Solutions To Skill-Assessment Exercises

SOLUTIONS TO
SKILL-ASSESSMENT EXERCISES
CHAPTER 2
2.1
The Laplace transform of t is
1
s
2
using Table 2.1, Item 3. Using Table 2.2, Item 4,
Fs
1
s 5
2
.
2.2
Expanding Fs by partial fractions yields:
Fs
A
s

B
s 2
C
s 3
2
D
s 3
where,
A
10
s 2s 3
2
S!0
5
9
B
10
ss 3
2
S!2
5
C
10
ss 2
S!3
10
3
; and D s 3
2
dFs
ds
s!3
40
9
Taking the inverse Laplace transform yields,

f t
5
9
5e
2t
10
3
te
3t
40
9
e
3t
2.3
Taking the Laplace transform of the differential equation assuming zero initial con-
ditions yields:
s
3
Cs 3s
2
Cs 7sCs 5Cs s
2
Rs 4sRs 3Rs
1
Collecting terms,
s
3
3s
2
7s 5Cs s
2
4s 3Rs
Thus,
Cs
Rs

s
2
4s 3
s
3
3s
2
7s 5
2.4
Gs
Cs
Rs

2s 1
s
2
6s 2
Cross multiplying yields,
d
2
c
dt
2
6
dc
dt
2c 2
dr
dt
r
2.5
Cs RsGs
1
s
2
s
s 4s 8
1
ss 4s 8
A
s

B
s 4
C
s 8
where
A
1
s 4s 8 S!0
1
32
B
1
ss 8
S!4

1
16
; and
C
1
ss 4
S!8
1
32
Thus,
ct
1
32
1
16
e
4t
1
32
e
8t
2.6
Mesh Analysis
Transforming the network yields,
+
V(s)
I
1
(s) I
2
(s)
V
1
(s)
V
2
(s)
I
9
(s)
+
1
1
s s
s
_
_
2 Solutions to Skill-Assessment Exercises
Now, writing the mesh equations,
s 1I
1
s sI
2
s I
3
s Vs
sI
1
s 2s 1I
2
s I
3
s 0
I
1
s I
2
s s 2I
3
s 0
Solving the mesh equations for I
2
(s),
I
2
s
s 1 Vs 1
s 0 1
1 0 s 2
s 1 s 1
s 2s 1 1
1 1 s 2
s
2
2s 1Vs
ss
2
5s 2
But, V
L
s sI
2
s
Hence,
V
L
s
s
2
2s 1Vs
s
2
5s 2
or
V
L
s
Vs

s
2
2s 1
s
2
5s 2
Nodal Analysis
Writing the nodal equations,
1
s
2
_ _
V
1
s V
L
s Vs
V
1
s
2
s
1
_ _
V
L
s
1
s
Vs
Solving for V
L
(s),
V
L
s
1
s
2
_ _
Vs
1
1
s
Vs
1
s
2
_ _
1
1
2
s
1
_ _
s
2
2s 1Vs
s
2
5s 2
or
V
L
s
Vs

s
2
2s 1
s
2
5s 2
2.7
Inverting
Gs
Z
2
s
Z
1
s
100000
10
5
=s
s
Chapter 2 Solutions to Skill-Assessment Exercises 3
Noninverting
Gs
Z
1
s Z
2
s
Z
1
s

10
5
s
10
5
_ _
10
5
s
_ _ s 1
2.8
Writing the equations of motion,
s
2
3s 1X
1
s 3s 1X
2
s Fs
3s 1X
1
s s
2
4s 1X
2
s 0
Solving for X
2
(s),
X
2
s
s
2
3s 1 Fs
3s 1 0
s
2
3s 1 3s 1
3s 1 s
2
4s 1
3s 1Fs
ss
3
7s
2
5s 1
Hence,
X
2
s
Fs

3s 1
ss
3
7s
2
5s 1
2.9
s
2
s 1u
1
s s 1u
2
s Ts
s 1u
1
s 2s 2u
2
s 0
where u
1
s is the angular displacement of the inertia.
Solving for u
2
s,
u
2
s
s
2
s 1 Ts
s 1 0
s
2
s 1 s 1
s 1 2s 2
s 1Fs
2s
3
3s
2
2s 1
From which, after simplication,
u
2
s
1
2s
2
s 1
2.10
Transforming the network to one without gears by reecting the 4 N-m/rad spring to the
left and multiplying by (25/50)
2
, we obtain,
1 kg
1 N-m-s/rad
1 N-m/rad
T(t)
q
a
(t)
q
1
(t)
s
2
su
1
s su
a
s Ts
su
1
s s 1u
a
s 0
where u
1
s is the angular displacement of the 1-kg inertia.
Solving for u
a
s,
u
a
s
s
2
s Ts
s 0
s
2
s s
s s 1
sTs
s
3
s
2
s
From which,
u
a
s
Ts

1
s
2
s 1
But, u
2
s
1
2
u
a
s:
Thus,
u
2
s
Ts

1=2
s
2
s 1
2.11
First nd the mechanical constants.
J
m
J
a
J
L
1
5
1
4
_ _
2
1 400
1
400
_ _
2
D
m
D
a
D
L
1
5
1
4
_ _
2
5 800
1
400
_ _
7
Now nd the electrical constants. From the torque-speed equation, set v
m
0 to nd
stall torque and set T
m
0 to nd no-load speed. Hence,
T
stall
200
v
noload
25
which,
K
t
R
a
T
stall
E
a
200
100
2
K
b

