the zero-state response o f the network
to F^^s). This says that the zero-state response to [A"j/j(r) + is l^^^^ce,
the transfer function model reflects the underlying linearit)^ o f the circuit.
The complete response has a second structural decomposition in terms o f the transient and steady-
state responses. T he notion o f a periodic signal is intrinsic to these classifications. A s ig n a l/r) is
periodic if there exists a positive constant T such thaty(r) =j{t + 7) for all / > 0. (The restriction
to r > 0 exists because our Laplace transform analysis implicitly constrains our function class to
those that are zero for t < 0.) If a signal is periodic, there are many positive constants for which
j{t) =J{t + 7) for all t> 0. For example, i f / / ) =J{t + T) for some T and for all t> 0, then it is true
for IT, 5T, etc. We define the fundam ental period, often simply called the period, o f / f ) to be
the smallest positive constant T for w h ic h /f ) = J{t + T) for all t > 0. Sinusoids are periodic sig
nals: sin(2Jtr) = sin(2Jt/^ + 2 k ) with fundamental period T= \. The square wave o f Figure 14.11 is
periodic with fundamental period 7"= 2.
Chapter 14 • Liplace I'ransform Analysis ill: Transfer Funciion Applications
f(t)
>k.
1
•
1 2 3 4 5
■ -1
This notion of periodicity and, by default, non-periodicity allows us to define the transient and
steady-state responses o f a circuit.
Steady-state response: Those terms o f the complete response that satisfy the definition o f
periodicity for r > 0. This includes a constant response.
Transient response: Those terms o f the complete response that are not periodic for / > 0 , i.e.,
that do not satisfy' the definition o f a periodic function for / > 0. Note that a constant response
satisfies the definition o f a periodic function.
A circuit response may have no transient part, as illustrated by the sustained sinusoidal oscillato
ry response o f the circuit given in Figure 1 4 .1 2a. Further, the steady-state part o f the response may
be zero, as in the circuit o f Figure 1 4 .12b, where is a damped sinusoid. If the a circuit is
unstable, the transient response may blow up, overwhelming the constant or periodic part o f the
complete response, as in the case o f = {e^‘ cos(10/) + 1 5)u{t) V, where the steady state is 15«(/)
V. Note that “transient” here does not mean something that diminishes in importance with time.
Most circuits have both a transient and a steady-state response. When the input is constant or peri
odic, the circuit response approaches the steady-state response asymptotically for large t, i.e., as the
transient dies out, only if the circuit is stable. For such circuits, the steady state is crucial. Further,
when the input is sinusoidal, the steady-state response is easily computed via the transfer function,
//(s ), or by the phasor method. Details o f the calculation are presented in section 4.
1F 1 H 1O 1F 1 H
(a) (b)
FlG U lll{ 14.12 (a) Unstable circuit illustrating the possibility of no transient response,
(b) Stable circuit having a zero steady-state response.
Chapter 14 • Laplace Transfomi Analysis III: Transfer Function Applications 6^)'
E X A M P L E 1 4 .5 . Com puting the response o f the circuit o f Figure 1 4 .1 3 provides a simple illus
tration o f the decomposition of the complete response into tiie sum of the zero-input and zero-
state responses. Also, some rearrangement o f the terms identifies the transient and steady-state
responses.
-O
— R . +
> c
ZJs)
P
—> 1
(a)
FIG U R E 14.13 /?Ccircuit for Example 14.5. (a) Time domain circuit.
(b) Frequency domain equivalent, accounting for initial condition.
S o l u t io n
Step 1. Computation o f the zero-state response. The input impedance can be viewed as a special type
o f transfer function. For Figure 1 4 .13a the output V(^s) is the voltage appearing across the input
current source. Hence,
__
R RC
Rl,, Rio
= RI, Ki t )
.v-i-
RC
Step 2 . Computation o f the zero-input and complete response. Now, supposing that ^(^0“) 0 , the
zero-input response is the inverse transform o f //(i)[C y^ '(0“)] as per Figure 1 4 .13b. Hence, by
superposition, the complete response is
\
RC u{!) + }e
Step 3 . Decomposition into transient and steady-state responses. As a final point, since the input is
dc, a step function, the complete response decomposes into its transient and steady-state parts as
Observe that from the above example the transient and zero-input responses are ordinarily different.
E X A M P L E 1 4 .6 . It is sometimes mistakenly said that the zero-input response contains only those
frequencies represented by poles o f the transfer function. To see the fallac)' o f this statement con
sider the y?C bridge circuit o f Figure 14 .1 4 . Com pute the zero-state and zero-input responses.
S o l u t io n
The transfer fiinction o f this circuit is H{s) = = 0> which has no poles. Hence, the
zero-state response is always zero. On the other hand, if 0, then
'o u t ( 0 = V c ( 0 )e x p
where = 1 Q. Thus, the zero-input response is a decaying exponential whenever v^^Qr) ^ 0 and
C > 0. Notice that the transfer function has no poles. As a side remark, in this case, the zero-input
response is also the transient response, with the steady-state response being zero.
T he phenomenon illustrated by Example 14 .6 occurs because the symmetry o f the resistor values
precludes excitation by the current source. Moving the current source to a different position, say,
in parallel with one o f the resistors, or changing the value o f one o f the resistors to 0 .5 Q will result
in a nonzero transfer function.
USO-) = 1
> m v
1 H
1Q
v.(t)
(a) . (b). . .
FIG U R E 1 4 .1 5 RL circuit for Example 14.7. (a) Time domain series RL circuit,
(b) Frequency domain equivalent, accounting for initial condition.
S o l u t io n
Using Figure 1 4 .15b and the principle o f superposition leads to the response
4s 1
4.V -2 2.V +2
(.s + l)(5 - + l) 5+1 r + 1
It follows that the zero-state response is -2e~^ti{t) + 2[cos(r) + sin(^)]«(/). Notice that both the zero-
input and the zero-state response contain a transient part, the part proportional to e~^. A little rear
ranging shows that the complete response is
which implies that the transient response o f the circuit is —e~^u{t) and that the steady-state
response o f the circuit is 2[cos(/) + sin(f)]/<(r) A.
Exercises. 1. An RLC network has transfer function H{s) = = l/(^ + 1). If an input
cos(?)«(t) A is applied, then for very large t, approaches a cosine wave o f what form?
A N S W E R : 0 .7 0 7 cos(/ - 45'")
Many books on elementary circuits contain two other notions o f response: tlie natural response and
forced response. To explain this we use the term “exponent” to mean \ in e^‘, where k is possibly
complex. For example, if sin(r) is part o f a response, then it comes from , in which case /. = j.
Natural response: Fhe portion o f the complete response that has the same exponents as the
zero-input response.
Forced response: 'I'he portion o f the complete response that has the same exponents as the
input excitation, provided the input excitation has exponents different from those o f the
zero-input response.
It would seem natural to try to decompose the complete response into the sum of the natural and
forced responses. Unfortunately, such a decomposition applies only when the input excitation is
(i) dc, (ii) real exponential, (iii) sinusoidal, or (iv) exponentially modulated or damped sinusoidal.
Further, the exponent o f the input excitation, e.g., a m f{t) = e'^‘ti{t), must be different from the
exponents appearing in the zero-inpur response. T he natural and forced responses are properly
defined only under these conditions.
T he decomposition o f a complete response into a natural response and a forced response is impor
tant for two reasons. First, it agrees with the classical method o f solving ordinary differential equa
tions having constant coefficients, where the natural response corresponds to the complementary
function and the forced response corresponds to the particular integral. Students fresh from a
course in differential equations feel quite at home with these concepts. The second reason is that
the forced response is easily calculated for any o f the special inputs— dc, real exponential, sinu
soidal, or damped sinusoidal. For example, if the transfer function is H{s) and the input is
then the forced response is simply
H{a)V^“ (1 4 .6 )
To justify equation 14.6, note that the Laplace transform o f the input is V!{s —a). Since the com
plete response is the sum of the zero-input and zero-state responses, we have
V
Complete response = [zero-inpnt response] +iT ' Hi s ) ------
s- a
The zero-input response terms all have exponents different from a, the exponent o f the input. The
second term, \H{s) Vl{s- //)], has only one term with exponent equal to a. Executing a partial
fraction expansion o f this term yields
K
H{ s) ------- = -------- + [terms corresponding to poles of H{s)]
s - (I s - a
Using the residue formula to calculate A'leads to K = H{a)V. Thus,
and L'^[H{s)VI{s - a)] has a term, which we identify as the forced response.
By using exacdy the same arguments, it is possible to show that if the input is a complex
exponential function Ve¥, where both Kand are complex numbers, then the forced response is
simply
H{s^) \W
R e[//W K ^ 1
This relationship o f the input to the forced response prompts some textbooks to define the trans
fer function H{s) as the ratio o f the forced response to the input, under the condition that the
input is a complex exponential Ve^‘. This, however, is not natural and makes one wonder at the
applicability o f such a definition to the broad class o f inputs for which the transfer function is
most naturally defined, as covered in Chapter 13 o f the text.
Suppose that a transfer function H{s) models a stable linear circuit containing a total o f n capaci
tors and inductors. In addition, suppose there are no comm on factors in the numerator and
denominator o f H{s) and that the degree o f the denortiinator ofH{s) is ri. (This means that the effect
o f each capacitor and inductor is included in H{s).) T he goal o f this section is to develop the fol
lowing formula: if H{s) satisfies the aforementioned assumptions, and the input to the circuit has
the form A cos(to/ + 0 ), then the steady-state circu it o u tp u t response has the form
B cos(o)r + (}))
(14.7a)
From an input-output viewpoint, these formulas imply that the frequency response is the steady-
state response o f a circuit to sinusoids o f varying frequencies. To construct this formula, suppose
again that a linear circuit has a stable transfer function model H{s). Suppose also that the circuit
input is a s i n u s o i d w h o s e Liplace transform is F{s) with zero-state response Y\s), as illustrated
in Figure 14.16.
Transfer Function
Since H{s) is stable, all poles lie in the open left half o f the complex plane. Assume that H{s) has
real, distinct poles labeled />j, ... , and complex poles labeled -(tj ± Consequently, H{s) will
have a partial fraction expansion containing only two types o f terms; those having real poles, p - <
0, and those having complex poles with negative real parts, i.e., a- < 0. Specifically,
Suppose now that the circuit is excited by a sinusoidal input o f the form
0.5 Aej^ 0 .5
^(•v) = --------:— + ---------:—
.V - yco s + jo i
Then a partial fraction expansion o f the Laplace transform o f the zero-state response, F(j) =
H{s)F{s), has the form
A\ At C|.v + D\ R\ R')
Y(s) = ----- ^
— +— — + . . . + --------- ^ ^----- T + . . . + ----- ^ +
s - p\ s - P2 (5 + a 1 j" + ( Pi j" - yw A--I- yco
In the steady state, i.e., for large t, the only residues o f interest are and /?2 , because the part o f
the time response due to the other terms decays to zero with increasing t. By the usual methods
o f complex variables, we obtain
Chapter 14 • Laplace Transform Analysis III: Transfer Funaion Applications 703
. J0.5Aej^ 0.5Ae~j^\
H(s)(s - j(o
A jw
s + jm j S = J(0
and
= 0.5A\H(-j(o)\e^^^^~j^^^^
/?2 = 0.5A\H(J(a) ^)
Consequendy, die Laplace transform o f the steady-state response when all initial conditions of the
circuit are zero is
s- yo) S+ jO)
In fact, this is the Laplace transform o f the actual steady-state response, provided that the zero-
input (nonzero initial conditions) response makes no additional contribution. The zero-input
response makes no contribution to the steady-state response when one or more o f the following
reasonable conditions on the circuit are met:
1 . The network has only practical passive elements, meaning that there are always stray
resistances present.
2 . The circuit may have active elements in addition to passive elements, but remains stable
in the sense that every capacitor voltage and every induaor current remains bounded for
any bounded circuit excitation.
3. The circuit contains a total o f n capacitors and induaors, and the stable transfer func
tion, H{s), has n poles.
Hence, if the input to the circuit lias the form A cos(o)r + 9 ), then the steady-state circuit output
response has the form B cos(ojr + (})), where the magnitude B = A\H{Jlo)\ and the phase shift ({) =
0 + //(» .
The next question concerns the numerical calculation o f B and One method is simply to eval
uate H{s) at s = JiO. W ith a calculator that easily accommodates complex numbers, this is quite
straightforward. An alternative method is to use the graphical technique o f the next section.
At this point, it is instructive to illustrate equation 14 .7 and at the same time compare it with the
phasor method studied in a first course on circuit theory.
E X A M P L E 1 4 .8 . In the circuit o f Figure 14.17, ji = 0.5 and = cos(2r) V. Find v^{t) for large t.
S o l u t io n
P art 1. Phasor method. From the principles o f phasor analysis detailed in Chapter 10, the phasor
domain circuit o f Figure 1 4 .1 7 at (O = 2 rad/sec is given by the circuit of Figure 14.1 8 .
1_
=0
./4 - y (14.9)
V ,= = 0 .8 Z - 3 6 .9 '
1 + yO.75
Therefore, for large t
v^{t) = 0 .8 c o s ( 2 r - 3 6 . 9 “) V
Chapter 14 • Laplacc Transform Analysis III: Transfer Function Applications OS
P art 2 . Laplace namform method. Tlie first step here is to construct the j-domain equivalent cir
cuit, which is given in Figure 14.1 9 .
V i-V ;- V i-0 .5 V |
+■ =0
2.V + 2 2
(1 4 .1 0 )
v,(s) (.v + ir
v^(t) = 0 .8 c o s ( 2 /- 3 6 .9 ° ) V
In this example, the two methods give the same answers, as expected. Since complex numbers are
easier to manipulate than rational functions, what is the motivation for such an analysis using the
transfer function H{s)^ W hy not stay with the phasor method? T he next example answers these
questions.
E X A M P L E 1 4 .9 . In Example 14.8 with the circuit o f Figure 1 4 .1 7 , let the value o f p be increased
to 1.5. Find y,(r) for large t.
S o l u t io n
P art 1. Phasor method. Since only the response for large t is desired, the problem appears to be one
involving sinusoidal steady-state analysis. The phasor domain circuit o f Figure 1 4 .1 8 yields the
single node equation
V, - 1 V, - 1.5V,
= 0
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications
V, = ------ !------ = 0 .8 Z 3 6 .9 °
' l - /).7 5
A beginner who has just learned sinusoidal steady-state analysis by the phasor method might
accept this answer. Unfortunately, the answer is not the voltage v^{t) for large r! T he reason is clear
from the Laplace transform analysis, which follows.
Part 2: Laplace transforyn method. From the ^-domain equivalent circuit o f Figure 14.1 9,
v j - y /- ^
25+2. 2
.9
Solving for the transfer function yields
%{s) -2s
His) =
Viis) (s-\ r
Since there are poles o f H{s) in the right half-plane, the circuit is unstable. As f becomes very large,
the magnitude o f v^{t) approaches infinit)', instead o f 0 .8 cos(2r + 3 6 .9 ") V, as calculated by the
phasor method.
This analysis demonstrates that the unstable behavior o f a circuit cannot be determined by the
phasor method. It is desirable to know when to use a particular method in order to avoid unnec-
essar)' complicated calculations. The following guidelines help:
1. W hen the stability o f the circuit has been assured by some means, and co has a specific
numerical value, the phasor method is the better method to use for computing the
response for large t, which is also the steady-state response in this case. Circuits whose
stability is guaranteed include those with only passive elements, such as resistors, capaci
tors, and inductors; and amplifier circuits o f well-established configurations.
2. The circuit is known to be stable, but CD is variable. In this case, the H{s) method is supe
rior to the phasor method. To say the least, we need only write sL instead ofy'coZ.. There
are other advantages to be gained from knowing the pole-zero plot o f H{s) that are not
possible with the phasor method. The examples o f frequency response calculations given
in the next section clearly demonstrate this point.
3. If the stability o f the circuit is not yet determined, then H{s) should be calculated and its pole
locations checked for stability. Then step 1 or step 2 should be referred to, as appropriate.
/Y(^) = _ S--0.5S + 5
V :Js) 5 - +0.5.V + 5. 7321
Chapter 14 • Laplace Transform Analysis III; Transfer Function Applications 707
is driven by a sinusoidal input = j2cos{2t + 45°)t^{t). Show that the steady-state response
is given by
L= 1 H +
v,(t)
C=1F
(a)
5. FREQ U EN CY RESPONSE
The frequency response o f a linear stable circuit having constant parameters characterizes the
input-output behavior o f the circuit to unit magnitude sinusoids, cos(oj/), as ca varies from 0 to
00 . This extremely important concept plays a major role in the analysis and design o f circuits and
systems. In terms o f the transfer function, the frequency response o f a stable circuit is the evalua
tion o f H{$) zi s =jw. In terms o f phasor analysis, studied in an introductory course, the frequen
cy response o f a circuit corresponds to the ratio o f the output phasor to the input phasor.
From the steady-state analysis perspective o f the previous section, if an input has the form
cos(ojQr), then the steady-state response (i.e., the response for large t, after all transients have died
out) has the form B cos(ojQr + (j)), where B = |//(/(Dq)| and the phase shift ({) = Z.//(/Wq). Here
|//(/Wq)| is the magnitude o f the complex number ^(/W q), and Z.A/(/0)q) is the angle o f the com
plex number //(/cOq). Thus, //(/(o ), for 0 s to < x , defines how a linear circuit adjusts the magni
tude and phase o f an input sinusoid to produce a steady-state output sinusoid o f the same fre
quency, but possibly with a different magnitude and phase.
An example o f practical importance is the specification o f a stereo amplifier. Here one specifies the
gain, gain-magnitude \H{jLo)\, to be more or less constant from 20 Hz to 2 0 kHz. W hy? Because
musical signals are composed o f sinusoids o f different frequencies. Accurate amplification o f the
music requires that all component sinusoids be amplified with equal gain.
’ OS Chapter 14 • Laplace Transform Analysis 111: Transfer Function Applications
Frequency response: The frequency response of a stable circuit or system represented by a transfer
function H{s) is the complex-valued funaion //(/w ) for 0 s OJ < oo. The magnitude (frequency)
response is |//(^)| for 0 to < co, and the phase (frequency) response is for 0 ^ co < co.
A complex-valued function H{jw) is a function such that for each value of to, Hijoi) is a complex
number. A complex number, a-^ + has a polar form, in which p , is the magnitude and
(j), the phase angle o f the number. Moreover, if + y' ^ 2 another complex number, then
(« j + y /?i)(« 2 +Jb2^ =
and
a2 + p2 P2
H(Ju)) = p(to)e-^‘^^‘'^^
is the angle or phase o f the frequency response. As in other books, magnitude response means the
magnitude o f the frequency response. Typically, frequency response computation requires a calcu
lator or computer. We now illustrate the idea o f frequency response with two so-called band-pass
transfer functions in which a band o f frequencies is passed with relatively little attenuation while
frequencies outside the band are significantly attenuated.
Q.25s
and .v^-h 0 .2 5 5 + 1
^ _________________ 0 -0 6 2 5 . v^_________________
^ ‘ / + 0.35355.s-^ + 2.0625.V - + 0.35355.V + I
»yIabcI(‘Magnitude response’)
»gtext(‘2nd O rder BP’)
»gtexr(‘4th O rder BP’)
O)
1/1
c
o
CL
t/i
O
u.
)
a;
-o
13
‘c
cn
(TJ
Frequency r/s
FIG U R E 14.21 Magnitude responses of a second-order and a fourth-order band-pass type transfer
function. Tlie fourth-order response has steeper sides (sharper cutoff) and a flatter top.
From the transfer function and by interpolation on the plots, one observes that at co = 0 the mag
nitude is zero and at co = x the magnitude is also zero; at (O = 1 the magnitude peaks, and this fre-
quenc)' is called the center frequency. This is characteristic o f a band-pass type o f response. T he
fact that the response peaks at OJ = 1 rad/sec means that the transfer functions are “normalized.”
Transfer functions o f practical band-pass circuits have much higher center frequencies. Such fre
quencies can be obtained by the technique o f frequenq' scaling, taken up in the next section.
A very important system theoretic relationship is that o f the poles and zeros o f the transfer func
tion to the magnitude and phase responses. In iJie above example for H^{s), the pole-zero plot is
given by Figure 1 4 .2 2 . One immediately notices that the poles are very close to the point yl on
the imaginary axis with the magnitude response peaked. Further, the zeros at w = 0 and co = x are
where the magnitude response is zero.
710 Chapter 14 • Laplacc Transform Analysis III: Transfer Function Applications
Pole-zero plot
- I --------------------- !--------------------- 1--------------------- :— ! 1 1^
1
0.8
: :
0.6
0 .4
S. 0.2
:
I 0
■
- 0.2
- -0 .4
- 0.6
:
- 0.8
-1
1 \
— I--------------------- --------------------- --------------------- : f 1 1
-1 .5 -1 -0 .5 0 0 .5 1 1.5
Real part
E X A M P L E 1 4 .1 1 . Iwo circuits have transfer functions H^{s) and / / 2 W, with the pole-zero plots
shown in Figures 14 .2 3 a and b, respectively, and gain constants o f 1. Qualitatively speaking (with
out doing any com putations), what can we deduce about the magnitude response o f each circuit?
To verify our qualitative deductions we will use M ATLAB to construct the exact magnitude
response.
JCO jO)
A A.
.. 12j X -• 12j
- 8j - 8j
- 4j - 4j
0 ----------© > a
II
- -4j ■ -4j
- -8j - -8j
- .- 1 2 j X - .-1 2 j
T >r
(a) (b)
FIG U R E 1 4 .2 3 Pole-zero plots for Example 14.11. (a) H^{s) (b) Hjis).
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications 11
S o l u t io n
P art 1. Qualitative analysis o f Figure 14.23a. This figure tells us that there are r\vo finite poles o f
H^{s) near ±\2j, but in the left half o f the complex plane, and that there is a finite zero at the ori
gin. So there is also a zero at y = x . Thus
(i) \H{jto)\ -♦ 0 as ( 0 -» 0
(ii) \H{j(o)\ -* 0 as (O -♦ oo
(iii) \H{jco)\ = maximum value as co -♦ 12
In the case o f point (iii), we can say in general that in some neighborhood o f to near the pole, the
transfer function peaks in magnitude.
P art 2 . Qualitative atialysis o f Figure 14.23b. This figure tells us that there are again rvvo finite poles
of Hjis) near ±12y, but in the left half o f the complex plane, and that there arc two finite zeros at
the origin. So there is no zero at ; = oo. Hence
(i) \H{jco)\ -♦ 0 as oj -» 0
(ii) \H{jo))\ constant as to -» oo
(iii) \H{jo})\ ~ maximum value as co 12
Again we cannot make a stronger general statement in point (iii) above.
P art 3 . Magnitude plots via MATLAB. Suppose the poles are at -0 .1 ± y l2 and we desire the actu
al magnitude response plot for A/, (y’co), 0 < co < 3 0 rad/sec. To construct the plot shown in Figure
1 4 .2 4 , we use the following M ATLAB code:
»w = 0:. 1:2 0 ;
»n = [1 0];
»d = poly([-0.1 + 12*j -0 .1 -1 2 * )]);
»h = freqs(n,d,w);
»plot(w, abs(h))
»grid
»xlabel(‘Frequency in rad/sec’)
»ylabel(‘Magnitude H 1 (jw)’)
Exercises. 1. Suppose the poles o f H^{s) are at -0 .1 ± j\ l and that we desire the actual magnitude
response plot for 0 < to < 3 0 rad/sec. Use iMATLAB to construct this plot. Verif)' the accuracy of
the qualitative predictions.
2. Show that the transfer function
■> 1
.V" +
H{s) = LC (1 4 .1 1 )
1
s~ + •.y+
RC LC
o f an unknown circuit with R= 1 Z. = 0.1 H , and C = 1 mF has the band-reject magnitude
response depicted in Figure 14 .2 5 . Use M ATLAB and the “freqs” command.
Frequency in r/s
FIGURH 1 4 .2 5 Plot of the magnitude response of the transfer function of equation 14.11.
The example below further illustrates this relationship with a pedagogicaily useful graphical tech
nique. Mastering this technique helps concretize the meaning o f magnitude and phase and rein
forces the qualitative discussion above on using pole and zero locations to compute the magnitude
and phase.
E X A M P L E 1 4 .1 2 . 'lb better grasp the ideas o f the magnitude, |/y(/'to)|, and the phase, Z.//(ja>),
o f a frequency response, suppose a transfer function has the form
(s-Zi)(s-Z2)
H{s) = (1 4 .1 2 )
is+]){s + \ + j ) { s + \ - j )
where Zj = 2j and = -2j. Figure 14 .2 6 a shows the pole-zero plot o f H{s).
Chapter 14 • Laplace Transform Analysis ill; Transfer Function Applications ~ 13
JO )
JO )
A
() 2j
X ■
■ j
-1
X I -J
-2j
(a)
FIG URE 14.26 (a) Pole-zero plot of H{s) as given by equation 14.12. (b) Measuring distances
graphically from zeros and poles to the point j\.
S o l u t io n
The plan o f this example is to compute the magnitude o f H{jio) graphically for oj = 1 , i.e., to com
pute \H{J\)\. This computation entails the following steps.
Step 1. Observe that j\ - defines a complex number. Think o f J\ - 2 , as a vector in the com
plex plane. Similarly,yi - Zj and J] - p-, where /> ] = —! and p 2 ,^ = - 1 ± 7 . define vectors. Each vec
tor has a length that can be determined either graphically, by physically measuring the distance
with a ruler, or by the Pythagorean theorem. Figure 14 .2 6 b illustrates the idea.
Step 2 . Following from step 1, the magnitude o f //(/I ) has the form
Step 3. Suppose we wish to compute the phase or angle o f H{j\) graphically. In Figure l4 .2 6 b ,
observe that each complex number viewed as a vector jio - z - ory ( 0 - p- can be represented in the
form x^\ where ip is the angle the vector makes with a horizontal line passing through its base.
For example, {j\ - J2) = -j\ = From basic complex number theory, the angle o f the prod
uct o f two complex numbers is the sum o f the angles, and the angle o f the ratio of two complex
numbers is the angle of the numerator minus the angle o f the denominator. Fience, from the
angles shown in Figure 14 .2 6 b ,
= (1 4 .1 3 )
{s - P\){S - P2 )---{s- P„)
Because the magnitude o f a product o f complex numbers is the product o f the magnitudes o f the
numbers, and because the magnitude o f the ratio o f two complex numbers is the ratio o f the mag
nitudes o f the numbers, the general form o f the magnitude response o f equation 14.13 is
- Zx |yto-C2 •••
Similarly, since the angle o f the product o f complex numbers is the sum o f the angles o f the num
bers, and since the angle o f the ratio o f two complex numbers is the difference in the angles o f the
numbers, the general form o f the phase response o f equation 1 4 .1 3 is
Thus, qualitatively speaking, H{jn)) tends to have a large magnitude for jco’s near poles and a small
magnitude for jco’s near zeros. As mentioned earlier, this can be used to advantage in estimating
the magnitude response and phase response o f a transfer function.
Exercise. Draw an estimate o f the general shape o f the magnitude and phase response o f the
Butterworth normalized low-pass transfer Kinction
M i> =
5“ + Jls -I- 1
Com pute the exact magnitude and phase at lo = 1. W hat happens to the magnitude and frequen
cy response if H{s) is changed to H^{s) = H{sl\Q) and H^{s) = / / ( / / 1 00)?
A N SW ER S: in random order: - 4 5 . O.TO"^, the general shape is the same wich H = /7 i(/]()0 )
= 0 .7 0 7 .
Design of filters and amplifiers often begins with a design template in which almost all parameter
values are normalized. In particular, a source or load resistance is often set to 1 t l . Also, a critical
or im portant frequenq^ is set to 1 rad/sec. Such circuits are called normalized. W ith the comple
tion o f a normalized design, engineers can frequency-scale to obtain realistic frequency responses
and magnitude-scale to obtain reasonable impedance levels as necessary to meet power and ener-
g)' restrictions.
Chapter 14 • Laplacc Transform Analysis III: Transfer Function Applications
FR EQ U EN C Y SCA LIN G
FIG U R E 1 4 .2 7 Band-reject type circuit that realizes the transfer function of equation 14.11.
S O L U T IO N
Step 1. Calculate the circuit transferfunction. By voltage division,
*
R LC
H{s) =
R +-
RC LC
Cs ------
Ls
Step 2 . Incorporate the effect o f /y.-
R R
R+ R+
C 1 J s\ I
■s +
Kf
I s V 1
+
LC
^ andA T
C ^
K K
with R unchanged, the new and old transfer functions are related in a ver>' simple way:
Chapter 14 • Laplacc Transform Analysis III: Transfer Function Applications
/ \
.y
A'
I )
Step 3. Plot the magjiitude response. Rather than go directly to M ATLAB, consider that
to
^old
Kf I
Table 14.1 depicts this relationship for specific values o f id.
TABLE 14.1
0
I \
1 ~ ^old J
\ ^ Jf /
In general, the new transfer function is related to the original H^[J,s) through the replacement of
s by s!K^. T he implication o f this relationship is that whatever happened at ^ = before scaling
must now happen at s= where the term frequency scalhig. In terms
o f the plot o f our specific transfer function, then, we obtain from Figure 14.25 the new plot o f
Figure 14. 28 by inspection.
For general linear circuits, vvc will only state the procedures for scaling and their effects, leaving a
rigorous mathematical justification to a more advanced text on circuit theory.
Procedure for frequency scaling: To frequency-scale a linear network by a positive scale factor Kj-
1. Divide all inductances and capacitances by / y
2. Leave all resistance values and controlled source parameters unchanged.
1-
which means that has zeros and poles at KjrZ- and for / = 1, . . . , w and k=
3. The magnitude and phase response cur\'es o f are those o f with the fre
quency' scale multiplied by / y Conversely, the magnitude and phase o f //(/to ) are the
same as
find the new pole and zero locations if the circuit is frequency-scaled by 1000.
AN SW I-RS: l0{)0/>,. !()()()/>„ and HHHb,
OAs
i*” + 0.1 1
Figure 14 .2 9 b shows the pole-zero plot and Figure 14 .2 9 c shows the magnitude response with
peak value at = 1 rad/sec.
(a) Frequency-scale the circuit by the factor /y = 10^ and compute the new transfer function.
(b) C om pute the new pole-zero plot and the new magnitude response curve.
/YY\
-I-
V J - ) > V
0.1 Q >
(a)
718 Chapter 14 • Laplacc Transform Analysis 111: Transfer Function Applications
JCO
FIG U R E 14.29 (a) Series resonant circuit, (b) Pole-zero plot, (c) Magnitude response.
S o l u t io n
(a) Dividing the capacitance and inductance by Kj-= 10*^ yields the circuit o f Figure 14. 30a.
JO )
(-0.05+ j0.999)10^ X
for
Hnew(s)
o—
(b)
FIG U R E 1 4 .3 0 Demonstration of the effects of frequency scaling, (a) Frequency-scaled circuit, (b)
New pole-zero plot, (c) New magnitude response curve.
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications
S 10^ /
10^/ / 5 ^ ( S
+ 0.1 +1
UO^i lio^/
(b) The ideas stated in part (a) are borne out by comparing the pole-zero plot and the magnitude
response curve for shown in Figures 14. 30b and c, with their counterparts for H{s) in
Figure 14 .2 9 . Notice that the pole locations have been scaled by Kj-^s per property 2. A direct
computation in M ATLAB verifies this:
Also, in Figure 1 4 .2 9 c for H{s), the peak response occurs at co^/^= 1 rad/sec, whereas in Figure 14.
3 0 c for the peak response occurs at = 10^ rad/sec.
At this point we round out our discussion o f frequency scaling by relating it to the time domain
via the time/frequency scaling property o f the Laplace transform:
or, equivalently.
( s \
K fl =H
\ j /
where H{s) is the circuit transfer function and h{t) is the circuit im pulse response.
'2 0 Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications
30 10 10
H(s) =
(5 + 2 ) ( 5 + 5 ) {s + 2) (5 + 5)
Find the impulse response when the transfer function is frequency-scaled by Kj:
S o l u t io n
10^ , ioa :,
f-lneAs) =
(5 + 2 /r y )(5 + 5/C|) (^ + 2A:/-) (5 + 5 /:^ )
/W (0 = - \QKfe~^^Au{t)=Kf h[Kfi)
Suppose for the sake o f argument that / y = 10^. Let us plot h{t) and as has been done in
Figures 1 4 .3 la and b.
t In mllli-sec
(a)
Chapter 14 • Laplacc Transform Analysis 111: Transfer Function Applications ^21
t in milli-sec
(b)
Observe that the plots are structurally the same. However, the time scale in Figure 14.30b is
in milliseconds and the magnitude has been scaled by Kj~= 10-^.
Exercises. 1. Use the Laplace transform time scale propert)' to show that in general
2. In Example 14 .1 5 , suppose the poles are changed from - 2 and - 5 ro - 3 and - 6 , respectively.
Suppose further that the transfer function is scaled by 10. Find the impulse response and the
new impulse response after scaling.
A N SW ER : /;(/) - - 10^--^’0//(/) and h, (/) = (lOOt’"-^"' - I
Frequency scaling has allowed us to obtain realistic frequency responses from normalized circuits.
Another technique in achieving a realistic design is magnitude or impedance scaling. A simple
example illustrates the idea.
Chapter 14 • Laplacc Transform Analysis III: Transfer Function Applications
1O lOOOQ
— — ------0 — —
+
^ 10000
1O
-O
(a) (b)
F K iU R E 1 4 .3 2 (a) A normali7.ed voltage divider circuit, (b) The voltage divider of part
(a) magnitude-scaled by 1000.
S olution
By direct calculation, the voltage ratio remains unchanged, i.e.,
which is 1000 times larger. T he elements in Figure 14 .3 2 b are said to be magniti(de-scaled (by
1000) from those o f Figure 14 .3 2 a .
The above example motivates a more general discussion. Suppose each impedance in Figure
•14 .3 3 a is scaled (multiplied) by to yield the circuit o f Figure 14 .3 3 b . As in Example 14.16,
the voltage ratios remain the same for both circuits:
. . («)
Further, the input impedances o f Figures 14 .3 3 a and b are related as
Z,(s) K .Z ,(s)
-o
+
V r A
Z,(s) J
Z ---w
-o
(a) (b)
FIG U R E 1 4 .3 3 . (a) A general voltage divider, (b) Magnitude scaling of the circuit in part (a) by
T he above discussion suggests that magnitude scaling can be achieved by multiplying resistances
and inductances by and dividing conductances and capacitances by This and several other
properties are stated next.
P rocedure for m agnitude scaling: To magnitude-scale a linear network by a scale factor K^.
1. Multiply all resistances and inductances by
2. Divide all capacitances and conductances by
3. For current-controlled voltage sources (C C V S ), i.e., the type, multiply the parameter
4. For voltage-controlled current sources (V C C S), i.e., the^^^ typ^> divide the parameter
by
5. Parameters for voltage-controlled voltage sources and current-controlled current sources
remain unchanged.
6. Ideal operational amplifiers remain unchanged.
The effect o f magnitude scaling on a transfer function is set forth below, with its verification left
for more advanced courses on circuit theory.
Effect o f m agnitude scaling on transfer functions: If H{s) is a voltage ratio or current ratio,
magnitude scaling has no effect on H{s). If H{s) has units o f ohms, then the magnitude-scaled net
work has If H{s) has units o f siemens or mhos, then the magnitude-scaled net
work has
, I s 2 + 0 .I ,v + l
Z{s) = — i-5-k 0.1 =
C s
and transfer function
0.l5
v;-,, Z(5) .r-hO.I.y+1
_JYY\
Im F 1000 H
V V.
Znew(s)
(b)
FIG U R E 1 4 .3 4 Magnitude scaling, (a) Original network, (b) Network magnitude scaled by 1000.
So lu t io n
(a) Following the procedure for scaling R’s, L's, and C s, we obtain the scaled network shown in
Figure 14 .3 4 b .
(b) For the simple circuit o f Figure 14.34b , the new input impedance is
Z„e,,.(s) = + 1 0 0 0 s + 1 0 0 = 1 0 0 0 ‘^ + 0 - 1 ^ + ^ ^ l O O O Z (s ) = K,„Z{s)
H 100______________ l l i = H(s)
.y“ + 0 .h + l .V-+0.1.V + 1
1000
Because the transfer function is a voltage ratio, there is no change. These results clearly illustrate
the stated effects o f magnitude scaling.
old
2. Two parallel resistors and /? 2 are magnitude-scaled by Verify that Rf,q =^m^eq
Hint. Use the formula for parallel resistance to compute and then relate
that to ■
Moving from normalized circuit design to realistic circuit design ordinarily entails both magni
tude and frequency scaling. This subsection provides several illustrative examples.
Chapter 14 • Laplace Transform Analysis III; Transfer Function Applications
EXAM PLE 14.18. The amplifier circuit shown in Figure 14.35a consists of r\vo stages. The first
stage is a Sallen and Key low-pass active filter. The transfer function of the Sallen and Key circuit is
V
^Ollt____-'>
-
H2(s ) =
V, .v +
V
^Olll _ V * out. V ^(1^ = - " ^ 1 -1
His) = ........................... — ------------------------------------------------------------ -
y. y y {s + 1) (.v“ + ^ + 11 5 ' + 2 . r + 25 + 1
FIGURE 1 4 .3 5 A third-order Butterworth low-pass filter, (a) The original network, (b) The scaled
network with K^= 2000ti and = 10,000; = 3 1.4 uS.
The overall transfer function o f the two-stage amplifier is a third-order, maximally flat
(Butterworth) low-pass filter, which will be studied in Chapter 19. The magnitude response o f the
transfer function has a 3 dB down frequency at OJ = 1 rad/sec, or l/(2ji) Hz = 0 .1 5 9 Hz— an
extremely non-useful audio frequency. In fact, when the 3 dB down frequenc)' is at (o = 1 rad/sec,
the magnitude response is said to be normalized. N ot only is 1 rad/sec not useful, but the element
values are unsuitable for practical applications. However, using both magnitude and frequency
scaling, this “reference” or normalized amplifier circuit can be made into a very practical filter.
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications
Suppose we wish to have the 3 dB down frequency a t^ ^ ^ = 5 0 0 Hz and the largest capacitor at
10 nF. Such a filter could be used to direct the low-frequency content o f a music signal to a woofer.
O ur goal requires that we frequency-scale the circuit by Kj-= IOOOtt and magnitude-scale such that
the 2 F capacitor (the largest) becomes 10 nF; i.e., must satisfy
10 X K )-'' = ^ ^
K jK ,„ lOOOrcK,,,
(5") cs
Exercise. A circuit has transfer function H{s) = — = -^ ------------- ,with a = Q.\, b = Aand c = ^ 0 .
Avj s + fl.v + h
After both magnitude and frequency scale changes o f = 4 and Kjr= 2, compute the new a, b,
and c.
A N SW E R : in random order: - 3 2 0 . 0 .2 , and 16
Examples 1 4 .1 5 and 1 4 .1 6 illustrate one o f several reasons for scaling a linear nersvork. Chapter
19, on elementary filter design, will derail further applications.
In system theory, and especially in control theory, engineers want the output signal o f a circuit or
system to track a given reference signal. T he idea behind the term nack is that for large t, the ref
erence signal and the circuit output are more or less indistinguishable. To accomplish this,
design/control engineers generate an error signal, e{i) = y(i) ~ yref^)^ where is the circuit out
put and is a given reference signal. Since much o f the analysis is done in the frequency
domain, one often knows E{s) = Y{s) - without knowing the related time functions.
Ordinarily, the design engineer needs to know the initial error, if(0), and the final error, f(co).
Available to engineers are the initial-value theorem and the final-value theorem, which permit the
computation o f these quantities in the frequency domain.
Initial-value theorem: Let = F{s) be a strictly proper rational function o f j, i.e., the numer
ator and denominator o f F a r e both polynomials in s, with the degree o f the numerator less than
that o f the denominator. Then
lim 5fCO = / ( 0 ^ ) (1 4 .1 5 )
Applying the definition o f the Laplace transform to the right-hand side and taking limits as s
approaches infinity implies that
l i m [ j f ( .v ) - / ( ( ) " ) ] = lim
,V -» X s-*»
(1 4 .1 6 )
where the dot over the function/(r) indicates the derivative o f the fimction. Observe that
Hence, the left-hand side o f equation 1 4 .1 6 equals the right-hand side o f equation 14.17.
Equation 14.15 follows from equating these two sides and canceling th e /(0 “) terms in both.
2 1
.V 5i + 2
S o l u t io n
By direct application o f the initial-value theorem,
EXAMPLE 14.20. Let the Laplace transform o f the velocity of a certain projectile be given by
, 500^ + 20
V{ s) = -----------------------------
i-(5 5 -h 2 0 )(1 0 .v + l)
Find the initial velocity, y(0'^), and the initial acceleration, //(O"^).
S o l u t io n
l b find the initial velocity, we directly apply the initial-value theorem:
5 0 0 ^ -f 2 0
v(O^) = lim sV(s) = lim =0
S-^ o o s —* x (55 + 20)(10. v+ 1)
Since acceleration is the derivative o f velocity, and the initial velocity is zero, from the time dif
ferentiation propert}'^ o f Table 12.2, assuming the velocity is continuous at f = 0,
•2S Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications
500.V + 2 0
/\(.v) = sV{s) - v(o^
(5.s + 2 0 )(1 0 .v + l)
20
5 () 0 i- + 20 s
f/(0 ) = Urn = Urn = 10
.V- (5.V + 20X10.S + 1) i-^oo 20 \
5.V + IO.V+-
.y / s!
T he initial-value theorem and the examples that follow it illustrate the computation o f initial val
ues. To compute final values vve use the next theorem.
Final-Value T h eorem : Suppose F{s) has poles only in the open left half o f the complex plane, with
the possible exception o f a single-order pole at ^ = 0. Then
lim sF{s) = Urn / ( / ) .
(1 4 .1 8 )
Proof:T\\Q condition o f the theorem, i.e.. the condition that F{s) has poles only in the left half of
the complex plane, with the possible exception o f a first-order pole at the origin, guarantees that
the limit on the right side o f equation 1 4.18 exists. This is because a partial fraction expansion of
F{s) leads to a time fu n c tio n //) that is a sum o f exponentially decaying signals and at most one
constant signal. Since the right-hand limit is well defined,
= { lim / ( / ) ''- / ( ( ) ■ ) .
\/-»X /
This implies equation 14.18.
500^ + 20
V ( .) =
.v(55 + 20)(10.v-h l)
500.V + 20
( 5 i + 2 0 ) ( 1 0 .v + l )
S o l u t io n
K(y) and A{s) have poles that meet the conditions o f the final-value theorem. Hence,
5005+20
lUn \'(t)= Urn ^ y ( 5 ) = lim = 1
I-*X .v-^0 ,v-*-0 (55 + 2 0 X 1 0 5 + 1 )
and
5 0 0 5 “ + 20 5
lini a(t)= lim 5y\(.0 = lim = 0
i-* x > .v-»() .v-^o ;5 5 + 2 0 )( 1 0 5 + 1)
Observe that a constant final velocity implies a zero acceleration as these expressions indicate.
E X A M P L E 1 4 .2 2 . W hat if the conditions o f the fmal-value theorem are not met? W hat would
go wrong? A simple example illustrates the problem. Let
Fis) = -
5” + 1
is undefined, i.e., it does not exist. The theorem, however, presupposes that both limits exist.
Again, the condition of poles in the leh half complex o f the plane with at most one pole at the
origin is necessary and sufficient for both limits to exist.
Exercises. 1. If F{s) = (6y + 10)/(2^^ + 4^), then, by the final value theorem, f{t) approaches what
value for large r?
A N S W E R : 2.5
1 5 t ^ - - - '( 5 - + 5 - 2 )
.v(5" + 5 5 + 6 )
8. BODE PLOTS
Section 5 described the use o f the poles and zeros o f N(s) to compute the frequency response o f a
circuit. In this regard, Hendrik Bode developed a technique for computing approximate or asymp
totic frequency response curves. These so-called Bode plots can be quickly drawn by hand. A
description o f the technique requires the introduction o f some terms widely used in the engineer
ing literature.
Let //(s) be a transfer function that is a dimensionless voltage ratio or a current ratio. As explained
in section 4 , for sinusoidal steady-state analysis, one replaces s by y’oj to study the circuit’s magni
tude response, |//(/w)|, and phase response, Z .//(/co). For convenience, let |//(/to)| be a voltage
gain, |K,/Kj|. T he ga/;/ in dB (decibels), denoted by is defined by the equation
= 2 0 log,„|//(/o)| (14.19)
For convenience, whenever we write log(.v), we will mean log,Q(>;). Solving for |//(/w)| in equation
1 4 .1 9 yields the inverse relationship
(1 4 .2 0 )
Table 14.1 presents some pairs o f |/y| and Thus, instead o f saying that |K,| is 10 times |Kj,
we may say that is 2 0 dB above K,, or that Kj is 2 0 dB below ¥2 - Similarly, to say that V-, is 3
dB above K, means that |K,| is 1.414 times |Kj|.
1 V2 2 10 100 1000
^dB 0 s3 ^6 20 40 60
Exercise. A certain amplifier has a dc gain o f 80 dB. W h at is the actual voltage gain?
A N S W E R : H)-'
O ne of the reasons for using the dB terminology is that it simplifies the analysis and design of
multistage amplifiers. Suppose an amplifier has three stages with voltage gains equal to 2 0 , 100,
and 10, respectively. The overall voltage gain is the product o f the gains of each individual stage,
which is 2 0 X 100 x 10 = 2 0 ,0 0 0 . Using the dB specification, the overall gain in dB is the sum of
the dB gains o f the individual stages. This is (2 6 + 4 0 + 2 0 ) = 8 6 dB. It is easy to justify this claim.
First,
This summation has pronounced advantages for repetitive calculations at many frequency points,
as when plotting a magnitude response such as equation I4 .l4 a . We could, of course, convert this
equation to an equation having all terms in dB. However, with an eye toward a further simplifi
cation, it is desirable to first rewrite H{s) in a slighdy different, but equally general, form, namely
+ 1
-Z 2 (14.21)
H(s) = Ks'‘ j '
^ - + l'
V -P i
where \p^, p 2 >—} are those poles o f H{s) that are not at the origin and {zp 2^, ...} are those zeros
of H{s) that are not at the origin; if a is positive (negative), then H{s) has a zeros (poles) at the
origin. For example, a transfer function
(^ + 5 0 ) (5 + 2 0 0 )
(5 + 8)( j + 6 0 0 )
has the equivalent form
+ 1 + 1
O ' H {s) = 2.5
^50 Uoo
(14.22)
+ 1
(f-l ^,600
Observing that = 10 log^^|//(70))p, setting s = /a) in equation 14.21, and noting that the mag
nitude squared o f a complex number is the imaginary part squared plus the real part squared yields
0.5 0.5
v2
(0
HdBi(o) = \K dB 0) dB + + 1
\Z2/ (14.23)
dB dB
2 ,0.5 -,0.5
( co\
dB
Equation 14.23 suggests that we may compute the dB vs. (O curve for each term on the right-hand
side and graphically sum the curves to obtain the desired H jg vs. (O curve. However, each indi
vidual curve is reasonably sketched by using log(o)) instead o f 0) as the independent variable. This
amounts to using semi-log paper to plot the dB vs. (O curves. Such a plot, with a linear scale for
the dB values on the vertical axis and a logarithmic scale for O) on the horizontal axis, is called a
Bode magnitude plot, in honor o f its inventor. Similarly, a plot of LHijiiS) vs. O), with a linear
scale for Z.//(/co) and a log scale for co, is called a Bode phase plot. Note that, because of the log
arithmic scale for the (o-axis, the actual distance on the paper between (O = 1 and O) = 10 is the
same as that between o) = 0.1 and O) = 1. (See Figure 14.36a.) Note also that the (O = 0 point will
not appear on the graph, because log(co) approaches -oo as O) approaches 0.
W
"3 2 Chapter 14 • LiplaceTransform Analysis III: IVansfer Function Applications
With log(oj) chosen as the independent variable, the plot of each term in equation 14.23 either is exact
ly a straight line or is a curve having two straight line asymptotes. I'his is illustrated in Figure 14.36.
FlG lIR Ii 1 4.36 Bode magnitude plots for three basic transfer functions,
(a) H{s) = K> 0. (b) H{s) = sin, a > 0. (c) H{s) - s!a +\, a > Q.
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications 733
In Figures 15.36b and 15.36c, the rising asymptote has a slope of 20 dB/decade, which means that
along this line, an increase in frequency by a factor of 10 causes an increase in gain of 20 dB. Here,
the word decade (abbreviated dec) simply means “10 times.” Another way to express the same
slope is to indicate the increase of gain in dB when the frequency is doubled, or increased by an
octave, in music terminology. It is easy to see that 2 0 dB/dec is equivalent to 6 dB/octave. In gen
eral, if a frequency, (Oj, is d decades above another frequency, coj, then, by definition, (cOj/cOj) =
lO'^. Conversely, if (0)2/t0 j) = r, then we say that (Oj is log(r) decades above (Oj.
In Figure 14.36c, the left asymptote is a horizontal line and hence has a slope of 0 dB/dec. The
point where the two asymptotes intersect is called the breakpoint, and the corresponding fre
quency is called the l>reakfrequency or comer frequency.
The derivations o f the true curves and asymptotes in Figures 14.36a and 14.36b are very simple
and are left as exercises. For figure 14.36c,
in which case
+1 + 1
■ B
and
| H (» | ^ = 201o g +1
( l) '
For 0) « a.
|H(j<o)|jB = 201o g [-y r] = 0
indicating that \H{j(a)\^g approaches the left asymptote in the figure. For co » a,
(I)'
|«Oo))|dB = 201og ,1 - ) +1 as 20 log - =201og(co)-201og(a)
V\a ) a
indicating that |^(/(o)|^^ approaches the right asymptote in the figure. The two asymptotes inter
sect at the point ( cd = a, = 0). At this corner frequency, the largest error, 3 dB, occurs
between the true value of |//(/co)|^^ and the value read from the asymptotic curve. The error at
twice or half the corner frequency is about 1 dB.
The following variations of the three basic Bode magnitude plots o f Figure 14.36 are easily
derived:
1. If H{s) = {sld)^y the Bode magnitude is similar to that shown in Figure 14.36b, except
that the slope is now 2 0 « dB/dec. The curve still passes through the point (co = a,
2. If H{s) = {sla + 1 ) ” , the Bode magnitude is similar to that shown in figure 14.36c, except
that the right asymptote has a slope o f 2 0 « dB/dec. The breakpoint is still at (o) = a,
|//(/co)|^^ = 0), and the error at the comer frequency is 3 « dB. If n is negative, the right
asymptote points downward.
734 Chapter 14 • Laplace Transform Analysis 111: Transfer Function Applications
Let us now consider the Bode plot for a general transfer function H{s). After expressing H{s) in
the form of equation 1 4 .2 1 , we can draw the asymptotes for each term in equation 14.23 with the
aid o f Figure 1 4 .2 6 . The asymptotes for the log(o)) curve can then be constructed very eas
ily by graphically summing the individual asymptotes. Since the asymptote for each term in equa
tion 14.2 3 is a piecewise linear curvey the graphical sum o f all the asymptotes is also a piecewise lin
ear curve. Accordingly, it is not necessary to calculate the sum o f dB values at a large number o f
frequencies. If there are n break frequencies, then the summation need only be carried out at these
frequencies and for the slopes o f the leftmost and the rightmost segments o f the piecewise linear
asymptote. T he following example illustrates this procedure.
E X A M P L E 1 4 .2 3 . Obtain the asymptotes for the Bode plot o f H{s) o f equation 1 4 .2 2 rewritten
below:
I s
— +1 -M
V50 / V200 /
H (s)= 2.5
+ 1
V8 / V6 0 0
Solution
There are four break frequencies. Rewriting H{s) as a product o f five factors with break frequen
cies appearing in ascending order yields
-1
■ .V
H{s) = 2.5 - + 1 — + 1 +1 +1
.8 .5 0 .2 0 0 .6 0 0
i}4.24)
= / / , X / / . X //3 X //4 X
Figure 14.3 7 a shows the asymptotes for the five individual terms in equation 1 4 .2 2 , and Figure
14 .3 7 b shows the asymptotes for
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications ^^5
CO (rad/sec)
(b)
FIGURE 14.37 Asymptote and Bode magnitude for H{s) of equation 14.22 or, equivalendy, equadon
14.24. (a) Bode plots for / / , through Hy (b) Bode plot for equal to sum of those for through
8 - 2 0 X 0 .7 9 6 = - 7 .9 2
- 7 .9 2 + 2 0 X 0 .4 7 7 = 1.63
6. Finally, consider the slope o f the rightmost segment. Since contributes an addi
tional-20 dB/dec to the slope for co > 6 0 0 , the resulting slope o f the rightmost segment
is zero.
T h e com p lete specification o f the piecewise linear asym ptote is shown in Figure 14 .3 7 b .
'3 6 Chapter 14 • Laplacc Transform Analysis 111; Transfer Function Applications
_ out _ K
«(■') =
^in 1 4 -£
show that (i) the dc gain is K and (ii) for s = ja, i.e., at co = a, the gain is 3 dB down from the dc
value. I his to is called the 3 dB frequency or 3 dB bandwidth of H{s).
2. Construct the piecewise linear asymptote for the transfer function H{s) = 40j-/[(i- + 2){s + 2 0 )].
O nce the piecewise linear asymptote for the Bode plot has been constructed, the true curve can
be sketched approximately by noticing that the error at each corner frequency is about 3w dB (pro
vided that the two neighboring corner frequencies are more than five times larger or smaller).
Figure 14 .3 7 b shows such a rough sketch. In the pre-com puter era, the ability to draw such a curve
by hand— even a crude approximation— was considered valuable. Nowadays, with the wide avail
ability o f personal computers and C A D sofuvare, one might just as well get the exact plot with
out bothering to look at the straight-line asymptotes. From the perspective o f circuit analysis, the
value of the ability to construct the asymptotes for a Bode plot is greatly diminished, but the tech
nique is still important for its application in the design o f feedback control systems. Such an appli
cation utilizes both the Bode magnitude plot and the Bode pha.se plot. Some background in con
trol systems is required for one to understand the use o f the Bode technique in this kind o f appli
cation. We will relegate the discussion of the topic to a more advanced course in feedback control.
O ur objective in mentioning it in this text is twofold: (1) to introduce the definitions o f some
com m oni)’ used terms, such as decibels, decade, and octave, and (2) to demonstrate a highly sys
tematic procedure for adding up several piecewise linear curves to obtain a desired curve, as
described in Figure 14.37.
The electromechanical characteristics o f the permanent magnet dc m otor have a simple circuit
like model amenable to Laplace transform analysis. The model is given in Figure 14.38 and con
sists o f an adjustable dc voltage source in series with a resistor an inductance and a device
labeled “m otor.” Here, represents the resistance present in the armature winding, and rep
resents the equivalent inductance o f the wire coil. The device labeled “m otor” has the current i J^t)
as an input and the angular velocity co(r) o f the rotating shaft as an output. T he interaction
Chapter 14 • Laplace Transform Analysis 111: Transfer Function Applications
between the electrical part o f the model and the mechanical part o f the model occurs at the loca
tion o f this symbol. T he voltage is an induced voltage proportional to the angular velocity
oj(r). Because co(/) is not a circuit variable, classical notions o f impedance, admittance, voltage
gain, etc., do not fit the problem, whereas the more general notion o f transfer function does, forc
ing us to move slightly beyond the confines o f circuit theory proper to analyze the system.
The torque J\t) produced on the rotating shaft by the current flowing through the armature coils
is proportional to the armature current /^(/), i.e..
Tit) = (1 4 .2 5 )
T he mechanical rotation o f the m otor affects the electrical portion of the system by inducing a
voltage This voltage is proportional to the m otor’s rotational speed, or angular velocity, co(r),
i.e.,
= (1 4 .2 6 )
Since the m otor converts electrical energy to mechanical energ^^ conservation of energy dictates
that the constant o f proportionalit)' be equal to the same constant that relates torque and current
for a lossless motor. Specifically, electrical power-in must equal mechanical power-out or, in equa
tion form.
= m coit)
which forces = k.
O ur first goal is to find the transfer function o f the motor, H{s) = £[o){t)]/V^{s). For convenience,
let Q(^) denote X [o)(r)]. As a first step, sum the voltages around the loop of elements in the cir
cuit model of Figure 14.3 8 . This results in the differential equation
V (/) = ) (1 4 .2 7 )
at
A ssum ing zero initial con ditions, the Laplace transform of equation 1 4 .2 7 is
From basic mechanics, the differential equation governing the mechanical portion o f the system
r(o = y „ ^ + B < o (o
at
where is the moment o f inertia of the combined armature, rotating shaft, and load; B is the
coefficient of friction; and 7\t) is the torque produced by the motor to turn the load. Recalling
equation 14.25, T\t) = the Laplace transform o f equation 14.29, assuming zero initial con
ditions, is
(14.30)
Solving equation 14.30 for IJis), substituting the result into equation 14.28, and then solving for
Q(j) leads to the expression
Q(s) = 2 V,{s)
(14.31)
Equation 14.31 characterizes the pertinent dynamics of the permanent magnet dc motor and
allows one to find the angular velocity o f the motor shaft as a function of time for given inputs.
To see how the motor responds to step inputs, suppose vjit) = Ku{t). As objectives, let us find (1)
the steady-state value, or final angular velocity, o f the shaft and (2) the steady-state value o f the
armature current. The final speed o f the shaft is important because, for example, one needs a fixed
speed for the rotation o f a compact disk or a fen. The final or steady-state current is important for
determining the power needed from the source.
If Vj(t) = Ku{t), then V(^s) = KIs. It follows from equation 14.31 that
k
K (14.32)
2\
RgS + k
S + S+
J
Applying the final value theorem to equation 14.32 implies that
k
(Ovv = lini co(r) = ------- y K
To isolate the armature current IJJ), again apply the final-value theorem to determine the steady-
state value of i {t). Combine equations 14.30 and 14.31 to obtain
1 B
— 5+
4 (^ ) = 2 ’'s ( 14.33 )
LaJm
Chapter 14 • Laplace Transform Analysis III: Transfer Function Applications
Equation 1 4 .3 3 allows us to find the armature current as a function o f time for a given input volt
age. As above, if the input is a step, i.e., if v^{t) = Ku{r), it follows that
1 B
— s + ---------
= - ^ ^ ------------ t t K. (1 4 .3 4 )
R,,B + k~\
S+
\ m 1
O nce again, application o f the final-value theorem to this expression leads to the value o f the
steady-state armature current:
Q
hi . ss ~ ^Cl ~ ^
t —►« B+ k
T he preceding analysis suggests the utility o f the Laplace transform as a tool for analyzing the
dynamic behavior o f electromechanical systems. In fact, system transfer functions o f the form of
equations 14.31 and 14.33 are often starting points for further analysis. Extensions of the above
analysis can be found in the homework exercises.
10. SUMMARY
This chapter has expanded the notion o f transfer function from its definition in Cliapter 13 into
a tool for modeling not only circuits, but other practical systems, such as a dc motor. The trans
fer function characterizes circuit or system behavior by the location o f its poles and zeros. For
example, if a transfer function has a pole on the imaginary axis or in the right half-plane, the asso
ciated circuit or system is said to be unstable, because there is an input or, possibly, simply an ini
tial condition (as in the case o f a second-order pole on the y'oj-axis) that will excite the pole and
cause the response to grow without bound. Further, the ubiquitous presence o f noise will always
excite poles on the imaginary axis and cause the response to be unstable.
This chapter categorized various types o f responses: zero-state and zero-input responses, natural
and forced responses, transient and steady-state responses, etc. Recall that the zero-state response
is the response to an input assuming zero initial conditions, which is the inverse Laplace transform
o f the product o f the transfer function and the Laplace transform o f the input excitation. Recall
that the zero-input response is due only to initial conditions on the capacitors and/or inductors o f
the circuit. The complete response for linear circuits having constant parameter values is simply
the sum o f the zero-input and zero-state responses. This decomposition generalizes to the broad
class o f linear systems studied in advanced courses. Under reasonable conditions, other decom po
sitions are possible, such as a decomposition into the natural and forced responses or transient and
steady-state responses. Other important responses are the impulse and step responses.
For stable circuits, the frequency response provides important information about the circuit.
Recall that frequency response is a plot o f the magnitude and phase o f //(/w ) as to varies from 0
to cc. T he Bode plot is a plot o f gain in dB vs. frequency represented using a log scale. In this con
text it is relatively straightforward to construct an asymptotic approximation using straight line
segments, from which the actual plot is easily sketched by hand. T he information in such a plot
tells us how a circuit behaves when excited by sinusoids o f different frequencies.
(0 Chapter 14 • Ijp la c c Transform Analysis III: Transfer Function Applications
Lastly, this chapter has introduced the initial- and final-value theorems, which provide a quick
means of computing the initial and final values o f a time function from knowledge o f its Laplace
transform. Such theorems have wide application in control system analysis, as evidenced in our
analysis o f the dc servo motor.
means all parameters having ohms as units would be multiplied by and those having
siemens as units would be divided by dimensionless parameters are left alone.
N atural response: the portion o f the complete response that has the same exponents as the zero-
input response.
Octave: a frequenq^ band whose endpoint is twice as large as its beginning point.
Op amp open-loop gain: the gain of the op amp when no feedback paths to the input terminals
are present.
Periodicy(/); function satisfying the condition that there exists a positive constant, T, such that
/ f ) =fijt+ 7) for all f > 0.*
Phase response: the angle o f the frequency response as a ftinction o f (O.
Piecewise linear curve: an unbroken curve composed of straight-line segments.
Pole (simple) o f rational function: zero of order 1 in denominator polynomial.
Poles (finite) o f rational {unction: zeros o f denominator polynomial.
Ramp function, K^)j integral o f unit step function having the form Ktu{i) for some constant K.
Rational fim cdon: ratio o f two polynomials; a polynomial is a rational function.
Stable transfer function: a transfer function for which every bounded input signal yields a
bounded response signal; i.e., all poles are in open left half o f the complex plane.
Steady-state response: that part of the complete response which either is constant or satisfies the
definition of periodicity for t> 0 .
Step response: the response of the circuit to a step function, assuming all initial conditions are
zero.
Transient response: those terms o f the complete response that are neither constant nor periodic
for ^ > 0; i.e., the transient response does not satisfy the definition of a periodic function
for f > 0.
Zero-input response: the response of a circuit to a set of initial conditions with the input set to
zero.
Zero-state response: the response of a circuit to a specified input signal, given that the initial con
ditions are all set to zero.
Zeros o f rational function: values that make the numerator polynomial zero.
O '
O '
* This (nonstandard) definition has been adapted for one-sided Laplace transform problems.
O '
742 Chapter 14 * Liplacc IVansform Analysis llh lr a n s fe r Function Applications
(double pole)
transient and steady-state responses.
jO)
A
-2 -1 )C4j
Figure P I4.1
-q>—►a
-2 1
2. Consider the pole-zero plot of a transfer
function H{s) given in Figure P I 4 .2 . -A]
(a) If the dc gain i s - 1 0 , find/yW -
(b) C om pute the impulse response. Figure P I4.3
(c) C om pute the step response.
C H E C K ; Your answer to (b) should be the 4. Consider the pole-zero plot o f a transfer
derivative o f your answer to (c), since the delta function H{s) given in Figure P i 4 .4 .
function is the derivative o f the step km ction. (a) If the gain H{s) is 4 .8 at ^ = 1, find
(d) If the input is find the pos His).
itive number a such that the response (b) C om p ute the impulse response.
does not have a term of the form Identify^ the steady-state and transient
Ke~‘^‘u{t). Find the zero-state response parts o f the response.
under this condition. (c) C om pute the step response. Identify'
jO) the steady-state and transient parts of
the response.
