ac cir cuit

analysis
with illustzative
pFoblenns

Phillip Cutler
Education Research Associates
and
Orange Coast
College
Costa Mesa
California

McGraw- Hill Book Company

New York
St. Louia
San Francisco
Düeseldorf
Johanneeburg
Kuala Lumpur
London
Mexico

New Delhi
Panama
Rio de Janeiro
Singapore
Sydney
Toronto
 A need frequently arises to examine the frequency re-
sponse (both amplitude and phase) of a circuit or ampli-
fier. This is particularly true in more ad vanced courses
w’here attention is focused on the behavior of feedback
amplifiers or control systems. Hence it is advisable to
introduce the subject at an elementary level so that we
will be on firmer ground in the future.

11 -1 The transfer function

Most any circuit or system that we might consider has
some kind of input and out put. Generally we are in-
terested in output voltage versus input voltage or voltage
‘/ ‘( the Bode plot gain as a function of frequency. but this need not be the
case. We could just as well be interested in current gain,
or out put voltage versus input current. and so on. When
dealing with control systems we might be interested in
motor speed versus armat ure voltage, or height of liquid
in a tank versus a potentiometer setting, etc. In any
event. the output versus input relationship as a function
of frequency is. for our purposes, called the transfer
function. The transfer functions of most concern to us
are exclusively electrical in nature and generally a
voltage gain.
There are a few different methods of plotting transfer
functions versus frequency. One of the simplest and most
useful is called the Bode plot. after a famous scientist who
did some outstanding research in feedback theory. It
turns out that to conveniently construct a Bode plot, the
transfer function is best expressed in certain algebraic
forms which are readily conveyed by example.

11-2 The low-pass filter

The first circuit we will develop the transfer function
for is the low-pass filter pf Fig. 11 -1. We seek the voltage

Ez 1

*’ 1+ i

f/ "
RC

Ficiunr 11- 1

gain A, —— E 2/E, as a function of frequency. Clearly we

have a voltage divider here and hence
E
z

the Bode plot 1 149

 Remember, we can transfery from numerator to denomi- As f increases, we would expect dp to decrease and
nator and vice versa simply by changing its sign, since exhibit some phase lag since E 2 is taken off the capaci tor
this corresponds to the similar transfer of an angle in and capacitor voltage lags capacitor current by 90-
polar form. Inspection of Eq. (I 1-lb) [or Eq. (11-la)] indicates that
Finding a common denominator and simplifying t I), whenJ: the frequency of the in put signal E, . is the same as
we obtain #, the quantity /' , = 1 and hence the magnitude of
t t A,. —— 1 ' = 0.707 and the phase angle is 0 = — tan '
2) 1 = — 45'
As f becomes very large, and in particular for J /{,
where the j term in Eq. ( 11-1s) approaches infin it y, so that
3) z = /tC A, I 'j x —— —y0, which means the gain magnitude
approaches zero and the gain phase is — 90‘. This is to be
Now it turns out that the product flC is called a t ime expected. since at high frequencies Xp <t fl. so that the
constant and it has the dimensions of seconds. since ' circuit impedance to E t is essentially resistive and hence
the current tends to be in phase with E t . However, this
V charge V A I sec = sec approximately zero-degree phase angle of the current is
A V A Y added to the 90° phase lag ( — 90‘) of the capacitor to
Now it is convenient to reason that the time constant r# yield the phase of E 2. Hence E 2 lags E , by approximately
is the period corresponding to a fictitious radian fre- 90’ at / Ji. Of course. with XC il pproaching zero in the
quency «a, so that high-frequency spectrum f ñ fz), X C <i A find hence the
magnitude of E, approaches zero. Thus the gain magni-
tude approaches zero.
or Now work through problems PS 11-1, PS 11-2. and
l PS 11-3.
6)
11-3 The high-pass filter
The use of the h subscript will be clarified later. S b' Another circuit commonly encountered is the high-pass
ing {ñ ) into (2) yields, for the low-pass filter,
filter of Fig. 11-2. This circuit is commonly used as a
1 ) coupling network in many amplifier circuits where E,
7) l f2
1
fh

Notice how simple this expression is and, in particular.

note its form. In polar form we should express (7) as
E2 f
8) r— 1
Ei — --‘-
i + (///›)‘ /tan* '///s
Juan
Hi@-pass filter
2016-08-17 20:05:55
--------------------------------------------
FIGURE 11-2
¿Cómo suele usarse un filtro pasa-
For reasons to become apparent later, the frequency f altas? RESPUESTA: Como red de
is called the corner or break frequency and it is extremely contains a dc as well as an ac component.
acoplamiento en los circuitosUsually we
important in constructing Bode plots. want only the ac amplificadores.
component to appear at the output.
lt might be interesting to correlate the mathematical The capacitor C, also called a blocking capacitor,
form of I I - I a (or I I - I b) with a qualitative analysis of the assures this, since only a time-varying voltage can cause
circuit. Surely at low frequencies c 'S so much larger capacitor current to flow. 2
than A that the magnitude of E 2 - E, . Correspondingly We can develop the voltage transfer function by the
in (11- l‹iJ for / approaching zero. or more significantly voltage divider relationship and writing
for f fi /„ the y term approaches zero and A,. 1 %.
1) ' A
' In a similar manner, the time constant 1./'A in an inductive circuit has
the dimensions of seconds. since
V
Recall that i, = C