E
a
v
noload
100
25
4
Substituting all values into the motor transfer function,
u
m
s
E
a
s
K
T
R
a
J
m
s s
1
J
m
_ _
D
m
K
T
K
b
R
a
_ _
1
s s
15
2
_ _
where u
m
s is the angular displacement of the armature.
Now u
L
s
1
20
u
m
s. Thus,
u
L
s
E
a
s
1=20
s s
15
2
_ _
2.12
Letting
u
1
s v
1
s=s
u
2
s v
2
s=s
in Eqs. 2.127, we obtain
J
1
s D
1
K
s
_ _
v
1
s
K
s
v
2
s Ts
K
s
v
1
s J
2
s D
2
K
s
_ _
v
2
s
Fromthese equations we can drawboth series and parallel analogs by considering these
to be mesh or nodal equations, respectively.
+
J
1
J
1
J
2
J
2
D
1
D
1
D
2
D
2
T(t)
T(t)
w
1
(t)
w
1
(t)
w
2
(t)
w
2
(t)
1 1
K
1
1
Series analog
Parallel analog
K
2.13
Writing the nodal equation,
C
dv
dt
i
r
2 it
But,
C 1
v v
o
dv
i
r
e
v
r
e
v
e
v
o
dv
Substituting these relationships into the differential equation,
dv
o
dv
dt
e
v
o
dv
2 it
(1)
We now linearize e
v
.
The general form is
f v f v
o

d f
dv
v
o
dv
Substituting the function, f v e
v
, with v v
o
dv yields,
e
v
o
dv
e
v
o
de
v
dv
v
o
dv
Solving for e
v
o
dv
,
e
v
o
dv
e
v
o
de
v
dv
v
o
dv e
v
o
e
v
o
dv
Substituting into Eq. (1)
ddv
dt
e
v
o
e
v
o
dv 2 it (2)
Setting it 0 and letting the circuit reach steady state, the capacitor acts like an open
circuit. Thus, v
o
v
r
with i
r
2. But, i
r
e
v
r
or v
r
ln i
r
.
Hence, v
o
ln 2 0:693. Substituting this value of v
o
into Eq. (2) yields
ddv
dt
2dv it
Taking the Laplace transform,
s 2dvs Is
Solving for the transfer function, we obtain
dvs
Is

1
s 2
or
Vs
Is

1
s 2
about equilibrium.
CHAPTER 3
3.1
Identifying appropriate variables on the circuit yields
+
C
1
i
L
i
C
2
i
C
1
i
R
C
2
R
L
v
o
(t) v
1
(t)
Writing the derivative relations
C
1
dv
C
1
dt
i
C
1
L
di
L
dt
v
L
C
2
dv
C
2
dt
i
C
2
(1)
Using Kirchhoffs current and voltage laws,
i
C
1
i
L
i
R
i
L
1
R
v
L
v
C
2
v
L
v
C
1
v
i
i
C
2
i
R

1
R
v
L
v
C
2
Substituting these relationships into Eqs. (1) and simplifying yields the state equations as
dv
C
1
dt

1
RC
1
v
C
1

1
C
1
i
L
1
RC
1
v
C
2

1
RC
1
v
i
di
L
dt

1
L
v
C
1

1
L
v
i
dv
C
2
dt

1
RC
2
v
C
1

1
RC
2
v
C
2
1
RC
2
v
i
where the output equation is
v
o
v
C
2
Putting the equations in vector-matrix form,
_ x
1
RC
1
1
C
1
1
RC
1
1
L
0 0
1
RC
2
0
1
RC
2
_
_
_
_
x
1
RC
1
1
L
1
RC
2
_
_
_
_
v
i
t
y 0 0 1 x
3.2
Writing the equations of motion
s
2
s 1X
1
s sX
2
s Fs
sX
1
s s
2
s 1X
2
s X
3
s 0
X
2
s s
2
s 1X
3
s 0
Taking the inverse Laplace transform and simplifying,
x
1
_ x
1
x
1
_ x
2
f
x
2
_ x
1
_ x
2
x
2
x
3
x
3
_ x
3
x
3
x
2
Dening state variables, z
i
,
z
1
x
1
; z
2
_ x
1
; z
3
x
2
; z
4
_ x
2
; z
5
x
3
; z
6
_ x
3
Writing the state equations using the denition of the state variables and the inverse
transform of the differential equation,
_ z
1
z
2
_ z
2
x
1
_ x
1
x
1
_ x
2
f z
2
z
1
z
4
f
_ z
3
_ x
2
z
4
_ z
4
x
2
_ x
1
_ x
2
x
2
x
3
z
2
z
4
z
3
z
5
_ z
5
_ x
3
z
6
_ z
6
x
3
_ x
3
x
3
x
2
z
6
z
5
z
3
The output is z
5
. Hence, y z
5
. In vector-matrix form,
_ z
0 1 0 0 0 0
1 1 0 1 0 0
0 0 0 1 0 0
0 1 1 1 1 0
0 0 0 0 0 1
0 0 1 0 1 1
_
_
_
_
z
0
1
0
0
0
0
_
_
_
_
f t; y 0 0 0 0 1 0 z
3.3
First derive the state equations for the transfer function without zeros.
Xs
Rs
1
s
2
7s 9
Cross multiplying yields
s
2
7s 9Xs Rs
Taking the inverse Laplace transform assuming zero initial conditions, we get
x 7_ x 9x r
Dening the state variables as,
x
1
x
x
2
_ x
Hence,
_ x
1
x
2
_ x
2
x 7_ x 9x r 9x
1
7x
2
r
Using the zeros of the transfer function, we nd the output equation to be,
c 2_ x x x
1
2x
2
Putting all equation in vector-matrix form yields,
_ x
0 1
9 7
_ _
x
0
1
_ _
r
c 1 2 x
3.4
The state equation is converted to a transfer function using
Gs CsI A
1
B
1
where
A
4 1:5
4 0
_ _
; B
2
0
_ _
; and C 1:5 0:625 :
Evaluating sI A yields
sI A
s 4 1:5
4 s
_ _
Taking the inverse we obtain
sI A
1
1
s
2
4s 6
s 1:5
4 s 4
_ _
Substituting all expressions into Eq. (1) yields
Gs
3s 5
s
2
4s 6
3.5
Writing the differential equation we obtain
d
2
x
dt
2
2x
2
10 df t
(1)
Letting x x
o
dx and substituting into Eq. (1) yields
d
2
x
o
dx
dt
2
2x
o
dx
2
10 df t
(2)
Now, linearize x
2
.
x
o
dx
2
x
2
o