C H ECK .- Your answer to (b) should be the
derivative o f your answer to (c), since the delta
-e —►o
-2 function is the derivative o f the step function.
(d) If the input is
:-4j 10.9
Find the zero-state response under this (c) If the input to the circuit is =
condition. Identify the transient and 2tu{t), use MATLAB to compute the
steady-state responses. partial fraction expansion and the
resulting time response. Sketch the
jw
ik approximate response for large t.
oj
(d) Repeat part (c) for v-J^t) = lp-u{t).
Figu re P I 4 .1 3
Chapter 14 • i^iplacc Transform Analysis 111: Transfer Function Applications
STABILITY PROBLEMS
14. For each of the circuits below, compute the
indicated transfer function and determine the
range or for which the circuit as m od
eled by the transfer function is stable in the
BIBO sense.
(a) For the circuit of Figure P I 4 .1 4a the
16. (a) The pole-zero plot o f a transfer func
Vi s )
transfer function is His) = , tion is given in Figure P i 4 . 16a.
R =4a, c = 0 . 2 5 F. Construct an input that will cause the
(b) For the circuit o f Figure P I 4 .1 4b the response to be unbounded (unstable),
V (.v) (b) Repeat part (a) for the pole-zero plot
transfer function is His) = .
o f Figure P i 4 .1 6b.
(c) For the circuit o f Figure P i 4 .1 4 c the Imag Axis
®-jco,
(a)
Figure P I4 .1 6
R. +
transfer function „
v;-(.v)
Figure 1P14.20
and the range o f a for which the transfer func
tion is stable.
2 1 . Find the complete response for r > 0 for the
circuit o f Figure PI 4 .2 0 for each o f the follow
ing circuit conditions. (If you properly utilize
the results o f Problem 20 and the principle o f
linearity, the answers to this problem can be
written down directly.)
(a) = 6 V, v-^j{t) = 5 sin(5r)w(r) V
(b) I 6 V > . , W = 8 f - '» M r t V
Figure P I4 .1 9
2 2 . Consider the network o f Figure P I 4 .2 2 .
Suppose C = 0.1 F, v-^j{t) = 10«(^) V, and ^f;(0")
= 4 V; a decomposition o f the complete
RESPONSES AND response has been found to be
CLASSIFICATIONS zero-state response = 10(1 - e~^^u{t) V
2 0 . Consider the circuit in Figure P I4 .2 0 . zero-input response = 4e~~^u{t) V
(a) Show that the transfer function is (a) Determine the transfer function.
(b) If v-^^{t) = 10 co s(10f)«(/) and I'^O”) is
changed to 8 V, com pute the complete
RC 8
H{s) = response for /■ > 0. Identify the
i + io
transient and steady-state parts o f the
response.
when = 50 /?, = 2 0 0 Q , R^^^= R^
C = 2 . 5 niF.
(b) Suppose Vj^^{t) = 10 sin(5/)//{/) V, and
v^Q-) = 8 V.
Chapter 14 • Laplacc Transform Analysis III: Transfer F u n aio n Applications 74
Pure
+ A’C H------ .v +
resistive C/?, .V+ 1.6
vJt) v,(t)
H (i) =
network
5-t- 2
sC + — + — .v-f-
/?1 /?2 CR.(1/
Write as the simi o f two com entiation and integration are linear oper
(b)
ponents: the zero-state response and ations when the initial conditions are
the zero-input response. zero and all circuit parameters are con
Separate the zero-state response into stant. For linear constant parameter cir
(c)
one term due to and another term cuits, then, a linear operation on the
Separate the complete response into on the response, provided that the initial
(d)
the steady-state response and the tran conditions are zero and the circuit
Utilize the answers o f previous parts to this concept with the structure o f the
(e)
write down the complete response for decomposition to obtain the answer.)
(c)
7AH Chapter 14 • Lapbcc IVansform Analysis 111: Transfer Function Applications
K.U'I
(b) Find the zero-state response.
(c) Find the zero-input response.
(d) Find the complete response. 2 7 . Consider the circuit o f Figure P i 4 .2 7 .
(e) Specify the general form of the natural (a) Show that the transfer function is
response.
Specify the transient and is) LC
(0 H{s) =
steady-state responses. KM .9“ -I- .v +
- v„ + L R2 C} R j) LC
16
r + 4.V + 20
when /?, = 1 Q , /? 2 =4 LX C = 0 .1 2 5
F, and L = 0 .5 H. ^
(b) Find the zero-state response to the
input = IO«(r) V.
(c) Draw an equivalent j-plane circuit
that accounts for the initial condi
tions. There are four possibilities, but
Figure 1M4.25 one is superior.
(d) Find the zero-input response for
2 6 . Consider the circuit in Figure P i 4 .2 6 . /Y-(0“) = 0 and /^(0“) = 1 A.
Suppose /e, = 10 Q, = 3 0 LX = 15 il, (e) Find the zero-input response for /^(0~)
C= 1 /40 F, and Z. = 8 H. Suppose = 0 and vc{0~) = 4 V.
+ i>2 (() = V (plot this input (0 Find the complete response for =
function so that you know its shape). 5u{t), = 2 V, and /^(O") = - 2 A.
(a) C om pute and /^(0“). Hint: Use linearit}'.
(b) Com pute and /^(f) for r > 0; this
is the complete response. ------- _ _ o
i,(t)
(c) Identify' the part o f the response due
only to the initial conditions at r = 0 “
that result from one could
think of this as the zero-input
Figure P I4 .2 7
response.
(d) Identify' the part o f the response due
28. The purpose o f this problem is to illustrate the
only to the part o f the input for r > 0,
computation o f the zero-input response using the
i.e., due to 25//(^) assuming zero initial
transfer Rmction concept. This requires proper
conditions; one could think o f this as
application of the equivalent initial condition cir
the zero-state response.
cuits and the use o f a transfer fiinction for each
(e) Identif)’ the transient and steady-state
initialized dynamic element. Consider the circuit
parts o f the complete response.
of Figure P I4.28a.
Chapter 14 • Liiplacc Iransform Analysis III: Transfer Function Applications
for r > 0.
Figure P I4.3 0
ANS\\T-:R: 0
/ Y Y V
2H
3 1 . Find the phase and magnitude in sinusoidal
,2 S 4S <
steady state for each o f the given transfer func
2F 4v.
tions when excited by the indicated sinusoidal
(a)
mput.
2s+ 6
(a) H{s) =
Vi„{s) .V--1-4.9-h i6
is excited by the input = 2 cos{2t
+ 4 5 °) V.
4S <
(b) Suppose has
5 ( 6’ + 4 )
output =3 cos(3f + 4 5 °) V.
(b)
Find the sinusoidal input that
Figure P i4.2 8 gives rise to this output.
(c) Let Z = 0 .2 H, C = 0.2 5 F, and /?j =
STEADY-.STATE CALCULATION = 7? = 1 n in the circuit of Figure
2 9. For the pole-zero diagram o f P i 4 .3 1 . Compute the transfer function
H{s) =
Vi„U) 9 2
.V“ -I-------5 -i-
1
+
V ,( t ) Figure P I4.32
n
Chapter 14 • biplacc Transform Analysis III: Transfer Function Applications 1
v^^is) 5 “ -h0 .5 5 + 5 .7 3 2 1
5^ - 5 . 5 5 - + 145-12
(f) ^6(*v) = —
The input to the circuit is =2^2 cos(2r + 5-’ + 5 .5 5 “ + 1 4 5 + 12
3 0 ° ) V. Knowing that the response has the A N SW E R : (b) I'h c result follows from the fol
form A cos{2t + (j)), find A and (j). lowing X'lATLAB code:
»din = [1 0 4];
3 8. A stable active circuit has the transfer func »nin = 8*[1 0 ];
tion »n = [2.5 -3]:
K,ur(^) 165^ + 4 4 5 + 1 2 8 »d = [l 2 51;
V:Js) + 85- + 365 + 80 »nn = conv(n,nin);
»dd = conv(d,din);
'5 2 Chapter 14 • Laplace Transform Analysis III: Transfer Funaion Applications
80
(b ) H2(s) = ^
43. The circuit shown in Figure P I4.43 has zero
initial conditions, y^O") = 0, and //^(0“) = 0).
80 (a) Find the transfer function H{s) =
(c )
s + 2 j + 160l5
(b) If = K(j«(r), find V({s) and
5 ^ + 8 5 + 1 616 and determine the largest voltage that
(d) H{s) = - , ------- g------- 2----------------- can appear across the capacitor. At
5^ + 155-^ + 8 3 5 ^ + 1 9 9 5 + 170
what times does this occur? This phe
y 4 l . In Figure 14.17 o f Example 14.8, let ji= nomenon has application in the
1.4 and recall that v-{t) = cos(2r) V. design o f insulators that support trans
mission lines. For example, a 100 kV
Chapter 14 • I^placc rransform Analysis III: Transfer Function Applications
FREQUENCY RESPONSE
4 8 . Draw to scale the pole-zero plot for H{s) =
(i + l)/[i(j^ + ^ + 10], and graphically compute
the magnitude and phase o f //(/co) for to = 0 .2 ,
0 . 5 , 1, and 10. W hat are the limiting values of
Figure P i4.45
the magnitude and phase, i.e., for co = ± 0 0 ?
4 7 . This problem illustrates the condition o f 5 0 . (This problem focuses on a qualitative under
steady state, but not sinusoidal steady state. For standing o f poles and zeros o f a frequenc)-
the circuit shown in Figure P 1 4 .4 7 , the op amp response.) You have determined the pole-zero plot
is assumed to be ideal. of a band-reject filter as shown in Figure P I4 .5 0
(a) Find the transfer function H{s). (or so you think, according to die qualitative
(b) If = 0 and y- (f) = 6(/) + 6(r - boardwalk suggestions o f the “professor”).
'^ 4 Chapter 14 » Laplace Transform Analysis III: Transfer Function Applications
jw jco
J
<>8j X 'l6j
X ■ 8j
1 1I II1fV ^W
---- 1------- -> a I
-0.2 -2 -1
X ■ -Bj
X *-8 j X -16j
(a) To make the gain o f the transfer func 5 2. Figure P 1 4 .5 2 shows the magnitude fre
tion 1 for large frequency you must quency response o f a transfer function H{s)
choose what value o f A" in whose numerator and denominator are both
( .9 - fourth order. Qualitatively determine the struc
ture o f H{s) approximately.
( . V - /?,)•■• •
Frequency-rad/s
Figure P i4.5 3
Chapter 14 • Laplacc Transform Analysis III; Transfer Function Applications 7SS
W hich o f the following is the best candidate for equivalent, the magnitude and phase
H{s)> responses 0 < to < IO.OOOti rad/sec for
s the second-order low-pass Butterworth
( 1) K
(^ + 0 .5 + ;5 ) ( 5 + 0 . 5 - y 5 ) filter
(2) K
l(s) = --------------- , --------^------------------------ .
(5 + 0 .5 + y5)(i- + 0 . 5 - y 5 ) + 1.414 + 1
\ 2 0 0 0 Ji/ 2000:1/
{s + j l ) { s - j 2 )
(3) K which has a 3 dB down [X )in t at 1000 Hz.
j(5 + 0 .5 + y5)(.? + 0 . 5 - j5)
(c) Referring to part (b), find
(s+j2){s-j2) ____ the steady-state phase and
(4) K
(5 + 0 .3 + y4)(.v + 0 . 3 - y 4 ) ( i + 0 .8 + y6)(.v + 0 .8 - y6) magnitude o f the output o f the
__________________ s { s + j 2 ) ( s - j 2 ) __________________ Butterw orth filter when the
(5) K input is (i) cos(5007rr), (ii)
(5 + 0 .3 + y 4 )(5 + 0 . 3 - y 4 ) ( 5 + 0 .8 + y 6 ) ( 5 + 0 . 8 - y 6 )
cos(20007rf), and (iii)
___________________ { s + j 2 ) { s - j 2 ) ___________________ cos(40007tr).
(6 ) K
5 (5 + 0 .3 + y 4 ) ( 5 + 0 . 3 - 7 ’4 )(5 + ().8 + y6)(.v + 0 . 8 - 7 6 ) Do your answers to part (c)
(s+j2)(s-j2) make sense? Why?
(7) K
5 ( 5 + 0 .8 + 7 6 X5 + 0 .8 - 7 6 ) Remark: T he gain, or the magnitude
of is t}^pically given in decibels
A N SW E R : (5) and computed as gain (dB) = 20
1Q H {S ) =
5 - + 1 .4 1 4 5 + 1
4H
57. Frequenc)' responses are typically plotred as malized high-pass Chebyshev filter
\H{j(S))\ vs. logjQicol or as gain = 2 0 log,()l//(yo))l transfer function
clB vs. logjQCO. These plots are com m only called 0 .5 0 1 2 .r
Bode plots. H{s) = ~T
0 .7 0 7 9 .S - 4 - 0 .6 4 4 9 5 + 1
(a) Repeat parts (a) and (b) o f Problem
55, except plot only the gain in dB vs. over the range 0 < f< 1 Hz.
logjolcol. (b) Repeat part (a) for the denormalized
(b ) Repeat parts (a) and (b) o f Problem high-pass Chebyshev filter
5 6 , except plot only the gain in dB vs.
logjolcol. 0 .5 0 1 2
\ 2 0 0 0 Jt/
H{s) =
5 8 .(a) Plot, using M ATLAB or the equiva 0 .7 0 7 9
/ Y Y V
R L
'.w Q
Figure P I4.60
+
V .Jt)
'" “ 6
Figure P i4.63
Figure P I4.61
+ 1 3 9 .9 0 f - ^ ) + 0 .9 9 8 9 f ;^ ) ^
^VlOO/ VlOO^ VlOO/^
H{s) = 5 4 3
5^ + 2 .7 9 9 6 f — ) + 1 3 .0 5 l f — ) + 1 7 .2 7 lf — ) + 1 3 .0 5 1 + 2.7991 ( — ) + 1
UOO/ VlOO/ \m J 100^ UOO/
Plot the magnitude frequency response in dB tie attenuation. This can easily be seen, since
on a semilog scale to verify a band-pass charac the transfer function has the form
teristic. Find an appropriate range of (0. o
1
Vo(s)
65. The circuit o f Figure P I 4.65 is the Sallen H{s) =
and Key low-pass filter, which can be used to + — 5+
RC In l
R^C
eliminate unwanted high-frequency noise. o
(a) Show that the transfer function is
___ K Observe that if s = ycOg = //RC, then
//(/COq) = 0, i.e., Vo(/(Oq) = 0 for
V. RxR'i C\C2
input sinusoids o f the form =
v:-in 1 1 \-K\ o
+• s+ y4sin(cOo^ + Cp).
/?2 Q ^2^2/
Figure P i4.65 Sallen and Key low-pass filter. = [2e^^+ 3cos(2/) - 1.5sin(2/)]«(r) V o
66, Figure P I4 .6 6 depicts the so-called twin-T (a) Compute the transfer function, H{s). n
RC circuit. This network is often used as a (b) Find all poles and zeros, and draw the
band-reject filter, i.e., a circuit that stops or pole-zero plot.
rejects signals at certain frequencies and allows (c) If there is no initial stored energy in
signals at all other frequencies to pass with lit- the network, i.e., no initial conditions.
Chapter 14 • Laplace Transrorm Analysis 111: Transfer Function Applications 759
com pute the network response to (b) Find the response o f this same relaxed
= sin(r)//(/) V. Does your answer circuit to the input = b(t) V.
make sense? W hy or why not? (c) Find the response o f this same relaxed
(d) Now suppose cos{lt)n{t) V.
= circuit to the input = 5b(t- 2) V.
For large t, show that Vg„,{t) ~ Kt
cos(2t + (j)) V for appropriate K and (J). 70. For this problem you are to use the proper
Does the response remain finite as r ties o f Laplace transforms as listed in Chapter
00 ? Is the network stable or unstable? 12. Suppose the Laplace transform o f the
impulse response, h{t), o f a circuit is given by
s+2
W(.v) =
(,s + l)(.v + 3)
v„Jt)
Find H^{s) and Hf{s) when
(a) h^{t)=h{At)
(b) h^it) = - 8)
Figure P i4 .6 7
68. A time-shift differentiator circuit has the INITIAL AND FINAL VALUE
property that its zero-state output is always PROBLEMS
equal to the derivative o f its input shifted in 7 1 . (a) The Laplace transform o f a signal y(r) is
time by some T > 0. Hence, its impulse
( 5 .v - l ) ( 4 .v - 5 ) ( 6 .v - 2 )
response is h{t) = h'{t - 7 ), a shifted version o f
the derivative o f the delta function. Suppose T 5 (2 5 -h l)(3 .s -F 2 )(5 i + 4 )
= 1 and the input to the circuit is given byy(f), Find the values o f /(O^) and /(co).
as sketched in Figure P I 4 .6 8 . Com pute the (b) Repeat part (a) when the Laplace
zero-state response^Cr) o f the circuit at ^ = 2 .5 transform o f a s i g n a l / r - 2)u{t- 2) is
and f = 6 seconds.
^_ 2 , (l(Xs - l ) ( 8 . v - 5 ) ( 1 2 . v - 2 )
^ .v(5.v+l)(3.s'-h2)(i- + 4 )
1 1 4s- - 4s - 4
s s s- + 2i' + 2
Figure P i4.6 8
A N SW ER ; - 2 , 1 Find the error, error(r) = for
large t. W hat was the initial error, error(0"’)?
6 9. The Laplace transform o f the response o f a C H E C K : for large t, error is 3
relaxed circuit to a ramp v^{t) = r ( f - 1) V is
(b )
7 8 . Obtain the asymptotes for the Bode plot o f
, ^ ^ a ) . l . v - l X 0 . 2 i + lJ + 1 + 1
(c ) zUU---------- ^----------------- 16 \ 1200
.y(.y“ + 144j H(s)= 10-
/ 5
+ I + 1
(1 6 5 -1 ) 25- 4 5 -1 \ -' 100 V 400
(d ) T +
(2 5 + 1 ) - 25+1 V25 + 1 V 25+ I/
10
(a) + 1 + 1
5(5 + 2) \ 20 2400
H(s}= 10
+1 +1
(b ) 200 800
5“ (25 + 2 5 )
Then verify your asymptotic plots by using
MA^’LAB to generate the true Bode plot.
(^ ■ -2 )
(c ) 20
5(25 + 25) “
Chapter 14 • Laplacc Transform Analysis 111: Transfer Function Applications "^61
> “80. Suppose a dc motor is modeled as per (c) W hat effect does R^ have on the
Figure 14.38 o f the text. The motor parameters steady-state speed for a step input of a
are = 25 Q . / „ = 0.005 kg-m^, k = 0.0 2 N - fixed amplitude?
m/A, B = e~^ N-m-s, and = 100 mH. (d) W hat effect does R^ have on the rate at
Calculate the steady-state angular velocity of which the armature approaches a
the motor and the steady-state armature cur steady state for a step input o f fixed
rent for the input 10«(r) V. amplitude?
> “81. Consider the dc motor modeled in > “8 2 .(a) For the dc motor of Figure 14.38 of
Figure 14.38 o f the text. the text, suppose k = 0.0 5 N-m/A and
(a) Using the same parameters as in R^ = 50 Q. Plot the steady-state cur
Problem 80 with = 5u{t) V, com rent as a function of 5 as 5 ranges
pute the armature current and the from 0 to c».
angular velocity as a function of time. (b) Using /?^ = 25 Q and B = e~^ N-m-s,
(b) Change to 50 Q, and recompute plot 0)^ as ^ ranges from 0 to oo. Recall
the armature current as a function of that increasing k increases the torque
time. per ampere o f armature current.
Explain why increasing k reduces co^^.
O '
w
C H A P T E R
Sometimes one must compute the average o f some quantity, sucii as the average vakie o f the light
intensity on a solar cell over the last fifteen minutes, or the average value o f the temperature in a
room over the last hour, or the average value o f a voltage over the last 50 milliseconds. W hen such
an average is updated continuously, it is termed a running average. The idea is that the readout o f
the device that averages these quantities always produces an updated value valid over a specified
prior time interval. If a voltage represents the value o f the quantity to be averaged, then one can
build a circuit whose output voltage is the required average. This is done by observing that an aver
age value o f a continuous time variable is simply the integral o f the variable over the proper time
interval, divided by the length o f the time interval. A device that integrates a variable over the last,
say, T seconds is called a finite time integrator. As an application o f the ideas o f this chapter, we will
look at a finite time integrator circuit and how it can be used to compute the average value o f a
quantit)^ The convolution concept directly leads to the required transfer function o f such a circuit.
CHAPTER OBJECTIVES
1. Introduce the notion o f the convolution o f two signals.
2. Using the notion of convolution, develop a technique o f time domain circuit response
computation that is the counterpart o f the transfer function approach in the frequency
domain, presented in Chapter 14. In particular, we seek to show that the convolution o f
an input excitation with the impulse response o f a circuit or system produces the zero-
state circuit or system response.
3. Develop objective 2 from r\vo angles: first, from a strict time domain viewpoint, and sec
ond, as a formal theorem relating convolution to the transfer function approach.
4. Develop graphical and analytical methods— in particular, an algebra— for evaluating the
convolution o f two signals.
5. Illustrate various properties o f convolution that are pertinent to block diagrams o f series,
parallel, and cascade interconnections o f circuits or systems modeled by transfer func
tions.
^6'i Chapter 15 • T im e Domain Circuit Response Computations: T h e Convokition Method
SECTION HEADINGS
1. Introduction
2. Definition, Basic Properties, and Simple Examples
3. Convolution and Laplace Transforms
4. Graphical Convolution and Circuit Response Computation
5. Convolution Algebra
6. Circuit Response Computations Using Convolution
7. Convolution Properties Revisited
8. Time Domain Derivation o f the Convolution Integral for Linear Time-Invariant
Circuits
9. Summary
10. Terms and Concepts
IL Problems
1. INTRODUCTION
At the beginning o f Chapter 12, we claimed that circuit response computations could take place
in either the time domain or the i-dom ain. Yet, except for the solution o f some very elementary
differential equations, circuit response computations have relied almost exclusively on the Laplace
transform technique. This chapter develops and explores the time domain counterpart o f the
Laplace transform method by introducing the notion o f the convolution o f two signals to produce
a third signal. We then show that the time domain convolution o f an input excitation with the
circuits impulse response yields the zero-state circuit response. Pictorially, the idea is expressed in
Figure 15.1, which is an update ot Figure 12.3.
Circuit Impulse
f(t) - > y(t) = h^f
Response h(t)
Input Signal Output Signal
h
T
Transform of
Circuit Transfer
Transformed- Output Signal
Function H(s)
Input Signal F(s) Y(s) = H(s)F(s)
riCnJRF, 15.1 Diagram showing the symmetric relationship of time and hequenc)'^ domain analyses.
The upper part of the diagram asserts that the convolution, denoted h * f, of the input signal/(f)
with the circuit impulse response h{t) produces a third signal, ^(r). This third signal is the zero-state
response, which equals
Chapter 15 • Time Domain Circuit Response Computations: The Convolution Method
Justification of the diagram o f Figure 15.1 begins with the formal introduaion o f the notion of
convolution. Mathematically, the convolution o f two fiinctions h{t) andy(f), denoted hy h * f or
h{t) * j{t)i results in a third function through the integration process,
The second step in verifying the diagram is proving that the convolution of the input signal with
the impulse response produces the zero-state circuit response. There are two approaches. One is
to work stricdy in the time domain and construct the actual zero-state response from the impulse
response and an arbitrary input excitation. The second approach is to prove that is
the convolution of the signal y(/) with the impulse response h{t). Because o f the Laplace transform
development o f the last three chapters, this direction seems the most painless and will be taken up
V—^ in section 3, after an introduction to the basic ideas o f convolution in section 2. Section 4 will
look at the graphical method of convolution, which helps in visually grasping the definition.
Section 5 will describe a convolution algebra, which yields a harvest o f shortcuts for evaluating the
convolution o f certain types o f signals. Section 6 looks at the computation o f circuit responses
using the convolution method, and here the averaging circuit is developed. Section 7 overviews
various properties o f convolution. The respective properties lend themselves to different structures
(e.g., parallel or cascade) for designing interconnected circuits. Section 8 describes the rudiments
o f the construaion o f the zero-state response o f a circuit working stricdy in the time domain
toward the development of a convolution integral.
Before closing this introduction, we should consider the question, why is convolution important^
One reason is that it allows engineers to directly model the input-output behavior of circuits and
general systems in the time domain, just as transfer functions model circuit behavior in the s-
domain.
A second reason is that circuit diagrams are sometimes unavailable or even get lost. How would
one generate a circuit model for analysis? One way is to display the impulse response on a CRT
and approximate by some interpolation function such as the staircase approximation illus
O ' trated in Figure 15.2. The process o f constructing an approximate impulse response o f an
unknown circuit or system is called system identification and is a vibrant area o f research. By stor-
ing the measured impulse response data as a table in a computer, one can numerically compute
the zero-state response of the circuit or system to an arbitrary input signal. This numerical simu-
lation process lets an engineer investigate a circuits behavior off-line. For example, simulating a
circuit destined for use in a hazardous environment provides a cost-effective means of evaluating
its performance in a simulated environment. Such performance evaluations often identify needed
design improvements prior to the constructing and testing of prototypes.
766 Chapter 15 * T im e Domain Circuit Response Computations: T h e Convolution Method
FIG UIIE 1 5 .2 A rectangular approximation to a hypothetical impulse response obtained from meas
ured data on an unknown circuit or system.
A third important reason for studying convolution in the context o f circuits relates to a deficiency
o f the one-sided Laplace transform: function segments that are nonzero for f < 0 are ignored by the
one-sided Laplace transform technique. Hence, time domain convolution offers a direct means of
computing the circuit response when signals are nonzero over the entire time interval, -<» < t< cc.
As mentioned, convolution is an integral operation between two functions to produce a third; i.e.,
the convolution o f two functions h(t) and J{t) produces a third function j = h* f. One might expect
that the convolution h * /e q u a ls the co n v o lu tio n /* h, i.e., the operation o f convolution is com
mutative. In fact, this is the case. To emphasize this property, we restate equation 15.1 in its more
general form:
CONVOLUTION
T he convolution o f two sig n a ls/r) and h{t) produces a third signal, ^(/^), defined according to
the formula
v (/ ) = J J A t - r)f{T)dT = - T)dx (1 5 .2 )
which is well defined, provided that the integral exists. This formula emphasizes the proper
ty that convolution is a commutative operation, i.e., h * f = f * h.
h * { r g ) = { h- f i * g (1 5 .3 )
and distributive,
(1 5 .4 )
Chapter 15 • T im e Domain Circuit Response Com putations: T h e Convolution Method 767
Another very useful property o f convolution is the so-called tim e shift property: if h{t) * j[t) =
then
h{t-T^)^j{t-T^)=g^t-T,-T,) (1 5 .5 )
Some simple examples serve to demonstrate the actual calculation process. These examples will
utilize the sifting p rop erty o f the delta function: if h{t) is continuous at / = T, then
So lu t io n
By the definition o f equation 15.2 and the sifting property o f the delta function,
Example 15.1 makes the point that 6 {t) acts like an identity' for convolution, i.e., it always returns
the function with which it is convolved. T he next example indicates that the convolution o f an
arbitrar)' continuous function y{/) with ad(t - T) produces a scaled and shifted version as
given below in equation 15.7.
So lu t io n
By the definition o f equation 15.2 and the sifting property o f the delta function.
Ler us a|3ply the results o f the above two examples to a simple staircase function.
EXA M PLE 1 5 . 3 . Com pute the convolution o f the functiony(/) shown in Figure 15.3a with h { t )
^ ld {t) -2 d {t - 1).
f(t)
iL
2 --
1-
H -------
1 2
(a)
So lu t io n
By the definition of equation 15.1 and the sifting propert)' of the delta function,
v(/) = / * / / = 2 j - t ) 6 ( t ) cI t - 2 j - T) d( T - l)r/r
( 15. 8)
= 2 / < r - T - 2 / ( / - T)]^,| = 2 / ( / ) + ( - 2 / ( / - 1))
Graphically combining the result given in equation 15.8 yields the relation in Figure 15.3b.
Exercise. C om pute the convolution o f the functiony(r) shown in Figure 15.3a with h{t) = 26{t) -
25{t-2).
AN SW 'FR: l f { t ) - l / { r - 2) =
S o lu tio n
By the definition o f equation 15.2,
since
T^O
|0 T< 0
We now break equation 1 5.9 into two regions: (i) t > r, in which case u{t- t) = 0 , making equa
tion 15.9 zero; and (ii) 0 s t s /, in which case u{t- t) = 1. W hen 0 t, equation 15.9 reduces
to
W e now apply the time shift propert}' o f convolution to the result o f Example 15.4.
u(t+ D - u(t-l)
>^ u(t) >k
1
1
W --- 1 1 1 ^ 1 > t
-1 1 2 1
(a) (b)
y(t)
FIG U R E 15.4 (a) h{t) = h(/), the step function. ( b ) / r ) = u{t + 1) - «(r - 1).
(c) Resulting y(r) for the convolution /;*/.'
Chapter 15 * T im e Dom ain Circuit Response Com putations: T h e Convolution M ethod
So lu t io n
Now using the time shift property o f convolution and equation 1 5 .1 0 , we conclude that
which is plotted in Figure 15.4c. We note that this result was achieved without any direct inte
gration.
E X A M P L E 1 5 .6 . Com pute the convolution^ = h */^of the two waveforms /j(t) and/(t) in Figure
15.5.
h(t)
(a) (b)
y(t)
> X
S o Lu n o N
We first note t h a t / r ) = ii{t) + «(r - 1 ) - lu{t - 2) and h{t) = u{t) - ii{t - 1). Using the distribu
tive law o f convolution.
Chapter 15 ‘ T im e Domain Circuit Response Com putations: T h e Convolution Method
Now using the time shift property o f convolution and equation 1 5 .1 0 , with a further application
o f the distributive law, wc conclude that
which is plotted in Figure 15.5c. We again note that this result was achieved without any direct
integration.
h(t) >k
1
w
(a)
FIG U R E 15 .6 (a), (b) Fu n ction s/r) and h{t) for the convolution of Example 15.7.
(c) The resulting convolution.
So lu t io n
Step 1. Apply definition and adjust limits o f integration. Applying the definition o f equation 15.2
yields
Since ;/(x) - z/(x - 7 ) is nonzero only for 0 ^ x :s T, the lower and upper limits o f integration
bccome 0 and T, respectively:
y{t) = h * f = ^ ^ e~^u{T-t)clT (1 5 .1 2 ) ^
Step 2. Detennirie the regions o f t over luhich the integral is to be evaluated. From the limits o f inte-
gration and the w(x - t), there are three regions o f interest: (i) r < 0, (ii) 0 s ^ s x < T, and (iii) T
^ t.
Step 3. Evaluate the convolution integral, equation 15.12, over the given regions.
Case I: t < 0. Here, r < 0 implies that x - f > 0 over 0 s x s 7". Hence, u(t - /) = 1 over 0 s x ^
Z and j.
y( 0 = e~'" c/t = 1- e~^
Case 2: 0 ^ t < T. For this c;ise, u { t - t) in equation 1 5 .1 2 is nonzero only when i >t. Hence, in
the region 0 ^ f < 7 , it must also be true that 0 < r s x < 7 for the integral o f equation 1 5 .1 2 to
be nonzero. Thus, the lower limit o f integration with respect to the variable x becomes f.
J . r _^ n 7 - j
v (/) = J ^ e '^u{T-t)dT = =e - e
Step 4 . Plot y{t). Combining the results o f step 3 with 7 = 1 implies that y(r) has the graph
sketched in Figure 15.6c.
Exercise. Find the convolution, say, y(/), o f the signal J{t) = e ^hi{t) with h{t) = Ku{-t).
A N SW E R : Ka '//(- /) f Ka
CONVOLUTION THEOREM
Supposey(/) = 0 and h{t) = 0 for / < 0. Then
(1 5 .1 3a )
i.e., convolurion in the rime domain is equivalent to multiplication o f transforms in the fre
quency domain; or equivalently,
Step 1. Given equation 15.13a, the first step is to write down the transform o f h[t) * J{t).
Specifically,
A couple of points are in order: (i) the inner integral, surrounded by parentheses, represents the
convolution h{t) * J{t); and (ii) the presence o f the step function u{t - r) is added as an aid to
emphasize the fact that h{t) = 0 for r < 0.
Step 2 . The goal at this point is to manipulate the integral o f equation 1 5.14 into a form that can
be identified as the product ol the Laplace transforms o f two functions, i.e., as H{s)F{s). The only
possible strateg)' is to interchange the order o f integration and group appropriate terms. Note that
the Re[^] must be chosen sufficiently large to ensure the existence o f the Laplace transforms of
both h{r) Under certain conditions that are t\'pically met, it is possible to interchange the
order o f integration within a com m on domain of convergence of H{s) and F{s). Interchanging the
order and regrouping the f-dependent terms inside a single parenthetical expression produces
Step 3 . Observe that the interior integral, surrounded by parentheses, in equation 15.15 is sim
ply the Laplace transform o f a time-shifted i.e., £[M{t - r)] = Substituting
e~^H{s) for the interior integral leads to the desired equivalence:
£ |/ i« ) « / ( O I = j ; ” f ( T ) H U ) e - ” 'dr = = H ( . v ) f ( ,v )
This theorem asserts the equivalence o f convolution o f one-sided signals with multiplication of
their transforms in the /-dom ain. For our purposes, h{t) assumes the role o f the impulse response
o f our circuit andy(r) the role of the input excitation. Accordingly, the convolution o f the impulse
response o f a circuit or system with an input signal, a time domain com putation, equals the
inverse transform o f the product o f the respective Laplace transforms. In other words, the diagram
o f Figure 15.1 is correct, as claimed under the conditions o f the theorem.
Chapter 15 * T im e Domain Circuit Response Com putations: T h e Convolution Method
Exercises. 1. An unknown relaxed linear system has impulse response b{t) = n{t) - u{t - 1). Find
the response to the input signaly(f) = l u { t - 1) using convolution and check your answer using
the Laplace transform method.
A N SW E R : h it - 1) - 2 r ( / - 2)
2. An unknown relaxed linear system has impulse response h{t) = (a - b)e~‘^‘u{t),aj? > 0 and input
sig n a l/r) = e~^‘u{t). Find the response ^(f) = h{t) * fit).
A N SW ER :
T he conditions o f the theorem are somewhat restrictive in terms o f computing circuit responses
strictly in the time domain. Specifically, it is the one-sided Laplace transform that does not rec
ognize function segments over the negative real axis— hence the condition on the input excitation
t h a t / r ) = 0 for /■< 0. This restriction does not lend itself to the computation o f initial conditions
and circuit responses due to input signals extending back in time to / = In general, the con
volution o f an input excitation with a circuits impulse response presupposes no such restriction.
However, justification o f the computation o f zero-state responses due to input excitations extend
ing back in time to / = -oo cannot be based on the convolution theorem o f the one-sided Laplace
transform. A justification o f the time domain convolution approach to computing circuit response
is reserved for the last section o f this chapter, due to its complexit)^ Nevertheless we will use the
result as necessary, such as in the next exercise.
2
H{s)^
.9+ 0.2
C om pute the convolution o f the impulse response with the input v{t), shown in Figure 15.7.
A N SW ER : lOlI - ^ + 1) - 1()[1 - ')],,(/ d
v(t)
J <
,1
-1
FIG U RE 15.7
simple, squarish waveforms. T he technique often leads to a more penetrating insight into the con
volution integral.
There are four key ideas in the graphical procedure: flip, shift, multiply, area. The following is a
detailed description o f the procedure.
To compute given by the convolution integral o f equation 15.1 or 15.2, for a specific value o f
t = T, perform the following steps:
1. Plot h{x) vs. T andy(T) vs. T curves.
2. Flip t h e / r ) curve about the vertical axis (t = 0) to obtain t h e / - r ) vs. T curve.
3. Shift th e y (- t) curve to the right by the am ount T to obtain t h e y ( r - t) v s. t curve.
4. Multiply. Plot the product h{z)j{T - r) vs. x curve.
5. Area. Calculate the area beneath the hir)j{T - r) vs. x curve for —x < t < oo. The result
is7(T ).
With regard to step 2, for each t, h{t-T) is a shifted horizontal flip o f/;(x): as t moves from to cc,
h{t - x) moves along the x-axis from x = -oo to x = oo; a simple illustration is h(t - x) = u{t - x), w'hich
is sketched in Figure 15.8. Part (a) o f the figure shows «(x), part (b) depicts u{-x), and part (c) plots
u{t-T), which slides to the right along the x-axis as t increases. For comparison, we can consider the
ftuKtion h{t-x) = u{t- 1 - t), whose three forms are given in Figures I5.8d through f
increasing t —
----w - w
---------------- ^ ^
increasing t
A u (x -l) u(-T-l) A
1 -1 t- 1
FIG URE 15.8 (a)-(c) For the function h{t-x) = « ( r - x ) , plots of (a) «(x), (b) « ( - t ) , and (c) « ( r - x ) .
(d)-(O For the function h{t-x) = u{t-x - 1), plots of (d) tt{x - 1), (e) //(-x - 1), and ( 0 «(^-x - 1).
Chapter 15 • T im e Domain Circuit Response Com putations: 'I'he Convolution Method
k.
f(t) " f(T)
I
h(t-T) = u (t-1 )
h(t) 1
increasing t —
— y
(a) (b )
So lu t io n
T he graphical solution to this convolution depends on a partitioning o f the time line into special
segments over which the graphical convolution is easily done. There are two regions to consider:
(1) -0 0 < / :s 0 and (2) t > 0.
Step 1. Consider the region r > 0. Figure 15.9b shows that h{t - t ) J { t ) = 0 for all r > 0 and all T .
Thus the convolution integral is zero and = 0, r > 0.
Step 2. Consider the region -oo < ^ ^ 0. From Figure 15.9b , the product h{t =1 x 2 = 2 for
r ^ T s 0 and zero elsewhere. Consequently, the area under the nonzero portion o f the product
functions h{t - is
Step 3. Combine the foregoing calculations into a plotted waveform. Figure 15.10 shows the func
tion j/(r) resulting from the convolution.
y(t)
Exercise. In Example 15.8, supposey(/) is changed ioj{t) = 2u{t) - 2ii{t —1). M ndj'(r) at r = 0 .5 ,
1.5, and 2 .5 sec by the graphical convolution method.
A N SW ER : 1 , 1 , 0
Another, more complex example will end our illustration o f the graphical convolution technique.
E X A M P L E 15.9. Com pute the convolution o f the triangular pulse h{t) with the square pulse
j{t) as sketched in Figure 15 .1 1 .
f(t)
(b)
/ \
/ \
/ \ /
/I \
/
-1
/
0 1 2 3 4
V .
5 6 7
Time (seconds)
(c)
T he goal, o f course, is to graphically evaluate the convolution integral o f equation 15.2 using the
following steps:
Chapter 15 • Tim e Domain Circuit Response Com putations: T h e Convolution Method
Step 1. Drauf h{t - x) audJ{T) on the T-axis for f < 0 and compute the area o f their product. Figure
15.12 shows h{t - x) andy(x) on the X-axis. From the figure, it is clear that h{t - x )fj) = 0 for ? <
0; hence, = 0 in the first region.
h(t-x)f(T)
A
-- 2T
Area
2T-t
t-2T t T
T he shaded area o f the figure is the difference between the area o f the large triangle, defined as
Area A = 0 .5 (2 7 )^
and the area o f the smaller triangle to the left o f the vertical axis, defined as
Area B = 0 . 5 ( 2 7 -
Chapter 15 •Tim e Domain Circuit Response Com putations: T h e Convolution Method
Hence,
j/(r) = Area = Area A - Area B = 2 Tr - 0 . 5 r
for 0 :s r < T. Alternatively, one may use the formula for the area o f a trapezoid, i.e., the average
height times the base, in which case one immediately obtains 0 . 5 ( 2 7 + I T - t)t = Area = ITt-
0.5/^.
h(t-T)f(T)
A
-- 2T
Area
■> I
t-2T T t
In this figure, the shaded area, which determines j'(f), is again the difference o f two triangular
areas, specifically.
y{t) = 0 . 5 [ r - { t - I T ) ] - - 0 . 5 [ 2 r - r]2 = 2 .5 7 ^ - Tt
for T^t<2T.
Step 4 . Figure 1 5 .1 5 shows the next region, 2T ^ t < 3T. Another straightforward calculation
yields
y{t) = 0 .5 ( 3 r - /)2 = 4.5T^ - 5 T t ^ 0.5t^
for 2T^ t < 3T.
h(t-T)f(x)
A
-■ 2T
Area
A > T
t-2T
F IG U R E 1 5 . 1 5 Graph o f h{t-T)j{T) on the x-axis ioxlT^ t < 5 T
780 Chapter 15 • T im e Domain Circuit Response Com putations: T h e Convolution Method
Step 5. Consider the region 5T^ t. Here, the product h{t-T )/(t) = 0 , in which caseyit) = 0 for r
>37:
In sum,
0, /< 0
2Tt-0.5r, 0^t<T
v(0 = 2.5T^-Tt, T^[<2T.
4.5T- -3T f + 0.5r, 2 T ^ t < 3 T
0.
Exercises. 1. Repeat the calculations o f the preceding example, except flip and shift J{t) instead of
h{t). Here, one looks zx. j{t - r) sliding through h{x). The calculations should be easier and the
result the same.
2, Find the output o f a linear and relaxed circuit with in p u t //) and impulse response h{t) (shown
in Figures 15 .1 6 a and b) at time r = 2 .5 seconds. Hint: Use graphical convolution.
A N SW E R : !
MGURF. 15.16
4. For h{t) andy{r) as sketched in Figure 1 5 .1 8 , fin d ^ (l) for the convolution ^(r) = h{t) j{t).
A N SW ER : 4
h(t)
^ t
-1
FIG U RE 1 5 .1 8
Sometimes the foregoing graphical techniques prove difficult to execute. Nevertheless, an under
standing o f them offers fundamental insight into the meaning o f the convolution integral. A use
ful set o f techniques for quickly evaluating convolution integrals arises from the properties o f a
convolution algebra, discussed in the next section.
5. CONVOLUTION ALGEBRA
A set o f functions, together with operations called addition and multiplication, is called an alge
bra, assuming certain conditions are satisfied. The set o f all functions that can be convolved with
each other also constitutes an algebra with respect to the operations o f addition (+), and convolu
tion (*). This set, together with the two operations, is called a convolution algebra. In this con
text, operations such as differentiation and integration are inverses o f each other. For example,
integrating a function and then differentiating the result returns the original function. W ithin the
convolution algebra, the co n v o lu tio n /* g is equivalent to the convolution o f the integral o f/w ith
the derivative o f The advantage here is that, by successive integration and differentiation, it is
often possible to reduce an apparently difficult convolution to a simpler one.
For a set o f functions to be an algebra with respect to + and *, several arithmetic operations must
hold. In particular, + and • must be both commutative and associative. The commutativit}^ and
Chapter 15 • Tim e Domain Circuit Response Com putations: T h e Convolution Method
For our purposes, the interesting aspects o f a convolution algebra o f functions rests with the inter
relationship o f convolution, differentiation, and integration. To map out this kinship, we use the
following notation
/ /(t)A (1 5 .1 6 )
and
/,< !)(,) = ^ (1 5 .1 7 )
dt
It can be shown that
/ * / 7 = / ( • ) * // - I ) (1 5 .1 8 )
if /(-o o ) = 0 and exists. T he co n stra in t/(-o o ) = 0 implies that the derivative o f / / ) is zero at
t= -oo and the constraint that exists means that the integral o f h{t) has finite area over the
semi-infinite interval (-oo, /] for all finite t. Similarly,
/* /,= /(- • ) (1 5 .1 9 )
So lu t io n
T'he goal is to use equation 1 5 .1 8 to evaluate the convolution, i .e .,/ * h =/^'^ , where/^*^(/)
= 5(r) + 6 ( / - I) -2 6 { t - 2) and = r{t) - K ^-1), as presented in Figure 15.19.
Chapter 15 * T u n c Domain Circuit Response Com putations: T h e Convolution Method /H3
(b)
Since f^\t) = + 6{t - \) -26 (/ - 2 ), a sum o f impulse functions, the sifting propert)-’ o f the
impulse function implies
yU) =f^ 1) - 1) (1 5 .2 0 )
W ith the picture o f g i v e n in Figure 15.19b , the right-hand side o f equation 1 5 .2 0 can be
interpreted as a graphical sum o f (shifted) versions o f //~ '^ (/), as illustrated in Figure 15.20.
Equations 1 5 .1 8 and 1 5 .1 9 , as illustrated in the preceding example, are special cases o f more gen
eral formulas. Specifically, let y =f * h- Then
where j and k are integers and the n o t a t i o n m e a n s theyth integral o f/o v e r [ - 0 0 , t] if j < 0, and
the y’th derivative if j > 0. O f c o u r s e , = /. An application o f this formula to the special case
where j = -k with y = 2 is given in the following example.
EXA M PLE 1 5 . 1 1. Find the convolution ^t) = :r cos{m)u{t) withy(f) = r{t) - r{t- 2 ), as sketched
in Figure 15 .2 1 .
S o lution
Some preliminary thought suggests that evaluation o f the convolution integral might proceed
more smoothly via equation 1 5 .2 1 ; i.e.,
where
Hence,
f*g =f^-'> * = [1 - cos(jrr)]K(r) - [1 - co s(7 r(r- 2))\ii{t - 2)
But 1 1 - cos(M ^~ 2))]u (r - 2) is just a right shift by one period o f [1- cos(;rr)]«(f). Therefore,
= [1 - c o s (;rr)]« (/)//(2 - t)
Tim e (seconds)
Convolution algebra and graphical convolution lend themselves to a second application o f the
convolution technique: the computation o f circuit responses from a staircase approximation to a
circuit impulse response. If a circuit schematic is lost, such an approximation could result from a
C R T readout o f the circuit impulse response measured in a laboratory. The following example
illustrates this application.
EXA M PLE 15.12. Suppose the schematic diagram o f a very old linear circuit is lost. However,
the circuit impulse response is measured in the laboratory and approximated by the staircase wave
form o f Figure 15.23. If the input to the circuit is = 100«(^), compute the output voltage,
at r = 0 and at r = 0.5 sec.
h t)
1.0
0 .8 --
J
0 .6 -
0.4 -
0.2 -
■f—I— h H — I— h H— I— I— I— I— h -l-
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
S olution
The objective is to convolve the input with the impulse response. The technique o f convolution
algebra where one differentiates the impulse response to obtain a sum o f shifted impulse functions
and integrates the input to obtain a ramp function seems to be the most straightforward approach
for this calculation.
Now, since the integral o f the input is 100r(^), we can compute the output voltage as
O f course, it is possible to obtain the solution just as easily in this case using the graphical method.
Simply flip the curve and slide it through the /?(f) curve. The area under the product curve
is simply the sum o f the rectangular areas, which are easy to compute.
2. As an alternative to the convolution algebra, find the convolution j{t) = u{t) with h{t) given
in equation 15.22. Hint: Use the result o f Example 15.4 or equation 15.10, and the time shift the
orem o f convolution, equation 15.5.
ANSWF.R: Replace the step functions in equation 15.22 with ramp functions.
This section contains a series o f examples that illustrate the convolution approach to computing
zero-state circuit responses.
Chapter 15 • T im e Dom ain Circuit Response Com putations: T h e Convolution Method /87
EXA M PLE 15.13. Consider the RC circuit o f Figure 15.24, whose impulse response is h{t)
e~^u{t). If the input is find when 7" = 0 and T > 0 for <7 = 0 and <7 > 0.
10
IF
So lu t io n
Case 1: a = 0. Here, = u{-t) V. Since the capacitor voltage, is the convolution o f the
input with the impulse response,
^ W (0 )= e~'^i((T)dT = 1
This makes sense because with a = 0, = u{-t) V is a 1 V constant, and the capacitor looks
like an open circuit at r = 0. On the other hand if r > 0,
^ W (^ ) = Jq - T)dx = e~^
This also makes sense physically because at 7’ = 0, the initial capacitor voltage is 1 V and the capac
itor discharges with a time constant o f 1 sec since there is no further nonzero input.
.-0^^)rdT =
1+ a
This initial voltage depends on the history o f the excitation and can be computed only by convo
lution because the circuit is not in steady state at T = 0.
Chapter 15 • Tim e Domain Circuit Response Com putations: T h e Convolution Method
•r e-'^
Vout^T) = Vc(0)e-^u{T) = -------u{T)
\+ a
AN SWF. R: ^
1+i/
3. In Figure 15.24, suppose the resistor has value /? > 0 and the capacitor C > 0. Show that the
impulse response o f the circuit is ^
h{t) = ^ e
aC
4. For the general impulse response computed in Exercise 3, suppose = e^‘'u{-t) V. Compute
1C
1
ANSWF.RS; 1, RC
1
----- + a
RC
E X A M PLE 15 .1 4 . The goal o f this example is to design a circuit that computes the running aver
age o f a voltage over the interval [/^- T, t] given a specification o f the necessary impulse
response, h{t).
S o l u t io n
A circuit that computes a running average must have the input-output relation
From our development o f convolution, such a circuit must have an impulse response h{t) satisfy
ing the relationship
Vnu,iO =
Now, h{t - t) must be a window function that captures the portion o f over the interval t -
Ts T ^ r. Figure 15.25 depicts the proper forms o f - t) and /;(?).
Chapter 15 * T im e Domain Circuit Response Compurations: T h e Convolution Method 789
h(t-i) >^h(t)
T 1
T
- -
---
t-T
FIG U RE 15.25 (a) The window function h{t - r). (b) The impulse response h{t).
The circuit design problem reduces to developing a circuit that integrates the function segment
V- {t) over t - T <, T To achieve this integration, note that
(15.23)
For the second integral on the right-hand side o f equation 15.23, let /I = r + 7 ’. Then =dr,
T = A + r , and
(15.24)
Since both X and x are dummy variables o f integration, we may replace A by ra n d rewrite equa
tion 15.23 as
(15.25)
where Vj^^{T - 7) is simply a delayed replica o f t^/„(x) and where, for practical reasons, we can
replace the lower limit of-oo by /q, the time the actual circuit turns on.
As a convenience, we will define a device called an ideal delay o f T seconds, whose input is yy,/^)
and whose output is T). Figure 15.26 shows the ideal delay as a device having infinite input
impedance combined with a dependent voltage source whose output is a delayed version o f the
input. Such a device can be achieved by storing the values of in a digital computer or, for
small T, by the use o f an analog delay line.
All the pieces are now available; it is merely a matter o f putting them together. Integration can
occur using an inverting ideal op amp circuit having a capacitive feedback and resistive input. The
input to this ideal op amp integrator can then scale and sum the voltages and T) to
produce the desired running average, by setting RC = T. This will guarantee the correct
scaling to achieve the desired average, since the transfer function o f the integrator will be M{RCs).
Figure 15.27 shows a circuit that will realize the desired running average.
FIG U RE 15.27 Op amp circuit that produces a running average of the input waveform,
provided that RC = T.
For the circuit o f Figure 15.27, the input drives the first inverting op amp circuit and also feeds
the ideal delay. T he second op amp circuit is an ideal inverting integrator whose inputs are
and T). With RC = T the output is the required running average.
From the perspective o f the impulse response theorem, the convolution properties o f commuta-
tivir)^ associativity, and distributivity have important implications in terms o f circuit and system
configurations. For example, if /?, {t) and hjit) are the impulse responses o f two systems, then com
mutativity says that means that the order o f a cascade o f circuits or systems is
mathematically irrelevant, provided that there is no loading between the circuits. The idea is illus
trated theoretically in Figure 15.28.
Chapter 15 * T im c Domain Circuit Response Com putations; T h e Convolution Method '9 1
h,(t)
} (a)
h,(t)
}
h,(t)
} (b)
h,(t)
}
FIGURE 15.28 Interchanging the order of impulse responses, in which the equivalence of parts
(a) and (b) follows from the commutativity of convolution.
Although mathematically parts (a) and (b) o f Figure 15.28 produce the same result, practical con
siderations often dictate a more careful realization process. For example, one circuit may have a
high input impedance w'hile the other has both a low input impedance and a low output imped
ance. In this case we would put the first high-input-impcdance circuit at the front end and the
other circuit at the output end.
100
H {s) = (15.26)
(.y-Kl)(.9-H20)
illustrates commutativity nicely. A designer may use either Figure 15.28a or Figure 15.28b to real
ize H{s). Magnitude scaling, say by = 10“^, will yield more realistic resistor and capacitor val
ues. Recall that in magnitude scaling, and
20 0.05 0
(a)
•92 Chapter 15 • I1 m c Domain Circuit Response Com putations: T h e Convolution Method
0.05 Q 20
(b)
Exercise. Verify that each o f the circuits o f Figure 15.29 realizes the transfer function o f equation
15.26. Magnitude-scale by = lO'’ to obtain more realistic parameter values. If only 1 [.iF capac
itors were available, you would need two scale factors— for the first op amp stage and
for the second op amp stage. What are the rwo scale factors?
AN SW ERS: Multiply each resistor by I O'’ and di\'ide each capacitor by 10'' to obtain the new val
ues. II only 1 uF capacitors are available, then for Figure 15.29b, = 10*^’ and = 0.5 x 10^’.
For Figure 15.29a, the two scale factors are interchanged.
Theoretically speaking, the associative propert)', /;, * (A, * = {h^ * h-y) * hy means that multi
ple cascades o f circuits or systems can be combined or realized in whatever order the designer
chooses. This assumes that there is no loading between the circuits or systems. Op amp circuits
called buffers or voltage followers having gains o f 1 can be used to isolate stages. On the other
hand, practical constraints may impose a condition on the realization o f a circuit for which the
mathematics o f the associative property does not account.
Finally, we consider the distributive property o f convolution: /;, * {hj + = {h^ h-^) +
(/?j * h^. One interpretation o f this property is that the superposition o f the input signals and
is valid. Flowever, Figure 15.30 presents rwo block diagrams with different interpretations.
Here one sees rwo possible topologies for realizing a system.
Chapter 15 • T im e Domain Circuit Response Com putations: T h e Convolution Method ~93
y(t)
FIGURE 15.30 Two possible block diagram interpretations of the distributive law For convolution.
Exercise. In terms o f reliability, i.e., possibly continued partial operation in the face o f a circuit
failure, suggest reasons why one realization in Figure 15.30 would be superior to the other. In
terms o f minimum number o f components, suggest reasons why one realization would be better
than the other.
As mentioned, a deficiency in the one-sided Laplace transform technique is its inability to deal
with signals whose time dependence may extend back to -<». This section develops the zero-state
system response as the convolution o f a not necessarily one-sided input excitation with the
impulse response o f the circuit or system. Throughout the development, we will assume that the
circuit or system under consideration is linear, i.e., composed of linear circuit elements. This
implies that the zero-state response o f the circuit satisfies the conditions o f linearity; i.e., if the
zero-state response to the cxcitation fj{t) \s yj{t) for / = 1 ,2 , then the zero-state response to the
input excitation addition, we assume that the circuit or sys
tem is time invariant, i.e., each circuit element is characterized by constant parameter values.
Mathematically, this means that j{t) is the input to a circuit element and y(/) the zero-state
response o f the circuit element, then y{t - 7) is the response 1 0 j{t - T) for all T and all possible
794 Chapter 15 • Tim e Domain Circuit Response Com putations: T h e Convolution Method
input signals y(^). intuitively speaking, time invariance means that if we shift the input in time,
then the associated zero-state response is shifted in time by a like amount. These properties under
lie the development that follows.
4- 0 < /< A
A
(15.27)
0 othenvise
u(t) - u(t-A) , , .
6A(t) = ---- 7---- for several As
1
A.
■> t
1 2 3
5^(t-kA)
1
A
kA (k-Hl)A
For convenience, let = M so that 6^ (r - M ) = 6/^(/‘ - tj^. Figure 15.33 shows a rectangular
approximation, v{t), to a continuous waveform v{t).
Chapter 15 * T im e Domain Circuit Response Com putations: T h e Convolution Method 79 S
^ v (t)
Expressing the rectangular approximation indicated in the figure analytically, using shifted ver
sions o f the pulse functions defined in equation 15.27, leads to the infinite summation
where we have interpreted the delta function as a limit o f short-duration pulses whose height is
inverse to the width so that the area is constant:
d(t)= lim d ^ ( { )
A^O
Observe that the right-hand integral o f equation 15.30 is precisely the convolution v{t) * S(r)
h i^ {t)= ^ ^ _ h {t -T )d f ^ {T )d T
Chapter 15 • Tim e Domain Circuit Response Com putations: T h e Convolution Method
follows from the convolution theorem. Equation 15.31 restates the law that the zero-state response
o f an input to a linear time-invariant circuit is the convolution o f the input with the impulse
response.
Now, by the assumption o f time invariance, kA) is the zero-state response o f a well-behaved
linear time-invariant circuit to - M ). Suppose further thatj'(f) is the zero-state response of
the same circuit to v{t). (See Figure 15.34.)
LinearTlme
8,(t)
Invariant Circuit
LinearTime
8.(t-kA)
Invariant Circuit
LinearTime
I (v(gA)8,(t-tJ
k = -cc
!< = -CC Invariant Circuit
h
LinearTime
v(t) y(t)- / v(t) h(t-x) di
Invariant Circuit /-oc
h
FlCirilF. 15.34 Zero-state responses to a particular linear time-invariant s)^stem, showing the framework
of the derivation. Note that the bottom condition follows because as A 0, A di, x, and H j.
It is now possible to use the approximation for v{t) given in equation 15.31 to generate an approx
imation to^(^) in terms o f a summation o f terms o f the form v{kA)hi^{t - kA)A. Taking the limit
as A 0 will yield as the convolution o f v{t) and h{t).
To derive this, note tliat for each k\ v{kA) = v{ti^ is a scalar. Hence, the zero-state response to
tf^A is tf^)A. By the linearity assumption, which implies superposition, the
zero-state circuit response to v{t), equation 15.29, is
(15.32)
k--x
In the limit, as A approaches zero, approaches a continuous variable t and A -♦ ch. Hence, if
the impulse response is sufficiently smooth— i.e., if it has sufficient continuous derivatives— then
Chapter 15 • Tim e Domain Circuit Response Com putations: T h e Convolution Method
Thus, we conclude that for a linear time-invariant circuit, the zero-state response to an input
excitation v{t) is the convolution o f the input v{t) with the impulse response h{t). We will refer to
equation 15.33 as the impulse response theorem,* which says that the zero-state response o f a lin
ear time-'mvariant circuit or system to a (possibly tivo-sided) input signal is the convolution o f the input
with the impube response o f the circuit.
9 . SUMMARY
This chapter has introduced the concept o f the convolution o f two signals. The convolution can
be evaluated analytically (by direct computation o f the convolution integral) or graphically. Often,
by applying the properties o f convolution algebra, it is possible to implement shortcuts for calcu
lating the convolution o f two signals, resulting in the simplification o f the analytical calculation
or o f the graphical calculation.
Using the notion o f convolution, the chapter developed a technique o f computing circuit respons
es in the time domain. This technique is the direct counterpart o f computing the transfer func
tion in the frequency domain, the approach presented in Chapter 14. Using the convolution
approach, one can compute the zero-state response o f a circuit excited by signals that extend back
in time to -oo, something not directly possible with the one-sided Laplace transform. However,
for one-sided signals— which constitute the great majority o f signals that are relevant to circuit
analysis— the convolution and Laplace transform approaches are completely equivalent, as
demonstrated by the convolution theorem. The chapter gave an example o f designing a circuit to
compute a running average. In addition, it presented an application of the convolution technique
to the computation o f circuit responses for a circuit whose impulse response is approximated on
a CRT. Future courses will expand the seeds planted in this chapter. For example, convolution is
pertinent to an understanding o f radar techniques, commonly used to identify speeding motorists.
Algebra: a set o f functions with respect to two operations, + and *, satisfying the commutative,
associative, and distributive laws. In addition, there must be an identit)' with respect to
each operation. For addition, the zero function serves as the identity. For convolution,
the delta function plays this role. The delta function is an identit)' element because o f its
sifting property.
Associativity: for convolution, the property h * (/* ^ = {h * f } * g.
Com m utativity: for convolution, the property h ' f = f * h.
Convolution: an integration process between two functions to produce a third, new function in
accordance with equation 15.1 or 15.2.
Convolution algebra: the algebra o f functions with respect to the operations o f addition (+) and
convolution (*).
798 Chapter 15 • Time Domain Circuit Response Computations: The Convolution Method
* T h e derivation o f this result is, o f course, not rigorous. A rigorous justification is given as theorem 4 o f Sandbergs
“Linear Maps and Impulse Responses,” IEEE Transactions, on Circuits and Systems, vol. 3 5 , no. 2, February 1988,
pp. 2 0 1 -2 0 6 .
Chapter 15 • T im e Dom ain Circuit Response Com putations: T h e Convolution Method 799
Problems A A
CONVOLUTION BY INTEGRAL
1. Leijit) = K^5{t - r , ) , > 0. Compute and -► t
3. Let//) = 25{t + 2) - 2 6 {t- 2). Compute and 7. Let f i t ) =K^e-^‘ii{t), a > Q, and f^{t)
plot the following convolutions: -K-,e~^Uiii), b > 0. Use the Laplace transform
(a) J{t) - cos{m)u{t) method to compute the following convolutions:
(b) J{t) - sin(;rr)«(r) (a ) fir)* fir)
(b) T iW / z W
4. Compute and plot the results o f each o f the Now use the results for parts (a) and (b) and the
following convolutions: various convolution properties to compute the
(a) u{t) * u(t - 2) following convolutions:
(b) n { t - 2) (c) f i r ^ T ) * f i t - T ) , T > 0
(c) u(^)*[u(t)-;^(t-4)] (d) fit^ T )*f^ [t-T )
(d) u(t^2)*lu(r)-u(r-2)] (e) fit+T^)-f^{r-T,),T^,T^>Q
(e) [ « ( f + 2 ) - ;v U ) ] * [ / / W - « ( r - 2 ) ]
8. Compute and draw the convolution y{t) =
5. Use the convolution integral to compute the f t ) * v{t) o f Figure Pi 5.8 with v{t) = 5{t + 4) -
convolutions in parts (a) and (b). Then use the 5 (f+ 2 ).
results for (a) and (b) to compute the remain
f(t)
ing convolutions. (Parts (a) and (b) can be used >k
to solve many o f the subsequent problems also.) 3 -
9. (a) Compute the convolution y{t) = f j ) * (b) Repeat the calculation in part (a)
v{t) for the functions in Figure PI 5.9 using graphical convolution for K = 2 .
using the fact that r{i) = «(/) * u{t) and
the properties o f convolution.
(b) Repeat the calculation in part (a)
using graphical convolution for 7’ = 1
and/r= 1.
. v(t)
2K
(a)
H------h
1 2 3
2K (b)
Figure P15.11
2T
■>t 12. (a) Compute the convolution j>(/) = j{t) *
-T v{t) o f the functions in Figure PI 5.12.
--K r\
(b)
Hint: Use the results o f Problem 5 for
u{t) * r{t).
Figure P I5.9 (b) Repeat the calculation in part (a)
using graphical convolution for =2
lO.(a) Compute the convolution oiy{t) = fj) and 7’ = 1.
* v{t) in Figure PI 5 . 1 0 using the fact
,f(t)
that lit) = «(/) * u{t) and the proper i. v(t)
/N
A
ties o f convolution. 2A
(a) v{t) = d{t) + 5(f + 2) (Hint: Use graph < r < 00.
ical convolution.) (b) yjit) = h{t) * and sketch for - o o <
^< 00.