the Bode plot |

 If z e let Now work through problems PS 11-4. PS 1 I -5. and
PS 11 -6.
Note that, in all the pro blems. we have tried to ex-
w’e can rewrite (1 I as press the t ransfer function so that it contains terms
of the form y.x and or 1 y . This will be possible
1
( 1 I -2‹i) w’henever we deal w’ith circuits containing a single-
energy storage element such as just one capacitor or one
inductor.
Now Eg. 13) is perfect lj i alid. h ut it is more desirable
I al so for reasons to become a pparent later l to get rid of
the negative denominator term. This may be accom- p! 11-4 The decibel
ished by m ultiply ing the numerator and denominator c›f The transfer functions encountered in practice w ill in
(3) by /J /}. lf j on do this and rearrange tc•rms. t he re general contain factors in the n umer:itor and deno mi-
nator. Therefore evaluation of the transfer function at a
particular frequency requires the process of multiplica-
tion and/or division, neither of w hich is conveniently
performed by graphic techniques. Since a graphic display
of the transfer function is our goal. we seek an alternative
approach.
You will recall that with the aid of logarithms. the
operations of multiplication and division are converted
to addition and subtraction, since
— tan ' / /,
1) log ab —— log a + log 6 ( 1 1-3‹i)
a
( 1 1-2c) ( 11-3h)
log
.Alt hough (4) appears more complicated than t 31. it b
is somewhat easier to Bode plot. which after all is our and raising a number to a power converts to multiplica-
goal. tion
To correlate (5) with a qualitative analysis of Fig. 11-2,
log a" x log a (1 l -3c)
we may reason as follows. In the low-frequency spectrum,
say for / ñ J}, the n umerator magnitude fi 0 and the Now graphic addition and subtraction is no problem
denominator magnitude - 1. Hence form <t /t, | A, | 0. since we need only count divisions or lay off distances
Physically we should expect this. since Np fl at low with a pair of dividers. For this reason, among others,
frequencies. The question is what constitutes low fre- the vertical scale in our graphic display will be expressed
quencies‘? The answer is that J} in the high-pass or fz in a logarithmic unit called the decibel, abbreviated dB.
in the low-pass filter forms a convenient spot frequency The decibel is formally defined as a unit of power gain
to separate the low.’ and high frequency spectrum. given by
Since at low frequencies Ap ñ A, the current leads
E , by’ al most 90" ; hence the output voltage which is taken
off the resistor should be in phase with the current. 4) A —— 10 log dB (11-4a)
Inspection of the phase portion of (5) indicates that for Equation (11 -4oJ is read as, “the power gain Ap, ex-
f /• ]) the tangent is 0‘ and hence the net angle is pressed in dB, is equal to 10 times the log of the power
90‘ — 0“ or fi 90°. ratio ’P› .” A more common form of expressing Eq.
At J‘ = f where f is the corner or break frequency, (11 -4a) is shown in Eq. (11-4b):
N, | = 1/p 2 and the phase is 90° — 45‘ = 45‘.
In the high-frequency spectrum (/ ñ Jj) we see from 5) (11-4b)
Eq. (4) lhat the reactive term in the denominator is
predominant, and hence A,. approaches 10. This is Although Eq. (11-4o) is in better mathematical form,
in agreement with the physical facts, for Np ۥ R. E 2 - Eq. ( 11 -4h), which is read and interpreted in the same
E . Furthermore, the current in this case is essentially way, is more prevalent and hence both forms will be
determined by R and hence is almost in phase with E, . used here.
Since E 2 is taken off the resistor which has no phase Symbols other than A are commonly used to designate
angle associated with it, it follows that E 2 is essentially the gain or ratio between two quantities. For example,
in phase with E t . the symbol R or G is often used, and as a further refine-

150 circuit
ment the symbol G is used if P2 > P or x if P < P . A —— 20 log } dB = 20 (log l — log 2) dB
We will, however, stick to the symbol .4 irrespective of Since the logarithm of 1 to any base is zero, the above

the Bode plot |

whether P2 O r P is the greater n umber. reduces to
Since electrical power is given by P —— E 2 'ft or 2
B,
we may express Eq. (4) as A —— 20 (0 — log 2) d B = — 20 log 2 dB
2
2 E 2 /It 2 = — 20 t0.301) dB = — 6.02 d B
6) A —— 10 log dB = 10 log dB
E /R Now work through problems PS 11-7, PS l 1-8. and
PS 11-9.
= 10 log
dB
11 -5 The Bode plot
= t0 log dB + 10 lo dB A most useful means of displaying the amplitude and
A,
E
phase response of a circuit or system is to plot the
2
= 20 log d B + 10 log ' dB magnitude (amplitude response) in decibels versus fre-
R quency as one curve and the phase versus frequency as
Similarly a separate curve. The same frequency axis is used for
2 both curves. and semilog graph paper is used so that a
!z z
7) A —— 10 log 2
dB = 10 log dB wide frequency range can be displayed. Curves with this
t ype of display are known as Bode plots, and they are
most useful in electronic circuit and control system
= 10 log dB analysis.
To familiarize ourselves with Bode plots, we will first
’2
= 10 lOg
’
dB + 10 log dB construct the Bode plots of some basic transfer function
R, factors. For the moment we are not concerned with how
these functions are physically realized or even if they
dB + 10 log ’2 dB can be. We just want to learn the mechanics of con-
structing them. Once these basic forms are mastered,
Equations (6) and (7) are particularly useful if one we can move on to more complicated forms which are
wishes to determine power gain (or loss) by measuring essentially combinations of the basic forms
E and A. or I and A, which is usually easier to do than a Bode also showed that the phase and amplitude
direct measurement of power. However. many electronic response for most transfer functions are uniquely related
circuits are primarily designed from a standpoint of and that the slopes of the amplitude response are
voltage or current gain. and hence we drop the resistive indicative of the phase response. Since it is usually
terms in (5) and (6) so that the definition of voltage and easier to measure amplitude rather than phase, it be-
current gain in decibels are respectively, hooves us to see how we can predict the phase shift by
observing the amplitude response.
8) A, —— 20 log dB = 20 log |A,.| dB (11-5o)
11-6 Basic forms
9) A; -— 20 log ’2
dB = 20 log |A ,| dB (11-5h) The first transfer function form (factor) we will consider
is that of A —— k where A is the general symbol we are
The ratio of the output over the input is correctly using to denote a transfer function and k is some arbi-
called gain when two quantities are of similar units. trary constant, a number which is independent of fre-
However, it is common practice to refer to the output- quency. Thus if A ltXl, we have A in dB given by
input ratio as gain or system gain even when they are A 20 log ltXI dB = 40 dB. Similarly if A 0.2, we
different dimensionally. For example, we might consider have A 20 log 0.2 dB — 14 dB. These functions
the gain of an optical system as being the output light Bode plot as straight lines, as shown in Fig 11-3. Note
intensity of a lamp versus applied voltage. It should that there is no phase shift when the amplitude response
also be clear that gain may be a number more or less is flat {A —— a constant).
than unity. If the gain is less than unity, it represents Next assume a transfer function of the form
attenuation or loss within the network or system.
Note that if the gain is less than unity, the gain ex- l) A= y )’
pressed in dB will be negative, since the logarithm of a
fraction is negative. For example, where in, is some arbitrary constant which we will refer

152 | ac circuit
 2 3 4 5 6 7 g g j 2 3 4 5 6 7 8 9 › 2 3 4 5 6 7 8 9 0

A = 100 or 40 dB

30

20
B

-10
A = 0.2 ors — 14 d B

' I I i i 1 11 i ' l,

Bode plot of a connant

A =k

TI ’.I: I i . I . . : I 'II ! I ’' I . I.. 'I ' I ! I‘ ’ I I ' I . i!