dx
2
dx
x
o
dx 2x
o
dx
from which
x
o
dx
2
x
2
o
2x
o
dx (3)
Substituting Eq. (3) into Eq. (1) and performing the indicated differentiation gives us the
linearized intermediate differential equation,
d
2
dx
dt
2
4x
o
dx 2x
2
o
10 df t
(4)
The force of the spring at equilibrium is 10 N. Thus, since F 2x
2
; 10 2x
2
o
from
which
x
o

5
p
Substituting this value of x
o
into Eq. (4) gives us the nal linearized differential
equation.
d
2
dx
dt
2
4
5
p
dx df t
Selecting the state variables,
x
1
dx
x
2

_
d x
Writing the state and output equations
_ x
1
x
2
_ x
2

d x 4
5
p
x
1
df t
y x
1
Converting to vector-matrix form yields the nal result as
_ x
0 1
4
5
p
0
_ _
x
0
1
_ _
df t
y 1 0 x
CHAPTER 4
4.1
For a step input
Cs
10s 4s 6
ss 1s 7s 8s 10
A
s

B
s 1
C
s 7
D
s 8
E
s 10
Taking the inverse Laplace transform,
ct A Be
t
Ce
7t
De
8t
Ee
10t
4.2
Since a 50; T
c

1
a
1
50
0:02 s; T
s

4
a
4
50
0:08 s; and T
r

2:2
a

2:2
50

0:044 s.
4.3
a. Since poles are at 6 j19:08; ct A Be
6t
cos19:08t f.
b. Since poles are at 78:54 and 11:46; ct A Be
78:54t
Ce
11:4t
.
c. Since poles are double on the real axis at 15 ct A Be
15t
Cte
15t
:
d. Since poles are at j25; ct A Bcos25t f.
4.4
a.
v
n

400
p
20 and 2zv
n
12; ;z 0:3 and systemis underdamped
.
b.
v
n

900
p
30 and 2zv
n
90; ;z 1:5 and systemis overdamped
.
c.
v
n

225
p
15 and 2zv
n
30; ;z 1 and systemis critically damped
.
d.
v
n

625
p
25 and 2zv
n
0; ;z 0 and systemis undamped
.
4.5
v
n

361
p
19 and 2zv
n
16; ;z 0:421:
Now, T
s

4
zv
n
0:5 s and T
p

p
v
n
1 z
2
_ 0:182 s.
From Figure 4.16, v
n
T
r
1:4998. Therefore, T
r
0:079 s.
Finally, %os e
zp
1
p
z
2
100 23:3%
4.6
a. The second-order approximation is valid, since the dominant poles have a real part of
2 and the higher-order pole is at 15, i.e. more than ve-times further.
b. The second-order approximation is not valid, since the dominant poles have a real
part of 1 and the higher-order pole is at 4, i.e. not more than ve-times further.
4.7
a. ExpandingG(s) by partial fractions yields Gs
1
s
0:8942
s 20
1:5918
s 10
0:3023
s 6:5
.
But 0:3023 is not an order of magnitude less than residues of second-order
terms (term 2 and 3). Therefore, a second-order approximation is not valid.
b. Expanding G(s) by partial fractions yields Gs
1
s
0:9782
s 20
1:9078
s 10
0:0704
s 6:5
.
But 0.0704 is an order of magnitude less than residues of second-order terms
(term 2 and 3). Therefore, a second-order approximation is valid.
4.8
See Figure 4.31 in the textbook for the Simulink block diagram and the output
responses.
4.9
a. Since sI A
s 2
3 s 5
_ _
; sI A
1
1
s
2
5s 6
s 5 2
3 s
_ _
: Also,
BUs
0
1=s 1
_ _
.
The state vector is Xs sI A
1
x0 BUs
1
s 1s 2s 3

