CS Soluiton Manual

Solution Manual
Control Systems
Principles and Design
4th Edition
M Gopal
Professor
Department of Electrical Engineering
IIT Delhi, New Delhi
Tata McGraw Hill Education Private Limited

NEW DELHI
McGraw-Hill Ofces
New Delhi New York St Louis San Francisco Auckland Bogot Caracas
Kuala Lumpur Lisbon London Madrid Mexico City Milan Montreal
San Juan Santiago Singapore Sydney Tokyo Toronto
Preface
This Solution Manual, offered as a supplement to the textbook: Gopal,

M., Control Systems: Principles and Design, 4th edition, Tata McGraw-Hill,
New Delhi, 2012, is an attempt of the author to share his experience and
thoughts with the instructors. In each chapter of the book (from Chapter 2
onwards), there are three types of problems:
Review Questions
This section is designed to help the students review control theory
introduced in each chapter by solving straightforward theoretical
questions whose answers directly follow from the chapter. For
instructors, this section may be helpful in setting theoretical
questions for various examinations.
Self-Appraisal Quiz
Multiple-choice questions in this section are designed as a self-
test for the students. Master key is provided. The questions are
directly related to concepts and methods introduced in the chapter.
A number of simple numerical questions are also included in the
Quiz for each chapter. For instructors, it is a Question Bank; they
may use these questions in modified form for Quiz tests. Also the
multiple-choice numerical questions can easily be converted to
simple numerical problems for examinations.
Problems
Each chapter ends with a variety of problems that vary in degree
of difficulty and complexity, and many chapters include several
practical problems to foster students motivation. This solution
manual is intended to assist instructors in this part of problem-
solving exercise. Each problem, particularly the design problems,
could be solved in many different ways. The authors presentation
in this manual is just one of the possible approaches. The instructors
may use the manual for guidance only, and would be required
to train the students with the possible alternatives as well. The
iv Preface
final answers given in the book may force the students to think
as per the authors solution. They have to be guarded against this
perception.
Throughout this manual, the numbers associated with referenced
Equations, Figures, Tables, Examples, Review Examples, and Sections are
given as pointers to the material in the textbook.
M GOPAL
mgopal@ee.iitd.ac.in
2
System Models and Response:
Transfer Function, Block Diagram,
Signal Flow Graph
Y ( s) s+3 s+3
2.1 = G(s) = 2 =
R( s) s + 7 s + 10 ( s + 2) ( s + 5)
1
(a) R(s) =
s+1
1 1 1
y(t) = e - t - e -2 t - e -5t m (t )
2 3 6
(b) (s2 + 7s + 10) Y(s) sy(0) y(0) 7y(0)

= (s + 3) R(s) sr(0)
Initial conditions before application of the input are
1
y(0) = 1, y(0) = , r (0) = 0
2
s + 15/2 s+3
Y(s) = +
( s + 2) ( s + 5) ( s + 2) ( s + 5) ( s + 1)
1 3
y(t) = e - t + e -2 t - e -5t
2 2
1000
2.2 Y(s) =
s( s + 34.5s + 1000)
2
w n2
= ; z = 0.5455, w n = 31.62
s( s 2 + 2zw n s + w n2 )
e -zw nt 1-z2
y(t) = 1 - sin w n 1 - z 2 t + tan -1 ;t 0
1-z2 z
= [1 1.193e17.25t sin (26.5t + 56.94)] m(t)

2 Control Systems: Principles and Design
Y ( s) 2s + 1 2s + 1
2.3 G(s) = = 2 =
R( s) s + 2 s + 5 ( s + 1) 2 + (2) 2
s+1 1 2
= 2 -
( s + 1) 2 + 2 2 2 ( s + 1) 2 + 2 2
R(s) = 1
Therefore
1
y(t) = 2e- t cos 2t - e- t sin 2t m (t )
2
1
2.4 (a) (i) R(s) = 1; Y(s) =
ts +1
1 -t /t
y(t) = e ; yss = 0
t
1 1
(ii) R(s) = ; Y(s) =
s s (t s + 1)
y(t) = 1 et/t yss = 1
1 1
(iii) R(s) = ; Y(s) =
s2 2
s (t s + 1)
y(t) = t t + tet/t; yss = t t
Aw KAw
(iv) R(s) = ; Y(s) =
s2 + w 2 (t s + 1) ( s 2 + w 2 )
KAwt - t /t KA
y(t) = 2 2
e + sin (w t + q )
t w +1 t w2 +1
2
q = tan1 (wt)
KA
y (t ) t = sin (w t + q )
t w2 +1
2
e-zw nt 1-z2
(b) (i) y(t) = 1 sin w d t + tan -1
1-z2 z
wd = w n 1 - z 2
System Models and Response: Transfer Function, Block Diagram, Signal Flow Graph 3
The final value theorem is applicable: the function sY(s) does

not have poles on the jw-axis and right half of s-plane.
yss = lim sY ( s ) = 1
s0
(ii) The following result follows from Review Examples 2.1.

2z e-zw nt 1-z2
y(t) = t + sin w d t + 2 tan -1
wn wd z
2z
yss = t
wn
The final value theorem is not applicable; the function sY(s)
has a pole on the jw-axis.
2.6 (a) g(t) = et et sin t

1 1 s2 + s + 1
G(s) = - = ;
s + 1 ( s + 1) 2 + 1 ( s + 1) ( s 2 + 2 s + 2)
1
DC gain = lim G ( s ) =
s0 2
s2 + s + 1
Y(s) = ; yss = lim sY ( s ) = 0.5
s ( s + 1) ( s 2 + 2s + 2) s0
s +1 1/2 1/4 1/4

(b) Y(s) = G(s)R(s) = = + -
s 2 ( s + 2) s 2 s s+2
1 1 1
y(t) = t + - e-2t
2 4 4
1 1
yss = t+
2 4
2.7 Revisiting Review Example 2.1 will be helpful.

wn = 100; z = 3; 2z/wn = 6/100 sec = 60 msec
2z
Steady-state error = 25t 25 t - 25
w n
= 1.5
2.8 M y (t ) + By (t ) + Ky (t ) = F (t ) = 1000 m (t )
1
Y(s) =
s ( s 2 + 10s + 100)
wn = 10, z = 0.5, wd = 5 3
2 -5t
y(t) = 0.01 1 - e sin (5 3 t + tan -1 3
3
2.9 (b) yss(t) = 2 |G(j2)| sin (2t + q)
2 1
|G(j2)| = = ; q = G ( j 2) = - 135
j 2 ( j 2 + 2) 2 2
yss(t) = 0.707 sin (2t 135)
E0 ( s) 1 1
2.10 = = ; t = 1.6 10 3 0.2 10 -6 = 0.8 10 -3
Ei ( s) RCs + 1 t s + 1
E0 ( jw ) 1
=
Ei ( jw ) jwt + 1
1 1
= ; Bandwidth w b = 1/t = 3125 rad/sec.
w b2t 2 +1 2
3125
fb = = 497 Hz
2p
2.11 The transfer function is
E0 ( s) R (1/sC )
G(s) = =
Ei ( s) sL + R (1/sC )
R / (1 + sRC ) R
= = 2
sL + R / (1 + sRC ) s RLC + sL + R
E0 ( jw ) R
= 2
= G( jw )
Ei ( jw ) R - w RLC + jw L
G(0) = 1, G() = 0; the circuit is a second-order lowpass filter.
R 1
G( jw ) w = w = =
b
(R - w RLC ) + w L
2 2 2 2
w = wb
2
This gives
2
R2 1 w L
2 2 2 2
= ; (1 - w b2 LC )2 + b = 2
(R - w b RLC ) + w b L 2 R
Substituting the values of R, L, and C, we obtain
16w4b 7w2b 1 = 0; wb = 742 rad/sec
2.13 Mx + Bx + Kx = K1 ( y - x)
Gravitational effect has been eliminated by appropriate choice of

the zero position.
X ( s) 0.1667
= G(s) =
Y ( s) (0.0033s + 1) (0.0217 s + 1)
w = 2p v/l = 1163 rad/sec
|G(jw)|w = 11.63 = 0.1615
x(peak) = 7.5 0.1615 = 1.2113 cm
2.14 (a) M x + Bx + Kx = F (t )
Gravitational effect has been eliminated by appropriate choice
of the zero position.
X (s) 1
= 2
F (s) Ms + Bs + K
Force transmitted to the ground
= K X(s) + Bs X(s)
B
K s + 1 F ( s )
=
M 2 B
s + s +1
K K
F(t) = A sin wt
Peak amplitude of the force transmitted to the ground at
steady-state
2
B
A 1 + w
K
=
2 2
Mw2 Bw
1 - + K
K
(b) Mx + B( x - y ) + K ( x - y ) = 0
X (s) Bs + K
= 2
Y (s) Ms + Bs + K
Peak amplitude of machine vibration
2
B
A 1+ g
K
=
2 2
Mg 2 Bg
1 - + K
K
2.15 M1 y1 + K1 ( y1 - y0 ) + B1 ( y1 - y0 ) + K 2 ( y1 - y2 ) + B2 ( y1 - y2 ) = 0
M2 y2 + K 2 ( y2 - y1 ) + B2 ( y2 - y1 ) = 0
Gravitational effect has been eliminated by appropriate choice of
the zero position.
2.16 (a)
+ + +
1 2 3+ 4

1 2
2 3+ 4
Y ( s) G1 (G2G3 + G4 )
=
R( s) 1 + (G2G3 + G4 ) (G1 + H 2 ) + G1 H1G2
(b)
4
Y ( s) G G G + G1G4
= 1 2 3
R( s) D
D = 1 + G1G2H1 + G2G3H2 + G1G2G3 + G4H2 + G1G4
2.17
+ + +
1 2 3
+
(1 1) 1
Y ( s) G1G2G3
= G4 +
R( s) 1 + G2G3 H 2 + G2 H1 (1 - G1 )
1 1 1 2 3 1 1
Y ( s) G G G D + G4 D 2
= 1 2 3 1
R( s) D
D = 1 + G2H1 G1G2H1 + G2G3H2
Removing loops that touch forward path (G1G2G3) yields D1 = 1.

But no loops touch forward path (G4); thus D2 = D.
Y ( s)
2.18 = (P1D1 + P2D2 + P3D3 + P4D4)/D
R( s)
P1 = G1G2; P2 = G1G3; P3 = G4H2G1G2;
P4 = G4H2G1G3
D = 1 (G1G3H1H2 G1G2H1H2)
D1 = D2 = D3 = D4 = 1
Y ( s) (1 + G4 H 2 ) (G2 + G3 )G1
=
R( s) 1 + G1 H1 H 2 (G2 + G3 )
1 1 3 1
2.19
3/ 3

+ + 3
1 2
1+ 3 2
Y ( s) G1 G2 G3
= M(s) =
R( s) W =0 1 + G2 H 3 + G3 H 2 + G1G2G3 H1
2 3
+
( + +
1 1 2+ 2) 3
Y ( s) G3 (1 + H 3 G2 )
= Mw ( s ) =
W ( s) R=0
1 + H 3G2 + G3 (G1 H1G2 + H 2 )
2.20 There are three forward paths:

P1 = G1G2G3G4G5; P2 = G1G6G4G5; P3 = G1G2G7
There are four individual loops:
P11 = G4H1; P21 = G2G7H2; P31 = G6G4G5H2;
P41 = G2G3G4G5H2
Loop P11 does not touch loop P21.
D = 1 (P11 + P21 + P31 + P41) + P11P21
By removing the loops that touch P1, we get D1 = 1. Similarly D2 = 1.
But loop P11 does not touch P3, so that D3 = 1 P11.
The closed-loop transfer function is
Y ( s) P D + P2 D 2 + P3 D 3
= 1 1
R( s) D
G1G2G3G4G5 + G1G6G4G5 + G1G2G7 (1 + G4 H1 )

=
1 + G4 H1 + G2G7 H 2 + G6G4G5 H 2 + G2G3G4G5 H 2 + G4 H1G2G7 H 2
t 2 s2 t 3 s3 t 4 s4 t 5 s5
2.21 e -t Ds = 1 tDs + D - D + D - D +
2! 3! 4! 5!
It is easy to calculate with long division that
1 - t D s /2 t 2 s2 t 3 s3
= 1 - t Ds + D - D +
1 + t D s /2 2 4
2 2 2 2
1 - t D s /2 + t D s /12 tD s t 3 s3 t 4 s4 t 5 s5
2 2
= 1 - t Ds + - D + D - D -
1 + t D s /2 + tD s /12 2 6 24 144
3
Development of Models for Industrial
Control Devices and Systems
3.2
er + if + 1 ia 1 w
KA Kg KT
Ra Js
et
Kb
Kt
w ( s) 50
=
Er ( s) s + 10.375
For Er(s) = 100/s,

100 50
w(s) =
s( s + 10.375)
w(t) = 481.8 (1 e10.375t)
3.3
qR + 1 1 qM 1 qL
KP KA KT
sLf + Rf Js2 + Bs 50
q L ( s) 1
=
q R ( s) s(0.1s + 1) (0.2 s + 1) + 1
Development of Models for Industrial Control Devices and Systems 11
3.4 TM = K1ec + K 2qM = JeqqM + Beq qM
K2 = slope of the characteristic lines shown in Fig. P3.4b

3-0
= = - 0.01
0 - 300
DTM 3-2
K1 = = = 0.1
Dec qM = 0
30 - 20
2
N1 N 3
N N J L = 0.0032
Jeq = JM +
2 4
2
N N
Beq = 1 3 BL = 0.00001
N2 N4
q M ( s) 10 q ( s) 1
= ; L =
Ec ( s) s(0.32 s + 1) Ec s s(0.32 s + 1)
q
2
3.5 (i) Jeq = JM + JL L = 0.6

qM
q
2
Beq = BM + BL L = 0.015
qM
K T = Kb
q M ( s) KT
=
Ea ( s) s[( Jeq s + Beq ) ( sLa + Ra ) + KT Kb ]
16.7
=
s( s + 100 s + 19.2)
2
(ii) A sketch of multi-loop configuration may be made on the lines

of Example 1.3.
(iii) It will behave as a speed control system.
(iv) Relative stability and speed of response may become
unsatisfactory.
3.6 J = JM + n2JL = 1.5 105; n = 1
B = BM + n2BL = 1 105
q M ( s) 13.33
= 2
q R ( s) s + 3.55s + 13.33
qR + + ec + TM 1 qM 1 qM
Ks KA K1 s
Js + B
K2
Kt
w ( s) K
3.7 (a) = ; K = 17.2, z = 1.066, wn = 61.2
Er ( s) 1 2 2z
s + s + 1
w n2 wn
(b) Modification in the drive system may be made on the lines of

Fig. 3.54
er + + 1 KT w
KA Kr
sLa + Ra Js + B
Kb
Kt
3.8
Tw
N(s)
qR + TM + qL
KP KA KT G(s)
TM = J M qM + B(qM - qL ) + K (q M - q L )
B (qM - qL ) + K (q M - q L ) = J LqL + Tw

(JMs2 + Bs + K) qM(s) (Bs + K) qL(s) = TM(s)
(Bs + K) qM(s) (JLs2 + Bs + K) qL(s) = Tw(s)

qL(s) = G(s) TM(s) N(s) Tw(s)
q L ( s) K P K A KT G( s)
=
q R ( s) 1 + K P K A KT G( s)
q L ( s) N ( s)
=
- Tw ( s) 1 + K P K A KT G( s)
Bs + K
G(s) =
s 2 [ J M J L s 2 + ( J M + J L ) Bs + ( J M + J L ) K ]
J M s 2 + Bs + K
N(s) =
s 2 [ J M J L s 2 + ( J M + J L ) Bs + ( J M + J L ) K ]
3.9 The required block diagram easily follows from Fig. 3.38
K1x Ay = K 2 Dp
A
(K 2 Dp)
= My
K2
(r y) Kp KA K = x
Y ( s) 2
= 2
R( s) s + 0.02 s + 2
3.10
Fw z z
i x
y
er + A + 1 1
KA Kv K1 + s estD
K2 Ms + B
e0
A
Ks
Let z(t) = displacement of the power piston

d
Y(s) = estDZ(s); tD = sec
v
Y ( s) G( s )
=
Er ( s) 1 + G( s ) H ( s )
(K A K v K1 A/K 2 )e - st D
G(s) = ; H ( s) = K s
s(Ms + B + A2 /K 2 )
dh r gh
3.11 A = q - qw -
dt R
AR
t=
rg
Qw(s)

Q(s) + R/rg H(s)
(ts + 1)
hr er + q h
Ks Ka Ke Kv

e0
Ks
dq q - qa
3.12 Ct = h-
dt Rt
qa
1/Rt
+
h + Rt q
ts + 1
t = RtCt
er + Rf eA h q
1 Kh
R
+
R2
Kt
R1
3.13 Heater:
d q1 q -q
C1 = h- 1
dt R1
dq q1 - q
Air: Vrc = - Qrc(q - q i )
dt R1
These equations give us
1 Q
t = R1C1, Kg = , wn =
Qr c R1C1V
R1C1Qrc + V rc + C1
2zwn =
VR1C1 rc
qi
ts + 1
Kg
+ Kg
h + q
1 2 2z
s + wn s + 1
wn2
qr er + eA h q
Ks KA Kh
e0
Ks
4
Use of Feedback for
Control of Uncertain Systems
1 s( s + 1)
4.1 SM
G= =
1 + G( s) H ( s) s( s + 1) + 50
SGM ( jw ) = 0.0289
w =1
-G( s) H ( s) -50
SM
H = =
1 + G( s) H ( s) s( s + 1) + 50
SHM ( jw ) = 1.02
w =1
Y ( s) PD
4.2 = M(s) = 1 1
R( s) D
K 3 K1 5 KK K K
P1 = ; D = 1- - 2 - 1 2 - 2 3 1 ; D1 = 1
s ( s + 1)
2
s ( s + 1) s + 1 s ( s + 1)
5K1
M(s) =
s 2 ( s + 1 + 5K1 ) + 5K1 + 5
M K1 s 2 ( s + 1 + 5K1 ) + 5 - 5K1 s 2
SKM1 = = 2
K 1 M s ( s + 1 + 5K 1 ) + 5K 1 + 5
5
SKM1 ( jw ) = = 0.5
w =0 5K1 + 5
4.3 For G(s) = 20/(s + 1), and R(s) = 1/s,
y(t ) open-loop = 20(1 et)
20
y(t ) closed-loop =(1 - e -21t )
21
For G(s) = 20/(s + 0.4), and R(s) = 1/s,
Use of Feedback for Control of Uncertain Systems 17
y(t ) open-loop = 50(1 e0.4t)
20
y(t ) closed-loop = (1 - e -20.4 t )
20.4
The transient response of the closed-loop system is less sensitive to
variations in plant parameters.
4.4 For G(s) = 10/(t s + 1), and R(s) = 1/s,
ess open-loop =0
1
ess closed-loop
= = 0.0099
1 + 10 K p
For G(s) = 11/(t s + 1), and R(s) = 1/s,
ess open-loop
= 0.1
ess closed-loop
= 0.009
The steady-state response of the closed-loop system is less sensitive
to variations in plant parameters.
4.5 The given block diagram is equivalent to a single-loop block diagram
shown below.
W(s) + Y(s)
0.4 1
s+1
10(K + s)
s+1
Y ( s) 0.4
=
W ( s) 11s + 10 K + 1
For W(s) = 1/s,
0.4
yss = lim sY ( s) = = 0.01
s0 10 K + 1
This gives K = 3.9
4.6 Case I:
Y ( s) s(t s + 1)
=
W ( s) s(t s + 1) + K
For W(s) = 1/s, yss = 0

Case II:
Y ( s) K
=
W ( s) s(t s + 1) + K
For W(s) = 1/s, yss = 1

The control scheme shown in the following figure will eliminate the
error in Case II.
W(s)
+ +
1 Y(s)
K c 1 + Plant
TI s
4.7 (a) Figure 4.16:

Y ( s) 1 50 50 50
= ; Y ( s) = = -
R( s) 10 s + 1 s(10 s + 1) s s + 0.1
y0(t) = 50(1 - e -0.1t ); y0 ss = 50
Figure 4.17:
56500 5000
Y ( s) 100 5000
= ; Y ( s) = = 101 - 101
R( s) 10 s + 101 s(10 s + 101) s 101
s+
10
-101
5000 10
t 5000
yc(t) = 1 - e ycss =
;
101 101
Time-constant has reduced from the value of 10 to 10/101 by

use of feedback.
(b) Refer Example 4.6.
4.8 (a) er = 50 0.03 = 1.5 volts

V ( s) 0.65Ke
(b) = M ( s) =
Er ( s) ( s + 1)(5s + 1) + 0.0195Ke
M K e ( s + 1)(5s + 1)
SKMe = =
Ke M ( s + 1)(5s + 1) + 9.75
325 19.5( s + 1)
(c) V(s) = Er ( s) - W ( s)
( s + 1)(5s + 1) + 9.75 ( s + 1)(5s + 1) + 9.75
Suppose that the engine stalls when constant percent gradient

is x%. Then
325 1.5 19.5 x
vss = lim sV ( s) = - =0
s0 10.75 10.75
This gives x = 25
The derivative term improves the relative stability; and the
integral term has the opposite effect.
4.9
wr + + 1 1 w0
Kt KA Kg KT
Ra Js
Kb
Kt
w o ( s) KA Kg
(a) = M(s) =
w r ( s) 2( s + 5) + K A K g
For wr(s) = 10/s,
125
wo(t) = (1 - e -130 t ); w oss = 9.61 rad/ sec
13
(b) When the feedback loop is opened,

w o ( s) 50 K A
=
w r ( s) 2( s + 5)
KA = 0.192 gives woss = 9.61 for wr = 10
wo(t) = 9.61 (1 e5t)

Time-constant in the open-loop case is 1/5 sec, and in closed-
loop case is 1/130 sec; system dynamics becomes faster with
feedback action
M K 2 s + 10
(c) SKMA = A =
K A M 2 s + 10 + K A K g
Kg = Kwg, where K is a constant.
M K g w g M K g 2 s + 10
SwMg = = =
K g w g M K g M 2 s + 10 + K A Kg
The open-loop transfer function of the system is
KAKg
G(s) =
2( s + 5)
Therefore
SKGA = 1 = SwGg
SKMA = SK A ; Sw g < Sw g
G M G
4.10 (a)
Tw qM = q C
1 1 1
KP + Kg + KT +
qR 1 qL
KA
sLf + Rf Ra Jeqs2 2
sKb
KP
q L ( s) 5K g
(b) = M(s) =
q R ( s) s( s + 1) ( s + 4) + 10 K g
M K g s( s + 1) ( s + 4)
SKMg = =
K g M s( s + 1) ( s + 4) + 10 100
SKMg ( jw ) 4 104
w = 0.1
q L ( s) ( s + 4) /2
(c) =
- Tw ( s) s ( s + 1) ( s + 4) + 10 K g
Wind gust torque on the load shaft = 100 newton-m