FIGURE
I! . . I
1 t-3
'I ‘ ’ I

the Bode plot |

 Amplitude

Phase 315°

60 ” n —— 3 - n
270’
60dB/dec
Slope ' › 8dB/ .
50

I I i I
40
40dB/dec*
SI ope = 12dB/cut -
30 135°

90°
w/w,0. I0.312S10 dB/de
dB
10 = 20 log(cu/cu,)— 20— d.0206.02I 3.9820 45°

— \ 35°
—30

-40

— 50

-60
01 10 100

F IGU RE 11 -4

and the phase is given by

d B = 40 log I ‹'I u›,)— 4012 04 I 2.0427 9640 $ = n90° (I l-dc)
Another commonly encountered form of transfer
function term is“
Note that because this time the transfer function is
the square of the previous case (n l), the slope is 1
l) A = (1 l-7o)
exactly twice as much and the phase angle is doubled.
Therefore for n = 2 and a transfer function of the form
given in Eq. (1), you should think in terms o( a positive
slope of 40 dB/dec or about 12 dB/oct and associate a Inspection of (1) reveals that an increase of en causes the
180‘ phase shift with this slope. These results are also denominator to increase, since a›, is some fixed value,
shown in Fig. l 1-4. and hence it causes A to decrease. Therefore, the ampli-
A Bode plot for n = 3 is also shown in Fig. 11-4. As tude response should have a negative slope. We also set
might be expected. this curve also goes through 0 dB that for any value of n, the amplitude must be 0 dB at
at in = ‹o, with a positive slope equal to 60 dB/dec or u› = en,. Since the j represents 90° of phase in the
approximately 18 dB/oct and the associated phase denominator of A, the net phase angle in the denominator
angle is 270°. must be n 90°. Consequently, the net phase of A must
As you can see from Fig. 11-4, the slope in general is be — n 90°.
given by Although we will not take the time to derive the data

slope = n (20 d B/dec) n (6 dB/oct) (l 1-6b)

“This transfer function is sometimes referred to as the integral function.

154 | ac circuit anaI'ysis

for plotting .4. the results are given in Fig. 11-5. The
magnitude response from (1) is as follows: l)
1
A = Let us first consider the case where n — 1 and calculate
the magnitude and phase. This yields the data in the
following table:
3) 1„ = 20 log l — n log

Since I he log 1 to any base is zero. w’e hax’e

5.71 *6.6 45 63.4 7S‘784.3
1.(J0 5. 1[ 0.1)
.4,B —›+ 2010 — (1.043 1 4. 120.0
c
0.969 3.II [ 6.99
Inspect ion of Fig. 11-5 indicates that except for being
negative. the slopes and phase for various ve I ues of n
are identical with Fig. 11-4. Therefore, for a function of A Bode plot of the above data is in the solid lirle
the form in ( 1). curves of Fig. 11 -6. Figure 11-6 also shows some dashed
!ines which represent straight-line approximations to
slope = — n t?0 dB/dec) or — n (6 dB/'oct) ( \ \ -7bj the actual curves. The beauty of the Bode plot (among
Q = — ii 90° ( 11-7c) other things) lies in the ease w ith which the straight-line
approximations to the actual characteristics may be
Now’ work through problems PS 11-10 and PS 11-11. constructed. For example, consider the upper magnitude
The next form of a transfer function factor we might response curve. Note that it may be approximated by
familiarize ourselves with is the two dashed straight lines which intersect at in,. We

F IGU RE 11-5

the Bode pI•t *

 25
A = 1 +j
20 LL ' I

dB
Amplitude 15

l db
10

0 ,

90 9Q

75

g 60

lead

I °
4s

30

6’

10 2 >, 2o‹ 10>,

FJGUR£ l I -6

call w, the break, corner, or cutoff frequency because of For m u›, the straight-line approximation and actual
the abrupt change of slope exhibited by these lines at amplitude curve merge for all practical purposes and
in,. For in < in, the approximated magnitude response the phase error is again about 6°. As you can see, the
of this transfer function is flat. At a decade below w, or amplitude and—to a somewhat lesser extent —the phase
in = 0.1 in, the error in this approximation is negligible response can be readily approximated by merely draw-
since the above table of values indicates an actual value ing straight lines with well-defined slopes. If greater
of only 0.043 dB. In the lower figure the dashed-line accuracy is desired, we can use the easily remembered
approximation to the phase curve is only about 6° below amplitude correction factors of the following table:
the true value.
At an octave below m„ or en = 0.5 w„ the error 0.1 0.5 10
between the actual and approximate amplitude curves
dB
is only about 1 dB and the phase error is only about 6‘,
which is essentially the largest phase error at any point.
At in, we have the largest amplitude error, which is
only 3 dB, and the phase error is zero since we draw the
The above correction factors and the general shape of
phase approximation curve (dashed line) so that it
the phase and amplitude curves should be committed
intersects 0° at u›,/10 and 90° at 10 in,. A straight line
to memory.
connecting these two points will intersect the point
(m/m, = 1, Q = 45°). Now let us consider other values of n in the function
At in, the approximation to the amplitude response ‘ w
changes slope abruptly from 0° dB/dec to 20 dB/dec or 1) A + j— (11-8a)
6 dB/oct. At an octave above m„ or 2 u›„ the amplitude 156 | ac circuit analysis
error is such that the true response is only about 1 dB
high and the phase error is about 6° low.
which has an amplitude given by
2) A —- n 20 log 1 + (w’ m,)2 (11-8fi)