2s
2
7s 7
s
2
4s 6
_ _
. The output is Ys 1 3 Xs
5s
2
2s 4
s 1s 2s 3
0:5
s 1

12
s 2
17:5
s 3
. Taking the inverse Laplace transform yields yt
0:5e
t
12e
2t
17:5e
3t
.
b. The eigenvalues are given by the roots of jsI Aj s
2
5s 6, or 2 and 3.
4.10
a. Since sI A
s 2
2 s 5
_ _
; sI A
1
1
s
2
5s 4
s 5 2
2 s
_ _
. Taking the
Laplace transform of each term, the state transition matrix is given by
Ft
4
3
e
t
1
3
e
4t
2
3
e
t
2
3
e
4t
2
3
e
t
2
3
e
4t
1
3
e
t
4
3
e
4t
_
_
_
_
:
b. Since Ft t
4
3
e
tt
1
3
e
4tt
2
3
e
tt
2
3
e
4tt
2
3
e
tt
2
3
e
4tt

1
3
e
tt
4
3
e
4tt
_
_
_
_
and
But
0
e
2t
_ _
; Ft tBut
2
3
e
t
e
t
2
3
e
2t
e
4t
1
3
e
t
e
t
4
3
e
2t
e
4t
_
_
_
_
:
Thus, xt Ftx0
_
t
0
Ft t
BUtdt
10
3
e
t
e
2t
4
3
e
4t
5
3
e
t
e
2t
8
3
e
4t
_
_
_
_
:
c. yt 2 1 x 5e
t
e
2t
CHAPTER 5
5.1
Combine the parallel blocks in the forward path. Then, push
1
s
to the left past the pickoff
point.
1
s
s
s
s
s
2
+
1
+

R(s)
C(s)
Combine the parallel feedback paths and get 2s. Then, apply the feedback formula,
simplify, and get, Ts
s
3
1
2s
4
s
2
2s
.
5.2
Find the closed-loop transfer function, Ts
Gs
1 GsHs
16
s
2
as 16
, where
and Gs
16
ss a
and Hs 1. Thus, v
n
4 and 2zv
n
a, from which z
a
8
.
But, for 5% overshoot, z
ln
%
100
_ _
p
2
ln
2
%
100
_ _
0:69. Since, z
a
8
; a 5:52.
5.3
Label nodes.
+ +
+
R(s)
s
s
N
1
(s) N
2
(s) N
3
(s) N
4
(s)
N
6
(s) N
5
(s)
N
7
(s)
s
C(s)
1
s
1
s
Draw nodes.
R(s) N
1
(s) N
2
(s)
N
5
(s)
N
7
(s)
N
6
(s)
N
3
(s) N
4
(s) C (s)
Connect nodes and label subsystems.
R(s)
C(s) 1
1
s
s
1
s s
1 1
1
1
1
s
N
1
(s) N
2
(s)
N
5
(s) N
6
(s)
N
7
(s)
N
3
(s)
N
4
(s)
Eliminate unnecessary nodes.
R(s) C(s) 1 s s
1
s
1
s
s
1
5.4
Forward-path gains are G
1
G
2
G
3
and G
1
G
3
.
Loop gains are G
1
G
2
H
1
; G
2
H
2
; and G
3
H
3
.
Nontouching loops are G
1
G
2
H
1
G
3
H
3
G
1
G
2
G
3
H
1
H
3
and
G
2
H
2
G
3
H
3
G
2
G
3
H
2
H
3
.
Also, D 1 G
1
G
2
H
1
G
2
H
2
G
3
H
3
G
1
G
2
G
3
H
1
H
3
G
2
G
3
H
2
H
3
:
Finally, D
1
1 and D
2
1.
Substituting these values into Ts
Cs
Rs

k
T
k
D
k
D
yields
Ts
G
1
sG
3
s1 G
2
s
1 G
2
sH
2
s G
1
sG
2
sH
1
s1 G
3
sH
3
s
5.5
The state equations are,
_ x
1
2x
1
x
2
_ x
2
3x
2
x
3
_ x
3
3x
1
4x
2
5x
3
r
y x
2
Drawing the signal-ow diagram from the state equations yields
1
s
1
s
1
s
1 1 1
1
5
4
2 3
3
r x
1
x
2
x
3
y
5.6
From Gs
100s 5
s
2
5s 6
we draw the signal-ow graph in controller canonical form
and add the feedback.
1
5
6
100
500
1
y
r
1
s
1
s
x
1
x
2
Writing the state equations from the signal-ow diagram, we obtain
x
105 506
1 0
_ _
x
1
0
_ _
r
y 100 500 x
5.7
From the transformation equations,
P
1
3 2
1 4
_ _
Taking the inverse,
P
0:4 0:2
0:1 0:3
_ _
Now,
P
1
AP
3 2
1 4
_ _
1 3
4 6
_ _
0:4 0:2
0:1 0:3
_ _
6:5 8:5
9:5 11:5
_ _
P
1
B
3 2
1 4
_ _
1
3
_ _
3
11
_ _
CP 1 4
0:4 0:2
0:1 0:3
_ _
0:8 1:4
Therefore,
_ z
6:5 8:5
9:5 11:5
_ _
z
3
11
_ _
u
y 0:8 1:4 z
5.8
First nd the eigenvalues.
jlI Aj
l 0
0 l
_ _
1 3
4 6
_ _

l 1 3
4 l 6
l
2
5l 6
From which the eigenvalues are 2 and 3.