(b) h{t) = +1) + 5(? - 1) (Hint: Use
graphical convolution.) (c) y^{t) = h{t) *^ (/ ). and sketch for - o o <
r < 00.
(c) /(^) = u{t + 3) - «(? + 1)
(d) y^{t) = h{t) *[/i(r) ^ and sketch
for -CO < t < 0 0 .
(e) y^U) = h{t) *[ u{-t) +^(^)], and sketch
for -oo < / < 00.
'" 6
Figure P I5.18
-► t
> t >«t -T
Figure PI 5.19
Figure P I 5.15
20. Consider the RLC circuit in Figure PI 5.20.
(a) Compute the transfer fiinaion //(j),
16. I f / / ) = K^e^ui-t), ^ > 0, and h{i) = K2 u(t),
compute ^(/) = h(t) and express H {s) = — — ——
s+ a s+ b
802 Chapter 15 • T im e Domain Circuit Response Com putations: T h e Convolution Method
using the indicated element values. the circuit is = ti{t) -u {t - 1) V, plot the
Then compute the impulse response response ^(r) using MATLAB as follows:
hit). tstep = 1;
(b) Compute y{t) = h{t) * where vin = [1];
= v{t) is given in Figure P I 5.19. h = [0. 2, 3, 1, 1];
(c) If the input voltage is V, y = tstep*conv(vin, h)
com pute y{t) = h{t) * y = [0 y 0];
(d) Repeat (c) if the input voltage is t = 0:tstep:tstep’ (length(vin)+length(h));
= 5e^^uH) V. plot(t,y)
grid
/ Y Y V —o
5Q 2H + h(t)
Vou,(t) 3 --
0.5F
—0 2-
Figure P I5.20
(a) Find the impulse response o f the cir 23. Repeat Problem 22, for the input =
cuit. u{t) + u { t - 2) - l u { t - A ) V. Hint: You need to
(b) Find the transfer function o f the cir change only the specification o f vin in the
cuit as a ratio o f polynomials by using MATLAB code o f Problem 22.
(c) Find the step response o f the circuit by have a finite nonzero duration.
the convolution method. (a) Let y{t) = h[t) * v{t) and let the time
(d) Compute the zero-state response o f step T = \. Using Example 15.4 and
the circuit to the inputy(r) = 8 « (-f) + the convolution properties, argue that
?,u{t-T)y T > 0. y{t) has a piecewise linear structure as
shown in Figure P i5 .2 4 c. Find the
breakpoint coordinates y i, y2 y3 and
GRAPHICAL CONVOLUTION y4, first in terms o f the literal levels vO,
22. The impulse response o f a particular circuit v l, ho, hi, and h2, and then in terms
is measured approximately on an oscilloscope, o f their actual numerical values.
as illustrated in Figure P I 5.22. If the input to (b) Define three polynomials in a-
Chapter 15 • T im e Dom ain Circuit Response Com putations: T h e Convolution Method 803
hO
p(x) = vO X + vl
q(x) = ho x“ + h 1 X + h2
h2
r(x) = yl x^ + y2 x^ + y3 x +y4
V = [ 2 4];
h= [ 3 - 2 1];
T = 1;
tstep = T;
Figure P I5.24
y = [0 conv(v,h)*tstep 0];
% The additional beginning and
A N SW ERS: (a) yl = vO x hO = 6, y2 = vO x h 1
ending zeros are added to indicate
+ vl X ho = 8 , v3 = vO X h2 + vl x h 1 = - 6 . and
% that the initial and final values o f
v4 = vl x h2 = 4
the convolution are zero, due to the
% finite duration assumption,
25. Repeat Problem 24 for the waveforms in
t = 0: tstep : tstep* (length(v) +
Figure PI 5.25. When using the MATLAB
length(h));
code, it is necessary to account for/j(/) being
plot(t, y)
nonzero for negative t. This is easily accom
grid
plished simply by shifting/J(r) to the right by
one unit and then doing the indicated convo
v(t)
A lution; the proper result is obtained by a left
5 -- shift o f one unit.
vl f,(t)
44-
vO
2
1 2 -4..
(a)
(a) (b)
Figure P i 5 .2 5
.so^ Chapter 15 * T im e Domain Circuit Response Com putations: T h e Convolution Method
Figure P i5.26
-T 2T
(c)
Figu re P i 5 .2 8
Chapter 15 * Time Domain Circuit Response Computations: The Convolution Method
29. Repeat Problem 28(a) w ith^(t) as given in 32. Consider Figure P I 5.32.
Figure P I 5.29. Hint: Graphically int^rate^(/). (a) Find the transfer function H{s) and
O '
fj(t) the impulse response h(t) of the circuit
in Figure PI 5.32a.
w
(b) Find the step response using convolu
tion algebra methods.
(c) Find due to the rectangular
■> t pulse input in Figure P I5.32b . Hint:
Again, use convolution algebra meth
ods, and sketch the output waveform
-4 -- (or T = 2n^l(LC) .■
w (a)
A(.)A
w 2
1
Vw^ > t
/ -► tse c
2n 4n 6n
Vw>
(b)
Figure P I 5 .3 4
'O
806 Chapter 15 • T im e Domain Circuit Response Com putations: T h e Convolution Method
35. Compute the convolution of^(/) = h{t) * 37. This problem repeats the trick o f Problem
j{t) for each j { t ) below, given the h { t) o f Figure 36 of using the Laplace transform method to
P 15.35. compute responses when the input is nonzero
(a) j{t) = o r cos{ojt)u{t) for - T < t < 0 when T > 0. Suppose a circuit
(b) J{t) = (ur s\n{(Ot)u{t) has the transfer function
(c) J{t) = a^e-‘^'u{i) V A
H(s)= -
Vin s +a
LAPLACE TRANSFORM
36. This problem illustrates a trick for using the
A
Laplace transform method for computing
responses when the input is nonzero for - T q <
/ < 0 when > 0. Suppose a circuit has the
■>t
transfer function -T
V
Figure P I5.37
Vin s+2
Suppose the input to the circuit is shown 38. The ideas presented in Problems 36 and 37
in Figure P I 5.36. can be generalized as developed in this prob
(a) Define (t) = V/„( t - T ) and compute lem. Starting with the definition of convolu
tion integral, equation 15.2, prove the time
shift properties for convolution as follows:
(b) Compute ) due to the input
Vjnis) and then (a) f i t ) * = - T^) * }>(t) f=t+T,
(c) To compute due to use
time invariance, i.e.,
(b) /«)«^,'(0 = [ / « - r , ) * g(r - r , ^
Vout(^)=^’out(l + T')-
39. Two active circuits with impulse responses
v.(t) h^{t) = 2e~^‘u{t) and h-,{t) = are cas
caded as shown in Figure P I 5.39. No loading
occurs between stages. Compute the zero-state
■>t response for = ti{t + 2) V. Would it be
-T
advantageous to use the convolution method
here?
Figure P I5.36
Figure P I 5.39
'« 6
(a) Determine the value o f from (d) Find the complete response, given
the input = \Qe^^‘u{-t) V. Can vour answers to parts (b) and (c).
this be done using Liplace transform were doubled and =
techniques? It so, how? 72te~^‘u{t) V, find the complete
(b) Find the transfer fijnction response without doing any further
calculations.
H(s) = C,
^e
and the impulse response h{t) in terms
o f R and C. v„(t)
6
(c) Suppose R = 10 n and C = 10 mF.
Given (0") computed in part (a),
find the response due to the input Fig u re P i 5 . 4 2
inside each box is the impulse response o f the 47. Repeat Problem 4 5 for h^(t) = 4«(/), ^2(^) =
sub-circuit or subsystem. Suppose A,(r) = Au{t), 56(^), h^it) = 5 6 { t - 2 ) , and h^{t) =
h-^i) = 4 « ( f - 2 ), and = k cos{,m)u{t).
(a) Using any of the convolution tech 4 8 . Figure P I 5.48 shows a configuration for an
niques you have learned, compute and interconnection o f active circuits. The function
sketch the overall impulse response. inside each box is the impulse response of the
(b) Using any of the convolution tech circuit. Suppose h^(t) = 46{t), hjit) = 4 8 {t - 1),
niques you have learned, find y{t) for h^it) = 4d{t - 2 ). h^(t) = 45{t - 4), and h^{t) =
j{t) = 6 «(r). 2 cos(;rf)tt(r).
(a) Using any of the convolution tech
f(t)-
h,(t)
h (b)
niques you have learned, compute and
sketch the overall impulse response of
each configuration.
Using any of the convolution tech
niques you have learned, find y{t) for
h,(t)
M = 6u{t).
Figure P I5.43
K t)= y 6{t-kT)
f(t) > y(t) itti
be a so-called impulse train. Find L
Note that if z, a complex variable, then
your answer should be of the form
Figure P I5.45
00
2 f(k T ) z -'‘
46. Repeat Problem 45 with h^{t) changed to it=0
h^{t) = 4[u{t) - u (t- 4)]. Having done Problem which is the so called Z-transform of the
4 5 , can you do this problem without any fur sequence \J{kT) |^ = 0 ,1 ,2 ,...}.
ther calculations?
C!liapccr 15 * Tim e Domain Circuit Response Com putations: T h e Convolution Method 809
50. The circuit o f Figure Pi 5.50 is initially at 51. Consider the circuit o f Figure P I 5.50
rest. The input again. Let the input be the impulse train
cc shown in Figure PI 5.51, i.e., analytically
'■ ,„«)= ' ^ d U - k T )
A=() v ,„ (0 = ^ 6 { t + kT)
is a periodic impulse train with T = 1 sec. k=0
(a) Show that the impulse response is h{t) with r = 1. Find for r > 0 and plot the
resulting waveform.
(b) Find the exact solution o f for 0
<t< \.
(c) Find the exact solution o f for 1
< /< 2
(d) Repeat part (c) for the interval 4 < r < (1) (1) (1) (1) (1)
-> t
5. You should wind up with the -4 -3 -2 -1 0
expression
Figure P I5.51
AN SW ER: i>Jr) = 1.27t'“*'-^^ for r > 0
VoJt) = 0.5^-0-5('-'^)(1 + ^-0-5 +
( ,.- 0 .5 ) 2 ^ (^ -0 .5 )3 ^ (^ -0 .5 )4 )
20
'"6 IF
Figure P i 5 .5 0
C H A P T E R
Calling friends and others by phone occurs daily. When we dial a number, the information is sent
to the central office by one o f two methods: fast tone dialing or the much slower pulse dialing. For
example, electronic processing o f the pulse-dialed long-distance number 555-555-5555 requires
about 11 seconds, while electronic processing o f the tone-dialed takes only about 1 second.
The keypad o f an ordinary touch-tone phone has 12 buttons arranged in four rows and three
columns, as shown in the following diagram:
row 1 697 Hz
row 2 770 Hz
row 4 941 Hz
“ Chapter 16 • Band-Pass Circuits and Resonance
Pressing any one button generates two tones, with the frequencies selected by an electronic circuit
inside the telephone set. For example, pressing the number 5 generates tones at 770 and 1336 Hz.
The row and column arrangements and the dual-tone method permit the representation o f 10
digits (0 ,... , 9) and two symbols (*, #) using only seven tones. These seven tones are divided into
two groups: the low-frequency group, from 6 9 7 to 941 Hz, and the high-frequenc)' group, from
1209 to 1477 Hz. "
Such tones are easily produced by an LC resonant circuit. The four tones in the low-frequency
group are produced by connecting a capacitor to four different taps o f a single coil (inductor). A
similar connection generates the three tones in the high-frequenq' group. When a button is
pressed to the halRvay point, a dc current from the central office is sent through the coil in the
tank circuit. When the button is fully pressed, the dc current is interrupted. This action initiates
sinusoidal oscillations in the LC resonant tank circuit at a frequency inversely proportional to
V L C . The presence o f small resistances causes the oscillations o f the tank circuit to die out.
However, pressing the button fully also connects the tank circuit to a transistor circuit that replen-
ishes the lost energy and sustains the oscillations.
At the central office, the equipment used to detect the presence ol the tones and to identify their
frequencies is much more sophisticated. Two filters are required, one for each ol the frequency ^
groups. Each filter must pass the frequencies within ±2% o f their nominal values (697 to 941 Hz
for one filter and 1209 to 1477 Hz for the other) and reject the signal if the frequencies are out- ^
side o f ±3% limits. The output tone from each filter is then proce.ssed digitally to determine its
frequency.
The concepts and methods developed in this chapter will allow us to understand the properties o f
resonant circuits and the design o f various basic t}^pes o f band-pass circuits, as used in the touch-
tone telephone system.
CHAPTER OBJECTIVES ^
1. Describe and characterize the ideal band-pass filter.
2. Understand band-pass circuits from the viewpoint o f transfer functions and pole-zero
plots.
3. Investigate the basic band-pass transfer function and its realization as a parallel or series ^
RLC circuit or op amp circuit.
4. Investigate band-pass circuits having practical capacitors and inductors in contrast to ^
ideal capacitors and inductors
5. Describe the phenomenon o f resonance and investigate the properties and applications
o f resonant circuits.
6. Investigate general second-order transfer functions having a band-pass type o f frequency
response.
Chapter 16 • Band-Pass Circuits and Resonance
SECTION HEADINGS
1. Introduction
2. The Ideal Band-Pass Filter
3. The Basic Band-Pass Transfer Function and Its Circuit Realizations
4. Band-Pass Circuits with Practical Components
5. The Resonance Phenomenon and Resonant Circuits
6. General Structure of the Band-Pass Transfer Function with One Pair o f Complex
Poles
7. Summary
8. Terms and Concepts
9. Problems
1. INTRODUCTION
How is it possible to listen to a favorite radio station by merely pushing a button or two or sim
ply turning a dial? Why do some very expensive receivers have very clear reception, while with
some cheaper models other stations chatter in the background? What circuitry inside the radio
makes this difference? The ability to clearly select a particular broadcast station depends on the
design o f an internal band-pass circuit. Such a circuit will pass signals within a narrow band o f fre
quencies while rejecting or significantly attenuating signals outside o f that band. To understand
why this is important, note that audio signals have significant frequency components up to about
3 kHz for voice and up to about 15 kHz for high-fidelity music. These frequencies are far too low
for wireless transmission. In (wireless) AM radio transmission, the audio signal modulates the
amplitude o f a carrier signal that is suitable for wireless transmission. The carrier signal is a high-
frequency sinusoidal waveform bet\veen 500 kHz and 1650 kHz. The modulated waveform con
tains many frequency components centered about the carrier signal frequency, but extending over
a range o f frequencies equal to twice the highest audio frequency. For example, the radio station
WBAA, at Purdue University, has a carrier frequency o f 920 kHz and occupies a band from
approximately 915 to 925 kHz. To select WBAA from all the carrier signals received by a radio
requires a good band-pass filter to pass the frequency band o f 915 to 925 kHz while rejecting sig
nals outside this band. This chapter introduces the idea and properties o f a band-pass filter.
In its simplest form, a band-pass circuit consists o f only one capacitor, one inductor, and one resis
tor, connected either in series or in parallel. In the first half o f the text, we analyze simple RLC cir
cuits where we emphasize (1) transient behavior under dc excitation and (2) sinusoidal steady-state
behavior at a single frequency. This chapter investigates the behavior o f circuits over bands o f sinu
soidal frequencies. iMany useful results may be obtained with the phasor and impedance concepts
studied earlier. However, rapid advances in technology have made it possible to have a band-pass
circuit without any o f the usual RLC circuit components. Therefore, a study o f the band-pass
propert)' o f a transfer function H{s) dominates the material o f this chapter. The resulting analysis
is readily applicable to general linear systems, whether they be electrical, mechanical, or otherwise.
Although transfer functions underlie our approach, circuit realizations with ideal and practical
components illustrate all the basic concepts and properties.
S14 Chapter 16 • Band-Pass Circuits and Resonance
This section sets forth the ideal band-pass characteristic and shows how a basic second-order trans
fer function can be used to approximate the ideal. Specifically, let H{s) be a voltage gain or some
other type o f network function. As we know, the curves for \H{j(a)\ vs. (O and /.//(/□) vs. OJ are
called the magnitude (frequency) response and phase response, respectively. Figure 16.1a shows an
ideal band-pass (rectangular) magnitude response curve. Here, “ideal” means that all frequency
components o f the input signal within the range 0)j < to < (o^ are amplified with equal gain (in
magnitude), and all frequency components outside o f the range are totally eliminated from the
output. Actually, for a band-pass circuit to pass a signal with frequency components in the range
0)j < CO < 0J 2 without distortion, there is also a requirement on the phase response that is ordi
narily studied in a course on signal analysis.
(b )
Chapter 16 • Band-Pass Circuits and Resonance S IS
(C)
FIG URE 16.1 Definitions o f peak frequency, co^, and bandwidth, (a) Ideal band-pass character
istic. (b) Approximate band-pass characteristic of simple RLC circuit, (c) Band-pass characteristic of a
more complex circuit.
Half-power frequencies are also called 3 dB (down) frequencies. Recall from Chapter 14 that dB
gain is defined as
dB {gain) = 201ogjQ |Myw)|
If a gain o f is reduced by the factor \/j2 , then the dB reduction in gain is given by
- 3 d B = 2 0 log 10
71
H en ce the terminolog}^ “3 dB d o w n .”
Chapter 16 • Rand-Pass Circuits and Resonance
For obvious reasons, 0)| is called the lower half-power frequency and the upper half-power
frequency, and their difference = CO2 - co, is the half-power bandwidth (or simply bandwidth)
o f the band-pass circuit/transfer function. Band-pass circuits are designed so that all frequencies
o f interest fall within the pass band [(Op co^].
One way o f categorizing and comparing different band-pass circuits/transfer functions is by their
selectivity, i.e., a circuits relative capability to discriminate between frequencies inside the pass
band and signals outside the pass band. The selectivit)' is measured by the quality factor, Q, o f a
band-pass circuit/transfer function. The quality factor, Q, is the ratio o f the (geometric) center fre
quency {yj(JJiC0 2 ) to the bandwidth, . For second-order circuits and transfer functions with a
bell-shaped magnitude response (Figure 16.1 b), the center frequency and peak frequency coin
cide, as we will show. In this case Q = =
where Q denotes the pole Q, to be defined in equation 16.1. For the circuit realization o f the
transfer function, Q is sometimes denoted by or A high-Q circuit passes only a very
narrow band o f frequencies relative to yJcOiOh or whereas a low-Q circuit has a broad (pass)
band and a less selective characteristic.
It is important to note that the concepts o f B^^, and Q o f a circuit are all based on the magnitude
function |//(/o)| and, therefore, on how the transfer function H{s) is defined. Even for the same circuit,
these values are different when the output is associated with different branches or when the input is
changed from a voltage source to a currcnt source. Further, for the investigation o f the frequency-selec
tive characteristic of the circuit, the foregoing definition o f Q is most appropriate because it directly
assesses the sharpness o f the magnitude response cur\'e. As such the definition allows Q to be experi
mentally determined. In certain other applications, where only one fixed frequency is o f interest, there
is another definition o f a circuits Q based on an energ)' relationship that is inadequate for general band
pass circuit design.
The most basic second-order band-pass transfer function having the bell-shaped curve o f Figure
16. lb has a pair o f complex poles and a single zero at the origin, i.e.,
JO)
FIGIJRK 16.2 Pole-zero plot of a band-pass H{s) with a single zero at origin as per equation 16.1.
Our next goal is to derive the peak frequency the maximum gain =|
the half-power frequencies tOj and t02, the circuit Q, and the frequency at which the angle of
/y(/to) is zero. The first step in this derivation is to write down
K
(16.2a)
2 o » + 7 ----------- -
(JO
m ax H(Jco) 1^1 _ \K
(I) / 2 l\~\ / ■) 9 \- 2a,
I CO., - 0 ) I (OZ, - U>~ \
mm 4a“+ (16.2b)
(0 0) CO
Note that since the numerator is a constant, the maximum occurs when the denominator is a min
imum, i.e., when co = (O^. Therefore we conclude that
m p (16.3)
COI,
Noting that Qp ~ , we further conclude
H(jco)
(16.4)
To find the half-power frequencies lOj and co-,, the maximum gain must be reduced by the factor
l/V 2 . Considering equation l6 .2 b , this occurs at those co’s for which
818 Chapter 16 • Band-Pass Circuits and Resonance
(col-co-)
= 4a“ (16.5)
CO
We denote the two positive solutions by cOj and a>2. with CO2 > o jj. Solving equation 16.6 using
the quadratic formula, we have
(
yJcTp+OJ-p = CO.
(16.7)
CO,
B,^=co2-C0^=20p=-^. (16.8)
Exercise. Show that for high-Q^^ transfer functions/circuits (say > 8),
(16.9)
We can deduce one more important property from the above equations. When equation 16.5
holds, the real and imaginary parts o f the denominator in equation 16.2a are equal. Hence the
denominator has an angle o f ±45^^. Thus the angles o f
H(jcOi) = ----- ^
(cor-col\
2CT. + 7
CO:
Exercises. 1. Given the transfer function H{s) = — --------- , find , (0,^ , , and a>i 2 .
5 ^ + 2 ^ -h 2 5 6
A N SW I'RS: = 16 r.id/sec, //,,, = 0.5, = 2, o), = 15.031. and (o^ = 17.031
2. Suppose that = 8 and 10^ = 1000 rad/sec. Find the exact to^ ^ from equation 16.7, the
from equation 16.8, and the approximate Wj 2 from equation 16.9. Now compute the magnitude
o f the percent errors between the approximate and exact values o f tOj 2-
A N SW ERS: in random order; 125. 939.^^. 1064.5, 1062.5. 937.5. 6 .1833, 0 .2077
3. Given equation 16.9, show that for a high-(^ transfer function/circuit (say > 8),
i.e., for the high-Q case the geometric and arithmetic means of 0)j and 0J 2 are approximately equal.
The next two questions are: (i) what is the geometric center frequency o f the magnitude response
and (ii) is it equal to to ? Here, using the square o f the geometric center frequency,
2 2 ^
CO1 CO2 = -O p + J a j - ¥ c o l Op + p+<4 = - Gp + Op + ( O p = ( O p
we conclude that for the above transfer function and associated circuit realizations, the geometric
center frequency, the pole frequency, and the peak frequency coincide, i.e..
As mentioned earlier, the circuit/transfer function Q is the ratio o f the center frequency to the
bandwidth. Given equations 16.8 and 16.10, we have
Q = Q ..- Q ,, = ^ (16.11)
Finally, the nonzero frequency for zero phase shift occurs when the transfer hmction is purely real,
i.e, when the imaginary part is zero. This occurs at CO = CO^; hence
.(b )
FICiURE 16.3 (a) A parallel /?/,C circuit, (b) Magnitude response curve.
So lu t io n
The key to the solution o f the example is the computation o f the transfer function as follows:
1
1 1
—s
C
Y(s) ^ 1 1 ^ 1 1
^ C5 + — + — .V + - — .y-l- —
R Li RC LC
Then, according to equation 16.1,
—s
H{s) = K — =K C
(16.13b)
and
2a^ ( 1 6 .1 3 c )
Chapter 16 • Bund-Pass Circuirs and Resonance i '- 1
li,„ = 2 0 p = - ^ (16.14)
Similarly from equations 16.7 and 16.9, the exact half-power frequencies are
1
tO\ —H---------- 1- 4 +
IR C V \1RC} LC (16.15)
while the approximate half-power frequencies for the high-(^ case are
Exercises. 1. Compute the exact values o f tOp O),, and Q for the circuit o f Example 16.1
when R - 2.5 Z, = 0.1 H, and C = 0.1 uF.
AN SW ERS: (0,,, = 10"^ rad/sec, o , = 12.198 x lO 'l to. = 8.198 x 10\ = 4 x \o\ and Q = 2.S
The preceding example and exercise demonstrate that for high-Q circuits (Q > 8), there is really
no need to use the exact equation 16.15 to compute cOj and co-,, as the much simpler estimates
given by equation 16.16 are sufficiently close to the true answers. Indeed, sometimes it is con
venient to use the approximate formulas when Q > 6.
In many practical circuits, the independent source could be a voltage source in series with a resis
tor. Before applying any o f the foregoing formulas, it is necessary to transform the circuit into the
form o f Figure l6.3a by the use o f the Norton equivalent circuit studied in a first course on cir
cuits. The resistance R in Figure l6.3a then is not a physical resistor, but rather, the equivalent
resistance of several resistances in parallel. The following example illustrates this reformulation.
EXA M PLE 16.2. In the circuit o f Figure 16.4a, an independent voltage source in series with
an internal resistance R^ = 40 kQ models a real-world sinusoidal excitation. Suppose L = 20 mH,
C = 0.05 l i F, and Rj = \0 kQ.
(a) Find the exact values o f co^^^, B^^, and Q.
(b) Estimate the values o f OOj and co-,.
(c) Find I
S22 Chapter 16 • Band-Pass Circuits and Resonance
(a)
So lu t io n
The solution proceeds by using the formulas developed in Example 16.1 after replacing the prac
tical source by its Norton equivalent and identifying R in Example 16.1 as the parallel com
bination o f and /?^. By the usual formula for two parallel resistors,
R,R, 4 0 .0 0 0 x 1 0 ,0 0 0 _
R= — =— :------ = 8 ,0 0 0 Q
Rs + R l 4 0 ,0 0 0 + 10,000
This results in the transfer function
—s
C R,C
\ R 0 \ 1
'm s " + -------s ' - + -------------------------------------- S + —
RC LC RC LC
1 1
W,„=(Op = = 3 1 ,6 2 2 .7 7 rad/s
^ V 0 .0 2 x 0 .0 5 x 1 0 " ^
1 1
Bco = = 2,500 rad/s
-6
RC 8,000x0.05x10
Chapter 16 • Band-Pass Circuits and Resonance 823
and
6 . ^ . 3 1 .6 2 2 .7 7 ^ , , ^
Z50 0
(b) Using the approximate equations for high-Q circuits, we obtain the half-power frequen
cies as
a . A . , , ,
2cr^ R,CB^ /?,
Notice that we did not use equation 1 6 .13c because the input in Example 16.1 is a current source
and not the voltage source o f Figure l6.4a. Hence the maximum value is not equal to R.
This example demonstrates that putting an external resistance in parallel with the LC tank circuit
reduces the value o f R, which in turn causes a larger bandwidth and thus a lower circuit Q while
keeping the peak frequency unaffected.
Exercise. Repeat Example 16.2 with the element values changed to R^ = 36 kQ, L = 40 mH, C =
0.25 pF, and R^ = 4 kH.
A N SW ERS: 10,000 rad/sec, 1111.11 rad/sec, 9, 9 4 44.44 rad/sec, 10,555.55 rad/sec.
The examples so far have illustrated only the analysis o f parallel IU.C circuits. In the design of a par
allel-tuned circuit, we must also pay attention to other factors, such as available component sizes,
desired voltage gain, and cost. In practice, design specifications ordinarily impose a small number of
constraints relative to the number o f circuit parameters to be determined. Consequendy, realistic
design problems usually do not have a unique answer, as illustrated in the next example.
E X A M PLE 16.3.
Design a parallel RLC circuit, as shown in Figure l6 .4 a , to have a magnitude response with^^^ =
200 kHz and a bandwidth o f 20 kHz. Only inductors in the range 1 to 5 mH are available. The
source has a resistance R^ = 50 kH.
So lu t io n
For the circuit o f Figure l6 .4 a there is a restriction on the available inductors. Hence, we keep L
as a variable, subject to the condition that 0.001 < L < 0.005 H. Using the specified peak fre-
.S2-» Chapter 16 • Band-Pass Circuits and Rcsonancc
2 2 ^
quency and rhc fact that =
LC
C = - r r = - -r = 6 .3 3 x 1 0 "'^ L
1 -_L _L
R ~ Ri_ R, '
or 1 I I
R[ R R,
Once a specific value o f L is chosen, we can calculate successively the values of C, R, and Rj. Since
R, which is the parallel combination ot R^ and R^, must be no greater than = 5 x lO'^ f2, the
upper limit for L is
R. 5 0 .0 0 0
^ma\ ~ = 0 .0 0 3 9 8 H .
1 .2 5 7 x 1 0 ' 1.257 x 1 0 '
Numerical values corresponding to the extreme values ot L are given in Table 16.1.
TABLE 16.1
L (mH) C (p F ) R(VQ) Rl i ^ )
Table 16.1 clearly shows that there is no unique answer to the design problem. 'Fhe freedom in
choosing a value for L in the range 1 to 3.98 mH can be utilized to accommodate another design
specification, such as a value for
Exercises. 1. In Example 16.3, find the maximum and minimum values o f |//(/to^^^)|.
AN SW ER: 0.251, 1.0
2. In Example 16.3, if the bandwidth requirement is changed to 10 kHz, determine the minimum
and maximum possible values o f L.
AN SW ER: 1.0 mH and 1.99 mH
Chapter 16 • Band-Pass Circuits and Resonance H2S
Dual to the parallel /?/.Cof Figure 16.3a is the series y?Z.Cof Figure 16.5, which has a voltage input
and a current output. Although we can use clualit)', to infer frequency response behavior, we pre
fer the direct transfer function approach.
E XA M PLE 16.4. For the series RLC circuit o f Figure 16.5, let 1^ be the desired output. Find
R L
SO L U T IO N
(a) Find the t r a n s f e r By inspection,
(16.17)
z ,„ ( .) 1 -)
.y“ + — .v-l- -
C ,.'
= K - =K
S~ + l O n S + Clj:
5“ + —^S + W~
Q „ '
1
Peak frequency: (o„, = Wp = (16.18)
J lc ,
_ „ _ ^p^S _ 1
C ircu it Q: =Qp = ( 1 6 .1 9 )
Cs
826 Chapter 16 • Band-Pass Circuits and Resonance
(0 ,2 = 0 3 ^ ,
iV 4Q - 2Qpj ^ 2L,
-B co
C0\ ') •=' (0,„ H--------— • ,
1
i -------
^ l^s /, /'
2 2L, 0 6 -2 2 )
From equation 16.4,
_ |/f| _ \K\Q,, _ 1
(16.23)
2<J,, (o,„ R,
Exercise. For Example 16.4, suppose = 5 O., = 1 mH, = 0.1 |.iF, and has a fixed mag
nitude. Find o)^^, and the approximate half-power frequencies.
AN SW ER: = 10’’ rad/scc. = SOOO rad/sec, Q.-^ = 20. 97,500 rad/sec, and 102,500 rad/sec.
The above derivations and calculations yield formulas and numbers. To add some meaning to the
concept o f circuit Q and its relationship to bandwidth, we provide a plot o f the normalized |//(yco)|
vs. (0 curve for different values o f in Figure 16.6. The ordinate is the ratio
|//(yco)|/|//(yo))|^^^^^, while the abscissa shows the ratio to / to T h e se curves are called universal res
onance curves because they are applicable to parallel RLC circuits, to series RLC circuits, or to any
system having a transfer function o f the form o f equation 16.1. Observe that as Q increases, the
bandwidth decreases, indicating a better selectivit)^
Chapter 16 • Band-Pass Circuits and Resonance 82'
(O/Wm
FIGURE 16.6 Normalized magnitude response for equation 16.1.
To conclude this section, we present an example o f an active band-pass circuit that avoids the use
o f inductors. The band-pass circuit illustrated here is only one o f more than a dozen configura
tions in use. This example illustrates the possibility o f eliminating inductances while producing
the same kind o f frequency response as the parallel or series RLC circuit. You can learn a lot more
about these active filters in a more advanced course.
From the assumption that the operational amplifier is ideal, we have K_ = 0 (virtual ground) and
/_ = 0 (infinite input impedance). Applying KCL to the inverting input node V_, we obtain
■ 'C , K , + - ^ = o ,
which yields A-7
v;. = - ou t
V
^'ci - V
* -in
+ iC | V „ + iC 2 (K „ -V „ „ ,)= ()
/e,
Substituting the previous expression for into this equation, regrouping terms, and solving for
^out in the desired transfer function:
-s
~RxC^
(16.24)
1
+
,/?2C, R iC iV I^\Ri C\C2
Ks Ks Ks
H{s) =
V'/„(^V) + Ba) S + 1 0 ; „ 2 2 .v^ + 100 .v + (1000)“
i-" + S + iO
Q }
Chapter 16 • Band-Pass Circuits and Resonance
We now frequency-scale s -* with = 1000 rad/sec. Thus uhe normalized transfer function
is
_ s K
= — X ------- r r ^ --------= TTrrr x
lei
?>.Computation o f element valuesfor normalized transferfiinction. Equating the circuit transfer func
tion with the normalized transfer function, we have
C, = C, = 1 F
we have, from equation 16.26,
Ri R2C\C2 R[20
Note also that
= I H U co,„ ) I = I I = ^ =^ = 200
RjC^ ^ RoSi
4. Frequency and magnitude scaling. As per the problem statement, we desire = 1000 rad/sec.
Hence we frequency-scale with Kj- = 1000. It is further required that = 1 liF. Hence,
we solve the following for K^^-.
The ratio o f /?, to R^ is very large and unrealistic. It turns out that this circuit is best suited to low-
Q transfer functions.
5. Verif' frequency response. To verify the frequency response for a realistic implementation o f the
above circuit, we consider a SPIC E simulation using the standard 740 operational amplifier as
shown in Figure 16.8. Observe that the maximum value o f the magnitude response is 200, as
expected, and that the bandwidth is about 16 Hz, which translates to about 100 rad/sec.
830 Chapter 16 • Band-Pass Circuits and Resonance
Exercises. 1. In Example 16.5, find the circuit Q and the approximate half-power frequencies
CO
1,2-
AN SW ER. 10,950 rad/sec and 1050 rad/sec
2. In Example 16.5, if Cj = 0.5 uF and C j = 1 f.iF, find the values o f /?, and Rj required to meet
the specifications on and
AN SW ER: A’, = 66.67 R,= 30 kU
Throughout this section we have assumed the use o f ideal inductors and capacitors. Practical
inductors have complex circuit models to account for real-world behavior. The next section takes
up an approximate analysis o f circuits containing simple models o f practical inductors and capac
itors.
How can we analyze band-pass circuits in the presence o f practical (non-ideal) inductors and
capacitors? Practical inductors and capacitors have models consisting o f their ideal cousins and
other ideal (“parasitic”) elements to account tor losses and coupling effects. Figure 16. 9a illustrates
Chapter 16 • Band-Pass Circuits and Resonance 8 31
a simple model o f a practical inductor for low to medium frequencies while Figure l6 .9 b is a rea
sonable model for high frequencies.
FIGURE 16.9 Two models of an inductor, (a) For low to medium frequencies, (b) For high frequencies.
The primary parameter is, o f course, the inductance L. The remaining elements, R^, and C^,
account for undesirable yet unavoidable practical effects and are called parasitic. Since an inductor
usually consists o f a coil o f wire, R^ represents the wires resistance. Also, a capacitance is present
between adjacent turns o f wire. Hence models this parasitic capacitance. The resistance R
accounts for the energy loss in the magnetic core material (if present) inside the coil. Complex mod
els such as Figure 16.9b, although important, if used for every inductor would unduly complicate
the analysis o f a band-pass circuit. Fortunately, for low to medium frequencies (up to a few mega
hertz), the simpler model o f Figure l6.9a suffices and hence underlies the material that follows.
Figure 16.10 shows two models o f a practical capacitor. Again, the primary parameter here is the capac
itance C; R , Rp and are “parasitics.” The leakage resistance, R^, accounts for the energy loss in the
dielectric; tne inductance and resistance R^ are due mainly to the connecting wires o f the capacitor.
At frequencies above \j , the capacitor actually behaves as an inductor! For frequencies of up to
a few megahertz, the simpler model o f Figure 16.10a suffices, and it is used for the analyses o f this text.
(a)
FIGURE 16.10 Two models of a practical capacitor, (a) For low to medium frequencies.
(b) For high frequencies.
832 Chapter 16 • Band-Pass Circuits and Resonance
How close is a pracrical inductor (Figure l6.9a) or a practical capacitor (Figure 1 6 .10a) to the
ideal? The so-called element quality factor provides a quantitative measure. To develop the quali-
t)' factor, consider that each practical inductor or capacitor has an impedance
For both the practical inductor and capacitor, the frequenc)-dependent reactance X is the primary
parameter of concern, whereas R represents the unavoidable “parasitic effect.” In the ideal case, R
is not present. Hence, the ratio
provides a measure of how close the network model is to an ideal inductor or capacitor. The larg
er the ratio, the better the element behaves as an ideal inductor or capacitor, (ideally, R = 0 and
the ratio is infinite.) This suggests defining the quality factor associated with a practical inductor
or capacitor having impedance Z(yco) as
^ , UCw)!
The inclusion o f co in equation 16.27 is to emphasize the fact that Qy depends on the frequency'
o f operation. The subscript Z indicates a generic impedance and may be replaced by more specif
ic descriptors such as “coil” or “capacitor.” Unlike the Q o f a circuit, which depends on the values
of the elements of the circuit and on the circuits configuration, the quality factor Qj/oi a practi
cal inductor or capacitor varies with the operating frequenc)' OJ and remains unchanged irrespec
tive o f its connection in the circuit. Any element with a finite ratio given in equation 16.27 is
termed a lossy component, as are all real-world components.
For the practical inductor model o f Figure l6.9a, equation 16.27 reduces to
^ R^ (16.28)
=S = (’6.29)
Chapter 16 • Band-Pass Circuits and Resonance S33
A higher implies a better-qualiry capacitor, in the sense that the energ)' loss in the device is
smaller and closer to ideal. Infinite Q^-represents an idealized, lossless capacitor modeled by a pure
capacitance. In practice, Q^;;is usually much greater than Q^, meaning that is more critical for
circuit performance, i.e., Q(;;is often assumed to be infinite. The reciprocal o f Q^^ is called the dis
sipation factor o f the capacitor and is denoted by cIq A lower dissipation factor means a better-
quality capacitor.
The determination o f and requires specification o f the operating frequency co. If the value
o f 0) is unspecified, the analysis o f a band-pass (or tuned) circuit proceeds under the assumption
that = Q^(cOq) and Q ^= Q^oOq), where C0q = M \[l C . We will discuss the meaning o f oJq in
section 5, on resonance.
When one uses the practical inductor and capacitor models o f Figures 16.9a and 16.10a, the series
RLC and parallel /?ZC band-pass circuits with practical sources have the more complex configura
tions o f Figures 16.1 la and b, respectively. Since they are no longer series or parallel RLC, their
transfer function is not the ideal band-pass type of equation 16.1. Hence the associated formulas
for peak frequency, bandwidth, etc. are not directly applicable.
FIG URE 16.11 (a) Model of a parallel tuned circuit using practical inductor and capacitor models,
(b) Model of a series tuned circuit using practical inductor and capacitor models.
834 Chapter 16 • Band-Pass Circuits and Resonance
Topologically speaking, the circuits o f Figure 16.11 are series-parallel because the input imped
ance “seen” by the source consists o f a sequence o f series connections and parallel connections o f
simple networks. Exact analysis o f such series-parallel band-pass circuits to obtain and is
cumbersome, especially in light o f a simpler, more efficient method widely used by engineers to
compute approximate solutions. The approximate analysis relies on the conversion between a
series circuit and an equivalent parallel circuit at a particular frequency. This conversion process
depends on the component quality factors developed in the previous subsection. See Problem 85
for the development of this equivalence. The next example illustrates the conversion process for
an inductor.
EXA M PLE 16.6. When the ideal components o f a parallel RLC circuit are modeled with a prac
tical inductor and a practical capacitor, the circuit is no longer parallel. However, by converting
the series inductor model to an “equivalent” parallel model, we can proceed w'ith our standard
analysis. To illustrate this conversion, consider the practical inductor model o f Figure 16.12a and
the “equivalent” parallel configuration in Figure 1 6 .12b. The goal o f this example is to find and
L in terms o f and at a particular O).
(a) (b)
FIG URE 16.12 Conversion o f an inductor model from (a) a series connection to
(b) a parallel connection at a fixed frequency.
So lu t io n
Y<ko) = _____!____
yrU (o) = — - j
Rp
Equating the real and imaginary parts o f the above t\vo admittances, we obtain
ly
Rp = = Rs['‘ +Q h<o)) (16.30a)
and
Chapter 16 • Band-Pass Circuits and Resonance 83^
^ co"-L, \ q I uo)]
where
(oL
At any particular frequency (O, if Qy (to) is sufficiently large (say Qj (o)) > 8), then equations 16.30a
and b suggest that
^p-^s (16.31b)
Conclusion: In a "paraller RLC with a practical inductor, we can replace the practical inductor by its
parallel counterpart valid in a neighborhood o f a single frequency and analyze the circuit to obtain
approximate values o f peak frequency, bandiuidth, etc.
Exercise. A 2 mH coil purchased from an electronic parts store has a 50 at 100 kHz. Find
the element values in the series representation and the parallel representation at 100 kHz as shown
in Figure 16,35.
A N SW ERS: l.^ = 2 mH, R^ = 25.13 LX s 2 mH, R^ = 62.83 \dl
A similar derivation can be done for the practical capacitor model o f Figure 16.13a. The details of
the derivation are left as a homework problem. The exact conversion equations for a specific to are
(, \ \
and
( 1 \
(16.32b)
As before, for high (say >8), these equations reduce to the simpler forms o f
(16.33a)
and
1
^ (1 6 .3 3 b )
Q c (c o )
Chapter 16 • Band-Pass Circuits and Resonance
(a) (b)
F-KiriU-. 16.13 Conversion of a capacitor model from (a) a parallel connection to
(b) a series connection exact at a fixed fi-equency.
Table 16.2 summarizes the various conversion formulas for both parallel to series and series to par
allel inductor and capacitor models.
Approximate
Exact Equivalent Equivalent Circuit, for High Q
Original Circuit
Circuit at (d,, {Q, > 8 and > 8)
and u) within (1 ± 0 .0 5 ) oJq
Q^R.
R (i+ Q l)
=Q,.o)oL
Hf-
Q,((i)^ = o)„R,C
Q; Q,
CL
1 + 1/Q'
1
Qc(®o) = (o„R,C
Chapter 16 • Band-Pass Circuits and Resonance S '- ‘
The next example illustrates the application o f the conversion formulas to a “practical” circuit to
obtain approximate band-pass characteristics.
EXA M PLE 16.7. The circuit o f Figure 16.14a contains practical components that make the cir
cuit not amenable to the formulas for ideal series and parallel RLCs developed earlier. For exam
ple, the sinusoidal source is represented by an independent voltage source in series with an
internal resistance = 40 The practical capacitor has a value o f C = 0.05 |.iF and a dissipa
tion factor o f 0.01 at cOq. Here cOq represents the undamped (no resistance) natural frequen
cy o f the circuit, i.e., ^
“'» = 7 Z c -
The practical coil has an inductance o f 20 mH and Q/(cOq) = 40. The external load resistance is
/?^ = 10 k n . Our goal is to find approximate values o f cOj, CO2, and
(a)
leakage
signal source
(b)
FIG URE 16.14 Approximate analysis of a high-Q circuit, (a) Original circuit,
(b) Approximate parallel /?Z,C circuit.
Solution
Step 1. Compute cOq.
Step 2. Find the parallel equivalent circuit values for the practical inductor. Since Q^(cOq) = 40, from
equations 16.31 and column 3 o f Table 16.1,
Step 3. Represent the capacitor model by a parallel RC. First, the dissipation fiictor tells us that
, ^ ^ ^^ ^ ------------ ^ 0 ---------- ^ ^ ^
‘ ^ clcCo^C o jj: 3 1 6 2 3 x 0 .0 5 x 1 0 " ^ ’
Step 4. Replace the practical source with its Norton eqidvalent and compute the eqtdvalent parallel
resistance, denoted as R. Replacing the practical source with its Norton equivalent and incorporat
ing the results o f steps 2 and 3 produces the network o f Figure 1 6 .14b. The parallel combination
o f R^, R^, ^leakage* and R^ is (using the notation “//” to indicate the parallel combination)
R = 40,000//25,298//63,247//10,000 = 5545 n
Step 5. Approximate analysis o f the circuit o f Figure 16.14a using Figure 16.14b. Our formulas from
the ideal parallel RLC case now apply, but the results are, of course, approximate for the circuit o f
Figure 16.14a:
1
= — =— — — ------- = 3 ,6 0 7 rad /s ,
RC 5 , 5 4 5 x 0 . 0 5 x 1 0 “^
3,607
- - ^ = 3 1 ,6 2 3 - 1 ,8 0 3 .5 = 2 9 ,8 1 9 rad/s,
0 )2 s + - ^ = 3 1 ,6 2 3 + 1,803.5 = 3 3 ,4 2 6 rad / s
Finally,
The term “resonance” has different meanings in different disciplines. From Webster's Collegiate
Dictionary, in the field o f engineering, resonance refers to the phenomenon o f “a vibration o f large
amplitude in a mechanical or electrical system caused by a relatively small periodic stimulus o f the
same or nearly the same period as the natural vibration period o f the system.” In this section, we
shall investigate this notion o f resonance and its manifestation in RLC circuits, in both cases uti
lizing the theories studied in previous chapters. The main applications o f resonant circuits are for
the filtering and tuning purposes. Additionally, resonant circuits can be used to transform a resist
ance from one value to another value (at a single frequency) to achieve maximum power transfer.
This “matching” application is discussed in the last subsection.
Let us approximate the pendulum motion o f the swing using a linear system with transfer function
+ (16.34a)
Mathematically, the periodic leg movements are modeled by the periodic impulse function
where T = 2 7 t/ o )T h is choice coincides with the period o f the natural response o f the system
transfer function. The build-up in magnitude can be seen quickly through superposition. The
contribution to the output, sayj/(r), due to the first Impulse at / = 0 in equation 16.35 Is
Figure 1 6 .15a shows the form o f this equation. Note that successive positive peaks {Kq, aV^, a^V^,
a^Vf^, ...} decrease geometrically by a = e~^P^. In fact, the waveform o f each period replicates the
waveform o f the prior period scaled by the factor a.
From time invariance, the contribution to the output ^(/) due to the second impulse at r = 7'in
equation 16.35 is simply J'q(/‘ - 7), as shown in Figure 16.1 5b. Similarly, for the third input
impulse zi t =1T, the response 1s7q(^- 2T), as shown In Figure 1 6 .15c.
8-10 Chapter 16 • Band-Pass Circuits and Resonance
From superposition, the response over [0, 3T) \s simply the sum
i.e., for 0 < r < 37'. This sum is illustrated in Figure I6.15cl.
a^)
F IG U R E 1 6 .1 5
Chapter 16 • Rand-Pass Circuits and Resonance
In Figure l6.15cl \vc observe thar successive positive peaks arc given by (1 + (1 + a +
In general one can show that the «th positive peak is given by
n = 1
rt-1 (16.37)
11 + ^/+ t/“ H------ ~ ^ “
k=0 l-« "
1 - a
1- « \-a
This means that the waveform reaches a steady-state periodic response in which the positive peaks
have value v
\-a
Exercises. 1. Consider the slightly damped second-order system with transfer function
5“ + 4
(a) Find the first five peaks for the input o f the previous exercise.
(b) Is the system stable? Why or why not?
ANSWHRS; (a) 4, 8, 12, 16, 20. (b) The system is unstable bccausc there is a pole on the imagi
nary axis of the complex plane.
Another interesting application o f resonance is in product security in stores. The securit)' tag is an
RLC circuit with very small R. The circuit is excited by a “radio” wave at its resonant frequenc}' by
the security panels in front o f the exit doors. If the circuit has not been destroyed at the checkout
counter, it begins to resonate as one approaches the exit and transmits a signal back to a detecting
device at an amplitude much higher than the original transmitted signal. Fhis sets ofi an alarm.
The easiest way to understand this phenomenon is by way o f frequency response. Let us consider
the second-order transfer function ^
H(s) = --------- ----------
(5 -h 0 .1 )-+ 4
84 2 Chapter 16 • Band-Pass Circuits and Resonance
The magnitude frequency response is given in Figure 16.16. Notice the sharp peak at about 2
rad/sec, which is 10 times the dc gain o f about 1. Thus sinusoidal inputs at frequencies close to 2
rad/sec produce a steady-state response with magnitude almost 10 times larger. This is precisely
the type o f resonance phenomenon that occurs with the security tags at stores.
Frequency in rad/s
F IG U R E 16.16
/Y Y ^
R L
(a)
:D >R
^ c
(b)
Using the frequency response approach o f Figure 16.16, we see that small inputs in these circuits can
produce large outputs when the conditions are right: (i) a high-Q circuit and (ii) input signal fre
quency and the peak frequency coincide. To develop a circuit-theoretic perspective on resonance we
take a more basic approach using the sinusoidal steady-state analysis ideas o f Chapter 10.
For any circuit containing one inductance L and one capacitance C, denote by (Oq the frequency
at which the two reactances = -1/oJqC and = (OqZ have equal magnitudes, i.e.,
(16.38)
in which case
0)^ =
J lc (16.39)
In Chapter 9, on second-order RLC circuits, cOq = I I JT C is called the undamped natural fre
quency. The name stems from the fact that if all resistive elements are absent (i.e., the circuit is
undamped), then a parallel or series connection o f L and C produces a natural response o f the
form K cos((Oq^ + 0). In the jargon, the parallel LC circuit o f Figure 1 6 .17b is called a tank cir
cuit, and 0)q is called the tank frequency.
Since and jXj^ = have opposite signs and equal magnitudes, if the elements L
and C are connected in series, then at (Oq the resulting impedance, =
0 and hence is equivalent to a short circuit. Similarly, if the elements L and C are connected in
parallel, then at (Oq the resulting admittance is zero and hence equivalent to an open circuit. These
properties are illustrated in Figure 16.18.
O-
Z(jO)o)
Z(jCOo)
O- a
(a)
o- O-
Y(jcOo)
Y(jcOo)
o- O-
(b)
FIG URE 16.18 (a) Series- and (b) parallel-connected ZCelements with coq =
>/Zc ■
With the short circuit and open circuit equivalents sho\vn in Figure 16.18, we can easily deduce die fol
lowing properties for the series and parallel resonant circuits of Figure 16.17 operating a t : coq = ■.
8 ''il Chapter 16 * Band-Pass Circuits and Resonance*
(1) The magnitude o f the impedance seen by the voltage source in the series RLC o f Figure
16.17a is minimum, and the impedance is a pure resistance equal to R.
(2) The magnitude o f the impedance seen by the current source in the parallel RLC o f Figure
1 6 .17b is maximum, and the impedance is a pure resistance equal to R.
(3) The magnitude o f the voltage across C o r I in the series /?ZCof Figure 16.17a is Q^./;.(Wq)
= oJqZ.//? times the magnitude o f the source voltage.
(4) The magnitude o f the current through Z, or C in the parallel RLC o f Figure 1 6 .17b is
O^.y^(ojQ) = cOq RC times the magnitude o f the source current.
Exercise. Derive property (4) using current division and property (2).
Properties (1) and (2) lead to a general definition o f resonant frequency, denoted to^.. Specifically
is the frequenq' at which the source sees an impedance or an admittance that is purely real, i.e., pure
ly resistive, despite the presence o f capacitors and inductors. For the series and parallel RLC circuits
Exercise. A sinusoidal voltage source at 1 MHz is applied to a series RLC circuit. If Z. = 300 uH,
/? = 5 n , and C is adjustable, what value o f C produces resonance?
A N SW ER: 84.4 p 1-
Using property (2) above, the calculation o f in many o f the earlier examples can be done with
very little effort. For example, in Example 16.2, at resonance.
RL 10'
= 0.2
R, + Ri^ 4 x 1 0 -^ + 10-^
Exercise. Use propert}' (2) to calculate in Example 16.4 for Figure 16.5.
A N S W l - K : 1//^ *
The resonance condition co^ = (O^ = 1/V LC can be achieved by vary'ing one o f the three parameters
oJq, C, or L. When C or Z, or both arc adjusted to achieve resonance, the circuit is often called a tuned
circuit. The next example illustrates how this tuning can be used in a practical application.
Chapter 16 • Band-Pass Circuits and Resonance 84^
EXA M PLE 16.8. Figure 16.19 displays an amplifier model containing a V CC S with g^ = 2 mS
(milli-siemens) and = 20 k fi. The applied sinusoidal voltage, has a magnitude o f 0.1
V at 10 MHz. The load is modeled by the parallel combination o f and the 40 pF capacitor;
the capacitance accounts for such real-world phenomena as wiring capacitance, the device input
capacitance, and other embedded capacitances. This capacitance cannot be removed from the cir
cuit and often has deleterious effects on the amplifier performance.
FIGURE 16.19 Amplifier circuit illustrating the application of the tuned circuit conccpt to elimi
nate undesirable capacitive effccts.
So lu t io n
1
3 p r = 3 9 7 . 8 Z - 8 8 .9 ° Q .
0 .00005 + y 2 jrx 10 X 4 0 x 1 0
I ^ 0-002 X 3 9 7 .8 = 0.0 7 9 6
Here the voltage gain is 0.0796/0.1 = 0.796 due to the low impedance o f C at the high operating
frequency.
(b) By tuning out the effect o f the capacitance, this poor gain response can be eliminated.
The inductance needed to tune out the capacitance is calculatcd from equation 16.39:
W ith a 6.33 uH inductor connected across the load, the parallel Z.C behaves like an open circuit
at 10 MHz and the load looks like a pure resistance o f 20 kQ to the amplifier. The new output
voltage magnitude is
846 Chapter 16 • Band-Pass Circuits and Resonance
I^out 1^ 0-002 X 2 0 ,0 0 0 = 4 V
wich a resultant voltage gain o f 4/0.1 = 40.
2. In Example 16.8, if the embedded capacitance is 63.3 pF instead o f 40 pF, what inductance is
needed to tune out the capacitive effects?
AN SW ER: 4 IIH
Series-ParalelResonantCircuits
For parallel and series RLC circuits, co^ = cOq. For series-parallel circuits containing only one L and
one C when it exists to^ ^ (jd^j in general, but it may not exist at all. The next example illustrates
the point.
E X A M P L E 16.9. Find the resonant frequency, and the input impedance at co^for the circuit
shown in Figure 16.20.
/YYY
L
Z(jo))
SO L U T IO N
Step 1. Calculate the admittance "looking into ” the input node pair o f the circuit o f Figure 16.20. By
the usual techniques.
\m{Y}= o ) ,C ----- ^ = 0 .
R~ -t- (oifL)
Solving for and then expressing it as a function o f Wq = \JyjLC yields
1 R- CR- (16.41)
0). = 1-
Chapter 16 • Band-Pass Circuits and Resonance 84‘
The rightmost term shows how the resonant frequenc)'’ is scaled away from the parallel or series
ideal cases where to^ = cOq.
Step 3. To obtain the values o f the admittance and impedance at resonance, substitute this value
o f (Of into equation 16.40 to obtain
RC
Yij0^r) = (16.42)
RC
Results similar to equations 16.40 through 16.42 can be derived for the circuit o f Figure 16.21.
/YY\
Z(jco) L
R
Exercises. 1. For the circuit o f Figure 16.21, let G = MR. Show that the input impedance is
./ coC
coZ-- —^------7
G " + ( ojC ) G~ + (coC)“
L ( 1 6 .4 4 )
Z(jo),) =
RC
848 Chapter 16 • Band-I’ass Circuits and Rcsonancc
When co-exists, i.e., there is a real solution to equation 16.41, then at to =co^
the source “sees” a pure resistance, the value o f which equals {URQ.
This property finds application in maximum power transfer from source to load in fixed-fre-
quency situations, as illustrated by the next example.
E XA M PLE 16.10. The output stage o f a certain radio transmitter is represented by a 1 MHz sinu
soidal voltage source having a fixed magnitude o f 50 and an internal resistance o f 100 Q as
shown in Figure 16.22a. A load resistance Rj models an antenna connected to the transmitter also
shown in the figure. The purpose of this example is to shov/ how a m atching network based on
the principle o f resonance can be designed to maximize the power delivered to the antenna.
(a) If Rj is adjustable, find the value o f yielding the maximum average power absorbed
by the load. What is the value o f (/*jr)niax^
(b) If Rj = 2Q Q. in Figure 16.22a, find the value o f P^.
(c) Suppose that Rj^ is fixed at 20 O., but a coupling net%vork consisting o f LC elements is
inserted between the source and load to increase the power P^ as shown in Figure 16.22b.
Choose values for L and C in the circuit so that {Pi),j,^j^ o f part (a) is again obtained.
lOOO 1000
/YYV
L
50 V
- 0 K
20 0
IM Hz 1MHz
(a) (b)
FIG URE 16.22 Matching load to source using a resonant circuit, (a) Load connected direcdy to
source, (b) A coupling network designed to maximize the load power.
So lu t io n
Ri = R^ = 100 n
and
2,500
= 6.25 W.
AR, 400
3 .4 7 2 w .
20
Chapter 16 • Band-Pass Circuits and Rcsonancc S h9
(c) If we can make the impedance at the input terminals o f the LC coupling network equal
to (100 + yO) LI, then maximum power will be drawn from the source. Since Z.Celements
consume zero average power, the same maximum power will be delivered to the load
resistance. The resonant circuit shown in Figure 16.22b provides a possible design. The
“LCR" load circuit is the one analyzed in Example 16.9. Hence, to calculate the element
values, we use equations 16.41 and 16.42 as follows:
7 6\“ 1 20“
(oT = 2 ji X 10 j -----------(from equation 16.41, squared)
and LC
Exercise. Redesign the coupling network in Example 16.10 if the resistors are = 300 Q and Rj
= 50 n .
AN SW ERS: 17.79 ul 1 .nui 1.18 nl-
A variation on the computation o f the resonant frequency is the design o f a circuit to achieve a
desired resonant frequency using a variable capacitor; this is the design that underlies the tuning
o f many AM radios.
E X A M PLE 16.11. Consider the series RLC o f Figure 16.23. Here, the voltage source has a fixed
magnitude |V^| and a fixedfrequency O). W ith R and L fixed, we seek the value o f the variable capac
itance C that maximizes the magnitude o f the voltage across the capacitor.
F I G U R E 1 6 .2 3 A d ju stin g C fo r m a x im u m o u tp u t voltage.
850 Chapter 16 • Band-Pass Circuits and Resonance
The first step is to compute the magnitude o f the voltage across the capacitor using vohage divi
sion:
I, , I ^
Vc
o / 1
R -+ U)L-
V \/
coC/
Maximizing |K j is equivalent to maximizing \V(^ or minimizing------- . To obtain this last expres
sion, we square the above expression for \V^ and rewrite it as
/ '( O = + 2 ( lo ^ L C - \ W L = 0
1
C=
R~ +(coL) Zxo‘
+1
Although this value o f C produces a maximum capacitor voltage, the circuit is not hi resonance., as
the value o f 1/VLC is not equal to the signal source frequency o). However, given this equation
for C, for a high-Q circuit (ojZ//? > 8), the condition is practically the same as l/VTC = to.
SO L U T IO N
(a) For L = 1 mH, = toZ,//? = 10^ x 10"^/5 = 20. This is a high-Q circuit. Therefore, C is
given approximately by
C = - \ - = — ^ ------ r = ( ) . l x l O ”^’ F ,
to -L lO '^ x lO " '
l^ c U « = Q Jn i = 2ov
n 7 T ^ ^ 10-'’ X 0 .0 9 9 7 5 X 10
J5 “ + 1 0 ^ x 0 .0 0 1 -
10^ X 0 .0 9 9 7 5 X 10"^ J
Plainly, the approximate solutions are very close to the exact solutions.
(b) For £ = 100 uH, = coZ//? = 10^ x 10“^/5 = 2. This is a low-Q circuit, requiring the
use o f equations 16.48 and 16.47. Here,
0-0001 ,^ _ 6 ^
C=— 5 0 .8 x 1 0 F,
5 2 + ( 1 0 ^ x 0 .0 0 0 ir
and
= 2 .2 3 6 V .
2 lO-'^ X 0 . 8 x 1 0 “^
5^+ iO-'^x 0 .0 0 0 1 -
1 0 ^ x 0 .8 x 1 0 " ^ /
Exercise. For part (b) o f Example 16.12, compute cOq = 1/ V lC . Is this value equal to the signal
frequency? Why or why not?
AN SW I:R. {.)(, = 1 .1 1 8 x 1 0 ^ rad/scc
W ith the experience gained from the analysis o f the first five sections, we now present a general
transfer function approach to the analysis o f band-pass type circuits. The circuits considered in
this section will contain only one inductance and one capacitance. Hence any associated network
function will have at most a second-degree polynomial in s as the denominator and numerator.
Thus, the general transfer fiinction H{s) (which includes the impedance function Z{s) and the
admittance fiinction F(j) as special cases) has the biquadratic form
u( A - +
d(s) s~ + 2 o pS + io~p (16.45)
A reasonably sharp band-pass characteristic requires that H{s) have complex poles, i.e., < O)^.
The finite zeros o f H{s), which are roots o f n{s) = 0, may be real or complex. The case o f complex
zeros, corresponding to more advanced filter characteristics such as the inverse Chebyshev or ellip
tic types, is beyond the scope o f this text. For practical reasons we focus on the case where H{s)
has one real zero or no zero.
When H{s) has one real zero and complex poles, then equation 16.45 reduces to
+ _ a\S + ao
s- + 2 o ^ s + (o l +
e„ "
SS2 Chapter 16 • Band-Pass Circuics and Resonance
jco
A
p, = -^ p + i« d
\“p
(0p2 = ap2 + C02 s-plane
-e -
-a -a
(0
Q =
P2 = -^ p -j« d -
FIG UIIE 16.24 Pole-zero plot of the transfer function H{s) of equation 16.46.
Equation 16.46 contains the usual qualit)' factor, Q , mathematically called the pole defined as
CO,
(16.47a)
= arcsin (16.47b)
\^P/
where the angle 0 is as shown in Figure 16.24. Relative to the pole-zero plot, measures how
close the pole is to the/o-axis: a higher Q means a smaller 0, implying a pole closer to they'co-
axis. As we already know, is related to the circuit Q, i.e., and serves as a quick estimate o f
the sharpness o f the response curve. For some special cases,
In Figure 16.24, another new quantity, appears. To ascertain the meaning o f to^y, recall from
Chapter 15 that the impulse response o f a system characterized by equation 16.46 has the form
The waveform h{t) is a damped sinusoid, and the quantit)^ (not co^ specifies the frequency o f
oscillation. For this reason, co^ is referred to as the damped oscillation frequency.
For a transfer function o f the form o f equation 16.46, our goal is to determine several key quan
tities that are indicative o f the circuits behavior: the bandwidth and the half-power
frequencies.
Chapter 16 • Band-Pass Circuits and Rcsonancc
Case 1. No finite zeros. When the transfer function o f equation 16.46 has no finite zero, then it
reduces to
K K
H{s) = (16.48)
.V" +
a
The pole-zero plot o f this H(s) differs from that o f Figure 16.24 only in that the single zero is now
absent. The series RLC circuit o f Figure 16.25a with the capacitor voltage as the output and the par
allel RLC o f Figure 16.25b with the inductor current as the output both fall into this category.
fYYV-
V.
(a) (b)
The transfer function o f equation 16.48 can represent either a low-pass or a band-pass character
istic depending on the value o f and the value o f H{0). In fact, the transfer function can display
both characteristics as well as intermediate behaviors. For example, if
o ,,, _________OOOO)-
1 1000
■V“ +
VI
then the maximally flat low-pass characteristic o f Figure 16.26a results.