the Bode pI•t *

and a phase angle You will notice that this expression is the reciprocal of
3) $ = n tan ' "’ ( 1 I-8r‘l Eq. ( I I -8a), which we just studied. The amplitude
response is given by
From (21 we see that the amplitude in d B and phase in 2) Adv- — n 20 log 1 + (‹ , e›,)' I i -9bi
decrees w ill be n times the values we just obtained for
plotting Fig. 11-6. Therefore, the corrPction factors in The phase response is given by
t he previous table must also be multiplied by ii.
A Bode plot of straight-line approximations to (l I ‹I› = — n tan ' t 11 -9r )
for different values of n is shown in Fig. 11-7. Since the
amplitude slope for ‹u < ‹ ›, was zero, it will still be zero A Bode plot for n = 1 is shown in Fig. 11-8 and, as you
for any '.al ue of n. Ho w ever, for m > e›, the slope must can see, the only difference in the amplitude response in
be multiplied by the appropriate n. Thus, for ii = 3. this case and that of Fig. I 1-6 is that for in > c›, the gain
the sl ope for m > t›, is 3 t 20 d B.'deck = 60 d B, dec or decreases. The slopes for en > ‹o, in Eq. (11-9s) for any
18 d B ‘oct. The corresponding phase asymptotically v alue of n are simply the negative of the slopes for Eq.
approaches 3 (90 ) — 270' For n = 3 at in = ‹u,. the t l 1-8n). The correction factors are also identical in
phase is n (45 ) = 135' and the amplitude error is ri magnitude except that now they lie below the straight-
(3 d B) — 9 d B. line approximations. The phase characteristic and its
No iv w ork problem PS l 1-1 2. correction factors for any value of n is also identical
Another com monl y encountered form of transfer w ith the single exception that we now have phase lag
function is rather than lead. Th us for in ñ co, and n — 1, the phase
angle approaches — 90°.
lJ Now work through problem PS 1 l- 13.
(11-9ol Transfer functions occasionally contain factors of
the form

dB
Amplitude
60
----Phase ”

55
8 = n tan ’ —
50
A## —— n 20 log J + (cu/ wr)2 —

45

40
(lead)
315 r- 35

270 r- 30 ,
ri = 3
225 25

J80 20

135 15

90 10

45 5

w”
0 0
0.1 05 10

FIGURE 1 1-7

the Bode plot |

 m/m
FIGU RE 11-8

Ct) 209 denominator form of Eq. (2) is more common, we will

A ——j —+ j26 — + 1 discuss it first.
Inspection of (2) indicates that for in 4 en, or (ui/w,)
I, the amplitude of A approaches 0 dB and the phase
= l y26 — (11-10c)
approaches 0°. For u› ñ u›„ the real term bein8 squared
will overwhelm the imaginary term, so that
O f.
1 1
l
2) A = 3) A - 1%8
j —+ j26 — + 1
Thus the phase angle approaches 180° and the amplitude
approaches zero. In general, for in i> m„ the amplitude is
1 ——+ j26 — given by
(11-10b)

These forms are called quadratic factors and they 4) A p -— 20 log log —
generally arise due to interaction between two energy
storage elements in a circuit or system.’ If the circuit *)

contains inductance and capacitance, so that some form If we let a›/u›, = 1 in Eq. (4), we have A$p —— 0, and
of resonance phenomenon occurs, we can expect to for a›/w, = 2 we obtain Apg — 12. Thus the slope in
encounter quadratic factors in their analysis. Since the the region of in a• w, tends to approach - 12 dB/oct,
which is essentially the same I —40 dB/dec. Exactly
The energy storage elements can be two capacitors, two inductors. or the same conclusions can be drawn for (1) except that
an inductor and capacitor. there the amplitude response rises and the phase leads

158 I ac circuit analysis

 +20
' 6 = 0.05
I
6 = 0. 1

-
6=0.2
•10
30 6 = 0.5

60
0

— 10 90
(
p

—20 120 ' '

a

— 2 !
p

1 — —* y26 —

— 30
n

— 40
0.1 10 180
¢ut0.1 ¢u,
|

—
Frequency, rad ians per sec F requency, radians per sec
u

FIGL RE 1 1 -9

as ‹o increases, whereas in (2) the amplitude decreases sketching denominator quadratic factors, we have Fig.
and the phase lags. In both cases the phase l 1-9. Note that as ñ tsometimes called damping factor)
asymptotically approaches 180‘. though from different increases, the peak or resonant point shifts below in,.
directions. The same curves may be used for numerator quad-
Now what happens at in = in, in {2)? Clearly the func- ratic factors if you remember that for u› ñ in, the
tion reduces to amplitude curve slopes upwards at 40 dB/dec; and at
5) .4 = 1 1
— 90 m„ instead o( a resonant rise, we may see a resonant dip
j2ñ 2J as show’n in Fig. 11-10a. The phase response is also
Thus the phase at ea = ‹u, is — 90° in (2) or + 90° in (l). similar except that it approaches 180° of phase lead for
The corresponding amplitude in (2) is in ir u›,. Fig. 11-10b illustrates the phase response in
this case.
6) A, p = — 20 log 26 (11-10r) Now work through problem PS 11-14.
Inspection of(6) indicates that if 6 = ,log 26 = log 1 =
0, so that E d = 0. If 6 < y, say 6 = } for example,
log 2ñ will be the log of a number less than 1 which is a
negative quantity, and in this case
' E ds — 20 log 2(,) = — 20( — 0.3979) z + 7.96 dB
Thus we see that the amplitude response actually rises
above the 0 dB level for 6 < 2 in (2). For 6 > 2, say
6 = 1, the log is positive and hence the number of dB is
now negative. Both the phase and magnitude of (1) and
(2) strongly depend upon the value of ñ in the vicinity of
u› = m„ and it is probably best to plot a few points for (al
an accurate representation. However, as an aid to
the Bode plot |
FIGURE l 1- 10
 11-7 Constructing the Bode plot
In general, a transfer function will contain various factors
in the numerator and denominator which consist of the
basic forms we just learned to plot. We can use this
background to construct transfer function Bode plots of ” 1 + j ’ ’z( t + R z)
practical networks. There are various ways of doing Let
this. all of which are best illustrated by example.
.Assume we wish to develop and Bode plot the voltage
gain transfer function of the so-called lag network of
Fig. 11- l la. First let us develop the transfer function