Now use Ax
i
lx
i
for each eigenvalue, l.
Thus,
1 3
4 6
_ _
x
1
x
2
_ _
l
x
1
x
2
_ _
For l 2,
3x
1
3x
2
0
4x
1
4x
2
0
Thus x
1
x
2
For l 3
4x
1
3x
2
0
4x
1
3x
2
0
Thus x
1
x
2
and x
1
0:75x
2
; from which we let
P
0:707 0:6
0:707 0:8
_ _
Taking the inverse yields
P
1
5:6577 4:2433
5 5
_ _
Hence,
D P
1
AP
5:6577 4:2433
5 5
_ _
1 3
4 6
_ _
0:707 0:6
0:707 0:8
_ _
2 0
0 3
_ _
P
1
B
5:6577 4:2433
5 5
_ _
1
3
_ _
18:39
20
_ _
CP 1 4
0:707 0:6
0:707 0:8
_ _
2:121 2:6
Finally,
_ z
2 0
0 3
_ _
z
18:39
20
_ _
u
y 2:121 2:6 z
CHAPTER 6
6.1
Make a Routh table.
Since there are four sign changes and no complete rowof zeros, there are four right half-
plane poles and three left half-plane poles.
6.2
Make a Routh table. We encounter a rowof zeros on the s
3
row. The even polynomial is
contained in the previous row as 6s
4
0s
2
6. Taking the derivative yields
s
7
3 6 7 2
s
6
9 4 8 6
s
5
4.666666667 4.333333333 0 0
s
4
4.35714286 8 6 0
s
3
12.90163934 6.426229508 0 0
s
2
10.17026684 6 0 0
s
1
1.18515742 0 0 0
s
0
6 0 0 0
24s
3
0s. Replacing the rowof zeros with the coefcients of the derivative yields the s
3
row. We also encounter a zero in the rst column at the s
2
row. We replace the zero
with e and continue the table. The nal result is shown now as
There is one sign change below the even polynomial. Thus the even polynomial
(4
th
order) has one right half-plane pole, one left half-plane pole, and 2 imaginary
axis poles. From the top of the table down to the even polynomial yields one sign
change. Thus, the rest of the polynomial has one right half-plane root, and one left
half-plane root. The total for the system is two right half-plane poles, two left half-
plane poles, and 2 imaginary poles.
6.3
Since Gs
Ks 20
ss 2s 3
; Ts
Gs
1 Gs
Ks 20
s
3
5s
2
6 Ks 20K
Form the Routh table.
From the s
1
row, K < 2. From the s
0
row, K >0. Thus, for stability, 0 < K < 2.
6.4
First nd
jsI Aj
s 0 0
0 s 0
0 0 s
_
_
_
_
2 1 1
1 7 1
3 4 5
_
_
_
_
s 2 1 1
1 s 7 1
3 4 s 5
s
3
4s
2
33s 51
Now form the Routh table.
There are two sign changes. Thus, there are two rhp poles and one lhp pole.
s
6
1 6 1 6
s
5
1 0 1 0
s
4
6 0 6 0
s
3
24 0 0 0 ROZ
s
2
e 6 0 0
s
1
144=e 0 0 0
s
0
6 0 0 0
s
3
1 6 K
s
2
5 20K
s
1
30 15K
5
s
0
20K
s
3
1 33
s
2
4 51
S
1
20:25
S
0
51
CHAPTER 7
7.1
a. First check stability.
Ts
Gs
1 Gs
10s
2
500s 6000
s
3
70s
2
1375s 6000
10s 30s 20
s 26:03s 37:89s 6:085
Poles are in the lhp. Therefore, the system is stable. Stability also could be checked via
Routh-Hurwitz using the denominator of T(s). Thus,
15ut: e
step
1
15
1 lim
s!0
Gs
15
1 1
0
15tut: e
ramp
1
15
lim
s!0
sGs
15
10
20
30
25
35
2:1875
15t
2
ut: e
parabola
1
15
lim
s!0
s
2
Gs
30
0
1; since L15t
2

30
s
3
b. First check stability.
Ts
Gs
1 Gs
10s
2
500s 6000
s
5
110s
4
3875s
3
4:37e04s
2
500s 6000
10s 30s 20
s 50:01s 35s 25s
2
7:189e 04s 0:1372
From the second-order term in the denominator, we see that the system is unstable.
Instability could also be determined using the Routh-Hurwitz criteria on the denomi-
nator of T(s). Since the system is unstable, calculations about steady-state error
cannot be made.
7.2
a. The system is stable, since
Ts
Gs
1 Gs
1000s 8
s 9s 7 1000s 8
1000s 8
s
2
1016s 8063
and is of Type 0. Therefore,
K
p
lim
s!0
Gs
1000
8
7
9
127; K
v
lim
s!0
sGs 0;
and K
a
lim
s!0
s
2
Gs 0
b.
e
step
1
1
1 lim
s!0
Gs
1
1 127
7:8e 03
e
ramp
1
1
lim
s!0
sGs
1
0
1
e
parabola
1
1
lim
s!0
s
2
Gs
1
0
1
7.3
System is stable for positive K. System is Type 0. Therefore, for a step input
e
step
1
1
1 K
p
0:1. Solving for K
p
yields K
p
9 lim
s!0
Gs
12K
14
18
; from
which we obtain K 189.
7.4
System is stable. Since G
1
s 1000, and G
2
s
s 2
s 4
,
e
D
1
1
lim
s!0
1
G
2
s
lim G
1
s!0
s
1
2 1000
9:98e 04
7.5
System is stable. Create a unity-feedback system, where H
e
s
1
s 1
1
s
s 1
:
The system is as follows:
+
R(s)
E
a
(s) C(s)
100
s + 4
s
s + 1
Thus,
G
e
s
Gs
1 GSH
e
s
100
s 4
1
100s
s 1s 4
100s 1
S
2
95s 4
Hence, the system is Type 0. Evaluating K
p
yields
K
p