(1000)^
H{5) =
.v“ + 5 + (1 0 0 0 )-
1.31
then something between a low-pass and a band-pass characteristic results, as shown in Figure
16.26b. Here low frequencies are still passed, yet the characteristic has a selectivit)' propert}^ result
ing from the pole o f 1.31. The ratio o f the maximum gain to dc gain is yfl . This means that
the peak is 3 dB above the dc gain. Now, with increasing pole Q the selectivity goes up, as does
the maximum gain, and the characteristic looks more and more like a pure band-pass. Finally, if
H{s) =
. (1000)*-
1000
.v“ + -------- s (1000)*
10
then the approximate band-pass characteristic o f Figure 16.26c results. Here the pole Q is 10 and
the ratio o f the maximum gain to dc gain is also 10. Although low frequencies are not attenuat-
854 Chapter 16 • Band-Pass Circuits and Resonance
ed, the characteristic is highly selective and frequencies near co^ are highly amplified, so that for
all practical purposes the characteristic is identified as band-pass.
Frequency in rad/sec
(a)
(b)
Frequency in rad/sec
(c)
FIGURK 16.26 (a) Low-pass characteristic, (b) Moderate Q^, resulting in characteristic exhibiting
both low-pass and band-pass behavior, (c) High-Q^ case, showing a marked band-pass characteristic.
Chapter 16 • Band-Pass Circuits and Resonance 8^5
The important aspects of the curves in Figure 16.26 are co^, and We first compute O)^
and then At the peak value, the derivative o f |//(/o))| is zero. Setting the derivative of |//(/to)|
to zero and solving for O) produces the exact formula ^
To compute consider
'm 2 2
^(Op - ct)^ co s(2 0 )j + (U p 4sin ^ (0)co s(20)
(o „ = o )p y Jco s(2 d )
K ____ H (0)
^/»= 2 . 0 6 .5 0 )
in(2d) sin(20
Similar derivations with more complex algebra yield the half-power frequencies,
^(o = « p ^ 2 (c o s ( 2 0 ) - 7 c o s ( 4 0 ) ) (16.52)
When is greater than l/V 2 or, equivalendy, if 0 < 4 5 °, the magnitude response of equation
W 16.48 starts fi’om a nonzero value at co = 0, rises to the peak value at co = O)^, and finally decreas
es to zero as CO -► 00. This behavior is illustrated in Figures 16.26b and c. If is only slighdy
greater than 0.707, then the magnitude response is essentially that of the low-pass type, with a
small hump in the pass band.
For the high-(^ case ( ( ^ > 8), the magnitude response near co^ approximates that of a pure band
pass circuit. The preceding exact expression for (o^ reduces to O)^ a o)^ for the high-(^ case.
856 Chapter 16 • Band-Pass Circuits and Resonance
t^\Qp
Similarly, for high Q^, s 2a^, . and ojj 2 - I'"* error o f these
approximations is less than 0 .5 % .
Exercise. For = 8, (o^ = 1000 rad/sec, and A^= in equation 16.48, computc the exact and approx
imate values o f and Then compute the percentage error in the approximation:
\exact - approximate
X 100
\exacf\
AN SW liRS: Exact values arc = 99 6 .0 9 rad/sec, B^^^ - 125.5 rad/sec, and - 8.015^.
Approximate values are 1000 rad/sec, B^^^ = 125 rad/sec. and = 8. Percentage errors are,
respectively, 0 .3 9 ‘^(), O-.V)'*/!), and 0.196% .
EXA M PLE 16.13. Consider the Sallen and K eyzcw e network o f Figure 16.27, which can be used
to realize the transfer function o f equation 16.48. As per a homework exercise in Chapter 14, the
transfer function is
1
' nut Ri R2C[C2
H(,v) =
1
U ,c , R2 C J R\R2<^iC2
FIGURE 16.27 Sallen and Key active net\vork for realizing the transfer
function o f equation 16.48.
Chapter 16 • Band-Pass Circuits and Rcsonancc 8^'
So lu t io n
'm + 5 + (1 0 0 0 )“
10
(a) By inspection (o^ = 1000 rad/sec, = 10, and K = = 10^’.
»n = wp^2;
»d = [1 w p / Q p w p ^ 2 ];
»Hm = abs(pol\^al(n,j*wm))/abs(pol)^al(d,j'wm))
Hm = 1.0013e+01
s 2 a ^ = 1 0 0 , 0 ), a co ^ - = 9 0 0 , CO 2^ = 1 1 0 0 ,
all in rad/sec.
(d) 'lb obtain the magnitude response for the given transfer function, as shown in Figure
16.28, we use the following MATLAB code:
»n = wp^2;
»d = [1 w p / Q p w p '^ 2];
»f = logspace(l ,3,600);
»w = 2*pi*f;
»h = freqs(n,d,w);
»semilogx(f,abs(h))
»grid
»xlabel(‘Frequency in Hz’)
»ylabel(‘Magnitude H(jw)’)
858 Chapter 16 • Band-Pass Circuits and Resonance
Frequency in Hz
FIG URE 16.28 Magnitude response o f high-Q^ active circuit of Figure 16.27.
Exercise. Suppose Cj = 0.2 j.iF, Cj = 0.5 nF, and /?j = ^2 ~ A"of equa
tion 16.48, along with exact values o f and Find approximate values o f B^, Wp and Wj.
A N SW ERS: = 10^ rad/sec, = 10, K = I0 ‘\ = 9975 rad/sec. /-/,,, = 10.013, /?,„ = 10-^
rad/sec, 10, 2 9000 rad/scc, to, = 1 1,000 rad/scc
Case 2. A single zero o jf the origin. In equation 16.46 with a^^ 0, i=- 0, and and <7, o f the
same sign, a zero is present in the left half-plane. A second form o f the band-pass transfer func
tion o f equation 16.16 is
s +a s+ a
H{s) = K =K
CO, s~ + 2 0 p.v + oj“
s +• (16.53)
Qn
Figure 16.32 sketches the pole-zero plot for this transfer function. Intuitively speaking, the closer
the zero is to the origin, the more the magnitude response resembles the response o f the case with
a single zero at the origin. For the transfer function o f equation 16.53, derivations o f values
of and the zero-phase-shift frequency are possible. (See the homework problems.) The results
are
No exact expressions are available for the half-power frequencies and the bandwidth For the
case o f high and a « approximate answers are
jc o + a jo3 + a
= A. 9 -) K
IS. 9 X (16.57a)
( » “ +j2oxj^^ + co“ (yco)“ + j 2 o i( j j , + co“ >
K
1+ (16.57b)
/ c o - - c o “ '\
2 o „ + ./
03
The second factor is approximately 1 for frequencies near co^ by virtue o f our assumption a « OJ^.
Hence the properties o f the magnitude response reduce approximately to those o f the first factor
in equation 16.57b. These properties are those o f a single zero at the origin. Hence for high-Q^^
circuits, the relations o f equation 16.56 approximate the single zero at the origin case.
(S60 Chapter 16 • Band-Pass Circuits and Resonance
1^1
Exercise. Use equation 16.57b to show that when > 8 and a « to , H ~
■P P ?Q ■
When exact values are desired, one must resort to SPIC E or MATLAB to obtain the frequency
response from which values can be graphically determined. SPIC E has the advantage o f not hav
ing to compute the transfer function o f the circuit; MATLAB requires this computation.
7. SUMMARY
This chapter began with a study o f simple series and parallel band-pass circuits. Because o f
its generality, we set forth a transfer function approach to the analysis and design o f these band
pass circuits. Specifically, we first investigated a second-order transfer function with a single zero
at the origin for which
Qp '
For this transfer function we derived various formulas for determining band-pass parameters, such
as the peak frequency,
a~p
These formulas have a special form for the parallel RLO.
to,,, = 0)^ =
4 lc
which is also the resonant frequency of the associated impedance; the bandwidth is found to be
^ (0 = t0 2 -0 J, = - ^
while the circuit Q is
0 = “ = u .„ R C = « ^
The resonance phenomenon of a second-order RLC circuit was then set forth from a frequency
domain perspective. One application was the design o f a matching network that produces maxi
mum power transfer, at a single frequency, from a source with fixed internal resistance to a fixed
resistance load.
Chapter 16 • Band-Pass Circuits and Rcsonancc iSOl
After this we cook up the consideration o f more general second-order band-pass transfer func
tions— for example, those with a single zero o ff the origin,
Qp "
Active bajid-pass circuit: a circuit containing operational amplifiers and no inductors that
achieves a band-pass characteristic.
Band-pass circuit: a circuit that passes signals within a band o f frequencies while rejecting other
frequency components outside o f the band.
Bandwidth (3 dB bandwidth), CO2 - cOp the difference between the two half-power fre
quencies.
Dam ped oscillation frequency, (0^ frequency given by the condition that if the transfer func
tion o f a second-order linear circuit has complex poles at j = -a ^ ± jio^, then the impulse
response has the form K cos(co^f + 0). The constant (jo^ is called the dayyiped namralfre
quency.
Half-power frequencies; see lower and upper half-power frequencies.
L C resonance frequency: frequenc)' at which the reactances o f L and Chave the same magnitude;
equals M-JLC rad/sec. (Sec also tank frequency.)
Lower half-power frequency, COj: the radian frequency below the center frequency at which the
magnitude response is 0.707 times the maximum value.
M atching network: an LC net\vork that transforms a resistance into a resistance o f a differ
ent, specified value at one frequenc)- or a band o f frequencies.
Peak frequency (center frequency), the radian frequency at which the magnitude response
curve reaches its peak.
Qp (pole Q): for a pair o f complex poles j = Qp = 1/(2 sin 0).
Quality factor Q (Qr/>) o f a band-pass circuit: the ratio o f the center frequency to the band
width, i.e., Q =
Quality factor Q£ {Qcoit) o f a coil: for a coil modeled by an inductance L in series with a resist
ance Rs y Qi = and is frequenc}^ dependent.
Q uality factor Q c iQcap) ® capacitor: for a capacitor modeled by a capacitance C in parallel
with a resistance R^^, Q c~ frequency dependent.
Q uality factor o f a reactive com ponent: for an impedance expressed zs Z = R + jX, Q^ =
Reactance: in sinusoidal steady-state analysis, the imaginary part o f an impcdance. For Z,, the reac
tance is for C the reactance is -l/(coC).
862 Chapter 16 • Band-Pass Circuits and Resonance
----------------------------------------------------------------------------------------------------------------------- ^
Resonance frequency, 00^ the unique radian frequenq^ at which the input impedance of a two-
terminal linear circuit becomes purely resistive.
Selectivity of a band-pass circuit: The circuit Q, defined as the ratio of the center frequency to
the bandwidth. A higher Q corresponds to better selectivity.
Susceptance: in sinusoidal steady-state analysis, the imaginary part of an admittance. For C, the
susceptance is coC For Z, the susceptance is -l/(coZ).
Tank circuit: the parallel connection of an inductor and a capacitor. In the idealized case (no
resistance), the total enei^ stored in a tank circuit remains constant, although there is a
continuous interchange of the energy stored in the various components.
Tank frequency, (Dq: defined as H j W in this text, regardless of the connection of the single L
and single C with other components in the circuit.
Tuned circuit: a second-order circuit containing one inductance and one capacitance, at least one
of which is adjustable to reach a condition of near resonance.
Undamped natural frequency: the natural fi-equency of a circuit consisting of lossless inductors
and capacitors. For the case of one inductor and one capacitor connected together, this
frequency is the same as the LC resonance frequency or the tank firequency and is equal
to m J l c .
Universal resonance curve: a normalized magnitude response curve of a band-pass transfer func
tion having one pair of complex poles and a single zero at the origin. The magnitude is
normalized with respect to the maximum gain, and the frequency is normalized with
respect to the center frequency.
Upper half-power frequency, (Oj: the radian frequency above the center frequency at which the
magnitude response is 0.707 times the maximum value.
r> .
* W e would like to thank Les Axelrod o f the Illinois Institute o f Technology for providing derivations that led to
this formula and those o f the bandwidth and half-power frequencies.
Chapter 16 • Band-Pass Circuits and Resonance 863
C H EC K ; poles are p, 2 = ±
'Problems
5 . Suppose a basic band-pass transfer function
BASIC BAND-PASS TRANSFER has = V s 200 rad/sec, = 10, and =
FUNCTION PROBLEMS 8 rad/sec.
1. Fill in the details o f the derivation of (a) Find H{s), its poles, and (X.
(b) Find approximate values For cOj and
1 (O2. Using MATLAB, plot the magni
0) 1.2 = ±a^; + p = ^ p i+ -
4G,-; 22, tude response and verify your calcula
tions.
from , find a parallel
K
N(Jco) = K — ^ T = / 2 _ 2\ circuit realization o f H{s).
i j o i ) + j l w o p+CO p I UJ uJ p \ _____ _
2 o „ -h y C H EC K : Q^= 11.32
2. Suppose the basic second-order band-pass 6 . Suppose a series RLC circuit has = 80
transfer function o f equation 16.1 has poles at rad/sec, = 8 , C = 1/80 F, and output l^{s)
-1 ± j5 and the gain at is 2 . Find K, H{s), with voltage source input
^1’ ^2- (a) Write down the transfer fijnction for the
quency response using MATLAB and verify series RLC circuit in terms o f R, L, and C.
your answers. (b) Find the values o f R and L.
C H EC K : cOj = 2.831 rad/sec (c) Find m
and approximate values for
OJj and 0)^.
3. Suppose the basic second-order band-pass (d) Use SPIC E to obtain the magnitude
transfer function o f equation 16.1 has poles at (frequency) response and verify your
- 8 0 ± y l5 9 9 and the gain at co^ is 6.25. Find answers to part (c).
K, H{s)y Q, exact and approximate C H EC K : /?= 0.125 n
values for tUj and Wj, and the relative percent
error in the approximate computation o f CO2. 7. Suppose a series RLC circuit has = 100
Using MATLAB, plot the magnitude response rad/sec, = 4, C = 0.01 F, and output
and verify your calculations. with voltage source input
C H EC K : t02 = 1681 rad/sec while the approx (a) Write down the transfer function for
imate value o f CO2 is 1679 rad/sec the series RLC circuit in terms o f R, L,
and C
4. Suppose a basic band-pass transfer function has (b) Find the values o f R and L.
to = V S 2 rad/sec, = 10, and B = 2 rad/sec. (c) Find and approximate values for
(a) Find H{s) and its poles. cOj and CO2.
(b) Find tOj and CO2. Using MATLAB, (d) Use SPIC E to obtain the magnitude
plot the magnitude response and veri (frequenqO response and verif}' your
fy' your calculations. answers to part (c).
C H EC K : I = 0.01 H and =4 S
Vguri-^) , find a parallel
■m(^)
8 . Consider the circuit o f Figure P I 6.8 for
circuit realization o f M{s).
which R^ = 40 k n , R, = 10 /• = 10 mH,
(d) Use SPIC E to obtain the magnitude
and C = 1 liF.
frequency response and compare it
with your answer in part (b).
S64 Chapter 16 • Band-Pass Circuits and Resonance
(a) Find the exact value o f the maximum P I 6.10, w h e r e = 2 mS, 100 Q, and
voltage gain and the corresponding = 10 \dl.
frequenc)' (in Hz). (a) Find H{s) in literal form.
(b) Find the exact 3 dB bandwidth (in (b) If A = 100 u H ,^ = 1040 kHz, and Bjr
Hz). = 10 kHz, find C, R, and
(c) Find the circuit Q. (c) If^^^ = 920 kHz, Bj= 10 kHz, and C
(d) Find approximate values o f the upper = 250 pF, find A, /c, and
(cOo) and lower ((o,) half-power fre
quencies (in Hz).
(e) Use iVIATLAB or its equivalent to v
plot |//(_/0))| vs. CO.
Using these components, find the A N SW ER: (a) 2.S mH .uui 530 12
(a)
range o f allowable inductance so that
the circuit can be tuned from 550 to 14. For the two-terminal circuit in Figure
Figure P i 6 .1 5
866 Chapter 16 • Band-Pass Circuits and Resonance
17. For the series resonant circuit shown in 20. For the circuit in Figure P I 6 . 2 0 1 kQ,
Figure P16.17, /?, = 40 ^ Z- = 0.8 H, C= 1.25 Z. = 10 niH, C = 0 .1 mF, = 100 and the
uF, and R j= 160 Q. op amp is ideal.
(a) Find the transfer function H{s) = (a) Construct the transfer function H{s) =
= -Zj(s)IZ,,{s) in terms
(b) Find the exact value o f the maxirnum o f the circuit elements R-^^, Rp L, and
voltage gain and the corresponding C, and put it in the general form
peak frequency (in Hz).;
(c) Find the exact 3 dB bandwidth (in s ~ ■ \-la p S + {H~p
Hz).
(d) Find the exact values o f the upper and (b) Find the values o f K, 0)_, the circuit
lower half-power frequencies. Q, and = |M;co,„)|.
(e) Find Q o f the circuit. (c) Compute the value o f the half-power
(0 Plot the magnitude response using bandwidth and the half-powcr fre
MATLAB or the equivalent. quencies CO, and oyj. (Approximate
values are acceptable.)
m v
(d) Sketch the pole-zero diagram that rep
R. L
resents the circuit, and note the exact
locations o f all the poles and zeros.
(e) If v-J^t) = lOOsin(lO^f) mV, determine
the magnitude o f Vg^^f{t) in steady
Figure P i6.17
state.
ANSWER: (b) w, - f£ ± r
V LC '
Figure P I6.23
''out
1
Figure P I6.21 + s +-
^2^2/ ^ 1^ 2Q ^ 2
22. For the circuit shown in Figure P I6.22: show' that
(a) Find the transfer function H{s) =
(i) H.„ =
C,
-hi and (ii)
Q- ..c,
(b) Find and Q.
Co R'
^2 Hi C,
(c) Now supposey^^^ = 200 Hz and Bj-=20
Hz are desired. One has available a 1 Hence, for high-Q transfer functions, this cir
H inductor, a 10 jxF capacitor, and cuit is undesirable.
arbitrary resistors.
(i) Determine the necessary value o f j.i. 25. Reconsider the active band-pass circuit o f
(ii) Determine the value o f R. Figure P16.23. The filter is to pick out the
midrange o f a typical audio speaker, in which
case/; = 500 Hz ,/2 = 3200 Hz, a n d ^ = 1265
Hz. The gain is to be 10. Find Bj'd.nd the
circuit parameters. Hint: Follow the procedure
described in Example 16.5. Use SPIC E or its
equivalent to generate the magnitude response
plot for 1 < / < 5000 Hz.
C H EC K : Q < 1 and R^ > 2, slightly
Figure P I6.22
A N SW ER S: (a) 12.S; (c) 10; (cl) = 20 (a) Find the capacitor Q at Wq = j-----.
(b) Find an approximate series
representation o f the capacitor near ( jO q .
(c) Find the approximate value o f Q.
(d) If the circuit is high Q, compute
approximate values for
0)j, and CO2.
, CYY\
Coil L
Figure P I6.29
T
31. Consider the RLC circuit in Figure P 16.31, 34 . Consider the circuit o f Figure P I 6.34.
in which = 20 L = 0.5 mH, = 20 Suppose R^ = 0.6 O., L = \ mH, R^- = 2.5 kT2,
Q., and C = 0.5 nF. ^ and C = 0.1 l i F. j
(a) Find the coil Q at wq = ------- (a) Find the capacitor Q at (Oq - -------
V LC
(b) Find an approximate series
(b) Find an approximate parallel represen
representation o f the capacitor near o jq .
tation o f the coil near cOq.
(c) Find the approximate circuit Q.
(c) Find the approximate circuit Q.
(d) If the circuit is high Q, compute
(d) If the circuit is high Q, compute
approximate values for
approximate values for
(Oj, and to^.
o jj, and t02-
R.
C H EC K ; 100
Figure P I6.31
3 5 . Repeat Problem 34 for R^ = 0.32 D., L = \6
32. Repeat Problem 31 for R^ = 40 k ii, L = 4 l i H , /?c= 800 a , and C = 0.25 uE
r>
;c
Suppose you have been asked by your supervi
sor to use this circuit in a band-pass design with
Figure P I6.36 f „ = ^ M Hz and a bandwidth o f 20 kHz and a
C H EC K : Q „ ,= 16 lossy capacitor with C = 2 0 0 pF and having
37. Again consider the circuit o f Figure PI 6 .36. (a) Determine the circuit Q, i.e.,
Verify that for high and high at cOq, (b) Find the necessary to achieve the
desired
^ Q coilQ cap
(c) F ind the inductance L o f the lossy coil
Qcoil + Qcap and then find R^ as shown in Figure
38. Now consider the circuit o f Figure P I 6.38. P16.38.
Verify that for high and high at (Oq, *(d) Suppose we now desire to double the
bandwidth by adding a resistor, R^g^„^
QcoilQcap in parallel with the lossy coil (i.e., in
a
Q coil Qcap parallel with the current source).
W hat is the proper value of R^source'
^ ^
Lossy Inductor
C H EC K : (c) R^<SQ.
^ QcoilQcap
3 9 . Consider again the circuit o f Figure Qdr = „ .f /
^coil *•*^cap
P I 6.38. Suppose L = 0.5 m H , C = 1.25 jiF,
= 30, and = 60 at cDq. Suppose you have been asked by your supervi
(a) Find R^ and R^. sor to use this circuit in a band-pass design with
(b) Conven the inductor to a parallel = 1 MHz, a bandwidth o f 2 0 kHz, and a
Chapter 16 • Band-Pass Circuits and Resonance 8 71
(cl) Cheek your results by doing a SPICE (or A N SW ERS: (b) 10.000 rad/scc. H).
equiv;ilent) simulation of your circuit. (i»^ s =£ 10^’ rad/sec. (O, a ‘)S0.0()0
rad/sec. (u-, s 1,050,000 rad/sec; (c) =
______ I. r~ coii
0.1 sin(oj,-r) V
0.01 H 750
10’“F
3000
RESONANT CIRCUITS WITH
SOOkO
APPLICATIONS
48. For each two-terminal circuit in Figure
capacitor P I6.48, the resonant frequency/^ = 10 kHz.
If C = 0.1 iiF and R = 1 kH, find the corre
Figure P I6.46
sponding values o f L. For each circuit, deter
mine the input impedance at resonance.
47. Consider the circuit in Figure P i6.47, which
contains a non-ideal capacitor, a non-ideal induc
tor, and a meter to measure the current response, Z(j(o)
Figure P i 6 .4 7
Chapter 16 • Band-Pass Circuits and Rcsonancc
(a)
(c) Plot the frequency responses o f the cir
cuit over 1 M H z < / < 10 M Hz with L
— fyy\- and without L.
L i „(t)
Z(j(o)
(!)
(b)
Figure P I6.49
amplifier load
50. Figure P I 6.5 0 displays an amplifier
model containingo a V CC S with ofn
?•,„ = 4 mS Figure P i6.51
capacitance, find the value o f L and Verify that at co^, the phase angle o f is
the resulting that will show that zero.
the amplifier gain at 10 MHz is great R
ly Increased.
(t)
Figure P I6.52
AN SW ERS: 480 rad/sec, 35.71 LI
Figure P i6.56
Figure P i6.57
Z{j(0,.) =
CR
58. A two-terminal network has input imped
ance
Chapter 16 • Band-Pass Circuits and Rcsonancc
Y(s)
Y„(s)
62. Equations 16.41 through 16.44, or those
derived in Problem 55, can be combined to
produce a set o f design formulas for a lossless
Figure P i6.59 network that matches two unequal resistances
at a single frequenc)^ The matching network
60. This problem uses an LC coupling network consists o f only one capacitance C and one
to maximize power to a load. The LC coupling inductance L, as shown in Figure P I6.62. At a
network always has a series connection for L\ specified frequency OJ, it is desired to have
the “parallel” capacitor is always closest to the matching at both ends, i.e.,
larger resistance— in this case the source resist
ance. Problem 62 will consider the general case. Z , (yo)) = /?, + /O and Zjijiyit) = R2 + P
Suppose the voltage source in Figure P I 6.60
has value = 100V 2cos(27t x 10^/) V. Let R^„j^ii denote the smaller o f (/?p /?-,) and
Compute the values o f L and C such that the the larger. Prove that
average power delivered to the load resistance (a) Cshould be connected in parallel with
is maximized. What is ^large' ^ should be connected
between the rsvo top terminals and
thus in series with R^„j^n-
(b) The element values are given by
large
—Rsmall
and
Figure P I6.60
A N SW l-RS: 17.79 uH, 1.186 pF, 8.3.^33 W C=
(oRla/ge Rsmall
8:^6 Chapter 16 • Band-Pass Circuits and Resonance
BAND-PASS TRANSFER
FUNCTIONS WITH NO ZEROS
Figure P i6.62 OR A SINGLE ZERO OFF THE
ORIGIN
63. The purpose o f this problem is to show that 65. Consider the circuit o f Figure P i 6.65.
the resonant frequenc)' (0^ depends on the (a) With V^{s) as the input and the
choice o f the input terminals and that (0^^ (Oq output, compute the transfer function
in general. Consider the circuit o f Figure His).
P 1 6 .6 3 , which has Wq = l/ V LC = 0.25 (b) I f / ? = 5 a , Z = 0.1 H ,an d C = 10|iF,
rad/sec. Find the resonant frequency (0^ if the verify that the circuit is high Q^.
input is connected across (c) Compute exactly and verify that
(a) A and B 0)^^ = CO^. Then compute the maxi
(b) B and C mum gain.
(c) A and C (d) With the values given in part (b),
compute approximate values for
40 n 10Q
CO,, and CO2.
'0 .8 H
2m F
Figure P I6.63
Figure P I6.65
ANSWURS: 2S, I2.5\/3 = 2 1 . 2 8 . 8 6 . all in
rad/sfc
66. Consider the circuit o f Figure P i 6.66.
(a) With as the input and I^{s) the
64. The circuit of Figure Pi 6.64 has
output, compute the transfer function
= l/ 'Tl C = 25 rad/sec. The purpose of this
His).
problem is to show that the resonant frequency (0,.
(b) I f / ? = 4 k a Z . = 0.1 H ,an d C = 10).lF,
depends on the choice ot the input terminals and
verify that the circuit is high
that (0^ 03^ in general. Find the resonant frc-
(c) Compute (0^^^ exactly and verify that
quenc)' 03^ and the equi\'alcnt impedance seen at
= CO^. Then compute the maxi
the terminals if the source is connected across
mum gain, H^^j.
(a) E and F
(d) With the values given in part (b),
(b) D and E
compute approximate values for
100
CO,, and 0) 2-
. 0.8 H
2m F 50 0 /
f— S +■
____1 UiC l ] LC
/?l/?2C|C2
H{s) = — _
1
+■
/?2Q/
1
s+ 1 (b) Recall that cOq = lly J h C . By the use of
equation 16.49, show that
o—
+
Figure P I6.69
Figure P I6.67
H{s) = K - 2
s+2OpS+(0p ------j -
71. Consider the circuit shown in Figure 73. The y?/.C circuit o f Figure PI 6.71 has trans-
P 16.71. fer function \/
(a) Show that the transfer function is s+—
L)
H {s) = Zi„{s) = ------- ^ ^
' R.
s+ + + ‘''S s + 11+
H {s) = Z;As) =
L} [X pC L \ LC
D \ 1
s + l+^.s_ Let /?. = 100 k^2, L = 0.225 H, R, = 1000 Q,
Rf,C L RPI LC and = 0.5 [.iF.
(a) Compute to and verify that the cir
(b) By the use o f equation 16.54 or part
cuit is low Q^.
(a) o f Problem 65, show that
(b) Compute exactly and verify that
is quite different from
1 1+- (c) Plot the frequency response for 0 < to
0).,, =
R r
< 40 0 0 rad/sec. Compute the maxi
mum gain and verify the result from
the plot. Also compute the zero o f the
transfer function as -Zy
(d) Define the coefficients o f the numera
tor and denominator o f your transfer
function in MATLAB as n = [1 zl]/C
and d = [1 ? ?]. Now use the com
mands below to compute the impulse
Coil
and step responses o f your transfer
Figure P I6.71
function.
1.0l28e+ 03
75. Consider the circuit in Figure P i 6.75. Let element values and repeat parts (c)
Z. = 1 H, C = 1 F, = 0.08 Q, R c = 0.02 a and (d) for the scaled circuit.
and Rs = 40 ^2.
R, = 0.05 n
(a) Find the transfer function H{s) =
o u r ''" ' nt' R, = 50 O
(b) Find and approximate answers for /Y Y V
L=1 H
" P “ ‘1 “ 2-
(c) O btain a magnitude frequency
response plot to graphically verify
,0 R, = 0.05 n
C =1 F
your answers.
Figure P I6.76
R
MISCELLANEOUS
77. Consider the idealized (tank) circuit o f
'■ 6 Figure P I 6,77. The moment the inductor cur
rent passes through zero with positive slope is
taken as the reference point, r = 0. At this time
instant the capacitance voltage is E volts.
Figure P i6.75 (a) Find V(\t) and i^{t) for / > 0 by the
Laplace transform method.
76. The analysis o f the non-series-paraliel cir (b) Find the energy stored in C as a func
cuit shown in Figure P i 6.76 requires writing tion o f t.
node or loop equations. Because there are no (c ) Find the energy stored in L as a func
series-parallel connections, one cannot apply tion o f t.
the equivalents o f Table 16.2. (d) Show that the total energy stored in
(a) Let r , = 1//?,, ^2 = i//?2, = i//?3, the LC tank is constant and is equal to
= \/L^s, = C 5J. Using nodal 0.5C £2.
analysis, show that the transfer func
tion is V
His) = -
hn
(a) Determine the range of the capaci 80. Consider the circuit o f Figure P I6 .80, in
tance C, i.e., Cq < C < C j, such that which R^= \ 0 n , L= 1 H, /e/= 0.8 n , and C
the circuit can be tuned to resonance = 1 F.
over the AM radio band (from 550 (a) Show that the transfer function H{s) is
kHz to 1650 kHz). that given in Problem 71.
(b) When the circuit is tuned to 550 kHz (b) Find the exact values o f to^, coy,,
and 1650 kHz, determine the circuit CO,, and co^^^. Since this is a low-Q cir
Q's, the two bandwidths, and the cuit, do not use the high-Q approxi
lower and upper halF-power frequen mations.
cies for each bandwidth.
sin (o )/ + 0)
where
to.
a = •all switches open when no button
and 2(2 is pressed
•pressing a button in column k
1 closes the switch SCK
1- •when any button is pressed halfway,
4(2^ S is closed to contact H; when the
Show that the peak amplitude o f the button Is fully pressed, S moves away
(b)
from H and makes contact with F
damped sinusoidal waveform of part
(a) decreases to Me = 0.368 of the
Figure P i6.84
highest peak approximately after Q/ji
cycles if Q is large (and, hence, (O^ es
'w ' 85. For the circuit o f Figure P I6.85a, we have
co^).
the impedance
r\
YijiO) =
■>G.
iBp
JX. r>
•-------- J
(a) (b) o
jCsO)
Most electronic equipment operates with dc power sources. For portable equipment, such as a
cordless phone and a cordless electric drill, batteries supply the dc power. Using non-rechargeable
batteries becomes expensive. Furthermore, replacing batteries in special equipment is a task not
easily handled by ordinary consumers. These two factors have prompted manufacturers to install
rechargeable lithium-ion batteries in portable equipment. By connecting several batteries in series,
the available dc voltage may range from 1.5 to 12 V. Whenever the battery runs low, it must be
recharged.
Recharging a battery requires a low dc voltage source (1.5—12V). An adaptor houses a device called
a transformer that changes the 110 V ac voltage at the household outlet to a much lower ac volt
age. The lower ac voltage is then rectified to become a dc voltage that charges the battery. Some
adaptors contain the transformer only, while others may also contain the rectifier circuit.
The concepts and methods developed in this chapter will allow us to understand how a trans
former works to change the ac voltage level and also to perform some other important functions
in electronic equipment.
8iS4 Chapter 17 • Magnetically Coupled Circuits and Transformers
CHAPTER OBJECTIVES
SECTION HEADINGS
1. Introduction
2. Mutual Inductance and the D ot Convention
3. Differential Equation, Laplace Transform, and Phasor Models o f Coupled Inductors
4. Analysis o f Coupled Circuits with Open-Circuited Secondary
5. Analysis o f Coupled Circuits with Terminated Secondary
6. Coefficient o f Coupling and Energy Calculations
7. Ideal Transformers
8. Models for Practical Transformers
9. Coupled Inductors Modeled with an Ideal Transformer
10. Summary
11. Terms and Concepts
12. Problems
1. INTRODUCTION
You may recall from your high school or grade school science class that if iron filings are sprinkled
on a piece o|- paper and a magnet is moved around beneath the paper, the iron filings move in con
cert with the magnet because the magnetic field induces a force on the iron filings. Similar to the
magnet and the iron filings, a changing current in one coil that is ver)' close to another coil induces
a voltage across the terminals o f the other coil.
Figure 17.1a shows rwo unconnected coils o f wire in close proximity. Figure 17.1b shows two
unconnected wire coils wound around a single ferromagnetic core, hi both cases, a voltage source
excites coil 1 while coil 2 is left open-circuited. Experimental evidence shows that a change in the
current z, generates a voltage Vj, called the induced voltage, across the open circuit; the induced
Chapter 17 • Magnetically Coupled Circuits and Transformers
voltage is proportional to the rate o f change of/ j. Each pair o f coils in Figures 17.1a and 17.1b is
said to be magnetically coupled.
Coll 1
-• +
V ,( t )
Coil 2
(a) (b )
FIGURE 17.1 Induced voltage in coupled coils, (a) Two coils in close proximity, (b) Two coils
wound on the same ferromagnetic core.
How does one quantitatively account for magnetic coupling? The strategv' is to introduce a new
circuit quantity called mutual inductance for coupled coils; specifically, similar to the v-i rela
tionship o f a single coil, the induced voltage satisfies the equation
cli^
V-7 = ± M 21 (17.1)
dt
where > 0 is the proportionality constant callcd the mutual inductance from coil 1 to coil
2, and the sign, here ±, depends on the relative winding directions o f the coils. Dot markings indi
cate the relative winding directions. With reference to Figure 17.1, a dot is placed on coil 1 for
reference; if the dot on coil 2 is in position A, the sign on equation 17.1 is +, and if the dot is in
position B, the sign is - . A description of the general dot convention is presented in the next sec
tion.
Experimental evidence demonstrates that if the two coils in Figure 17.1 are stationary, the induced
voltage, is proportional to the rate o f change o f i . e . , the induced voltage
du{t)
\>2{t) = ±M2\
dt
as set forth in equation 17.1. Note, however, that coil 1 with inductance Z,, continues to act as an
inductor for which
di\{t)
v,(/) = Li
dt
There are two effects present in the circuits o f Figure 17.1: an induced effect and the usual v-i
relationship o f an inductor. Similarly, Figure 17.2 shows the reverse coupling to that o f Figure
17.1. For the circuit o f Figure 17.2, with the reference dot placed at the top o f coil 2,
d ilit)
v, ( 0 = ± M ,2 (17.2)
dt
where + would be used if the dot on coil 1 were in position A and - if in position B. As in figure
17.1, the dots indicate the relative directions o f the windings o f the two coils. Also as before, the
second coil continues to act as an inductor, for which
d ilit)
viit) —Lo
dt
Secondary
Coil 2
v,(t)
(l )
FIGURE 17.2. Coupling from coil 2 to coil 1 (winding directions not explicidy shown).
As will be verified in a later section, = ^ \ 2 > Hence we designate the positive constant
(17.3)
Figure 17.3 shows a composite o f Figures 17.1a and 17.2 where currents are present in both coil
1 and coil 2. Four effects are now present in the circuit: two induced effects and the tw'O usual self
inductance effects. Linearly superimposing the effects (superposition), we obtain the equations o f
the mutually coupled inductors:
Chapter 17 • Magnetically Coupled Circuits and Transformers 887
(17.4a)
(It dt
dt ~ dt (17.4b)
Two questions remain: (i) When is the sign positive and when is the sign negative? (ii) How is the
value o f M determined (experimentally)?
Primary Secondary
+ A
i,(t) 0 v,(t) v,(t)
B •
FIGURE 17.3. Coupled coils with current excitations present on primar}' (coil 1) and secondar}' (coil 2).
The following rule, identified with equation numbers, governs the choice o f sign lor the induced
voltage.
The voltage drop across one coil, from the dotted terminal to the undotted
terminal, equals M times the derivative o f the current through (17.5a)
the other coil, from the dotted terminal to the undotted terminal.
Or, equivalently,
The voltage drop across one coil, from the undotted terminal to
the dotted terminal, equals M times the derivative o f the current (17.5b)
through the other coil, from the undotted terminal to the dotted terminal.
With reference to Figure 17.3, if the dot is in position A, all signs are positive, whereas if the dot
is in position B, the sign on M is negative. This rule gives the voltage drop due to the mutual
inductance. To obtain the total voltage drop across an inductor that is coupled to another, one
must add in the voltage drop induced by the self-inductance o f the individual coil, which depends
on whether the labeling is consistent with the passive sign convention.
8 S8 Chapter 17 • Magnetically Coupled Circuits and Transformers
EXA M PLE 17.1. For the configurations o f Figure 17.4, determine the pair of equations that spec
ify the relationship betAveen the voltages and currents.
M M
1 2
+ +
•-
S o l u t io n
First we consider Figure 17.4a. The voltage z^,(r) and the current /, (r) as well as the voltage V2 U)
and the current ijit) satisfy the passive convention. For each coil acting alone,
cji^
(It
for k = 1 ,2 . However, the voltage induced in coil 1 by the current in coil 2 is negative relative to
the indicated polarity on v^{t) as per rule 17.5b, i.e., /-,(r) enters the dotted terminal so that-/'2W
can be viewed as entering the undotted terminal. Fience
‘ ‘ dt dt
On the other hand, the voltage induced in coil 2 by iAt) satisfies rule 17.5b. Hence
Chapter 17 • Magnetically Coupled Circuits and Transformers 8S9
Exercise. For the configurations o f Figure 17.5, determine the pair o f equations that specify the
relationship between the voltages and currents.
M M
+
V,
(a)
EX A M P L E 17.2. This example presents the procedures for marking the dots on an unmarked
pair o f coupled inductors and for determining the value o f M. Consider the configuration o f
Figure 17.6, in which a current source is exciting terminal A ot coil 1 o f the coupled inductors
with unknown M.
DVM
Also marked
com, neg, low
increasing current
Step 2. Apply an increasing current /j(r), i.e., a current for which > q fo r all t.
clt
For example, one could set /,(/) = 10/«(/) mA for 10 seconds. We know that /,(/) induces a volt
age at the terminals o f coil 2 according to
(It
where M > 0. If we put the leads o f a voltmeter across the terminals C -D as suggested in Figure
17.6, the reading will either be Vjit) > 0 or v^{t) < 0. Suppose the reading is V2 {t) < 0. If we reverse
the leads o f the voltmeter by putting them across the terminals D -C , the reading will have the
opposite sign, i.e., v-,{t) > 0.
Step 3. Reconnect the voltmeter leads until V2 {t) > 0, i.e., the reading is positive. Place a dot on the
terminal o f coil 2 for which the voltmeter lead is marked + {or "plus" or “high” or “pos'), i.e., at the
terminal o f higher potential.
For the situation described above, the dot would be placed at terminal D. However, the dots on
coils 1 and 2 could be simultaneously moved to the opposite terminals o f each coil without chang
ing the relative information they convey. (Problem 1 confirms this statement.)
Part 2: Determining M
v-){t) = ±M
Again, \'2(0 where M > 0. Taking absolute values o f both sides o f this equation
implies that
(If
^2(0
M=
clidt)
dt
Exercise. In figure 17.6, /j(/) = 2tu{t) A and a voltage = \Ou{t) mV is measured. Determine
the placement o f the dots and the value o f M. j
AN SW ERS: Dots are at terminals A and or at B and D: M = ------ = 0.005 H.
Compared with the ramp o f Example 17.2, a more practical input signal and measurement
scheme uses a triangular waveform, as displayed in Figure 17.7.
Chapter 17 * Magnetically Coupled Circuits and Transformers 891
E X A M PLE 17.3. Figure 17.7 shows a circuit and two waveforms, /,(r) and V2 {t), as might be dis
played on an oscilloscope. Determine the placement o f a dot either at position C or at position
D, and the value o f the mutual inductance.
i,(A)
So lu t io n
For the time interval 0 < r < 0.5 msec, /,(/) is a ramp function and Vjit) is constant. The infor
mation is similar to that given in Example 17.2 and the solution method is the same. First we place
the dot at terminal A as the current enters A. The current /, is increasing, and ^2 is positive. We
must now determine if the dot goes at terminal C or D. Since the current i^{t) enters the dotted
terminal and is increasing over 0 < r < 0.5 msec, its derivative is positive over 0 < r < 0.5 msec.
Also, the voltage iQ^t) is positive for 0 < r < 0.5 msec with the indicated polarities. Hence accord
ing to rule 17.5a, the dot goes at terminal C and
di^
dt
To determine M, consider that the measured values during 0 < r < 0.5 msec give Vj = and
di^ldt= 1/0.0005 = 2000 A/sec. Thus, M = 2/2000 = 0.001 H.
EXA M PLE 17.4. In the circuit o f Figure 17.7, suppose the dot positions are at A and C. If /jW
= 2(1 A, find Vjit).
So lu t io n
v’9(/) = M — = 0 .2 e “ ’^'//(/) V
dt
Chapter 17 • Magnetically Coupled Circuits and Transformers
Exercise, In the circuit o f Figure 17.7 with the dots in positions A and C, if /j(t) = 0.01
sin(1000?)«(^) A, find v~,{i) tor r > 0.
ANSW ER: = 0.0 Tcosi 1000.^)/,u) A
The preceding treatment o f the mutual inductance has not referred to the physical construction
o f the coils, although we have described procedures for measuring M when the coupled inductors
are assumed to be enclosed in a sealed box. However, for designing a pair ot coupled inductors, or
for a better understanding o f mutual inductance, one must relate the coil construction to the val
ues o f Z.J, Ljy and M. A rigorous study o f this problem requires a background in field theory and
magnetic circuits, which are covered in advanced texts or physics courses. Nevertheless, we set
forth here a few basic properties with reference to Figure 17.8.
Coil 1
Coil 2
(b)
FIGURE 17.8. (a) Coupled coils in close proximity, (b) Coils couplcd through magnetic core.
1. In Figures 17.8a and b, the number o f turns for each o f coils 1 and 2, respectively, is TVj
and Nj. Then the self- and mutual inductances have approximately the ratio
L, :L 2 :M = n } \NI\N^N2
2. If two coils/inductors are placed in a nonmagnetic medium (e.g., air), bringing the
inductors closer together increases the value o f M.
3. If one inductor o f a pair is rotated, then a larger value o f M results when the axes o f the
inductors are parallel to each other. The smallest value o f M occurs when the axes are per
pendicular to each other.
4. Changing the core on which the two inductors are wound from a nonmagnetic material
(e.g., air, plastic) to a ferromagnetic material may increase the values o f Z,j, L-,, and A/by
a factor o f several thousand.
Chapter 17 • Magnetically Coupled Circuits and Transformers
As a final note, the development above presupposes linearity. If the rvvo inductors are placed in a
nonmagnetic medium, this holds true. If the inductors are coupled through a ferromagnetic medi
um (e.g., an iron core), then the linear relationships o f equation 17.4 hold only if both currents
are sufficiently small that the magnetic medium avoids saturation, a phenomenon discussed in
other courses or more advanced texts. Our investigations consider only the linear case.
Figure 17.9 shows a pair o f coupled inductors. As developed in section 2, since each coil contin
ues to act as an inductor, the terminal voltage depends on the derivative o f the current through
the coil plus an additional induced voltage due to the changing current in the other coil. This led
to the set o f differential equations 17.4, repeated below as 17.6, where the plus sign is used when
the secondary dot is in position A and the minus sign is used when the secondary dot is in posi
tion B.
(17.6a)
dt dt
dh dh
V2(t) = ± M —^ + L 2 ^ (17.6b)
dt dt
Assuming that there is no internal stored energ}^ in the coupled inductors at time r = 0, then the
Laplace transform o f equations 17.6 yields
Equations 17.7 represent the x-domain model o f the coupled inductors. Further, if one is con
cerned w'ith the sinusoidal steady state, then replacing s by jui yields the following equations for
analysis using the phasor method:
894 Chapter 17 • Magnetically Coupled Circuits and Transformers
V 2 = ± y to M j + (17.8b)
For equations 17.7 and 17.8, the plus sign is used when the secondar)' dot is in position A and
the negative sign otherwise. These three sets o f equations constitute the core o f the examples and
analyses presented in the remainder o f the chapter. Throughout, the j-domain method is the pre
ferred method, but each pair o f equations has its specific uses.
EXA M PLE 17 . 5 . A pair o f coupled inductors are connected in two different ways, as shown in
Figure 17.10. Find the input impedances, and -^,„2(^)> and the corresponding equivalent
inductances, and L^^2 '
(b)
FIGURK 17.10 Impcdanccs and equivalent inductances of two series-connected inductors.
(a) + Z2 + 2 ^ - (b) + >^2 - 2 ^ -
S o lution
For Figure 1 7 .10a, label the inductor voltages j and with positive reference on the left side
o f each inductor. Observe that enters the dotted terminals o f both coils. Hence = I^{s)
= I-jis). Direcdy applying equations 17.7, we obtain
^in (-^) = (-^) + ^L2 (^) = (^ 1 + Ms) Ii,j (^) + {Ms + ) I i n (^)
= {L ^ + L 2 + 2 M )sIi„ {s)
It follows that
im plying that
Chapter 17 • Magnetically Coupled Circuits and Transformers 89S
^eq\ - + ^2
Similarly, for Figure 17.10b, with the same voltage definitions, we observe that again = I^{s)
= Ijis). Also note that /2W enters the undotted terminal. Applying equations 17.7, we obtain
It follows that
Z,»2(^>= = [L\ + h - 2 « ) s =
ImU)
implying that
Exercise. Suppose the circuits o f Figures 17.10a and b are connected in parallel. If Z,, = 40 mH,
Z,2 = 60 mH, and A/ = 25 mH, find the equivalent inductance o f the parallel connection.
AN SW ER: . r . 5 ml I
= 0 7 -9 )
This relationship suggests another way o f determining M and dot markings from measurements.
If an instrument for measuring self-inductance is available, we can use the instrument to measure
Z.^^j and L^^2} from equation 17.9, the difference is AM.
In the next example we compute the voltage across the secondary o f a coupled inductive circuit
using a high-resistance voltmeter.
EXA M PLE 17.6. In the circuit o f Figure 17.11, assume that the meter resistance is ver)' large
and looks like an open circuit to the secondary. The switch S is closed at r = 0. Find ij{t) and Vy{t).
8% Chapter 17 • Magnetically Coupled Circuits and Transformers
FIGURU 1 7 .11 A coLiplcd induccor circuit that might be lotind in a lab to determine dot position.
So lu t io n
Since the secondary looks like an open circuit, /-> = 0 and K, = Msly Hence if we find /j we can
easily find K,.
Applying KYL ro the primary loop, using equation 17.7a and the fact that /, = 0 produces
.V sms+R) s +
Therefore
R
/l(0 = \ -e
R
Since
£q _ MZiq _ iW/To/Li
V/^ = yW.s7, = Ms
s { L^s +R) ~ m s + R ) ~
It follows that
= H(/)
h
Note that u-y(t) > 0 by virtue o f the way the meter leads were connected to the secondary. Hence
we conclude that the dot is indeed on the upper terminal of the secondary.
Chapter 17 • Magnetically Coupled Circuits and IVansformers 89'
Exeixise. Suppose in the circuit o f Example 17.6 = 10 V, /^= 2 Q, the time constant o f the circuit
is 1 msec, and the voltage at = 1 msec is 3679 V. Find L^, M , and the ratio M : L^io see the
factor by which the small, safe voltage is amplified when the switch is closed. The point here is that
when switching with inductors occurs, a small voltage ma)' produce a dangerously iiigh voltage.
AN SW ERS: /., = 2 mi l. M = 2 H, and M : A, = 1000
An interesting application to older car ignition systems also uses an open-circuited secondary to
produce a very high voltage from a small one to fire a spark plug.
EXA M PLE 17.7. Figure 17.12a shows an automobile ignition system found on older cars while
figure 17.12b shows a simplified equivalent circuit model. Todays ignition systems use electronic
switching. Specifically, the block with the condensor (capacitor) and ignition point is replaced by
something referred to as an “ignition module.” The module contains a power transistor circuit to
perform the switching action electronically, with the trigger timing typically actuated by a sensor
that measures the position of the cam shaft cither optically or magnetically.
CONDENSOR
SPARK
PLUG
fL AL
IGNITION f i
SWITCH IGNITION IGNIT10N(f o)
POINT COIL
+ 12V
---- n n---- DISTRIBUTOR
H I
t J
BATTERY
(a)
point
to spark plugs
— through a
I mF + distributor
primary secondary
(b)
In Figure 17.12a, the ignition coil is a pair o f inductors wound on the same iron core, which cre
ates a strong coupling between the coils. The prim ary coil is connected to the battery, while the
secondary coil is connected to the spark plugs or load. The primary has a few hundred turns o f
heavy wire, the secondary about 20,000 turns o f ver)' fine wire. When the ignition point (or con
tact) opens by cam action, a voltage exceeding 20 ,0 0 0 V is induced across the secondary, causing
the spark plug to fire. The generation o f a high voltage to cause the spark plug to fire is accom
plished by a basic RLC circuit containing a switch that represents the point o f the ignition system.
Since the secondary is open-circuited, it has no effect on the solution for the primary current. Let
us do the analysis using the equivalent circuit model o f Figure 17.12b. Suppose that the switch
has been closed for a long time. Accordingly, at r = 0~, we have /j = ElR= 12 A. Using the model
for an initialized inductor given in Figure 14.18 results in the ^-domain equivalent circuit o f fig
ure 17.13.
The primary current is simply the net driving voltage divided by the total impedance in the series
circuit o f Figure 17.13, i.e..
R
— t- L i — ------
(17.10)
R
R+ + sL .V“ + .9 +-
sC ‘ Li L ,C
Suppose now that E = 12 V, /? = 2 Q, Z.j = 2 mH, M = 0.5 H, L~, = 130 H, and C = 5 l^iF.
Substituting the given component values into equation 17.10 yields
6 (5 + 1 0 0 0 ) 6 (.V + 5 0 0 ) + 0 . 0 0 5 X 9 , 9 8 7 . 5
(17.11)
.v“ + 10()0i + 10^ (.v + 5 0 0 )- + ( 9 ,9 8 7 .5 ^
s cos(lO.OOOr) (1 7 .1 2 b )
Chapter 17 • Magnetically Coupled Circuits and Transformers 899
Having obtained we calculate V2 {t) from the basic relationship o f equation 17.6, using the
plus sign for the dot positions:
di^
V2(/) = M
(It
From this expression, the voltage reaches a magnitude o f 3 0 ,0 0 0 V in about 157 psec (one-
fourth o f a cycle o f the oscillations). This voltage is high enough to cause the spark plug to fire.
After the spark plug fires, the secondary is no longer an open circuit; the above methodology ceas
es to hold during the firing o f the spark plug.
Exercise. Compute the energy stored in just prior to the switch across the capacitor opening.
AN SW ER. 0.036 1
In going from equation 17.12a to 17.12b, we neglected the second term in 17.12a, retaining only
the first. W ith the practical component values used in ignition circuits, this approximation is usu
ally valid. In terms o f equation 17.11, the approximation is as follows:
R \ R R
S H------- I H---------------- 5 +
I 2 lJ 2L , E (17.13)
“ — X
R 2 R 1 R 1 R 1
S + ----- 5 + --------- s~ + — s + ------
L] L,C L, L]C
where
R \ I r V I
and ^d =
z .,c U tJ ” ^/z;c
For the first few cycles o f oscillations, the value o f is nearly zero and the value o f the expo
nential factor is nearly 1. Therefore, /j(/) may be further simplified to
')()() Chapter 17 • Magnetically Coupled Circuits and Transformers
£ ( r \
/|(/) s — COS for small t > 0
R
-1 . i t \
\'o{t) = M — = M — ^sm Hit)
dt R ■ Ju c
E 1 M
V-7(/) a A/----- f = = — QE
- Z.| (17.14)
Q=
R
is the quality factor o f the series RL\ C circuit. (See Chapter 16.) Later we will show that M!L\ is
approximately equal to the ratio o f the number o f turns o f the secondary to that o f the primary.
Equation 17.14 is a simple formula for estimating the maximum voltage that occurs at the sec
ondary. It shows that although the battery voltage is only 12 V, what appears at the secondary for
a brief moment is very much higher due to the switching action. The voltage is stepped up due to
two factors: (1) the Q o f the series RL^C circuit and (2) the turns ratio o f the ignition coil. From
equation 17.14, a smaller R produces a higher voltage across the secondar)\ But a small R causes
a larger current to flow in the primary circuit and therefore shortens the life o f the breaker point.
In practice, when the engine is running, a resistance wire or an actual resistor is placed in the pri
mary circuit to limit the amount o f current flow through the breaker point. The capacitor (con
denser) serves a similar purpose— that o f protecting the breaker point by suppressing the arc that
results when the point opens.
Exercise. The ignition circuit o f Figure 17.12b has £ = 12 V, Z,j = 1 mH, and C = 0.01 [J.F. If the
total resistance in the primary circuit is /? = 8 Q and M/Zj = 50, estimate the maximum voltage
appearing at the open-circuited secondary when the s\vitch opens.
AN SW ER; Alniui l .s r k\'
Chapter 17 • Magnetically Coupled Circuits and Iransform ers ')()!
To begin this section, we would like to develop some general formulas that make the analysis of
doubly terminated coupled inductors straightforward. By “doubly terminated” wc mean the con
figuration o f Figure 17.14 where there is a source and source impedance connected to a load
impcdance through a pair o f coupled inductors.
EX A M PLE 17.8.
Consider the circuit o f Figure 17.14. Show that
M~s~
(i) Z i„ {s)= L ^ s-
L'fS + Z-)
Z.v ZoM
(v) the overall voltage gain Gy = — =
An + 2 ] (^1^2 - M -).v + L,Z2
So lu t io n
'1 V /
Vi(^)
Part 1: Find Z,„(5) = ------- by two loop equations in the y-domain, converting to matrix form,
and solving.
In matrix notation,
Step 2. Solve for /j and /j. Using the well-known formula for the inverse o f a 2 * 2 matrix, we
have
-1 p
■ /r L,.v Mv
/2 A/s + Z-)
Hence
L 2 S + Z2
^1 ------------------------------- T T M
LiS{L2S + Z 2 )-M \ s-
V'l , W^.v-
/| L 2 S + Z2 (17.15a)
-M
(17.16)
L\(L2S Z 2 ) - M~s
Part 3: Find C ^.2 = — •Using expression 17.16 and the fact that Vi = - •^2-^2’ conclude that
“ Vi
^ V2 _ -Z y l. Z-yM Z^M
^v2 - 7 7 - (17.17)
Ki V, m L2S + Z 2 ) - M - s ( l , L 2 - M “ ).v + L,Z2
V'l
Part 4: Find Gj.i = — .B y voltage division
Z;., Z,„
z,-„ + z , (17.18)
-M v
Note that the answer can be obtained directly be writing a loop equation for the secondar\'.
The next example applies the foregoing development to a specific case and at the same time
expands the developed formulas to new cases.
So lu t io n
First we note that co = 10 rad/sec. AJso, using the effective value for the phasor voltage o f the source
yields W = 60 Z. 0 " V. Thus all calculated voltage and current phasors are effective values.
10
Z2(yiO ) = — 7 = 5 - 5 ./
CylO +
}_ O.ly + 0.1 1+ 7
R,
From equation 17.15 and the fact that M = 0.5 H, we have
= i ,S l o T 4 . = 5 4 ^ jl2 ^ 2 1 .8 7 Z 1 2 .5 .r
ZyM 0 .5 ( 5 - 5 ./ )
V 2 = G .oV , = V, = - ( 5 4 + j\ 2)= 1 8 - 6 7 = 1 8 .9 7 / 1 -1 8 .4 4 ''
- M ‘ )./10+L|Z2 7.5y + 10 - IO7
-M -0.5
u = V, = ------- ^----- (54 + y l2 ) = - 2 . 4 - 1 . 2 7 = 2 .6 8 Z .- 1 5 3 .4 "
(L ,L 2 - M - ) 7 lO + L ,Z 2 10
1 0 -— 92.57
S ;■
^1 -"54+12./
I. = = 1 . 2 - 2.4./ = 2.683Z. - 63.44'^
Z,„(ylO ) 5 + y20
Step 6 . Fitid the average power delivered to the load. The complex power delivered to the load is
Sometimes the application o f the formulas developed in the previous rsvo examples, although
straightforward, is not the simplest route to the answer. The following example is a case in point.
EXA M PLE 17.10. Find the steady-state components o f z/j(r) and V2 {t) at the frequency 1 rad/sec
for the circuit o f Figure 17.16, in which ^^^^(t) = cos{t)u{t) V. Note that because a resistance is pres
ent, the circuit responses will contain both a transient and a steady-state component.
1Q M = 2H
------ O
+
V
V, 2F
IH 5H -
-o
So lu t io n
Since only the sceady-stare responses are required, we use rlie pliasor method and write two loop
equations given that co =1 rad/sec and = 1 Z. 0^^ V.
and
It follows that
J4.5 0. 5 1
v,=v,,-i, =i- = 0 .1 1 0 4 3 Z .8 3 .6 6 "
- 0 .5 + ./4.5 0.5 - ./4.5 -./•9
-j2 1 _9
v, = - - L i , = _ _ x = 0.22086Z . - 96.34"'
jl “ j2 - 0 .5 + 74.5 - 0 .5 + J4.5 1 - ./9
Observe that
V , = -2 V ,
in which case
= 0 .2 2 0 8 6 c o s (r - 9 6 .3 4 °) V and = 0 .1 1 0 4 3 cos(; + 8 3 .6 6 °) V
Exercise. In Example 17.7, write the simultaneous equations in matrix form and then solve by
Cramer’s rule. Which method is easier?
Example 17.10 is a drastic case contrived to bring up an underlying propert)': the dot markings
for coupled inductors (/.,, A,, M) determine the ± sign in the equation relating v-^ to eli^ldt and
the equation relating to di-)ldt. No intrinsic relations ber^veen the polarities o f and v-, are
conveyed by the dot convention. For most practical circuits, hoAvever, it is true that the voltage
drops o f coupled inductors from the dotted terminal to the undotted terminal are in phase or
nearly in phase.
906 Chapter 17 • Magnetically Coupled Circuits and Transformers
As a practical example ro end this section, we analyze the circuitry typical o f the front end o f an
AM radio receiver.
EXA M PLE 17 .1 1 . Figure 17.17a shows circuitry typical o f the front end o f an AM radio. Figure
17.17b shows a simplified RLCM model o f this circuitr}'. Here, and the resistance = 300 ^2
together represent the antenna. Typical parameter values might be /?2 = Rj„ = 14,7 k fl, Z,j = 50
nH, L-, = 2450 nH, M = 350 nH, and C = 104.5 pF. Find the transfer function
H {s) =
antenna
RF
Amplifier
M
•
coupling ■ tuning
SR
X in
'■ 6 C
coils 1capacitor
(a) (b)
R = 300 Q
R = 14.7 kO
L c / 2
'■ 6
= 98nH
= 2612 pF
(c)
FIG U RE 17.17 Circuits for Example 17.11. (a) The original circuit, (b) A simplified circuit model
in which L^L-y = M~. (c) A design without couplcd inductors.
So lu t io n
The first step in the solution is to observe that L^L-, = Afi. Then from equation 17.19, the trans
fer function is
Z-yM M
V ,{s) Z ,„ + Z , (z,,L 2 - M “ ) j + L , Z 2 ^m + ^l ^1 (1 7 .2 0 )
where Z , = /?, and Z^y = R-y I !C = ----- — . Also, from equation 17.15,
s+
R-yC
Chapter 17 • Magnetically Coupled Circuits and Transformers 90
2 ,2
V Loi + Z j L2 S + Z 2
Substituting the expression for Z-^^{s) in equation 17.21 into equation 17.20, then dividing
through by Z j and simplifying, we obtain
M
s
CR^Lo _ 4 .5568x10^ 9
H (s) =
_y2 ^ / _ _ _ _ h ^ _ L _ \ ^ ^ _ _ L 5^ + 1 .3 0 2 0 X 10^^ + f 6 .2497 x 1 0 ^ ( 1 7 . 2 2 )
\CRiL2 R2 C ) L2 C '
Ks Ks
S + B^^.S + fO,„ ^2 ^ ^^ ^ 2
Q
By inspection, then,
co„,
TU
= 6.2497 X 10"^ rad/s, W
= 13.02 x 10^ rad/sec, =—
P
= 3.5 ’, 0
XI.
= 4 8 ,'
and w
Instead o f the coupled inductors o f Figures 17.17a and b, the same and can be obtained
with the parallel resonant circuit o f Figure 17.17c. Following the design method described in
Example 16.3, we find the required element values to be Z, = 98 nH and C = 2612 pF. However,
the maximum voltage gain would have a much lower value, l^(;w,„)| = 14,700/(14,700 + 300) =
0.98, compared with 3.5 for the coupled circuit o f Figure 17.17a.
The higher voltage gain achieved in Figure 17.17b can be explained by the concept o f nuiximum
power transfer. A routine analysis would reveal that at to = to^^, the input impedance “seen” by the
source is a pure resistance equal to the source resistance R^, i.e..
Z ,v ,0 ,„ ) = = 300 Q
Hence the sources impedance o f /?j sees a load equal to itself. Thus, maximum power has been
extracted from the source. Since the inductors and the capacitor together form a lossless coupling
net\vork (see the homework problems), the same maximum power is transferred to the load resis
tor Rj. In Figure 17.17c, the reflected load impedance at/o^^^ is not 300 and hence maximum
power is not transferred.
Figure 17.17b contains two inductors and one capacitor. Accordingly, one would normally con
sider it a third-order circuit. Yet the denominator o f equation 17.22 is only a second-degree poly
nomial in s. This simplification is due to the condition M~ = The significance o f this con
dition will be discussed in the next section.
The general analysis given above is quite useful for complex circuits. However, for many applica
tions the rwo coils have a common terminal, as shown in Figure 17.18a. Figures 17.18b and c show
908 Chapter 17 • Magnetically Coupled Circuits and Transformers
rhe T- and Tt-equivalent circuits char often allow the application o f series-parallel techniques to
simplify the analysis. With the common terminal the coupled inductors have only three accessible
terminals instead o f four. Such an arrangement is called a three-terminal device. The reader might
observe that if A/> Z, or M > L~,y one o f the three inductances in the T- or Tt-equivalent circuit may
have a negative value. This negative inductance appears in a mathematical model and is not the
inductance o f a physical component. Also, the equivalent circuits shown in Figures 17.18b and c
are for the specific dot locations indicated in Figure 1 7 .18a. A change o f one dot location in Figure
17.18a will result in a change in the sign in front of M in Figures 17.18b and c.
2
L ,- M
L^-M L ,- M
A = L,L^-M2
Verification o f the parameter values in the equivalent circuits in terms o l l j , Z.,, and M is left to the
homework problems. We now illustrate their use with a simple equivalent inductance example.
EXA M PLE 17.12. In the circuit o f Figure 17.19, all initial conditions are zero. If i-^(^t) = u { t - 1)
A, find the response,
2H
0.5H
10 IH
2.5H 3H
SO L U T IO N
The key is to find the input impedance seen by the current source. Apply the T-equivalent circuit
to the coupled inductors o f Figure 17.18b. Using series-parallel techniques, the equivalent induc
tance in parallel with the 1 H resistance is = 2 H. T'hus
Exercises. 1. Repeat Example 17.12 with the dot on the 3 H inductor moved to the bottom.