1 and inspect it for time constants ( RC products or L R

ratios) whose reciprocals denote corner frequencies. If Equation ( 1) may be resolved into two factors
several I ime constants arise in the analysis, we can
correspondingly ex pect to have several corner fre- l
‹!v
2) quencies to consider.
Starting with the v‘oltage divider relationship, we may l + —
derive
The first factor was studied in conjunction with Eq.
( 11-8) and the second with Eq. ( 11-9). A Bode plot of the
amplitude response of the first factor (numerator) is
shown as curve .V in Fig. 11- 11b. The amplitude response

N
0
A=
¢
cat = 10ca,
dB 0
' C2R z
' ' ' ' t , t i

p2
O , ” ,,
E„

90°

45°

0°

—45°

—90°

the Bode plot |

 FIGURE 1 l- 11

160 | ac circuit analysis

 of the second factor is shown as curve D. Since the PROBLEM S WITH SO LUTION S
amplitude scale is logarithmic (dB). multiplication of the
amplitude of these factors is equivalent to addition. PS 11 -1 What is the corner frequency of the low-pass
Hence. lhe composite amplitude response is simply the hl ter in Fig. PS I 1- I '’ What is the 'oltage guin trmeter
point-by-point addition of curves iV and D as shown in function at the corner frequency and at 1 10. 1 2. tw ice.
curx’e .4d B = ,V -r D. A pair of dividers is most useful in and ten times the corner frequency?
this regard. Correction factors can be inserted so that a
smooth curx’e can be drawn, but generally the st raight-
line approximations are adequate.
The phase of the tirst and second factors has’ also been
approximated w’ith straight-line segments as d, and d›p
respectively. The phase angles may also be algebraicall y
..dded with a pair of dividers or by counting squares in a
manner similar to the amplitude response.
With a lift le practice you will in most cases be able to N IGL RE PS 1 1 - 1
construct the amplitude response of .X without having
to construct the individual factors. For example, with SOL L“ TIO!¥ The time constant is
‹.'›› > ‹ a, in (1 ). we know the first break (corner) occurs at
‹ › as «’e go from the low to the high-frequency range. I) r, = RC —— l0 kf2(0.01 QF) = 1 x 10 “ sec
Since ‹'a, is in the denominator. .4dB w’ i ll break downward The corresponding radia n corner frequency is
at ru , w ith a slope of 20 d B/dec and continue doing
so
until ru› . w’hereupon the numerator starts climbing at 2) ‹›, = ' = 10“ rad/sec
20 d B, dec. Thus the positive numerator slope cancels the
negative denoniinator slope so t hat the slope of *d The corresponding corner frequency is
levels off at 0 dB;’dec.
Although it is somewhat trickier. the phase response 3) = 1592 Hz
may be similarly reasoned. To a first approximation the
denominator phas’e breaks at 0.loa, and decreases at The general expression for the transfer function by
it bout 45' dec. Thus at ‹'a, the denominator angle is Eq. (11- 1) is
45 . but note the numer‹itor phase is comine into play.
since m , = 0. 1 ‹o2 aFld the numerator phase is increasing A ——
at 45' dec. Thus at cr, the numerator and denominator
phase Slopes cancel, sO that the phase remains constant
at the — 45‘ level it acquired at a›, . This holds true until Plugging in the given values of f/f —— 0.1, 0.5, 1, 2, and
t which point the denominator phase has gone 10 into (4), we have
••
•
through its maximum possible shift of — 90°, so that the
slope of the denominator phase angle is zero. The numer-
10
ator phase slope. however, continues to climb at 45‘,/dec 0.0995
for a decade abo ve 2 This pulls the net phase slope 995 0.894 0.707 0.447
— 84
(Q) up at 45‘ dec above ro, until 10 2 at which point
the n umerator has gone through its maximum possible
phase shift of + 90‘. Thus for in > both numerator PS 11 -2 Prove that the corner frequency of a low-pass
and denominator phase slopes are zero and the net filter such as shown in Fig. l 1-1 is that frequency at
phase s’lope levels off at 0° due to the — 90’ in the which .YC —— R.
denominator canceling the + 90‘ due to the numerator. SOLU TIO N Applying the voltage divider relationship
From a physical viewpoint we can reason in Fig. to Fig. 11- 1, we have
1 1- 1 la that at low frequencies C 2 looks like an open
circuit and hence Eg = E¡ or A g - 0. As in increases, 1) A,. =
C 2 comes into play so that it starts to load A , , which
decreases E,. This accounts for the break at u›, . As in Dividing numerator and denominator by —jX C find
increases further, <2 eventually looks like a short and rearranging terms, we obtain
E, = E;, A 2/(ft , + A 2). This accounts for m„ where C 2
fades out of the picture. 1
Now work through problems PS 11-15, PS 11- 16, PS 11- 17.
l—
the Bode plot |
 R
1+ j

162 | ac circuit
the Bode plot |
 c

162 | ac circuit
If we let R — • c in Eq. (2). f

) A. = 4)
g' 2 = 0.707/ — 45'

which is precisely what we showed in the previous

problem w hen we set J = J§. Setting / JJ equal to 0.1. 0.5, 1, 2. and 10 yields the
PS 11 -3 What would be the corner frequency of the following values
circuit in Fig. PS 1 1-3?
/ f, O. 1 0.5 1 2 I t)
1.4 , | 0.9950 0.447 0.707 0.594 0.995
/0 b4.3' 63.4' 45 26.6' 5. 7 I '

PS lJ -5 Determine the voltage transfer function of

the coupling network shown in Fig. PS 11-5. Resistor
A t is analogous to the output resistance of a voltage
source driving a high-pass filter.
FIGURE PS I I -3

SOLU TIO!V By inspection we see that ft, and fl_ may

be thevenized with the voltage source to simplify the
equivalent circuit that the capacitor looks back into.
Th us, by analogy with Fig. 11- 1, intuition. or both.
‹› - - ‹ )C. The radian corner frequency is
then in, = 1/r .
PS 4 4 -4 What is the corner frequency of the high-pass
filter in Fig. PS l l -4? What is the voltage gain transfer
function at the corner frequency and at 1/ 10, 1/2. twice, FiGURE PS 1 1 -5
and ten times the corner frequency’?