100
4
25
The steady-state error is given by
e
step
1
1
1 K
p
1
1 25
3:846e 02
7.6
Since Gs
Ks 7
s
2
2s 10
; e1
1
1 K
p
1
1
7K
10
10
10 7K
:
Calculating the sensitivity, we get
S
e:K

K
e
@e
@K

K
10
10 7K
_ _
107
10 7K
2

7K
10 7K
7.7
Given
A
0 1
3 6
_ _
; B
0
1
_ _
; C 1 1 ; Rs
1
s
:
Using the nal value theorem,
e
step
1 lim
s!0
sRs1 CsI A
1
B lim
s!0
1 1 1
s 1
3 s 6
_ _
1
0
1
_ _
_ _
lim
s!0
1 1 1
s 6 1
3s s
_ _
s
2
6s 3
0
1
_ _
_
_
_
_
lim
s!0
s
2
5s 2
s
2
6s 3
2
3
Using input substitution,
step
1 1 CA
1
B 1 1 1
0 1
3 6
_ _
1
0
1
_ _
1 1 1
6 1
3 0
_ _
3
0
1
_ _
1 1 1

1
3
0
_
_
_
_
2
3
CHAPTER 8
8.1
a.
F7 j9
7 j9 27 j9 40:0339
7 j97 j9 37 j9 6
5 j93 j9
7 j94 j91 j9
66 j72
944 j378
0:0339 j0:0899 0:096 < 110:7

b. The arrangement of vectors is shown as follows:
jw
s
s-plane
X X
2 4 1 3 6 5 7
X
M
1
M
2
M
3
M
4
M
5
(7+j9)
0
From the diagram,
F7 j9
M
2
M
4
M
1
M
3
M
5
3 j95 j9
1 j94 j97 j9
66 j72
944 j378
0:0339 j0:0899 0:096 <; 110:7
8.2
a. First draw the vectors.
jw
s
s-plane
X
X
2
3
0 1
j1
j2
j3
j1
j2
j3
From the diagram,
angles 180
tan
1
3
1
_ _
tan
1
3
1
_ _
180
108:43
108:43
180
:
b. Since the angle is 1808, the point is on the root locus.
c. K
Ppole lengths
Pzero lengths
1
2
3
2
p

1
2
3
2
p

1
10
8.3
First, nd the asymptotes.
s
a

poles zeros
#poles #zeros
2 4 6 0
3 0
4
u
a

2k 1p
3

p
3
; p;
5p
3
Next draw root locus following the rules for sketching.
8 7 6 5 4 3 2 1 0 1 2 3
5
4
3
2
1
0
1
2
3
4
5
Real Axis
I
m
a
g

A
x
i
s
8.4
a.
j3
s
jw
s-plane
X
X
O
2
2
j3
0
b. Using the Routh-Hurwitz criteria, we rst nd the closed-loop transfer function.
Ts
Gs
1 Gs
Ks 2
s
2
K 4s 2K 13
Using the denominator of Ts, make a Routh table.
We get a row of zeros for K 4. From the s
2
row with K 4; s
2
21 0. From
which we evaluate the imaginary axis crossing at
21
p
.
c. From part (b), K 4.
d. Searchingfor the minimumgain tothe left of 2 onthe real axis yields 7 at a gainof
18. Thus the break-in point is at 7.
e. First, draw vectors to a point e close to the complex pole.
jw
s
s-plane
X
X
2
2 0
j3
j3
At the point e close to the complex pole, the angles must add up to zero. Hence, angle
fromzero angle frompole in 4
th
quadrant angle frompole in 1
st
quadrant 180
,
or tan
1
3
4
_ _
90
u 180
. Solving for the angle of departure, u 233:1.

s
2
1 2K 13
s
1
K 40 0
s
0
2K 13 0
8.5
a.
jw
4 3
X
X
s-plane
o
0
z = 0.5
j4
j4
s
2
o
b. Search along the imaginary axis and nd the 1808 point at s j4:06.
c. For the result in part (b), K 1.
d. Searching between 2 and 4 on the real axis for the minimumgain yields the break-in
at s 2:89.
e. Searching along z 0:5 for the 1808 point we nd s 2:42 j4:18.
f. For the result in part (e), K 0:108.
g. Using the result from part (c) and the root locus, K < 1.
8.6
a.
s
jw
z = 0.591
2 4 6
X X X
0
s-plane
b. Searching along the z 0:591 (10% overshoot) line for the 1808 point yields
2:028 j2:768 with K 45:55.
c. T
s

4
jRej
4
2:028
1:97 s; T
p

p
jImj
p
2:768
1:13 s; v
n
T
r
1:8346 fromthe
rise-time chart and graph in Chapter 4. Since v
n
is the radial distance to the pole,
v
n