AN SW ER: Same as in example.
M
2. Find the input impedance for the circuit o f Figure 17.20. In the circuit, k = .
is called the coupling coefficient, to be studied in the next section. v ^ l^ 2
ANSW ER:
2.V+ 0.5
0.50
3H
3H
o— o-
FIGURE 17.20
The first part o f this section justifies our assumption that One can justify that
A/j2 = A/t] = M by the principles o f magnetic circuits, but this is beyond the scope o f this text.
Our justification stems from the physical property that a pair o f stationary coupled coils cannot
generate average power. We will also show that the mutual inductance M has upper bound
’■> i.e., the mutual inductance can never exceed the geometric mean o f the self-inductances.
Justification o f Myj = ^ 2 1 ~ ^
This property is a consequence o f the principles o f electromagnetic field theory, which are beyond
the scope o f this text. To make our approach accessible to the beginning student, we build our jus
tification on the passivit)' principle for inductors.
, du d ii
(17.23a)
„ = + (17.23b)
~ dt ~ dt
Let us apply /j = sin(/) and /, = cos(r) to the inductors. From equations 17.23, the terminal volt
ages are
d d
Vj = Li — s in (/ )+ M p — c o s ( / ) = Li c o s ( / ) - M p s in (0
dt " dt
and
Vo = yU^i — s i n ( 0 + c o s ( 0 = M ^ i c o s i t ) - Losin(t)
~ ~ dt ~ dt
The total instantaneous power delivered to the coupled inductors is the sum o f the powers deliv
ered to the inputs, i.e.,/>(r) = v^{t)i^{t) + V2 {t)i2 it)- Therefore
To calculate the average power delivered to the coupled inductors, we use the identities
sin(r) cos(f) = 0.5 sin(2f), sin^(r) = 0.5[1 - cos(2r)], and cos^(r) = 0.5[1 + cos(2/)]. It follows imme
diately that the first and the last terms in equation 17.24 make no contribution to whereas
the terms involving ^o
= (17.25)
This result shows very clearly that if M j 2 > •^21’ excitations /, = sin(f) and ij = cos(r) will
lead to a negative violating the passivity principle. Similarly, if M p < then the new
(transposed) excitations Zj = cos(f) and 12 = sin(r) will again lead to a negative violating the
passivity principle. Therefore, we conclude that A/p = J^ 2 r
W ith = A/-7] = M, the average power, P^^^, is always zero for arbitrary sinusoidal excitations.
Having proved that A/j2 = = M, we shall now show that there is a limit to the value o f M
that is attainable once I j and L 2 are specified. Again we use the passivity principle and the fact
that stored energy is the integral o f the instantaneous power.
Chapter 17 • Magnetically Coupled Circuits and Transformers 11
Consider the coupled inductors shown in Figure 17.9. The voltage-current relationships at the ter-
minals are given by
and
^ V2(/) = ± M ^ + t 2 ^ (17.26b)
with the upper sign (+) for the dot in position A and the lower sign ( -) for the dot in position B.
O ' Let us assume that the inductor currents are initially zero (at r = 0). In this state, there is no ener
gy stored in the system.
o
Recall that the energy stored by any device over a time period [0, 7] is the integral o f the instan
taneous power over the interval. For a pair o f coupled inductors with no initial stored energy, we
o
have
o
dt
o
= i L „ f (r )+ i z „ | (r ) ± j [ M (/,f + ,-,5 ),
o
^ 4 L ,f ( r ) 4 i ,/ | ( r ) ± M j ;'3 '^ ( ,V ,)
^ 2 “ 2 J/,(0)/2(0)
^ = U ,i f ( T ) + ^ L ^ il ( T ) ± M i t { T M T ) (17.27)
where ^(/|/2) is the total derivative o f the product /j /2 and is equal to /jc/f^ +
O '
For the specific scenario above, we apply driving sources to the inductors to bring the currents up
VO to /|(7) = /j and i2{1) = /j . Then, the energy delivered to the inductors during the time interval
(0, 7) is, by equation 17.27,
W (J) = ^ W \ ± ^ hh (17.28a)
To grasp property 2, we may adjust the sources so that the currents are brought back from 7j and
/ j at / = r to zero at some t = T > T. Then the energy delivered by the sources to the inductors
during T < t < T may be calculated in a similar manner, to obtain
o '
912 Chapter 17 * Magnetically Coupled Circuits and Iransformers
with the upper (-) sign for the dot in position A and the lower (+) sign For the dot in position B
in Figure 17.9. Equation 17.28b is precisely the negative ol equation 17.28a. Thus, all o f the ener
gy delivered by the sources during 0 < t < 7’ has been returned to the sources during T ^ t ^ T\
For this reason, the energy given by equation 17.27 is called the stored energy. Another way of
recovering the stored energ)' is described in Problem 48. The physics o f the situation shows that
the energ)- is stored in the magnetic field produced by the currents in the inductors.
M ^ 7 ^ ,L 2 (17.29b)
2
W(T) = - l 2 1 ± 2M.X + L ^
L|.v“ i
(17.30)
U2J [h ) - 2 “
where x = ^\l^2 ^ current ratio. Equation 17.30 shows that Wi'f) is negative whenever^(a;) is
negative. Now,y(,v) is a second-degree polynomial in a: with a positive coefficient for the term.
Consequendy, the curv'e ofy(.v) vs. a: will be a parabola opening upward. From analytic geometr)',
depending on the sign o f the discriminant D = {M~ - L^L-,), the curve may or may not intersect
they{.v) = 0 axis, as illustrated in Figure 17.21. From the figure, it is obvious that if D > 0, there
will be some current ratio that yields a negative^(a:) and hence a negative W{T), again violating
the passivity principle. Therefore, D = {M~ - L^L^) < 0, which yields equation 17.29.
f(x)
Chapter 17 • Magnetically Coupled Circuits and Transformers 13
The degree co which M approaches its upper bound yJL\l^ is expressed by a positive number
called the coefficient o f coupling, defined as
M
k ^ - j= . (17.31)
yjL\L2
From equations 17.30 and 17.31,
(17.32)
When /^ = 0, Af is also zero, and the inductors are uncoupled. When ^ = 1, the inductors have
uuity coupling, an idealized situation impossible to realize in practice.
S o lu tio n ^ ^
1. For both dot positions, k = - 7- -------- = , = 0.8 .
V 5x20
2. For the dot in position A in Figure 17.9,
Stored energy =
EX A M PLE 17.1 4 . In the circuit o f Figure 17.22, /j(7) = /j= 6 A. Find the minimum value o f
the stored energy and the corresponding value o f /2(7) = A.
M = 3H
9H
4H
FIGURE 17.22 Coupled inductors for calculating the stored energy in Example 17.14.
S o lu tio n
From equation 17.27,
W = 0 .5 L ,/ f+ 0 .5 L 2 / | -M / ,/ 2
= 4 .5 / 2“ - 18/2 + 7 2 = /(/2 )
Following the standard method in calculus for finding the maximum and minimum, we set
dW Idlj to zero and solve for Ij-
d l2 clh
This yields 12 = 2 A, and the corresponding minimum stored energy = 18 - 36 + 72 = 54 J.
7. IDEAL TRANSFORMERS
Two coupled dijfereutial equations containing three parameters Z.j, Z<2, and M characterize the cou
pled coils of Figure 17.9. By imposing two idealized conditions on these parameters, the pair o f
differential equations can be approximated by a pair o f algebraic equations.
Idealization 1: The coupled inductors have unity coupling; i.e., M~ = and the cou
pling coefficient ^ = 1.
Effect o f Idealization 1: With ^ = 1, the coupled coils have the voltage transform ation
property
chapter 17 * Magnetically Coupled Circuits and Transfonners 915
where a\ s 2i constant and both v-^ and V2 are the voltage dropsfrom dotted {position A) to
undotted terminals o f the coils, as per figure 17.9.
To derive the condition of equation 17.33, note that the constraint = ZjZ2 implies that Lj/M
= M /Zj. Let the ratio
M
Then Zj = oM and M = aLj. Substitute these relationships into equations 17.7. This leads to the
following sequence of equalities:
W ith unity coupled coils, equation 17.33 indicates that the voltages v^{t) and V2 {t), both from dot
ted {position A o f Figure 17.9) to undotted terminals, always have the same polarity. W ith coupling
less than unity, it is possible for v-^ and V2 to have opposite polarities at some time instants, as
shown in Example 17-10.
Idealization 2: In addition to unity coupling, the coupled coils have infinite mutual and
self-inductances.
Effect o f Idealizations 1 and 2: W ith k= \ and idealization 2, the pair o f coils has the cur
O
rent transformation property
where both / j and are the currents entering the dotted terminals of the coils, as per Figure
17.9, dot in position A.
To derive equation 17.34, we again make use of the unity coupling condition L^/M = M/L 2 = a^
and rewrite equations 17.7 as
d(Ly. .
-------- ^/l + /9 (17.35a)
dt\M ^ ^
916 Chapter 17 • Magnetically Coupled Circuits and Transformers
Reasonably assuming that the voltage v^it) is bounded, i.e., there is some finite constant such
that |i'|(f)|s for all t, then
lim
M-*oo \M
V2 (17.35b)
lim
L^-*o
In other words, as Zj “ ►<», Z2 <», and Af -► <», equations 17.35a and 17.35b each lead to
^ (a ii + / 2 ) - 0 (17.36a)
whose solution is
ai^{t) + i2 {t) = C (17.36b)
for some constant C. Now assume that the coils are unenergized prior to the application o f exci
tations; i.e., at some time f = Zq in the past, /j(Zq) = which must be true for any real cir
cuit. It follows that the constant C in equation 17.36b is zero, and consequently /2(^) = -^^|(^)-
We remark that this derivation is valid for non-dc voltages and currents. The negative sign in equa
tion 17.33 implies that at any time, if a current enters one coil at the dotted terminal, then the
current in the other coil must leave the dotted terminal.
The condition o f infinite inductances (Z,p Lj, M) is another idealization that is not realizable in
practice. However, this condition can be approximatedhy using a magnetic material with very high
permeability as the common core for the two coils.
r\
Chapter 17 * Magnetically Coupled Circuits anti Transformers 917
V’2 (0 N2 /tCO Ni a
are said to be an ideal transformer, as shown in Figure 17.23a.
FIG U RE 17.23 Symbol and defining equations for an ideal transformer, (a) Both currents enter
the dotted terminals, (b) enters the dotted terminal and leaves the dotted terminal.
In Figure 17.23 two vertical bars serve as a reminder o f the presence o f a ferromagnetic core in the
physical device. The word “ideal” may or may not appear in the schematic diagram. Again, the
mathematical model o f an ideal transformer depends only on the turns ratio a : 1 and the relative
dot positions. To avoid the negative sign in the current relationship, an alternative labeling o f volt
ages and currents as shown in Figure 17.23b may be used. The subscript p stands for the primary
coil, which is connected to a power source, and s for the secondary coil, which is connected to a
load. Note that is entering at the dotted terminal and is leaving the dotted terminal. The nota
tion o f Figure 17.23b is more commonly used in the study o f electric power flow.
One important simplification resulting from the idealizations is that an ideal transformer is char
acterized by two algebraic cc[\.\7iX.\ons in terms o f its terminal voltages and currents through a sin
gle parameter a, the turns ratio. This is to be contrasted with a pair o f coupled coils, characterized
by two differential equations containing three parameters, Z,p L-,, and M.
018 Chapter 17 • Magnetically Coupled Circuits and Iransformers
EXA M PLE 17 .1 5 . For the circuit o f Figure 17.24, suppose /?^ = 10 0.,a = 0.1,/ ?^ = 1000 Q, and
v^{t) = 20 cos(300;rr) V. Find /?,, /j(r), and /-,(/).
R i,(t) ijlt)
a: 1
SO LU TIO N
From the definition o f resistance,
„ V ,(0 a V 2 (/ ) 2 V’2 ( 0 2n n n . in n n ,n
= —— = — = a — ----- = a Ri = 0.01 x 1000 = 10 Q
'l(') -'2 (')
R
V](/) = ------ -— = — X 20cos(300> t/) = 1 0 co s(3 0 0 ;r/) V
/?_, + /?, • 2
It follows that
vAr) 10cos(300;rr)
/,(/) = -^— ----------------------- = cos(300;r/) A
10
2. If in Example \7.\5 R^= 10 ^2 and = 400 0 find /?,, /j(r), V2 {t), and ijit).
A N SW ERS. 40 £2. 16 cos(300.t/) V, O.-i cos(300.Tr) A. U)0 a)s(3()0;r/) V,-0.04cos(300.t/) A
E X A M P L E 17.1 6 . For the circuit o f Figure 17.25, find input impedance Z-JS) and then the
transfer functions
a:1
— <—
• +
Z,(s)
:
ideal
SO L U T IO N
Using the definition o f an impedance and the properties o f the ideal transformer.
^ ^ ^ V^(s) aVjis) 2 .
Z:,,{s) = —— = — = a — ---= a ZAs)
/|Cv) J i i s ) -hAs) ^
c rz ^ is)
Vi„(s) Z,(>v) + Zi„{s) z,{s) + a -Z ^s )
V|(5) a
in which case
Finally,
, , , ( , ) . Z2<£) = = ----------=1^^---------
Z i,(.?)V j(s) Zi_U) - Z,(.s) + a-Z^(s)
920 Chapter 17 • Magnetically Coupled Circuits and Transformers
Exercises. 1. Repeat the derivation Z-^p) in Example 17.16 For the case where the dot on the
secondary is at the bottom.
ANSW ER: Z J s ) = a"Z,[s)
2. In Example 17.16, suppose Z/ (.v) = and Z^{s) = R + L^. Find the frequency oJq at
which the primary is resonant, i.e., the source sees a pure resistance.
a
ANSW ER.
J lc
3. In Example 17.16, suppose a = 2, Z^is) Zp) = 10 12, and =— •
Find/jW and/j(r). ^ ^
Z Js ) = a^~Z^{s) (17.37)
where a is the turns ratio taken in the direction from source to load. (See Figure 17.26; the
dots are not marked on the figure because their positions are immaterial for this application.)
Chapter 17 • Magnetically Coupled Circuits aiul Transformers 921
10.V
Exercise. In Figure 17.26, suppose n = 0.1 and Find
ANSW ER.
.s- + 10
E X A M PLE 17.17. For the circuit o f Figure 17.25, find the instantaneous power delivered by the
ideal transformer and the energ)' stored by the ideal transformer.
SO L U T IO N
With conversion o f the voltages and currents to the time domain, the power delivered to ideal
transformer is by definition
i't
pit) = V j/| + v’o/^ = (c /i'-)) — =- + = 0
~ \ a}
Hence, considered as a single unit, an ideal transformer neither generates nor consumes instanta
neous power; whatever instantaneous power is received at one side must transfer to the other side.
Furthermore, since p{t) is identically zero, its integral with respect to t is also zero. Thus, an ideal
transformer cannot store any energy.
In summary, the instantaneous power dehvered to one side o f an ideal transformer is transferred
to whatever loads the other side, and as a result no energy is stored in the ideal transformer.
Exercises. 1. In Figure 17.26 suppose the primar)' o f the transformer has <7 = 0.1 and is connect
ed to a source having = 10 Q. Find Zj{s) for maximum power transfer to the load.
ANSWHR. Z^U) = 1 kLl
2. In Figure 17.26 suppose the primary o f the transformer again has a = 0.1 and is connected to
a source having Z^^(/100) = 10 + J5 0 Cl. Find Z^iJlOO) for maximum power transfer to the load.
AN SW ER. Z^(/100) = 1000 -y '5000 LI
Divide rhe first equation by Ms and the second equation by L 2 S. To observe the frequenc)'-depend-
ent behavior, set s =j(x) to obtain
= ^ I \ i j ( o ) + I- , ( j co ) = a I i { j c o ) + 12 (j co )
Mjw M
Practical transformers have rsvo customary properties: (i) the frequency content o f the voltages and
currents are band limited, meaning that the transformer is guaranteed to operate only over a
restricted frequency range, 0 < ^ O) :S that docs not include dc; and (ii) maximum volt
ages are specified, i.e., |Vj(/to)| :s and |K,(/a))| rs Under these practical conditions we
see that it is necessar)' to make M and L-> (and since a = I M, also) sufficiendy large over the
frequency range ^ co :s so that
Viij(o) V2 O )
0 and sO
Mco Li(o
Thus we will achieve the current transformation property /|{/(jd) = over the frequency
range o f interest provided the voltage magnitudes are less than
E X A M PLE 17 .1 8 . In Figure 17.27 an ideal transformer steps down the voltage o f a 2400
source to pow'er 10 incandescent lamps in parallel, each drawing 0.5 A.
(a) Find the voltage (magnitude) across each lamp, and
(b) find the current (magnitude) delivered by the source.
transformer load
source
2400 V
rms 0 10
FIGURE 17.27 A transformer is used to step down a source voltage to meet the lamp specs.
S olution
Since only magnitudes are involved in this problem, dot positions on the transformer and refer
ence directions for voltages and currents arc immaterial. The turns ratio is a =1760/88, Since 2400
= the voltage across each lamp is
Chapter 17 * Magnetically Coupled Circuits and Transformers 923
88
^’load 2 .4 0 0 = 120 V,
1,760
Since 10 lamps use 0.5 A each, the total current used is 5 A. Using the current transformation
property o f the ideal transformer,
/.V X 50 = 0.25 A
1,760
Exercise. In the circuit o f Figure 17.26, represent each lamp by a 200 H resistance. Suppose the
source is limited to delivering a maximum o f 3 kVA. How many lamps in parallel can be con
nected to the secondary o f the transformer?
ANSWER: 41
E X A M PLE 17.19. Figure 17.28 shows a simplified model o f an audio amplifier containing an
ideal transformer. The input voltage is 1 at 2 kHz. T he load is a loudspeaker, represented by
z 4 Q resistance.
(a) Find the average power delivered to the 4 load if it is connected directly to the ampli
fier (i.e., with the transformer removed and the resistor connected across A-B).
(b) With the transformer and load connected as per Figure 17.28, with turns ratio a= 5, find
the average power delivered to the load.
(c) If the turns ratio a is adjustable, what value allows maximum power transfer to the load?
W hat is the value o f the maximum power?
a:1
speaker
So lu t io n
(a) No transformer and resistor connected across A-B. First observe that the magnitude o f the current
through the 4 Q resistor is
200
4Q = 0 .2 2 1 2 A
9 0 0 -^ 4
From the voltage divider formula, the voltage across the primary is
100
pri (200|V ,„|)=20 V
900+100
= 351=4v
5
Therefore,
4“
— = 4 W
(c) For the maximum power transfer, the turns ratio a should match the secondary impedance to
that o f the primary; i.e., the 4 resistance reflected back to the primary should be 900 to match
the internal resistance o f the amplifier. Hence
900 = X 4 ^ = 15
W ith = 15, the reflected impedance is 900 Q. and |= 100 V, meaning that
100 f A ^
= 6.667 V. Thus --------= 1 1 .1 1 W.
15
Exercise. Repeat all parts o f Example 17.19 for a loudspeaker with a resistance o f 16 ohms.
ANSW ERS: i).-(,28 W. W, 11.11 W
The ideal dc voltage source was one o f the first basic circuit elements studied in this text. “Ideal”
means that it maintains a constant voltage for arbitrary loads connected across its terminals. Such
a voltage source does not exist in the real world. However, because an ideal dc source in series with
a resistance will approximately represent practical sources such as a battery, its use is very impor
tant. Similarly, although an ideal transformer does not exist in the real world, it can approximate
ly represent a practical transformer with the addition o f some other ideal circuit elements.
Chapter 17 * Magnetically Coupled Circuits and Transformers 92 S
Because o f the ferromagnetic core used in its construction, a practical transformer is inherently a
nonlinear device. Nevertheless, a first level o f approximation represents a practical transformer, by
the circuit o f either Figure 17.29b or Figure 17.29c. Figure 17.29a is ideal, whereas Figure 17.29b
contains extra inductances to account for leakage fluxes and other magnetic phenomena. These
inductances give rise to a finite usable bandwidth for a practical transformer. Figure 17.29c shows
additional resistances to account for internal power losses. Thus the ideal transform in conjunc
tion with other circuit elements can be used to approximate a real transformer. Specifically, trans
formers do not work for dc or for very high-frequency signals. The circuit model o f Figure 17.29b
makes this behavior clear: the inductance causes a short circuit at dc, and the inductances
and produce open circuits at high frequencies. A second level o f approximation would begin to
account for nonlinearities, a topic left for more advanced courses.
N ,:N ,
+ -I-
N. N,
(a) V.
N.
ideal
N ,:N ,
L» /Y Y \
-h +
(b) V,
ideal
• L. R +
V,
Ideal
F IG IJR F 17.29 Linear transformer models, (a) Ideal case, (b) Extra inductances to account for leak
age fluxes and other magnetic phenomena, (c) Extra resistances to account for internal power losses.
In Figures 17.29b and c, the parameters and are called primary and secondary’ winding leak
age inductances, respectively. They are due to that part o f the magnetic flux that links w'ith one
winding but not the other. is called the magnetizing inductance and is due to that part o f the
magnetic flux that links both windings. Practically speaking, and
926 Chapter 17 • Magnetically Coupled Circuits and Transformers
The model o f Figure 17.29c incorporates the resistances and which are the resistances o f the
primary and secondary windings, respectively. R^^ accounts for the power loss in the iron core due
to hysteresis and eddy currents. These resistances as well as the inductances L^, and can be
calculated from knowledge o f the physical layout and materials properties used in the construc
tion o f the transformer. This task usually requires a large set o f design formulas and (empirical)
charts. After a transformer is constructed, these parameters can be experimentally determined.
Books on ac machinery set forth such experimental techniques.
From a circuits perspective, once the model parameters in Figure 17.30 are known, the steady-
state analysis o f a circuit containing practical transformers requires no more than the phasor
method or the Laplace transform method for its analysis. l\vo examples will now be given.
Example 17.20 shows the effect these parameters have on frequency response.
E XA M PLE 17.2 0 . Example 17.19 considered an audio amplifier with an ideal coupling trans
former. Figure 17.30 shows the same amplifier with a practical transformer in place o f the ideal
one. O ur objective is to illustrate the effect a practical transformer has on the power delivered to
the 4 Q. speaker as a function o f frequency. We assume our turns ratio ^ = 15, which produced
maximum power transfer in Example 17.19. We use the model o f Figure 17.29c with = 40 Q
(equivalent winding resistance), ^ ^ = 8 mH (effective leakage inductance), 1 M ii (core loss),
and = 1 H (magnetizing inductance). For simplicit)^ we set Z,^= /?^= 0 in the model o f Figure
17.29c.
Lp= 8mH
1 5 :1
40
S o l u t io n
Our analysis begins with the writing o f two loop equations in terms o f /,(;) and l 2 {s):
-20()K „ = [ « , + /? . + L,,^ + ] /, - - V V L / ,
" M ‘ " R ,„ ^ L ,„ s } R,„ + L,„s -
1 -I- y
R,„
Chapter 17 • Magnetically Coupled Circuits and Transformers 92-7
- 2 0 W ,„ .( R , .
In matrix form
- 2 0 0 L„, 5
h = ■Vin
LpL,„S^ +[{R\+ Rp)L,„ + Req{i-p + + (Ri + f<p)Req
Our purpose is to show how average power to the load varies as a function o f frequency in a prac
tical transformer. To this end, for v,„(/) = yj2cos(cot) V, the power delivered to the load is the
power delivered to the primary o f the ideal transformer, which is
Recall that the effective value o f the input voltage is 1 V. To obtain the desired plot o f P iif) (dis
played in Figure 17.31), consider the following MATLAB code.
TD
TO
O
—I
o
4-'
$
0
Q.
q;
cn
(Z
I—
01
>
<
Frequency in Hz
(a)
"O
fD
o
OJ
5
o
Cl.
0)
cn
2
oj
<
Frequency in Hz
(b)
FIG URE 17.31 Plot of the power delivered to the speaker load as a function of frequency in Hz. (a)
Overall response, showing sharp decrease with increasing frequency, (b) Low-frequency response,
showing significant reduction in power transfer.
The plot shows that the power is down from the peak by about 0.5 W at 4 5 0 0 Hz, and, at 10
kHz, the power delivered to the load is down about 2.5 W. Also note that a practical transformer
does not operate at dc and at very low frequencies, as demonstrated in Figure 17.31b.
In the next example we consider voltage drops in commercial power lines using practical trans
formers.
Chapter 17 • Magnetically Coupled Circuits and Transformers ‘)2 9
EX A M PLE 17.21. Fhe pniccical transformer in Figure 17.32a is designed for operation at 60 Hz
and 1100/220 V, i.e., ideally a 5:1 step-down transformer. Using the model o f Figure 17.29c, we
represent the circuit in Figure 17.32b with the following parameters: = 0 .050 H (the primar)^
winding resistance), = 0.002 1^2 (secondary winding resistance), = coZ,^ = 0.4 Q (primary leak
age reactance), = (oZ,^ = 0 .016 ^2 (secondary leakage reactance), and = 250 Q (mag
netization reactance). R^^^is very large and its effect is neglected. If the load draws 100 A at a power
factor o f 0.6 lagging, i.e., I 2 = 100 Z. -53.13^^ A, and the load voltage is V , = 220 Z. 0^’ V, what
is the magnitude o f the needed source voltage to achieve the desired load voltage-current values in
the presence o f non-idealities o f the transformer?
ac Low
High Voltage Voltage
Source Load
(a)
Load
5:1
ac
© High Voltage
Source
2
ideal
(b)
FIGURE 17.32 Analysis of a circuit containing a practical transformer, (a) A practical transformer
for stepping down ac voltage, (b) A circuit model utilizing an ideal transformer.
So lu t io n
c- ^1 J’ lOO .. .
Since — !- = --------= 5 (eiven turns ratio), we have
yV2 220
Voltage across the primary o f the ideal transformer
= 5 X 2 2 1 .4 ^ 0 .2 0 7 ° = 1 1 0 7 ^ 0 .2 0 7 ° V
= 0 .2 x ( 6 0 - ; 8 0 ) = 1 2 - y l 6 A
1 107Z 0
= ^------- = 4 .4 2 8 Z. - 8 9.8° = 0 .015 - y 4.428 A
J2 5 0
= 2 3 .7 ^ -5 9 .5 4 ° x (0.05 + ;0 .4 ) = 9 .5 5 ^ 2 3 .3 4 ° V
Thus, the magnitude o f the source voltage is 1115.8 V, which is 15.8 V higher (to overcome the
non-ideal effects) than what would be needed for an ideal transformer.
When a circuit contains coupled inductors, loop analysis is a natural way to analyze the circuit.
One then solves the resulting simultaneous equations by any o f the techniques studied earlier.
Although very general and systematic, such methods have extensive mathematical operations that
obscure the essential physical properties o f the circuit. In this section, we shall present some mod
els for a pair o f coupled inductors that utilize an ideal transformer. Since the three basic prop
erties (voltage, current, and impedance transformations) o f an ideal transformer are relatively easy
to comprehend, using such models in place o f a pair o f coupled inductors helps us to more easily
understand the circuit behavior, without complicated mathematics.
Chapter 17 • Magnetically Coupled Circuits and Transformers 931
k=1
1:
1
o-
+ +
V,
ideal
(b)
T he proof o f the equivalence is easy. The v-i relationships for the circuit o f Figure 17.33a are given
by differential equations 17.6. We need only show that the circuits o f Figures 17.33b and 17.33c
have the same v-i relationships. Consider Figure 17.33b first. In this case, A/= JL]L-> . Using the
current transformation property o f the ideal transformer, we obtain
d Zo . ■
Vj = L — i\ + , — h —L\------ JL\L-t — = L i -------- h M —
' ' dt V cit ^ dt ‘ di dt
^+U — = M — + U -
“ dt dt dt
These two equations are exacdy the same as equations 17.6. Hence, the circuits o f Figure 17.33a
and Figure 17.33b are equivalent. A similar derivation proves the equivalence o f the circuits o f
Figures 17.33a and 17.33c.
Chapter 17 • Magnetically Coupled Circuits and Transformers
Exercises. 1. Verify the equivalence o f the circuits o f Figures 17.33a and 17.33c.
2. In Figures 17.34a and b, let Z, = 2.4 H, Z-2 = 6 mH, L j = 3 mH, and = 0.2 H. Find
for each circuit using the equivalences in Figure 17.33.
ANS WK KS . O.Ss, 0 . 1 6s.
k=1 k=1
z in ^ --
— >
L , C L , L , C L .
(a) (a)
FIGURE 17.34
E X A M PLE 17.2 2 . The circuit o f Figure 17.35a has a unity coupling coefficient. Find the band
width, the center frequency, and the maximum voltage gain.
300 Q M = 350nH
2450 nH
« .© 50 nH
104.5 pF 14.7 kn
(a)
soon 1 :7
Chapter 17 • Magnetically Coupled Circuits and Transformers
300 0
5 1 2 0 .5 pF
+
V
© 5 0 nH
V. 300 0
(0
FIGURE 17.35 Analysis of a unit)' coupled circuit, (a) A circuit containing unit)' coupled inductors,
(b) An equivalent circuit utilizing ideal transformers, (c) Equivalent circuit after reflecting load
impedances to primary.
So lu t io n
The coefFicient o f coupling is = M l -y/^1^2 = 3501 yjSO x 2 ,4 5 0 = 1. Replacing the coupled coils
with the equivalent circuit o f Figure 17.33b yields the circuit o f Figure 17.35b. From the imped
ance transformation property o f an ideal transformer, looking into the primary we see an imped
ance 1/49 times the load impedance. Therefore, looking into the primary, we see a resistance o f
14,700/49 = 300 Q. in parallel with a capacitance o f 49 x 104.5 pF = 5120.5 pF.
Vl(s)
Figure 17.35c captures the new equivalent circuit. Th e band-pass characteristics o f H {s) =
Vsis)
follow the analysis done in Example 16.2. Consistent with the notation o f Example 16.2, we have
Q _ ___ _ ___________
RC 150 X 5 ,1 2 0 .5 x l 0 ” ' “
\H(/
I 'io)\
'huav = —
ID = 0 .5
From the voltage transformation propert)' o f an ideal transformer, is simply 7 times Vy So the
only quantities affected for the transfer function
The above example illustrates how a simple substitution o f an equivalent model for coupled
inductors can reduce a circuit to a simpler form for which the analysis is straightforward.
Exercise. Solve Example 17.22 again using the equivalent circuit o f Figure 17.33c.
ANSW'HRS: I he same as given in Example 17.22, of course.
M :L M :L
■<— o
+ I
k'L.
ideal ideal
(b) (c)
L.:M
O— o— /YY\ ■<— o
+ + i. +
V,
k'L
ideal ideal
(d) (e)
FIG U RF 17.36 Four different models for coupled coils with /'< 1, cach consisting of one ideal
transformer and rwo self-inductances, (a) Coupled inductor with k < 1. (b) One model using an
ideal transformer and two inductances, (c), (d), (e) Three alternative models.
(a) (b)
FIGURE 17.37
The equivalent circuits of Figure 17.36 are derived as follows. Since it is possible to subtract
a small inductance from Z,j such that the remaining inductance, Z,'|, satisfies the condition L ’ = M".
In other words, the new inductance, L' ^ together with L-, and M, forms a unity coupling system. Since
Z,'j = M^IL) = k~L^, the induaance to be subtracted is equal to Zj - Z .'j = (1 - This inductance
must be added back in series with Z,', to obtain a model for the original coupled inductors. The mod
els o f Figures 17.36b and c result. Repe-ating the process on yields the models o f Figures 17.36d and
e. A total o f four equivalent circuits is possible. Each consists o f two uncoupled inductors and one ideal
transformer. Clearly, when k= 1, the models in Figure 17.36 reduce to those in Figure 17.33.
E XA M PLE 17.23. Find the steady-state components o f and Vjit) at the frequency 1 rad/sec
for the circuit o f Figure 17.38, in which = cos{t)u{t) V. Note that because a resistance is pres
ent, the circuit responses will contain both a transient and a steady-state component.
1Q
M = 2H
----------O -----------
+ •
(a)
i 2 :5
+
V, 2F
(b)
F I G U R E 1 7 .3 8 Stead y -state analysis o f co u p led in d u cto rs u sing th e ideal tra n sfo rm er m od el.
Chapter 17 * Magnetically Coupled Circuits and Transformers
for the coupled inductors in Figure 17.38a, we obtain the equivalent circuit o f Figure 17.38b.
From the impedance transformation property o f the ideal transformer,
It follows that
z . , , a ) = 0 .8 y , i z , ,, ,( ;) = 5 g ; j ^ = - 0 .0 8 8 9 ,Q
^ 0.2y 7 0 .0 8 8 9 ^ ^ ^ I0 4 2 Z 8 3 .6 6 ^ V
‘ 1 + 0.2./ - 7X).0889
V , = 2.5V^, = 0 .2 2 0 8 6 ^ - 96.34^’ V
SUMMARY
This chapter has examined the phenomenon o f induced voltage in one inductor caused by a
change o f current in another inductor. A new parameter called the tnutttal inductance (A/) between
the coils was introduced. M was defined as a constant and is present in the coupled differential
and 5-domain equation models o f the coupled inductors. As illustrated in the chapter, M can be
determined experimentally. This treatment has avoided digressing into the study o f magnetic cir
cuits, which in fact underlies a rigorous development o f the concepts in this chapter. From the cir
cuit analysis perspective, this mathematical treatment is adequate. However, for a deeper under
standing o f the physical phenomena, one must study the principles o f magnetic circuits.
A proof o f M p = = M was given that made use o f the passivit}' principle and some trigono
metric identities. With the establishment o f this equalit)', computing the energ)' stored in the cou
pled inductors follows from simple integration. The expression for the stored energ)' and the pas-
siyitv principle then led to an upper bound for the value ot the mutual inductance, namely M <
yj^ ^2 • 1 he coefficient of coupling was then defined as k = M l .
An ideal transformer was defined as a device satisfying both the voltage transformation and cur
rent transformation properties. For practical transformers, these two properties hold only approx
imately— for example, over specified voltage and current ranges as well as over specified frequen
cy ranges.
Transformers have broad applicability, h'or example, in power engineering, transformers are used
to step up or step down the voltages. In communication engineering, transformers are used to
change a load impedance for the purpose o f maximum power transfer.
Although ideal transformers can only be approximated in the real world, they nevertheless remain
an important basic circuit element because a practical transformer or a pair o f coupled inductors
can be modeled by an ideal transformer and some additional R, L, and/or C elements. The use o f
such models simplifies many analysis problems and gives physical insight into the operation o f a
circuit.
71-equivalent for coupled inductors: if two inductors have one terminal in common, then the
three-terminal coupled inductors are equivalent to three uncoupled induaors (one of
which may have a negative inductance) connected in the 7t-form.
Primary: The winding (coil) in a transformer that is connected to a power source.
Secondary: The winding (coil) in a transformer that is connected to a load.
T-equivalent for coupled inductors: if two inductors have one terminal in common, then the
three-terminal coupled inductors are equivalent to three uncoupled inductors (one of
which may have a negative inductance) connected in the T-form.
Transformer: a practical device that satisfies approximately the voltage transformation and cur
rent transformation properties.
Unity coupling: coefficient of coupling k= I.
Voltage transformation property: for unity coupled inductors, the ratio |i/j(/) : V2 {t)\ is a con
stant equal to the turns ratio N^INj =
r>
r>
Chapter 17 • Magnetically Coupled Circuits and Transformers 939
20 M = 2H
(It (It
V9 (r) = ± M ^ ± L9 ^
dt ~ dt
4H 3H
Figure P i7.1
Figure P I 7.2
940 Chapter 17 * Magnetically Coupled Circuits and Transformers
i,(t) ■>
Figure P i7.5
D
(a)
i,(A)
C H E C K : Coefficients for derivative terms in
1>
t (msec)
Kc loop a, ( I , , - Zp - loop b,
\ ^ (-Z p - Z j, - A/), loop c,
1\ / ^ 2
-1
(b)
v,(V)
COMPUTING EQUIVALENT
t (msec) INDUCTANCES, Z,.^(s), AND
---- ►
-2 RESPONSES FOR SIMPLE
(c) CIRCUITS
Figure P I7.6
9. (a) Find the equivalent impedance o f the
circuit o f Figure PI 7.9a. Use the result
7. Consider the circuit shown in Figure P i 7.7.
to find the impulse and step respons
I'h e 2 H inductor is short-circuited. The cur
rents in the coupled inductors for t> 0 are /,(r) es.
(b) Repeat part (a) for the circuit o f Figure
= 2e~‘ A and i^{t) = e~^ A.
P I 7.9b .
(a) Find the mutual inductance M (in
henries).
Chapter 17 • Magnetically Coupled Circuits and Transformers ‘M l
M = 2H
j T Y V -^ Y Y V -
3H ^ SW
v.(t)
6 30 n
M = 2.5H
r m
2H
. ''.w © IF
Figure P I7.11
(a)
(D 30 0
M = 2.5 H
/Y Y V
2H
12. Consider the circuit shown in Figure
P17.12.
(a) Determine the input impedance and
input admittance.
(b) C H E C K : admittance poles at - 0 .5 and - 3
Figure P I7.9 (b) Determine the transfer function
2H
/Y Y \
Figure P I7.12
Z„(s)
(a) (b)
Figure P I7.14
RL.„s
-cir (a)
AN SW ERS: (a) Z,„(.v) =
(a) L . + 7-3
independent ol dot positions
Figure P I7.15
18. Find the input admittance o f each o f the
AN SW ERS: (a) 25 rad/sec, 0 12; (b) I rad/sec, coupled inductor circuits in Figure P I7.18.
Z j j i u ) ^ 0.5 Q
M = 4H
16. A handy dandy henry counter is used to O-
measure the inductance o f a pair o f coupled Z,„(s)
A N SW ER:
M = 2H C.v-(. v -!-4) + .V+2
O- C = 1 / 6 F ; Z;„(.v) =
Cv(.v + 2)
Z,„(s)
22. For the circuit o f Figure PI 7.8, find
3H 4H Zin(s) =
(b)
Figure P I7.18 when = /?^ = = 1 n , Z,j = Z,-, = 2 H, and
A/ = 1 H. Hint: After writing loop equations
AN SW ERS: (;,) — , (b) — with numbers, consider using the symbolic
4.S i.v toolbox in MATLAB (or the equivalent) to
19. Find the input impedance o f the cir evaluate determinants in a Cramers rule solu-
cuit shown in Figure P I 7.19. Does the answer non.
9.V- +6s+\
depend on the positions o f the dots? AN SW ER: Z,„(.v) = - ^
3.S- + 10.V + 3
GENERAL ANALYSIS OF
CIRCUITS WITH COUPLED
INDUCTORS
23. For the circuit shown in Figure P i 7.23,
= 1 V.
M- (a) If a dot is placed at position A, com
AN SW ER; .v
Li + Ljj pute the zero-state response.
(b) If a dot is placed at position B, com
independent o f the dot positions
pute the zero-state response.
M= 1
20. Find the zero-state response to the
input V for the circuit o f Figure
P17.20.
Figure P i7.24
= 10 V, find the voltage gain 28. For the circuit shown in Figure P i 7 .28,
compute the zero-state response for
V2(.v)
G ,{s) = (a) = 30u{t) mA
(b) = 30e-^',Ki) mA
and then compute Vjit), t > 0. Use the formu (c) /^^(/) = 60(1 - e~^^u{t) niA. Hint;
las developed in Example 17.8 or, alternatively, Apply linearit)' to the answers to parts
write two mesh equations and solve by (a) and (b).
Cramers rule. (d) Suppose Vf^Q~) = 30 V and all other
9000.V
C H EC K : C^.= initial conditions are zero. Find the
( a + 500)(5 + 9000)
zero-input response.
26. Consider the circuit shown in Figure
PI 7.26.
(a) Suppose the circuit is relaxed at r = 0~.
If = \2n{t) V, find the time con
0.625 H
(b)
stant o f the circuit and
Suppose
for / > 0.
= - \lu{-t) V; find v^{t)
-L
0.1 F
<^10
0
for r > 0.
(c) Suppose = - 6 u{-t) + Guit) V;
find V2 U) for /^> 0. Hint: Use lineari Figure P I7.28
ty. AN SW ER rd): v^ Jt) = 1.6<---'//(r) - GAe-^'u{t)
\
200 0
M = 0.5 H
—o
+
29 . In the circuit shown in Figure P I7.29, all
vJt)
6 0.2 H'
5H
v,(t)
—o
initial conditions are zero at r = 0~.
(a) If = 2u{t - 1) A, compute the
response, Hint: Use the result
Figure P i7.26 o f Problem 19.
(b) Find /,(r).
27. For the circuit shown in Figure P i 7.27, (c) Find /,(r)
i j t ) = 100«(r) niA.
(a) If a dot is placed at position A, com
pute the zero-state response.
(b) If a dot is placed at position B, com
IH
pute the zero-state response.
M = 0 .i H
o-
+ A ANSWl-R: (b) /,(/) = 2(1 - '»)Ma-1 ) A
v Jt) 0.1 H \0.2 H ''
B 1• 50 mF 30 , Consider the circuit shown in Figure
P I 7 .30 , having zero initial stored energ)'. Let
Figure P i7.27 /?! = 9 a
(a) Find
A N S W E R (b ): = 0 .1 sin ( l()/)//(/) \’ (b) F in d Z J ^ ) = /? + Z .,W .
(c) Find
Chapter 17 •Magnetically Coupled Circuits and Transformers 945
Y = ..........
(d) F in d G „ 2 = ^ H = det(Y)/det(A)
simple(H)
(e) Find Gy = -3 s
C H EC K : H (s) =
(^ + 1X^ + 3)
(f) Find the response, to input
(b) Find the impulse response, ^(f), again
v j t ) = \6 .8 le^^^u{t) V. using the symbolic toolbox in MAT
R. LAB:
+ syms t h
'.w Q V,(t) 200
h = ilaplace(H)
0.2 H
+
V,(t)
33. Consider the RLC circuit shown in Figure 3 5 . Consider the circuit o f Figure P 17.35, in
P 17.33. which Z, = 0.9 H , Z2 = 0 .4 H , Af = 0.2 H , and
(a) Find andZ.„W . T?2 = 80 Q. Suppose = 120«(-r) - I 2 0 u{t)
(b) Find the poles and zeros of Z-JJ). Find V.
the impulse response. (a) Compute the currents in the two
(c) Find the zero-input response if inductors at ^ = 0“ .
= 10 V with all other initial conditions (b) Write two differential equations for
zero. the coupled circuits in terms o f /j(r)
(d) Find the zero-state response if = and
cos(100/)«(/) A. (c) Take the Laplace transform of the
(e) Find the complete response of the cir equations computed in part (b),
cuit. accounting for the initial condition.
(d) Find V2(t) for t > 0 .
(e) Now suppose = \2u{-t) -
• j V, Find for / > 0.
/2 5 0 -
L (t)( -----
S 0 .2 m F 0.2 H '
z L
v jt )i
Figure P I7.33
Figure P I7.38
.ANSWI-R: k = 0.2S
30
fY Y Y jy Y V
39. Consider the circuit o f Figure PI 7.39. Find
Q 20
the value o f Csuch that the voltage gain is zero
at CO = 3 3 3.33 rad/sec. Use the 7i-equivalent
circuit o f Figure 17.18c.
no coupling betw een inductances
(b) c
Hh k=0.5
Figure P i7.36
20 0 40 0
AN SW ER; L^.= 2 H. A., = 3 H, and = 1H
”0 0 .1 2 H .
0.27 H
200 0 ,
COUPLING COEFFICIENT
Figure P I7.39
PROBLEMS AND ENERGY
A N SW l'R: 3 3 .33 ul-
CALCULATIONS
37. Consider the circuit in Figure P i 7.37. Find 40. For the circuit shown in Figure P I 7.40,
the value o f Csuch that the voltage gain is zero find the coupling coefficient k such that at co =
at CO = 333.33 rad/sec. Hint: Use theT-equiva- 10"^ rad/sec, the voltage gain is zero.
lent circuit. Use the Ti-equivalent circuit o f Figure 17.18c.
k=0.5 0.5 pF
20 0 40 0
0.27 H
0.12
0.6 mF
T l
38. For the circuit shown in Figure P i 7.38, 41. Consider the circuit o f Figure P 17.41.
determine the coupling coefficient k such that (a) Find the mutual inductance M so that
at CO =10"^ rad/sec, the voltage gain is the coupling coefficient k = \.
zero. Hint; Use theT-equivalent circuit for cou (b) Obtain the transfer function using the
pled coils and recall that a series LC behaves as formulas o f Example 17.8. Make use
a short circuit at w = \NLC. o f your answer to part (a).
9i.S Chapter 17 • Magnetically Coupled Circuits and Transformers
40 n >V (s)
Figure P i7.42
(d) Determine the coefficient o f coupling k. currents are /j(0) = 7| and /-,(0) = Ij- Also sup
pose /j(7) = 0 = Between 0 and 7', /,(a)
and i-y(t) have arbitrary waveforms. Show that
the energy delivered by the inductors to the
current sources during the interval 0 < t < T
4H 9H equals ( 0 .5 7 ,7j2 0 .5 ^ 2 ^ +
dots are now moved to A-C, show that the final
result is (0.57.j7,2 + Q.5L,l2^ -
Figure P i7.45
48. In contrast to Problem 47, where the cou
ANSWHRS; (a) 6 6 .S j . (h) 2/3 A and 6 J, (d) pled inductors delivered power to the sources,
U.S we can connect two resistors 7?, and Rj to the
inductors and let the current decrease exponen
46. For the coupled inductors o f Figure tially to zero. In particular, for the circuit o f
P I7.46, /,(r) = A cos(o)/ +6) and = B Figure P I7.48, it is possible to show that the
cos(o)/ + q)). What is the period T in terms o f energ)' delivered by the inductors to the resis
CO? Regardless o f whether the dots are at A-B or tors during 0 < ^ < x equals (0.57,,7,^ +
A-C, sliovv that the average power delivered to ± A/7,7-,) depending on the position
the coupled inductors is zero; i.e., show that o f the dots. Demonstrate the validity o f this
assertion for the specific parameter values and
initial currents 7.j = 10 H, 7-, = 2 H, A/ = 3 H,
-Lii\ ( 0 +
/?, = /e, = 1 Q., /,(0) = 1 A,“uid /2(0) = - 3 A,
with dots in position A-B, as follows:
= 0 (a) Calculate the stored energ)^ at r = 0.
(b) Calculate /, (r) and for / > 0.
(c) Evaluate the integral
Hints:
1. Write down the coupled inductor +V'2/2)^//
equations in the time domain with ±
for dots in B and C. and comparc with the result o f part (a).
1 J
2. Recall that = — f (vi/, + v^h)dt\ CHECK.- (a) 5 J, (b) /,(r) = e Ui{t) A and i^{t)
“ = -?>e-‘u{t) A
substitute and simplify.
JL R, > v,(t) L,
,{t) L, L, v,(,) 0 1 , ( 0
Figure P I7.48
ANALYSIS OF CIRCUITS
CONTAINING IDEAL
TRANSFORMERS vJt)
50. In the circuit o f Figure P i 7.50,
R^=\50 Q, /?, = 600 D .,R^=\2 kQ,
and v,„(/) = 5 yj2 c 0 s{(0 t) V,where O)
= 200071 rad/sec.
(a) Find the turns ratio, Wj : ft-, for maxi Figure P I7.52
mum power transfer to Rj^.
Given your answer to part (a), find ANSW l-R; (a) Z , = 10 k il, Z, = 500 k il: (h)
(b)
0 .625. 0.1, 0.1, 0.0 0 6 2 5 ; (c) 5/1, -50/1, (d)
v^it) and v^ Jt).
Given your answer to part (a), find 0.5
(c)
and /2(r). 53. Repeat Problem 52 for ?n = 20 and n = 5.
(d) Find the power delivered to the load. CH EC K : 0.2
Figure P I7.55
A N S W E R : (a) 2 0 , 5, 100 W ; (b) 6 4 W
O '
Chapter 17 •Magnetically Coupled Circuits and Transformers 95 1
o
56. The circuit shown in Figure P I 7.5 6 crude such, it can be represented approximately by an
ly models an audio amplifier circuit. Set ideal transformer in conjunaion with some
20 inductances and resistances. This problem illus
K =— .
3 trates the use of such a model for analysis pur
Sw/
poses.
Each 8 Cl resistor models a tweeter, and each 16
n resistor models a woofer. Suppose the left A certain calculator that normally uses four 1.5
and right speakers each consume 80 watts of V batteries comes with an adapter whose
average power. Determine: approximate model is shown in Figure
(a) The turns ratios a and b such that the P I7.57a, where = 0.9 H, = 9.6 H, « = 20,
w average power delivered to the speak and R = 200 Q.
ers is as large as possible. (a) If the calculator is not connected to
(b) The voltage across and current the adapter output, but the adapter is
w
through (rms values) each woofer and plugged into a household ac outiet of
tweeter. 110 V ^ , what average power is con
(c) i and then (rms values). sumed by the adapter?
(d) The power delivered by the dependent (b) If the output o f the adapter is acciden
source. tally short-circuited, what are the
(e) The power delivered by the independ magnitudes (rms) o f the ac current
ent source. drawn from the wall outlet and
(f) The power gain o f the circuit, i.e., the through the short circuit?
ratio of the power delivered to the (c) If a typical load, represented by a 50 Q
speaker load to the power delivered by resistor, is conneaed to the output of
O '
the independent source. the adapter, what are the approximate
Right magnitudes (rms values) of
Speaker the voltage across and the
current through the load?
Source (d) If the adapter is
'O treated as a pair of coupled
inductors as shown in
O ' Figure P I 7 .57b, deter
mine the parameters Zj,
O L2> and M.
Figure P I7.56
O
C H EC K : i = 4 . . ^ ^ = 8 V ^ ,
Pl.m = 27.778
O
952 Chapter 17 • Magnetically Coupled Circuits and Transformers
i l = ^ = « = O.I
M L2
ideal
The circuit is in the sinusoidal steady state and
Figure P I7.58 has known phasor values V| = 1 V and I| = —j
A.
59. Consider the circuit shown in Figure (a) If A'f is infinite (i.e., an ideal trans
P 17.59. former) and 0) = 10 rad/sec, find the
(a) Compute the impulse response. magnitude o f the current ratio
(b) Compute the step response. Compare your answer with the value
(c) Compute the zero-state response to = 0.1.
= 26 cos(2t) V. Identify the tran ANSW ER: 0.1
sient and steady-state responses.
(b) Repeat part (a) for M = 10 H and to =
1:10
10 rad/sec. Hint: Use the phasor equa
20
tions in section 7 to calculate the cur
v„(s)
© 1F 0.01 F
rent ratio I^/Ip
AN SW ER: 0.09
Figure P I7.59
(c) Repeat part (a) for A/ = 10 H and (O =
60. Consider the circuit o f Figure P I 7.60. 1207T rad/sec.
Suppose the transformer dots are in positions A N SW liR: 0.0 9 9 7 3 5 . praciicall\' the same as
A-B. for ^/ = 0 . 1
(a) Find the transfer function
H {s) =
V,„U)
(b) C o m p u te the zero-state response,
Chapter 17 • Magnetically Coupled Circuits and Transformers 953
Q =
e . ' A 6a = arcsm
and ■
lo , 2sin(^^) to PJ
and for exact answers to part (c) recall equa
tions 16.49 through 16.52.
9S4 Chapter 17 • Magnetically Coupled Circuits and Transformers
B D
Figure P I7.67
B ideal d
transformer
67. In the circuit o f Figure P 17.67, suppose ANSWL'.R: (a) 0.1479 watt, (b) 0.64 watt, (c)
= 10 n , = 40 U, and R^ = 2 0.. for - 2, = 0.1463. 0.87671 watt, (d) 1 watt
(a) If the load is identified as the termi
nals A-B, find a for maximum power 69. Consider the network configuration in
transfer. Figure PI 7 .6 9 . The source voltage
(b) If the load is identified as terminals C- v^(t) = y / l C 0 S ((0 / ) V, where to = 10^ rad/sec.
D (i.e., is the load), find a for max (a) Design a lossless network N such that
imum power transfer. maximum average power is transferred
to the 10 load at to = 10^ rad/sec for
Chapter 17 • Magnetically G>upled Circuits and Transformers 955
O '
each o f the cases cited below.
Case 1 : N consists of one inductor L
and one ideal transformer. You will
need to determine the configuration
of the inductor and ideal transformer
as well as the values o f L and the turns
ratio a.
Case 2 : N consists of a pair of coupled Figure P I7.70
inductors, in which case you must
determine the values o f L^, and M. 71. The front end o f a radio receiver uses a
ance over the range o f frequencies and response o f this circuit and verify that it is a
(p
(b) Find the transfer fimction H{s) =
Lossless ^ 100 Suggestion: Use one o f
O '
''.W 10 nF N
the models of Figure 17.33 for the
coupled coils.
(c) Compute (O^, J| , and Q of
Figure P I7.69
the circuit.
O '
70. Consider the resonant circuit shown in
Remark: We could have designed a parallel res
W Figure P I7 .7 0 containing a pair of coupled
onant circuit to achieve the frequency-selective
inductors. Suppose = 20 Zj = 1.5 mH, M
property of the circuit in Figure P I 7 .7 la.
= 3 m H, Z.2 = 6 m H , and R i = 320 Q.
O However, because of the mutual inductance, we
(a) Find the coupling coefficient k.
can achieve a much larger voltage gain than
(b) Model the coupled inductors using
O that achievable by a simple parallel RLC. This
one o f the equivalent circuits in Figure
becomes clear when one uses a model for the
17.33. Find the impedance yj(f) and
coupled inductors o f Figure P I 7 .7 lb; this
O then Kj(/o)) as a function of C
model contains an ideal (step-up) transformer
(c) Determine the value o f C that leads to
having a turns ratio of 1:10.
O a resonant frequency of O)^ = 10
krad/sec.
(d) Find the voltage gain at resonance;
O
i.e., find
< J
Chapter 17 • Magnetically Coupled Circuits and Transformers
APPLICATIONS
72. Recently, coupled coils have been manufac
tured on printed circuit boards by printing two
spiral coils on opposite sides o f a board. Such
coils are called coreless PCB transformers and
can be used to effect electrical isolation as well as
energ}' and signal transfer. This problem illus
trates the use o f the models developed in this
chapter for an approximate analysis o f the fre
quency response o f coreless PCB transformers. AN SW ERS; k - 0.5. (b) l,slmni= 600 n il.
l.sfrics= 4^0 nH, turns ratio a = 2, (c) =
Figure P i 7.72 shows an equivalent circuit o f a 6 .0 8 6 Mrad/:,CC. 7^,., = 3.333 .Vlrad/j,ec. Q =
PCB transformer for which Z., = /-2 “ 18.26. = 9.13
M = 300 nH, and = Rj = 1.2 which rep
resent the winding resistances. Suppose tiie sec 73. Figure P I 7.73 shows two series RLC cir
ond coil is terminated in a capacitance o f C = cuits physically isolated but magnetically cou
600 pF in parallel with a resistance of /?y = pled. The input is a sinusoidal voltage sotirce
2000 a . cos(cor). We will investigate some
steady-state voltage gains. This investigation
(a) Find the coupling coefficient k. has apphcation to biomedical implants, where
(b) Construct a model for the coupled the coupled inductors represent a magnetic
inductors using Figure 17.35 of the coupling between an implant and an external
text. circuit.
(c) For a first analysis, assume that the
effect o f Rj^ and the shunt inductance
Chapter 17 • Magnetically Coupled Circuits and Transformers
V;„
Figure P I7.73
tance M needed to maximize |VJ.
What is the maximum value o f |Vout\'
ANSWTRS: (c) Q = 20. M = 0 . 05 mi l .
If the coupling coefficient ^ = 0 (or Af = 0),
r nttijnux = 1
then the steady-state output magnitude is
= 0. Further, as shown in Chapter 16, [V qI =
74 . Example 17.7 provided an approximate
Ql V^. As k increases from zero to its maximum
analysis o f a car ignition system. This problem
value o f 1, |Vq| decreases monotonically,
asks for a SPIC E simulation o f the example.
whereas lV^,J increases to a maximum value
Reconsider the circuit o f Figure 17.12b with
and then decreases. Exact formulas for several
out the switch. Show that the initial current
critical values will now be derived.
through the primary o f the coupled inductors is
/■^l(0~) = 12 A. Since the secondary is open-cir-
(a) Show that if the rsvo series RLC cir
cuited, there is no initial current there.
cuits in the primary and secondary are
Generate a SPIC E simulation for 0 < f < 1
identical, i.e., Z,j = Z2 = Cj = C2 =
msec. Verify that |^'2(^)|,„^v - 36 x 10^ V.
C, 2nd
when
k=~,
Q
and at this degree o f coupling
C H A P T E R
Two-Ports
An actress speaks into a microphone. Speakers instantly replicate her voice, which resounds
throughout the auditorium. W hat happens between the speaking and the hearing? A microphone
produces a voltage signal that changes in proportion to the tenor and loudness o f the voice o f the
actress. Amplifiers magnify this changing voltage signal perhaps a hundred or a thousand times to
drive speakers whose cones reverberate in proportion to the changing voltage signal. The cones
then cause the air to vibrate intensely, also in proportion to the tenor and loudness o f the actress’s
voice. Her words become heard by thousands because o f the amplifier.
Amplifier circuits are found in instrumentation and a huge number o f appliances. In radios, radio
frequency amplifiers first magnify signals from an antenna. Special circuitry then extracts the
audio portion from these antenna signals. Other circuitry amplifies the audio to drive speakers.
Video signals from a video cassette recorder are amplified by special circuits for connection to a
monitor. Amplifier circuitry enhances signals coming from sensors in various manufacturing
processes. Repeater circuits, among other things, amplify phone signals whose magnitudes have
attenuated during microwave transmission. There are a large number o f other applications o f
amplifiers.
From the preceding discussion, one can surmise that an amplifier circuit has an input signal and
an output signal. This configuration is represented by a device called a two-port that has an input
port for the input signal and an output port for the output signal. The following figure represents
the idea (one o f the homework problems asks for an analysis o f this amplifier):
960 Chapter 18 • Two-Ports
Input Output
Signal Signal
Often the circuit berween the ports is highly complex. This chapter looks at shorthand methods
for analyzing the input-output properties o f two-ports without having to deal directly with a pos
sibly highly complex circuit internal to the two-port. The chapter will provide a variet)- o f meth
ods for analyzing amplifier circuits.
CHAPTER OBJECTIVES
SECTION HEADINGS
1. Introduction
2. O ne-Port Networks
3. Tw o-Port Admittance Parameters
4. Parameter Analysis o f Terminated Two-Ports
5. Two-Port Impedancc Parameters
6. Impedance and Gain Calculations o f Terminated Two-Ports Modeled by z-
Parameters
Chapter 18 • 'Iwo-Ports %1
7. Hybrid Parameters
8. Transmission Parameters
9. General Relations am ong Two-Port Parameters
10. Reciprocity
11. Parallel, Series, and Cascaded Connections o f Two-Ports
12. Summary
13. Terms and Concepts
14. Problems
1. INTRODUCTION
Figure 18.1 shows a general one-port, whose two terminals satisfy the propert)' that for any volt
age Kj across them, the current entering one terminal, say, /,, equals the current leaving the sec
ond terminal. A resistor is a one-port: the currcnt entering one terminal equals the current leav
ing the other terminal. A capacitor and an inductor are also one-ports. K general one-port contains
any number o f interconnected resistors, capacitors, inductors, and other devices. In a one-port,
only the relationship between the port voltage and current is o f interest. For example, the port
voltage and current in a resistor, capacitor, and inductor satisfy the relationships Cdv^Jdt
- Iq and Ldi^ldt= respectively. Practically speaking, one-ports are macroscopic device models
emphasizing input-output properties rather than detailed internal models. Thevenin and Norton
equivalent circuits determine one-port models when only a pair o f terminals o f a network is o f
interest.
O- O- -O
+ One + Two +
V. V, V,
Port Port
-o
(a) (b)
A two-port is a linear network having two pairs o f terminals, as illustrated in Figure 18.1b. Each
terminal pair behaves as a port; i.e., the current entering one terminal o f a port equals the current
leaving the second terminal o f the same port for all voltages across the port. Coupling networks
such as transformers have two pairs o f terminals, each o f which behaves as a one-port. Hence,
transformers are two-port devices. In modeling a rsvo-port, one must define a relationship among
four variables. Different groupings o f current and voltage variables lead to different t}^pes o f char
acterizing parameters. For example, adm ittance parameters (termed ^-parameters) relate the two
(input) voltages, and V-,, to the (output) port currents, /j and 7^. Impedance parameters
(termed 2-parameters) relate the two-port (input) currents to the two-port (output) voltages.
Other t}’pes o f parameters investigated in this chapter are hybrid or A-parametcrs, and trans
mission or /-parameters.
962 Chapter 18 • Two-Pons
Source One-Port
FIG U RE 18.2 A general rwo-port connccced to a source one-port (represented by a Thevenin equiv
alent) and terminated in a load one-port (represented by a Thevenin equivalent impedance).
In practice, two-ports often represent coupling devices in which a source delivers energy to a load
through the two-port network as suggested in Figure 18.2. For example, stereo amplifiers take a
small low-powcr audio signal and increase its power so that it will drive a speaker system.
Determining and knowing ratios such as the voltage gain, current gain, and power gain o f a two-
port is ver\' important in dealing with a source that delivers power through a t\vo-port to a load.
This chapter develops various formulas for computing these gains for each type o f two-port
parameter.
2. ONE-PORT NETWORKS
EXA M PLE 18.1. The circuit o f Figure 18.3 is pertinent to a simplified model o f a common-col-
lector stage o f a transistor amplifier circuit. Specifically, the one-port o f Figure 18.3 models the input
impedance. In a common-collector amplifier stage, the impedance Z j is very large and is often neg
lected, i.e., we assume |Z,| = =c over the useful bandwidth o f the circuit. The goal o f this example is
to compute Z and interpret Zy^^ in terms o f a transistor current gain parameter denoted by (3.
So lu t io n
We attach a hypothetical vohage source, say, across the port terminals, top to bottom being
plus to minus, to induce a hypothetical current ly Neglecting Z ,, from KCL,
Since the input impedance is the ratio o f the input voltage to the input current,
Zi„=-^=(P+l)Zi,
Thus, for a large P, say, 150, the input impedance can reach very high levels for reasonably sized
impedances Z^. When amplifying voltage signals, we desire to have the ratio o f the internal source
impedance to the amplifier input impedance be small. Conversely, the amplifier input impedancc
should be large relative to the internal source impedance.
Exercises. 1. In the circuit o f Figure 18.3, suppose P = 100, Z j = oo, and the load resistance Z^ =
16 Q. If a sinusoidal voltage = 15^ 2 sin(5000r) V is connected to the input terminals, find
the average powers and and the power gain.
AN SW ER: 139 mW. 14.06 W, 101
2. Consider the circuit o f Figure 18.4. T he one-port shown here models the output impedance of
a common-collector stage o f a transistor amplifier circuit. Show that
7 _ 'o i l ! _
Io„, (P + l ) '
3. In the circuit o f Figure 18.4, suppose P = 99 and Z j = 1 kH. If a current source = 200 niA
is connected to the one-port terminals, find and |/j|.
ANSW ER: 2 V. 2 mA
Chapter 18 • Two-Ports
A second example that is common to basic electronic analysis depicts a circuit used in the analy
sis o f field-cfFect transistors (FETs).
E X A M P L E 18.2. Figure 18.5a represents a simplified model o f the input impedance o f a field-
effect transistor circuit. As in Example 18.1, we neglect the large impedance in parallel with^^^K,.