SOLU TION Applying the voltage divider relationship

25pF
we have

l)

Dividing numerator and denominator by ft , + R1 yields

FIGURE PS 11-4 e, + e,” — j

SOLU TION The time constant from Eq. (11-2) is 1
1) r, = fiC = 1 kf2(25 pF) = 25 x 10* 3 sec
The corresponding rodian corner frequency is
If we let the corner frequency be

the Bode plot |

M ultiplying n umerator and denominator of (4) by yes ‹o, Therefore
and introducing / yields
7J A, —— 20 log d B = 20 log 115.6 d B

5) ¿ _ — 20t 2.0631 d B = 41.26 d B

PS 11-8 E xpress the v olt‹iac gain of the netv•’ork in
Fiy. PS i 1-8 in dB.
Note that for / / e just sce the fi t, R volt:ige 18 k
divider. since .Yt 0.
PS 11 -6 Prove that the corner frequency is the
frequency at w hich .¥, = R It ; in Fig. PS 11-5.
SOL I’ T 10 ¥ If we divide numerator and denominator
of Eq. ( 1) by R , + fi , in the solution of the pres ious
problem. we obtain
FIGU RE PS 1 1 -
b

SOL TIO N
12 kCl
When .¥p = R , + R the abox’e equation reduces to .4, 20 lO d B = 20 lo dB
12 kC1 -r 18 kf1
= 0.707 R, = 20 log (0.4J d B = 20( — 0.3979) d B
R , + R _›
7‘9588 dB
This is exactly the same result that would be obtained by
setting / = J} in t5) of the previous problem. Thus when PS 4 4 -9 The voltage gain in d B of the four-terminal
/ - J, Jp = fi , -t- fi 2 . network in Fig. PS 11-9 is A, = — 5.67 d B. What is the
VO ltage gain ratio £,¿ E t?
PS 1 4 -7 Express the voltage gain of the circuit in
Fig. PS 11-7 in d B at / = /„ the resonant frequency.

270 It 39 m H
I’
E,
dBa ——
— 5.67 E 2

4O pfi
dB 4, = —5.67
FIGURE PS 11-9
SOLU TION
FIGURE PS 11-7
11 A —— 20 logp 2
dB

SO LU TION E2
f 2) — 5.67 dB = 20 1 dB
l) = 127.4 kHz
_ , — 5.67
2 #z
3) =1 ) og (—02835)
2) en, = 800.6 x 10° rad/sec 2
Therefore we seek the number which represents the ratio
31 N, = in, L = 31.22 kf2 EI and whose logarithm is —0.2835 (the notation
Xz 31.22 kf2 log ' implies the antilog). You will recall that in log
4) fi 0.27 kfl = 115.6 tables positive mantissa values only are usually tabu-
5) E, = Q E; . lated, and hence to obtain a positive mantissa we add
and subtract 1, so that
6) 4) = log ' (1 — 0.2835 — 1) = log ' (0.7165 — 1)

the Bode plot |

Looking in a table of logarithms (or on a slide rule) for The phase angle is always zero, s’ince A equals a constant.
the number corresponding to 0.7165, we have 5.206 and Remember that a constant amplitude response implies
with a characteristic of — 1.‘ zero phase shift.
f,
51 = 0.5206 2-) d = 20 log_ 0.004 — 48 d B
E, The phase angle is again zero at all ‹u.
PS 11 -10 Construct a Bode plot for the amplitude 3) The corner frequency is in, = 50 rad/sec. Since n = 1,
response for each of the following transfer functions: this Bode plot will have a positive slope of 20 d B/dec
(or 6 dB oct) and pass through 0 dB at in — m,. The
phase shift is a constant 90‘.
1) A = 3200 4) The corner frequency is in, = 300 rad/sec. Hence the
Bode plot will pass through 0 dB at in, = 300. Since the
2) A = 0.004 in term is in the denominator and ii — 2, the slope will be
negative and equal to — J0 dB/dec (or — 12 dB;'oct). For
n = 2, Q = — n 90° = — 2(90°) = — 180°.
Determine the phase response in each case by inspection The resultant curves are plotted in Fig. PS 1 l-10.
of the above transfer functions.
SOL C’TION PS 11 -11 What is the transfer function corresponding
to the amplitude response curves labeled o, b, c. ti in
1) J = 20 log 3200 - 70.1 dB Fig. PS 11- 11?
• Negative mant issas for negative loearit h ms may be con venient I y SOLU TION Curve a crosses 0 dB at a corner fre-
ev a1 uated on a sl ide r ule w ith the L and CI scales. quency f, —— 40 Hz and a positive slope 6 dB/oct.

FIGURE PS 1 l- 10

164 I ac circuit
 60

40

20

dB 0 ”

20
1 I I

—4

I
i
[
—60 IIIII I I I I IIIIII l. I. I I
0.01
01 1.0 10
7(k Hz)

FIGURE PS l l - 1 1

Since the 6 d B oct slope implies 90‘ of phase shift, PS 11 -12 Construct a Bode plot for the function
A.
f
40 500 Hz
First show the straight-line approximations for the
Curve b is constant at — 43 dB. Therefore
amplitude and phase response and then insert the correc-
— 43 d B = 20 log A tion factors so that a smooth curve may be constructed.
43 d B SOLUTION The amplitude response will be flat from
A, = log ' — = 0.00708 f —— 0 Hz until/ = f, = 500 Hz. Forf > /„ the slope rises
20 at 40 dB/dec or 12 d B/oct. Since n = 2, the amplitude
Curve c has f 200 Hz and a positive slope of 18 correction factors are 2, 6, and 2 dB at 0.5/„/„ and 2/,
d B 'oct. Hence the phase shift is 3(90°) as n = 3. Therefore respectively. The phase corrections are approximately
+ 12° and — 12° at 0.1/, and 10/, respectively and — 12°
and + 12° at 0.5 f, and 2f, respectively. The phase curve
will go through 90° at /, and approach 0° and 180°
asymptotically for f ñ f and f a• f, respectively. The
Curve d has f, —— 3000 Hz and a negative slope of 24 resultant plot is shown in Fig. PS I l-12.
dB oct. Therefore Ps ) 1 -I 3 Construct a Bode plot of the transfer
i 1
f function A = f 3