2:028
2
2:768
2
p
3:431. Thus, T
r
0:53 s; since the system is Type 0,
K
p

K
2
45:55
48
0:949. Thus,
e
step
1
1
1 K
p
0:51:
d. Searching the real axis to the left of 6 for the point whose gain is 45.55, we nd
7:94. Comparing this value to the real part of the dominant pole, 2:028, we nd
that it is not ve times further. The second-order approximation is not valid.
8.7
Find the closed-loop transfer function and put it the form that yields p
i
as the root locus
variable. Thus,
Ts
Gs
1 Gs
100
s
2
p
i
s 100
100
s
2
100 p
i
s
100
s
2
100
1
p
i
s
s
2
100
Hence, KGsHs
p
i
s
s
2
100
. The following shows the root locus.
s
jw
j10 X
O
s-plane
0
X j10
8.8
Following the rules for plotting the root locus of positive-feedback systems, we obtain
the following root locus:
s
jw
2 4
X
s-plane
1
X X o
0 3
8.9
The closed-loop transfer function is Ts
Ks 1
s
2
K 2s K
. Differentiating the
denominator with respect to K yields
2s
@s
@K
K 2
@s
@K
s 1 2s K 2
@s
@K
s 1 0
Solving for
@s
@K
, we get
@s
@K

s 1
2s K 2
. Thus, S
s:K

K
s
@s
@K

Ks 1
s2s K 2
:
Substituting K 20 yields S
s:K

10s 1
ss 11
.
Now nd the closed-loop poles when K 20. From the denominator of Ts; s
1;2

21:05; 0:95, when K 20.
For the pole at 21:05,
Ds sS
s:K
DK
K
21:05
1021:05 1
21:0521:05 11
_ _
0:05 0:9975:
For the pole at 0:95,
Ds sS
s:K
DK
K
0:95
100:95 1
0:950:95 11
_ _
0:05 0:0025:
CHAPTER 9
9.1
a. Searching along the 15%overshoot line, we nd the point on the root locus at 3:5
j5:8 at a gain of K 45:84. Thus, for the uncompensated system,
K
v
lim
s!0
sGs K=7 45:84=7 6:55.
Hence, e
ramp uncompensated
1 1=K
v
0:1527.
b. Compensator zero should be 20x further to the left than the compensator pole.
Arbitrarily select G
c
s
s 0:2
s 0:01
.
c. Insert compensator and search along the 15%overshoot line and nd the root locus at
3:4 j5:63 with a gain, K 44:64. Thus, for the compensated system,
K
v

44:640:2
70:01
127:5 and e
ramp compensated
1
1
K
v
0:0078.
d.
e
ramp uncompensated
e
ramp compensated
0:1527
0:0078
19:58
9.2
j5:8 at a gain of K 45:84. Thus, for the uncompensated system,
T
s

4
jRej
4
3:5
1:143 s:
b. The real part of the design point must be three times larger than the uncompensated
poles real part. Thus the design point is 33:5 j 35:8 10:5 j 17:4. The
angular contribution of the plants poles and compensator zero at the design point
is 130:8
. Thus, the compensator pole must contribute 180
130:8
49:2
.
Using the following diagram,
p
c
s
jw
s-plane
10.5
j17.4
49.2
we nd
17:4
p
c
10:5
tan 49:2
, from which, p
c
25:52. Adding this pole, we nd
the gain at the design point to be K 476:3. A higher- order closed-loop pole is
found to be at 11:54. This pole may not be close enough to the closed-loop zero at
10. Thus, we should simulate the systemto be sure the design requirements have
been met.
9.3
6:83 at a gain of K 58:9. Thus, for the uncompensated system,
T
s

4
jRej
4
3:5
1:143 s:
b. For the uncompensated system, K
v
lim
s!0
sGs K=7 58:9=7 8:41. Hence,
e
ramp uncompensated
1 1=K
v
0:1189.
c. In order to decrease the settling time by a factor of 2, the design point is twice the
uncompensated value, or 7 j 13:66. Adding the angles fromthe plants poles and
the compensators zero at 3 to the design point, we obtain 100:8
. Thus, the
compensator pole must contribute 180
100:8
79:2
. Using the following

diagram,
p
c
s
jw
s-plane
79.2
7
j13.66
we nd
13:66
p
c
7
tan79:2
, fromwhich, p
c
9:61. Adding this pole, we nd the gain
at the design point to be K 204:9.
Evaluating K
v
for the lead-compensated system:
K
v
lim
s!0
sGsG
lead
K3=79:61 204:93=79:61 9:138:
K
v
for the uncompensated system was 8.41. For a 10x improvement in steady-
state error, K
v
must be 8:4110 84:1. Since lead compensation gave us
K
v
9:138, we need an improvement of 84:1=9:138 9:2. Thus, the lag com-
pensator zero should be 9.2x further to the left than the compensator pole.
Arbitrarily select G
c
s
s 0:092
s 0:01
.
Using all plant and compensator poles, we nd the gain at the design point to
be K 205:4. Summarizing the forward path withplant, compensator, and gain
yields
G
e
s
205:4s 3s 0:092
ss 79:61s 0:01
:
Higher-order poles are found at 0:928 and 2:6. It would be advisable to simulate
the system to see if there is indeed pole-zero cancellation.
9.4
The conguration for the system is shown in the
gure below.
1
s(s+ 7)(s +10)
R(s)
C(s) +
K
+
K
f
s
Minor-Loop Design:
For the minor loop, GsHs
K
f
s 7s 10
. Using the following diagram, we nd
that the minor-loop root locus intersects the 0.7 damping ratio line at 8:5 j 8:67. The
imaginary part was found as follows: u cos
1
z 45:57
. Hence,
Im
8:5
tan 45:57
,
from which Im 8:67.
s
jw
s-plane
7
z = 0.7
X X
10
8.5
(-8.5 + j8.67)
q
Im
The gain, K
f
, is found from the vector lengths as
K
f