Therefore, on the motor shaft

Tw(s) = n 100/s = 50/s
qLss = slim
0
sq L ( s) = 0.1 rad
dV
4.11 (a) = qi q; V = Ch(t), q = h(t)/R
dt
(b)
hr + qi R h 1 q
K
RCs + 1 R

H ( s) KR
(c) = M(s) =
H r ( s) RCs + 1 + KR
M R 1
SM
R= =
R M RCs + 1 + KR
1
SRM =
s=0 1 + KR
KR 1
(d) For Hr(s) = 1/s, hss = ; ess =
1 + KR 1 + KR
4.12 (b) Open-loop case:
q ( s) KR1
(i) = G(s) = ; SKG = 1
Er ( s) R1Cs + 1
qi
1/R1
er + R1 q
+ +
K
R1Cs + 1
et
Kt
q ( s) 1
(ii) = . For qi(s) = 1/s, qss = 1
q i ( s) RiCs
(iii) t = R1C
Closed-loop case:
q ( s) KR1 R1Cs + 1
(i) = M(s) = ; SKM =
Er ( s) R1Cs + 1 + KKt R1 R1Cs + 1 + KKt R1
q ( s) 1 1
(ii) = . For q i ( s) = 1/s, q ss =
q i ( s) R1Cs + 1 + KKt R1 1 + KKt R1
R1C
(iii) t=
1 + KKt R1
dq q - q i q - q a 1 RRC
(c) K(er et) = C + + ; R2 = ,t = 1 2
dt R1 R2 UA R1 + R2
qi 1 + + 1 qa
R1 R2
+
er + + R1R2 q
K
(R1 + R2) (ts + 1)
Kt
q ( s) R1
= ; q ss = R1 / (R1 + R2 )
q a ( s) open-loop
(R1 + R2 ) (t s + 1) open-loop
q ( s) R1
=
q a ( s) closed-loop
(R1 + R2 ) (t s + 1) + KKt R1R2
R1
q ss closed-loop
=
R1 + R2 + KKt R1R2
4.13
r + ef + 1 if KT w
Kt KA
sLf + Rf + 1 Js + B
Kt
Time-constant (without current feedback) = Lf /(Rf + 1)
I f ( s) KA
=
E f ( s) sLf + Rf + K A + 1
Lf
Time-constant (with current feedback) =
Rf + K A + 1
Position control system is shown below.
+ + + KA if KT w w
r
KP 1
sLf + Rf + 1 Js + B s
Kt
KP
4.14 Fig. P4.14a:

q ( s) 25 K
= M(s) = 2
q r ( s) s + 5s + 25 K
M M K s( s + 5)
SK = =
K M s 2 + 5s + 25
SKM ( jw ) = 1.41
w =5
Fig. P4.14 b:
q ( s) KK 2 25 K
= M(s) = 2 = 2
q r ( s) s + (1 + KK1 ) s + KK 2 s + 5s + 25 K
This gives K2 = 25 and K1 = 4

25 K
M(s) =
s + (1 + 4 K ) s + 25 K
2
s( s + 1)
SM
K = 2
s + 5s + 25
SKM ( jw ) 1
w =5
5
Concepts of Stability and
the Routh Stability Criterion
5.1 (a) All the elements in the first column of the Routh array are
+ve. Therefore, all the roots are in the left-half plane.
(b) Two sign changes are found in the first column of the Routh
array. Therefore, two roots are in the right-half plane and the
rest in the left-half plane.
(c) s3-row of the Routh array has zero pivot element, but the
entire row is not all zeros. We replace the pivot element by
e and then proceed with the construction of the Routh array.
As e 0, two sign changes are found in the first column of the
Routh array. Therefore, two roots are in right-half plane and
the rest in the left-half plane.
(d) s1-row of the Routh array is an all-zero row. Auxiliary
polynomial formed using the elements of s2-row is given by
A(s) = s2 + 1
We replace the elements of s1-row with the coefficients of
dA( s)
= 2s + 0
ds
and proceed with the construction of the Routh array.
There are no sign changes in the resulting Routh array; the
characteristic polynomial does not have any root in the right-
half plane. The roots of the 2nd-order auxiliary polynomial
are therefore purely imaginary.
The given characteristic equation has two roots on the
imaginary axis and the rest in the left-hand plane.
(e) Since all the coefficients of the given characteristic polynomial
are of the same sign, the system is unstable. The Routh array
formation is required only if the number of roots in the right-
half plane are to be determined.
Only one sign change (the s0-row has ve pivot element) is

found in the Routh array. Therefore one root is in the right-
half plane and the rest in the left-half plane.
(f) s3-row of the Routh array is an all-zero row. The auxillary
polynomial is
A(s) = 9s4 + 0s2 + 36
We replace elements of s3-row with the coefficients of
dA( s)
= 36 s 3 + 0 s
ds
The resulting Routh array has s2-row with zero pivot element
but the entire row is not all zeros. We replace the pivot
element with e and proceed with the construction of the
Routh array. Two sign changes are found in the first column;
the characteristic polynomial D(s) has two roots in the right-
half plane. Three possibilities exist: (i) the 4th-order auxiliary
polynomial has all the four purely imaginary roots and the two
right-half plane roots are contributed by the factor D(s)/A(s),
(ii) the 4th-order auxiliary polynomial has complex roots with
quadrantal symmetry (two roots in the right-half plane) and
the factor D(s)/A(s) has all the roots is the left-half plane,
and (iii) the 4th-order auxiliary polynomial has two real-axis
roots and two imaginary-axis roots with quadrantal symmetry
and the factor D(s)/A(s) has one root in the right-half plane.
Examining A(s) we find that auxiliary-polynomial roots are
s = 1 j1, 1 j1.
The given system, therefore, has two roots in the right-half
plane and the rest in the left-half plane.
5.2 (a) s1-row of the Routh array is an all-zero row. The auxiliary
polynomial is
A(s) = s2 + 9
By long division
D(s)/A(s) = (s4 + 2s3 + 11s2 + 18s + 18)/A(s) = s2 + 2s + 2
Therefore
D(s) = (s2 + 9) (s2 + 2s + 2)
= (s + j3) (s j3) (s + 1 + j1) (s + 1 j1)
(b) s3-row of the Routh array is an all-zero row. The auxiliary

polynomial is
Concepts of Stability and the Routh Stability Criterion 27
A(s) = s4 = 24s2 25
D(s)/A(s) = (s5 + 2s4 + 24s3 + 48s2 25s 50)/A(s)
=s+2
D(s) = (s + 2) (s4 + 24s2 25)
= (s + 2) (s2 1) (s2 + 25)
= (s + 2) (s + 1) (s 1) (s + j5) (s j5)
3
(c) s -row of the Routh array is an all-zero row. The auxiliary
polynomial is
A(s) = s4 + 3s2 + 2
D(s)/A(s) = (s6 + 3s5 + 5s4 + 9s3 + 8s2 + 6s + 4)/A(s)
= s2 + 3s + 2
D(s) = (s2 + 3s + 2) (s4 + 3s2 + 2)
= (s2 + 3s + 2) (s2 + 1)(s2 + 2)
= (s + 1) (s + 2) (s + j1) (s j1) (s + j 2 ) (s j 2 )
5.3 With s = s 1, the characteristic equation becomes

s 3 + s 2 + s + 1 = 0
s1 -row in the routh array is an all-zero row. The auxiliary polynomial

is
A( s) = s 2 + 1 = ( s + j 1) ( s - j 1)
We replace the elements of s 2 -row with the coefficients of
dA( s)
= 2 s + 0
ds
and proceed with the construction of the Routh array. There are no
sign changes in the resulting Routh array. The s-polynomial does
not have roots to the right of the line at s = 1, and there are two
roots at s = 1 j1. The largest time-constant is therefore 1 sec.
5.4 (a) Unstable for all values of K

(b) Unstable for all values of K
(c) K < 14/9
(d) K > 0.528
5.5 With s = s 1, the characteristic equation becomes

s 3 + 3Ks 2 + (K + 2) s + 4 = 0
From the Routh array, we find that for K > 0.528, all the s-plane
roots lie to the left of the line at s = 1.
4K 20 / ( s + 1)
5.6 G(s) =
2 s + 1 1 + 20 0.2/ ( s + 1)
0.05
H(s) =
4s + 1
1 + G(s) H(s) = 0 gives
8s3 + 46s2 + 31s + 5 + 4 K = 0

From the Routh array, we find that the closed-loop system is stable
for K < 43.3.
5.7 (i) s3 + 5s2 + 9s + K = 0; K < 45

(ii) s3 + 5s2 + (9 K) s + K = 0; K < 7.5
(iii) s4 + 7s3 + 19s2 + (18 K) s + 2 K = 0; K < 10.1
5.8 (a) D(s) = s3 + 10s2 + (21 + K)s + 13 K = 0
(i) K < 70 (ii) K = 70
(iii) for K = 70, A(s) = s2 + 91 = 0
D(s) = (s2 + 91) (s + 10) = (s + j9.54) (s j9.54) (s + 10)
(b) D(s) = s3 + 5s2 + (K 6)s + K = 0

(i) K > 7.5
(ii) K = 7.5
(iii) For K = 7.5, A(s) = s2 + 1.5 = 0
D(s) = (s2 + 1.5) (s + 5) = (s + j1.223) (s j1.223) (s + 5)
(c) D(s) = s4 + 7s3 + 15s2 + (25 + K) s + 2 K = 0

(i) K < 18.1, (ii) K = 28.1
(iii) For K = 28.1, A(s) = s2 + 7.58 = 0
D(s) = (s2 + 7.58) (s2 + 7s + 7.414)
= (s + j2.75) (s j2.75) (s + 5.7) (s + 1.3)
Concepts of Stability and the Routh Stability Criterion 29
5.9 (a) D(s) = s4 + 12s3 + 69s2 + 198s + 200 + K = 0

From the Routh array we find that the system is stable for
K < 666.25.
For K = 666.25,
A(s) = 52.5s2 + 866.25 = 0
This gives
s = j4.06
Frequency of oscillations is 4.06 rad/sec when K = 666.25
(b) D(s) = s4 + 3s3 + 12s2 + (K 16) s + K = 0
From the Routh array, we find that the system is stable for
23.3 < K < 35.7.
For K = 23.3,
A(s) = 9.57s2 + 23.3 = 0; s = j1.56
For K = 35.7,
A(s) = 5.43s2 + 35.7 = 0; s = j2.56

Frequency of oscillation is 1.56 rad/sec when K = 23.3; and 2.56 rad/
sec when K = 35.7.
5.10 D(s) = s3 + as2 + (2 + K) s + 1 + K = 0
From the Routh array we find that for the system to oscillate,
(2 + K) a = 1 + K
1+ K
Oscillation frequency = =2
a
These equations give a = 0.75, K = 2
5.11 D(s) = 0.02s3 + 0.3s2 + s + K = 0

(a) K < 15
(b) For K = 15, the auxiliary equation is A(s) = s2 + 50 = 0. The
oscillation frequency is 7.07 rad/sec at K = 15.
(c) With K = 7.5 and s = s 1,
D(s) = 0.2 s3 + 2.4 s 2 + 4.6 s + 67.8 = 0
Two sign changes are there in the first column of the Routh
array; therefore two roots have time-constant larger than
1 sec.
5.12 (a) Unstable for all Kc
(b) For stability, TD > t.

5.13 (a) D(s) = s3 + 15s2 + 50s + 25 K = 0; K < 30
(b) D(s) = t s3 + (1 + 5t)s2 + 5s + 50 = 0; t < 0.2

2
(c) D(s) = t s3 + (1 + 5t)s2 + 5s + 2.5 K = 0; K < + 10
t
1
5.14 D(s) = s4 + 5s3 + 6s2 + Kcs + Kc a = 0; a =
TI
From the Routh array, we find that K < (30 25a) results in system
stability. Larger the a (smaller the TI), lower is the limit on gain for
stability.
5.15 D(s) = s3 + 15s2 + (50 + 100 Kt)s + 100 K = 0
From the Routh array, we find that for stability K < (7.5 + 15 Kt).
Limit on K for stability increases with increasing value of Kt.
6
Performance Specifications
on System Time Response
6.1 (b) The characteristic equation is

s2 + 10s + 100 = s2 + 2zwns + w2n = 0
(i) wn = 10 rad/sec; z = 0.5; wd = w n 1 - z 2 = 8.66 rad/sec
p - cos- 1 z p
(ii) tr = = 0.242 sec; tp = = 0.363 sec
wd w d
1-z2 4
Mp = e - pz / = 16.32%; ts = = 0.8 sec
zw n
(iii) Kp = lim
s0
G( s) = ; K = lim sG( s) = 10;
v s0
Ka = lim s 2G( s) = 0
s0
1 1 1
(iv) ess = = 0; ess = = 0.1; ess = =
1 + Kp Kv Ka
6.2 Characteristic equation is
s2 + 10s + 10KA = s2 + 2zwns + w2n = 0
(a) KA = 2.5 gives z = 1, which meets the response requirements.
(b) Kp = , Kv = 2.5
5 1
ess = + = 0.4
1 + K p Kv
6.3 (a) Using Rouch criterion, it can be checked that the close-loop
system is stable.
10 10 10 0.2 10 0.2
(i) ess = = = 0; (ii) ess = + = + =2
Kv K v Ka 0.1
(b) The closed-loop system is unstable.
6.4 The closed-loop system is stable. This can easily be verified by Routh
criterion.
5
G(s) =
(0.1s + 1)( s + 1)(0.2 s + 1)
10 10
ess|r = 10 = =
1 + Kp 1+ 5
Y ( s) 0.1s + 1
=
W ( s) ( s + 1)(0.2 s + 1)(0.1s + 1) + 5
For W(s) = 4/s, yss = 4/6

ess|total = 7/3
6.5 (a) The characteristic equation of the system is

900s3 + 420s2 + 43s + 1 + 0.8 KA = 0
Using Routh criterion, we find that the system is stable for KA

< 23.84
(b) Rearrangement of the block diagram of Fig. P6.5 is shown
below.
qr + 50 q
KA 0.016
3s + 1 30s + 1
1
10s + 1
0.016 50K A
G(s) =
(3s + 1)(30 s + 1)
1
H(s) = ; the dc gain of H(s) is unity.
10 s + 1
Kp = lim G( s) = 0.8 KA; ess = 10 =1

s0 1 + Kp
This gives KA = 11.25
Performance Specications on System Time Response 33
Y ( s) 5K w n2
6.6 = 2 =
R( s) s + 20 s + 5K s 2 + 2zw n s + w n2
w2n = 5K, 2zwn = 20
For K = 60, wn = 17.32, z = 0.58

4
ts = = 0.4 sec.
zw n
- pz / 1-z2
Mp = e 100 = 10.68%
Y ( s) 1 1 300
= - 2 = - 2
W ( s) s + 20 s + 300 300 s + 20 s + 300
1 w n2
= - 2 2
w n2 s + 2zw n s + w n
The term inside the bracket is same as Y(s)/R(s) for which peak
overshoot to unit step input is 0.1068. Therefore, maximum value of
1.1068
|y(t)| to unit step disturbance = - = 3.69 103.
300
Y ( s) K w n2
6.7 = 2 = 2
R( s) s + ps + K s + 2zw n s + w n2
For Mp = 9.5%, z = 0.6

p
w2n = K, wn = K ; 2zwn = p, z = = 0.6
2 K
Therefore p = 1.2 K
Y ( s) 1 1
= 2 ; W(s) =
W ( s) s + ps + K s
1
yss = lim sY ( s) = = 0.01. This gives K = 100
s0 K
p = 1.2 100 = 12
K 100
Kv = lim sG( s) = =
s0 p 12
ess = 0.12
Y ( s) 1 9
6.8 = . For F(s) = 9/s. yss = = 0.03
F ( s) Ms 2 + Bs + K K
This gives K = 300 newtons/m

/ 1-z2 0.003
e - pz = = 0.1
0.03
This gives z = 0.59
p p
tp = = =2
wn 1 - z 2 0.81w n
This gives wn = 1.94 rad/sec

From the equation
B K
s2 + s+ = s2 + 2zwns + w2n
M M
we obtain
M = 79.71 kg, B = 182.47 newton-m/(m/sec)
6.9
q ( s) KT /J
w n2
= 2 B KT = 2
q R ( s) s + s+ s + 2zw n s + w n2
J J
Tw

qR + + 1 q
KT
Js2 + Bs
(ln M p )2
z2=
(ln M p )2 + p 2
For Mp = 0.25, z = 0.4
2z
Steady-state error to unit-ramp input = = 0.04
wn
This gives wn = 20
q ( s) 1
= 2
- Tw ( s) Js + Bs + KT
10
qss = = 0.01
KT
From these equations, we obtain the following values of system
parameters.
KT = 1000 newton-m/rad
B = 40 newton-m/(rad/sec)
J = 2.5 kg-m2
6.10 Jw + Bw = KT(wr w)
w ( s) 45
=
w r ( s) 100 s + 50
2p
wr = 50 rad/sec = step input
60
w(t) = 1.5p (1 e0.5t)
wss = 1.5p rad/sec = 45rpm; ess = 5rpm
6.11 Jq + Bq = KT (qr q)
q ( s) 2400
= 2
q r ( s) 150 s + 600 s + 2400
p p 16
qr(s) = ; q(s) = 2
3s 3 s( s + 4 s + 16)
p e - zw nt
q(t) = 1 - sin (w d t + cos- 1 z )
3 1-z2

p - 2t p
= 1 - 1.1547e sin 3.46t +
3 3
- pz / 1 - z 2
Mp = e = 16.3%
6.12 (a)
qR + e ec + 1 qM 1 qL
Ks KA K1
Js2 + Bs 50
sK2
4K A
(b) G(s) =
s( s + 100.5)
Kv = 4KA/100.5
p
For a ramp input qR(t) = p t, the steady-state error =
Kv
5p
The specification is 5 deg or rad.
180
Therefore
100.5 p p
= ; this gives KA = 904.5
4K A 180
The characteristics equation is
s2 + 100.5s + 4 904.5 = 0
wn = 60.15; z = 0.835; Mp = 0.85%; ts = 0.0796 sec
(c) With PI control, the system becomes type-2; and the steady-state
error to ramp inputs is zero provided the closed-loop system is
stable.
1
4K A 1 +
s
G(s) = ; KA = 904.5
s( s + 100.5)
The characteristic equation is
s3 + 100.5s2 + 4 904.5s + 4 904.5 = 0
It can easily be verified that the closed-loop system is stable.
(d) The characteristic equation can be expressed as

(s + 1.0291) (s2 + 99.4508s + 3514.6332) = 0
or
(s + 1.0291) (s + 49.7254 + j32.2803) (s + 49.7254 j32.2803) = 0
The dominance condition is satisfied because the real pole is very

close to the zero. Therefore, the transient response resembles that
of a second-order system with characteristic equation
s2 + 99.4508s + 3514.6332 = 0
wn = 59.28; z = 0.84; Mp = 0.772%; ts = 0.08 sec
6.13 (a)
Ket et
Ia
(const)
Rf
KA Lf
JL
qR qL
6.13 (b)
qR + e+ 1 qM 1 qL
KP KA KT
Jeqs2 50
sKKt
KP
(c) The characteristic equation of the system is

s2 + 100 KKAs + 6KA = s2 + 2zwns + w2n = 0
This gives KA = 2.67; K = 0.024
6.14 (a)
qR + + KT 1 1 qL
KP KA
Ra Js2 + Bs 10
sKb
20K A
(b) Open-loop transfer function G(s) =
s(4 s + 202)
20K A 1
Kv = = ; this gives KA = 1010.
202 0.01
The characteristic equation becomes
s2 + 50.5s + 5050 = 0 = s2 + 2zwns + w2n
This gives z = 0.355; Mp = 30.33%
20(K A + sK D )
(c) G(s) =
s(4 s + 202)
20K A
Kv = = 100; ess = 0.01
202
Characteristic equation of the system is
s2 + (50.5 + 5KD)s + 5050 = 0 = s2 + 2zwns + w2n
For KD = 6.95, z = 0.6
For z = 0.6, Mp = 9.5%
- KA
(d) System zero: s = = 145.32
KD
Real part of closed-loop poles = zwn = 42.636.
The zero will affect the overshoot. The peak overshoot will be
slightly more than 9.5%.
3K A
6.15 (a) G(s) =
s(0.3s + 1)(1 + 2 s)
0.03 0.01
Kv = 3KA; ess = =
3K A KA
Y ( s) 0.3(1 + 2 s)
(b) =
Fw ( s) s (0.3s + 1)(1 + 2 s) + 3K A
0.1
yss =
KA
(c) Characteristic equation is

0.6s3 + 2.3s2 + s + 3KA = 0
For stability, KA < 1.28.
Minimum value of error in part (a) = 7.81 103
Minimum value of error in part (b) = 0.078
6.16 (a) System-type number = 2
H ( s) 0.5s (0.1s + 1)
(b) = 2
- Qw ( s) s (0.1s + 1) + 0.5(K P + sK D )
0.5
hss = - ; KP > 10
0.5K P
0.1s3 + s2 + 0.5KDs + 0.5KP = 0
For stability, KD > 0.1KP
6.17
qr + K2 1 q
K1 + s s+1
Characteristic equation is
s2 + (1 + K1)s + K2 = 0 = s2 + 2zwns + w2n
For K2 = 19.237, K1 = 3;
z = 0.456, wn = 4.386, Mp = 20%, ts = 2
7
PID Control
7.1 (a) Dyss = 5C; Du = 10%

Dyss 5C
K = = = 0.5C /%
Du 10%
t = 10 min
0.5
G(s) =
10 s + 1
t 10
tCL = =
1+ K 1.5
1 1
Kp = slim
0
G(s) = K; ess = =
1+ K 1.5
(b) D(s) = KP
t 7
= 3 gives KP =
1 + KK P 1.5
1
ess = = 0.3
1 + KK P
7.2 A unity-feedback configuration of the given system is shown below.
qr + e 200 0.02 q
D(s) G(s) =
(s + 1)(s + 2)
E ( s) 1 1
= ; e = lim sE(s); qr(s) =
q r ( s) 1 + D( s) G( s) ss s 0 s
(i) ess = 1/3

(ii) ess = 0
(iii) ess = 1/3
PID Control 41
The integral term improves the steady-state performance and the

derivative term has no effect on steady-state error.
7.3 (i) s2 + 1 = 0; oscillatory
(ii) s2 + 2s + 1 = 0; stable
(iii) s3 + s + 2 = 0; unstable
7.4 (a) s2 + Kc TDs + Kc = (s + 1 + j 1.732) (s + 1 j 1.732)

This gives Kc = 4, TD = 0.5
(b) Alternative control scheme is shown below.

qr + + 1 q 1 q
K1 s s

K2
q ( s) K1
= 2 ; K1 = 4, K2 = 2
q r ( s) s + K 2 s + K1
7.5 (a) The system is stable for KP < 9.