Compute the input impedance in terms o f Z j, Z 2, Z 3, and
FIC'il’RE 18.5. Simplified input impedance model of a field-effect transistor (FET) circuit.
S o lu tio n
As illustrated in Figure 18.5, assume a hypothetical voltage source (plus to minus is right to
left) has been attached across the port terminals. At the bottom node, /, + g„,V-[ + ^ = 0. Since
K, = Z,/, and /2 = + K,)/Z2, it follows that
V-
Now, since Z,„ =
Avi
Z ;„ = Z , + (1 + g „ Z | ) Z2 = Z j + (1 + Z j) z ,.
Exercise. In the circuit o f Figure 18.5, suppose = 0 .002 S, Z-,= 500 Q, and Z j is the imped
ance o f a 0.1 l-iF capacitor. Find the frequency (in Hz) at which the magnitude ofZy^^ is 707 Q..
ANSW i-R; 1(I03.J 11/
EXAMPLE 18.3. Consider the circuit ofFfigiire 18.6. Find the Tlievenin and Norton equivalent circuits.
I.
I.
1 :a
----->---------
+ • + C A +
9 J > ■
S o l u t io n
The voltage across A-B satisfies
Kw = - ^ I a + V 2 = ^ I a * « V i
Cs Cs
To determine the Thevenin equivalent, we express Kj in terms o f and /^. From Ohm’s law and
the current relationship /j = -nl^,
Therefore,
yAB = + R cr
Cs
1 + Ra^ 1
h aRCs + Ra2
RCa^s + 1 Cs
J
(a) (b)
FIGURE 18.7 (a) Thevenin and (b) Norton equivalent circuits of Figure 18.6.
Exercise. Suppose R = 100 = 5, C = 10 mF, and = Au{t) mA. Find and the poles
and zeros o fZ ,,,
AN SW hR: I 'j. = lO/.v; pole at x = 0 ,ukI zero ai = U.04.
966 Chapter 18 • Two-Ports
A technique known as matrix partitioning often simplifies the calculation o f Thevenin and
Norton equivalent circuits. Example 18.4 illustrates the technique.
EXA M PLE 18.4. Find the Thevenin equivalent circuit for the network o f Figure 18.8.
So lu t io n
Step 1. Construct the loop equations. By inspection, the loop equations for the circuit satisfy
• v, ■ — + 18 -8 -10 h
2.V
13 6
-8 18 -4 h (18.1)
s
0 -10 -4 16 /3 .
The goal is to find in terms o f /p the matrix equation 18.1 is partitioned accordingly. To avoid
dealing with numbers, let us rewrite this partitioned matrix equation as
r\v,i \V,2
vvv> (18.2)
where = 136/j, = 0, and the ^Vjjs represent their analogous (partitioned) counterparts in
equation 18.1. In general, Vj is the net voltage in loop 2 and is the net voltage in loop 3.
Step 2. Solve the partitioned matrix equations for K, in terms o f /j arid the vector [ K,, The
matrix equation 18.2 may be rewritten as a set o f two equations, namely
h
V| +'^12
h (1 8 .3 )
and
Chapter 18 • Two-Ports 967
V2 I
= W21/1 + VV22 (18.4)
^3
Since is invertible (for all passive networks), solving equation 18.4 for the vector [/j
yields
V2 I
(18.5)
[V2
U/,, - W^2W22^V2^ ] /i + U'12^^22' (18.6)
^3
MATLAB or its equivalent, or possibly a symbolic manipulation software package, allows the
matrices in this equation to be conveniently computed. Comparing equation 18.6 with the struc
ture o f a Thevenin equivalent circuit,
V,
(18.8)
-1
■18 -4 ‘ •- 8 . _ .V+ 0 .0 9 5 ^
- Z ,„ ( .) = -1 -8 -1 0 = 5 .2 6 5 -------------- Q
Is -4 18 -1 0 s
-1 •1361
— 18 -4 ‘ 84
- 10 ] s
-4 16 s
0
s + 0.095
5.3
h + \- M
s
Exercise. Suppose the capacitor in Example 18.4 is changed to a 2 H resistor. Compute the new
Z,,[s).
AN SW l-ll: Z^,,{s) = 7.2647 i l via the MA'I'LAB code
AVI 1 = 20; W 12 = [-8 -10]:
»\V21 = [-8: -101: \V22 = [18 -4:-4 16]:
vZth = \V11 - \Vl2'inv(\X/22)*\X'21
1. L
O- -O
+ Two +
V. V,
Port
o -o
FIG URE 18.10 Standard two-port configuration having four external variables: /j, I^, Kj, and V-,.
Chapter 18 ‘ Two-Ports 969
ADMITTANCE PARAMETERS
'rhroughout this and later discussions, assume that the two-port ot Figure 18.10 has no inter
nal independent sources and that all dynamic elements are initially relaxed, i.e., have no ini
tial conditions. Under these assumptions, the admittance parameters or ^-parameters o f a
two-port are expressions for the terminal currents, /j and I 2 , in terms o f the port voltages,
and ^2’
Using either o f these sets o f equations, one can define each admittance parameter, as fol
lows:
( 18. 10)
-II V,
vs=o Vl=0
V2 I =
V, =0
Since each admittance is defined with regard to a shorted terminal voltage = 0, the^^y are
often called short-circuit adm ittance parameters. I'he unit o f an admittance parameter is S.
Some examples will illustrate convenient methods for computing the ^-parameters.
EXA M PLE 18.5. Compute the short-circuit admittance parameters o f the circuit in Figure 18.'
S o l u t io n
The overall strateg}' is to write equations for /, and /, in terms o f and using nodal analysis.
Accordingly, by Inspection,
* y 2 V ._ * K, ( K , - K,) = ( g ^ - r ,) . (K j * K,)
in which case
y\ \ - ^1 + ^3>^12 - ^V>'21 - ~ ^3’ ‘'*^‘^>'22 " ^2 ^3'
Exercise. Suppose the controlled current source in Figure 18.11 is reconnected across with the
arrow pointing to the left. Find the new ^-parameters,
A N S W E R : n , = K, + = g „ ,~ and y ,, = )\ +
In the next example, we combine the method o f matrix partitioning with the use o f nodal equa
tions to compute the y-parameters.
EXA M PLE 18.6. Compute the ^-parameters for the circuit o f Figure 18.12.
^TT
S o l u t io n
Computation o f the j'-parameters will again proceed by the method o f nodal analysis. Consider,
for example, node 1, in which the current /, must be determined, /j has the form
where each variable is understood to be a function o f s. The coefficient K j, is simply the sum o f
the admittances incident at node 1. The coefficient is simply the negative o f the sum o f the
admittances between nodes 1 and 2, and ^ is the negative o f the sum o f the admittances between
nodes 1 and 3. Similarly, I 2 = V-y] + Y-^^Vy where Yjj is the sum o f the admittances
incident at node 2, etc. Hence, in matrix form, the node equations o f the circuit o f Figure 18.12
are
Chapter 18 •Two-Ports 97
w .v+ 3 - .V : -2
r\y,i
11 : M/pi
iM
h = -s 5 + 21 -1 a V2
VV2V V w ji
0’
This nodal equation matrix is symmetric, because the /^LC network contains no dependent
sources. Using the method o f matrix partitioning introduced by equation 18.6 yields
■/r -li
-\/,- ( ■5 + 3 - s ' 1 T •V,-
[2 1]]
/2. - ■ .^ 2 . [ - 5 5 + 2 ” 5 1
i - + 2.2 -(5 + 0 .4 )
-(5 + 0 .4 ) 5 + 1.8
Exercise. The circuit o f Figure 18.12 is modified by adding a 2 resistor in parallel with the
capacitor. Find the new ^/-parameters.
r 5 + 2 .7 - ( 5 + 0 .9 )!
.ANSWFR:
- ( .s + 0 .9 ) .s + 2.3
S o l u t io n
Find the y-parameters by using nodal analysis in conjunction with the ideal transformer equations.
First, at port 1,
/, = K, + (V , - V,) = 2Vi - = 2V , - - ^ 2
(1 8 .1 2 )
Now considering that al 2 = -/, a node equation at the primary o f the transformer yields
Chapter 18 • Two-Ports
h =— h = -- (V i-V i) = - - V ,+ 4 - V - ,. (18.13)
a a a a~
where the last equality uses the relationship aV^ = V2. Putting equations 18.12 and 18.13 in matrix
form yields the ^'-parameter relationship
r\/,i
a a
Exercise. In the circuit o f Figure 18.13, the top resistor is changed from 1 Q. to 0.25 With the
turns ratio a =2 , find the new ^/-parameters.
A N SW I- R ; .S.
-2 1.25
1 =.^11^1 +>'12^2
One circuit-theoretic interpretation of this equation has the port current /j equal to the port volt
age Kj times an a d m i t t a n c e ] in parallel with a voltage-controlled current source^, 2 ^^2- ^
ilar interpretation o f the equation
A =^21 +J'22 ^2
These interpretations lead to the tw o-dependent source equivalent circuit o f a rwo-port repre
sented by the short-circuit admittance parameters. (See Figure 18.14b.) This equivalent circuit
aids the computation o f input and output impedances and voltage gain formulas. Note: in this
chapter and later chapters the standard resistance symbol often is used to designate a general
impedance or admittance.
Chapter 18 • Two-Ports
O- -o
+ +
y„ y.2
V. V,
L V2, V22 J
o-
(a)
FIGURE 18.14 (a) Short-circuit admittance parameters, (b) Their two-dependent source equivalent
circuit. The dotted line at the bottom of (b) indicates that the two halves are not necessarily connected.
This section takes up the task o f analyzing term inated tw o-ports modeled by ^-parameters. A
rvvo-port is terminated when it has source/load admittances. Any circuit or system in which a
source provides an excitation signal to an interconnection network that modifies the signal and
drives a load impedance can be represented by a terminated rwo-port. Such a scenario is common
to numerous real-world systems. For example, the utility industry delivers power to a home from
a generating facility through a transmission network, and a telephone system delivers voice infor
mation from the phone through a transmission network to a receiver.
O -o
+ +
’’ Vn y ,2
V,
L V21 V22 J
O ■O
The first method for computing is a matrix method. Recall equation 18.9a,
■/l' Vll >'12 ■\V
h .'21 V22 V2
Incorporating this terminal condition into the two-port ^-parameter equation yields
■/l >’12
det
0>’22 + y'L (>’22 + ^l ) ^1
V. =
det ■>’11 >’12 >’l 1(>’22 + " >’12>’21
>’21 >’22 +
Because is the ratio o f /, to the input adm ittance o f the two-port o f Figure 18.15 is
Vl2>'21 (18.14)
>’22 +
E X A M P L E 18.8. Derive the input admittance o f equation 18.14 using the two-dependent
source equivalent o f Figure 18.16. Here we avoid the solution o f simultaneous equations while
increasing insight into the operation o f the two-port.
I. L
FIG URll 18.16 Input admittance calculation using two—dependent source equivalent circuit.
S o l u t io n
W ith regard to the right side o f Figure 18.16, the current ^^21 iTiust equal the current through
the parallel admittances j /22
->'21^. = - ( 7 2 2 - ^ / .) ^2
Hence,
.V2 I
v-, = -
(18.15)
1
Equation 18.15 says that the voltage V2 equals the current, -^21 impedance, + K^)
Now consider the left side o f Figure 18.16. Here
>'12>'2I Vi
>’l I -
>’22 + /
Exercise. Let denote a source admittance. Show that the output admittance o f Figure 18.17 is
J 12J 2 I
(18.16)
V-
-O- -o
+ +
Vu Vl2
V. V,
V2, V22
-o
G ain C alculations
Our objective now is to derive a formula for the voltage gain of a doubly terminated r\vo-port, as
illustrated in Figure 18.18. Again, t h e s y m b o l s denote general admittances rather than the tra
ditional conductances.
+ yi2^2 +
V. V,
y„
V2/, '22
FIG URE 18.18 Doubly terminated rwo-port driven by voltage source, symbols denote general
admittances.
{ Ys \ ->’21 (18.17)
The overall gain calculation breaks down into two cascaded gain calculations as follows:
^ V, V/. V/,
Gi/ = — = — X — (18.18)
v; V, V/,
(1 8 .1 9 a )
Chapter 18 • Two-Forts
VS
To compute the gain G ^.2 = — from directly from equation 18.15,
V /.=-------
>’22 + Yl
which implies
^ V21
^ v2= — = - (I8 .1 9 b )
M >22 +
Equation 18.17 follovv.s by substituting equations 18.19a and 18.19b into equation 18.18.
Exercises. 1. In the circuit o f Figure 18 . 18, 7 ,, = 5 ,^ p = -0 .2 , y 21= 50, j/2-) = 1 (all *n mS), =
1 k n and /?^ = 2 kQ. Find Z-„ and G y = VJV^.
A N SW ER: = 8 5 .' U. 6\ - 2.(>32
ANSW^HR;
IMPEDANCE PARAMETERS
For the nuo-port o f Figure 18.10, the z-parameters zij relate the port currents to the port voltages
according to the mati ix equation
-1 2 -
h h=0 /,=0
(18.21)
Z2\ = ^22
h=0 /,=0
Since each zij is defined with one o f the ports open-circuited, i.e., /j = 0 or /-, = 0, the z ip are
called open-circiiit im pedance param eters. Their unit is the ohm.
Chapter 18 • Two-Ports
S o l u t io n
Rather than apply the z-parameter definitions o f equation 18.21, we will use mesh analysis to
obtain equation 18.20 directly.
(18.22)
.y “ .V \ s/
V9 = + — (/] + h ) = — /i + f 3 + — ] h
~ “ s ~ s \ s / (18.23)
10
’^11 - 12' s .y
Z=
]0 ay + 10
^21 ~22
s s
In the next example, we utilize the technique o f matrix partitioning to compute the z-parameters
o f a n:-network.
S o l u t io n
l/ ,= Z ,/ ,- Z ,/ 3
= Z 2/2 + Z 2/3
and
0 = - Z / i + Z 2/2 + (Z, + Z 2 + Z^)I^
Putting these equations in matrix form and partitioning the matrix appropriately yields
‘ z, 0 1 -z , h
= 0 Z, j Z2
^2 ^2
0 —Z| Z2 j Z| + Z2 + Z3 h
Hence, using the matrix partitioning formula derived in Example 18.4, we obtain
1 0 ■ 1 ■
- Z iZ i'
0 ^2 Z] + Z 2 + Z 3 -Z ,Z 2 ^2
Z | (Z 2 + Z 3 ) ^ 1^2
z, + z , + z . Z|Z2 Z 2 (Z ]+ Z 3 )
2. Now suppose that Z , and Z 2 are changed to 1 F capacitors and Z 3 is a 1 H inductor. Find the
new 2-parameters.
’-II - 12' 1 ’ .V - +1 I ■
A.N'SWER:
"21 -22 .v(.v-+2] 1 .V - + 1
Chapter 18 • Two-Ports 979
•V,- ^12' w
V2_ Z21 ^22 h
then perhaps it follows that
-1 ■Vi'
h' ’-11 ^12' 'y\\ 3’12‘ -Vi-
h >21 Z22 ^2 .>’21 .''22 .^2.
212' 1 ■5 + 1 1 ■
Exercise. A certain two-port has z-parameters ’^11
^21 ^22 .9-^2 1 5+ 1
Find the ^-parameters. Can you construct a three-element passive circuit that has these ^/-parameters?
Despite this inverse relationship, some circuits have z-parameters but not ^-parameters, and vice
versa, as illustrated by the following example.
EXAM PLE 18.11. Compute the z-parameters o f the circuit of Figure 18.21. Do the/-parameters
exist?
O- -O
+ +
V. V,
o- -o
S o l u t io n
By inspection, uhe 2-parameters o f the circuit o f Figure 18.21 are
V,- R R-
A
vs. R R ^2 I2
The 2-parameter matrix, [z-^, is singular, since d e t[2 j = = 0. Because the [z-^ matrix does
not have an inverse, the circuit has no y-parameters. One can check j/, j directly to verify this claim.
Consider Figure 18.22. Because Vj = 0, there is also a short circuit across Kj, making the ratio
vi 1 =
Vi
^-,=0
undefined.
O-
V, R v ,= o
O-
FIGURE 18.22. Equivalent circuit for c o m p u t i n g i n which port 2 is shorted, so that = 0.
Here, equals the sum o f two voltages: z^ j/j plus the voltage due to a current-controlled volt
age source given by z^-^Ij- This is precisely the left-hand portion o f Figure 18.23b. A similar inter
pretation follows for Vj = -221A ’ yielding the right-hand side o f Figure 18.23b.
Mr- 0 — o
O -0 + ^22 +
+ +
^12 V, V,
V, V,2
- ^2, ^22 .
-0 - "211
a
(a) (b)
F IG U R E 1 8 .2 3 (a) z-paramcters for modeled nervvork. (b) Tw o-dependent source equivalent circuit.
Chapter 18 • Two-Ports 98
This equivalent circuit proves useful for computing voltage gains and input and output imped
ances o f terminated two-ports. It should also be noted that there are other, equivalent circuits that
interpret the 2-parameters. A similar remark can be made for j/-parameters.
-12-21
^oiit ~ ^22 (18.25)
! 1 + •^v
In calculating which is the Thevenin impedance seen by the load, the independent voltage
source, is set to zero. Hence,
Solving for /, and substituting into equation 18.26 yields the output impedance formula o f equa
tion 18.25.
Chapter 18 • Two-Ports
Exercises. 1. Repeat the preceding derivation using a matrix method and Cramer’s rule.
2. Derive the following formula for the input impedance o f a terminated rwo-port:
V ^12^21
A n = ^ \ ---------(18.27)
<•22 +
Gain Calculations
The next phase o f our rwo-port analysis is to repeat the ^'-parameter derivation o f the voltage gain
o f the rwo-port in the context o f 2-parameters. Specifically, our aim is to compute
r ^2 ^2 ^ ^
C 8-28)
V^l _ Z/,
Gvi = ^ = (18.29)
^2 -----^21^1
+ ^22
n + <-22 ^in
Substituting equations 18.29 and 18.30 into equation 18.28 yields the voltage gain:
-21
Gy = Gy2Gy\ = (18.31)
\Z22 + ^ L / \ ^in + /
An application o f this formula and its derivation to a cascaded network of two-ports (two transis
tor amplifier circuits) appears in the next example.
E X A M PLE 18.12. Consider the nerwork o f Figure 18.25, which represents a two-stage (transis
tor) amplifier configuration. Each stage utilizes the same transistor in a difi^erent circuit configu
ration. The first stage is an amplification stage that will amplify a small source voltage to a much
larger one. The second stage is an impedance-matching stage used to match the load to the out
put impedance o f the amplifier circuitry to achieve maximum power transfer, at least approxi
mately. The 2-parameters, in ohms, for each stage in Figure 18.25 are given by
Chapter 18 •Two-Ports 983
V 160
S o l u t io n
6 6 7 9 1 x 1 .0 2 5 8 x10^
An! = ^11----- = 1.0262 X 10^ - (18.33)
+ Z, 6794+16
The gain here is small, but remember that this stage’s real purpose is impedance matching, not
amplification. By proper choice o f the z-parameters, the output impedance will approximately
match that o f the load. This allows us to dispense with an expensive and bulky impedance-match
ing transformer.
Step 3. Compute fo r stage 1. Here, observe that (i-’quation 18.33) in parallel with the 2
kD resistor between the two stages acts as a load to stage 1. Z-^^j parallel with 2 kQ becomes
the load to stage 1, denoted Zj^ = 1239.8 H.
The input impedance seen at the front end o f stage 1 follows from equation 18.27:
4 , = - ^ 1 2 ^ 1 - = 350 + ^ = 6 8 7.3 n
Z22+^ li 6 . 6 6 7 + 1 ,2 3 9 .8 (1 8 .3 5 )
Chapter 18 • Two-I’orts
Step 4. Compute the voltage gain, G,,, = V ,/VsJ o r stage 1. Using the result o f equation 18.35 and
applying equation 18.29 yields
j" ^21 ^
^vl - T 7 -
\ Z^. + Zj,j) \^/J + ^22/ (18.36)
i -10^’ ^( 1,239.8
= -2 0 5 .7 .
[ 7 5 + 687 .3 ] 1 1 ,2 3 9 .8 + 6.6 6 7 ;
Here the large gain of stage 1 leads to significant amplification o f the input signal. For example,
a —40 mV sine wave would be amplified to a little over 8 V, which can drive a small speaker.
Step 5 . Compute the overall voltage gain, Gy= The desired gain is simply the product o f
equations 18.34 and 18.36, i.e., G’^/= x = “ 2 0 5.7 x 0.7388 = - 1 5 2 , which remains fair
ly large. Indeed, a —40 mV sine wave would be amplified to approximately 6 V. Other amplifica
tion stages could be added to further increase the overall gain.
Step 6. Verify that the load matches the amplifier circuitjy to a reasonable degree. In this task, one
first computes the output impedance, o f stage 1 using equation 18.25. The parallel combi
nation of with 2 k n , denoted Z p = 1.732 kH, becomes the source impedance to stage 2. It
is then easy to compute the output impedance o f stage 2, again using equation 18.25. The answer
is = 17 Q. The details are left as an exercise.
Exercises. 1. For the circuit o f Figure 18.25, verify that the output impedance equals 17 Q.
2. In Figure 18.25, call the 2 kQ resistor R-,. Find a new value of Rj so that exactly equals 16 Q.
ANSW ER: 1.783 k il
3 . Suppose the 75 O. source resistance in Example 18.12 is changed to 300 which would represent a
twin line connection bet\veen an ideal voltage source and the first amplifier stage. Redo the example.
ANSW ER: Numerical values are obtained using the h)llowing M AFI.AB code;
1^ = 300;
z z ll = 1.0262e6: zzl2 = 6791;
7.7.21 = 1.0258e6; /.z22 = 6794;
x.ll = 350; vM = 2 .6 6 "; z21 = -le 6 ;
z22 = 6667; R2 = 2e3; ZL = 16;
Zin2 = zzl 1 - z/.I 2'/.z21/(/z22 + ZI.)
(;2 = (zz21/Zin2)-ZL/(Zl.+zz22)
ZLl = R 2 ’ Zin2/(R2+Zin2)
Zin = z ll - zl2'z21/(z22 + Z L l )
G1 = (z21/(Rs+Zin))-ZLl/(ZI.Uz22)
Gv = G 1 *G2
Zoutl = z22-zl2*z21/ (7ll+ R s)
Zs2 = R 2*Zoutl/(R 2+Zouil)
Zout = zz22-zz 12' zzl I /(zz 1 1+Zs2)
Chapter 18 • Two-Ports 985
7. HYBRID PARAMETERS
As we have seen, some circuits have ^-parameters but not z-parameters, and vice versa. A circuit
element that has neither is the ideal transformer.
EXA M PLE 18.13. This example shows that the ideal transformer o f Figure 18.26 has neither z-
nor/-parameters. From the definition o f an ideal transformer (Chapter 17), V2 = a V and /j =
-al-). Clearly, K, and cannot be expressed as functions o f /j and I 2 , nor can /, and be
expressed as functions o f K, and Vj- Hence, an ideal transformer has neither j/- nor z-parameters.
1 :a
o-
+ +
V,
Two-port circuits having neither z- nor/- parameters require an alternative modeling technique.
The hybrid parameters offer one o f several alternatives.
HYBRID PARAMETERS
Hybrid parameters, h-jy are a cross between y- and 2-parameters: a voltage and a current
are outputs, with /, and Vj as inputs. Specifically, if the rwo-porr o f Figure 18.10 contains no
internal independent sources and has no initial stored energy, then the hybrid parameters are
defined by the matrix equation
Vi- •/'II im -
(18.37)
h /'21 hi V2
As with both/- and 2-parameters, we interpret equation 18.37 as a two-dependent source equiv
alent circuit, as illustrated in Figure 18.27.
‘)86 Chapter 18 'Two-Ports
'2
a
-o +
+ +
■ h,2~
V. V.
_ h21 h22 _ -
-o
o-
(a)
FIGURE 18.27 (a) Hybrid parameters, (b) Two-dependent source equivalent circuit.
Exercise. Justify the two-dependent source equivalent circuit interpretation o f Figure 18.27; i.e.,
apply KVL and KCL to the circuit o f Figure 18.27b to derive equation 18.37.
The definition o f each /^-parameter follows directly from either the preceding equivalent circuit or
from equation 18.37. For example,
V'-,=0 (18.38a)
Because Aj| is the ratio o f an input voltage to an input current, it is an input impedance. Since
^2 = 0, /?], is termed the short-circuit input impedance. Notice, however, that A, j is simply
related to both the and the z-parameters as follows:
I -12^21
/ ,„ = ^ 'I I
(18.38b)
The second //-parameter, />2 p is called the short-circuit forward current gain, since it is the ratio
o f I j to /] under the condition K, = 0, i.e.,
V^=0 (18.39a)
From the 2-parameter equation Vj = ^21A ^'>2-^2’ ^2 “ *2 , has a simple z-parameter inter
pretation.
(18.39b)
^-,=0 ^22
The third //-parameter is
(18.40a)
/,=0
Since it is the ratio o f Kj to Vj under the condition that port 1 is open-circuited, i.e., /j = 0, it is called
the reverse open-circuit voltage gain. Interpreting //p in terms o f the y-parameters, we obtain
Chapter 18 • Two-Ports 987
>'12
(18.40b)
I,=0 '"'I
Finally, we note that the open-circuit output adm ittance is
h (18.41a)
l' n = -
/,=0
and has units o f S. The word “open-circuit” suggests a z-parameter interpretation. Considering
that V-y = 221-^1 h ~
1 _ V12J 21
— - >'22------------
:22 3’11 (18.41b)
/,=0
This relationship is similar (notice the subscripts) to equation 18.38b, which determines the short-
circuit input impedance.
We will return to these equivalences later, after we gain some computational experience.
E XA M PLE 18.14. Consider the two-port o f Figure 18.28, whose front end is a short circuit and
whose secondary is an open circuit. Thus it has neither z- nor ^-parameters. Our objective is to
compute the /^-parameters.
I L
O- -O
+ +
V,
-o
FIG URE 18.28 Simple two-port with /^-parameters but neither z- nor ^-parameters.
SO L U T IO N .
By inspection, the A-parameters are
/2. 0 0 V2
T he second example illustrates the computation o f /^-parameters for an ideal transformer circuit.
Chapter 18 • Two-Porcs
EXA M PLE 18.15. Find the /^-parameters o f the c%vo-port in Figure 18.29.
O---------
1 :a ------o
+ • *+ R +
A
V. V, V,
S o l u t io n
Step 1. Construct an equation fo r K,. From the primar}' and secondary voltage relationship o f an
ideal transformer,
V, = - v - ,
n 1
V, + - / o
a ~ a ~ (18.42)
Step 2. Construct an equation fo r in terms o f the other variables. From the primary and second
ary current relationship o f an ideal transformer,
/ 2 = -/ ,
Cl (18.43)
Step 3. Write equations 18.42 an d 18.43 in matrix form, and solve fo r Kj and I-, in terms ofl^ and
Vj. In matrix form, equations 18.42 and 18.43 are
rv/,1 0 i
a /ll (18.44)
/o - - 0
a
Solving for the vector [ Kj produces the /^-parameter equation
-1
1 r 0n \R \'
-V,- -M —r —
a a h' a~ Cl
h VS ^2
- - 0 -1 0
0 1 a a
Chapter 18 • Two-Ports
L
R.
O-----
+
V, '2 V,
o- -o
S o l u t io n
Step 1. Find A, j using equation 18.38a. With V2 = 0, i.e., port 2 shorted,
= R^
v-,=o
Step 2. Find using equation 18.39a. Again with = 0, i.e., port 2 shorted, the current I 2 =
-/ j since all current flows through the short circuit. Hence
V-,=0
Step 3. Find A,2 using equation 18.40a. W ith /j = 0, i.e., port 1 open-circuited, K, = V2 since
there is no current through Ry Hence,
/,=0
Step 4. Find A22 using equation 18.4 la. Again with /j = 0 , /2 = (C^ + in which case
= Cs + Gt
/,=0
In summar\-,
\Rx 1
-1 Cs + G^
990 Chapter 18 • Two-Ports
Exercises. 1. If all parameters in Example 18.16 are 1 with proper units, find the /;-parametcrs.
ANSW ER:
2. If the capacitor in Example 18.16 becomes an Z. H inductor, find the new A-parameters.
ANSW ER:
-1 — + G-.
Lv
This subsection analyzes the impedance and gain properties o f a terminated two-port character
ized by /^-parameters as in Figure 18.31, similar to the analysis done with both the z- and )'-param-
eters.
Vi
Recall that Zj„{s) = — . From the right half o f Figure 18.31, ~ ~
h
It follows immediately that
/hi
(18.45)
From the left-hand side o f Figure 18.31, = A|j/| + Substituting for Vj using equation
18.45 implies that the input impedance is
/, 1,22+ Yl (18.46)
Knowledge o f the input and output impedances/admittances permits us to derive various gain for
mulas in terms o f the //-parameters. For example, consider again the left half o f Figure 18.31. Since
the input impedance is known from equation 18.46,
h=^V \ (18.48a)
^2 = ------ ^ / | = ------ V,
1122+Yl ^in (18.48b)
^ V2
^v2 = 7 7 =
^1 + W (18.49)
Voltage division at the front end o f Figure 18.31 yields the other voltage gain formula,
G, _
Zi„ + Z^ (18.50)
The overall voltage gain is the product o f equations 18.49 and 18.50, i.e.,
^ ^ 1 /?21
7 7 = C j.|Gi,2 = -
A N SW ER: ^2 ^ ^lM\
/1 h->'^ + )' /
8. TRANSMISSION PARAMETERS
Transmission or f-parameters were first used by power system engineers for transmission line
analysis and are still so used today. They are sometimes called ABCD parameters.
V92 Chapter 18 • Two-I’orts
TRANSMISSION PARAMETERS
T he /-parameter representation has the matrix relationship
Vj- ■^11 hi ■V2 ■
( 18.52)
/ l. fl\ t il -I2
with the matrix, T = called the /-parameter matrix. As with th e ^ , z-, and /^-parameters,
the entries t-j are defined as follows:
^ " ■ ^ 2 /, = 0 ^-,=0
(18.53)
/,=0 v^=o
The matrix equation 18.52 leads directly to the relationships o f equations 18.53 by setting
the appropriate quantity, I 2 or Vj, to zero.
In computing a single t-j vs'ith equations 18.53, some care must be exercised in exciting the circuit.
By definition,
h= ()
The ordinary interpretation o f this equation is: apply an input V-, and find an output under
the condition that = 0. 'I'hen j is the ratio o f the Laplace transform o f the response to that
o f the input K,, i.e., a reverse voltage gain when port 2 is open-circuited. This situation causes a
predicament: an independent voltage source for Vj causes a current to flow. To circumvent this
predicament, we use the slightly modified formula
1
^11
/7 = 0
The quantit)' ^ forward voltage gain when port 2 is open-circuited) suggests that
we excite port 1 by Kj with port 2 open-circuited, which forces /-, = 0. It is then straightforward
to calculate /j j as the inverse o f Similar interpretations must be made with regard to
the other defining formulas in 18.53.
EXA M PLE 18 .1 7 . Consider again an ideal transformer circuit, shown in Figure 18.32. Here V-,
= and /j = This leads to the /-parameter matrix
0 \ V2
a
0 a
-/o
Chapter 18 •Two-Ports
O 1 :a -o
+ +
V.
Input and output impcdance calculations for terminated two-ports modeled by f-parameters do
not follow the usual pattern. Nevertheless, the t\s'o-port of Figure 18.33 has input impedance
7 _ h\'^L +^12
(18.54)
and output impedajice
7
^out
= ^22^v *^^12
^21^5 +^11 (18.55)
V,
Exercise. The rwo-port o f Figure 18.33 has r-parameters r, j = 0 .0025, ^p= 500
^ ,^ 21= 3 .1 2 5 x 1 0 -8 S, and = 0.00625. IfZ^ = 200 kQ. and = 20 kH, find
AN SW ER: 80 kL X 200 k L l
One o f the most important characteristics o f /-parameters is the ease with which one can use them
to determine the overall /-parameters o f cascaded rwo-ports, as illustrated in the next example.
994 Chapter 18 • Two-Ports
EXA M PLE 18 .1 8 . Compute the f-parameters o f the cascaded two-port o f Figure 18.34 in terms
of and Tj, the t^vo-port parameter matrices o f the first and second sections, respectively.
S o l u t io n
From the definition o f the f-parameters for each two-port,
1 r ^21 o [ K;*T 1
1 and = 7’-,
/ l. .-^ 2 . h .-^ 4 .
1
r^ ii
• ■ ■V3- ■V 4 ■
= 7 ’i1 = T\1 = T\Tl
1 L
-h h .-^ 4 .
implying that the f-parameter matrix o f the cascade o f Figure 18.35 is simply Tj.
Exercise. The f-parameters o f the two-ports in Figure 18.34 are (in standard units)
0.1 0 • •2 10'
T\1 = and Ti =
0 10 0.1 1
Find the f-parameters o f the cascade (T j comes first, followed on the right by Tj )• Then find Vj
when a voltage source o f 4 V is applied to port 1 o f the cascade and port 2 o f the cascade is open-
circuited.
0.2
A N S W I- R : T =
10
The h-.y-, t-, and z-parameters are interrelated. This subsection derives several relationships, with
others left to the homework problems. The complete set o f results is summarized in Table 18.1 for
reference purposes.
The trick is to rewrite the jzf-parameter equations so that Vj and I2 are on the left with /j and Vj
on the right:
~ ^ \ lh - ^\\h
Z2 2 I 2 - ^ \ h ~ ^2
Solving equation 18.56 for the vector [V^ /j]^under the proviso that ^^2 0 yields
-1
T ,- ■1 -Zl2 ■Zii O' h'
h 0 Z22 -Z2\ 1 V2 (18.57)
Thus, we have used matrix methods to direcdy compute the A-parameters in terms of the 2^-param-
eters under the condition that 222 ^
All other relationships are derived in a similar manner. For example, to express j^-parameters in
terms of /^-parameters, one must rewrite the A-parameter equations so that and /j appear on the
left-hand side with and on the right. Then using matrix form and inverting the appropri
ate matrix under the condition of a nonzero determinant produces the desired result.
Vw>
Table 18.1 specifies the interrelationships among all the parameters studied thus far.
w
Chapter 18 * Two-Ports
A/; Ar ■
>'22 ~>'12 !h i [^11
^11 ^12 Aj Ay I1 2 2 ^21 ^21
z-Parameters
Z21 Z2 2 -> ’21 >’11 /!21 1 1 ti2
Av Av hn. /222 /21 h\
J_ _- h 12 [^22 -A t ■
^22 ■12
A- Ac >’11 >’12 /Ml ^12 ^12
y-Parameter
>’21 >22 //2i ^ -1 ^11
-■^21 ^11
A:: Az fh 1 1 /12 ^12
Exercises. 1. Use the code below to create an m-file in MATLAB for conversion o f 2-parameters
to /^-parameters. Verify that Z = [1 2;3 4] produces H = [-0 .5 0 .5 ;-0 .7 5 0.25].
2. Use the code below to create an m-file in MATLAB for conversion o f z-parameters to ^-param-
eters. Verify that Z = [1 2;3 4] produces T = [1/3 -2/ 3;l/ 3 4/3].
%converting z to t paramters
function [t,tl I,tl2 ,t2 1 ,t2 2 ] = ztot(z)
zl 1 = z (l,l); z l2 = z (l,2 ); z21=z(2,l); z22=z(2,2);
deltaz = z ll * z 2 2 - z l2 * z 2 1 :
tl 1 = zll/z21;
t l2 = deltaz/z21;
t21 = l/z21;
t22 = z22/z21;
t = [ t i l tl2 ; t21 t22];
3. Write m-files for the remaining items in the conversion table for your own future use.
10. RECIPROCITY
Writing node equations for an ordinar)' linear circuit leads to a matrix equation having the form
Often the node admittance matrix Y = [y^^] is symmetric, i-e-.7,y = yjj for ; ^ J. Such networks are
termed “reciprocal.”
998 Chapter 18 •Two-l’orts
RECIPROCAL NETWORKS
Any circuit that has a symmetric coefficient matrix either in a nodal equation or loop equa
tion representation is said to be reciprocal. Further, a two-port represented by either ^-param
eters or ^-parameters is said to be reciprocal if Z j 2 = Z21 ~^2r
From Chapter 3 we know that circuits without dependent sources have symmetric coefficient
matrices in both the nodal and loop equation representations. On the other hand, the symmetr)'
o f the z-parameters and j'-parameters is typically lost when dependent sources are present. In gen
eral, we can prove that a circuit containing R ’s, L's, C ’s, and transformers, but no dependent sources or
independent sources, is a reciprocal network. From the definition o f a reciprocal two-port, we can
conclude further that if the hybrid and/or f-parameters exist, then from Table 18.1
= “ ^21
and/or
Conversely, any two-port that has parameters satisfying these conditions is said to be reciprocal.
Exercise. Recall that the relationship between the ^-parameters and z-parameters is
'II A/
-11 -12 ^21 ^21
^21 1 ho
^21 ^21
Show that if 2 j 2 = ^21’ Ar = 1, where A denotes “determinant.” Then show the converse, i.e.
if A/ = 1, then Zp = z^j.
Proving that any two-port created from a reciprocal network has symmetric z-parameters is
straightforward. We write the loop equations in matrix form with /p and /, being the volt
ages and currents o f ports 1 and 2. Since the underlying netw'ork is reciprocal by assumption, its
loop equation has a symmetric coefficient matrix partitioned as shown:
where the W-j matrices are defined in the obvious way. j is symmetric. Since the inverse o f the
symmetric matrix Wj-y is symmetric, and since the sum o f two symmetric matrices is symmetric,
the resulting z-parameters are also symmetric. The symmetry o f the ^-parameters follows by the
symmetry o f the inverse o f a symmetric matrix.
FIG URE 18.35 Equivalence o f short-circuit zero-state responses induced by voltage sources for
reciprocal netw'orks.
What reciprocity interpretation 1 says is that if one applies a voltage at port 1 and measures the
short-circuit current at port 2 with an ideal ammeter (zero meter resistance), then applying the
same voltage at port 2 would result in measurement o f the same short-circuit current at port 1.
Conversely, we can show that if reciprocit}' interpretation 1 is true, then the j'-parameters are sym
metric. To see this, observe that the configuration o f Figure 18.35a implies
1()()() Chapter 18 •Two-Ports
^<>“1
Vo I —-----
„ _ ha
_ ^Ollt
» 2 - —
^in r,„=o ^2h
Reciprocit)' interpretation 2 says that if one injects a current at port 1 and measures the voltage at
port 2 with an ideal voltmeter (infinite input resistance), injecting the same current at port 2
would result in measurement o f the same voltage at port 1.
Exercise. Assuming reciprocit)' interpretation 2 is true, show that it follows that ^2V
2-parameters are symmetric.
Chapter 18 ‘ Two-Ports 1001
Exercise. Show that if reciprocity interpretation 3 is assumed true, it follows that A|, = \i.e.,
the forw ard short-circuit current gain, , equals the reverse open-circuit voltage gain, — .
^in V;«
The three reciprocity interpretations have rigorous proofs. They however, are beyond the scope ol
the text.
Finally, a two-port that is reciprocal has an equivalent circuit representation with no dependent
sources. For example, suppose a reciprocal two-port has the z-parameter representation
V\ - 2|I /j + 2|2 A
and
~22 ^2
1002 Chapter 18 • Two-Ports
in which case
(18.59a)
and
(18.59b)
liquations 18.59 have the so-called T-equivalent circuit interpretation given by Figure 18.38.
I.
^22’ ^12
-o
+ +
V. V,
o-
FIGURE 18.38 2-parametcr T-equivalent circuit of a reciprocal 2-port.
The resistors represent general impedances.
Exercise. Suppose the r\vo-port o f Figure 18.39 is reciprocal and modeled by ^param eter equa
tions. Compute K, ^ Yj, and in terms o f the ^/-parameters, |, j/p,
AN SW ER: 1 .J'l2’ -^22 -'12
O- -o
-I- -I-
V. V,
o-
A general linear two-port has four external terminals for connection to other networks as illus
trated in Figure 18.40a. When the input and output ports have a common terminal, only three
external terminals are available, as indicated in Figure 18.40b. Such a rwo-port is typically called
a common-ground tivo-port, although the common terminal is not necessarily grounded in the
sense o f being connected to earth.
Chapter 18 ‘ Two-Ports 1003
+ -• •- + +
A
'^Common'^
A +
General
V. V, V. ground V,
two-port
two-port
W
1. L
(a) (b)
An interconnection o f rwo-ports has a new set o f z-, y-, /;-, or ^-parameters obtained very simply
from the individual t^vo-port parameters. The interconnected rwo-ports o f Figure 18.41 have new
parameters computed from those o f and as follows:
(18.60)
(18.61)
(18.62)
(a) (b)
1004 Chapter 18 • Two-Porrs
-• •-
Na
-# #■
N
(c)
FIGURE 18.41 Three typical interconnections of rvvo-ports. (a) Parallel connection,
(b) Series connection, (c) Cascade connection.
where Y, Z, and T denote the admittance, impedance, and transmission parameter matrices,
respectively, and the subscripts a and b refer to the networks and N^, respectively.
A derivation o f these formulas is straightforward. For example, from the definitions o f and N^,
where
Vi. r/i.
ha
and similarly for the voltage and current vectors o f N^. From Figure 1 8 .4 la, ^“
K ^ h - ^ substitution yields I - (K^+ Y^Vy w'hich verifies equation 18.60.
To verify equation 18.61, consider Figure 1 8 .4 lb. Here and = Z^/^. But V=
and = /. By direct substitution V = (Z^+ verifying equation 18.61.
The derivation o f equations 18.60 and 18.62 is easily extended to the case o f more than two rwo-
ports;
K= k. h- - (18.63)
2. If two or more two-ports, common-ground or not, are connected in cascade, then
r = r ,r ,r ,... (18.64)
Equations 18.62 and 18.64 for the cascade connection hold whether or not the component rwo-
ports are o f the common-ground t)'pe. However, equations 18.60 and 18.63 for parallel connec
tions in general \\o\(\ only for common-ground two-port connections, as shown in Figure 18.4 la.
Similarly, the series connection equation 18.61 holds only for the case illustrated in Figure 1 8 .4 lb.
Series connection o f two general rwo-ports (Figure 18.40a) or series connection o f more than t\vo
common-ground two-ports (Figure 18.40b) requires an ideal transformer for coupling, as demon
strated in the homework problems. Examples 18.19 and 18.20 explain why equations 18.60 and
18.61 fail when two non-common-ground two-ports are connected together.
Chapter 18 •Two-Ports 100=)
E X A M PLE 18 .1 9 . This example illustrates the difficult)^ with a non-common-ground series con
nection. Consider the two-port shown in Figure 18.42, which is a series connection o f two com
ponent two-ports. The z-parameters o f the individual two-ports are given by
■2 r ■Rx+2 1 ■
and Zf,u -
1 2 1 /?3 + 2
S o l u t io n
(a) Observe that the parallel connection o f the 6 H and 3 ^ resistors is 2 Q. Thus, by direct cal
culation, the z-parameter matrix o f the interconnected two-port is
■6 4' •2 r ■8 r 10 2-
Z= ^ Zn
£i + Z;,
I) = -1- —
4 6 1 2 1 5 2 7
(b) With R ,= R . = 0,
Z,, =
On the other hand, by direct calculation, the z-parameter matrix o f the interconnected two-port is
•4 2- •4 1
Z= = + Ay =
2 4 2 4
(c) In part (a), Z ^ Z^ + Z^ because, after the interconnection, neither nor acts as a two-
port, as defined in Figure 18.40. This can be understood by inspecting Figure 18.43, with the
indicated the loop currents.
1006 Chapter 18 •Two-Ports
------- O-
0
-o-
V V,
o ---------^ s / s A - -------o-
FIGURE 18.43 A nonzero leads to /j ^ /*.
W ith y?j and Rj nonzero, the mesh current is in general nonzero. Observe that 1*^ = 1^ -
/|. Hence, for the left terminal pair o f the current entering the top terminal does not equal
the current leaving the bottom terminal. With unequal terminal currents, no longer has a z-
parameter characterization, because the 2-parameter definition requires equal currents entering
and leaving the terminal pair, as per Figure 18.40. On the other hand, if = /?, = 0 in Figure
18.43, then the third mesh equation is satisfied for arbitrary values o f the mesh currents /j, Ij, and
7^, as there is no resistance at all in the third mesh. In particular, let = 0, in which case /|j = 7j,
and the 2-parameter characterization o f is valid. Similar arguments hold for N^, If = R^ =
0, equation 18.61 holds because Figure 18.43 now has the same interconnection as depicted in
Figure 18.4 lb.
E X A M PLE 18.20. This example illustrates the problem o f a non-common-ground parallel con
nection. Figure 18.44 shows rwo two-ports connected in parallel. Before the connection, each
two-port has ^-parameters
0.7 - 0 .2
S.
- 0 .2 0.7
After the connection, by direct calculation, the new two-port has ^-parameter matrix
r 1.625 -0 .6 2 5 -
Y= S.
-0 .6 2 5 1.625
FIG URE 18.44 New rvvo-port N subject to invalid application of equation 18.60 for non-common-
ground rwo-ports in parallel.
The reason for the failure o f equation 18.60 under these circumstances is the same as for the cir
cuit o f Example 18.19. If a voltage source is applied to port 1 o f N, we find that the currents
and /j'^^ are not equal. Thus, cannot continue to be characterized by a set o f rwo-port^-param
eters in forming N.
There are, however, some special cases o f non-common-ground two-ports for which equation
18.60 holds for a parallel connection. The following is one example.
E X A M PLE 18.21. fhis example illustrates how to achieve a parallel interconnection o f two gen
eral rwo-ports o f Figure 18.40a so that equation 18.63 remains valid. Reconsider the tw'o-ports
and o f Example 18.20, which have y-parameter matrices and Y^, respectively. Suppose
a 1:1 ideal transformer is placed at the front end o f in Figure 18.45; call this new two-port
FIG URE 18.45 Equation 18.63 holds for these non-common-ground rwo-ports. (a) Parallel connec
tion of rwo general rvvo-ports. (b) Justification of equal currents at rw'o terminals of each port.
S o l u t io n
(a) The )'-paramcrers o f N j arc rhe same as for because rhe ideal 1:1 transformer with indicat
ed dor positions forces N * and to have the same port currents and voltages at port 1. In fact,
if N * and were enclosed in a box with only the leads observable, the two-port properties would
be identical.
(b) To show that the ^/-parameters of N (the interconnection o f N * and are Y = + Y^, we
must first show that
(ii) and 12 1 , - / ,j
Chapter 18 • Two-Ports l()()‘)
To confirm these equalities, consider Figure 18.45b, which represents Figure 18.45a with all the
nonessential components removed to avoid overcrowding figure. Four Gaussian surfaces are
drawn, 5,^, and 5,^^. Note that all Gaussian surfaces go through the core o f the trans
former. Recall that KCL holds for a Gaussian surface: the algebraic sum o f the currents entering
(leaving) the surface is zero. This immediately asserts the validit)' o f statements (i) and (ii) above.
Hence, within the interconnection, the two-ports N * and continue to act as individual two-
ports. Therefore, equation 18.60 remains valid, i.e.,
n .4 - 0 .4
Y = Y a -fY b =
- 0 .4 1.4
This example extends directly to multiple parallel interconnections. Although the precise condi
tions for the applicability o f equations 18.60 and 18.61 to the parallel and series connections o f
non-common-ground two-ports are known, they are not practical enough to be included here.
Our emphasis is on interconnections o f common-ground two-ports, which occur most often in
practice.
12. SUMMARY
This chapter presented a unified setting for one-port analysis while providing a comprehensive
extension to two-ports. Two-ports are common to numerous real-world systems such as the utili
ty power grid that delivers power to a home from a generating facility through a transmission net
work. Another representative two-port is a telephone system that delivers a speaker’s voice to a lis
tener by sending a converted electrical signal through a transmission network. The characteriza
tion o f a two-port for such systems is done through their input-output properties. Four sets o f
characterizing parameters were developed: impedance or z-parameters, adm ittance or ^-parame
ters, hybrid or //-parameters, and transmission or f-parameters. In order to analyze various
aspects o f a system characterized by a two-port, formulas for computing the input
impedance/admittance, the output impedance/admittance, the voltage gain, etc. were derived.
Quantities such as voltage and power gain are very important aspects o f amplifier analysis and
design, as illustrated in Example 18.12, which depicts a two-stage transistor amplifier configura
tion. Although Example 18.12 utilized the medium o f z-parameters, the more customar)' medi
um for transistor amplifier design is //-parameters.
Conditions and formulas for parallel connection o f two-ports were presented in terms of^-param-
eters while series connections were studied using 2-parameters. Formulas for determining the
transmission parameters o f cascades o f rwo-ports were also developed. In addition, the chapter
introduced and interpreted the notion o f reciprocit)' in terms o f the different two-port parame
ters. Reciprocal circuits generally contain only R's, Z.s, Cs and transformers. Under certain restrict
ed conditions a reciprocal network may contain a dependent source, as the homework problems
will investigate.
101 U Chapter 18 •Two-Ports
Admittance or j-parameters: descriptive two-port parameters in which the port currents are
functions o f the port voltages.
Hybrid parameters or /^-parameters: descriptive two-port parameters in which Kj and are
expressed as functions o f /j and K,.
Impedance or z-parameters: descriptive two-port parameters in which the port voltages are func
tions o f the port currents.
Input admittance; the admittance seen at port 1 o f a possibly terminated two-port.
Input impedance; the impedance seen at port 1 o f a possibly terminated two-port.
Matrix partitioning: the partitioning o f a matrix set o f equations into groups to obtain a simpli
fied solution in terms o f the partitioned submatrices.
Norton equivalent of one-port: a current source in parallel with the Thevenin impedance.
Open-circuit impedance parameters; the impedance or z-parameters.
Open-circuit output admittance: the hybrid parameter hj-,-
Output admittance: the admittance seen at port 2 o f a rvvo-port possibly terminated by a source
impedancc.
Output impedance: the impedance seen at port 2 o f a two-port possibly terminated by a source
impedance.
Partitioned matrix; a matrix that is partitioned into submatrices for easier solution o f sets o f
equations.
TU-equivalent circuit: equivalent circuit o f a reciprocal two-port containing three general imped
ances in the form o f k, as in Figure 18.39.
Reciprocal network: a network whose node equations or loop equations have a symmetric coef
ficient matrix.
Reciprocal two-port; Z]2 = ^21 ~J2V
Reverse open-circuit voltage gain; the hybrid parameter A jj.
Short-circuit admittance parameters; the admittance or ^'-parameters.
Short-circuit forward current gain; the hybrid parameter
Short-circuit input impedance; the hybrid parameter
T-equivalent circuit; equivalent circuit o f a reciprocal two-port having three general impedances
in a T shape as in Figure 18.38.
Terminated two-port: a two-port attached to a load impedance and a source with, in general, a
nonzero impedance.
Thevenin equivalent of a one-port: Voltage source in series with the Thevenin impedance.
Transmission or ^-parameters: parameters where and are expressed as functions o f Vj and
-h -
Two-dependent source equivalent circuit: equivalent circuit tor a two-port containing imped
ances/admittances and two dependent sources.
Chapter 18 • Two-Ports
P r o b le m s + V -
ONE-PORTS
1. In Figure PI 8.1, suppose 100 V,Z^ = 20
n , /?^ = 1 k n , and (3 = 49. Find o f rhe one-
port. Then find and the power to the one-
port. Does Z j have any effect on the answers?
Figure P I8.3
C H EC K : (a) 1 1 k n
'»(t)
Figure P i8.4
the initial inductor current is zero. Identify the steady-state part. I f the
input cosine frequency were changed
to 100 rad/sec, what would happen to
the steady-state magnitude?
R
-------syx/'------
b :1
Figure P i8.6
Figure P I8.8
ANSWKR; (a) Z,;, = .s- + 0.8. I = -2 /
hR
— I— "\/V^
in I 20
40 '2 0 J_ '
16
Figure P I 8.7
CH ECK: = 0.5
Figure P i 8.9
Chapter 18 • Two-Ports 1013
.'^’11 .’^’12' •4 - o .r
which m S,
He He >’21 .'’22. 60 1
2F 1F
= 1 k n , and = 2 kn .
10 2F in
(a) Compute K,,,. Z^. Y^,, and Z„,„.
(b) Find the voltage gain
Figure P I8.14
G,. = ^ .
K,
15. The terminated two-port configuration in
Figure PI 8.15 is characterized by ^-parameters. (c) If v^{t) = 9u{t) V, find the power
absorbed by R^.
Suppose V/ - + ANSW FRSM a) );„ = 8 m S ,Z ,,= 1 2 5 a
= 2.2 mS; (b) - 4 .4 4 ; (c) 0.8 wart
(a) Find^, j i f = 10 ^^,721 = - 1 S .^,2
= 0.03 S ,y 2 2 = 0.2 S, and R ^ = \ O n . 18. Reconsider the two-port o f Figure P I 8.15,
(b) If v^{t) = 3 0 i4 {t) V, find the power, in which R^= 10 H ,^ 2i = 2 S ,^ p = 0.02 S ,^22
p^U), absorbed by the 10 Q load at = 0.2 S, R^ = i o n , and the voltage gain is
port 2.
^ = 0 .6 .
(a) Findj/j,.
(b) If v^{t) = 10«(r) V, find the power
absorbed by Rj^.
C H EC K : (b) 160 watts
.(s)
Figure P i 8.16
20. In a laboratory, you are asked to determine
the admittance parameters o f a circuit. You
Chapter 18 •Two-Ports 1015
decide to shon-circuit port 2, place a unit step network assuming y2 (/) is the output.
current source at port 1, and measure the port (d) If vj^t) =10 s'm{2t)u{t) V, compute the
voltage, t/j = (1 - V, and the pon 2 steady-state and transient responses
current, /2 (^) = A. Knowing that this is assuming V2(t) is the output. Use the
sufficient to determine at most two of the residue command in MATLAB to
parameters, you then break the short circuit compute the partial fi^ction expansion.
and terminate port 2 with a 1 resistor and
measure the new step responses as = (1 -
+ V and ij{t) = A.
(a) Compute the/-parameters of the two-
port.
(b) If port 2 is terminated in a 1 Q resis
tor, find the input impedance seen at
port 1.
(c) If port 2 is terminated in a 1 resis Figure PI8.22
tor and driven at port 1 by a current CHECK: (b) 0.25 / (s + 1), (d) transient
source i^{t) = cos{t)u{t) A, compute response e~^u{t) V, steady-state response
the steady-state magnitude of the gain, (-cos(2f) + 0.5 sin(2/))«(f) V
Yz
23. Figure P I8.23 represents a two-stage ampli
h
fier. Suppose
21. This problem shows that one can simulate
•2 -0.64' 0.4 - 0 .0 0 r
an inductor using an active two-port terminat r, = mS, Y2 =
ed by a capacitor. For the circuit of Figure 25 1 7.5 0.025
P18.21, Cj = 125 juF, C2 = 0.8 F,/u =>^22 = 0’ = 150 Q, = 2 kn, and /?! = 2 kQ. Find
-JVi2 = 4 S. the volt^e gain
(a) Compute the input impedance Z-JJ).
z-' _ VOM/
What is the equivalent L seen at port 1?
(b) Determine the resonant frequency CO^
(c) If /? = 2 5 is placed in parallel with Cj, Although the solution may be obtained by solv
find the new resonant fi^juency 0)^ ing a set of six simultaneous equations, a much
better method that also gives more insight into
o-n e- the performance of the amplifier, and works for
y.. y.j
Y= any number of stages, is to proceed as follows:
(a) Find the input admittances y.nl and
i^»i-
Figure PI 8.21
(b) Find the voltage gains of the stages
CHECK: (a) = 50 mH, (b) 400 rad/sec, (c) successively, starting from the source
240 rad/sec end. Use this information to find the
overall voltage gain,
22. For the circuit of Figure PI 8.22, R^= I Q,
=1 Q, and q = 1 E
G, = ^ .
(a) Find the ^parameters. ^2 (5 )
^in
(b) Find the voltage gain, Gy2 = y •
(c) Compute the impulse ^
and step responses of the terminated
101(> Chapter 18 • Two-Ports
z-PARAMETERS
(a) With the switch in position A, find 26. Consider the circuit o f Figure P i 8.26.
(a) Compute the open-circuit impedance
V2(-v)
parameters.
V'|(5)
(b) If port 2 is terminated in Z-,(s), find
(b) If = -4u{t) V, find 0 < r< the input impedance,
0.25 sec. (c) If port 1 is open-circuited, Zj(s) is a 1
(c) Is rhe circuit stable? Explain your rea H inductor, Z-,(s) is a parallel combi
soning. nation o f a I Q resistor and a 0.5 F
id) Find i/-)(0.25~). capacitor, and
(e) For t > 0.25, after the switch has
moved to position B, draw and label 2K
the frequency domain equivalent cir S-+4
cuit that will allow one to compute find f,(r) and v>y(r) in steady state.
y^(s).
(0 Compute Zj(s) if ^ = 2.
1 :a
(g) Compute for /> 0.25. -^ s/ V -
z.
I, A B b;1
•-
Y=
W
(a) Compute the z-parameters. ------ ^ -------
(b) Find the ^parameters by matrix inver 1£1
r\
sion using, for example, MATLAB, r It *
\f
1 < 1
h 0.5 F _ 0.25 F +
(c) If port 1 is open-circuited and r >
1 0.5 F
8
D----- ----- 0
Figure P i8.29
O '
1018 Chapter 18 •Two-Ports
32. Consider rhe circuit o f Figure P i 8.32. output o f the two-port to the load.
Suppose Vy(0 = 20 V 2 co.s(w/);/(0 V, = 1 (b) What is the maximum average power
Q, = 1 Q, and the z-parameter matrix is absorbed by = 2 when
/^(/) = 2>/2 cos(2/) A, a = 2 , and/?Q
0.5 0.5
s+ = 50
s s
Z=
0.5 0.5 b :l
,v +
aR„
respectively.
(a) Compute the input impedances, Z -^2
andZ;„,.
(d) Find the power gain of the circuit, i.e.,
(b) Compute the voltage gain, VgJVj-
the ratio
(c) Compute the power gain,
p _ ^ o u t^ o u t
p _ ^ o u t^ o u t S o in - ..
vi/,
37. This problem shows
' ^ ^ equivalent circuits for a pair of coupled
put impedance of the amplifier circuit.
inductors using z-parameters. Each equivalent
I. I, L circuit consists of one ideal transformer and
two inductances. The analysis of a coupled
circuit with the use of such equivalent circuits
'len is very often more illuminating than writing
and solving simultaneous equations.
(a) Find the z-parameters of the two-port
Nj of Figure PI 8,37a.
(b) Show that two-port N j of Figure
Figure P18.35 P I 8 .3 7 b has the same z-parameters as
N j.
CHECK: m = 2 kii and ^y = -9.8147 (c) Show that two-port N3 of Figure
''s
P18.37C has the same z-parameters as
36. Consider the cascaded two-port in Figure N,.
P I8.36, in which = 10 Q, /?| = 20 Q, and Rj^ (d) Use the equivalent circuit N2 of pan (b)
=4Q . and the properties of an ideal trans
former to find 0)^, the bandwidth, and
Vi„ max
approximate values for the half-power
frequencies for the coupled tuned cir
cuit of Figure P18.37d.
Figure P I8.36
10 2 0 Chapter 18 •Two-Ports
1 :n
(d)
(a)Two-port N,
Figure P I8.38
(b) Two-port Nj 1 :n
k y ~ L ,: >/l .
o -fY Y \
(1-k^) L,
NT N2 Ideal Transformer
Figure P I8.37
(a)
/i-PARAMETERS
A A
111I
(0
Chapter 18 •Two-Ports 1021
2000 0
H ,=
50 0.05x10-3
nooo 0.966 ■
Ho =
Figure P18.41 -51 0.8x10-3
42. The /f-parameters of the two-port of Figure (a) Find the input impedances of the
W P18.32 are = 250 SI, h -^2 - 2-5 x 10“^, />2 i stages successively, starting from the
= 125 , h22 = 2.25 mS, -2^ = 1 kH, and = load end.
500 a (b) Find the output impedances of the
(a) FindZ.„ andZ^^,. stages successively, staning firom the
source end.
(b) Find the gain = (c) Find the overall voltage gain,
^ _ ''out
(c) Find the power gain of the circuit, i.e., C .- — .
the ratio
(d) A 1 pF capacitor is inserted in series
p - with Z^ to prevent dc voltage in the
V,/, ■
power supply (not shown in the dia
(d) Suppose a capacitance of 5 |AF is con gram) from entering the signal source.
nected across Dort 2, Suppose Because of this capacitance, low-fre
V j(/ ) = 10V2cos(400f) V. FindKjW quency signals will be amplified less.
in steady state and the average power Determine the frequency (in Hz) at
absorbed by Z^ = 500 Q. Hint: Obtain which the magnitude
the Thevenin equivalent circuit seen by
^out - 0.707 X Max Value
the Z^ and C combination.
O
1022 Chapter 18 •Two-Ports
p y s/ V -o - - 0- Figure PI8.46
2-port 2-port CHECK: />2i = 200, = 1800
N1 N2 •L Q ,Z }„ = 2 k Q
G, = -^ = 100;
and (iii) the source resistance is Z^ = 8 Idi. when a source is applied to port 2.
(ii) There must be maximum power trans
fer from the amplifier output to the
Chapter 18 •Two-Ports 1023
load under the condition that - (b) Verify that the /-parameters o f the cir
800 Q.. cuit o f Figure PI 8.49b arc
..... V. 25 1 O'
(ill) _L = — T =
V, 26 >2 1
Vo
(iv) The voltage gain — = -1 0 0 . (c) Compute the /-parameters o f Figures
P I8 .4 9 c and d.
Given these specifications;
(a) Compute Z/j 2- -o o-
(b) Compute ^^id the turns ratio
a.
(c) Compute//jj.
-O O-
Compute the input impedance (a) (b)
(d) Compue *2 ,
(e) Compute the ratio o f the power deliv -o o-
ered to the load to the power delivered
to Rjj
-o o-
(C ) (d)
Figure P i8.49
ANSWHRS:
■1 2i
: (d)
>2 1 Z.
2 xlO -’ 0.01 n: 1
AN SW ER: l/?,/l =
4W .3S 3.6538 x lO " '
(in standard units)
(a)
n: 1
t-PARAMETERS
4 9 .(a) Verify that the /-parameters o f the cir
cuit o f Figure P i 8.49a are
(b)
T =
0 1
1024 Chapter 18 • Two-Ports
n: 1
c^ JY Y Y -6
4H
(0
Figure P i8.50
Figure P I8.53
51. Find the /-parameters o f the ner\vork in
Figure P I 8.51. 54. The rwo-port o f Figure PI 8.54 is described
by r-parameters.
4; 1
2H
(a) Derive the input impedance relationship
^ll^L +^12
ti\^L ^22
-4 0 ■ t~>2^s hi
’1 2.v^ -4 -S.v ■ ^out ~
1
ANSWlUl:
0 — 0 1 0 -0 .2 5
4_ V,
(c) Derive the voltage gain Gj.] = — .
Kv
5 2 .(a) Find the r-parameters o f tiie cascaded
network in Figure P I 8.52.
(d) Derive the voltage gain = Ir.