3000 5tXi Hz

the Bode plot |

 2 345 6 7 8 91 2 345 6 7 8 91 2 345 6 7 8 91
60

iÎ 1
40
F*
“ 'A
“soo HT
‘ d8 Î1 .1_
pe =' 40d tB/dec

20

6 dB
- 2 dB—
0

0ᵉ
18

135ᵉ 0º 5º
0ᵉ

‘ ’’°

0. J I 0. f 0.0 f, 2f, 10 f

FIGURE PS 11- 1
2

SOLE’TION The approximated amplitude response

1
will be flat for / > f, until/,. At /, the slope changes from
zero to — 60 dB/dec. Correction factors for sketching the jwCE + j2(w'LC -
actual response are 3, 9, and 3 dB at 0.5 f„ f„ and 2 /, 1)
respectively and + 18° and — 18° at 0.5f, and 2f, respec- Now the trick is to algebraically wrestle the final form of
tively. The phase curve will go through — 135° at /, ( 1) into the quadratic form of Eq. (11-10b), repeated
here as (2)..
and approach 0° and — 270° asymptotically for f ñ 2) 1
A =
and f ñ /, respectively. The resultant plot is shown in
i— — ¢u
2
+ j26 —
Fig. PS 11- 13.

PS 11 -I 4 Determine the voltage gain transfer func- If we let

tion and construct a Bode plot for the circuit of Fig.
2
PS 1 l- 14c. 3) 83‘333 10’ (rad/sec)2 or m = 9129
xc
SOLU TION First we must develop the transfer func- rad/sec and therefore obtain
tion. Thus p 1
4) C = 2
l)
and then substitute (3) and (4) into (1), wc have

1
166 | ac circuit
 3 4 5 6 7 8 9 1
0 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 10
-3dB’
-9dB-*- -

3dB

—40 _ Slope = 60 dB/dec w.

/l , ’

— 60

—0 5 / 10 f,
0’

— 90°

I
i7|b 1 1 l ‹

— 180°
iI

'I II
I

—270°’!

FIGURE PS 11- 1 3

1 for in in, a straight line along 0 dB until m„ at which

z+ point it breaks downward with a slope of —40 d B/dec
— 1 — for in ñ in,. The response in the vicinity of in, is deter-
mined by the damping factor 6. Turning to Fig. 11-9 we
Since
see that for 6 = 0.043 we are off the plot. Hence we turn
6) Q = p —— 11.65 to Eq. (11- l0c).
10) dB peak = — 20 log 2d = 21.3 d B
we may substitute (6) into (5) to obtain
which yields the height of the resonant peak at in,.
1 For a more accurate plot, we might as well evaluate the
7) A,=
transfer function. Thus we have
l —
1
Equation (7) is in the same form as (2), with 11) A =

1 9129
8) 26 = 1 j2(0.0429)
+ 9129
Of 1
2
1 — - ’(9.399 x 10 ‘)m
g ' 0’0429 > 0‘043 83.33 x 10‘ -
9) 8'
2
and plugging in various values of in yields the following
In constructing the Bode plot, we know we will have table of data, which is plotted in Fig. PS 11-14b:
 the Bode plot 1g7

166 | ac circuit
 4izi, |0‹'„

mrad,sec 912.9 4564 7303 9129 18258 36516 91290

Q —0.497' —3.27’ —10.8 —90‘ —174‘ —177 —178.7' —179.5‘
.4dg 0.087 2.48 8.72 21.3 —1.98 —9.56 —23.5 —39.9

R
470 n 0.6 H

0.02«F

+ 20 g —— + 21.3dB

1
+10 A=
Amplitude + /2t0.0431 g)
" 9129 “
IB 0

—10

—20

-30

—40

0°

—45°

—90°
Phas
— 135°

— 180°

O T 05 08 10 4

FIGURE PS 11- 14

the Bode plot |

PS 4) -I 5 Develop the transfer function and construct
a Bode plot of the amplitude response of the network
shown in Fig. PS I I - I 3a. Assume ft t - 2 2)

2. Analyze the transfer function to ascertain what con-

tributions the various factors will make in the Bode plot.
Equation t2), the transfer function, is of the form
E, b
E„ —— a
W here a —— R ›, ( R + fi 2 }, b —— I + ym e› i , ri nd c — 1 +
y‹o, m2. Therefore,
dB = 20 (log n + log b — log c)
The log a term will contribute a component which is
independent of frequency, while the b and c terms are
frequency-dependent. We have shown in previous ex-
amples that the b term will break at in, and slope up-
F IG L’RE PS 11 - 1 5 ward at 20 dB per decade, while the c term breaks at m2
and slopes downward at 20 dB per decade. Inspection
of the circuit indicates that at low frequencies, where
SOL D’TIO.N c, is very large compared with R „ the out put will equal
1. First develop lhe transfer function. R; R + 21! E„. At frequencies where +c, is small
compared with 2 the output will approach the input.
’2 3. Construct the Bode plot. For the purpose of sketch-
ing, let us assume that - Then /( ‹ + 21'
}, or — 6 d B. This fixes the ratio of r, •2

2
" R , + /t 2

The Bode plot is sketched in Fig. PS 11-

15b. PS 41 -1 6 Bode plot the transfer
function
’2 10 1 +20

SOLU TION In the very low frequency spectrum or for

Dividing numerator and denominator by fl , + u› approaching zero, we see that A = 10, or 20 dB. This
means we have a netwoJ capable of providing an
2 yields amplification of ten times down to dc or zero frequency,
as shown in Fig. PS l l-16. Electronic amplifiers can do
this. As a› increases, the first break occurs at ci = 5
rad/sec, and since this is in the denominator, the gain
where (amplitude) rolls off at — 20 dB/dec from the 20-dB
level. The next break at w = 20 rad/sec is in the numera-
tor. which tries to pull A up at + 20 dB/dec. However,
’eq
A, +R, this is canceled by the previous — 20 dB/dec slope, so
that the net slope is now zero and it stays that way until
168 I ac circuit
 FtGURE PS 1 l -16