1
Kp = slim
0
D(s) G(s) = KP; ess =
1 + KP
ess = 0.1 (10%) is the minimum possible value for steady-state
error. Therefore ess less than 2% is not possible with proportional
compensator.
KI
(b) Replace KP by D(s) = 3 + . The closed-loop system is stable
s
for 0 < KI < 3. Any value in this range satisfies the static accuracy
requirements.
1000 K P
7.6 (a) Kv = = 1000
10
10 + 1000 K D
z= = 0.5
2 100
These equations give KP = 10 and KD = 0.09

(b) The closed-loop poles have real part = zwn = 50. The zero is
present at s = KP/KD = 111.11. The zero will result in pronounced
early peak. z is not an accurate estimate of Mp.
7.7 (a) Kv = KI = 10
(b) With KI = 10, the characteristic equation becomes
D(s) = s3 +10s2 + 100(1 + KP) + 1000 = 0
With s = s 1,
D(s) = s3 + 7 s2 + [100(1 + KP) 17]s + 1009 100 (1 + KP) = 0

In Routh array formation, KP = 0.41 results in auxiliary equation
with purely imaginary roots. Therefore KP = 0.41 results in dominant
s-plane poles with real part = 1.
(c) Routh array formation for D(s) gives the following auxiliary
equation.
A(s) = 7s2 + 868 = 0; s = j 11.14.
The complex-conjugate s-plane roots are 1 j 11.14.
wd = w n 1 - z 2 = 11.14; zwn = 1
These equations give z = 0.09.
(d) D(s) = (s + 7) (s + j 11.14) (s j 11.14) = 0
The third s-plane closed-loop pole is at s = 8. The dominance

condition in terms of third closed-loop pole is reasonably satisfied.
The zero is at s = KI/KP = 24.39. The zero will not give pronounced
early peak. Therefore, z approximately represents Mp.
7.8 (a) With TI = ,

80(1 + TD s)
D(s)G(s) = 2
s + 8 s + 80
The characteristic polynomial becomes
D(s) = s2 + (8 + 80TD)s + 160
TD = 0.216 gives z = 1
PID Control 43
(b) The zero is at s = 1/0.216 = 4.63. Real part of the closed-loop pole
= zwn = 12.65. Therefore, small overshoot will be observed.
(c) The characteristic equation becomes
80
s 3 + 25.28 s 2 + 160 s + =0
TI
TI > 0.0198 for stability.
K
7.9 (a) G(s) = ; Kv = slim
0
sG(s) = K
s( s + 1)
K = 10 will give steady-state unit-ramp following error of 0.1.

D(s) = s2 + s + K = s2 + 2zwns + w2n
K = 10 gives z = 0.158.
10
(b) G(s) =
s( s + 1 + 10 Kt )
D(s) = s2 + (1 + 10 Kt)s + 10 = s2 + 2zwns + w2n
Kt = 0.216 gives z = 0.5.
10 10
Kv = slim sG(s) = =
0 1 + 10 Kt 3.16
ess = 0.316
10 K A
(c) G(s) =
s ( s + 1 + 10 Kt )
10 K A
Kv =
1 + 10 Kt
D(s) = s2 + (1 + 10 Kt)s + 10 KA = s2 + 2zwns + w2n
From these equations, we find that
KA = 10 and Kt = 0.9 give ess = 0.1 and z = 0.5.
(ln M p )2
7.10 (a) z2=
(ln M p )2 + p 2
Mp = 0.1 gives z = 0.59
4
ts = = 0.05. Therefore, wn = 135.59
zw n
5 K1
G(s) =
s (0.2 s + 1 + 100 K 2 )
D(s) = 0.2 s2 + s(1 + 100 K2) + 5 K1 = 0
Therefore s2 + 5(1 + 100 K2)s + 25 K1 = s2 + 2zwns + w2n
This equation gives K1 = 735.39 and K2 = 0.31

5 K1
(b) Kp = , Kv = = 114.9, Ka = 0
1 + 100 K 2
q ( s) 4
7.11 =
Tw ( s) s ( s + 1) + 10 K + 10 Kt s
4
For Tw(s) = 1/s, qss = = 0.05
10 K
This gives K = 8
D(s) = s2 + (1 + 10 Kt)s + 10 K = s2 + 2zwns + w2n
Kt = 0.79 gives z = 0.5
- R4 - R1 de
7.18 e01 = ei ; e02 = ei = R1 Csei = - R1C i
R3 1/Cs dt
0 - e01 0 - e02 0 - e0
+ + =0
R2 R2 R2
R4 de R
ei + R1C i = e0; D(s) = KP + KDs; KP = 4 , KD = R1C
R3 dt R3
- R4 - 1/sC 1 1
7.19 e01 =
R3
ei ; e02 =
R1
ei = -
R1Cs
ei = -
R1C e
i dt
0 - e01 0 - e02 0 - e0
+ + =0
R2 R2 R2
R4 1
R3
ei +
R1 C e dt = e0
i
PID Control 45
KI R 1
D(s) = KP + ; KP = 4 , KI =
s R3 R1C
E0 ( s) R4 R R6 1/sC2
7.20 = + +
Ei ( s) R3 1/sC1 R5 R2
R R 1
= 4 + R1C1 s 6 +
R3 R5 R2C2 s
- R4 R RCs
7.21 e01 = ei ; e02 = - ei = - ei
R3 1 R1Cs + 1
R1 +
sC
R RCs K1 s
e0 = 4 + ei ; D(s) = K 2 +
R3 R1Cs + 1 t 1s + 1
R4
K2 = ; K1 = RC; t = R1C
R3
K1 K
s t s +1+ 1 s
K2 1 K2
(b) D(s) = K 2 1 + = K2
t 1s + 1 t 1s + 1

K1
t1 + s + 1
K2
= K2
t 1s + 1

K1 K t1
Kc = K2; t = t 1 + ; at = a t 1 + 1 = t1; a =
K2 K2 K
t1 + 1
K2
1
- R4 R1 sC
7.22 e01 = ei ; e02 = - R ei
R3 R1 + 1
sC
- R1RC
= ei
R1Cs + 1
R R1RC
e0 = 4 + ei
R3 R1Cs + 1
K1 R
D(s) = K 2 + ; K2 = 4 , K1 = R1RC, t = R1C
t 1s + 1 R3
K /K (t s + 1 + K1 /K 2 )
(b) D(s) = K 2 1 + 1 2 = K 2 1
t 1s + 1 t 1s + 1
K2 t1
= s + 1 (t 1 s + 1)
1 + K1 /K 2 1 + K1 /K 2
K2 t1 t1
Kc = ;t= ; bt = t1 = b ;
1 + K1 /K 2 1 + K 1 /K 2 1 + K1 /K 2
b = 1 + K1/K2
8
Root Locus Plots and System Stability
8.1 (a) sA = 2; fA = 60, 180, 300; intersection with jw-axis at

j5.042; angle of departure fp from 1 + j4 = 40.
(b) sA = 1.25; fA = 45; 135, 225, 315; intersection with
jw-axis at j1.1; angle of departure fp from 1 + j1 = 71.6;
multiple roots at s = 2.3.
(c) Angle of arrival fz at 3 + j4 = 77.5; multiple roots at s =
0.45.
(d) sA = 1.33; fA = 60; 180, 300; intersection with jw-axis at
j 5 ; angle of departure fp from 2 + j1 = 63.43; multiple
roots at s = 1, 1.67.
(e) sA = 1.5; fA = 90, 270
(f) sA = 5.5; fA = 90, 270; multiple roots at s = 2.31, 5.18.
(g) sA = 0.5; fA = 90, 270; intersection with jw-axis at j2.92.
8.2 (i) sA = 2; fA = 60, 180, 300; intersection with jw-axis at

j 10 ; angle of departure fp from 3 + j1 = 71.5.
Two root loci break away from s = 1.1835 at 90. Two root
loci approach s = 2.8165 at 90.
jw
j 10
3 + j1

s
2

(i)
(ii) Intersection with jw-axis at j 2 3 ; angle of departure fp

from - 3 + j 3 = 60.
Three root loci approach the point s = 2, and then break away
in directions 120 apart.
jw jw
j23
3 + j3

60

s
9 4 s
2

(ii) (iii)
(iii) sA = 4; fA = 90, 270

The characteristic equation has three roots at s = 3; the three
root loci originating from open-loop poles approach this point
and then breakaway. Tangents to the three loci breaking away
from s = 3 are 120 apart.
(iv) sA = 1; fA = 45, 135, 45, 135; intersection with the
jw -axis at j1.58; angle of departure fp from 1 + j2 = 90
There are two roots at s = 1, two roots at s = 1 + j1.22, and
two roots at s = 1 j1.22. The break away directions are shown
in the figure.
(v) There are four roots at s = 1. The break away directions are
shown in the figure.
jw jw
1 + j2 1 + j1

45

2 s 2 45 s
(iv) (v)
Root Locus Plots and System Stability 49
8.3 sA = 1; fA = 60, 180, 300; intersection with the jw-axis at j 2 ;

The z = 0.5 loci passes through the origin and makes an angle of q =
cos1 z = 60 with the negative real axis. The point sd = 0.33 + j0.58
on this line satisfies the angle criterion. By magnitude criterion, the
value of K at this point is found to be 1.04. Using this value of K, the
third pole is found at s = 2.33. Therefore,
1.04
M(s) =
( s + 0.33 + j 0.58)( s + 0.33 - j 0.58)( s + 2.33)
8.4 sA = 3; fA = 45, 135, 225, 315; intersection with the jw-axis at

j3.25; departure angle fp from 4 + j4 = 225; multiple roots at s =
1.5.
At the intersection of the z = 0.707 line with the root locus, the
value of K by the magnitude criterion is 130. The remaining pair of
complex roots for K = 130 can be approximately located graphically.
It turns out that real part of complex pair away from jw-axis is
approximately four times as large as that of the pair near jw-axis.
Therefore, the transient response term due to the pair away from
the jw-axis will decay much more rapidly than the transient response
term due to the pair near jw-axis.
8.5 sA = 2.67; fA = 60, 180, 300; intersection with jw-axis at j 32 ;

departure angle fp from 4 + j4 = 45.
The point sd = 2 + j3.4 on the z = 0.5 line satisfies the angle
criterion. By magnitude criterion, the value of K at this point is
found to be 65. Using this value of K, the third pole is found at
s = 4; the dominance condition is not satisfied.

The least value of K to give an oscillatory response is the value of

K at the multiple roots (K = 1.128).
The greatest value of K that can be used before continuous
oscillations occur is the value of K at the point of intersection with
the jw-axis. From the Routh array formulation, it is found that the
greatest value of K is 30, and at this gain continuous oscillations of
frequency 5 rad/sec occur.
8.7 The proof given in Example 8.1.
8.8 The proof follows from Example 8.1.

Draw a line from the origin, tangential to the circular part of
the locus. This line corresponds to damping ratio for maximum
oscillatory response. The tangential line corresponds to z = 0.82 and
the value of K at the point of tangency is 2.1.
8.9 The proof given in Example 8.2.
8.10 s = s + jw
w +1 w -1 w w
tan -1 + tan -1 = tan -1 + tan -1 180 (2q + 1)
s s s s +2
Taking tangents on both sides of this equation, and noting that
w w
tan tan -1 180 = ,
s + 2 s +2
we obtain
w +1 w -1 w w
+ +
s s s s +2
=
w + 1 w - 1 w w
1- 1-
s s s s + 2
Manipulation of this equation gives

2
1 5
s - + w 2 - = 0
2 4
There exists a circular root locus with centre at s = 1/2, w = 0 and the
radius equal to 5 /2 .
8.11 Use the result in Problem 8.7 for plotting the root locus. Complex-
root branches form a circle.
The z = 0.707 line intersects the root locus at two points; the values
of K at these points are 1 and 5. The point that corresponds to K = 5
results in lower value of ts; therefore we choose K = 5(sd = 3 + j3).
- pz 1-z2
For z = 0.707, Mp = e = 4.3%
4 4
ts = = sec
zw n 3
K ( s + 4)
G(s) =
( s + 2)( s - 1)
1 ( s + 2)( s + 1)
E(s) = R(s) = R( s)
1 + G( s) ( s + 2)( s - 1) + K ( s + 4)
For R(s) = 1/s (assuming selection of K that results in stable closed-
loop poles)
-2
ess = lim sE(s) =
s0 -2 + 4K
For K = 5, ess = 1/9
Therefore, steady-state error is 11.11%.

8.12 Breakaway points obtained from the solution of the equation
dK/ds = 0, are s = 0.634, 2.366. Two root loci break away from
the real axis at s = 0.634, and break into the real axis at s = 2.366.
By determining a sufficient number of points that satisfy the angle
criterion, it can be found that the root locus is a circle with centre at
1.5 that passes through the breakaway points.
Both the breakaway points correspond to z = 1. For minimum
steady-state error, we should select larger value of K.
s = 0.634 corresponds to K = 0.0718
s = 2.366 corresponds to K = 14
8.13 The characteristics equation is

0.8K
1+ =0
(10 s + 1)(30 s + 1)(3s + 1)
K
or 1+ = 0; K = 0.000888 K
1 1 1
s + 10 s + 30 s + 3
sA = 0.155; fA = 60, 180, 300; intersection with the jw-axis at
j0.22; multiple roots at s = 0.063
The point on the z = 0.707 line that satisfies the angle criterion is sd
= 0.06 j0.06; the value of K at this point, obtained by magnitude
criterion, is 0.00131. Therefore K = 1.475.
8.14 (a) Ideal feedback sensor:

K 10 K
1+ = 1+ =0
s(0.1s + 1) s( s + 10)
K = 10 gives z = 0.5.
(b) Sensor with appreciable time constant:

K
1+ =0
s(0.1s + 1)(0.02 s + 1)
K
or 1 + = 0: K = 500 K
s( s + 10)( s + 20)
Locate a root locus point on the real axis that gives K = 5000 (i.e., K
= 10). Complex-conjugate roots can then be found by long division.
The process give z = 0.4.
8.15 Characteristic equation is given by

D(s) = s3 + 6s2 + 8s + 0.1K(s + 10) = 0
which can be arranged as
K ( s + 10)
1+ = 0; K = 0.1K
s( s + 2)( s + 4)
sA = 2; fA = 90, 90; intersection with the jw-axis at j 20 ;
The point on z = 0.5 line that satisfies the angle criterion is
sd = 0.75 + j1.3. The value of K at point, obtained by magnitude
criterion, is 1.0154. Hence K = 10.154.
The third closed-loop pole is found at s = 4.5
8.16 Characteristics equation of the system is

s3 + s + 10Kts + 10 = 0
which can be rearranged as
Ks
1+ 2
= 0; K = 10Kt
s + s + 10
Notice that a zero is located at the origin and open-loop poles are
located at s = 0.5 j3.1225. As per the result of Problem 8.9, a
circular root locus exists with the centre at zero and radius equal to
10 .
The point on the z = 0.7 line that satisfies the angle criterion is sd =
2.214 + j2.258.
The gain K corresponding to this point is 3.427. Hence the desired
value of velocity feedback gain Kt is 0.3427.
8.17 Characteristic equation of the system is

D(s) = s(s + 1) (s + 4) + 20 Kts + 20 = 0
which may be rearranged as

Ks
1+ = 0; K = 20Kt
( s + j 2)( s - j 2)( s + 5)
sA = 2.5; fA = 90, 90; departure angle fp from s = j2 is 158.2.
The following two points on z = 0.4 line satisfy the angle criterion:
s1 = 1.05 + j2.41, s2 = 2.16 + j4.97
The value of K at s1 is 0.449, and at s2 is 1.4130.
The third pole corresponding to K = 0.449 is found at s = 2.9, and
for K = 1.413 at 0.68.
From the root locus plot, one may infer that the closed-loop pole at
s = 0.68 is close to system zero at the origin. This is not the case. In
fact, the closed-loop system does not have a zero at the origin:
Y ( s) 20
=
R( s) s ( s + 1)( s + 4) + 20 (1 + Kt s)
In the root locus plot, the zero at the origin was introduced because
of the process of modifying the characteristic equation so that the
adjustable variable K = 20Kt appears as a multiplying factor.
Kt = 0.449 satisfies the dominance condition to a reasonable
extent. For Kt = 1.413, complex-conjugate poles are not dominant.
8.18 The characteristic equation of the system is

s(s + 1) (s + 3) + 2s + 2a = 0
K
1+ 2
= 0; K = 2a
s( s + 4 s + 5)
sA = 1.3333; fA = 60, 60, 180; intersection with the jw-axis at
j 5 ; departure angle fp from complex pole in the upper half of
s-plane = 63.43; multiple roots at s = 1, 1.666.
The point on z = 0.5 line that satisfies the angle criterion is sd =
0.63 + j1.09. The value of K at this point, obtained by magnitude
criterion, is 4.32. Therefore a = 2.16.
The third pole for K = 4.32 is at s = 2.75.
8.19 The characteristic equation of the system is

s3 + 9s2 + 18s + 10a = 0
K
1+ = 0; K = 10a
s ( s + 3)( s + 6)
sA = 3; fA = 60, 60, 180; intersection with the jw-axis at
j 3 2 ; multiple roots at s = 1.268.
The point on z = 0.5 line that satisfies the angle criterion is sd = 1 +

j1.732. The value of K at this point is 28. Therefore a = 2.8. The third
pole for K = 28 is at s = 7.
( s - 2)
8.20 e s = -
s+2
K ( s - 2)
1- = 0 = 1 F(s)
( s + 1)( s + 2)
K (s + jw - 2)
F(s) =
(s + jw + 1)(s + jw + 2)
K (s - 2 + jw )
=
(s + 1)(s + 2) - w 2 + jw (2s + 3)
w w (2s + 3)
F(s) = tan -1 - tan -1
s -2 (s + 1)(s + 2) - w 2
w w (2s + 3)
- 2

s - 2 (s + 1)(s + 2) - w
= tan - 1
1 + w w (2s + 3)
2
s - 2 (s + 1)(s + 2) - w
F(s) = 0 if
w w (2s + 3)
- =0
s - 2 (s + 1)(s + 2) - w 2
Manipulation of this equation gives
(s 2)2 + w2 = 12
Centre = (2, 0) and radius = 12 = 3.464
It can easily be verified that at every point on this circle,
F(s) = 0.
Revisiting Example 8.7 will be helpful.
9
Compensator Design
Using Root Locus Plots

The value of K at the point of intersection of the root locus and
z = 0.3 line is 7.0.
Kv = slim
0
sG(s) = K/5 = 1.4; ts = 4/zwn = 11.6 sec.
With the compensator, the characteristic equation becomes
K (10 s + 1)
1+ =0
s (100 s + 1)( s + 1)( s + 5)
or
K ( s + 0.1)
1+ = 0; K = 0.1 K
s ( s + 0.01)( s + 1)( s + 5)
The value of K at the point of intersection of z = 0.3 line and root
locus is 6.0. Therefore K = 60.
Kv = slim
0
sD(s)G(s) = K/5 = 12; ts = 4/zwn = 12.7 sec.
K ( s + 2.5)
9.2 1+ =0
s ( s + 1)( s + a )
Desired closed-loop pole; sd = 1.6 + j4. Angle criterion at sd is
satisfied when a = 5.3. The value of K at sd is 23.2. The third pole is
at s = 3.1.
K
9.3 (a) 1 + =0
s ( s - 2)
System is unstable for all K > 0.
(b) In practice, the poles and zeros can never exactly cancel since they
are determined by two independent pieces of hardware whose
numerical values are neither precisely known nor precisely fixed.
We should therefore never attempt to cancel poles in the right-
half plane, since any inexact cancellation will result in an unstable
closed-loop system.
K ( s + 1)
(c) 1+ =0
s ( s - 2)( s + 8)
From the Routh array, we find that the loci cross the jw-axis when
K = 19.2. For K > 19.2, the closed-loop poles are always in the left-
half plane and the system is stable.
E ( s) 1 s ( s - 2)( s + 8)
= =
R ( s) 1 + G ( s) s ( s - 2)( s + 8) + Ks + K
For R(s) = 1/s and K > 19.2 (required for stability),
ess = slim
0
sE(s) = 0
For R(s) = 1/s2 and K > 19.2,
-16
ess = slim
0
sE(s) =
K
K ( s + 0.1)
9.4 G(s) = ; K = 4000
s ( s 2 + 0.8 s + 4)
s 2 + 0.8 s + 4
D(s) =
( s + 0.384)( s + 10.42)
D(s) improves transient response considerably, and there is no effect
on steady-state performance.
K
9.5 (a) 1+ =0
s ( s + 2)
4
z = 0.707; ts = = 2. Therefore, zwn = 2
zw n
Proportional control does not meet this requirement.
K P + sK D K ( s + K P /K D )
(c) 1+ = 1+ D
s ( s + 2) s ( s + 2)
At the points of intersection of z = 0.707 line and zwn = 2 line, the
angle criterion is satisfied when KP/KD = 4.
Compensator Design Using Root Locus Plots 57
The root locus plot of

K D ( s + 4)
1+ =0
s ( s + 2)
has circular complex-root branches (refer Problem 8.7). At the
point of intersection of z = 0.707 line and zwn = 2 line, we find by
magnitude criterion that KD = 2. Therefore KP = 8.
9.6 z = 0.45 (specified). Let us select the real part of the desired roots as
zwn = 4 for rapid settling.
The zero of the compensator is placed at s = z = 4, directly below
the desired root location. For the angle criterion to be satisfied at
the desired root location, the pole is placed at s = p = 10.6. The
gain of the compensated system obtained by magnitude criterion, is
96.5. The compensated system is then
96.5( s + 4)
D(s)G(s) =
s ( s + 2)( s + 10.6)
Kv = slim
0
sD(s)G(s) = 18.2
The Kv of the compensated system is less than the desired value of

20. Therefore, we must repeat the design procedure for a second
choice of the desired root.
9.7 Desired root location may be taken as
sd = 1.588 + j3.152
It meets the requirements on Mp and setting time. With D(s) =
s+1
, angle criterion is met. Magnitude criterion gives K = 147.
s + 10
The compensated system has Kv = 2.94.
9.8 z = 0.707; 4/zwn = 1.4 zwn = 2.85
wn = 4; w n 1 - z 2 = 2.85
sd = 2.85 + j2.85
( s + 1.5)2
With D(s) = , angle criterion is satisfied at sd. The value of
( s + 3.5)2
K, found by magnitude criterion, is 37.
9.9 From the root locus plot of the uncompensated system, we find
that for gain of 1.06, the dominant closed-loop poles are at 0.33
j0.58. The value of z is 0.5 and that of wn is 0.66. The velocity error
constant is 0.53.
The desired dominant closed-loop poles are 0.33 j0.58 with a