(b) If a 14 volt source is applied at port 1,
^1
compute Vjit)-
7 12
ANSW l-R: (a) 55. Suppose a t\vo-port has both 2-parameters
4 7
and /“-parameters. Compute the r-parameters in
terms of the z-parameters. Hint: Rewrite the z-
53. Use the results o f previous problems to parameter equations in the form
obtain the r-parameters o f two-ports A, B, C,
and D in Figure P I 8.53. Then use matrix mul -h
Ml = Mo
tiplication to obtain the r-parameters o f the Vi
overall two-port. and invert the appropriate matrix to obtain
<•21 -21
1 o.
'•21 -21
Chapter 18 • Two-Ports 102S
(a)
PARAMETER CONVERSION
AND INTERCONNECTION OF
TWO-PORTS
2-port
58. A two-port N has known 2-, y-, and h-
N1
parameters. A new two-port is formed by
adding one single impedance Z or admittance
(b) Y to N in various ways. Prove the following
I. relationships bervveen the old and new -rwo-
'
w
port parameters.
+ L
2-port (a) If Z is connected in series with port 1
V,
- N1 o f N, then Zn = 2Ml
,, + Z and
^^\\,new = + Z. Other z- and h-
parameters remain the same.
(0
(b) If Z is connected in series with port 2
o f N, then 2^2 n rtv = -2=22 +
Figure P I8.56
parameters remain the same.
(c) If K is connected in parallel with port 61. Compute the^-parameters of the two-port
2 of N, then =^22+^ of Figure Pi 8.61.
^21,new = ^22 + ^ ■ ^ther y- and h-
parameters remain the same.
(d) If Y is connected in parallel with port
1 of N, then;»„ + K Other
^parameters remain the same.
r )
60. Assume standard labeling in Figure P I8.60, 63. Assume standard port labeling for the cir
in which the inner two-port labeled N has «- cuits in Figure P I8.63. Find the z-parameters
parameter matrix of each of these circuits. For the interconnec
T4 5' tions shown in Figures PI8.63c and d, when do
Q. the overall ^-parameters equal the sum of the
3 4
individual two-port parameters? When this is
(a) Find the /-parameters of the overall not the case, explain why not.
two-port N* and then find the z- o-^ys/ V —o o------------ 1------------o
parameter matrix. 1n 1o
(b) Now suppose the ports are connected
1o
to current sources, in which case /j(/)
= 12(f) = I5u(t) A. Assuming zero ini (a) (b)
tial conditions, find and
What is the response if ij(t) = = o —\/\y\/—I— —o o— —I—s/X/v—o
5u(t) A? Hint: Use linearity. 10 ) 10
10
2F
port 1 port 2 1O
------- •-------
N‘
10 < 10
O O—n / \ / V ^ O
(0 (d)
Figure P I8.60 r ^
Figure PI 8.63
Chapter 18 • Two-Ports 102'
yYY\
160 1 H Ij
0.5 F
A '
1# m \
Figure P I 8.64 "4 -1 “
Y=
.-2 1 _
2 1
Z = a.
0 -2
Figure P I 8.65
Figure P I 8.69
[3.2 l.S
A N SW ER: Q.
!' //■Im'ii- ~ (^)
- 0 .8
72. Two rwo-ports and N^^are connected in
70. The rwo-port is connected to a 1:1 ideal series to form a new two-port N , as shown in
transformer to form N a *. Then N^*
a and N 0/ are Figure P I 8.72. Find the new z-parameters.
connected in series to form a new rwo-port N
as shown in Figure P I 8.70, in which /?j = 2
/?2 = 2 Q , /?3 = 3 and R^ = 1Q .. Find the z-
parameters of N .
Figure P I 8.72
mine rhe voltage and power gain of A N SW ER: (h) /'| ^^{[) = 2.505 a)s(5()0r -
rlie overall c\vo-port. 30.15“) V
(cl) Plot the frequency response o f the ^ .. , . • r t- n,o-rc •
r ■ r r ^ r Consider the circuit or Figure P I S . 75, m
ampliher as a function ot / = 271(0 for jn
^ ^ , ,, . ^ . which u = 0.2, /?, = 1 Q, and R . = 4 Q.
0 < /< 20,000 kHz using Spice. /\ c l i c • l
° ^ (a) For what value or is the rwo-port
reciprocal?
(b) For the value of found in part (a),
compute tlie circuits z- and />parame-
ters.
i,(t)
■ 7.5 5.0 ■ ■ 1.25 1.25
v/t) 2-Port vjt) (b)
5.0 4.0 -1 .2 5 0.25
—•
(a)
;c' r,: r) = 10 co.s(200(); - n “) \’.
v,W 2-Port
(b)
Figure P i 8.74
1030 Chapter 18 •Two-Ports
r>
R.
yyy\
L
Figure P I8.76
R
r\
R.
can be derived from another with
Rc a
n
C H EC K : L = 3 .3 4 6 6 H or I = 0.8 9 6 6 H in
series R^ with AT= 2
n
C H A P T E R
In a stereo system, a power amplifier is connected to a pair of speakers. The most common type of speak
er system consists of one or more drivers enclosed in a wooden box. The amplifier feeds a signal to each
drivers voice coil. The voice coil sits within a field produced by a permanent magnet and is attached to
a heavy paper or plastic cone. When excited by a current from the power amplifier, the coil interacts with
the magnetic field of the permanent magnet and vibrates. The coil pushes and pulls the speakers cone,
which, in turn, proportionately moves the air, making sound waves.
The high and low frequencies that make up music place opposite requirements on a loudspeaker.
A good low-frequency speaker should be large, in order to push a lot of air. A good high-frequen-
cy speaker should be light, in order to move back and forth rapidly.
A rwo-way speaker system consists of a small, light tweeter to handle the treble signals and a large
woofer to handle the bass. A better system is a three-way system, with a third, midrange speaker
to handle the f-requencies in the middle. The magnitude response o f a typical rwo-way system is
shown in the following figure.
l^O’ ltoa;
U
II
OI
:= CT
fO 'O
e
k.
E
O C
C OJ
1032 Chapter 19 • Principles of Basic Filtering
A crossover netiuork (low-pass or high-pass filter) separates the frequencies so that the woofer
receives the low-frequency content of the music and the tweeter the higher-frequency content. In
the magnitude response plot illustrated in the figure, both curves have the same 3 dB frequency
at 2000 Hz. This frequenq’’ is called the crossover frequency. This chapter explores the basic design
principles and realizations o f low-pass and high-pass filters. Some band-pass filtering is also dis
cussed. Some simple crossover circuits are considered in the problems section.
CHAPTER OBJECTIVES
1. Introduce the meaning and (brickwall) specification of low-pass filters in terms of dB
loss.
2. Set forth the maximally flat Butterworth magnitude response and associated Butterworth
transfer function.
3. Present a step-by-step design algorithm for finding the filter order and associated
Butterworth transfer function.
4. Present basic passive and active circuits that realize a Butterworth transfer function.
5. Set forth the properties of the Butterworth transfer function.
6. hitroduce high-pass filter design through frequency transformation.
7. Present basic passive and active circuits that realize a Butterworth high-pass filter.
8. Introduce an algorithm for the design of a band-pass filter using frequency transformations.
SECTION HEADINGS
1. Introduction and Basic Terminology
2. Low-Pass Filter Basics
3. Butterworth Solution to the Approximation Problem
4. Butterworth Passive Realization
5. Active Realization o f Low-Pass Butterworth Filters
6. Input Attenuation for Active Circuit Design
7. Properties of the Butterworth Loss Function
8. Basic High-Pass Filter Design with Passive Realization
9. Active Realization o f High-Pass Filters
10. Band-Pass Filter Design
11. An Algorithm for Singly Terminated Butterworth Low-Pass Networks
12. Summary
13. Terms and Concepts
14. Problems
Switched capacitor networks/filters, an idea introduced in Chapter 13, are something of a hybrid
between analog and digital filters. Both are beyond the scope o f this chapter, which takes a cir
cuits viewpoint on some basic analog filtering concepts and techniques.
Analog filters process the actual input waveform with circuits composed of discrete components
such as resistors, capacitors, inductors, and op amps. Analog filters are of two types, passive and
active. Passive analog filters are composed only of resistors, capacitors, and inductors. Active ana
log filters consist o f resistors, capacitors, and op amps or other types of active elements.
B a s ic T erm in olog y
A filter is a device (often an electrical circuit) that shapes or modifies the fi^quency content (spearum)
of a signal or waveform. We represent a filter by a transfer ftmction H{s) whose frequency response is
The gain, gain magnitude, or frequency response magnitude is |//(^)|. If is normal
ized so that its maximum gain is 1, then the gain in dB (decibels) is Gjg{vi) = 20 log]Q|//(/ci))|. In this
chapter we will assume that the maximum value of |//(/o))| has been normalized to 1.
An important frequency called the cutofF frequency (or frequencies), also called the half-power
point (or points), is that frequency, denoted O)^ for which
2 1 1
H (j(Oc) = 2 ’ '
Since power is proportional to voltage squared or current squared (gain squared) for a fixed load
resistance, we arrive at the terminology of half-power. This is also called the 3d B down point
O ' because = 10 logiol//(/copP = 10 logjo(0.5) = - 3 dB.
Often, design specifications for a filter are expressed in terms of attenuation or loss rather than
gain. This results in several definitions that are dual to the gain-related definitions above. The loss
fiinction, denoted H (/w), is the reciprocal o f the transfer ftmction. Thus, the attenuation, or loss
magnitude, is
H (jw ) '
HiJco)
It follows that the filter loss or attenuation in dB is
With these definitions established, the chapter will examine low-pass (LP), high-pass (HP), and-
band pass (BP) filters, leaving the study o f band-reject filters to higher-level texts. A low-pass fil
ter is a device (typically a circuit) offering very little attenuation to the low-frequency content
(low-frequency spectrum) of signals while significandy attenuating (blocking) the high-frequency
content o f those signals. High-pass filters do the opposite: they block the low-frequency content
and allow the high-frequency content of a signal to pass through. Finally, band-pass filters, as
described in Chapter 16, allow a band o f frequencies to pass while significantly attenuating those
outside the band. Interestingly, the general practical design of all such filters, especially for the
high-order case, is generally based on a low-pass prototype design; the low-pass prototype is trans
formed into a high-pass or band-pass type using a frequency transformation.
H )34 Chapter 19 • Principles o f Basic Filtering
As a firsr step in exploring the design o f such filters, vve describe Butterworth (low-pass) transfer
functions. Butterworth was the name o f the English engineer who first developed this special class ^
o f transfer functions in his paper “ O n the Theory of Filter Amplifiers.” ' The next step is presen
tation of an algorithm that adapts these basic Butterworth transfer functions into ones that meet
a given set of L P filter design specifications. Once the transfer function is known, an engineer
must implement this transfer function as a passive or active circuit. This chapter will also outline /—
some techniques for generating passive and active reaHzations.
As mentioned, a low-pass filter allows the low-frequency content o f a signal to pass with little
attenuation while significantly blocking the high-frequency content. The immediate question is,
why use a low-pass filter? One possible answer is that the noise in a noisy signal often has most of
its energy in the high-frequency range. For example, so-called white noise has a constant frequen-
q- spectrum. Hence, low-pass filtering a sinusoidal signal corrupted by white noise will generally
result in a “cleaned-up” information signal, as illustrated in Figure 19.1. In Figure 19.1, the thick-
ness o f the curves represents the infiltration of noise. By using the low-pass filter, the curve is
“sharpened” by reducing the high-frequency noise content.
The suspension system o f a car is a low-pass filter: slow, rhythmic (low-frequency) road variations
are permitted, while the effects o f chuckholes and bumps (high frequencies) are “ filtered out.”
Lowpass Filter
Whenever LP filters are needed, an engineer must provide design specifications. Historically this
was done using a low-pass filter (brickwaJl) specification, as illustrated in Figure 19.2. Two pairs
of numbers (d)^,, and (to^, characterize this brickwall specification, where
1. equals the pass-band edge frequency,
2. 0 < 0) < is called the pass-band,
3. niaximum dB attenuation permitted in the pass-band,
4. to^ is the stop band edge frequency.
5. 0)^ < CD d e f i n e s t h e s t o p b a n d ,
The shaded region in Figure 19.2 represents a brickwall. The attenuation of the filter in dB, i.e.,
/l(co) =-20 log,Q|//(/w)| = 20 log,o|/^(/w)|, must reside outside the shaded region. Finding a (nor
malized) filter transfer function that meets the brickwall specs is callcd the approximation prob
lem. Once the approximation problem is solved by identification o f the proper normalized trans
fer function, the next step is to construct a circuit realization of the normalized transfer function
and then to frequency-scale to obtain the proper pass-band edge frequency, and finally to magni-
tude-scale to obtain the proper impedance levels.
The simplest technique for solving the approximation problem is with 3dB normalized
Butterworth transfer functions whose squared magnitude responses arc given by
H ( jc o )
( (ti ) (19.2)
1+
U c/
(tip
where is the 3 dB down point, or cutoff frequency, o f the filter, and e is a to-be
specified constant. As mentioned earlier, the term “3 dB down point” arises here because, for all w,
( coY
10 lO g lQ 1+ = 10 1og,ol2j = 3 dB
i.e., there is 3 dB of loss at (O = co^ W hen e = 1 in equation 19.2, the pass-band edge fre
(0
quency and cutoff frequency coincide. If we now define the normalizedfrequency as Q =
CO,.
the magnitude response of equation 19.2 becomes
^3dBNLP ( J ^ ) ~ =
1 + Q^ ( 1 9 .3 )
1036 Chapter 19 * Principles o f Basic Filtering
Equation 19.3 denotes the wtli-order 3 dB normalized Butterworth magnitude response. The words
“ 3 dB normalized” refer to the hict that at Q = 1, the loss is 3 dB (the gain is -3 dB). Remember
that the actual filter transfer function depends on a proper choice of or f , which we will clar
ify shortly.
Before proceeding further, we ask a critical question: Does this kind o f representation make sensed To
answer this question, note that the dB loss of equation 19.3 is
Plotting this function for various ;/s as a function of normali/xd frequency, Q, for a normalized
stop band edge frequency, = 3.5, = 3 dB, and = 20 dB yields the polynomial curves
in Figure 19.3. Clearly, for h i g h e r w e can have higher values o f These cur\'es can be made
to lie outside the brickwalls o f Figure 19.2 and hence can validly be used to meet a L P filter spec
ification.
-XD
c
Normalized Frequency, Q
FIGURE 19.3 Plot of the normalized Butterworth magnitude responses for n = 2, 3,
Chapter 19 • Principles o f Basic Filtering 1()3‘
From Figure 19.3, vve see that as n increases, the pass-band magnitude response becomes flatter
and the transition to stop band becomes steeper. This property suggests that there is an appropri
ate value of w to meet a given set of brick\vall specifications (ci3^, and (o)^, The idea
is to use equation 19.2 and these specs to determine a proper value for the filter order «, a prop
er cutoff frequency (o^., and the normalized transfer function.
1 -I-
(^ s\
(19.5a)
2: 1 0 ^ - 1
/'" p V _ 1
(19.5b)
Dividing the left and right sides of equation 19.5a by the left and right sides o f equation 19.5b
(this maintains the inequalit)' o f equation 19.5a) yields
(w X
(19.6)
•«gio
na (19.7)
( COs^
lo g lO
(Or, J
\ ^ P
Thus n can be any integer satisfying inequalit}' 19.7. Usually one takes the smallest such n.
103S Chapter 19 • Principles o f Basic Filtering
(M,
LOcjnm CO. = COc,max
2 ,^ A A min __ j (19.8)
To derive the range given in equation 19.8, we reconsider equations 19.5. Taking the 2«th root of
both sides of equation 19.5a and solving for yields the right side of equation 19.8. On the other
hand, taking the 2«th root of equation 19.5b and solving for to^ yields the left side of equation 19.8.
E xercise. Verify the mathematical details in the above paragraph on the derivation of equation
19.8.
EXA M PLE 19.1. Suppose we are given the brickwall specs = (500 Hz,2 d B ) and
- (2000 Hz,3 dB). Find (a) the minimum filter order «, (b) and co^ (c) and
the normalized L P squared magnitude function, and then (d) plot the magnitude responses for
the cases where O)^ = and a)^ = over the frequency range 0 :s I n x 2500.
Step 1. Find the minimum filter order n from equation 19.7. Here
/
1 0 ^ ^ - _1 ' ( l O - '- l ^
iogio log 10
_ I 10° “ - 1
- = 2.68
1 •ogio(4)
•ogio
\ P/
implying that the minimum filter order is 3. This can also be accomplished in M A T L A B as follows:
»n=buttord(wp,ws,Amax,Amin,V)
n =3
CD, I n X 500
COc \ m m = 2;r X 547 = 3435 rad/s
2 ^ 1 A "max _ I
and
2 1
^ (> )
(O
1+
co^
Chapter 19 * Principles o f Basic Filtering 1039
Step 4. Plotting this function over the frequenc)' range 0 ^ 2jvf^ 2 ttx 2500 can be achieved with
the following M A T L A B code. The resulting plot is displayed in Figure 19.4.
»f= 0:4:2500;
»hl = sqrt(l. ./(I + (2*pi*f ./wcmin).^6));
»h2 = sqrt(l. ./(I + (2*pi*f ./wcmiix),^6));
»plot(f,-20*logl 0(h 1),f,-20*logl 0(h2))
»grid
Frequency in Hz
Observe that with the choice o f = (jo^ the magnitude response curve passes through the
pass-band edge frequency with loss exactly equal to leaving more than adequate loss at (O^.
O n the other hand, with the choice of co^ = co^ the magnitude response curve passes through
the stop band edge frequency with loss exactly equal to with less than the maximal allowable
loss at . In practice, to allow for element tolerances, one would choose O)^ somewhere in
between.
Note that rvvo zeros are in the right lialf complex plane and two in the left. The two in the left
half plane determine the Butterworth loss function. For illustration we provide the zero plot of
Figure 19.5. Again, note the symmetry o f the zeros with respect to the imaginary axis.
0.2 i/ !
* J
0
I**•
•
-0.2
-0.4
/ •
i \ i
-0.6
: ! X
-0.8
i !
-1
1 1
-1 -0.5 0 0.5 1
Real part
As per our earlier statement, from stability considerations we pick the left half plane zeros to form
i-e.,
W e conclude that
r + y[2s+\
For higher-order cases, the procedure of Example 9.3 can again be used. The results are given in
Table 19.1, which presents the 3 dB normalized Buttenvorth loss functions. In practice, one never
computes the actual transfer function o f a filter that meets a set of non-normalized specs. Rather,
one realizes the normalized loss or transfer function and then magnitude- and frequency-scales to
obtain the proper circuit. This process is illustrated in the next section.
1042 C h ap te r 19 • Principles o f Basic Filtering
Lo s s F u n c t i o n s , « = 1, 5.
n ^3 d B N L p (^)
1 S+ 1
2 S~ + y / l s + 1
3 {s + l ) ( j ^ + s + \ ) = + 2 r + 2s + 1
{ r + 0 . 7 6 5 3 7 s + 1)(^2 + 1 . 8 4 7 7 6 i + 1 )
4
= / + 2 .6 1 3 1 P + 3 .4 1 4 2 j2 ^ 2 . 6 1 3 b + 1
{s + l)(j^ + 0 . 6 1 8 0 3 ^ + l ) ( j ^ + 1 . 6 1 8 0 3 j + I)
5
= P + 3 .2 3 6 1 / + 5 .2 3 6 1 P + 5.2 3 6 1 ^ 2 + 3 . 2 3 6 b + 1
E X A M P L E 19.4
The circuit of Figure 19.6 must realize a Butter\vorth filter meeting the low-pass brickwall speci
fication of Figure 19.7. In the final design we require R^= 100 Q.
R.
LC
out
Network
-o
F IG U R K 19.6 Structure of a filter driven by a practical source.
40 db
2db
>• f(Hz)
100 Hz 1000 Hz
So lu t io n
From the brickwall specification of Figure 19. 7 , we have (oj^ = 2 0 0 k , = 2 d B) and (o)^ =
20007T, = 40 dB).
_ 1
loglO 1 0 ® - _ 1
- = 2.1164
2 0 0 0 .T
•o g lO
200 jt
COn
CO.
(O, 2 0 0 JZ
0j^=(0 c.min = 687 = 21 SJjT rad/s ec
Step 3. Looking up the third-order normalized Butterworth loss fiinction and inverting to obtain
the transfer function yields
H MBNLP
(19.12)
is)
. Vouti^)
Step 4. Choose a candidate passive circuit and obtain its transfer function,
W e wind up with the circuit o f Figure 19.8. Note that we will adjust the Vin(s)
source resistance later with magnitude scaling.
/VY\
L
in c
V. V
W e obtain the transfer function using nodal analysis. At the top node o f C j,
Vc - - ! - K u, = v,„ (19.13a)
Ls
A t the o u tp u t node.
1044 Chapter 19 • Principles of Basic Filtering
Solving equation 19.13b for Kq and eliminating K q from equation 19.13a yields
• . 1 , . Cl +
— = 2 , ------ = 1, and — ---- = 2
.
C| LC^C') LC\C~)
2
Clearly, Cj = 0.5 F and = 1 implies LC^ = 2. Then
LCi
C 1+ C 2 0 .5 + C 9 1c r- 1r ■^11
— ---- = 2 => C ? = 1.5 F and L = — H
LC ,C 2 1 ^ 3
Step 5. Frequency-scale the circuit to obtain the desired, a n d magnitude-scale to obtain the specified
source resistance o/lOO Q. W e frequency-scale by Kj-= co^ = = 687 = 218.7TT and magnitude-
scale by = 100. Thus, = 100 /?^ = 100 Q
E xercise. Reconsider Example 19.4. I f is to be 50 Q and O)^ = recompute the final ele
ment values.
.\ N S W F R S : 7.38~2 iiK 22.162 uF. and 0.0492 H
In the above example, there is no load attached to the filter. A situation in which there is no source
resistance but a load resistance is given in the problems at the end of the chapter. In addition, there
is a problem containing both a source and a load resistance.
Here we have used a coefficient matching technique that is manageable for orders 1, 2, and 3. For
higher orders the method is unwieldy and we must use the methods o f nerwork synthesis studied
in other courses or resort to normalized filter tabulations given in filter handbooks.
Chapter 19 • Principles o f Basic Filtering 1045
K. K
H?iclBNLp{^) = —-- — = H„{s)Hi,{s) (19.15)
S~ + .V + 1 i’ + 1
where = 1. The gains and /iQ, are present in equation 19.15 to allow the individual stages
to have dc gains different from 1, as is sometimes necessary. The overall transfer function is real
ized as a cascade of ^ ^
= — — -------- a n d H ij { s ) =
s~ + .s -I" 1 .v+1
with the constraint = 1. All transfer functions beyond first order require one or more sec-
ond-order stages in cascade for their realization. The use of a basic second-order active circuit is
key to the realization of any of these second-order sections.
To derive note that the properties of an ideal op amp force the voltage across 7?^ to be
Using voltage division,
■Vo, u t
o r equivalently,
1046 Chapter 19 • Principles o f Basic Filtering
This observation permits us to simplify the diagram of Figure 19.9 to the equivalent circuit of
Figure 19.10.
Construction o f H { s ) now proceeds by nodal analysis. At node a— the node between C , and
Ri,
1 ( 1
(19.16)
Similarly at node b,
(19.17)
V ( 5)
Eliminating and solving for the ratio ^ yields the transfer function of the Sallen and
Key low-pass filter:
K
K ,A ^ ) R\R2C\C2
(19.18)
Vi„U) 2 M 1 ]-K \ 1
S + 1 ------------ h — H------- -I H------------------------
U ^ lC , R 2 C 2 ) R\Ri C iC 2
w h e r e / r = l / , , / l / = 1 + /? ^ //? ^
K
H [s) =
s^ + — s+ \ (19.19)
Q
The design template derives from equating the denominator coefficients of equations 19.18 and
19. 19. The dc gain can be adjusted using other methods such as input attenuation (to be described
later). In particular, equating the denominator coefficients o f equations 19.18 and 19.19 requires
that
Chapter 19 • Principles o f Basic Filtering 104'
1 1 1- K
(19.20a)
Q /?,C, /?2Ci R 2C 2
and
(19.20b)
R\R2C\C2
The solutions to equations 19.20a and 19.20b are not unique because there are t%vo equations in
five unknowns. This means that we can impose up to three additional constraints to produce dif
ferent solutions. Different solutions produce the proper filtering action, but have different behav
iors in terms o f the sensitivit)' of the frequenq' response to variations nominal resistance and
capacitor values. Also, different designs have different ratios of clement values that may have prac
tical significance.
One popular and robust design is the Saraga design. For the (normali/ed) Saraga design, the three addi-
R. Q
tional constraints imposed on the solution of equations 19.20 are C, = 1, Cj = >/3Q, and — = ,
R] v3
which were chosen to minimize certain sensitivities in the circuit performance. Solving equations
19.20a and 19.20b for R ,, /?2, and K using these additional constraints yields
C 2 = IC \ = S q . R] = Q~\ R2 = 3
I/V , K = 4 / 3 , Rs = R^/3. The final circuit realization of
equation 1 9 . 1 9 is given in Figure 1 9 . 1 1 . Again we note that the dc gain K = 413 will need to be
adjusted by other means, to be described. Frequency and magnitude scaling are necessary to
achieve proper cutoff frequencies and impedance levels.
F IG U R F. 19.11 Normalized Saraga design of Sallen and Key circuit for realizing the
transfer function of equation 19.19.
W ith this design in hand, we now redo Example 19.4 using an active realization.
1048 Chapter 19 • Principles of Basic Filtering
K K
^7,clBNLp{^) = -------------- ^ ------- - H^,{s )H ij{ s )
S~ + .V -I- I 5 + 1
where = 1. Later, we will frequency- and magnitude-scale to obtain proper cutoffs and
impedance levels. For
S~ + .S• + I
S~ + .V +
Q
we have Q = 1. The Saraga design element values are C-, = 1 F, Cj = >/3 F, /?, = I Q, 7?2 “ ^
Q., K = K j = 4/3, and These element values realize the transfer function
.V“ + .V+ 1
To obtain the correct realize the transfer function
K,, _ 3/4
(19.21)
s -t-1 .V+
Observe that the choice o f /Q, makes A'/Q, = U the correct overall dc gain. The transfer function
of equation 19.21 is realized with the leaky integrator circuit o f Figure 19.12, whose transfer func
tion set equal to H(^{s) is
1
R ,C 3/4
Hl,^ar(s) =
1 5+1
.y-l-
R'yC
To complete the normalized design, choose C = \ V, Rj - \ and /^| = 4/3 Q.
R,
Combining the second-order and first-order sections o f the normalized design produces the cir
cuit of Figure 19.13.
In the last step o f our design we frequenc}'-scale by /y = 218.7;r. If the smallest capacitor is to be
100 nF, then
c o ld _
K,n = = 14.555 x 1 0 ^
K jC ,,,,, 2 I 8 . 7 ; r l O “ ^
Hence the final parameter values for the second-order section are
Q = 100 tiF, q = 173.2 uF, = 14.56 kQ, Rj = 8.4 kQ, R^ = 30 kQ, ami Rf^ = 10 K2
Similarly, for the first-order section C = 100 nF, / ?2 = 14.55 and R^ = 19.4 kQ. This leads to
the final circuit design given in Figure 19.14.
14 .5 5 kO
173.2nF
1050 Chapter 19 • Principles o f Basic Filtering
This design assumes the fiirer is driven by a volrage source with a very small resistance; otherwise
the source impedance must be considered part o f the front end series resistance. Also, we could
have magnitude-scaled each section separately (but chose not to).
K C\C2
s~ + — .V +1 + -- .V + -------- (19.22)
Q C, C ,C 2
Matching the denominator coefficients produces Cj = 2Q and C j = 1/2Q. Since the dc gain with
K = 1 is 1, there is no need to modify the overall circuit gain. Figure 19.15 illustrates design A.
C, = 2Q
Notice, however, that for Design A the ratio C^/C, = 4Q^ is 400 for a circuit with Q = 10. Such
a large variation may be undesirable in a practical circuit.
= Q L2. This design is computed using the constraints = I Q (yielding /f= 2), C, = 1 F,
and equal time constant, = ^2^2' these choices,--------- = 1implies that /?| = 1
and /^2 - Q- Thus, /?jCj = 1 = RjC2 implies that C , = — . Here the maximum parameter ratio is
= C j/ G ) = Q, w h ic h is better than the 4Q^ ratio of D esign A . Figure 19.16 shows the result
in g circu it.
Chapter 19 * Principles o f Basic Filtering lO r
C =1
F IG U R E 19.16 Design B for realizing the normalized transfer function of equation 19.19,
assuming standard units.
In design A of the previous section, the gain o f the filter is A"= 1. In design B, the filter gain is K
= 2. In Example 19.5, the gain of the Saraga design {K = 4/3) was corrected by the subsequent
first-order section. I f the filter were second order with a dc gain of 1, then the circuit gain would
need to be modified for both Design B and the Saraga design. In order to reduce a high gain,
maintain the filtering properties, and keep as many parameter values as possible at their original
design values, we use a technique known as input attenuation. In this technique, the front-end
resistor is replaced with a voltage divider circuit, as illustrated in Figure 19.17.
(b)
F IG U R E 19.17 Illustration of input attenuation concept, (a) Original circuit,
(b) Circuit with input attenuation.
10 ^ 2 Chapter 19 • Principles o f Basic Filtering
Figure 19.17a represents the original active network, while Figure 19.17b represents the modified
network. For the filtering characteristics to remain invariant, the impedance looking into the
dashed boxes must remain at /?, for both circuits. Thus, the parallel combination of and
must equal /?j, i.e.,
' + (19.23)
^4
a =
/?3 + /?4 (19.24)
=^ and
■ a ^ \ -a (19.25)
Thus, one can reduce the gain o f the Sallen and Key low-pass circuit via the simple technique of
input attenuation.
As mentioned earlier, the Butterworth loss function has a maximally flat pass-band response in the
sense that as many derivatives as possible o f the loss magnitude response are zero at Q = 0. This is
often termed m axim ally flat. In the case o f the ;zth-order Butterworth loss function, it is possible
to show that
at Q = 0 and
cf
H (jQ )\ = 0 .5 { 2 n ) \ * 0 ( , 9.27)
dQ^’’
at Q = 0. This is consistent with the notion of being maximally flat. To verify equations 19.26 and
19.27, observe that if « 1, then
Equations 19.26 and 19.27 follow after differentiation of equation 19.28 2u times and evaluation
of each derivative at Q = 0.
Several closing remarks are now in order. First, the cutoff frequency is the half-power point, or the
3 dB down point. The terminology follows because |MyO)| = 1 (the gain at dc is unity) and
Chapter 19 • Principles o f Basic Filtering 10^ 3
H (j\ ) = - ^ = 0.707,
V2
yielding half power for normalized Butterworth gain functions H{s). Second, for large Q,
-n [20 logjQ(Q)]
indicating that the gain rolls off with a slope proportional to the number of poles, n, o f the gain
function. Specifically, the slope equals -20« dB per decade. These statements are also valid if to
replaces Q.
High-pass filters invert the frequency characteristics of low-pass filters. A high-pass filter is a
device— usually a circuit— that significantly attenuates the low-frequency content o f a signal while
passing the high-frequency content with minimal attenuation. Figure 19.18 illustrates typical
brickwall specifications for a high-pass filter. Here frequencies above the pass-band edge frequen-
cy co^ have little attenuation, while frequencies below are significantly attenuated— precisely
the inverse function o f a low-pass filter. In fact, low-pass and high-pass specifications are related
by a simple inversion of O). In particular, we define a H P (high-pass) to L P (low-pass) frequen
cy transformation as
^ = — (19.29)
This frequency transformation applied to the briclc\vall specifications o f Figure 19.18 yields the
set of normalized low-pass specifications given by Figure 19.19.
10 V i Chapter 19 * Principles o f Basic Filtering
A(co)
min
3dB
A_..-
> (0
0 CO. %
F IG U R E 19.18 Typical brickwall characteristic of a high-pass filter.
A(co)
Given the normalized L P (N L P ) equivalent specs, one finds the Buttenvorth filter order accord
ing to the usual formula, equation 19. 7, reproduced as equation 19.30:
] 1''VniH _ I ( 10
log
clO •o g lO (19.30)
_ 1
( Q ,\
lo2 10
By convention we take the filter order to be the smallest integer n satisfying equation 19.30. For
a passive realization, once n is known, one would realize the resulting N L P Butterworth transfer
Chapter 19 • Principles of Basic Filtering I05S
J_ J_
Cs _ LC
1 1
Ls + /?<• + S + — S+ -
Cs LC
As with the development o f equation 19.8, once n is picked, there is a permissible range of nor
malized given by
Q.
Qc ju i n (19.31)
where Once is chosen, the Butterworth transfer function that meets the N L P
equivalent specs is found as
/YY\
u o .C
V, V
C/Q,
The actual H P transfer function is related to the other “ normalized” transfer functions as follows:
10„
H H p{s) = H N Lpi— \ = H
This is equivalent to doing a special frequenq' transformation on the circuit elements o f Figure
19.21. Specifically
1056 Chapter 19 • Principles o f Basic Filtering
(O.
(19.32)
Cs (0), C oj,
C
A similar substitution shows that an Inductor L with impedance Ls becomes a capacitor of value
Figure 19.22 illustrates the transformation. ^
-O /YY\ ■O
1
L=
Cco
o- fY Y \ -o -o
1
c = Leo.
Hence the L P circuit of Figure 19.21 becomes the H P circuit o f Figure 19.23.
Cnew= —
L CO
F IG U R H 19.23. H P circuit derived from the N L P reali/aiion of L'igure 19.21. Observe that
to.
It remains only to magnitude-scale this circuit to obtain an acceptable passive H P filter. W e illus
trate the above ideas in the following example.
Chapter 19 • Principles of Basic Filtering lOS"
>0)
200n 2000n
Solution 9000 t
Step 1. Compute the equivalent N LP specificatiom . Using Q = the H P to N L P trans
formation, the N L P specifications are easily computed as
(H/. ' ' ■ = 2 dB) and (Q^ = 10, A„,.„ =40 dB)
Step 2. Detennine the Butterworth transfer function. As these specifications correspond to those of
Example 19.3, one computes the filter order n = 3. From Tiible 19.1, the third-order Butterworth
transfer function is
Step 3. Realize the third-order 3 dB N LP transfer fin ctio n . According to Example 19.4, the 3 dB
N L P Butterworth transfer function can be reali7.ed by the passive RLC circuit o f Figure 19.8, with
values assigned as in Figure 19.25.
/YY\ -O
1n +
C^=1.5F
V, V
-y
C =0.5 F
---- c-o
F I G U R E 19.25 Third-order 3 d B N L P filter circuit o f Example 19.4 having 3 d B down point at 1 rad/sec.
1058 Chapter 19 * Principles of Basic Filtering
Step 4. Compute the norm alized 3 dB down frequency. The 3 dB frequency of the actual N L P
equivalent filter is calculated from equation 19.31, using the left inequality:
= 1.0935
1/6
Step 5. Frequency-scale by to obtain the N L P circuit. The N L P circuit o f Figure 19.25 is fre-
quency-scaled by the factor Kj-= 1.0935 to produce the network o f Figure 19.26.
/YY\ -O
1o 1.219H +
V. V
- J '' ■N r
0.4573 F 1.3717 F
t
Step 6. Apply the LP to H P circuit element transformation detailed in Figure 19.22. Converting the
circuit of Figure 19.26 to the required H P circuit yields the network of Figure 19.27.
0.1305 mF
-O
1Q +
V. V
0.116mH
0.3481 mH
13.1 mF
-O
lOO +
V. V
1.16 mH
3.481 mH
-o
F IG U R I: 19.28 Passive circuit realizing the high-pass specifications of Figure 19.24 with 10
source impedance.
E x e r c is e . Use SPIC E to confirm the high-pass characteristic o f the circuit o f Figure 19.28.
Chapter 19 • Principles o f Basic Filtering 10^‘)
The preceding example completes our basic discussion o f passive high-pass filtering.
I
H 3dBNLP {s) =
( i - +l ) |. v“ ij + 2 a’“ + 2.S + 1
One approach to H P design first realizes this 3 dB N L P transfer function as a cascade o f a second-
order active circuit with a first-order circuitin Example 19.5, this resulted in the circuit o f Figure
19.13, redrawn as Figure 19.29.
1O
The second step in the completion o f a H P design is to replace each resistor with a capacitor whose
value is the reciprocal of the resistance; likewise, each capacitor is replaced by a resistor whose value
is the reciprocal o f the resistance value:
R
(19.33)
C
1060 Chapter 19 • Principles of Basic I*iltering
Applying the transformation o f equation 19.33 to the circuit of Figure 19.29 produces the circuit
o f Figure 19.30, with one exception to the rule: the feedback resistors on the second-order section
that connect to the inverting input node are not changed to capacitors; these resistors only set up
a voltage division, so it is unnecessar)' to replace them.
In terms of transfer functions, the operation o f equation 19.33 changes s in the 3 dB N L P trans
fer function to \/s. Specifically,
2
s s
HM BNLP W = H 7.CIBNHP M BNLP X
(^ + 0 ^“ -f-5-f-l
Notice that each section of the transfer Rinction has a (normalized) H P characteristic. For example,
5 5
l i m ------ = 1 and lim ------- = I whereas li m ---- = 0 and lim = 0.
5 —*o c s+ 1
In
-1 5 - 1 (19.34)
and K j- = 5745.9 .
Chapter 19 • Principles of Basic Filtering 1061
^ = 13.05
t<fCne.
13.3 nF
7.54 kO
The circuit o f Figure 19.31 presumes that an ideal voltage source is input to the filter. I f the source
has an internal impedance, this impedance will affect the pole locations and the overall perform
ance o f the high-pass circuit. To circumvent this problem, one may insert a voltage follower cir
cuit between the practical source and the input to the filtering circuit o f Figure 19.31, as illus
trated in Figure 19.32.
Voltage
Follower
Practical Circuit
Band-pass circuits were first considered in Chapter 16, mostly from an analysis perspective. In this
section we will briefly oudine the design o f a B P filter to meet a given set o f brickwall specifica
tions as shown in Figure 19.33.
A Loss(dB)
>► CO
0). 0) 0)
P2
In Figure 19.33 O)^, and 10^2 pass-band edge frequencies while and 0)^2
band edge frequencies. The center of the B P filter is defined as the geometric mean of pass-band
edge frequencies, i.e., = yjojp\(0p2 •
Given the B P brickwall specs, the design procedure begins with a frequency transformation to
convert these specs to equivalent N L P specs, i.e., (Q^ = 1, and (Q^., To illustrate how
to generate these N L P specs we use a frequency transformation as follow's: define a bandwidth
w ith respect to as Z? = 0)^2 “ define two potential N L P stop band edge frequencies
as
^.v/ = (19.35)
Iko,i
for / = 1,2. Given these two numbers, the N L P equivalent specs are
To illustrate the need for the magnitude sign in equation 19.35 and the need for the minimum
function in equation 19.36, let us do a simple example by breaking apart the formula o f equation
19.35 and putting it back together. Suppose = 2 rad/sec and 0)^2 = ^ rad/sec with to^j = 1
Chapter 19 * Principles of Basic Filrcring 1063
_ [-6 i=\
a ; ,, I 6 i=l
[19.2 i= 2
One can view this partial transformation as the generation of two distinct L P filters: one defined
on negative frequency (Q^ = - 6, =-15) and one defined on positive frequency =
6, = 19.2). O f course, we cannot really have two distinct L P filters, one for negative fre
quency and another for positive frequency. O u r mathematics requires that the magnitude respons
es of each filter be symmetric because magnitude is an even function of (o. Thus the only way to
properly interpret the above transformation is that it implicitly generates two distinct low-pass fil
ters, one of which is more stringent than the other. Specifically, the “ negative” filter is really a LP
filter with edge frequencies (Q^, = 6, Q^, = 15), and the positive filter is really a L P filter with
edge frequencies ~ “ 19.2). Observe that the ratio
^,v/
Q^, |3.2 /= 2
differs for the rwo filters. The first filter is more stringent than the second, prompting the need
for a minimum fiinction in equation 19.36. Since our edge frequencies are always specified on the
positive axis and since the magnitude response is symmetric, w'e insert the absolute value signs in
equation 19.35. Finally, in order to have Q^- = 1 we divide by - to^j so that the edge fre
quencies correspond to a N L P filter.
EXA M PLE 19.8. A B P filter has the specs = 26 krad/sec, = ^7 krad/sec, = 30.6
krad/sec, 0)^2 = 32.6 krad/sec, = 3 dB, and = 14 d k Design a minimum-order
Butterworth passive B P filter meeting these specs, assuming the filter is driven by a voltage source
in series with a 10 kQ source resistance.
S o l u t io n
W si =
6.1838e+00 5.0195e+00
»W s = m in(W si)
W s =
5.0195e+00
Step 3. Find a realiztition o f the NLP filter. Since = 3 dB, the first-order Butterworth N L P
has transfer f^unction Hf^j^p{s) = ----,whose passive circuit realization is given in Figure 19.34.
.V+ 1
-O
1Q +
1F
Step 4. Realize the given BP filter. The algorithm to generate the B P filter from the N L P filter is
based on the frequency transformation
.v“ +
(19.37)
Bs
This frequency transformation causes a change in the capacitive and inductive circuit elements o f
the N L P circuit. From equation 19.37, the impedance o f an inductor changes as follows:
( s - + K\ L 1
Ls = —s +
Bs B B (19.38)
\LK I
rhe expression on the right of equation 19.38 is a sum of impedances. This sum, then, is a cir
cuit composed o f an inductor in series with a capacitor as shown in Figure 19.35.
/YY\
L L 3.
B LK
= — .V+ (19.39)
Bs B f
XK)
The expression on rhe right is a sum of admittances. Hence under the frequency transformation,
a capacitor becomes a parallel L C combination as illustrated in Figure 19.36.
C
CK
Applying the transformation of Figure 19.36 to the N L P circuit of Figure 19.34 yields the circuit
o f Figure 19.37, where
------- O
1O +
1 B
B K '^out
------- O
Step 5. Miignitudc-scale to obtain the correct source resistance. W ith = 10"^, we conclude that
= 50 nF and = 20 m H . The final circuit realization is illustrated in Figure 19.38.
Km
-o
I _ BK +
bp “ k"
-o
F IG U R E 19.38. Final BP design.
Bs
/Y gp (.v)-
This transfer function and circuit show a clear similarity to the material developed in Chapter 16.
In contrast to the above procedure, the development o f Chapter 16 works only for w = 1 in the
equivalent N L P circuit, which results in a B P circuit o f order 2. I f the order o f the N L P equiva-
'0 6 6 Chapter 19 * Principles o f Basic Filtering
lent were 2 or higher, rhe B P order would be at least 4 and the background from Chapter 16
would prove inadequate. The next example illustrates how design is done for higher-order circuits.
E X A M P L E 19.9. A B P filter has the specs to^j = 1500 rad/sec, = ^^00 rad/sec, = 3000
rad/sec, to^^ = 4000 rad/sec, = 2 dB, and = 20 dB. Design a passive B P filter meeting
these specs that is driven by a voltage source having an internal resistance of 100 Q.
So lu t io n
Part 1. F in d equivaletn N L P specs. Define K = = 12 x 10^ as the square o f the center fre
quency of the filter. Define the bandwidth with respect to as 5 = co^2 ~ ~ 1000 rad/sec.
W e use the following M A T L A B code to obtain the necessary numbers.
The M A T L A B output is
B = 1000
K = 12000000
W si = 6.5000e+00 4.0000e+00
Ws =4
" T T T ir r T (19.40)
Chapter 19 • Principles of Ba.sic Filtering 1067
This docs not have the correct at Q = 1. W e will adjust this by frequency-scaling the circuit
that realizes ^ circuit that realizes the second-order Butterworth transfer function is
given in Figure 19.39.
yvYV -o
1n L +
'- 0
1
LC
(19.41)
2 1 1
S H----- A + ------
L LC
1
c,w/« ■= 1.1435 rad/sec
- W - \A^mat _ I
L C
in which case =— = 0.61838 H and =— = 1.2368 F.
Thus the circuit that realizes the N L P characteristic with a loss of at Q = 1 or, equivaiendy,
3 dB of loss at Q^= 1.1435 is given in Figure 19.40.
IY Y \ -O
IQ 0 .6 1 8 3 H +
V out
1 .2 3 6 8 F
o , - H},dBNLP
u I 1.3076
n^LPkS) ■ (19.42)
+ 1.617 Ly-f-1.3076
-1- +1
U .1 4 3 5 J \ \ A 435}
Part 3. Realize the passive band-pass filter. To compute the desired B P transfer function, one would
replace s in as follows:
1068 Chapter 19 • Principles o f Basic Filtering
+Wp\COp2
' ^ Bs
One could then attempt to realize the B P transfer function directly. This is numerically unwise
and, fortunately, unnecessary. A simple, numerically sound procedure is to replace each inductor
and capacitor in the N L P realization by an equivalent circuit representing the above frequency
transformation. These equivalent circuits were developed in equations 19.38 and 19.39, and illus
trated in Figures 19.35 and 19.36, respectively. The substitution scheme is repeated in Figure
19.41.
/YTV
L L _B
B LK
(a)
(b)
Given the transformations of Figure 19.41, the following M A F L A B code produces the desired B P
circuit parameter values, assuming the final source resistance is 100 O..
» % Insert L, C, B, K, and Km
»L = 0.61838; C = 1.2368;
»B = 1000; K = 12000000; Km = 100;
» % Compute K1
» K f = sqrt(K)
K f= 3.464 le+03
The final circuit is given in Figure 19.42, where R^= 100 ^2, = 61.8 m H , - 1.35 |.iF,
~ ^Ibp ~
LI bp
-O
R. Cl bp C2bp +
V.
^out
L2bp ^
Active B P design is beyond the scope of this text, but coverage can be found in more advanced
texts.
In 1937 E. L. Norton published a paper in the Bell Systems Technical Jou rn a l? The paper contains
explicit formulas for the L C element values in a singly terminated 3 dB normalized Butterworth
L P filter where the load or source resistance is 1 ohm. O u r goal in this section is to present these
formulas without derivation. Similar formulas have been developed by other researchers in the
context of network synthesis, a subject for which many texts are available for further reference.
For an «th order 3 dB normalized Butterworth filter with a single 1H termination, the sequen
tial formulas for the L C elements are
' 7T \
= s in (19.43a)
\I n )
/3 > t ^
S in s in
\ 2nj I In ]
a-) = (19.43b)
K
co s*
sm sm
2/1
= ( 1 9 .4 3 c )
In
1070 Chapter 19 • Principles o f Basic Filtering
JT
a„ = ;jsin (19.43d)
\In }
»buttLC(4)
Elvalues =
3.8268e-01
1.0824e+00
1.5772e+00
1.5307e+00
Given these values, it remains to interpret them as inductances or capacitances. Two rules govern
the realization:
(i) There is never a shunt element in parallel with a voltage source.
(ii) There is never a dangling element at the load end.
Hence for the fourth-order filter above there are two possible circuit realizations, given in Figure
19.43.
0.3827 H 1.577 H
/YYY. /YY\ -O
1Q +
V.
1.0824 F 1.5307 F
(a)
Chapter 19 • Principles o f Basic Filtering 1071
0.3827 H 1.577 H
/YYY /YY\ -O
+
1Q
1.0824 F 1.5307 F
-O
(b)
F IG U R E 19.43. (a) R ^ = \ Q . (b) 7?/ = 1.
All that remains in the design procedure is frequency and magnitude scaHng, which are covered in
earher sections o f the chapter.
12. SUMMARY
This chapter has covered the basics of Butterworth LP, HP, and B P fiher design. Such design tech
niques build on a set of filter specifications requiring that the desired filter magnitude response lie
outside certain briclcvvall regions. Finding transfer functions that meet a set o f briclcwall con
straints is called the approxim ation problem . This chapter developed algorithms for finding
Butterworth transfer functions for the LP, HP, and B P cases. In addition, basic passive realizations
were presented as well as active circuit realizations, except in the B P case. In the active case, the
focus was the Sallen and Key low-pass and high-pass topologies. Both the passive and active t)^pes
of realization are built around the coefficient-matching technique, which associates the coeffi
cients in the circuit transfer function with the coefficients of the desired transfer function; one
then solves the resulting equations for appropriate circuit parameter values.
There are, of course, many other types of filter transfer functions; Chebyshev, inverse Chebyshev,
and elliptic are other well-known t)'^pes. Also, in addition to low- and high-pass filters, there are
band-pass, band-reject, and magnitude and phase equalizers. To add to the richness of the area of
filtering, there are analog passive, analog active, recursive digital, non-recursive digital, and
switched capacitor implementations o f all o f these filter types. The preceding exposition is mere
ly a drop in a very large and fascinating bucket of filter design challenges.
3 d B bandwidth o f B P filter: the difference between the two frequencies of a B P filter at which
the gain is 3 dB down from the maximum gain; if the maximum gain is 1, these fre
quencies correspond to a gain of M -Jl this is the most common reference for the mean
ing o f bandwidth.
3 d B bandwidth o f L P filter: the 3 dB down frequency, to
Active realization: a realization consisting of op amps, /?’s, and C s.
Approxim ation problem: the problem of finding a transfer function having a magnitude
response that meets a given set o f brickwall filter specifications.
1072 Chapter 19 • Principles of Basic Filtering
Attenuation (dB); the loss magnitude expressed in dB, i.e., y4((o) = - 2 0 logjQ|//(/(D)|.
BP to LP frequency transfonnation: the generation of equivalent normalized LP specs from a
given set o f BP specs.
Bandwidth with respect to the difference between the two frequencies o f a BP filter at
which the gain is down by a value of from the maximum gain.
Band-pass filter: a filter that passes any frequency within the band 0)^j s O) s while signifi-
cantiy attenuating frequency content outside this band.
Coefficient-matching technique: a method o f determining circuit parameter values by matching
the coefficients of the transfer funaion of the circuit to those of the desired transfer func
tion and solving the resulting equations for the circuit parameters.
Cutoff frequency: the frequency at which the magnitude response o f the filter is 3 dB down from '
its maximum value.
Filten a circuit or device that significandy attenuates the frequency content o f signals in certain
frequency bands and passes the frequency content within certain other, user-specified fre
quency bands in the sinusoidal steady state.
Frequency response magnitude: magnitude of the transfer function as a function of j(a, i.e.,
^
Gain in dB: 2 0 logjo|//(/w)| or 10 log,o|//(/co)p.
Gain magnitude: frequency response magnitude.
Half-power point: the point at which the magnitude response curve is 3 dB down.
High-pass fUten a filter that significantly attenuates the low-frequency content of signals and
passes the high-frequency content. ^
HP to LP frequency transformation: a transformation that converts a given set o f high-pass ■ '
brickwall specifications to an equivalent set o f low-pass specifications.
1
Loss magnitude: the inverse of the gain magnitude, i.e., ^ ^ ^ .^ y •
Low-pass (brickwall) filter specification: a filter specification requiring that the desired filter
magnitude response lie outside certain “brickwall” regions.
Low-pass filter: a filter that passes the low-fi-equency content o f signals and significantly attenu
ates the high-frequency content.
LP to BP fi%quency transformation: a technique for converting a passive LP filter to a BP filter
by converting Vs to series LC circuits and Cs to parallel LC circuits.
LP to HP frequency transformation: a technique for converting a passive LP filter to a HP fil
ter by converting Vs to Cs and Cs to Vs.
Maximally flat: property o f a filter at a point (0 wherein the magnitude response has a maximum
number o f zero derivatives.
Normalized Butterworth loss functions: a set o f Butterworth transfer functions, ordered by ' '
degree, having 3 dB loss at the normalized frequency s= jQ = j\ .
Passive analog filter: a filter composed only of resistors, capacitors, inductors, and transformers.
R C to CR transformation: a technique for translating a low-pass active filter to a high-pass active
filter in which resistors become capacitors and capacitors become resistors.
Radian Frequency
(a)
Radian Frequency
(b)
Figure P i 9.2
Figure P I 9.1
. r -h 0.803816435 + 0.82306043
10 7 4 Chapter 19 • Principles of Basic Filtering
10. Find a minimum-order Butterworth filter (a) Find the minimum filter order meet
meeting the specs = (1 dB, 75 Hz) ing these specs.
1076 Chapter 19 * Principles of Basic Filtering
(b) If R = 2 Q and Rj = 8 Q, find two solu 17. Consider the circuit of Figure P I 9.17.
tions {L, Q so that H{s) has a second- (a) Show tiiat the transfer function is
order Butterworth (maximally flat) I
response with 3 dB down (from its
R ,L C
maximum gain) frequency equal to 1 H{s) =
T
rad/sec. .V“ + s +■ + ! / « ,
(c) Magnitude- and frequency-scale both LC
net\vorks resulting from part (b) so (b) If Rj= ^ find two sets of parameter
that in the new networks the smaller values that realize the second-order
resistance is 2 kQ, and the 3 dB fre normalized Butterworth gain func
quency is 5 kHz. tion.
(c) Magnitude- and frequency-scale both
networks resulting from part (b) so
that in the new nervvorks the smaller
resistance is 2 k il, and the 3 dB fre
quency is 5 kHz.
R,> 1
Figure P I 9.15 / Y Y \
L
Figure P I 9.17
LC
H .A s) =
2
s~ + ---- f- -- y -I------ .A N S W E R S : (b) (C, L) = (0.4483, 3.3460) or
\C U LC
(1.6730, 0.8966) in F and H; (c) ( C L) = (26.6
which can be used to realize the normalized sec nF, 57.1 m H ) or (7.1 nF, 0.213 H ). becomes
ond-order Butterworth characteristic 4 [d l
K
H i s ) = —T- 18. Consider the second-order low-pass trans
.V -I- J 2 :s +
fer function
for K = 0.5. Suppose such a filter is to have cut
K
off f r e q u e n c y = 3500 Hz. (In this case, the fil H(s) = ^
ter is to have magnitude 3 dB down from its ^in ( s
maximum gain at f^.) I—
V'^p I
(a) Find values of L and C to realize the
normalized transfer function. where (0^^= 10^ rad/sec.
(b) I f the final value o f the resistors is to (a) Find |Myw)| and co^ (the 3 dB down
be 1 kt^, compute the new values o f L frequency). Is it true that co^ = O)^ in
and C that realize the desired filter. this case?
(c) Alternatively, suppose the final value (b) Find the impulse response and step
o f the capacitor is to be 10 nF. response.
Compute the new values of L and C (c) Using the RLC topolog)' shown in
that realize the desired filter. Figure P i 9 .18a, realize H{s) given that
/?£ = 10 k n in your final design. W hat
is the value of K in vour solution?
1078 Chapter 19 * Principles of Basic Filtering
Hint: First set O)^ = 1 and then fre gain is down 3 dB from its maximum
quency- and magnitude-scale to value), find the new element values.
obtain the final answer. A N S W E K . C, „ „ „ = C , „ „ „ = 0 .106 u F and
(d) Using the R L C topology shown in = 0.2122 H.
Figure P i 9.18b, realize H{s) given that
R^= 100 in your final design. W hat 20. Consider Figure P i 9.20.
is the value o f A" in your solution? (a) Show that the transfer function is
1
lY T v
L V L^LjC
I \ n 2 ( L\ + Lo + C\ 2
p . ----- 1— + —------ =------ j + ----------
^ ^1 ^2/ L 1L2C } L^LjC
(a)
yvYY (b) Determine values so that the circuit
L
realizes a third-order Butterworth gain
6 (c)
characteristic.
Find parameter values o f a third-order
(b) low-pass Butterworth filter having
cutoff 20 kHz and resistor values o f 1
Figure P I 9.18 kn.
(d) Use S P IC E to verify the frequency
19. Consider the doubly terminated filter cir response.
cuit o f Figure P I 9.19
10
/Y Y \ /Y Y \
L, I
—0
+ K
p r 10 >
+
Vo„, c >
Figure P i 9.20
Figure P I 9.19
A N S W F R S : (b) A, = L , = ] H, C = 2 1-: (c)
(a) Show that the transfer function of the 15.9 nF and 15.9 niH.
circuit is
L C .C .
« c > ( .0 = - ^ =
5 4-
c ,l c'-2 L C .C t l '-2
L C .C
(b) By equating the coefficients o f the 2I. Consider the two-port configurations
denominator o f with the shown in Fig ure P I 9.21.
denominator o f the third-order 3 dB (a) Prove that for Figure P I 9.2la
N L P Butterworth loss function, show
^oiit _ ^21
that the only possible realization is
= C2 = 1F and Z. = 2 H.
(c) If the source and load resistance are to (b) Prove that for Figure P I 9 .2 lb
be 1 k n an d^ = 1500 Hz (meaning the
Chapter 19 • Principles of Basic Filtering 1079
associated with Figure 19.21b, design (b) Realize the first-order normalized
a second-order Butterworth low-pass Butterworth transfer function
filter having co^= 1 rad/sec, R = 0 Q
and R^= \ Q. ^ 3dBNLP^^'^) - ~rr~r7 - — r
V in is ) S+\
(e) Frequency- and magnitude-scale the
circuits o f parts (c) and (d) so that the i.e., compute values for R , R-,, and C.
new filters have to^ = 5000 rad/sec, Hint: Let C= I E
and the single resistance in the circuit (c) I f the 3 dB down frequency is to b e ^
is 1 k n . = 3500 Hz, find so that the capac
itor is 1 nF in your design. Compute
the new resistor values and then deter
mine /?j if the source resistance is R^ =
500 Q.
C H E C K : /?, = 45 kQ.
Figure P I9.25
n
Chapter 19 • Principles oi Basic Filtering 1081
27. Realize rhe filter o f Problem 8 as a cascade (b) W lia t is the dc gain, K, o f the circuit?
o f a Sallen and Key circuit (Saraga design) and M odify the circuit to achieve the
a first-order active circuit such as the leaky inte desired gain of the transfer function.
grator. The 3 dB down point should be at (c) If the actual pass-band edge frequency
^'^cmhr largest capacitor should be 50 nF. is to be 7 kHz and C j is to be 0.05 uF,
(H int: It is often useful to generate an Excel determine the new clement values.
spreadsheet to do the relevant calculations, (d) Plot rhe loss magnitude response of
especially for multiple designs.) Verify the fre /V(s) in part (a) in dB. Determine
quency response o f your design using S P IC E .
C H E C K :^ „„„,= 2 d B
28. Repeat Problem 27, except use design A
instead ol rhe Saraga design for the Sallen and 32. Repeat Problem 31 using design A of the
Key portion. Sallen and Key circuit.
29. Realize the filter o f Problem 10 as a cascade 33. Repeat Problem 31 using design B of the
of two Sallen and Key circuits. The 3 dB down Sallen and Kev circuit.
point should be at The largest capacitor
should be 0.1 uF.
HIGH-PASS PASSIVE DESIGN
30. Repeat Problem 29, except use design B 34. A second-order Buttenvorth H P filter has 3
instead of the Saraga design for the Sallen and dB down point at^. = 2000 Hz. 'Fhe second-
Key portion. order Butterworth N L P prototype is given in
Figure P I 9.34 and has transfer function
31. In addition to Butterworth filter transfer
functions, there are Chebyshev filters, which
LC
have a faster transition from pass-band to stop
~> 1 1
band. This problem investigates the implemen s~ -I- -------.V-f"
R jC LC
tation o f a Chebyshev LP filter transfer function.
Recall that a second-order transfer function (a) W ith R^= ] Q., compute L in H and
C in F to realize the normalized sec-
ond-order prototype.
C H ECK: C = l/ V 2 F
(b) Using the results o f part (a), construct
can be realized by the Saraga design o f a Sallen a N H P circuit with \ Q. and
and Key circuit. ^^cHP~ ^ rad/sec. This is the so-called
(a) Determine the new values o f /?,, 3 dB normalized H P filter (3 dB
Cj, and C j that will realize the poles N H P ).
of a normalized second-order (c) Now construct the final H P design
Chebyshev L P transfer fiinction with R j = 100
(d) Do a S P IC E simulation to verif}- your
design.
^ ------------
0.803825+ 0.82306
1
r -W --/Y Y \ -o
+
10 •- 1o
LC jC j vJt)(
Hr
3 \ 2 + Co 1
------ — + ----------
S
Cj
+ — S + —
LC\C2
S
LC\C2
Figure P I9.36
Now construct the final HP design if (b) Compute O)^ for the desired HP filter.
(c)
the source resistance is to be 100 Q. Does your answer make sense? Think
Do a SPICE simulation to verify your about this carefully.
(d)
(c) Realize the 3 dB NLP transfer func
design.
tion using the circuit of Figure
P 19.37. Set I = 2 H and make Cj =
C2. Recall from Problem 19 that the
transfer function of the circuit is
VquA^) LC1C2
Vin(s) c, +c
36. A second-order high-pass c,c
1^2 LC 1C 2 LC 1C2
Butterworth loss function has stop band edge fre
quency fs= i kHz, pass-band edge frequency =
(d) Construct the appropriate high-pass
7 kHz, and cutofi^fijequencyf(. = 5500 Hz.
circuit with the largest capacitor equal
(a) Determine the attenuation in dB 2x fp
to 100 nF.
and^.
(e) Verify the frequency response o f your
(b) Realize the second-order 3 dB NLP
design using SPICE.
Butterworth filter using the circuit of
Figure P I9.36; i.e., compute values for L
and C.
Chapter 19 • Principles o f Basic Filtering 1083
A A(w)
Figure P I 9.41
HIGH-PASS ACTIVE
REALIZATION
42. Consider the set of H P specs for which
J - (2400 Hz, 3dB) and =
(600 Hz, 30dB). Following the procedure o f
Example 19.7, design an active H P filter meet
-► 0)
ing these specs so that the smallest capacitor in
the final design is 10 nF.
Figure P I 9.38
80 tweeter
Figure P i 9.47
AN.SWl-.R: L = 63(> u H , C = 9.95 uF'
MISCELLANEOUS
47. A certain audio amplifier has a very low
internal resistance. It is therefore approximate
ly represented by an ideal voltage sourcc. The
8 woofer and t\veeter each may be approxi
mately represented by a resistance of 8 ohms.
Design the simple crossover network shown in
Figure P i 9.47 such that the crossover frequen
cy is 2000 Hz.
C H A P T E R
Brief Introduction to
Fourier Series
CHAPTER OBJECTIVES
1. Introduce the concept o f and calculation procedure for the Fourier series of a periodic
signal.
2. Describe the relationship between the complex and the real Fourier series representations.
3. Set forth and discuss basic properties of the Fourier series.
4. Show how the basic Fourier series properties can be used to compute the Fourier series
of a wide range of signals from a few basic ones.
SECTION HEADINGS
1. Introduction
2. The Fourier Series: Trigonometric and Exponential Forms
3. Additional Properties and Com putational Shortcuts for the Fourier Series
Representation
4. Summary
5. Terms and Concepts
6. Problems
1. INTRODUCTION
Non-sinusoidal periodic waveforms are an important class of signals in electrical systems. Some
prominent examples are the square waveform used to clock a digital computer and the sawtooth
waveform used to control the horizontal motion of the electron beam of a cathode ray T V pic
ture tube. Non-sinusoidal periodic functions also have importance for non-electrical systems. In
fact, the study of heat flow in a metal rod led the French mathematician J. B. J. Fourier to invent
the trigonometric series representation o f a periodic function. Today the series bears his name.
The Fourier series o f a periodic signal exciting a linear circuit or system leads to a simplified
understanding o f the effect o f the system on the original periodic signal. This idea is briefly
explained in the next few paragraphs and illustrated in Example 20.1.