in = 100, at which it breaks down again at — 20 dB/dec SOL U TION The above transfer function may be re-
and remains so. If correction factors are inserted or written as
Aaa is actually computed, the smoothed amplitude
response curve results. j
The phase response is approximated by lines sloping 2) A = '
plus or minus 45‘/dec, depending upon whether or not 1+ y 1+ j
the term is in the numerator or denominator. The phase
approximations start at 0° at 0.1 ui„ go through 45° Inspection of (2) indicates that the first factor is that of a
at m„ and terminate at + 90° at 10 ni,. The resultant phase high-pass ftC type filter and the second factor is that of a
approximation is illustrated by the ste d phase curve. low-pass filter with identical corner frequencies. There-
For more accuracy, the various correCtlOR factors can be fore, in the frequency range ‹u < ni t , we have the low-
inserted or the actual phase angle of A may be computed pass response climbing at 20 dB/dec while the high-pass
to yield the smoothed phase curve of Fig. PS 11-16. response is flat at 0 dB. At m = in „ the high-pass re-
PS 11 -I 7 Bode plot the transfer function sponse breaks to zero slope while the low-pass response
breaks downward at — 20 dB/dec. This yields the ap-
j proximated amplitude response of Fig. PS 11-17.
A = The approximate phase response may be sketched by
1)
CU
2 first drawing the high-pass filter phase response as a line
starting at w/a› t = 0.1, sloping down 45°/dec from + 90°

the Bode plot |

 to 0° at elm, = 10, and then sketching the low-pass high- and low-pass phase curves. Should greater accuracy
filter phase with the same slope and in values but starting be required. correction factors may be inserted or actual
at 0' and terminating at — 90°. The approximate phase values computed in (2). The actual phase and magnitude
may then be sketched by algebraically adding both the are also shown in Fig. PS 11- 17.

' 250

20
A= 200

150

10
100
$ High pass
50
Approximated a
Actoal phas
0 0°
e Low pass
—50
Actual A
—10
— ioo

dB low pass
-150
dB h igh pass

— 20
—200

—250

— 30 -3X

0.1 10

FlGUnE PS 1 1- 17

170 | ac circuit
PROBLEMS WITH ANSWERS
PA 4 4 -I Develop the transfer function and Bode plot PA 11 -4 Develop the transfer function and Bode plot
t he amplitude response of the circuit in Fig. PA I 1- Ia. the amplitude response of the circuit in Fig. PA I 1-4o.

—dB -2OOB
dec

Ib I

See Fig. PA 11- lb. See Fig. PA 11-4ò .

PA 11 -2 Develop the transfer function and Bode plot PA 1 -5 Develop the transfer function and Bode plot
the amplitude response of the circuit in Fig. PA 11-2s. the amplitude response of the circuit in Fig. PA 11-5a.

60 k0

0 50
30 kZ
—954 — 20D8
dec
— dB

—9.54———
-4008
dec

— dB

ANS W ER AN S W ER
See Fig. PA 11-2b. See Fig. PA I I -5b.

PA 41 -3 Develop the transfer function and Bode plot PA 11 -6 Develop the transfer function and Bode plot
the amplitude response of the circuit in Fig. PA l 1-3a. the amplitude response of the circuit in Fig. PA 11-6O.

“’ “ RC
— g.s4 —— -
20dB

— 6.02 -dB
— dB

the Bode plot |

ANS WER ANS WER
See Fig. PA 11-3b. See Fig. PA 11-db.

172 | ac circuit
 PA 11-7 Evaluate A at cv = cJ, if J = , in the follow-
ing transfer function :
I + ‘ j oa
—100
2
Iy —1 — —OU + y2ñ
‹- , «
43 d B / — JS

PA 11-8 Bodc plot .4 = t 1

. Show
. ,..«. —-›
both phase and amplit ude.

*60

0 Iag
3

1* y 0.5 X 10—
+40 0

45

90

135

180

225

—20

315

— 40
360

dB
—60

01 10 100

ANS W ER See Fig. PA 11-8.

the Bode plot |

 PA 4 ) -9 Estimate the transfer function applicable to
Fig. PA 11-9.

+20

dB
—20

—40

60 0.01 0. 1.0 0 100

ANS W ER

1+
A = 10 0.01 100
1+ l + j
0.01 ' + ' 10

174 I ac”circuit
PA 11-10 Bode plot

1 1 +y
+0.1

i
- de c

Y
6

—90
0 dB/

dB
— 180

40 dB/d

—t70
20 c B/dec

-360

—60 dB/dec***"^

01 01 1 10 100 1000

AN S W E R See Fig. PA l 1-10.

th• Bode plot 1 175

PROBLEMS WITHOUT ANSWERS

P ) -I Bode plot the transfer function A = 28.000. P 11 -8 Develop the transfer function and construct
the Bode plot for the circuit of Fig. P 11-8.

93 en

7 kf2

P )1 -2 What is the corner frequency of a transfer

function expressed as 1 + ju› T where T is a time con-
stant?
P 1 -9 Construct a Bode plot for the circuit of Fig.
P 11-9.
I kf2

P 1 -3 Bode plot the transfer function A = y —. 1 qF

P 1 ) -10 Construct a Bode plot for the circuit of Fig.

P 11- 10.

P 11 -4 Bode plot the transfer function A = j

P 4 1 -5 Bode plot the transfer function A =

1 P 14 -11 Develop the transfer function and construct a
J ‘ CO 2 Bode plot for the circuit of Fig. P 11-11.
10 mH

1 kf2

P J1 -6 Bode plot the transfer function

A = l + j
f
5 kHz P 11 -12 Develop the transfer function and construct a
Bode plot for the circuit of Fig. P l 1-12.

P 11-7 Bode plot the transfer function

1

- 80 kilz

176 | ac ”circuit
P ) -13 Sketch the Bode plot for the circuit of Fig. P 11 -4 5 Estimate the transfer function of the Bode
P 11-13. plot in Fig. P 11- 15.

too kn

P 11 -14 With regard to the transfer function

_ 100(l + yw 1) P 14 -I 6 Estimate the transfer function of the Bode
( I + yea5)' tju›1) plot in Fig. P 1 l-16.
lill in the blank spaces.
(a) The factor 100 in the numerator represents
dB of gain. I
I
(b j The l + yen 1 factor will break at en = and
slope dB/dec or d B/oct. It will 40 d'
_ dB
contribute degrees of phase at 40— -
dB \ d
the corner frequency. + 20 dec i
(c) The ( 1 + yo›5)2 factor breaks at en = and
slopes d B/dec. lt contributes 1000
degrees of phase at the corner frequency and
the corresponding amplitude response is
(‹?) The jm 1 factor contributes degrees of
phase and slopes

the Bode plot | 177