Kv of 5. The lag compensator
s + 0.1
D(s) =
s + 0.001
gives Kv boost of the factor of 10. The angle contribution of this
compensator at 0.33 + j0.58 is about seven degrees. Since the angle
contribution is not small, there will be a small change in the new root
locus near the desired dominant closed-loop poles. If the damping
of the new dominant poles is kept at z = 0.5, then the dominant poles
from the new root locus are found at 0.28 j0.51. The undamped
natural frequency reduces to 0.56. This implies that the transient
response of the compensated system is slower than the original.
The gain at 0.28 + j0.51 from the new root locus plot is 0.98.
Therefore K = 0.98/1.06 = 0.925 and
0.925( s + 0.1)
D(s) =
s + 0.01
K K
9.10 (a) 1 + = 1+ =0
(2 s + 1)(0.5s + 1) ( s + 0.5)( s + 2)
From the root locus we find that no value of K will yield zwn =
0.75.
s + K I /K P
(c) D(s) = KP + KI/s = K P
s
K P ( s + K I /K P )
D(s)G(s) =
s ( s + 0.5)( s + 2)
At the point corresponding of z = 0.6 and zwn = 0.75 lines, the angle
criterion is satisfied when KI/KP = 0.75.
The gain at the desired point is KP = 2.06. This yields KI = 1.545.
9.11 (a) The dominant closed-loop poles of uncompensated system are

located at s = 3.6 j4.8 with z = 0.6. The value of K is found as
820.
Therefore Kv = slim
0
sG(s) = 820/200 = 4.1.
s + 0.25
Lag compensator D(s) = gives a Kv boost of the factor
s + 0.025
of 10. The angle contribution of this compensator at 3.6 + j4.8 is
1.8, which is acceptable in the present problem.
(b) PI compensator will turn the system to type-2: therefore specification

on Kv will be satisfied.
K ( s + K I /K P )
D(s) = P
s
Placing zero of the compensator near the pole at the origin will
nearly preserve the transient response. Let us place the zero at s =
0.1. At the point of intersection of the root locus with z = 0.6, KKP
= 8.36. Therefore
836 ( s + 0.1)
D(s)G(s) = 2
s ( s + 10)( s + 20)
9.12 (a) Kv = 20 demands K = 2000. However, using Routh criterion, we

find that when K = 2000, the roots of the characteristic equation
are at j10. Clearly the roots of the system when Kv requirement
is satisfied are a long way from satisfying the z requirement. It will
be difficult to bring the dominant roots from the jw-axis to the z =
0.707 line by using a lead compensator. Therefore, we will attempt
to satisfy the Kv and z requirements by a lag compensator.
Form the uncompensated root locus we find that corresponding
to z = 0.707, the dominant roots are at 2.9 j2.9 and value of K is
236. Kv of the uncompensated system = 2.36. Therefore necessary
ratio of zero to pole of the compensator is
z 20
= = 8.5
p 2.36
We will choose z = 0.1 and p = 0.1/9. For this choice, we find that the
angle contribution of the compensator at 2.9 + j.2.9 is negligible.
(b) PI compensator will turn the system to type-2; therefore specification
on Kv will be satisfied. Placing a zero near the pole at the origin will
nearly preserve the transient response.
K P ( s + K I /K P )
D(s) =
s
KK P ( s + 0.1)
Placing a zero near the pole at s = 0.1 gives D(s)G(s) = .
s ( s + 10)2
At the intersection of the root locus with z = 0.707 line, KKP = 245.
K ( s + 1/t 1 )( s + 1/t 2 )
9.13 D(s)G(s) =
s ( s + 1)( s + 5)( s + 1/at 1 )( s + a /t 2 )
1 1 1 a
= 0.05; = 1; = 0.005; = 10
t1 t2 at 1 t2
K ( s + 0.05)
D(s)G(s) =
s ( s + 0.005)( s + 5)( s + 10)
For z = 0.45, we obtain
sd = 1.571 + j3.119, K = 146.3, Kv = 29.3, wn = 3.49
K
9.14 D(s) = KP + KDs = K P 1 + D s
KP
K
K P 1 + D s
KP
(a) D(s)G(s) =
s2 + 1
s2 + 1 + KP + KDs = ( s + 1 + j 3) ( s + 1 - j 3)
This equation gives KP = 3 and KD = 2

Kp = slim
0
D(s)G(s) = 3.
(b) With PID controller

K ( s + 0.2)( s + 1.3332)
D(s) =
s
the angle criterion at sd = -1 + j 3 is satisfied. The gain K is 2.143.
9.15 (a) Uncompensated system:
Mp = 20% z = 0.45
At the point of intersection of z = 0.45 line and the root locus, zwn =
1.1 and K = 56
(b) ts = 4/zwn = 3.6 sec
Kv = slim
0
sG(s) = 2.07
(c) Compensated system:

Mp = 15% z = 0.55
3.6
ts = = 1.45; zwn = 4/1.45 = 2.75
2.5
2.75
wn = = 5; sd = 2.75 + j4.18
0.55
Kv 20
Lead-lag compensated system:
K ( s + 3)( s + 0.9)
D1(s)D2(s)G(s) = ; K = 506
s ( s + 0.15)( s + 3)( s + 9)( s + 13.6)
D1(s) = KP1 + sKD = KD (s + KP1/KD)

Zero at s = 4.9 satisfies angle criterion. Therefore
KK D ( s + 4.9)
G(s)D1(s) =
s ( s + 3)( s + 9)
At the sd point, KKD = 37.5

37.5 4.9
Kv = = 6.8
39
Desired Kv = 20. Including integrator will meet this requirement.
Place the zero at s = 0.1 (near the pole at the origin).
( s + K I /K P 2 )
D2(s) = K P 2 + K I = K P 2
s s
( s + 0.1)
= KP 2
s
KK D K P 2 ( s + 4.9)( s + 0.1)
G(s)D1(s)D2(s) =
s 2 ( s + 3)( s + 9)
At the point of intersection with the z-line, KKD KP2 = 38.2

K
9.16 G(s) =
s( s + 2 + KKt )
E ( s) 1 s ( s + 2 + KKt )
R ( s) = 1 + G( s) = s ( s + 2 + KK ) + K
t
For R(s) = 1/s2,
ess = 2 + KKt 0.35

K
z 0.707; ts 3 sec.
The characteristic equation of the system is
s2 + 2s + KKts + K = 0
KKt s
1+ 2
=0
s + 2s + K
The locus of the roots as K varies (set KKt = 0) is determined from
the following equation:
K
1+ =0
s ( s + 2)
For K = 20, the roots are 1 4.36.
Then the effect of varying KKt is determined from the locus
equation
KKt s
1+ =0
( s + 1 + j 4.36)( s + 1 - j 4.36)
Note that complex root branches follow a circular path.
At the intersection of the root locus and z = 0.707 line, we obtain
KKt = 4.3. The real part of the point of intersection is s = 3.15, and
therefore the setting time is 1.27 sec.
AK
9.17 G(s) =
s ( s + 1)( s + 5) + KKt s
s(s + 1) (s + 5) + KKt (s + A/Kt) = 0
KKt ( s + A/Kt )
1+ =0
s ( s + 1)( s + 5)
The angle criterion at s = 1 + j2 is satisfied when A/Kt = 2.5 Let us
take A = 2 and Kt = 0.8.
AK
Kv = lim sG( s) =
s0 5 + KKt
with K = 10, Kv = 1.54

The closed-loop transfer function of the system is
Y ( s) 20
=
R( s) ( s + 1 + j 2)( s + 1 - j 2)( s + 4)
10
Nyquist/Bode Frequency Response
Plots and System Stability
10.1 (a) Revisiting Example 10.1 will be helpful.

t 1t 2
(i) 1 0 (ii) 0 180 (iv) - 90
t1 + t2
(b) Revisiting Example 10.2 will be helpful.
(i) t j (ii) 0 180
(c) Revisiting Example 10.3 will be helpful.

(i) 180 (ii) 0 270
(d) (i) (t1 + t2) j (ii) 0 270

t 1t 2
(iii) - 180
t1 + t2
t 1t 2 (t 1t 2 )
(e) (i) 180 (ii) 0 0 (iv) - 90
t1 + t2
10.2 (a) Number of poles of G(s) in right half s-plane, P = 1

Number of clockwise encirclements of the critical point, N = 1
Z = number of zeros of 1 + G(s) in right half s-plane = N + P = 2
The closed-loop system is unstable.
(b) P=0
N = (Number of clockwise encirclements of the critical point
number of counterclockwise encirclements of the critical
point) = 0
Z = N + P = 0; the closed-loop system is stable.
10.3 The polar plot of G(jw) for w = 0+ to w = + is given in Fig. P10.3b.

Plot of G(jw) for w = to w = 0 is the reflection of the given polar
plot with respect to the real axis. Since
G(jw)|w = 0 = 180,
G(s) has double pole at the origin. The map of Nyquist contour
semicircle s = rejf, r 0, f varying from 90 at w = 0 through 0 to
+90 at w = 0+, into the Nyquist plot is given by 2f (an infinite
semicircle traversed clockwise).
With this information, Nyquist plot for the given system can easily
be drawn.
(i) P = 0, N = 0; Z = N + P = 0
The closed-loop system is stable
(ii) P = 1, N = 0; Z = N + P = 1
The closed-loop system is unstable.
(iii) P = 0, N = 0; Z = N + P = 0
The closed-loop system is stable.
10.4 (a) The key points to the polar plot are:

G(jw)H(jw)|w = 0 = 180
G(jw)H(jw)|w = = 0270
The intersections of the polar plot with the axes of G(s)H(s)-
plane can easily be ascertained by identifying the real and
imaginary parts of G(jw)H(jw). When we set Im [G(jw)H(jw)]
to zero, we get w = 4.123, and
G(jw)H(jw)|w = 4.123 = 1.428
Similarly, setting Re [G(jw)H(jw)] to zero, we get intersection

with the imaginary axis.
Based on this information, a rough sketch of the Nyquist plot
can easily be made. From the Nyquist plot, we find that N = 2.
Since P = 0, we have Z = N + P = 2, i.e., two closed-loop poles
in right half s-plane.
(b) The key points of the polar plot are (refer Problem 10.1d):
G(jw)H(jw)|w = 0 = 3 j
G(jw)H(jw)|w = = 0 270
Intersection with the real axis at w = 1 2 ;
4
G( jw ) H ( jw ) w = 1 = -
2 3
The map of Nyquist contour semicircle s = rejf, r 0, f varying
from 90 at w = 0 through 0 to +90 at w = 0+, into the
Nyquist/Bode Frequency Response Plots and System Stability 65
Nyquist plot is given by f (an infinite semicircle traversed

clockwise).
Based on this information, a rough sketch of Nyquist plot can
easily be made.
P = 0, N = 2. Therefore Z = N + P = 2
(c) G(jw)H(jw)|w = 0 = 2 180
G(jw)H(jw)|w = = 0 90
No more intersections with real/imaginary axis. From the
Nyquist plot, we find that N = 1. Since P = 1, we have Z = N
+ P = 0; the closed-loop system is stable.
(d) G(jw)H(jw)|w = 0 = 180
G(jw)H(jw)|w = = 0 90
No intersections with real/imaginary axis.
The map of Nyquist contour semicircle s = rejf, r 0, f
varying from 90 at w = 0 through 0 to + 90 at w = 0+,
into the Nyquist plot is given by 2f (an infinite semicircle
traversed clockwise).
With this information, Nyquist plot for the given system can
easily be drawn.
P = 0, N = 0; Z = N + P = 0
1
(e) G(jw)H(jw)|w = 0 = 0
100
G(jw)H(jw)|w = 10 = 0
Consider the Nyquist contour shown in figure below. For the

semicircle s = j10 + rejf, r 0, f varying from 90 at w = 0
through 0 to + 90 at w = 10+,
1 1
G(s)H(s) = lim f
= lim
r 0 ( j 10 + re ) + 100
j 2 r 0 re ( j 20 + re jf )
jf
1
= lim e - jf
r0 r
It is an infinite semicircle from w = 10 to w = 10+ traversed

clockwise.
G(jw)H(jw)|w = = 0180
The Nyquist plot is shown in the figure below.
The Nyquist plot passes through 1 + j0 point. The Nyquist
criterion is not applicable.
jw
Im
j
1
100
j10
s Re
w = 10+ j 0+
j10 w= w=
Nyquist Nyquist
contour plot
(f) G(jw)H(jw)|w = 0 = 8 + j
G(jw)H(jw)|w = = 090
G( jw ) H ( jw ) w = 3 = 2 + j0
No intersection with the imaginary axis.
The map of Nyquist contour semicircle s = rejf, r 0, f
varying from 90 at w = 0 through 0 to +90 w = 0+, into the
Nyquist plot is given by f (an infinite semicircle traversed
clockwise).
With this information, a rough sketch of Nyquist plot can
easily be made (refer Fig. P10.2a).
P = 1, N = 1. Therefore Z = N + P = 0
- K (1 + w 2t 1t 2 ) jw K (t 1 - t 2 )
10.5 G(jw)H(jw) = +
w 2
(1 + w 2t 12 ) w 2 (1 + w 2t 12 )
(i) For t1 < t2, the polar plot of G(jw)H(jw) is entirely in the third
quadrant as shown in the following figure.
P = 0, N = 0, Therefore Z = N + P = 0; the closed-loop system
is stable.
(ii) For t1 > t2, the polar plot of G(jw)H(jw) is entirely in the
second quadrant, as shown in the following figure.
P = 0, N = 2. Therefore Z = N + P = 2; the closed-loop system
is unstable.
Im
(i) (ii)
10.6 (a) G(jw)|w = 0 = 40; G(jw)|w = = 0 270
Intersection with the real axis at 0.8.
This critical point 1 + j0 will be encircle by the Nyquist plot if K >

1/0.8. Since P = 0, we want net encirclements of the cirtical point to
be zero for stability. Therefore K must be less than 5/4.
4(1 + s) K (t 2 s + 1)
(b) G(s) = = ; t1 < t2
s 2 (1 + 0.1s) s 2 (t 1 s + 1)
The Nyquist plot for this transfer function has already been given in
the Solution of Problem 10.5. From this plot we see that the closed-
loop system is stable for all K.
4(1 + 0.1s) K (t 2 s + 1)
(c) G(s) = = 2 ; t1 > t2
s 2 (1 + s) s (t 1 s + 1)
The Nyquist plot for this transfer function has already been given in
the Solution of Problem 10.5. From this plot we see that the closed-
loop system is unstable for all K.
e -0.8 jw
(d) G(jw) =
jw + 1
1
= [(cos 0.8w w sin 0.8w) j(sin 0.8w + w cos 0.8w)]
1 + w2
The imaginary part is equal to zero if
sin 0.8w + w cos 0.8w = 0
This gives
w = tan 0.8w
Solving this equation for smallest positive value of w, we get w =
2.4482.
G(jw)|w = 0 = 1 0; G(jw)|w = 2.4482 = 0.378 + j0
G(jw)|w = = 0
The polar plot will spiral into the w point at the origin (refer
Fig. 10.39).
The critical value of K is obtained by letting G(j2.4482) equal 1.
This gives K = 2.65. The closed-loop system is stable for K < 2.65.
10.7 The figure given below shows the Nyquist plot of G(s)H(s) for K
= 1. As gain K is varied, we can visualize the Nyquist plot in this
figure expanding (increased gain) or shrinking (decreased gain) like
a balloon.
Since P = 2, we require N = 2 for stability, i.e., the critical point
must be encircled ccw two times. This is true if 1.33 K < 1 or K >
0.75.
1 1
10.8 G( jw ) H ( jw ) = 1 + j0; G( jw ) H ( jw ) = 0.05 + j 0
K w =0
K w =
1
G( jw ) H ( jw ) = 0.167 + j0
K w = 0.62
The Nyquist plot is shown in the figure below. If the critical point
lies inside the larger loop, N = 1. Since P = 1, we have Z = N + P = 2
and the closed-loop system is unstable.
If the critical point lies inside the smaller loop, N = 1, Z = N + P =
0 and the closed-loop system is stable. Therefore, for stability,
0.167 K < 1 or K > 6.
Im
w = 0+
Re
w=
10.9 Eliminating the minor-loop, we obtain forward-path transfer

function of an equivalent single-loop system.
K ( s + 0.5)
G(s) =
s3 + s2 + 1
By routh criterion, we find that the polynomial (s3 + s2 + 1) has two
roots in the right half s-plane. Therefore P = 2.
1 1
G( jw ) = 0.5 0; G( jw ) = 0 - 180
K w =0 K w =
1
G( jw ) = - 0.5 + j 0
K w = 1.4
The polar plot intersects the positive imaginary axis; the point of
1
intersection can be found by setting real pat of G(jw), equal to
K
zero. With this information, Nyquist plot can easily be constructed.
From the Nyquist plot we find that the critical point is encircled
twice in ccw if 0.5K < 1 or K > 2. For this range of K, the closed-
loop system is table. The figure given below illustrates this result.
Im
j1.5
j1 1
G(jw)-plot
K
j0.5
w = 1.4 w=
w = 0+
0.5 Re
0.5
10.10 Figure P10.10a:
Ke -2 s
G(s) =
s( s + 1) (4 s + 1)
1 1
G( jw ) = 90; G( jw ) = 2.55 + j0
K w =0 K w = 0.26
1
G( jw ) =0
K w =
The polar plot will spiral into the w point at the origin (refer
Fig. 10.39).
It is found that the maximum value of K for stability is 1/2.55 = 0.385
Fig. P10.10b:
Ke -2 s
D(s)G(s) =
s 2 (4 s + 1)
1 1
D( jw )G( jw ) = 180; D( jw )G( jw ) =0
K w =0
K w =
The polar plot will spiral into the w point at the origin. A rough
sketch of the Nyquist plot is shown in figure below. It is observed
that the system is unstable for all values of K.
Im
w = 10+
Re
10.11 Revisiting Example 10.15 will be helpful.

1 + G(s) = 0
may be manipulated as
e - st D 1
G(s) = = 1 or G1 ( s) = = - e st D
s( s + 1) s( s + 1)
The polar plot of G1(jw) intersects the polar plot of -e jw t D at a

point corresponding to w = 0.75 at the G1(jw)-locus. At this point, the
phase of the polar plot of - e jw t D is found to be 52. Therefore,
0.75 tD = 52 (p/180)
This gives
tD = 1.2 sec
10.12 G(jw)|w = 0 = 2.2 j; G(jw)|w = = 0 270

G(jw)|w = 15.9 = 0.182 + j0; G(jw)|w = 6.2 = 1 148.3
GM = 1/0.182 = 5.5; FM = 31.7
wg = 6.2 rad/sec; wf = 15.9 rad/sec.
10.13 (a) This result can easily proved using Routh criterion.
(b) General shape of the Nyquist plot of given transfer function is
as shown in Fig. P10.2b. When K = 7, the polar plot intersects
the negative real axis at the point 1.4 + j0. The resulting
Nyquist plot encircles the critical point once in clockwise
direction and once in counter-clockwise direction. Therefore
N = 0. Since P = 0, the system is stable for K = 7. Reducing the
gain by a factor of 1/1.4 will bring the system to the verge of
instability. Therefore GM = 0.7.
The phase margin is found as +10
10.14 (a) The system is type-0; so the low-frequency asymptote has a

slope of 0 dB/decade, and is plotted at dB = 20 log 25 = 27.96.
Asymptote slope changes to 20 dB/decade at the first corner
frequency wc1 = 1; then to 40 dB/decade at wc2 = 10, and to
60 dB/decade at wc3 = 20.
Asymptotic crossing of the 0 dB-axis is at wg = 16.
(b) The system is type-1; so the low-frequency asymptote has a
slope of 20 dB/decade and is drawn so that its extension
would intersect the 0 dB-axis at w = K = 50. The asymptote
slope changes to 40 dB/decade at the first corner frequency
at wc1 = 1; then to 20 dB/decade at wc2 = 5, and back to 40
dB/decade at wc3 = 50. The asymptotic gain crossover is at wg
= 10.2.
(c) The system is type-2; the low-frequency asymptote is drawn
with a slope of 40 dB/decade, and is located such that its
extension would interest the 0 dB-axis at w = K = 500 =
22.36. The asymptote slope changes to 60 dB/decade at the
first corner frequency wc1 = 1; then to 40 dB/decade at wc2 =
5, to 20 dB/decade at wc3 = 10 and to 40 dB/decade at wc4 =
50. The asymptotic gain crossover is at wg = 10.
(d) The system is type-1, and the low-frequency asymptote has
a slope of 20 dB/decade and is drawn so that its extension
would cross the 0 dB-axis at w = K = 50. The asymptote slope
changes to 40 dB/decade at wc1 = 10; then to 20 dB/decade at
wc2 = 20; to 40 dB/decade at wc3 = 50; and to 80 dB/decade
at wc4 = 200 (note that the corner frequency for the complex
poles is wc4 = wn = 200).
The asymptotic gain crossover is at wg = 26.
10.15 (a) The low-frequency asymptote has a slope of 20 dB/decade

and passes through the point (w = 1, dB = 20). The asymptote
slope changes to 40 dB/decade at wc = 10.
Compensate the asymptotic plot by 3 dB at w = 10, by 1 dB
at w = 5 and by 1 dB at w = 20. The compensated magnitude
plot crosses the 0 dB line at wg = 7.86.
The phase is computed from
G(jw) = 90 tan1 0.1 w
The phase shift at wg = 7.86 is 128.2, and the resulting phase
margin is 128.2 (180) = 51.8. Because the phase curve
never reaches 180 line, the gain margin is infinity.
(b) Low-frequency asymptote has a slope of 20 dB/decade and
passes through the point (w = 1, dB = 20)
The asymptote slope changes to 60 dB/decade at the corner
frequency wc = 10. Compensate the asymptotic plot by 6 dB
at w = 10, by 2 dB at w = 5 and by 2 dB at w = 20. The
compensated magnitude plot crosses the 0 dB line at wg = 6.8.
The phase is computed from
G(jw) = 90 2 tan1 0.1 w
The phase shift at wg = 6.8 is 158.6. Therefore, the phase
margin is 180 158.6 = 21.4. At a frequency of wf = 10, the
phase shift is 180; the gain margin is 6 dB.
(c) The low-frequency asymptote has a slope of 20 dB/decade
and passes through the point (w = 0.1, dB = 20 log 200 =
46). The slope changes to 40 dB/decade at the first corner
frequency wc1 = 2, and back to 20 dB/decade at the second
corner frequency wc2 = 5. Compensate the asymptotic plot by
3 dB at w = 2, by 1 dB at w = 1, by 1 dB at w = 4, by +3
dB at w = 5, by +1 dB at w = 2.5 and by +1 dB at w = 10. The
compensated magnitude plot crosses the 0 dB line at wg = 9.
The phase shift at this frequency is 106.6. Therefore the
phase margin is 73.4.
Phase never reaches 180 line; the gain margin is infinity.
(d) Reconsider the Bode plot for system of Problem 10.15c.
The time-delay factor e0.1s will not change the magnitude plot;
it will change only the phase characteristics. For the system
with dead-time,
wg = 9 rad/sec; FM = 22; wf = 13.65 rad/sec; GM = 4.2 dB
40 1
(e) G(s) = =
( s + 2) ( s + 4) ( s + 5) 1 1 1
2 s + 1 4 s + 1 5 s + 1
The low-frequency asymptote coincides with 0 dB line; its slope