1086 Chapter 20 • B rief Introduction to Fourier Scries
W hen a periodic input excites a linear circuit, there are many ways to determine the steady-state
output. Using a Fourier series method of analysis, the input is first resolved into the sum o f a dc
component and infinitely many ac components at harmonically related frequencies. For example,
a 1 kHz square wave voltage with zero mean and a 0.5n V peak-to-peak value, w ithy{f) = {u{t)
-2u{t -0 .57) + u {t- 7)), T = 1 msec,
Vy(/) = 0 .2 5 jt ^ /(/ - n T )
( 20. 1a)
H=0
2/2 + 1
/i=0
( 20. 1b)
s in (w „r) - ^sin(3cu^/) + ^ sin(5w ^/) - ... u ( t )
w h e re ^Q = 2jtx 1000 rad/sec. In the Fourier series representation (which we will later develop), we
observe that the signal has a zero average dc value and harmonically related frequency components
For steady-state calculations, let be a phasor representing the (2;/ + l)tli harmonic o f the
Fourier series representation of Then, in steady state, the effect o f the system on this term
is given by = H {j{2n + 1)Wq)V^ , assuming that the transfer function H{s) is stable. Again by
linearity, the actual time domain output is then computed for each The resulting time func
tions are then summed to obtain the steady-state part o f the complete response. Example 20.1
illustrates the particulars.
E X A M P L E 20.1. Figure 20.1b shows a simple R C circuit (/? = 1 C = 1 F) and a square wave
input voltage (Figure 20.1a) with E = 3 0 jt V and T = 4 sec. Find the first four components o f the
output voltage v^{t) in the steady state.
>
r R
b ■ +
• • •
v Jt) v jt)
(a) (b)
S o l u t io n
Step 1. Determine the Fourier series representation o f v-^^it). The fundamental frequency o f v-^^{t) is
Jq = 1/7’ = 0.25 Hz or C0q = I jifo = 0.5jt rad/sec. As shown in die next section, u-^^{t) has the
infinite (Fourier) series representation
” s in ((2 « + ^
nil)
«=0n 2// + 1
For this signal the average dc value is 15 ti. The remaining terms in the Fourier series are harmon
ically related sinusoids of decreasing magnitude.
Step 2. F in d the circuit transfer fim ction. The stable transfer function o f the circuit is
^ s) =
Vi, . v+1
Step 3. D eterm ine the magnitude an d phase o f H{s) at s = j( 0 . For sinusoidal steady-state analysis,
set s =jo ) to obtain
Jd
H(jco) = - =H„/
JOJ + 1
in which case
, 6 = - tan“ ^(w )
yjco^ + 1
Step 4. Find steady-state responses to a ll components o f Vj^^{t). Table 20.1 lists the steady-state
responses to several components of Vj^Jit).
0) 0 Wo 5«o
Step 5. Apply superposition to obtain the steady-state portion o f the complete response. Neglecting har
monics o f scventii order and higher, the approximate steady-state solution is
Vg{t) = 15j*r + 3 2 .22sin((/jyr- 57.52") + 4.15 sin(3w Q /-7 8 .0 2 ‘’) + 1.516 sin(5wQ - 8 2 ,7 4 ”) + ...
This response shows the effect of the system on each component of the input signal and how it in
turn affects the overall steady-state output response. Since the time constant o f the circuit is 1 sec,
this steady-state response more or less constitutes the actual response for r > 5 sec.
Exercise. Use MATLAB to plot the approximate waveform for vj^t) for 0 < r < 8 sec, based on
Tiible 20.1. From the plot identif)' Alternatively, you can use the max
and min commands in MATLAB.
AN SW l-R: 82.013 V and 12.235 \'
In Example 20.1, the first step o f the solution was to represent a periodic waveform as a sum o f
sinusoidal components, called the Fourier series. Section 2 covers the definition and basic proper
ties o f Fourier series. Section 3 describes several shortcuts for computing the Fourier coefficients
and identifies other important properties. Since, in practice, only a finite number o f terms can be
considered, the Fourier series method yields only an approximate solution.
Because many mathematical and engineering handbooks have extensive tables o f Fourier series o f
different waveforms, it is convenient to use these tables in much the same way as one uses a table
o f integrals or a table o f Liplace transforms. The Fourier series o f some basic signals are provid
ed later, in Table 20.3. The use o f this table, together with some properties and shortcuts dis
cussed in section 3, make the study o f the Fourier series method much more palatable to begin
ning students o f circuit analysis.
B a s ic s
/ r + 7) = P ) (20.2)
r i s the period o f the signal. The fundamental period is the smallest positive real number Tq for
which equation 20.2 holds; ^ = l/7j, is called the fundamental frequency (in hertz) o f the sig
nal; cOq = I k/ q = 271/7*0 is the fundamental angular frequenc)^ (in rad/sec). T he sinusoidal wave
form o f an ac power source and the square wave form used to clock a digital computer are com
mon periodic signals. Figure 20.2 shows a portion o f a hypothetical periodic signal.
Chapter 20 • Brief Introduction to Fourier Series 1089
00
/ (0 = Y + 2 ! (« «co s(m v )f/ 7„sin (/ ;fo „/ )) (20.3)
//= 1
Equivalently,
where
(20.5a)
and
(20.5b)
Both infinite series, equations 20.3 and 20.4, are called the trigonometric Fourier series repre
sentations We note that equations 20.5a and 20.5b follow fi-om the trigonometric identit)'
\
A cos(.v) + fisin(.v) = + B~ cos .Y + tan ^ ----- with due regard to quadrant.
\ I Aj f
From equation 20.4, observe that is the average value ofy(r). In electric circuit analysis,
refers to the dc component ofJ{r) . The first term under the summation sign, <^jCos(tOQt + ^^,), is
called the fundamentaJ com ponent (or first harmonic) ofy(r), with amplitude and phase angle
0 j. The term d-, cos(2o;(jt + 6-,) is called the second harmonic o^J{t), with amplitude d-y and phase
angle and similarly for the other terms d^^cos(;/WQt + 0^^), which are the ;/th harmonics as indi
cated by the term n^o^y
Chapter 20 • Brief Introduction to Fourier Scries
As illustrated in Example 20.1, given any periodic function/r), it is important to determine the
coefficients in equation 20.3 or equation 20.4. For the purpose o f easier calculation, it is advan
tageous to introduce the equivalent complex Fourier series representation o f the periodic func
tion y(f).
/ (0 = i (20.6a)
n = -ao
where it can be shown that
In + T,
c„ = (20.6b)
Since we have two (allegedly) equivalent forms, let us now develop the relationship between the
real and complex forms o f the Fourier series. Recall the Euler identities:
/ (')= 2 ‘V E
/I = - oc “ /I = I
= -y + ^ ^----------- + h
n = \\ 2j
= 2 R e (0 (20.8a)
b„ = -21m(^-„) (20.8b)
Chapter 20 • Brief Introduction to Fourier Series 1091
^0 = ^0 = ^ (20.8c)
= ^ c „ ,n = 1, 2, 3 ,... (20.8e)
In the real trigonometric Fourier series equations 20.3 and 20.4, the summation is over positive
int^er values o f«, whereas in the complex exponential Fourier series the summation extends over
int^ers n such that -oo < « < oo. While each term in equation 20.3 or equation 20.4 has a wave
form displayable on an oscilloscope, each individual term in the complex exponential Fourier
series lacks such a clear physical picture. However, two conjugate terms in the complex exponen
tial Fourier series always combine to yield a real-time signal d^cos{no)Qt + 6^.
To develop equation 20.6b for the coefficient we multiply both sides of equation 20.6a by
r > " 0 'to obtain
* = -00
Integrating over one entire period, [/q,/q + T^y produces
Jto k = -oo
(20.9)
* = -00 it = -00
tn + T„
c„ =
Tn (20.11)
The lower limit of integration, ?q, can be any real number, but is usually chosen to be 0 or - T qH,
whichever is more convenient. In addition, Tq will sometimes be written simply as T.
A hand computation of coefficients would proceed by first computing for « = 0, 1,..., using equa
tion 20.11. One would then obtain and using equations 20.8a and b, and and 0^ by equa
tions 20.8c, d, and e. Other formulas are available for obtaining the Fourier coefficients by integrals
involving sine and cosuie functions. However, equation 20.11 is preferred because an integration
involving exponential fimctions is often simpler than an int^ration involving sinusoidal flmctions.
y w
w
Chapter 20 • Brief Introduction to Fourier Series
EXA M PLE 2 0 .2 . Find the trigonometric Fourier series for the square wave signal o f Figure 20.3.
S olution
The fundamental period o f/ f) is Tq = T. By inspection, the average (dc) value is
L L L
(20.12)
A .( JT^
— sin n — , / ; = l , 2 , . . .
Jtn 2}
Thus our signal has the complex Fourier series of equation 20.6 with given by equation 20.12 and
a „ = 2 R e ( c , J = — sm n — ,h , = 0
jTn
Substituting these coefficients into equation 20.3 yields the following trigonometric Fourier series
for the square wave o f Figure 20.3:
A 2A / I I
/ ( 0 = T ‘*'— c o s ( o ; o / ) - - c o s ( 3 a ;o O + 7 C o s ( 5 a ;o / ) - .. . (20.13)
2 7T \ 3 5
After computing the Fourier series of a periodic signal y(^), it is straightforward to find the Fourier series
o f a related periodic signal g(t) whose plot is a translation o f the plot of/f). A translation o f a plot is
Chapter 20 • Brief Introduction to Fourier Scries 1093
a horizontal and/or vertical movement o f the plot without any rotation. A translation o f the waveform
in the vertical direction causes a change in the dc level and affects only the coefficients d^y and Cq.
A translation o f the waveform in the horizontal direction causes a time shift that changes only the
angles 0,^ and has no effect on the amplitude We state the relanonship formally as follows.
^ 0= do = CQ + K = dQ + K = — + K (2 0 .l4 b )
/; = ± 1 , ± 2 , . . . (20.14c)
=1 , /? = ± l ± 2 , . . . (2 0 .l4 d )
The proof o f this property is straightforward and is left as an exercise. Note that equation 2 0 .1 4c
indicates that a time shift o f the signal affects only the phase angles o f the harmonics; the ampli
tudes o f all harmonics remain unchanged.
Exercise. Supposey(f) in Example 20.2 has A = 30 and suppose ^ f) =J{t) - 10 . Find the coeffi
c i e n t s a n d o f the Fourier series o f ^ r).
AN SW I'R: = 0.5^/,, = 3; all other coefficients arc unLh.mgcd, i.e., = — sin n —
because h.. = 0.
We will now use the translation property to obtain the Fourier series o f a square wave that is a
translation o f Figure 20.3.
EXA M PLE 2 0 .3
The square wave g{t) shown in Figure 20.4 is anti-symmetrical with respect to the origin. Find
the complex and real Fourier series for^^) given the Fourier series (equation 20.12) ofJ{t) depict
ed in Figure 20.3.
u rn Chapter 20 • Brief Introduction to Fourier Series
g(t)
>
' A/2
w
-I -0.5T 0 0.5T T
-A/2
FIG URE 20.4 A square wave anti-symmetrical with respcct to the origin.
S olution
The curve is a translation o f the waveform y(f) o f Figure 20.3. Specifically,
S ( 0 = / ((-/ ,/ ) + a: = /
COS ncoQt
(
= cos ncoQf - n —
\ 4 \ 2}
2A (20.15)
sin(wQ/) + —sin(3wQ/)-i- jsln (5 w ()/ ) + ...
7Z
For the case o f a square wave signal, by choosing the time origin and dc level properly, the result
ant plot displays a special kind of symmetry that results in the disappearance o f all sine terms or
all cosine terms. The square waves o f Figures 20.3 and 20.4 are special cases o f the periodic func
tions amenable to such simplifications. The general case is given by the following statement.
The plot o f an even function is symmetrical about the vertical axis. Examples o f even functions
include cos(cl)/) and the square wave o f Figure 20.3. The plot o f an odd function is anti-sym-
metrical about the vertical axis. Examples o f odd functions are sin(o)/) and the square wave o f
Figure 20.4. The proofs o f the symmetry properties are left as homework problems.
Chapter 20 * Brief Introduction to Fourier Scries 1095
Exercise. Suppose in Figure 20.4 From Example 20.3 is shifted into q{t) = g { t - 0 .2 5 7 ). Find
the coefficients and o f the Fourier series o f q{t).
lA -T\
AN SW ER: (h) = (i{) = 0. and hir n ^ 0. d„ = a„ = -------sin n — .h„ = 0 .
.111 ~)
To simplify the calculation o f the Fourier coefficients, we should attempt to relocate the time ori
gin or change the dc level so that the new function ^t) displays even or odd function symmetry.
This may not be possible for an arbitrary periodic signal. When it is possible, we will calculate
the Fourier coefficients o f the new function ^t), which has only cosine terms or sine terms, and
then use the translation propert)' to obtain the Fourier coefficients for the original function//) .
A waveform o f particular importance in signal analysis is the periodic rectangular signal shown in
Figure 20.5. The fundamental period is T, and the pulse width is fiT. The constant (5 is called
the duty cycle, usually expressed as a percentage o f T. The square wave o f Figure 20.3 is a spe
cial ca.se o f the rectangular wave o f Figure 20.5 with a 50% duty cycle.
EXA M PLE 2 0 .4
Find the trigonometric Fourier series for the rectangular waveform o f Figure 20.5.
S olution
T he procedure is almost identical to that used in Example 20,2 for a square wave. By inspection,
the average value is
(20.16a)
To calculate the other Fourier coefficients, choose fg = -7 7 2 . Equation 2 0 .10 then yields
pT fW
~E.
^ /\
= 2 R e (c -„ ) = — sin(/?^.T) (20.16c)
jzn
b.. = 0 (2 0 .l6 d )
Specifying the coefficients completes tine determination ot the Fourier series, i.e.,
2A ^ (sin(nl^jr)\
(1,1 ) = p/\ 1------- y COS(nw„t)
n= \ /I = l^
A very important conclusion about the rectangular wave can be drawn by examining equations
20.16: as the ratio o f the pulse width to the period becomes very small, the magnitudes o f the
fundamental and all harmonic components converge to twice the average (dc) values. To see this,
recall that sin(.v)/x approaches 1 as .v approaches 0. From equation 20.16c, we may rewrite as
a„
'n = — sin(///i.;r)= 2ftA
jTn nftji
To give some concrete feel to this property, lable 20.2 gives the ratios o f = —— and —
^'(ly ^av
for w = 1, ..., 9 for the case o f ft = 0 .0 1 . Answers are rounded o ff to three digits after the decimal
point. Note that when the periodic rectangular signal is shifted vertically, the ratio =
TABLE 20.2
A m p lit u d e s o f t h e F i r s t N in e H a r m o n i c s f o r t h e C a s e P = 0 .0 1
n 2 3 4 5 6 7 8 9
The constant property holds approximately true when the pulse width is a very small fraction
o f the period 7q. Even i f a waveform is not rectangular, if the pulse width is very small compared
Chapter 20 * Brief Introduction to Fourier Series 109:
to its period, then the nearly constant propcrrv' o f —~ iind ^ continues to hold, as long as the
«I
pulse is o f a single polarit)'. For example, consider the periodic short pulse shown in Figure 2 0 .6 .
In calculating the Fourier coefficients f;; o f this pulse using equation 2 0 .1 1 , the limits o f the inte
gral, originally (/q, + 7 ), are changed to (-(^ 7 7 2 , ^ 7 7 2 ). As approaches zero, the factor
in the integrand has a value very close to 1 in the new time interval, as long as n, the
harmonic order being considered, is not very high. Therefore, for pulses o f narrow width we have
(xT aT
(20.17a)
rhus in terms of the coefficients, from equation 20.8, again for pulses o f narrow width we
have
(20.17b)
This result is pertinent to the approximate analysis o f a rectifier circuit covered in other texts or
in the second edition o f this text.
To this point, we have calculated the Fourier coefficients only for some very simple periodic sig
nals. The evaluation o f the integral in equation 20.11 becomes much more involved when the
signal j{t) is not rectangular. Fortunately, many mathematical and engineering handbooks now
include comprehensive tables o f Fourier scries. From a utility point o f view, one may use these
tables much the same as one uses a table o f integrals or a table o f Laplace transforms. In effect,
the need to carrj' out the integration in equation 20.11 is not compelling in practice.
In many applications, it is important to know the average power o f a (periodic) signal and the
magnitude o f its various harmonics. From equation 20.5, the effective value o f the dc compo
nent is |<^q|, and those for the fundamental and various harmonics are d^j! yfl , /; = 1 , 2 , . . . . It is
easy to show that the effective value, or the rms value, o f/ r) is
(20.18a)
where the d- coefficients are from the Fourier series o f equation 20.4. If fit) (current or voltage) drives
a 1 ohm resistor, then the average power absorbed by the resistor is ~ ^eff • Hence we
say that the average power o f a periodic signal y(/) represented by a Fourier series is given by
1098 Chapter 20 • Brief Introduction to Fourier Series
r~\
1 1 *
P av eJ{t)= F }ff= dl+ -di + - d l - ^ ..= 2 k„| (20.18b) O
2 n=-oo
The relationships indicated in equations 20.18 are often termed Parsevals theorem.
The information on the phase angle is important when one wishes to construct the time domain
response in steady state. For the time domain problem, the Fourier series method yields only an
approximate solution, because one can only sum a finite number of terms in the series.
C on v erg en ce o f th e F o u r ie r S eries
Convergence of the Fourier series is an intricate mathematical problem, the details of which are
beyond the scope of this text. On the other hand, it is important to be aware of the ways in which
the Fourier series may or may not converge to a giveny(/). Our discussion is not comprehensive,
but is adequate for our present purposes.
To begin, we define a partial sum of terms of the complex Fourier series of a function ^(/) as
sn « )= 2
k= -N
From our experience thus far, 5yy(f) must in some way approximatey(r). The difference between
j{t) and its approximation, 5yy(r), is defined as the error
^at( 0 = / ( 0 - 5 ^ ( 0
r\
One way to get a handle on how well Sj^t) approximates is to use the so-called integral
squared-error magnitude over one period, [/gi T\, defined as
e ^ (l) dt
This is often called the energy in the error signal, as energy is proportional to the integral of the
squared magnitude of a function. It turns out that for ftmaions having a Fourier series, the choice
of the Fourier coefficients minimizes E[^ for each N. Further, for such fiinctions, 0 as
00, i.e., the energy in the error goes to zero as JV becomes large. This does not mean that at each r,
fij) and its Fourier series are equal; it merely means that the energy in the error goes to zero.
o
Continuous and piecewise continuous periodic functions have Fourier series representations. A
piecewise continuous fiinction, such as a square wave, is a function that (1) has a finite number
of discontinuities over each period but is otherwise continuous, and (2) has well-defined right- r\
and left-hand limits as the function approaches a point of discontinuity. For piecewise continu
ous functions, it turns out that the Fourier series converges to a value halfway between the values
of the left- and right-hand limits of the function around the point of discontinuity. Even so,
-* 0 as TV-» 00 for piecewise continuous functions.
There are many functions that are not piecewise continuous and yet have a Fourier series. A set
of conditions that is sufficient, but not necessary, for a function to have a Fourier series repre
sentation is called the Dirichlet conditions.
Chapter 20 • Brief Introduction to Fourier Series 1099
D irichlet conditions
Condition 1. Over any period, [fg, tQ+T\,J{t) must have the property that
In the language o f matliematics, this means that J{t) is absolutely integrable. The conse
quence o f this property is that each o f the Fourier coefficients is finite, i.e., the exist.
Condition 2. Over any period o f the signal, there must be only a finite number o f min
ima and maxima. In other vv^ords, functions like sin(l/^) are excluded. In the language o f
mathematics, a function that has only a finite number o f maxima and minima over any
finite interval is said to be o f hounded variation.
Condition 3. Over any period, y(r) can have only a finite number o f discontinuities.
As mentioned, at points o f discontinuity, the Fourier series will converge to a value midway between
the left- and right-hand values o f the function next to the discontinuity. There may be other differ
ences as well. Despite these differences, the energy between the function and its Fourier series
representation is zero; i.e., with N approaching oo, goes to zero. Thus, for all practical purpos
es, the functions are identical. This practical equivalence allows us to analyze how a circuit responds
to a s i g n a l b y analyzing how the circuit responds to each o f its Fourier series components.
If a periodic function y(f) is known only at some sampled points, e.g., by measurements, then its
Fourier coefficients must be calculated by the use o f numerical methods. On the other hand, if
j{t) has an analytic expression, then its Fourier coefficients can often be calculated from equation
20.11, The properties discussed below are o f great value in simplifying the calculation o f Fourier
coefficients. Their proofs are fairly straightforward and arc left as homework problems.
The linearity property: Letyj(/) and f^{t) be periodic w'ith fundamental period T. \^j{t)
= KJ'^{t) + K^fyU), then the Fourier coefficients o^j{t) may be expressed in terms o f those
of/j(f) and ^ (r) according to the following formulas:
(20.19a)
(20.19b)
and
(20.19c)
In general,
n s
This property allows us to easily obtain the Fourier series of a sum of periodic signals when the
Fourier series of the individual signals is already known.
DEFINITION
A periodic function y(/) is said to be half-wave symmetric if
In words, y(r) is half-wave symmetric if a half-period shift of the plot combined with a flip about
the horizontal axis results in the identical function y(/). Some simple examples of half-wave sym
metric functions are sin(ft)r), cos{(Ot), and the square wave of Figure 20.4,
c<*)
c„ =
Except for the constant term, all terms derive from those ofy(r) by differentiating k times;
conversely, all terms oij{t) derive from those of/^^(/) by ^-fold (indefinite) integration. The exclu
sion of the constant term in the relationships poses no difficulty, since the constant is simply the
average value of the periodic function.
Again, these properties help to simplify the calculation of Fourier coefficients. In fact, these prop
erties make it possible to compute the Fourier series for the waveforms given later in Table 20.3
without carrying out the integration of equation 20.11. Achieving this, however, depends on first
finding the Fourier series of a waveform for which computing by equation 20.11 is extremely
easy. We will illustrate several such calculations in Examples 20.5 through 20.8. During this devel
opment, the following trigonometric identities will prove useful;
/ ^ y\ / _ y'
cos(jc)-cos(y) = - 2 s i n ( ^ ^ - j sin (^ -Y -j (20.21a)
n
Chapter 20 • Brief Introduction to Fourier Scries !01
E X A M PLE 2 0 . 5 . Find rhe Fourier series for rhe periodic impulse trainy§(^) shown in Figure 20.7.
A jk A k. A
(A) (A) (A) (A) (A)
-21 -T 2T
S olution
Using the sifting property' o f an impulse function together with equation 20 .1 0 yields
T
A (20.22a)
T
Equation 20.22b states that, for a periodic impulse train, all harmonic components have magni
tude equal to twice the average value. This is the limiting case o f the short pulse propert)- stated
in section 2. The next example uses the derivative and integral properties o f the Fourier series to
develop an alternative derivation o f the Fourier series for a periodic rectangular pulse train, derived
earlier by the use o f equation 20.11.
E X A M PLE 2 0 .6 . Find the Fourier series for the periodic rectangular pulses^(f) shown in Figure
20.8a. (This corresponds to item 2 in Table 20.3.)
S olution
Figures 20.8a and b show ^(r) and its derivative,^ (t). The latter may be written as the sum o f
two shifted impulse trains:
/ /(5T
«T’ \\
{ t-
f n( t ) = f6 t + — -fa (2 0 .2 3 )
\ ^ r \ ~ r;
102 Chapter 20 • Brief Introduction to Fourier Series
(b)
FIG U RE 20.8 (a) Periodic rectangular pulse crain and (b) its derivative. The parenthetical values,
{A) and {-A), next to each impulse in part (b) denote the weight of that impulse, i.e., its area.
Using the time shift property (translation in the horizontal direction), together with equations
20.22b and 20.21a,
2A
f'p ( 0 = Y, - cos{na)Qt - n^jv))
(20.24)
X 4/\
" E (sin(/7o;oOsin(«/3;r))
« = I ^
00
(2 0 .2 5 )
E XA M PLE 2 0 .7 . Find the Fourier series for the half-wave rectified sine wave fhsi^) shown in
Figure 20.9a (item 10 in Table 20.3).
A f.(t)
(0
FIGURE 20.9 (a) Half-wave rectified sine wave as the product of two functions: (b) a cosine wave
and (c) a square wave.
S olution
The periodic fiinction j{t) o f Figure 20.9a may be viewed as the product o f the sinusoidal wave
y4cos((0Qr) and the square wave^(r), shown in Figures 20.9b and c, respectively. Using the Fourier
series for^(^) given by equation 20.13, we have
//i.v(0= [ ^ c o s ( w o O lA ( 0
M 9 / 1 1 \1 (20.26)
= y4cos(o;o0 ~ cos(ojo^) — cos(3w()/) + —cos(5o;o/) - ...
2 ;r ^ 3 5
104 Chapter 20 • Brief Introduction to Fourier Series
A lA
= - + - cos(w()/) + — - c o s ( 2 ft ;,)0 - — cos(4coqi)
IT z JT \ J Id
A
= - + - co.s((OoO + — 2 T T — -co.s( 2 mu„/ (20.27)
^ 2 JT 4 ,r _ 1
Note that the fundamental component is present in and the remaining terms are all even
harmonics.
E X A M PLE 2 0 ,8 . Find the Fourier series for the full-wave rectified sine wavej^(r) shown in Figure
20.10a (item 9 in 'I'able 20.3).
Chapter 20 • Ekicf Introduction to Fourier Series lOS
■> t
FIGURK 20.10 (a) Full-wavc rcctificd sine wave as the sum of the two signals in (b) and (c).
S olution
One approach is to apply the same technique as in Example 20.7. Specifically,
where f^{t) is the square wave o f Figure 20.9c. Alternatively, to avoid repeating all the arithmetic,
note that is the sum o f the two waveforms shown in Figures 20.10b and c; i.e.,
(20.28)
Substituting equation 20.27, the Fourier series into equation 20.28 yields the desired
Fourier series:
lA 4/4
(20.29)
4 ;r -I
The derivative propert}' is particularly useful for tackling periodic piecewise linear waveforms.
Piecewise linear waveforms consist o f straight-line segments. From differentiating once, or at most
twice, impulses appear. The integration given by equation 20.11 is trivial if the integrand contains
a shifted impulse function. This fact, together with the derivative and integral properties, reduce
the task o f calculating the Fourier coefficients for piecewise linear waveforms to some complex
number arithmetic. These examples and several other commonly encountered periodic waveforms
have Fourier series as given in Table 20.3. Engineering and mathematical handbooks contain
much more comprehensive tables. O f course, when a waveform does not appear in a table, the
Fourier coefficients must be computed manually or numerically.
106 Chapter 20 • Brief Introduction to Fourier Series
TABLE 20.3
Chapter 20 • Brief Introduction to Fourier Series 107
^/o = F „ ,= 0 .5 A
n (a + P)ji
b„= 0
A A
/ (/ ) = - + - c o s (w o / ) +
Jl L
n+1
M v H L _ COS (inojQt)
^ k 4n^ - 1
ave=A/n rms=A/2
108 Chapter 20 • Brief Introduction to Fourier Series
4. SUMMARY
Given that many mathematical and engineering handbooks have extensive lists o f the Fourier
series o f common signals, this chapter has taken a practical approach to the calculation o f the
Fourier series and its application to circuit analysis. The idea is to use tables such as Table 20.3 in
the same way engineers have come to use integral tables. The keys to using such tables for the com
putation o f the Fourier series o f a waveform are the various properties that allow one to convert a
known series into one that fits a new waveform. The idea here is to express the new waveform as
a translation of, a linear combination of, a ^-fold derivative of, or a ^-fold integral o f signals with
known Fourier series as in Table 20.3, or any mixture o f these operations. The Fourier coefficients
o f the new signal can then be expressed in terms o f the Fourier coefficients o f signals with known
Fourier series.
Knowledge o f the Fourier coefficients o f a signal such as the output o f an audio amplifier allows
one to investigate phenomena including the distortion introduced by the amplifier. In the case o f
a dc power supply, such knowledge allows us to characterize the degree o f unwanted ripple in the
output o f a rectifier circuit. In addition, the application o f Fourier series plays an important role
in the computation o f steady-state circuit responses to periodic input signals.
Average value o f periodic/(^): the term in equation 20.4; also referred to as the dc compo
nent ofy(r).
Derivative/integration property: Let/^^H^) denote the ^ h derivative o f a periodic function y(^).
Then the Fourier coefficients o f the ^ h derivative, satisf}' = (/>zoJq)^c^^ for all
and conversely, J{t) is the /’th integral of/^^^(^), then for all n except
0.
D uty cycle: for a rectangular signal having fundamental period T as illustrated in Figure 20.5, the
duty cycle is the constant (3 that determines the pulse width ftT.
Even function: y(/) = The plot o f an even function is symmetrical about the vertical axis.
Exponential (complex) Fourier series; decomposition into a sum of complex exponentials
as given in equations 20.6.
Fundamental com ponent (first harm onic): the first term under the summation sign in equa
tion 20.4, cos(o)q^ + 0 j), having amplitude d^ and phase angle 0 j.
Fundamental frequency (in hertz): /o = — . where T^^ is the fundamental period. Note:
_ 2:7r . ^0
Wq = 2;r/() = — is the fundamental angular frequency in rad/sec.
Fundamental period: the smallest positive real number for w h i c h + Tq) =J{t).
Chapter 20 • Brief Introduction to Fourier Series 1109
f(t)
e**
6. Find the Fourier coefficients and 6^ o f the
periodic function shown in Figure P20.6. Flint:
Make use o f the result o f Problem 5.
Figure P20.2
Chapter 20 • Brief Introduction to F-'ouricr Series 1111
-1 0 1 2 3
Figure P20.6
>
Figure P20.10
8. Consider the isosceles triangular wave shown
in Figure P20.8, which is item 4 ofTable 20.3. 11. Consider the asymmetrical triangular wave
Compute the Fourier series o f this waveform shown in Figure P 2 0 .l l, which is item 7 o f
utilizing the results for items 2 and 3 ofTable Fable 20.3. Let 7 = 1 and u = 0.25. Find the
20.3. Hint:y(^) is the product o f (1) the wave Fourier coefficients and d-^. Hint:
form o f item 2 with proper height and (2) the Differentiating J{t) twice results in periodic
waveform o f item 3 shifted down by a suitable impulse trains.
amount. f(t)
> t
Figure P20.8
12. Find the Fourier series coefficients Tq, c’p
9. Consider the clipped sawtooth wave shown and C-, for the periodic waveform shown in
in Figure P20.9, which is item 6 ofTable 20.3. Figure P20.12. Hint: The period T = 4, and
Compute the Fourier scries o f this waveform you should express J{t) as the sum o f several
utilizing the results o f items 2 and 5 ofT able shifted square waves.
20.3. Hint:y(r) is the product o f (1) the wave
1 112 Chapter 20 • Brief Introduction to Fourier Scries
A
oj= 10^ rad/secand 10-^ rad/scc. For
the frequency range990 x 10-^ < w< 1010
X 10-^ rad/sec, assume that the band-p;iss
amplifier has the following ideal magni
tude and phase characteristics;
\H{joj)\ = 10
-2 and
Figure P 20.12
/LH ( jco) = - —— X 45^'
5000
jy r \ v„(t)
10H
6 4 pF
Fig P20.14
Figure P20.13
ANS\X'ER; (a) = 0.2 cos(((/;. - 10 Jt) +
ANSW ER; (a) 0.1 cos((<o + 2 cos(a;^./) + 0 .2 c o s ((o ;.
+ + 0.1 cos(((/^ + V; (b) =
r.... (M = 200 + >/2 42.6cos((o/ - 1 7 5 .4 ") + 2 cos((oj. - + 4.5°) + cos((«;. - +
y 9°) + 20 cos(oj^j') + 2 cos((o>. + -
1.192cos(3n;/ - 1 4 S .7 " )+ 0 .5 6 COM5o>/ - 129.2^)
4.5") + cos((o;,+ 2 r o J / - 9 " ) V
with u) - 377 rad/sec; lb) 204.5 V. 4.18 W
15 . Consider the circuit o f Figure P20.15, in
14. Consider the circuit o f Figure P20.14. The which C = I F (initially relaxed), R = 1.443 =
input to this ideal band-pass amplifier is an \Un{l) n , and is a sequence o f impulse
amplitude-modulated waveform given by currents Q b[t—nT), T - 1 sec, Q = 1 coulomb,
=J{t)x cos(cu^r) V, where and ;; = 0 , 1 , 2 , —
(a) Show that the response due to the first
J{t) = 2[1 - 0.2 cos(w,/) + 0.1 cos(2(t;,/)] hnpitlse current alone is =
Chapter 20 • Brief Introduction to Fourier Series 13
16. Consider again the circuit of Figure 17. The LC resonant circuit in Figure P20.17a
P20.15. If the impulse current train has been is initially relaxed and the impulse train
applied for a long time (theoretically since t =
- oo), that is, « = - 00,..., -2, -1,0, 1,2, ...
(a) Show that for the time interval (0, T)
/i=0
14 Chapter 20 • Brief Introduction to Fourier Scries
shown in Figure P20.17b is applied. The natu shown in P 2 0 .I8 b is applied. This problem
(and circuit) differs from the previous one only
ral frequency o f the tank circuit is f.. = ------\-----
I 2 h /lC in the presence o f the resistance R. When the
and the period is Tq - — = 2vtVZc . value o f R is large enough, the impulse response
Jo is a damped sinusoid. The purpose o f this prob
(a) Show that if the frequency o f the lem is to show that if the damping effect o f i s
slight and the input impulse train “synchro
impulse train, / = — , is the same as
7' , nizes” with the damped sinusoids o f the
the resonant frequency / q = impulse response, then the output voltage mag
In y fL C ' nitude, although finite, can reach a ver\' large
then /y(r) grows without bound. value.
Specifically, show that the inductor (a) ¥\n6H{s) = - ^ ---- .If//(^) has a pair
current in the time interval {n - \ ) T <
o f poles at ^= - CJ ± then the
t < 7i T \s given by i^{t) = n x
impulse response has the form h{t) =
sin(w^r).
(,-ot cos{ioj) + 5sin(w^)]//(f). Find
(b) Show that iff=fQlm for any integer m,
CJand in terms o f R, L, and C.
the inductor current again grows
unbounded. Note that (i)w j Wq =
(c) Show that if/ = 2/q, then i^{t) is peri
(ii) the waveform o f every cycle (time
odic and remains bounded. Sketch
2ji
two cycles o f /^(r). interval o f length Tj = ---- ) differs
(X)j
(d) Show that in general if T = (2w +
from its adjacent cycle by only a mul
1)(0.57'q) or, equivalently,
tiplicative constant a = .
2/o
(b) Show that if the input impulse train syn
2m + 1 chronizes with the damped oscillations,
then /^(/) is periodic. Sketch t\vo
meaning that T = T j = 2nUo^ then
cycles of
the output due to the impulse train
o f Figure P 20.18b is
Further, if Q = 1 coulomb and T = 1 (a) If the input has been applied for a long
sec, show that v^JJ) in steady state is time, then the circuit reaches steady
given by (consider item 18 of Table state and as shown in Figure
12.1) P20.19c. Using this fact and equation
8.19a, show that in steady state.
■'.»,W = 2<'-«®3I<[cos(2 w ) -
0.1103 sin(2;»)] V for 0 < r < 1 L, — 1/ I V
''max - V*mm •
^oul,mm ~ *min o .5 7
ik
(C8 (Q) (Q)
-T 2T
(a)
(b) .v jt)
V
Figure P20.18 ...
fflin
CHECK: (d) //(5) =
r + 1.3863^ + 39.96
's-^ s -0.5T
(e)/ / (5 ) = _ ---------- 0.5T T 1.5T
r +0.102586j + 34.481049 (b)
and = 20tf“°*°5’^'[cos(2;«) -
0.0082 sin(2;r^)] V for 0 < ^< 1
w ■
16 Chapter 20 • Brief Introduction to Fourier Series
ANS\X^:RS:(h) 1 1.235V. 83.013 V, (c) H.9 % . 22. Consider the circuit o f Figure P20.22, in
-1 .2 % which C = I F, 7? = \Un{2) = 1.443 T=A
sec, I/, = 15 V. = 0.25, v^Q-) = = 1
20. The method o f Problem 19 can be extend V, and is a sequence o f rectangular pulses
ed to computation o f the steady-state response as shown in Figure P20.22b.
o f any first-order linear network to a square (a) Show that for the first cycle, 0 < r < 4
wave input. Let the transfer function o f the sec, the complete response is V(^t) =
first-order linear network be [15 - 1 5 (0 .5 )1 + (0 .5 )r = 15 -
14(0.5)^ V for 0 < / < 1 sec and V(^t) =
/ ’[Output(/)] H{0) 8(0.5)^ V, for 1 < t < A sec. Sketch the
H(s) =
/ ’[InputC/)] 1+T5 v^t) waveform for the first c)'cle 0 <
and the input be the square wave of Figure r < 4 sec.
P 20.19b. Show that in steady state, (b) Show that v^JA') =
= 1 V.
w , V
* max - V^min (c) Use the result of part (b) to verify that
Output(/),„,„ = H{0)
0.57 the V(\t) waveform for any subsequent
l+ e ^ cycle, n T < t < {n + U T is identical to
and that o f the first cycle except for a time
Figure P20.21
-•S
A.\S\X ER,S: (a) H (s) = : ( b ) ± 0 .2 4 5 V
i()()0.v +
Chapter 20 • Brief Introduction to Fourier Series
R R. R.
'. W © v jt )
'» '« 6
O '
------ o
(a) Figure P20.23
Avjt)
■> t
0 pT T T+pT
(b)
Figure P20.22
Vw/
^ w
I N D E X
W
122 Index
■^ w
W
24 Index
n
phasor relationships for resistors, nodal and loop analysis in i^-domain, 6 3 4
capacitors, and, 4 4 9 - 4 5 5 SSS analysis with phasors, 4 8 3 - 4 8 8 n
series and parallel inductors and statements of, 6 1 , 6 3
capacitors, 2 9 3 - 3 0 2
Initial-value theorem, 7 2 6 - 7 2 9 , 7 4 0 Ladder network, 2 0 6 - 2 0 8 /n
Input admittance, 9 7 4 , 10 1 0 Lagging, 4 5 1 , 5 1 8 , 5 2 6
Input attenuation, 1 0 5 1 - 1 0 5 2 Laplace, Pierre Sim on, 543
n
Input impedance, 9 6 2 , 9 9 0 , 1010 Laplace transform analysis
Instability, 3 6 2 basics, 5 4 3 - 6 0 2
Instantaneous power, 2 2 - 2 5 , 4 1 , 2 7 9 , 3 0 7 , basic signals, 5 4 8 - 5 5 4 r>
5 0 0 - 5 0 5 , 5 3 2 , 533 basic signals and signal
Instantaneous stored e n e ^ , 2 8 0 , 291 representation, 5 9 2 - 5 9 5
o
Insulators, 3 , 41 circuit responses, applied to
In t^ ra tin g factor m ethod, 3 2 7 , 3 6 4 differential equations, 6 0 0 - 6 0 2
Integration, Laplace transform m ethod and, frequency shift property, 5 7 5 - 5 7 6 n
5 7 9 -5 8 0 , 5 7 9 -5 8 2 integration, 5 7 9 - 5 8 0
Integrator, 3 5 9 , 3 6 4 , 6 2 9 , 6 5 8 inverse Laplace transform, 5 6 5 - 5 7 5
ideal, 3 6 0 Laplace transform pairs, 564
leaky, 3 5 9 , 3 6 4 one-sided, 5 5 4 - 5 6 3
Int^ro-difFerential equations, 5 8 5 - 5 9 0 overview, 5 4 7 - 5 4 8
Internal resistance, 6 4 by partial fraction expansion, 5 9 8 -6 0 0
Inverse Laplace transform, 5 6 5 - 5 7 5 , 591 properties, 5 8 4
Inverting amplifier, 159, 160, 173, 177 properties, finding, 5 9 5 - 5 9 8 o
Isolation amplifier, 164 second order tim e domain methods,
5 4 4 -5 4 7
Joule, 12 solution o f integro-difierential
equations, 5 8 5 - 5 9 0
KUo, 2 9 time dif&rendadon formula, 5 7 6 -5 7 8 r\
Kilowatt-hour (kW h) meter, 5 2 6 time-/frequency-scaling property,
KirchhofTs current law (K C L ), 5 1 - 1 0 6 , 53, 5 8 2 -5 8 3
5 5 - 6 0 , 89 V‘i relationship o f capacitor, 5 8 0 - 5 8 2
definitions, 5 3 - 5 4 circuit applications, 6 0 3 - 6 8 1
explained, 5 5 - 6 0 design o f general summ ing
first order R L and R C circuits, 3 4 8
n
in t^ rators, 652-^57
for Gaussian curves/surfaces, 59 equivalent circuits for initialized
Laplace transform analysis circuit inductors and capacitors, 6 1 8 - 6 2 5 o
applications, 6 0 9 fluorescence, 6 0 3 - 6 0 4 , 6 2 3 - 6 2 5
Laplace transform m ethod and, 5 5 7 , 591 impedance, admittance, voltage
linearity, superposition, source division, source transformations, r\
transformation and, 191 Thevenin, N orton equivalents,
nodal and loop analysis and, 1 0 7 -1 0 8 6 5 9 -6 6 2
o
nodal and loop analysis in ^-domain, 6 3 4 impedance and adm ittance, 6 0 5 - 6 0 9
SSS analysis with phasors, 4 8 3 - 4 8 8 impedance and admittance,
statements of, 55 manipulation, 6 0 9 - 6 1 7
KirchhoflPs voltage law (K V L ), 5 1 - 1 0 6 , impulse and step responses, 6 3 1 - 6 3 4
5 3 - 5 5 , 6 0 - 6 4 . 90 nodal and loop analysis in x-domain,
capacitors and, 2 8 7 6 3 4 -6 4 0
n
defm itions, 5 3 - 5 4 op amp integrator design and, 6 8 0
explained, 6 0 - 6 4 response calculation with initial
first order R L and R C circuits, 3 4 8 conditions, 6 6 8 - 6 7 6
Laplace transform analysis circuit sawtooth waveform generadon and,
applications, 6 0 9 6 8 0 -6 8 1 o
Laplace transform m ethod and, 5 5 7 , 591 switched capacitor circuits and
linearity, superposition, source conservation o f charge, 6 4 5 - 6 5 2
transform ation and, 191
r
Index 112^
w
1126 Index
r\
diflferential equation, Laplace transform, Modulus, 4 3 5 , 4 8 2
phasor models of, 8 9 3 - 8 9 5 M ultiplication-by-r property, 5 5 9 - 5 6 0 r^
dot placement, M, and basic equations, Mutual inductance (M), 8 8 5 , 9 3 7 . See also
9 3 9 -9 4 0 M agnetically coupled circuits
general analysis o f circuits with coupled dot convention and, 8 8 6 - 8 9 3
inductors, 9 4 3 - 9 4 7 rule for induced voltage drop due to, 8 8 7
ideal transformers, 9 1 4 - 9 2 4
magnetically coupled, defined, 883 Nano, 2 9
modek for practical transformers, 9 2 4 -9 3 0 Natural fi-equency o f a circuit, 3 6 5 , 3 8 8 ,
mutual inductance and dot convention, 417, 687 o
8 8 6 -8 9 3 Natural modes o f vibration, 3 2 6
M agnitude (modulus), 4 3 3 , 4 8 2 Natural response, 3 5 2 , 3 6 5 , 7 0 0 , 741
M agnitude response, 7 0 8 - 7 0 9 , 7 4 0 Network equivalency, 3 7 3
r>
M agnitude scale factor, 6 3 0 , 7 1 4 - 7 2 6 , Network fim ction, 6 2 6
7 4 0 -7 4 1 Networks r s
M atching networks, 8 4 8 , 861 equivalent networks, 2 1 2 - 2 1 3
M A TLA B two-terminal networks, 2 0 9 , 2 1 2 , 2 1 5 ,
first order R L and R C circuits, 3 2 4 , 3 3 2 , 2 2 8 ,2 5 7 r\
341 with V/I division, SSS analysis, 4 9 0 - 4 9 1
for fi«quency response plots, 4 7 2 —4 7 6 Nodal analysis, 1 0 7 -1 5 4 , 110, 141
n
Laplace transform and partial fraction concepts of, 1 1 0 -1 1 1
expansion, 572 defined, 110, l 4 l
linearity, superposition, source floating voltage sources, 1 2 3 -1 2 8 ,1 4 6 -1 4 8 o
transformation and, 192 general SSS analysis, 4 9 2 - 4 9 4
residue comm and, 671 grounded voltage sources, 1 1 1 -1 2 2
series-parallel resistive circuits and, 1 0 4 -1 0 6 history of, 1 0 7 -1 0 8 r>
M atrix inverse, 141 Laplace transform analysis circuit
M atrix notation, 6 3 6 applications, 6 3 4 - 6 4 0
M atrix partitioning, 9 6 6 , 101 0 linearity and, 195
M axim dly flat, 1 0 5 2 -1 0 5 3 , 1072 loop analysis and, 1 2 8 - 1 3 9
M axim um Power Transfer theorem, o f pressure-sensing device, 4 7 7 —481 o
2 4 9 -2 5 6 , 257, 5 3 9 -5 4 2 . See also N orton steady-state circuit analysis using phasors
theorem; Thevenin theorem and, 461
r>
for A C circuits, 5 2 6 terminology, 1 0 9 -1 1 0
defined, 5 3 2 Node, 54, 90
Maxwell, Jam es Clerk, 107 Node voltage, 6 0 , 9 0 , 109, I 4 l r\
Maxwell’s equations, 2 7 3 , 3 0 7 Non-ideal constant current source, 87
Mega, 2 9 Non-ideal constant voltage source, 87
M em ory Non-ideal dependent current source, 8 7 n
capacitors and, 2 8 6 Non-ideal dependent voltage source, 87
inductors and, 2 7 5 Non-inverting amplifier, 162, 163, 177
Memoryless device, 41 Nonlinear, 8 0
Mesh, I 4 l Nonlinear controlled source, 80
Mesh analysis, 128, 141, 461 Normalized Butterworth loss functions, 1072
Mesh current, 128, 141 Normalized circuits, 7 1 4
Metastable, 6 9 0 Normalized fi^quency, 1 0 3 5 - 1 0 3 6
m ho, 3 0 , 4 0 , 41
o
N orton, E . L , 2 2 7 , 1069
M icro, 29 N orton theorem, 2 2 7 - 2 6 8 . See also
Microwaves, second order linear circuits and, M axim um Power Transfer theorem;
o
3 7 9 -3 8 0 Thevenin theorem
MilU, 2 9 corollary to Thevenin and N ortons
Modified loop equations, 136 theorem’s for passive networks, 2 3 6 o
Modified n o ^ analysis (M D A ), 110, 123, equivalent o f one-port, 1010
140, 141
Index 1127
w
n
1128 Index
n
distinct complex poles, 5 7 3 - 5 7 5 R C circuit. See RL/RC circuits
Laplace transform analysis transfer R C to C R transformation, 1062, 1072
n
function applications, 6 8 5 - 6 9 3 , Reactance, 4 5 6 , 4 8 2 , 8 4 3
7 4 2 -7 4 4 Reactive power, 5 1 0 - 5 1 5 , 5 3 2
pole frequency, 8 1 9 Real part, 4 3 5 , 4 8 2 r\
pole Q. 8 5 2 , 861 Real power, 5 3 2
pole-zero, 8 1 6 Reciprocity
r>
o f rational fiinction reciprocal networks, 9 9 8 , 1010
fmite, 5 9 1 ,6 8 5 ,7 4 1 reciprocal two-port, 1010
simple, 5 9 1 , 741 two-ports and, 9 9 7 - 1 0 0 2 , 1 0 2 9 -1 0 3 0 o
repeated, 5 6 9 - 5 7 2 , 6 8 5 Rectangular approximations to signals,
Potential difference, 12, 16 7 9 4 -7 9 5
Power, 1 6 - 2 5 , 4 1 . See abo Sinusoidal steady Rectangular coordinates, 4 3 6 , 4 8 2
state (SSS) Region o f convergence (R O C ), 5 56, 591
apparent, 5 1 0 - 5 1 5 , 531 Repeated pole, 685
available, 528 Residue theorem, 5 65, 5 6 6
average power and SSS power Resistance, 3, 3 0 , 4 1 , 4 8 2
calculation, 5 0 0 - 5 0 5 , 5 33, 5 3 4 - 5 3 6 converter, 2 4 7
complex, 5 1 0 - 5 1 5 , 5 3 1 , 5 3 6 - 5 3 7 equivalent, 6 5 , 6 9 , 70
conservation o f power, 19, 2 0 , 3 6 , 4 0 , internal, 64
5 1 5 . 5 3 1 ,5 3 7 negative, 6 8 , 2 4 6
half-power frequencies, 815 output, 2 5 0
half-power points, 1033, 1072 parallel, 6 8 - 7 3 r>
instantaneous, 2 2 - 2 5 , 4 1 , 2 7 9 , 3 0 7 , phasor admittance/impedance, 4 5 6
5 0 0 - 5 0 5 , 53 2 , 533 transfer, 79
lower half-power frequency, 8 1 6 , 861 Resistivity, 3 3 , 41 r\
M axim um Power Transfer theorem, Resistors. 15, 16, 2 2 , 3 0 , 3 1 , 4 1
2 4 9 -2 5 6 , 257, 526, 532, 5 3 9 -5 4 2 ideal (linear), 15, 37
r\
power gain, 80 impedance of, 6 0 7 - 6 0 8 , 6 5 8
reactive, 5 1 0 - 5 1 5 , 5 3 2 nonlinear, 41
real, 5 3 2 phasor relationships for inductors, r\
upper half-power frequency, 8 1 6 , 8 6 2 capacitors, and, 4 4 9 - 4 5 5
Power factor (pO. 5 1 7 - 5 2 6 , 5 3 2 . 5 3 7 - 5 3 9 Resonance. See abo Band-pass circuit
Power gain, 80 phenomenon, 8 3 9 - 8 4 2 r\
Practical inductors, 8 3 0 resonance frequency, 8 6 2
Practical source, 80 series and parallel resonant circuits,
r\
Practical transformers, 9 2 4 - 9 3 0 8 4 2 -8 4 6
Primary coils, 8 9 8 , 9 1 7 , 9 3 8 series-parallel resonant circuits, 8 4 6 -8 5 1
Product over sum rule, 71 Response n
Proportionality, 2 0 1 - 2 0 8 classifications, 3 5 2
natural frequency o f a circuit and, 215 complete, 3 5 2 , 3 6 4
property, 215 forced, 3 5 2 , 3 6 4
proportionality property, 2 0 1 - 2 0 8 natural, 3 5 2 , 3 6 5
Pulse function, 549 source-free, 3 2 8 - 3 3 6 , 3 6 5
step, 3 3 6 - 3 4 7 , 365
Quality factor, 8 1 6 unstable, 3 3 5 , 3 6 5
o f a band-pass circuit, 861 zero-input, 3 2 8 - 3 3 6 , 3 5 2 , 3 6 5
o f a capacitor, 861 zero-state, 3 5 2 , 365
o f a coil, 861 Reverse open-circuit voltage gain, 9 8 6 - 9 8 7 ,
r\
o f L and C components, 8 3 0 - 8 3 3 1010
o f a reactive com ponent, 861 Reverse voltage gain, 9 9 2
R L C circuit. See abo Second order linear n
Ramp function, 55 0 , 5 6 0 , 5 9 1 , 741 circuits
Rational function, 5 6 5 , 6 8 5 , 741 computing SSS response and, 4 4 2 —444
Index 1129
r>
using phasors, 4 6 1 —4 6 6 Square wave, 3 4 5 , 3 4 6
power calculation, 4 9 9 - 5 4 2 Stable circuit, 4 3 3
complex power and, 5 1 0 - 5 1 5 defined, 4 8 2
complex power conservation, 5 1 3 -5 1 7 stability and boimdedness, 6 9 0
effective value o f signal and average Stable transfer function, 741 n
power, 5 0 6 - 5 0 9 Steady state, 4 3 3 . See also Fourier series
instantaneous and average powers, Steady-state analysis, 6 0 4 , 6 5 8
5 0 0 -5 0 5 Steady-state calculation, 7 4 9 - 7 5 3
m aximum power transfer in SSS, Steady-state circuit output response, 701
5 2 6 -5 3 0 Steady-state response, 6 9 6 , 741
power factor/power (actor correction, Step function, 4 1 7
5 1 7 -5 2 6 Step response, 3 6 5 , 4 1 7 , 6 3 1 - 6 3 4 , 741
single-frequency analysis with Stored energy, 3 7 6 , 9 1 0 - 9 1 4
effective values, 5 0 9 - 5 1 0 Stray capacitance, 3 6 1 , 3 6 2 , 3 6 5
Sm oothing, 2 6 9 - 2 7 0 , 3 0 3 - 3 0 5 Supermesh, 136
Supernode, 123
n
Solar cell, 3 7 6
Source, 16 Superposition, 1 9 1 -2 2 5 , 3 6 5 . See also
controlled, 2 8 , 2 9 , 7 8 , 80 Linearity; Source transformation n
current, 56 o f average power, 5 0 3 - 5 0 5
current-controlled ciurent (C C C S), 28, 39 linearity and, 1 9 3 -2 0 0 , 3 4 7 - 3 5 2
current-controlled voltagp (C C Y S), 28, 39 proportionality and, 2 0 1 - 2 0 8 n
floating, 1 2 3 -1 2 8 , 140, 1 8 8 -1 8 9 source transformations and, 2 0 8 - 2 1 2
ideal current, 2 7 , 38 source transformations and equivalent
independent current, 2 7 , 38 networks, 2 1 2 - 2 1 3
independent (ideal) current, 2 7 , 38 superposition property, 2 0 1 , 215
independent (ideal) voltage, 2 5 - 2 7 , 3 7 Susceptance, 4 8 2 , 8 6 2 r\
nonideal, 8 5 - 8 8 Switched capacitor circuit, 6 4 5 - 6 5 2 , 6 5 8
nonlinear, 80 Switching
nonlinear controlled, 80 first order R L and R C circuits, 3 2 3 , 3 4 5
practical, 2 6 , 7 9 , 8 0 , 81 in R L C circuit, 6 4 0 - 6 4 5
v-i characteristics, 8 0 , 86 Sym m etric matrbc, 141 n
voltage, 6 4 Sym m etry properties o f Fourier series, 1094,
Source free, 4 1 7 1109
Source-free response, 3 2 8 - 3 3 6 , 365 System identification, 7 6 5 r>
Source-free second order linear circuits,
3 8 5 -4 0 0 Tank circuit, 8 4 3 , 8 6 2
r\
Source transformation, 1 9 1 -2 2 5 . See also Tank frequency, 8 4 3 , 8 6 2
Linearity; Superposition Temperature measurement, nodal and loop
defmed, 2 1 5 analysis, 143 r\
explained, 2 0 8 - 2 1 2 T-equivalent circuit, 9 0 8 , 9 3 8 , 1010
Laplace transform analysis circuit Tera, 29
Term inated secondary, magnetically coupled
r\
applications, 6 5 9 - 6 6 2
linearity and, 1 9 3 -2 0 0 circuits and, 9 0 1 - 9 0 9
linearity and superposition, 2 0 1 - 2 0 8 Term inated two-ports, 9 7 3 , 1010
source transformation theorem and Thevenin, M . L ., 2 2 7
equivalent networks, 2 1 2 - 2 1 3 Thevenin theorem, 2 2 7 - 2 6 8 . See also
source transformation theorem for M axim iun Power Transfer theorem;
independent sources, 2 0 9 N orton theorem
Source transformation property, 6 1 5 , 6 5 8 corollary to Thevenin and N ortons
S P IC E (Simulation Program with Integrated theorem s for passive networks, 2 3 6
n
Circuit Emphasis), 178 equivalent o f a one-port, 1010
first order R L and R C circuits, 3 4 6 general approach to finding Thevenin
for frequency response plots, 4 7 2 - 4 7 6 and N orton equivalents, 2 3 6 -2 4 1
W ien bridge oscillator and, 4 1 4
Index 1131
O
Laplace transform analysis circuit SSS for stable networks and systems,
applications, 6 5 9 - 6 6 2 7 0 1 -7 0 7
M axim um Power Transfer theorem and, stability problems, 7 4 5 - 7 4 6
2 4 9 -2 5 6 steady-state calculation, 7 4 9 - 7 5 3
N orton theorem and Transfer impedance, 6 2 6
active networks, 2 4 1 - 2 4 5 Transfer resistance, 7 9
linear passive networks, 2 2 9 - 2 3 6 Transformers. See also M agnetically coupled
op amp circuits, 2 4 6 - 2 4 9 circuits
SSS analysis, 4 9 1 - 4 9 2 defined, 885
(Thevenin’s equivalent resistance), 2 5 7 ideal, 9 1 4 - 9 2 4 , 9 3 0 - 9 3 6 , 9 5 0 - 9 5 6
SSS power calculations and, 4 9 9 practical, 9 2 4 - 9 3 0
steady-state circuit analysis using phasors Transient analysis, 5 4 4 , 6 5 8
and, 4 6 4 - 4 6 6 Transient response, 4 3 3 , 6 9 6 , 741
superposition and, 3 4 9 Transistor photo timer, 3 7 7
Thevenin equivalent, deBned, 2 2 9 , 2 4 2 Translation o f a plot, 1 0 9 2 - 1 0 9 3 , 1109
Thevenin equivalent resistance, 231 Translation property o f Fourier series,
Thevenins equivalent circuit, 2 5 7 1 0 9 3 -1 0 9 4 , 1109
Thevenin theorem for passive networks, Transmission {t-) parameters, 9 6 1 , 9 9 1 - 9 9 4 ,
231 1010, 1 0 2 3 -1 0 2 5
two-ports and, 9 6 4 - 9 6 8 Triangular waveform, 3 5 5 - 3 5 7
3 dB bandwidth, 8 1 5 - 8 1 6 , 861 Trigonom etric Fourier series, 1089, 1109
o f B P filter, 1071 Tuned circuit, 8 4 4 , 8 6 2
o f LP filter, 1071 Two-dependent source equivalent circuit,
3 dB down point, 1033 9 7 2 - 9 7 3 , 1010
3 dB firequency, 7 3 6 , 7 4 0 Two-ports, 9 5 9 - 1 0 3 0
Three-term ind device, 8 0 , 9 0 8 admittance parameters, 9 6 8 - 9 7 3
T im e constant, 3 2 9 , 3 6 5 amplifiers, 9 5 9 - 9 6 0
L >
T im e differentiadon formula, 5 7 6 - 5 7 8 general relations am ong two-port
T im e dom ain circuit response com putation. parameters, 9 9 4 - 9 9 7
See Convolution A-parameters, 9 6 1 , 9 8 5 - 9 9 1 , 1 0 2 0 -1 0 2 3
T im e invariance, 6 5 8 , 7 9 3 , 7 9 8 impedance and gain calculations o f
Tim e-scaling property, 5 8 2 - 5 8 3 terminated two-ports modeled by
T im e shift property, 5 5 8 - 5 5 9 , 5 9 1 , 7 6 7 z-parameters, 9 8 1 - 9 8 4
Touch-tone phones, 8 1 1 - 8 1 2 impedance parameters, 9 7 6 -9 8 1
Transconductance, 7 9 one-port networks, 9 6 1 , 9 6 2 - 9 6 8 ,
Transfer admittance, 6 2 6 1 0 1 1 -1 0 1 2
Transfer conductance, 7 9 parallel, series, cascaded co n n eaio n s of,
Transfer function, 6 2 5 - 6 3 0 , 6 5 8 , 6 8 3 - 7 6 1 1 0 0 2 -1 0 0 9
applicadons and bode techniques, 7 6 0 -7 6 1 parameter conversion and inter
band-pass circuits, 8 6 3 - 8 6 8 , 8 7 6 - 8 7 9 connection o f two-ports, 1 0 2 5 -1 0 2 9
bode plots, 7 3 0 - 7 3 6 reciprocity, 9 9 7 - 1 0 0 2 , 1 0 2 9 -1 0 3 0
classification o f responses, 6 9 3 -7 0 1 ^-parameters, 9 6 1 , 9 9 1 - 9 9 4 , 1 0 2 3 -1 0 2 5
D C m otors and, 7 3 6 - 7 3 9 transmission parameters, 9 9 1 - 9 9 4
electric m otor analysis and, 6 8 3 ^ p aram eter analysis o f terminated
frequency response, 7 0 7 - 7 1 4 , 7 5 3 - 7 5 9 two-ports, 9 7 3 - 9 7 6 , 1 0 1 2 -1 0 1 6
frequency scaling and magnitude scaling, z-parameters, 1 0 1 6 - 1 0 2 0
n
Undamped natural frequency, 8 4 3 , 8 6 2 Wattage, 19, 41
Underdamped circuit, 3 9 0 , 3 9 2 , 4 1 7 Waveform r\
U nit impulse function, 5 5 2 , 591 effective value of, 505
U nit step fu n a io n , 3 2 4 - 3 2 5 , 3 6 5 , 5 4 8 , 591 periodic current, 2 7 9 - 2 8 0
U nity coupling, 9 3 1 , 9 3 8 period of, 4 3 3 n
Universal resonance curves, 8 2 6 , 8 6 2 sawtooth, 3 2 2 - 3 2 3 , 3 6 5
Unstable response, 3 3 5 , 3 6 5 scaled sum o f waveforms, 4 1 7
n
Upper half-power frequency, 8 1 6 , 8 6 2 second order linear circuit and, 3 8 3 - 3 8 5
sinusoidal voltage, 4 1 2
v-i characteristic, 2 6 , 3 7 - 3 8 , 41 triangular, 3 5 5 - 3 5 7 , 6 8 0 -6 8 1
v-i relationship o f capacitor W ien bridge oscillator, 4 1 2 , 4 1 4 , 4 7 5 —476
Laplace transform m ethod, 5 8 0 - 5 8 2 W ire, resistance o f copper wire, 33. See also
Virtual ground, 160, 178 Resistivity
V inually grounded, 160
Virtual short circuit model, 1 5 8 ,1 6 0 , 1 7 7 ,1 7 8 j^-parameters
n
V 257 analysis o f terminated two-ports,
Voltage, 2, 16 9 7 3 - 9 7 6 ,1 0 1 2 - 1 0 1 6
branch, 107 z-parameters and, 9 7 6 - 9 8 1
gain, 117
node, 6 0 Zero
regulators, 3 0 3 finite zero, 6 8 5
Voltagp division, 9 0 ,2 9 5 ,2 9 9 ,3 0 0 ,4 8 2 ,6 5 8 Laplace transform analysis transfer
formula, 6 6 , 4 5 6 function applicadons, 6 8 5 - 6 9 3 ,
Laplace transform analysis circuit 7 4 2 -7 4 4
applications, 6 5 9 - 6 6 2 Zero-input response, 3 2 8 - 3 3 6 , 3 5 2 , 3 6 5 ,
Voltage drop, 10 4 8 2 , 6 9 3 , 741 n
Voltage follower, 1 64, 178 Zeros o f rational function, 5 6 6 , 5 91, 741
Voltage gain, 80 Zero-state response, 3 5 2 , 3 6 5 , 6 9 3 , 7 4 1 , 7 9 8
Voltage (potential difference), 41 Z Js). 9 4 0 - 9 4 3
Voltage regulator, 3 0 7 2-parameters, 9 7 6 - 9 8 1 , 1 0 1 6 -1 0 2 0
Voltage source, 4 1 , 6 4
nonideal, 8 6
v-i charaaeristics, 8 0 , 8 6
Voltage-to-current converter, 188
n
Voltage transformation property, 9 1 4 , 9 3 1 ,
933, 937, 938 n
Voltmeter, 14, 7 6
sensitivity, 7 7
n
r s
o o