changes to 20 dB/decade at wc1 = 2, to 40 dB/decade at wc2
= 4, and to 60 dB/decade at wc3 = 5. Compensated magnitude
plot can easily be obtained by applying corrections.
The phase plot crosses the 180 line at wf = 7 rad/sec, and
the gain margin is 20 dB. Since the gain never reaches 0 dB
(wg = 0), the phase margin is infinity.
(f) From the Bode plots of G(jw) we find that gain crossover
frequency wg = 3.16 rad/sec, and the phase margin is 33. The
phase plot is asymptotic to the 180 line in the low-frequency
range; it never reaches 180.
Therefore the GM =
K /5
10.16 G(s) =
1
s( s + 1) s + 1
5
From the Bode plot of the system sketched for K = 10, we find that
FM = 21, GM = 8 dB.
Increasing the gain from K = 10 to K = 100 shifts the 0 dB axis down
by 20 dB. The phase and gain margins for the system with K = 100
are
FM = 30, GM = 12 dB
Thus the system is stable for K = 10 but unstable for K = 100.
K /40
10.17 (a) G(s) =
1 1 1
2 s + 1 4 s + 1 5 s + 1
Since G(s) has all poles in left half s-plane, the open-loop
system is stable. Hence the closed-loop system will be stable
if the frequency response has a gain less than unity when the
phase is 180.
When K = 40, the gain margin is 20 dB (refer Problem 10.15e).
Therefore, an increase in gain of +20 dB (i.e., by a factor of
10) is possible before the system becomes unstable. Hence K
< 400 for stability.
(b) From the Bode plot of G(jw) with K = 1, we find that the
phase crossover frequency wf = 15.88 rad/sec and gain margin
= 34.82 dB. This means that the gain can be increased by
34.82 dB (i.e., by a factor of 55) before the system becomes
unstable. Hence K < 55 for stability.
(c) From the Bode plot of G(jw) with K = 1, we find that wf =
0.66 rad/sec and GM = 4.5 dB. Thus the critical value of K for
stability is 1.67, i.e., K < 1.67 for the system stability.
10.18 From the Bode plot of G(jw), it can easily be determined that
(a) when tD = 0, the GM = 12 dB, FM = 33; and
(b) when tD = 0.04 sec, the GM = 2.5 dB and FM = 18. The
relative stability of the system reduces due to the presence of
dead-time.
(c) The value of tD for the system to be on verge of instability is
obtained by setting the phase margin equal to zero, i.e.,
w gt D 180
G1 ( jw ) w = w - = - 180
g p
where G1(jw) is the system without dead-time, and wg is the

gain crossover frequency. From the Bode plot of G1(jw), we
get wg = 7.4 rad/sec and G1(jwg) = 147
Therefore,
7.4 t D 180
- 147 - = - 180
p
or tD = 0.078 sec
10.19 Since the system have minimum-phase characteristics, zeros (if any)
are in left half s-plane.
(a) The low-frequency asymptote has a slope of 20 dB/decade
and its extension intersects the 0 dB-axis at w = 4. The system
is therefore type-1; 4/s is a factor of the transfer function.
The asymptote slope changes from 20 dB/decade to 0 dB/
decade at w = 2 (this corresponds to the factor (1 + s/2)), then
to 20 dB/decade at w = 10 (this corresponds to the factor
(1 + s/10)). The transfer function
1
4 1 + s
2
G(s) =
1
s1 + s
10
(b) The low-frequency asymptote has a slope of +20 dB/decade

and it intersects 0 dB-axis at w = 0.2. The system is therefore
type-0; Ks is a factor of the transfer function with 20 log 0.2K
= 0 or K = 5.
The low-frequency asymptote has a magnitude of 20 dB at a
frequency wc1, where
20 log (5 wc1) = 20. This gives wc1 = 2.
The transfer function
5s
G(s) =
1 1 1
1 + 2 s 1 + 10 s 1 + 30 s
(c) The low-frequency asymptote has a slope of 6 dB/octave

with a magnitude of 9 dB at w = 1. The system is therefore
type-1; K/s is a factor of the transfer function with 20 log (K/w)
= 9 at w = 1. This gives K = 0.355.
1
0.355 (1 + s) 1 + s
20
G(s) =
1
s1 + s
40
(d) The low-frequency asymptote has a slope of 20 dB/decade
with a magnitude of 40 dB at w = 2.5. The system is therefore
type-1; K/s is a factor of the transfer function with 20 log
(K/2.5) = 40. This gives K = 250.
250
G(s) =
1 1
s 1 + s 1+ s
2.5 40
(e) The low-frequency asymptote has a slope of +12 dB/octave;
the system is therefore type-0 and Ks2 is a factor of the transfer
function. The first corner frequency wc1 = 0.5. The slope of the
asymptote changes to +6 dB/decade at this corner frequency.
Therefore 1/(2s + 1) is a factor of the transfer function.
The asymptotic plot of Ks2/(2s + 1) is shown in the figure

below.
dB
32
20 20 log Kw 2
10
0
0.1 1 w
0.5
From this figure, we find that

(20 log Kw2 20 log 2w)|w = 1 = 32
This gives K = 79.6
79.6 s 2
G(s) =
(1 + 2 s) (1 + s) (1 + 0.2 s)
10.20 The low-frequency asymptote has a slope of 20 dB/decade. Ks is

a factor of the transfer function with 20 log K = 30. This gives K =
31.623. The transfer function
31.623s
G(s) =
1 1
(1 + s) 1 + s 1 + s
5 20
(i) 20 log (31.623wg1) = 0; this gives wg1 = 0.0316
1
(ii) 20 log (31.623 wg2) 20 log wg2 20 log w g 2
5
1
20 log w = 0; this gives wg2 = 56.234
20 g 2
10.21 The low-frequency asymptote has a slope of 6 dB/octave and has a

magnitude of 11 dB at w = 3. K/s is therefore a factor of the transfer
function with 20 log (K/3) = 11. This gives K = 10.64.
10.64
G(s) =
1 1
s s + 1 s + 1
3 8
1 1
G(jw) = 90 tan1 w - tan -1 w
3 8
G(jw) = 180 at w = 4.9
The gain obtained from asymptotic magnitude plot at w = 4.9 is

10.64 4.9
20 log - 20 log = 2.5 dB
4.9 3
Therefore, GM = 2.5 dB
10.22 Revisit Review Example 10.2
1
5 1 + s
10
G(s) =
1 0.6 1 2
s 1 + s 1 + s+ s

2 50 2500
10.23 From the asymptotic magnitude plot we find that the system is
type-0 with a double pole at s = 3. The transfer function obtained
from the magnitude plot is
0.1
G(s) = 2
1
3 s + 1
G(jw) vs w may be compared with the given phase characteristics to
check the accuracy of identification of corner frequency at wc = 3.
11
Performance Specifications on
System Frequency Response
11.4 (a) The relations

1
z4 -z2 + = 0; wr = w n 1 - 2z 2
4Mr2
yield z = 0.6 and wn = 21.8.
Note that (Mr, z) relation gives two values of z for Mr = 1.04; z =
0.6 and z = 0.8. We select z = 0.6 as damping ratio larger than 0.707
yields no peak above zero frequency.
The characteristic equation of the given system is
s2 + as + K = s2 + 2zwns + w2n = 0
This gives K = 475 and a = 26.2

4
(b) ts = = 0.305 sec
zw n
1/2
wb = w n 1 - 2z 2 + 2 - 4z 2 + 4z 4

= 25.1 rad/sec
11.5 From the response curve, we find

Mp = 0.135, tp = 0.185 sec.
(ln M p )2
z2 = gives z = 0.535 for Mp = 0.135
(ln M p )2 + p 2
p
= 0.185 gives wn = 20
wn 1 - z 2
The corresponding frequency response performance indices are:
1
Mr = = 1.11; wr = w n 1 - 2z 2 = 13.25
2
2z 1 - z
1/2
wb = w n 1 - 2z 2 + 2 - 4z 2 + 4z 4 = 24.6

11.6 z = 0.01 fm
The characteristic equation of the system is
t s2 + (1 + KKt)s + K = 0
This gives
1 + KKt
wn = K /t ; z = = 0.01fm
2 Kt
1 2 10 -4 f 2 t - Kt 2 10 -2 f t 10 -4 f 2 t - K
K= m m m t
Kt2
11.7 The polar plot of G(jw) crosses the real axis at w = 1/ t 1t 2 . The
magnitude |G( jw)| at this frequency is given by
Kt 1 t 2
|G(jw)| =
t1 + t2
Therefore
Kt 1 t 2
Gm =1
t1 + t2
This gives
1 1 1
K = +
Gm t 1 t 2
11.8 (a) It can be determined from the Bode plot that wg = 1 rad/sec
and FM = 59.2
2z
(b) FM = tan -1
4z 4 + 1 - 2z 2
FM = 59.2 z = 0.6
wg = w n 4z 4 + 1 - 2z 2
This gives wn = 1.4

Therefore,
1-z2
Mp = e -pz / = 9.48%
4
ts = = 4.76 sec.
zw n
Performance Specications on System Frequency Response 81
11.9 The low frequency asymptote has a slope of 20dB/decade and

intersects the 0 dB-axis at w = 8; 8/s is a factor of the transfer function.
The forward path transfer function can easily be identified as
1 1
8 1 + s 1 + s
2 4
G(s) =
1 1 1
s 1 + s 1 + s 1 + s
8 24 36
The phase curve can now be constructed.
From the phase curve and the asymptotic magnitude curve, we
obtain
FM = 50; GM = 24 dB
Phase margin of 50 gives z = 0.48 which corresponds to Mp 18%.
11.10 (a) From the Bode plot, we get FM = 12

(b) For a phase margin of 50, we require that G(jw)H(jw)
= 1130 for some value of w. From the phase curve of
G(jw)H(jw), we find that G(jw)H(jw) = 130 at w = 0.5. The
magnitude of G(jw)H(jw) at this frequency is approximately
3.5. The gain must be reduced by a factor of 3.5 to achieve a
phase margin of 50.
(c) FM of 50 gives z = 0.48 which corresponds to Mp = 18%.
K ( jw + 2)
11.11 G(jw) =
( jw )2
w
G(jw) = tan -1 - 180 = 130
2
This equation gives w = 2.3835
The magnitude of G(jw) must be unity at w = 2.3835.
K ( jw + 2)
=1
( jw )2 w = 2.3835
This equation gives K = 1.826.

Since the phase curve never reaches 180 line, the gain margin = .
11.12 (a) From the Open-loop frequency response table, we find that wf
1
= 10 rad/sec and GM = 20 log = 3.88 dB; wg = 8 rad/sec
0.64
and FM = 10.
(b) For the desired gain margin of 20 dB, we must decrease the
gain by (20 3.88) = 16.12 dB. It means that magnitude curve
must be lowered by 16.12 dB, i.e., the gain must be changed by
a factor of b, given by 20 log b = 16.12. This gives b = 0.156.
(c) To obtain a phase margin of 60, we must first determine the
frequency at which the phase angle of G(jw) is 120, and
then adjust the gain so that |G(jw)| at this frequency is 0 dB.
From the Bode plot, we find that lowering the magnitude plot
by 16.65 dB gives the desired phase margin. It equivalently
means that the gain should be changed by a factor of a where
20 log a = 16.65. This gives a = 0.147.
11.13 A 9.48% overshoot implies z = 0.6. For this damping, required phase
margin is 59.2.
100K
G(s) =
s ( s + 36)( s + 100)
K / 36
=
1 1
s s + 1 s + 1
36 100
Make a Bode plot for say K = 3.6.
From the plot we find that at w = 14.8 rad/sec, G(jw) = 120.8.

At w = 14.8, the gain is 44.18 dB. The magnitude curve has to be
raised to 0 dB at w = 14.8 to yield the required phase margin. The
gain should be changed by a factor of a where 20 log a = 44.18. This
gives a = 161.808.
Therefore,
K = 3.6 161.808 = 582.51
wg = 14.8 = w n 4z 4 + 1 - 2z 2
This gives
wn = 20.68
p
tp = = 0.19 sec
wn 1 - z 2
11.14 (a) From the Bode plot, it can easily be determined that GM =
5 dB and FM = 30. The gain crossover frequency wg = 0.83
rad/sec.
(b) The magnitude plot should be lowered by 5 dB to obtain a

gain margin of 10 dB, This is achieved by reducing the gain by
5 dB.
(c) The gain at w = 1.27 is 4.8 dB. Therefore, the gain should be
increased by 4.8 dB to obtain a gain crossover frequency of
1.27 rad/sec.
(d) Gain at the frequency which gives G(jw) = 135, should be
0 dB.
From the Bode plot we find that this requires reducing the
gain by 3.5 dB.
11.15 The requirements on FM and wg are satisfied if we increase the gain

to the limit of zero GM with a phase margin 45. From the Bode
plot we find that K = 37.67 meets this requirement.
11.16 (a) From the Bode plot we find 0 dB crossing at a frequency of

0.47 rad/sec with a phase angle of 145. Therefore, the phase
margin is 35. Assuming a second-order approximation, FM =
35 z = 0.33, Mp = 33%.
(b) The zero dB crossing occurs with a phase angle of 118.
Therefore the phase margin is 62. FM = 62 z = 0.64, MP =
7.3%.
11.17 The 0 dB crossing occurs at w = 0.8, with a phase angle of 140
when tD = 0, and 183 when tD = 1. Therefore phase margin of
the system without dead-time is 40; and with dead-time added, the
phase margin becomes 3. We find that with the dead-time added,
the system becomes unstable. Therefore, the system gain must be
reduced in order to provide a reasonable phase margin.
We find from the Bode plot that in order to provide a phase
margin of 30, the gain would have to be decreased by 5 dB, i.e., by
a factor of b where 20 log b = 5. This gives b = 1.78.
We find that the dead-time necessitates the reduction in loop
gain is order to obtain a stable response. The cost of stability is the
resulting increase in the steady-state error of the system as the loop
gain is reduced.
11.18 Make a polar plot of
50/18
G(jw) =
jw (1 + jw / 3)(1 + jw /6)
25
G(jw)|w = 0 = - - j ; G(jw)|w = = 0 270
18
The polar plot intersects the negative real axis at w = 4.24.

The polar plot is tangential to the M = 1.8 circle (refer Eqn. (11.22)).
Therefore, the resonance peak Mr = 1.8.
The bandwidth of a system is defined as the frequency at which
the magnitude of the closed-loop frequency response is 0.707
of its magnitude at w = 0. For the system under consideration,
closed-loop gain is unity at w = 0; therefore, bandwidth is given
by the frequency at which the M = 0.707 circle intersects the
polar plot. This frequency is found to be 3.61 rad/sec. Therefore
wb = 3.61
1 - pz 1-z2
z4 -z2 + 2
= 0; Mp = e 100
4 Mr
wb = w n (1 - 2z 2 ) + 4z 4 - 2z 2 + 2

ts = 4/zwn
The (Mr, z) relation gives z = 0.29, 0.96 for Mr = 1.8. We select z =

0.29 because damping ratio larger than 0.707 yields no peak above
zero frequency, z = 0.29 corresponds to Mp = 38.6%.
For z = 0.29 and wb = 3.61, we get wn = 2.4721. The setting time ts
= 5.58 sec.
11.19 The polar plot is tangential to M = 1.4 circle (refer Eqn. (11.22)) at
a frequency w = 4 rad/sec. Therefore Mr = 1.4 and wr = 4.
1
z4 -z2 + =0
4 Mr2
wr = w n 1 - 2z 2
Mr = 1.4 z = 0.39.
The corresponding Mp = 26.43%.

wr = 4 and z = 0.39 wn = 4.8. The setting time ts = 2.14 sec.
The (Mr , z ) relationship has been derived for standard second-
order systems with zero-frequency closed-loop gain equal to unity.
The answer is based on the assumption that zero-frequency closed-
loop gain for the system under consideration is unity.
11.20 From the Bode plot, we find that
wg = 8.3 rad/sec; wf = 14.14 rad/sec;
GM = 8.2 dB; FM = 27.7

For each value of w, the magnitude and phase of G(jw) are transferred
from the Bode plot to the Nichols chart. The resulting dB vs phase
curve is tangential to M = 6.7 dB contour of the Nichols chart at
w = 9 rad/sec. Therefore Mr = 6.7 dB and wr = 9. The dB vs phase
curve intersects the 3 dB contour of the Nichols chart at w = 13.6.
Therefore bandwidth wb = 13.6 rad/sec.
The following points are worth nothing in the use of Nichols chart.
(i) The frequency parameters (wr or wb) can be easily determined
by transferring the dB and/or phase data form the dB vs phase
curve to the Bode plot.
(ii) The open-loop dB vs phase plot may not be tangential to one
of the constant-M contours of the Nichols chart. One must do
a little interpolation. In the present problem, Mr = 6.7 dB has
been obtained by interpolation.
11.21 Plot the given frequency response on a dB scale against phase angle.
From the dB vs phase curve, we obtain,
GM = 7 dB; FM = 17
The dB vs phase curve when transferred on Nichols chart gives Mr

= 10 dB; wr = 2.75 rad/sec; wb = 4.2 rad/sec
Note that the 3 dB bandwidth definition is applicable to systems
having unity closed-loop gain at w = 0. We have made this
assumption for the system under consideration.
11.22 For parts (a) and (b) of this problem, refer Problem 11.10.
(c) Since the data for 3 dB contour of the Nichols chart has been
provided, we don't require the Nichols chart for bandwidth
determination. On a linear scale graph sheet, dB vs phase
curve of the open-loop frequency response is plotted with data
coming from Bode plot. On the same graph sheet, the 3 dB
contour using the given data is plotted.
The intersection of the two curves occurs at w = 0.911 rad/sec.
Therefore bandwidth wb = 0.911.
2z
(d) FM = tan -1
4z 4 + 1 - 2z 2
wb = w n 1 - 2z 2 + 2 - 4z 2 + 4z 4

FM = 50 z = 0.48; Mp 18%
wb = 0.911 and z = 0.47 wn = 0.7025; ts = 11.93 sec.

11.23 The frequency-response data from the Bode plot of G(jw) when
transferred to the Nichols chart, gives the following result.
Mr = 1.4; wr = 6.9 rad/sec
From the relations
1
Mr = ; wr = w n 1 - 2z 2
2z 1 - z 2
we obtain
z = 0.39; wn = 8.27 rad/sec
Note that the given type-0 system has zero-frequency closed-loop

gain = 54/(25 + 54) = 0.6835. The approximation of the transient
response of the system by z = 0.39 and wn = 8.27 has large error since
the correlations (Mr, wr) (z, wn) used above, have been derived
for systems with unity zero-frequency closed-loop gain.
11.24 (a) Transfer the frequency-response data from the Bode plot of
G(jw) to the Nichols chart.
Shifting the dB vs phase curve down by 5.4 dB makes this
curve tangent to the M = 20 log1.4 contour. Therefore gain
must be changed by a factor of b, where 20 log b = 5.4. This
gives b = 0.54.
(b) The gain-compensated system intersects the 3 dB contour of
the Nichols chart at w = 9.2. Therefore bandwidth wb = 9.2 rad/
sec.
Note that the 3 dB bandwidth definition given by Fig. 11.12
is applicable to systems having unity closed-loop gain at
w = 0. The given system does not satisfy this requirement. The
answer, therefore, is an approximation of the bandwidth.
11.25 (a) From the Bode plot we find the phase crossover frequency wf
= 2.1 and GM = 14 dB. For the gain margin to be 20 dB, the
magnitude plot is to be brought down by 6 dB. The value of K
corresponding to this condition is given by the relation
20 log K = 6
which gives K = 0.5
(b) From the Bode plot it is observed that FM = 60 will be
obtained if gain crossover frequency wg is 0.4 rad/sec. This
is achieved if magnitude plot is brought down by 7 dB. This
condition corresponds to K = 0.446.
(c) From the dB-phase plot on the Nichols chart, we find that for
this plot to be tangent to M = 1 dB contour, the dB-phase curve
must be brought down by 4 dB. This condition corresponds to

K = 0.63. The corresponding value of wr (read of from Bode
plot) is 0.5 rad/sec.
The value of bandwidth (wb) is the frequency at which the
dB-phase plot intersects the 3 dB contour. It is found that wb
= 1 rad/sec.
(d) The dB vs phase plot on the Nicholo's chart when raised by
2.4 dB, intersects the 3 dB contour at w =1.5 rad/sec. This
condition corresponds to K = 1.35.
11.26 Plot the given frequency response on a dB scale against phase angle.
From this dB vs phase curve, we obtain
GM = 16.5 dB; FM = 59
Transfer the dB vs phase curve on the Nichols chart. We find that

this curve when raised by about 5 dB. touches the M = 20 log1.4
contour of the Nichols chart. Therefore the gain should be increased
by a factor of b, where 20 log b = 5. This gives b = 1.75.
The gain-compensated system has GM = 11.6 dB and FM = 42.5.
11.27 The dB vs phase plot becomes tangential to 20 log1.4 dB contour on
Nichols chart if the plot is lowered by about 11 dB. This means that
the gain K must be reduced by a factor of a where 20 log (1/a) = 11.
This gives a = 3.5.
Phase margin of the gain-compensated system is 42.5
Mr = 1.4 z = 0.387; FM = 42.5 z = 0.394
The following comments may be carefully noted.

(i) Zero-frequency closed-loop gain has been assumed to be
unity.
(ii) One is usually safe if the lower of the two values of z is utilized
for analysis and design purposes.
12
Compensator Design Using Bode Plots
12.1 (a) From the Bode plot of G(jw)H(jw), we find that GM = 6 dB

and FM = 17.
(b) Now the gain and phase of the compensator with Kc = 1 are
added to the Bode plot. From the new Bode plot we find that
the gain must be raised by 1.5 dB to have a gain margin of
6 dB.
20 log Kc = 1.5 gives Kc = 1.2
12.2 (a) From the Bode plot of G(jw)H(jw) we find that wg = 4.08 rad/
sec; FM = 3.9; GM = 1.6 dB
(b) Now the gain and phase of the compensator are added to the
Bode plot. From the new Bode plot, we find that
wg = 5 rad/sec; FM = 37.6; GM = 18 dB
The increase in the phase and gain margins implies that lead
compensation increases margin of stability. The increase in wg
implies that lead compensation increases speed of response.
12.3 (a) Kv = lim sD(s)G(s) = K = 12

s 0
(b) From the Bode plot and Nichols chart analysis of the system
with K = 12, we find that FM = 15; wb = 5.5 rad/sec: Mr = 12
dB; wr = 3.5 rad/sec.
(c) The phase margin of the uncompensated system is 15. The
phase lead required at the gain crossover frequency of the
compensated magnitude curve = 40 15 + 5 = 30
1 - sin 30
a = = 0.334
1 + sin 30
The frequency at which the uncompensated system has a
magnitude of -20 log (1/ a ) = 4.8 dB is 4.6 rad/sec. Selecting
Compensator Design Using Bode Plots 89
this frequency as gain crossover frequency of the compensated

magnitude curve, we set
1
= 4.6
at
The transfer function of the lead compensator becomes
12(0.376 s + 1)
D(s) =
0.128 s + 1
(d) FM = 42; wb = 9 rad/sec; Mr = 3 dB;
wr = 4.6 rad/sec
10(0.5s + 1)
12.4 Lead compensator D(s) = meets the phase margin and
0.1s + 1
steady-state accuracy requirements. These requirements are also
10(10 s + 1)
met by the lag compensator D(s) =
100 s + 1
In this particularly simple example, specifications could be met
by either compensation. In more realistic situations, there are
additional performance specifications such as bandwidth, and there
are constraints on loop gain. Had there been additional specifications
and constraints, it would have influenced the choice of compensator
(lead or lag).
12.5 (a) The settling time and peak overshoot requirements on

performance may be translated to the following equivalent
specifications;
z = 0.45; wn = 2.22
z is related to the phase margin by the relation
z
FM = 45
0.01
wn is related to the gain crossover frequency wg by the
relation
wg = w n 4z 4 + 1 - 2z 2
For a closed-loop system with z = 0.45, we estimate from this

relation
wg = 0.82 wn
Therefore, we require a gain crossover frequency wg 1.82
The phase margin of the uncompensated system is 0 because

the double integration results in a constant 180 phase lag.
Therefore, we must add a 45 phase lead at the gain crossover
frequency of the compensated magnitude curve. Evaluating
the value of a, we have
1 - sin 45
a= = 0.172
1 + sin 45
To provide a margin of safety, we select a = 0.132
The frequency at which the uncompensated system has a
magnitude of 20 log (1/ a ) = 8.77 dB is 1.60 rad/sec. This
falls short of the required gain crossover frequency. We should
choose still lower value of a, say 0.099. The frequency at which
the uncompensated system has a magnitude of 20 log (1/ a )
= 10.03 is 1.8 rad/sec. This may be acceptable.
1
wm = = 1.8. This gives t = 1.76
at
The closed-loop system with the compensator
1.76 s + 1
D(s) =
0.174 s + 1
satisfies the performance specifications.
(b) The gain crossover frequency of the compensated system

is required to be at wg= 1.82 rad/sec. At w = 1.82, the
uncompensated Bode plot has the gain 10.4 dB. For the PD
compensator to provide a gain of +10.4 dB at w = 1.82, we
require that D(s)= (1 + sTD)
10.4 = 20 log 1 + (1.82)2 TD2
This gives TD = 1.73

The phase angle contributed by a PD compensator with TD =
1.73 at w = 1.82 is 72.4. Therefore D(s)= (1 + 1.73s) gives a
FM of 72.4 and gain crossover frequency of 1.82.
A A
12.6 Kv = = = 20
ess 0.05 A
K = 20 realizes this value of Kv.
From the Bode plot of
20
G(jw) =
jw ( j 0.5w + 1)
we find that the phase margin of the uncompensated system is 18.

The phase lead required at the gain crossover frequency of the
compensated magnitude curve = 45 18 + 3 = 30.
1 - sin 30
a= = 0.334
1 + sin 30
The frequency at which the uncompensated system has a magnitude

of 20 log (1/ a ) = 4.8 dB is 8.4 rad/sec. Selecting this frequency
as the gain crossover frequency of the compensated magnitude
curve, we set
1
= 8.4
at
Therefore the compensator
0.2 s + 1
D(s) =
0.0668 s + 1
The phase margin of the compensated system is found to be 43.7 . If
we desire to have exactly a 45 phase margin, we should repeat the
steps with decreased value of a.
From the Nichols chart analysis of the compensated and
uncompensated systems, we find that the lead compensator has
increased the bandwidth from 9.5 rad/sec to 12 rad/sec.
12.7 From the Bode plot of

20
G(jw) =
jw ( j 0.5w + 1)
we find that the phase margin of the uncompensated system is 18.
To realize a phase margin of 45, the gain crossover frequency
should be moved to wg where the phase angle of the uncompensated
system is: 180 + 45 + 5 = 130.
From the Bode plot of uncompensated system, we find that wg = 1.5.
The attenuation necessary to cause wg to be the new gain crossover
frequency is 20 dB. The b parameter of the lag compensator can
now be calculated.
20 log b = 20
This gives b = 10. Placing the upper corner frequency of the

compensator a decade below wg, we have 1/t = 0.15.
Therefore, the lag compensator

6.66 s + 1
D(s) =
66.6 s + 1
As a final check, we numerically evaluate the phase margin and the
bandwidth of the compensated system. It is found that that FM =
45 and wb = 2.5 rad/sec.
12.8 From the steady-state requirement, we set K = 100. From the Bode
plot of
100
G(jw) =
jw ( j 0.1w + 1)( j 0.2w + 1)
we find that the phase margin is 40, which means that the
system is unstable. The rapid decrease of phase of G(jw) at the
gain crossover frequency wg = 17 rad/sec, implies that single-stage
lead compensation may be ineffective for this system. (The reader
should, in fact, try a single-stage lead compensator). For the present
system, in which the desired Kv is 100, a phase lead of more than
85 is required. For phase leads greater than 60, it is advisable to
use two or more cascaded stages of lead compensation (refer Fig.
12.15). The design approach may be that of achieving a portion of
the desired phase margin improvement by each compensator stage.
We first add a single-stage lead compensator that will provide a
phase lead of about 42.5, i.e., the compensator will improve the
phase margin to about 2.5.
Since the phase curve of the Bode plot of the uncompensated system
at the gain crossover frequency has a large negative slope, the value
of a = 0.19 given by
1 - sin 42.5
a=
1 + sin 42.5
will not yield a phase margin of 2.5. In fact, when we set a = 0.08
and t = 0.1, the single-stage compensator improves the phase margin
to 3.
0.1s + 1
Therefore D1(s) =
0.008 s + 1
Notice that the single-stage lead compensator not only improves
the phase margin, but also reduces the slope of the phase curve
of the gain crossover frequency. By adding another stage of the
compensator with D2(s) = D1(s), we obtain
100(0.1s + 1)
D2(s)D1(s)G(s) =
s (0.2 s + 1)(0.008 s + 1)2
Note that pole and zero at s = 10 cancel each other.
The final compensated system has a phase margin of 45.
2500 K
12.9 G(s) =
s ( s + 25)
K = 1 satisfies the steady-state performance requirement. The phase
margin of the uncompensated system, read at the gain crossover
frequency wg = 47 rad/sec, is 28.
1 - sin 25
a= = 0.405
1 + sin 25

of 20 log (1/ a ) = 3.93 dB is 60 rad/sec. Selecting this frequency
as gain crossover frequency of the compensated magnitude curve,
we set
1
= 60
at
The transfer function of the lead compensator is
0.026 s + 1
D(s) =
0.01 s + 1
The phase margin of the compensated system is found to be 47.5. By

Nichols chart analysis, it is found that the lead compensator reduces
the resonance peak from 2 to 1.25 and increases the bandwidth from
74 to 98 rad/sec.
12.10 The phase margin of the uncompensated system
2500 K
G(s) = ;K=1
s ( s + 25)
read at the gain crossover frequency wg = 47 rad/sec is 28. To realize
a phase margin of 45, the gain crossover frequency should be
moved to wg where the phase angle of the uncompensated system is:
180 + 45 = 135.
From the Bode plot of the uncompensated system we find that wg
= 25. Since the lag compensator contributes a small negative phase
when the upper corner frequency of the compensator is placed at

1/10 of the value of wg, it is a safe measure to choose wg at somewhat
less than 25 rad/sec, say, 20 rad/sec.
The attenuation necessary to cause wg = 20 to be the new gain

crossover frequency is 14 dB.
The b parameter of the lag compensator can now be calculated.
20 log b = 14. This gives b = 5
Placing the upper corner frequency of the compensator a decade

below wg, we have 1/t = 2. Therefore the lag compensator
0.5 s + 1
D(s) =
2.5 s + 1
From the Bode plotNichols chart analysis we find that the

compensated system has
FM = 51; Mr = 1.2 and wb = 27.5 rad/sec.
12.11 (a) Kv = slim

0
sG(s) = 5
From the Bode plot we find that the uncompensated system
has a phase margin of 20; the system is therefore unstable.
We attempt a lag compensator. This choice is based on the
observation made from Bode plot of uncompensated system
that there is a rapid decrease in phase of G(jw) near wg. Lead
compensator will not be effective for this system.
To realize a phase margin of 40, the gain crossover frequency
should be moved to wg where the phase angle of the
uncompensated system is:
180 + 40 + 12 = 128
From the Bode plot of the uncompensated system, it is found
that wg = 0.5 rad/sec. The attenuation necessary to cause wg to
be the new gain crossover frequency is 20 dB. The b parameter
of the lag compensator can now be calculated.
20 log b = 20. This gives b = 10
Since we have taken a large safety margin of 12, we can place
the upper corner frequency of the compensator at 0.1 rad/sec,
i.e., 1/t = 0.1. Thus the transfer function of the lag compensator
becomes
10 s + 1
D(s) =
100 s + 1
From the Bode plot of compensated system we find that
the phase margin is about 40 and the gain margin is about
11 dB.
(b) Let us design for a FM of 40 + 15 = 55. Uncompensated
system phase is 125 at w = 0.44 rad/sec. Gain at this frequency
is 21 dB.
The PI compensator must provide an attenuation of 21 dB.
1
20 log b = 21. Therefore b = 11.22 =
0.09
1 ( s + 0.115)
Trial-and-error; = 0.115. Therefore D(s) = 0.09
t s
12.12 It easily follows that K = 30 satisfies the specification on Kv. From
the Bode plot of
30
G(jw) =
jw ( j 0.1w + 1)( j 0.2w + 1)
we find that gain crossover frequency is 11 rad/sec and phase margin

is 24. Nichols chart analysis gives wb = 14 rad/sec.
If lead compensation is employed, the system bandwidth will
increase still further, resulting in an undesirable system which will be
sensitive to noise. If lag compensation is attempted, the bandwidth
will decrease sufficiently so as to fall short of the specified value of
12 rad/sec, resulting in a sluggish system. These facts can be verified
by designing lead and lag compensators. We thus find that there is
need to go in for lag-lead compensation.
Since the full lag compensator will reduce the system bandwidth
excessively, the lag section of the lag-lead compensator must
be designed so as to provide partial compensation only. The lag
section, therefore, should move the gain crossover frequency to
a value higher than the gain crossover frequency of the fully lag-
compensated system. We make a choice of new gain crossover
frequency as wg = 3.5 rad/sec. The attenuation necessary to cause
wg to be the new gain crossover frequency is 18.5 dB. This gives the
b parameter of the lag section as
20 log b = 18.5; b = 8.32, say, 10
Placing the upper corner frequency of the lag section at 1/t1 = 1, we
get the transfer function of the lag section as
t 1s + 1 s+1
D1(s) = =
bt 1 s + 1 10 s + 1
It is found that the lag-section compensated system has a phase
margin of 24.
We now proceed to design the lead section. The implementation of
the lag-lead compensator is simpler if a and b parameters of the lead
and lag compensators, respectively, are related as a = 1/b. Let us first
make this choice. If our attempt fails, we will relax this constraint on
a.
The frequency at which the lag-section compensated system has a
magnitude of 20 log (1/ a ) = 10 dB is 7.5 rad/sec. Selecting this
frequency as gain crossover frequency of the lag-lead compensated
magnitude curve, we set
1
= 7.5
at 2
The transfer function of the lead-section becomes
t2s + 1 0.422 s + 1
D2(s) = =
at 2 s + 1 0.0422 s + 1
The analysis of the lag-lead compensated system gives FM = 48 and
wb = 13 rad/sec.
(ln M p )2
12.13 z 2 =
(ln M p )2 + p 2
Mp = 0.2 z = 0.456
p
tp =
wn 1 - z 2
tp = 0.1 wn = 35.3
wb = w n [1 - 2z 2 + 2 - 4z 2 + 4z 4 ]1/2
= 46.576
2z
FM = tan - 1
4z 4 + 1 - 2z 2
48
In order to meet the specification of Kv = 40, K must be set at 1440.

From the Bode plot of
144000
G(jw) =
jw ( jw + 36)( jw + 100)
we find that the gain crossover frequency is 29.7 rad/sec and phase
margin is 34.
1 - sin 24
a = = 0.42
1 + sin 24
of 20 log (1/ a ) = 3.77 dB is 39 rad/sec. Selecting this frequency
as gain crossover frequency of the compensated magnitude curve,
we set
1
= 39
at
The transfer function of the lead compensator becomes
1440(0.04 s + 1)
D(s) =
0.0168 s + 1
It can easily be verified by Nichols chart analysis that the bandwidth

of the compensated system exceeds the requirement. We assume the
peak time specification is met. This conclusion about the peak time
is based on a second-order approximation that should be checked
via simulation.
12.14 K = 2000 meets the requirement on Kv. We estimate a phase margin
of 65 to meet the requirement on z.
(a) From the Bode plot we find that the uncompensated system
with K = 2000 has a phase margin of zero degrees. From the
Bode plot we observe that there is a rapid decrease of phase at
the gain crossover frequency. Since the requirement on phase
lead is quite large, it is not advisable to compensate this system
by a single-stage lead compensator (refer Fig. 12.15).
(b) Allowing 10 for the lag compensator, we locate the frequency
at which the phase angle of the uncompensated systems is:
180 + 65 + 10 = 105. This frequency is equal to 1.5 rad/
sec. The gain crossover frequency wg should be moved to this
value. The necessary attenuation is 23 dB. The b parameter of

the lag compensator can now be calculated.
20 log b = 23. This gives b = 14.2
Placing the upper corner frequency of the compensator
one decade below wg, we have 1/t = 0.15. Therefore the lag
compensator
6.66 s + 1
D(s) =
94.66 s + 1
The phase margin of the compensated system is found to be
67.
(c) The bandwidth of the compensated system is found to be wb =
2.08 rad/sec.
12.15 Approximate relation between FM and z is z 0.01FM
z = 0.4 FM = 40
Let us try lag compensation.

The realize a phase margin of 40, the gain crossover frequency
should be moved to wg where the phase angle of the uncompensated
system is: 180 + 40 + 10 = 130. From the Bode plot of the
uncompensated system we find that, wg = 6. The attenuation
necessary to cause wg to be the new gain crossover frequency is
9 dB.
20 log b = 9. This gives b 3
Placing the upper corner frequency of the lag compensator two
octaves below wg, we have 1/t = 1.5.
The lag compensator
0.67 s + 1
D(s) =
2s + 1
Nichols chart analysis of the compensated system gives wb = 11 rad/
sec. Using second-order approximation
1/2
wb = w n 1 - 2z 2 + 2 - 4z 2 + 4z 4

we get wn = 10
Therefore ts = 4/zwn = 1 sec
Note that the zero-frequency closed-loop gain of the system is
nonunity. Therefore, the use of 3 dB bandwidth definition, and the
second-order system correlation between wb and (z, wn) may result
in considerable error. Final design must be checked by simulation.
12.16 (a) From the Bode plot of uncompensated system we find that
FM = 0.63
(b) Phase margin of the system with the second-order compensator
is 9.47. There is no effect of the compensator on steady-state
performance of the system.
(c) The lead compensator
1 + 0.0378 s
D(s) =
1 + 0.0012 s
meets the requirements on relative stability.
13
Digital Control Systems
13.2 Taking sampling frequency twenty times the closed-loop natural

frequency (which is a measure of bandwidth), we have
2p
ws = 20wn = ; this gives T = 0.5 sec
T
Settling time ts = 4/zwn = 13
1
th of ts is 1.3 sec; T = 0.5 sec is therefore a safe choice.
10
U ( s) 2.2( s + 0.1)
=
E ( s) s + 0.01
2 1 - z-1
s =
T 1 + z-1
This gives
U (z) 2.2[4.1 - 3.9z-1 ]
=
E(z) 4.01 - 3.99z-1
B( s) Kt s
13.3 =
Y ( s) ts + 1
t b (t) + b(t) = Kty (t)

t
1 K
b(t) = b(0) - b(q )dq + t [ y(t ) - y(0)]

t t
0
T b(k ) + b(k - 1) Kt
b(k) = b(k - 1) - + t [ y(k ) - y(k - 1)]
t 2
1 - T /2t K t /t K t /t
= b(k - 1) + y(k ) - y(k - 1)
1 + T /2t 1 + T /2t 1 + T /2t
Digital Control Systems 101
13.4 (a) From the tuning rules given in Table 7.2,

Kc = 3.75; TI = 5 min; TD = 0.8 min
1 1
(b) tCD = t D + T = 2 min + 8 sec = 2.067 min.
2 2
Replacing tD by tCD in the tuning formulas, we obtain
Kc = 3.63; TI = 5.17 min; TD = 0.83 min
13.5 From tuning rules given in Table 7.1,

Kc = 22.5; TI = 28.33 sec.
Correction to account for sampling is not required since the test has
directly been conducted on the digital loop.
13.6 (a) x1: output of the unit delayer

1
x1(k + 1) = - x1 (k ) + r(k ) ;
2
y(k) = 2x1(k) x1 (k + 1) = 2.5x1(k) r(k)
y(k + 1) = 2.5x1(k + 1) r(k + 1) = 0.5y(k) + 2r(k) r(k + 1)
Y (z) -z + 2
=
R(z) 1
z+
2
-z 2
(b) R(z) = 1; Y(z) = +
1 1
z+ z+
2 2
k k-1
-1 -1
y(k) = m (k ) + 2 m (k - 1)
2 2
-5 2
z Y ( z) 3
(c) R(z) = ; = + 3
z-1 z 1 z-1
z+
2
k
5 -1 2
y(k) = - m (k ) + m (k )
3 2 3
1
13.7 2Y(z) 2z1Y(z) + z2Y(z) =
1 - z-1
z3
Y(z) =
(z - 1)(2z2 - 2z + 1)
z - z2 + z
= + 2
(z - 1) 2z - 2z + 1
z 1 (z2 - 0.5z) 1 0.5 z

= - 2
+ 2
z - 1 2 (z - z + 0.5) 2 z - z + 0.5
Therefore,
k k
1 1 kp 1 1 kp ; k 0
y(k) = (1)k - cos + sin

2 2 4 2 2 4
13.8 Taking z-transform of the given equation:

Y (z) z2 z2
= 2 =
R (z) z - 3z + 2 (z - 1)(z - 2)
z+1
R(z) = 1 + z1 =
z
Y (z) (z + 1) 3 2
= = -
z (z - 2)(z - 1) z - 2 z - 1
y(k) = 3(2)k 2(1)k; k 0
Final value theorem will not give correct value since a pole of Y(z)
lies outside the unit circle.
2z - 3
13.9 (a) R(z) = z1; Y(z) =
z (z - 0.5)(z + 0.3)
Y (z) 2z - 3
= 2
z z ( z - 0.5)(z + 0.3)
40 20 10 50
= - + - +
z z2 z - 0.5 z + 0.3
y(k) = 40d(k) + 20d(k 1) 10(0.5)k + 50(0.3)k; k 0
z - 6 z2 + z
(b) R(z) = ; Y(z) =
z -1 1 1 1 1
(z - 1) z - + j z - - j
2 4 2 4
Y (z) -16 8 + j4 8 - j4
= + +
z z-1 1 1 1 1
z- + j z- - j
2 4 2 4
16 16z - 6
= - +
z-1 5
z2 - z +
16
y(k) = 16(1)k + (0.56)k [7.94 sin (0.468k) + 16 cos (0.468k)];
k 0
z2
13.10 (a) R(z) = 1 + z2 + z4 + =
z2 - 1
Y (z) z
=
z ( z - 1)( z + 1)( z - 0.5)(z + 0.3)
-0.833 -0.41 0.476 0.769
= + + +
z - 0.5 z + 0.3 z + 1 z - 1
y(k) = 0.833(0.5)k 0.41 (0.3)k + 0.476(1)k + 0.769(1)k;
k 0
z
(b) R(z) =
z-1
Y (z) 1
= 2
z ( z - 0.5) ( z - 0.1)(z - 1)
-5 2.5 -6.94 4.44
= 2
+ + +
(z - 0.5) z - 0.5 z - 0.1 z - 1
y(k) = 10k(0.5)k + 2.5(0.5)k 6.94(0.1)k + 4.44(1)k; k 0
1 1 - e - sT
13.11 G(s) = ; Gh0(s) =
s+1 s
Gh0G(z) = Z{Gh0G(s)}
-1 1 1 - e -T
= (1 - z ) Z = -T
s( s + 1) z - e
z Y (z) 1 - e -T 1 -1
R(z) = ; = = +
z-1 z -T
(z - e )(z - 1) z - 1 z - e -T
y(k) = 1 e kT; k 0
10 1
10
13.12 G(s) = ; Gh0G(z) = (1 z ) Z s( s + 2)
s( s + 2)
Tz z z
Gh0G(z) = 5(1 z1) 2
- + -2T
(z - 1) 2(z - 1) 2(z - e )
10(z + 0.76)
For T = 0.4 sec, Gh0G =
16(z - 1)(z - 0.46)
1 q (z) Gh0G(z)
13.13 G(s) = ; L =
s ( s + 1) q R (z) 1 + Gh0G(z)
1
Gh0G(z) = (1 z1) Z 2
s ( s + 1)
z (T - 1 + e -T ) + (1 - e -T - Te -T )
=
(z - 1)(z - e -T )
For T = 0.25 sec,
0.0288(z + .092)
Gh0G(z) =
(z - 1)(z - 0.7788)
13.14 The filter is described by the following difference equation.

U (z) z
u(k) = u(k 1) + 0.5 e(k); = 0.5
E(z) z -1
1
Gh0G(z) = (1 z1) Z 2
s ( s + 1)
(T - 1 + e -T )z + (1 - e -T - Te -T )
=
(z - 1)(z - e -T )
1
fs = 5 Hz = = 0.2 sec
T
0.019z + 0.0175
Gh0G(z) = ;
(z - 1)(z - 0.819)
Y (z) Gh0 G(z) D(z) 0.0095z (z + 0.92)

= = 3
R(z) 1 + GhoG(z) D(z) z - 2.81z2 + 2.65z - 0.819
Y (z) Gh0G(z)
13.15 =
R(z) 1 + Gh0GH (z)
1 1 - e -1
Gh0G(z) = (1 z1) Z = -1
s( s + 1) z - e
1 e -1z - 2e -1 + 1
Gh0GH(z) = (1 z1) Z 2 = -1
s ( s + 1) (z - 1)(z - e )
Y (z) (1 - e -1 )(z - 1) z 0.632z

= 2 -1
; R(z) = ; Y(z) = 2
R(z) z - z+1- e z-1 z - z + 0.632
(a sin W)z
ak sin (Wk) 2 2
; a2 = 0.632;
z - (2a cos W)z + a
1
con W = = 0.629; W = 0.80 rad; y(k) = 1.02(0.795)k sin (0.89k)
2a
1 0.632 Y (z) 0.632
13.16 Gh0G(z) = (1 z1) Z = ; = ;
s ( s + 1) z - 0.368 R ( z) z + 0.264
z
R(z) =
z-1
0.632z
Y(z) = ; y(k) = 0.5 [1 (0.264)k]m(k)
(z - 1)(z + 0.264)
13.17 Steady-state error can be calculated from the corresponding

continuous-time system as sampling does not affect steady-state
performance of a continuous-time system.
() + () + 1 1 q( )
1

q ( s) K1
G(s) = =
E ( s) s( Js + K 2 )
Kp = lim G( s) =
s0
K1
Kv = lim sG( s) =
s0 K2
Ka = slim
0
s2G(s) = 0
1
13.18 Kp = zlim
1
Gh0G(z) = ; Kv = lim (z - 1) Gh0G(z) = 3.041
T z1
1
Ka = lim (z - 1)2 Gh0G(z) = 0
T2 z1
ess (unit step) = 0; ess (unit ramp) = 0.33; ess (unit acceleration) =
13.19 z2 1.9z + 0.9307 = 0; z= 0.95 j0.168 = 0.965 e j 0.175 = re jq
ezwnT = 0.965 = r; zwnT = ln r; wnT 1 - z 2 = q
From the above equations, we get

z - ln r - ln r
= ;z=
1-z 2 q (ln r )2 + q 2
1
wn = (ln r )2 + q 2 ; z = 0.199; wn = 8.93
2
13.20 Consider the corresponding continuous-time system.
() + + + ()
() 1
+1
K2 sK1 + K 2
(i) D(s) = K1 + ; D(s)G(s) =
s s ( s + 1)
1 1
Kv = K2; = 0.01 =
Kv K2
Y ( s) s
(ii) = ;
W ( s) s ( s + 1) + (K1 s + K 2 )
1
Y(s) = ; yss = 0
s ( s + 1) + K1 s + K 2
Thus a PI compensator meets the requirements.
K (z - 0.9048)
13.21 Gh0G(z) =
(z - 1)2
Sketch a root locus plot; complex roots lie on a circle. Using
magnitude condition, we obtain the value of K at the closed-loop
pole z = 1. It is 2.1. Therefore, the system is stable for 0 < K < 2.1.
There is a double pole at z = 0.81. The value of K at this point is
obtained as 0.38.
K [(T - 1 + e -T )z-1 + (1 - e -T - Te -T ) z-2 ]
13.22 Gh0G(z) =
(1 - z-1 )(1 - e -T z-1 )
T = 1 sec.
0.3679K (z + 0.7181)
Gh0G(z) =
(z - 1)(z - 0.3679)
The root locus is a circle; the breakaway points are at z = 0.6479 and
z = 2.0841.
At the point of intersection of the root locus with unit circle, we find
by magnitude condition, K = 2.3925. This gain results in marginal
stability.
1.1353K (z + 0.5232)
(b) (i) T = 2 sec, Gh0G(z) =
(z - 1)(z - 0.1353)
Breakaway points are at z = 0.4783, 1.5247; critical gain
K = 1.4557.
3.0183K (z + 0.3010)
(ii) T = 4 sec, Gh0G(z) =
(z - 1)(z - 0.0183)
Breakaway points are at z = 0.3435, 0.9455; critical gain
K = 0.9653. The smaller the sampling period, the larger
the critical gain for stability.
13.23 For T = 1 sec.
(0.368z + 0.264)
Gh0G(z) =
z (z2 - 1.368z + 0.368)
The system is stable for 0 < K < 0.785
K (z + 0.717)
13.24 Gh0G(z) =
(z - 1)(z - 0.368)
The root locus is a circle with breakaway points at z = 0.648, 2.08.
(a) At the point of intersection with the unit circle, K = 0.88 and
z1, 2 = 0.244 j0.97 = 1 1.33 rad = ejwT;
w = 1.33 rad/sec.
(b) The value of K at this point is 0.072.
eT/t = 0.648, t = 2.3 sec.
(c) At the point of intersection of the root locus with z = 0.5 locus,
we get K = 0.18. The line through intersection point makes an
angle of 32 with real axis. Therefore
w nT 1 - z 2 = 32 = 0.558 rad; wn = 0.644 rad/sec
0.2 A (z - 0.5)
13.25 z2 + 0.2Az 0.1A = 0; 1 + = 0;
z2
K (z - 0.5)
Gh0G(z) = ; K = 0.2 A
z2
The root locus is a circle. At the point of intersection with unit circle,
we find by magnitude criterion, K = 0.666. Therefore, A = 3.33.
0.01873K (z + 0.9356)
13.26 Gh0G(z) =
(z - 1)(z - 0.8187)
(a) z = 0.5; wn = 4.5; z = ezwnt wnT(1 z2)
This give z1, 2 = 0.6354 45 = 0.4493 j0.4493
Angle deficiency at point P corresponding to z1 is 72.25 deg.
We choose the zero of the controller to cancel the pole at z =
0.8187. Then the pole of the controller is determined to satisfy
the angle condition at P. This gives
z - 0.8187
D(z) = ; |D(z)Gh0G(z)| = 1 requires K = 13.934
z - 0.1595
13.934(z - 0.8187)
D1(z) =
(z - 0.1595)
1
(b) Kv = lim (z - 1) D(z)Gh0G(z) = 3.
T z1
0.2K (z + 0.368)
13.27 Gh0G(z) =
(z - 0.368)(z - 0.315)
The root locus of the uncompensated system is a circle. At point P

corresponding to the intersection of root locus with z = 0.5 locus, we
get 0.2K = 0.3823 or K = 1.91.
At this point,
wn = 1.65, Kp = 0.957.
We will use lag compensator. Kp is to be increased by a factor of
7.5 z - b1 1 - b1
= 7.837; D(z) = ; = 7.837
0.957 z - b2 1 - b2
Let b2 = 0.98; then b1 = 0.84

wn is slightly decreased with lag compensator (as seen from root-
locus plot of the lag-compensated system), which is acceptable.
14
Control System Analysis Using
State Variable Methods
14.1
u if ia qM qM qL
1 Kg + 1
KT
1 1
n
sLf + Rf s(La + Lg ) + Ra + Rg sJe + Be s

x4 x3 x2 x1
Kb
Je = n2J = 0.4; Be = n2B = 0.01
x 1 = x2; 0.4 x 2 + 0.01x2 = 1.2x3
0.1 x 3 + 19x3 = 100x4 1.2x2; 5 x 4 + 21x4 = 4
or x = Ax + bu
0 1 0 0 0
0 -0.025 3 0 0
A = ;b=
0 -12 -190 1000 0

0 0 0 -4.2 0.2
y = qL = nx1 = 0.5x1
14.2
u 1 if 1 q 1 q
KA KT
sLf + Rf + 1 x3 Js + B x2 s x1
. .
x1 = x2; 0.5x2 + 0.5x2 = 10x3
Control System Analysis Using State Variable Methods 111
.
20x3 + 100x3 = 50u; y = x1
0 1 0 0
A = 0 -1 20 ; b = 0 ; c = [1 0 0]

0 0 -5 2.5
.
14.3 x1 = qM; x2 = q M, x3 = ia; y = qL
.
x1 = x2
.
2x2 + x2 = 38x3
.
2x3 + 21x3 = ea 0.5x2
. k
ea = k1(qR qL) k2 qM, = k1qR 1 x1 k2x2
20
0 1 0 0
A= 0 1
-0.5 19 ; b = 0 ; c =
0 0
20
k1 (k + 0.5) 21 k1
- - 2 -
40 2 2 2
14.4 x1 = w; x2 = ia
Jw + Bw = KT ia
dia
Ra ia + La = ea Kbw
dt
ea = Kcec = Kc [k1(er Ktw) k2 ia]
B KT
-
J J
A= ;
- ( k1 t c + Kb )
K K - (Ra + k2 Kc )
La La

0
b = k1Kc ; c = [1 0 ]
La
-11 6
14.5 A = P-1 AP =
-15 8
1
b = P1 b = ; c = cP = [2 -1]
2
Y ( s) PD s -2 1
= 1 1 = = 2
U ( s) D -1 -2
1 - (-3s - 2 s ) s + 3s + 2
1 X2 1
U X1 = Y
Y ( s) P D + P2 D 2 + P3 D 3 + P4 D 4
= 1 1
U ( s) D
2 s -1 (1 - 8 s -1 ) + 15s -2 (1) + (-2 s -1 )[1 + 11s -1 ] + 2 s -1 (6 s -1 ) 2[1]
=
1 - [-11s -1 + 8 s -1 + (-15s -1 ) (6 s -1 )] + [(-11s -1 ) (8 s -1 )]
1
=
s 2 + 3s + 2
X1
1 2
U Y
6
2
X2
0 1 0
14.6 A= ;b=
0 0 1
1 1 -1 0
A = P-1 AP =
-1 -1; b = P b = 1

|lI A| = |lI A| = l2
14.7 X(s) = (sI A)1 x0 + (sI A)1 bU(s)

= G(s)x0 + H(s) U(s)
s( s + 3) s+3 1 1
1 1
G(s) = -1 s( s + 3) s ; H( s) = s
D D 2
-s -1 s 2 s
D = s3 + 3s2 + 1
14.8
0 0 0
x3 x2 x1
u + +x +x +x
+ 3 + 2 + 1
+
3
+
14.9 Taking outputs of integrators as state variables, we get (x1 being the
output of rightmost integrator),
.
x1 = x2
.
x2= 2x2 + x3
.
x3 = x3 x2 y + u
y = 2x1 2x2 + x3
0 1 0 0
A = 0 -2 1 ; b = 0 ; c = [2 -2 1]

-2 1 -2 1
14.10 (a) G(s) = c (sI A)1 b

s+3
=
( s + 1) ( s + 2)
1
(b) G(s) =
( s + 1) ( s + 2)
.
14.11 x1 = 3x1 + 2x2 + [2x1 1.5x2 3.5x3]
.
x2 = 4x1 5x2
.
x3 = x2 r
-5 0.5 -3.5 0

A = 4 -5 0 ; b = 0 ; c = [0 1 0 ]

0 1 0 -1
14
G(s) = c(sI A)1 b =
( s + 1 ( s + 2) ( s + 7)
14.12 (a) x1 = output of lag 1/(s + 2)

x2 = output of lag 1/(s + 1)
. .
x1 + 2x1 = x2; x2 + x2 = x1 + u
y = x2 + (x1 + u)
-2 1 0
A =
-1 -1 ; b = 1 ; c = [-1 1]; d = 1

(b) x1 = output of lag 1/(s + 2)

x2 = output of lag 1/s
x3 = output of lag 1/(s + 1)
. .
x1 + 2x1 = y; x2 = x1 + u
.
x3 + x3 = x1 + u; y = x2 + x3
-2 1 1 0
A = -1 0 0 ; b = 1 ; c = [0 1 1]

-1 0 -1 1
Y ( s) s+3 2 -1
14.13 (i) = = +
U ( s) ( s + 1) ( s + 2) s + 1 s + 2
-1 0 1
A =
0 -2 ; b = 1 ; c = [2 -1]

u + x1
2
+
+
y
1
+
+ x2
1
+
Y ( s) b3 5
(ii) = 3 = 3
U ( s) 2 2
s + a 1 s + a 2 s + a 3 s + 4 s + 5s + 2
From Eqns (14.51), the second companion form of the state
model is given below.
0 0 -2 5
A = 1 0 -5 ; b = 0 ; c = [0 0 1]

0 1 -4 0
u + x1 + x2 + x3 = y
5

2 5 4
Y ( s) b s 3 + b1 s 2 + b 2 s + b 3 s 3 + 8 s 2 + 17 s + 8
(iii) = 03 = 3
U ( s) s + a 1s2 + a 2 s + a 3 s + 6 s 2 + 11s + 6
From Eqns (14.49), the state model in first companion form is
given below.
0 1 0 0
A= 0 0 1 ; b = 0 ; c = [2 6 2 ]

-6 -11 -6 1
+ + + y
+ + +
1 8 17 8
u + x3 x2 x1

6 11 6
+ +
+ +
Y ( s) s+1 b2 s + b3
14.14 (i) = 3 = 3
U ( s) s + 3s + 2 s s + a 1 s 2 + a 2 s + a 3
2
From Eqns (14.51):

0 0 0 1
A = 1 0 -2 ; b = 1 ; c = [0 0 1]

0 1 -3 0
Y ( s) 1 b3
(ii) = 3 2
= 3 2
U ( s) s + 6 s + 11s + 6 s + a 1 s + a 2 s + a 3
From Eqns (14.49):
0 1 0 0

A= 0 0 1 ; b = 0 ; c = [1 0 0 ]

-6 -11 -6 1
Y ( s) s 3 + 8 s 2 + 17 s + 8 -1 2 1
(iii) = 3 =1+ + +
U ( s) s + 6 s 2 + 11s + 6 s+1 s+2 s+3
-1 0 0 1
A = 0 -2 0 ; b = 1 ; c = [-1 2 1]; d = 1

0 0 -3 1
Y ( s) 1000 s + 5000 b2 s + b3
14.15 (a) = 3 = 3
U ( s) s + 52 s + 100 s s + a 1 s 2 + a 2 s + a 3
2
From Eqns (14.49),

0 1 0 0
0 ; b = 0 ; c = 5000 1000 0
A= 0 1 [ ]
0 -100 -52 1
Y ( s) 1000 s + 5000 50 -31.25 -18.75

(b) = 3 = + +
U ( s) s + 52 s 2 + 100 s s s+2 s + 50
0 0 0 50
A = 0 -2 0 ; b = -31.25 ; c [1 1 1]

0 0 -50 -18.75
Y ( s) 2s2 + 6s + 5 1 1 1
14.16 = = + +
U ( s) ( s + 1) ( s + 2) ( s + 1)
2 2
s+1 s+2
From Eqns (14.60),

-1 1 0 0
A = 0 -1 0 ; b = 1 ; c [1 1 1]

0 0 -2 1
t
A (t - t )
x(t) = e bu(t ) dt
0
e - t te - t 0
-1
-t

eAt = L 1 [( sI - A) ] = 0 e 0

0 0 e -2 t
t - (t - t )
(t - t ) e dt

0 1 - e - t - te - t
t t
A (t - t )
e bu(t ) dt = e - ( t - t ) dt = 1 - e - t
0 0
t 1
(1 - e ) -2 t
e -2 ( t - t ) dt 2
0
y = x1 + x2 + x3 = 2.5 2et tet 0.5 e2t

+
+
1 1
u + x2 + x1
1
+
1 1
+ y
+
+ x3
1
+
14.17
1
0 0
x2 x2
s1 s1
1 s1 1 s1
u
x2 x1
2 2
G ( s) G12 ( s) H 1 ( s)
G(s) = 11 ; H ( s) =
G21 ( s) G22 ( s) H 2 ( s)
X 1 ( s) s -1 (1 + 2 s -1 )
= G11(s) =
x10 1 - (-2 s -1 - 2 s -1 + s -2 ) + (- 2 s -1 ) (-2 s -1 )
s -1 (1 + 2 s -1 ) 1/2 1/2
= = +
D s+1 s+3
X 1 ( s) s -2 (1) 1/2 - 1/2

= G12(s) = = +
x20 D s+1 s+3
X 2 ( s) s -2 (1) 1/2 - 1/2

= G21(s) = = +
x10 D s+1 s+3
X 2 ( s) s -1 (1 + 2 s -1 ) 1/2 1/2
= G22(s) = = +
x20 D s+1 s+3
X 1 ( s) s -2 (1) + s -1 (1 + 2 s -1 ) 1
= H1(s) = =
U ( s) D s+1
X 2 ( s) 1
= H2(s) =
U ( s) s + 1
Zero-input response:
x(t) = eAt x0 = L 1 [G(s)x0]

-t -3t
1 e + e e - t - e -3t x10
= -t
2 e - e -3t e - t + e -3t x20
Zero-state response:
t
A (t - t ) -1
1 - e - t
x(t) = e b u (t ) dt = L [ H ( s ) U ( s )] = -t

0 1 - e
0 1
14.18 A= ; eigenvalues are l1 = - 1, l2 = - 2
-2 -3
Y(s) = c(sI A)1 x0 + c(sI A)1 b U(s)
s+4 1
= +
( s + 1) ( s + 2) s ( s + 1) ( s + 2)
3 2 1/2 1 1/2
= - + - +
s+1 s+2 s s+1 s+2
1 3
y(t) = + 2e - t - e -2 t
2 2
.
14.19 x1 = 3x1 + 2x2 + [7r 3x1 1.5 x2]
.
x2 = 4x1 5x2
-6 0.5 7
A= ; b = ; c = [0 1]
4 -5 0
eAt = L 1 [sI A)1]
1 -4 t 2 -7 t 0.5 -4 t 0.5 -7 t
e + 3e e - e
= 3 3 3

4 e -4 t - 4 e -7 t 2 -4 t 1 -7 t
e + e
3 3 3 3
t
4 4
y(t) = x2(t) = 7 e -4 ( t - t ) - e -7 ( t - t ) dt
0 3 3
28 1 1
= (1 - e -4 t ) - (1 - e -7 t )
3 4 4
e - t + te - t te - t
14.20 eAt = -1 -t
-t
-te e - te
x(t) = e A ( t - t0 ) x(t0 )
Given: x1(2) = 2: t0 = 1, t = 2
Manipulation of the equation gives
x1(2) = 2e1 x1(1) + e1 x2(1) = 2
If x2(1) = 2k, then x1(1) = e1 k
Thus
e 1 - k x1 (1)
is a possible set of states x (1) for any k 0
2k 2
e -2 t -t
a b 1 e a b 1
14.21 -2 t
= ;
-t =
- 2e c d -2 -e c d -1
This gives
a b 2e - t - e -2 t e - t - e -2 t
eAt = = -2 t -t

c d 2e - 2e 2e -2 t - e - t
= L 1 [(sI A)1]
2 1 1 1
s + 1 - s + 2 -
s + 1 s + 2
(sI A)1 =
2 2 2 1
s + 2 - s + 1 -
s + 2 s + 1

s -1
(sI A) =
2 s + 3
0 1
A=
- 2 - 3
c
cA is a triangular matrix with diagonal elements equal
14.22 V =
to unity; | V | = (-1)n for all a s. This proves the result.
i
n-1
cA
14.23 U = [b Ab An 1b] is a triangular matrix with diagonal elements

equal to unity; |U| = (1)n for all ai s. This proves the result.
1 -2
14.24 (i) U = [b Ab] = ; r (U) = 2
0 1
Completely controllable
c 1 -1
V= = ; r (V ) = 1
cA -3 3
Not completely observable.
(ii) The system is in Jordan canonical form; it is controllable but not
observable.
(iii) The system is in controllable companion form. The given system
is therefore controllable. We have to test for observability
property only.
c

r(V) = r cA = 3
cA 2

The system is completely observable.
(iv) Given system is in observable companion form. We have to test
for controllability property only.
r(U) = r[b Ab A2b] = 3
The system is completely controllable.
14.25 (i) Observable but not controllable

(ii) Controllable but not observable
(iii) Neither controllable nor observable.

Controllable and observable realization:
A = 1 ; b = 1 ; c = 1
1
14.26 (i) G(s) = c(sI A)1 b =
s+2
Given state model is in observable companion form. Since there
is a pole-zero cancellation, the state model is uncontrollable.
s+4
(ii) G(s) =
( s + 2) ( s + 3)
The given state model is in controllable companion from. Since
there is a pole-zero cancellation, the model is unobservable.
14.27 (a) |lI A| = (l 1) (l + 2) (l + 1)

The system is unstable.
(b) G(s) = c(sI A)1 b
1
=
( s + 1) ( s + 2)
The G(s) is stable
(c) The unstable mode et of the free response is hidden from the
transfer function representation
10 b2
14.28 (a) G(s) = 2
= 2
s + s s + a1s + a 2
0 1 0
A= ; b = ; c = [10 0 ]
0 -1 1
(b) Let us obtain controllable companion form realization of
10 ( s + 2) 10 s + 20 bs s + b3
G(s) = = 3 = 3
s( s + 1) ( s + 2) s + 3s + 2 s s + a 1 s 2 + a 2 s + a 3
2
0 1 0 0

A = 0 0 1 ; b = 0 ; c = [20 10 0 ]

0 -2 -3 1
(c) Let us obtain observable companion form realization of
10 ( s + 2)
G(s) =
s ( s + 1) ( s + 2)
0 0 0 20
A = 1 0 -2 ; b = 10 ; c = [0 0 1]
0 1 -3 0
2 -5
14.29 F = 1 ; g = 1 ; c = [2 0 ]
0
-1
2
1
zI - F = z2 - z + =D
2
2z + 2
G(z) = c(zI F)1 g =
2 1
z -z+
2
2e -T - e -2T e -T - e -2T
14.30 eAt = -T -2T
;
-2e + 2e -e -T + 2e -2T
T
Aq
0.5 - e -T + 0.5 e -2T
g= e bdq =
0 e -T - e -2T
0.6 0.233 0.2

For T = 1 sec, eAT = ; g = 0.233
-0 . 465 -0 . 0972
Discrete-time model of the plant is,

x(k + 1) = eAT x(k) + gu(k); y(k) = x1(k)
For the feedback system,
e(k) = r(k) x1(k) = u(k)
x(k + 1) = eAt x(k) + g[r(k) x1(k)]

Discrete-time model of the closed-loop system is,
x(k + 1) = Fx(k) + gr(k); y = cx(k)
0.4 0.233 0.2

F= ;g= ; c = [1 0 ]
-0.698 -0.0972 0.233
0.696 0.246 -0.021

14.31 F = eAt = -1
; g = (e - I) A b = 0.747 ;
AT
0. 123 0. 572
c = [2 4]; d = 6
0 1 0
14.32 (a) A =
0 -2 ; b = 2 ; c = [1 0 ]

The closed-loop continuous-time system:

x 1 0 1 x1 0
x = + r
2 -2 -2 x2 2
Characteristic equation:
l2 + 2l + 2 = 0
System is stable.
1
1 (1 - e -2T )
(b) F = e AT
= 2

0 e -2T
1 1
T 1 - e -2q T - + e -2T
g= -2q dq = 2 2

0 2e 1 - e
-2T

Closed-loop digital system:
1 1 1
1 - (T - + e -2T (1 - e -2T
x(k + 1) = 2 2 2 x (k )

-(1 - e -2T ) e -2T

1 1 -2T
T - + e
+ 2 2 r (k )
-2T
(1 - e )

Characteristic equation:
1 1
l2 + - (1 - e -2T ) + T - + e -2T l
2 2
1 1 -2T
+ e -2T + - e - Te -2T = 0
2 2
T = 3 sec:
l2 + 1.4987l + 0.4939 = 0; Unstable
T = 0.4 sec:
l2 1.325l + 0.545 = 0; Stable
1 1
14.34 (a) |lI F| = l - l -
4 2
1 1
Eigenvalues of F are: l1 = , l2 =
4 2
1
(b) G(z) = c(zI F)1 g =
1
z - 4
(c) Controllable but not observable
15
Control System Design Using
State Variable Methods
15.1 (a) k = [k1 k2 k3]

0 1 0
A bk = 0 0 1

-6 - k1 -11 - k2 -6 - k3
Characteristic equation is
s3 + (6 + k3)s2 + (11 + k2)s + 6 + k1 = 0
Desired characteristic equation is
(s + 5)(s + 2 + j3.464)(s + 2 j3.464) = s3 + 9s2 + 36s + 80 = 0
k = [74 25 3]
(b) x = x x
x = (A mc) x ; m = [m1

T
m2 m3]
- m2 1 0

A mc = - m2 0 1
-6 - m3 -11 -6
|sI (A mc)| = s3 + (6 + m1)s2 + (11 + 6m1 + m2)s + 11m1

+ 6m2 + m3 + 6
(s + 5)(s + 2 + j3.464)(s + 2 j3.464) = s3 + 9s2 + 36s + 80
m1 = 3, m2 = 7, m3 = 1
15.2 (a) k = [k1 k2 k3]

-k1 - k2 -6 - k3
A bk = 1 0 -11
0 1 -6
|sI (A bk)| = (s + 5)(s + 2 + j3.464)(s + 2 j3.464)

k1 = 3, k2 = 7, k3 = 1
(b) mT = [m1 m2 m3]
|sI (A mc)| = (s + 5)(s + 2 + j3.464)(s + 2 j3.464)
m1 = 74, m2 = 25, m3 = 3
0 9 9
15.3 (a) A=
1 0 ; b = 0 ; c = [0, 1]

(b) k = [k1 k2]
-9k1 9 - 9 k2
(A bk) = ; |sI (A bk)| = s2 + 9k1s + 9k2 9
1 0
Desired polynomial: (s + 3 + j3)(s + 3 j3) = s2 + 6s + 18

k = [2/3 3]
(c) m = [m1 m2]T
0 9 - m1 2
A mc = ; |sI (A mc)| = s + m2s + m1 9
1 - m2
Desired polynomial: (s + 6 + j6)(s + 6 j6) = s2 + 12s + 72.
81
m=
12
15.4 y = w20y = u
. 0 1 0
(a) x= 2 x + u
-w 0 0 1
(b) wn = 2w0; z = 1; s = 2w0
Characteristic polynomial: (s + 2w0)2 = s2 + 4w0s + 4w20
0 1
A bk = 2 ;
-w 0 - k1 -k2
|sI (A bk)| = s2 + k2s + w20 + k1

k = [3w20 4w0]
(c) m = [m1 m2]T

20w 0
|sI (A mc)| = (s + 10w0)2 gives m = 2
.
99w 0
15.5 (a) k = [k1 k2]
0 1
(A bk) =
20.6 - k1 - k2
Characteristic polynomial: |sI (A bk)| = s2 + k2s 20.6 + k1
Desired polynomial: (s + 1.8 + j2.4)(s + 1.8 j2.4) = s2 + 3.6s + 9
k = [29.6 3.6]
(b) M = [m1 m2]T
16
|sI (A mc)| = (s + 8)2 gives m =
84.6
0 1 0
15.6 (a) A= ; b = ; c = [1 0 ]
0 0 1
1
(b) wn = 1; z = ;
2
|sI (A bk)| = s2 + 2zwns + w20 gives
k = [k1 k2] = 1 2
(c) wn = 5; z = 0.5; s2 + 2zwns + w20 = s2 + 5s + 25

|sI (A mc)| = s2 + 5s + 25 gives
m 5
m = 1 =
m2 25
15.7 (a) Desired polynomial: (s + 2)(s + 1 + j1)(s + 1 j1)
= s3 + 4s2 + 6s + 4
. . .
x1 = x2; x2 = x2 + x3; x3 = 2x3 + u; y = x1
x 1 = x - r; x 2 = x2 ; x 3 = x3
x = Ax + bu
0 1 0 0
A = 0 -1 1 ; b = 0

0 0 -2 1
~ = k x k x k x + k r; N = k ;
u = -kx 1 1 2 2 3 3 1 1
0 0 1
U = [b Ab A b] = 0 1 -3
2
1 -2 4
p1 = [1 0 0]
p1 1 0 0

P = p1A = 0 1 0 ; | sI - A |= s 2 + 3s 2 + 2 s
p A 2 0 -1 1
1
k = [4 6 - 2 4 - 3]; k = [4 3 1]
(b) Desired polynomial: s3 + 8s2 + 24s + 32

.
z ATz + cTh; h = mTz
1 0 0
U = [cT ATcT (AT)2 cT] = 0 1 -1
0 0 1
p1 0 0 1

p1 = [0 0 1]; P = p1AT = 0 1 -2

p (AT ) 2 0 -3 4
1
3 2
|sI A| = s + 3s + 2s
mT = [32 24 2 8 3]; mT = [5 7 8]
x1 -1 0 0 x1 1 0

15.8 x2 = 0 -2 0 x2 + 1 u + 0 r
x 3 -1 -3 0 x3 0 1

Desired polynomial: s3 + 4s2 + 8s + 8
|sI A| = s3 + 3s2 + 2s; k = [1.4 2.4 1.6]
u(t) = 1.4x1 2.4x2 + 1.6 (r - y)dt
-3 2 1
15.9 A= ; b = ; c = [0 1]
4 -5 0
(a) |sI (A bk)| = (s + 4)(s + 7) gives k = [3 1.5]

. .
(b) x1 = 3x1 + 2x2 3x1 1.5x2 + r; x2 = 4x2 5x2
1
x2s() = y() = r
7
.
(c) x3 = r y = r x2
x1 -3 2 0 x1 1 0
=
x2 4 -5 0 x2 + 0 u + 0 r
x 3 0 -1 0 x3 0 1
k = [2 1.5 3.5]
0 1 0 0
0 0 1 ; g = 0
15.10 F=
-4 -2 -1 1
Desired characteristic polynomial:

1 1 1 1 3 2 1
z z + + j z + 2 - j 2 = z + z + 2 z
2 2
|zI (F gk)| = z3 + (1 + k3)z2 + (2 + k2)z + 4 + k1
3
k = -4 - 0
2
0 1 0 0
0 0 1 ; g = 0

15.11 F =
-0.5 -0.2 1.1 1
|zI (F gk)| = z3 + (k3 1.1)z2 + (k2 + 0.2)z + k1 + 0.5
Desired Polynomial: z3
k = [0.5 0.2 1.1]
2 -1 4
15.12 (a) F=
-1 1 ; g = 3

4 5
U = [g Fg] =
3 -1
1 p 1 3 -4
p1 = [3 -4 ]; P = 1 =
19 p1 F 19 10 -7
1 1 1
|zI F| = z2 3z + 1; |zI (F gk)| = z + j z - j = z2 +
2 2 4
-3 111 -18
k = 4 3 ; k =
76 19
(b) x = (k + 1) = (F mc) x (k) + gu(k) + my(k); c = [1 1]
z(k + 1) = FTz(k) + cTh(k); h(k) = mTz(k)
1 1
U = [cT FT cT ] = ; p1 = [1 -1]
1 0
p1 1 -1
P = T
=
p1F 3 -2
|zI FT| = z2 3z + 1
|zI (FT mTcT)| = z2
mT = [1 3]; mT = [8 5]
15.13 Double integrator plant:

x(k + 1) = Fx(k) + gu(k);
y(k) = cx(k)
1 T T 2 /2
F = ; g = ; c = [1 0]
0 1 T
1-z2
Regulator; z = 0.5; wn = 4, z1, 2 = e -zw nT e jw nT = 0.77 j 0.278
Characteristic polynomial; z2 1.54z + 0.67
T2 T2
|zI (F gk)| = z2 + Tk2 + k1 - 2 z + k1 - Tk2 + 1
2 2
k = [13 3.95]
State observer:
Result follows from Review Example 15.3:
2
m =
10
1 0.0952 0.00484
15.14 x(k + 1) = x (k ) + u(k ); y(k ) = [1 0] x(k )
0 0.905 0.0952
k = [105.1 14.625]
Observer design:
z(k + 1) = FTz(k) + cTh(k); h(k) = mTz(k)
mT = [1.9 8.6]
.
15.15 (a) y + y = u; T = 1 sec; A = 1, b = 1;
T
eAT = 0.368; g = -t
e dt = 0.632
0
y(k + 1) = 0.368 y(k) + 0.632u(k)

-T
(b) Time constant = 0.5 sec; z = e t = 0.135
Desired characteristic equation: z 0.135 = 0

F gk = 0.368 0.632K
|z (F gk)| = z (0.368 0.632K) = 0

Comparison with desired characteristic equation gives,
K = 0.3687
(c) With u(k) = 0.3687 y(k) + r, we have

y(k + 1) = 0.135 y(k) + r
At steady-state,
1
ys = 0.135 ys + r; ys = r
0.865
(d) Integral control:
z = 0.5, wn = 4
1-z 2
rejq = e -zw nT e jw nT = - 0.128 j 0.043
(z + 0.128 + j0.043)(z + 0.128 j0.043) = z2 + 0.256z + 0.018
Desired characteristic polynomial: z2 + 0.256z + 0.018

= z2 + a1 z + a2
x2(k) = x2(k 1) + r x1(k)
x1 (k + 1) 0.368 0 x1 (k ) 0.632 0
x (k + 1) = x (k ) + u) k ) + r
2 -0.368 1 2 -0.632 1
0.368 0 0.632
F= ;g=
-0.368 1 -0.632
zI - F = z2 1.368z + 0.368 = z2 + a1z + a2
0.632 0.2326
U = [g F g ] =
-0.632 -0.8646
p1 = [0 1] U1 = [1.58 1.58]
p -1.58 -1.58
P = 1 =
p1 F 0 -1.58
k = [K1 K2] = [a2 a2 a1 a1]P=[0.553 2.013]
x1s x1s 0
(e) x = [F - gk ] x + r
2s 2 s 1
Solving the above set, we get
x1s = ys = r
0 1 0 2 0
15.16 A = ; b = ; u = - kx; Q = ;R=2
0 0 1 0 2
{A, b} pair is controllable, and Q is positive definite.

Matrix Ricatti equation is: ATP + PA PbR1 bTP + Q = 0
Solving for P, we get
2 3 2
P= ; k = R -1 bP = 1 3
2 2 3
0 1 0 2 0
15.17 A = ;b= ;Q= ;R=2
0 -1 1 0 2
Solving Ricatti equation, ATP + PA PbR1 bTP + Q = 0, we get
4 2
; k = R b P = [1 1]
-1 T
P=
2 2
15.18 x(k + 1) = 0.368x(k) + 0.632u(k)

J= [ x 2 (k ) + u2 (k )]; F = 0.368, g = 0.632, Q = 2, R = 2
k =0
From matrix Ricatti equation,
P = Q + FTPF FTPg(R + gTPg)1gTPF, we get
0.3994P2 + 0.9304P 4 = 0
P = 2.207; K = (R + gTPg)1gTPF = 0.178;
u(k) = 0.178x(k)
16
Nonlinear Systems Analysis
16.2 Revisiting Example 16.2 (Fig. 16.16) will be helpful.

4 1 -p E
G(jw) = 2
;- =
jw (1 + jw ) N ( E ) 4M
90 2 tan1w1 = 180
This gives w1 = 1 rad/sec
1
- = |G(jw1)| = 2
N (E1 )
This gives E1 = 8M/p
8M
y(t) = e(t) = - sin t
p
5 2
G(jw) = ; N(E) = 4 0.1
1-
jw (1 + j 0.1w )2 pE E
90 2 tan1 0.1w1 = 180
This gives w1 = 10 rad/sec.

For a stable limit cycle, 2 D < E < ; D = 0.1
1
- = |G(jw1)| = 0.25
N (E1 )
This gives E1 = 0.3

Amplitude of limit cycle is 0.3 and frequency is 10 rad/sec.

For a limit cycle to exist, 2 D < E < ; D = 0.2
G(jw1) = 180 for w1 = 10.95. |G(w1)| = 0.206
1
| G ( jw 1 ) | - for any value of E
N (E )
The intersection of the two curves does not occur; therefore no limit
cycle exists. Intersection of the two curves occurs for
pD
< 0.206 or D < 0.131
2
K
16.5 G(jw) =
jw (1 + jw )2
2 p
2
1 1 1
N(E) = - sin -1 - 1-
p 2 E E E

2 -1 1 1 1
= 1-
sin + 1 - 2 = 1 Nc(E)
p E E E
The function Nc(E) is listed in Table 16.3
For any K > 0, one of the following can happen.
-1
(i) G(jw)-plot does not intersect locus
N (E )
-1
(ii G(jw)-plot intersects the locus; but the point of
N (E )
intersection corresponds to an unstable limit cycle.
16.6 Revisiting Review Example 16.2 (Fig. 16.40) will be helpful.
10
G(jw) =
jw (1 + jw )(1 + j 0.5w )
2 -1 1 1 1
N(E) = sin + 1- = Nc(E)
p E E E2
The function Nc(E) is listed in Table 16.3
G(jw1) = 180 gives w1 = 2
-1
|G(jw1)| = gives E1 = 4.25
N (E1 )
The limit cycle with amplitude 4.25 and frequency 2 rad/sec is a
stable limit cycle.
Nonlinear Systems Analysis 137

K
G(jw) =
jw (1 + jw )(1 + j 0.5w )
-1
For K = 1, G(jw)-plot does not intersect locus. For K = 2,
N (E )
two intersection points are found. One corresponds to unstable limit
cycle and the other corresponds to a stable limit cycle of amplitude
3.75 and frequency 1 rad/sec.
16.8 (a) Characteristic equation has two real, distinct roots in the left
half of s-plane. The origin in the (y, y ) plane is a stable node.
(b) d 2 ( y - 1) d ( y - 1)
2
+5 + 6( y - 1) = 0
dt dt
The singular point is located at (1, 0) in the (y, y ) plane. The
characteristic equation is
s2 + 5s + 6 = 0 = (s + 3) (s + 2)
The singular point is a stable node.
(c) d 2 ( y - 2) d ( y - 2)
2
-8 + 17( y - 2) = 0
dt dt
The singular point is located at (2, 0) in the (y, y ) plane. The
characteristic equation is s2 8s + 17 = 0 with complex roots in
the right half s-plane. The singular point is an unstable focus.
16.9 x1 = q, x2 = q
x2
x = - g = f1 (x)

sin x1 f2 (x)
l
f1 f1
x 0 1
x2 ; x = 0, p
A= 1 = -g
10
f2 f2 cos x10 0
l
x1 x2
For q0 = x10 = 0,
0 1
A = - g ; singular point is Centre
0

For q0 = x10 = p,
0 1
A = g ; singular point is Saddle
0

d 2 ( y - 1) d ( y - 1)
16.10 2
+ 2z + y-1 =0
dt dt
(i) z = 0; singularity (1, 0) on (y, y ) plane is a center
(ii) z = 0.15; singularity (1, 0) on (y, y ) plane is a stable focus.

Jq + T sgn (q) = K K e; e = q q
c A 1 R
Therefore
K T

e+ e + c sgn(e) = 0; K = KAK1
J J
T

e + w n2 e + c sgn(e) = 0; wn = K /J
K
With x1 = e and x2 = e/wn, we have
x1 = wnx2
Tc
x2 = -w n [ x1 + sgn(w n x2 )]
K
For Tc = 0, we have
x1 = wnx2; x 2 = wnx1
dx2 x
= - 1 ; this gives x12 + x22 = x12 (0)
dx1 x2
The effect of the Coulomb friction is to shift the centre of the circles
in phase plane to +Tc/K for x2 < 0 and to Tc/K for x2 > 0. The
steady-state error found from phase trajectory is 0.2 rad. From the
phase trajectory, we see that
(a) the system is obviously stable for all initial conditions and
inputs; and
(b) maximum steady-state error = 0.3 rad.
16.17 Revisiting Review Example 16.3 will be helpful.

With x1 = y and x2 = y , we get
x1 = x2
- x2 ; Region I (-1 < x1 < 1)

x 2 = - x2 - 2( x1 - 1) ; Region II ( x1 > 1)
- x - 2( x + 1) ; Region III ( x1 < - 1)
2 1
From the phase portrait it is seen that the system is stable in the
equilibrium zone defined by
1 x1 1; x1 = 0
With zero deadzone, we get

x1 = x2; x2 = x2 2x1
The isocline equation becomes

-2 x1
x2 =
m+1
A typical phase trajectory of such a system oscillates about the
origin with decreasing amplitude. Deadzone helps to reduce system
oscillations; but causes steady-state error.
16.18 The system equation in the linear range is
e + e + e = 0
The characteristic equation
s2 + s + 1 = 0
has complex roots in left half of s-plane. Therefore, on the (e, e )
plane, the origin is a stable focus.
The system equation in saturation region is
e + e = 0.4 sgn (e)
With e = x1 and e = x2, the isocline equation in the linear range

( 0.4 < e < 0.4) becomes
- x1
x2 =
m+1
For e < 0.4, the isocline equation is
0.4
x2 =
m+1
For e > 0.4, the isocline equation is
- 0.4
x2 =
m+1
The trajectories are asymptotic to the ordinates 0.4 and +0.4

depending on whether the error is +ve or ve.
From the phase trajectories (shown below) plotted from the given
initial conditions (with and without saturation), we find that velocity
is always greater for trajectories without saturation; thus the error
saturation has a slowing down effect on the transient.
e
0.4
0 1.0 1.6 m =
e
m = 3 m=1
m = 2 0.4 m=0
With saturation
m = 1 Without saturation
3 1
0 1 T 2 2
16.19 A = ; A P + PA = I gives P =
-1 -1 1 1
2
P is positive definite matrix. Therefore equilibrium state is

asymptotically stable.
0 1 T
16.20 A = ; A P + PA = I
-1 1
-3 1
2
Solving for P, we get: P = 2
1 -1
2
P is not a positive definite matrix; the system is unstable.
16.21 x1 = x2; x2 = (1 |x1|) x2 x1

Take V(x) = x21 + x22
Then V(x) = 2x1 x1 + 2x2 x2 = 2x22 (1 |x1|)
For V(x) to be negative definite, (1 |x1|) > 0 for x2 0 or |x1| < 1
V(x) is not identically zero anywhere other than the origin. Therefore,
for |x1| < 1, origin is the equilibrium state and is asymptotically
stable.
16.22 x1 = x2; x2 = x1 x12 x2
V(x) = 2x12 x22 < 0
Therefore origin is the equilibrium state and is asymptotically

stable
16.23 V(x1, x2) = ( x12 + x22 )
V = 2 ( x12 + x22 )( x12 + x22 - 1)

Thus V is locally negative definite in the region defined by x12 + x22 < 1;
origin is locally asymptotically stable.
16.24 x1 = x, x2 = x
x1 = x2
x2 = K1x2 K2x32 x1
Origin is the equilibrium point.

V = x21 + x22; V(x) as ||x||
V = 2(K1 x22 + K2 x42)

(i) Globally asymptotically stable
(ii) Unstable
(iii) Asymptotically stable if 0 < x22 < K1/K2

CS Soluiton Manual

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CS Soluiton Manual

Uploaded by

Copyright:

Available Formats

Solution Manual

Tata McGraw Hill Education Private Limited

This Solution Manual, offered as a supplement to the textbook: Gopal,

(b) (s2 + 7s + 10) Y(s) sy(0) y(0) 7y(0)

= [1 1.193e17.25t sin (26.5t + 56.94)] m(t)

The final value theorem is applicable: the function sY(s) does

(ii) The following result follows from Review Examples 2.1.

2.6 (a) g(t) = et et sin t

s +1 1/2 1/4 1/4

2.7 Revisiting Review Example 2.1 will be helpful.

2.11 The transfer function is

G(0) = 1, G() = 0; the circuit is a second-order lowpass filter.

Substituting the values of R, L, and C, we obtain

16w4b 7w2b 1 = 0; wb = 742 rad/sec

Gravitational effect has been eliminated by appropriate choice of

|G(jw)|w = 11.63 = 0.1615

x(peak) = 7.5 0.1615 = 1.2113 cm

Removing loops that touch forward path (G1G2G3) yields D1 = 1.

2.20 There are three forward paths:

G1G2G3G4G5 + G1G6G4G5 + G1G2G7 (1 + G4 H1 )

For Er(s) = 100/s,

w(t) = 481.8 (1 e10.375t)

3.4 TM = K1ec + K 2qM = JeqqM + Beq qM

K2 = slope of the characteristic lines shown in Fig. P3.4b

3.5 (i) Jeq = JM + JL L = 0.6

(ii) A sketch of multi-loop configuration may be made on the lines

(b) Modification in the drive system may be made on the lines of

TM = J M qM + B(qM - qL ) + K (q M - q L )

B (qM - qL ) + K (q M - q L ) = J LqL + Tw

(JMs2 + Bs + K) qM(s) (Bs + K) qL(s) = TM(s)

(Bs + K) qM(s) (JLs2 + Bs + K) qL(s) = Tw(s)

Let z(t) = displacement of the power piston

These equations give us

4.3 For G(s) = 20/(s + 1), and R(s) = 1/s,

y(t ) open-loop = 20(1 et)

y(t ) open-loop = 50(1 e0.4t)

For W(s) = 1/s, yss = 0

For W(s) = 1/s, yss = 1

4.7 (a) Figure 4.16:

Time-constant has reduced from the value of 10 to 10/101 by

4.8 (a) er = 50 0.03 = 1.5 volts

Suppose that the engine stalls when constant percent gradient

(b) When the feedback loop is opened,

wo(t) = 9.61 (1 e5t)

Wind gust torque on the load shaft = 100 newton-m

Therefore, on the motor shaft

Time-constant (without current feedback) = Lf /(Rf + 1)

4.14 Fig. P4.14a:

This gives K2 = 25 and K1 = 4

Only one sign change (the s0-row has ve pivot element) is

= (s + j3) (s j3) (s + 1 + j1) (s + 1 j1)

(b) s3-row of the Routh array is an all-zero row. The auxiliary

D(s)/A(s) = (s5 + 2s4 + 24s3 + 48s2 25s 50)/A(s)

D(s) = (s + 2) (s4 + 24s2 25)

= (s + 2) (s2 1) (s2 + 25)

D(s)/A(s) = (s6 + 3s5 + 5s4 + 9s3 + 8s2 + 6s + 4)/A(s)

D(s) = (s2 + 3s + 2) (s4 + 3s2 + 2)

= (s2 + 3s + 2) (s2 + 1)(s2 + 2)

5.3 With s = s 1, the characteristic equation becomes

s1 -row in the routh array is an all-zero row. The auxiliary polynomial

5.4 (a) Unstable for all values of K

5.5 With s = s 1, the characteristic equation becomes

8s3 + 46s2 + 31s + 5 + 4 K = 0