1:5
2
8:67
2
_
1:5
2
8:67
2
_
77:42
Major-Loop Design:
Using the closed-loop poles of the minor loop, we have an equivalent forward-path
transfer function of
G
e
s
K
ss 8:5 j8:67s 8:5 j8:67
K
ss
2
17s 147:4
:
Using the three poles of G
e
(s) as open-looppoles toplot a root locus, we searchalong
z 0:5 and nd that the root locus intersects this damping ratio line at 4:34
j7:51 at a gain, K 626:3.
9.5
a. An active PID controller must be used. We use the circuit shown in the following
gure:
+
Z
1
(s)
Z
2
(s)
I
1
(s)
V
1
(s)
V
o
(s)
V
i
(s)
I
a
(s)
I
2
(s)
where the impedances are shown below as follows:
C
1
R
1
Z
1
(s) Z
2
(s)
C
2
R
2
Matching the given transfer function with the transfer function of the PID controller
yields
G
c
s
s 0:1s 5
s

s
2
5:1s 0:5
s
s 5:1
0:5
8

_
R
2
R
1
C
1
C
2
_ _
R
2
C
1
s
1
R
1
C
2
s
_
Equating coefcients
1
R
1
C
2
0:5 (1)
R
2
C
1
1 (2)
R
2
R
1
C
1
C
2
_ _
5:1 (3)
In Eq. (2) we arbitrarily let C
1
10
5
. Thus, R
2
10
5
. Using these values along with
Eqs. (1) and (3) we nd C
2
100 mF and R
1
20 kV.
b. The lag-lead compensator can be implemented with the following passive network,
since the ratio of the lead pole-to-zero is the inverse of the ratio of the lag pole-to-
zero:
R
1
C
1
R
2
C
2
+
v
o
(t) v
i
(t)
Matching the given transfer function with the transfer function of the passive lag-lead
compensator yields
G
c
s
s 0:1s 2
s 0:01s 20
s 0:1s 2
s
2
20:01s 0:2
s
1
R
1
C
1
_ _
s
1
R
2
C
2
_ _
s
2
1
R
1
C
1
1
R
2
C
2
1
R
2
C
1
_ _
s
1
R
1
R
2
C
1
C
2
Equating coefcients
1
R
1
C
1
0:1 (1)
1
R
2
C
2
0:1 (2)
1
R
1
C
1
1
R
2
C
2
1
R
2
C
1
_ _
20:01 (3)
Substituting Eqs. (1) and (2) in Eq. (3) yields
1
R
2
C
1
17:91 (4)
Arbitrarily letting C
1
100 mF in Eq. (1) yields R
1
100 kV.
Substituting C
1
100 mF into Eq. (4) yields R
2
558 kV.
Substituting R
2
558 kV into Eq. (2) yields C
2
900 mF.
CHAPTER 10
10.1
a.
Gs
1
s 2s 4
; G jv
1
8 v
2
j6v
Mv
8 v
2
2
6v
2
_
For v <
8
p
, fv tan
1
6v
8 v
2
_ _
:
For v <
8
p
, fv p tan
1
6v
8 v
2
_ _ _ _
:
b.
Frequency (rad/sec)
P
h
a
s
e

(
d
e
g
)
;

M
a
g
n
i
t
u
d
e

(
d
B
)
Bode Diagrams
10
1
10
0
10
1
10
2
200
150
100
50
0
100
80
60
40
20
0
c.
Real Axis
I
m
a
g
i
n
a
r
y

A
x
i
s
Nyquist Diagrams
0.05 0 0.05 0.1 0.15 0.2
0.08
0.06
0.04
0.02
0
0.02
0.04
0.06
0.08
10.2
2
0

l
o
g

M
120
100
80
60
40
0.1
1
10 100 1000
40 dB/dec
20 dB/dec
20 dB/dec
40 dB/dec
Frequency (rad/s)
Actual
Asymptotic
Frequency (rad/s)
P
h
a
s
e

(
d
e
g
r
e
e
s
)
150
100
50
200
45
o
/dec
90
o
/dec
45
o
/dec
90
o
/dec
45
o
/dec
45
o
/dec
0.1 1 10 100
Actual
Asymptotic
1000
10.3
The frequency response is 1/8 at an angle of zero degrees at v 0. Each pole rotates 90
in going from v 0 to v 1. Thus, the resultant rotates 180
while its magnitude

goes to zero. The result is shown below.
Re
Im
0 1
8
w =
w = 0
10.4
a. The frequency response is 1/48 at an angle of zero degrees at v 0. Each pole rotates
90
in going from v 0 to v 1. Thus, the resultant rotates 270
while its
magnitude goes to zero. The result is shown below.
Im
Re
1
48
w = 0 w =
0
w = 6.63
1
480
b. Substituting jv into Gs
1
s 2s 4s 6
1
s
3
12s
2
44s 48
and
simplifying, we obtain G jv
48 12v
2
j44v v
3
v
6
56v
4
784v
2
2304
. The Nyquist diagram
crosses the real axis when the imaginary part of G jv is zero. Thus, the Nyquist
diagram crosses the real axis at v
2
44; or v
44
p
6:63 rad=s. At this fre-
quency G jv
1
480
. Thus, the system is stable for K < 480.
10.5
If K 100, the Nyquist diagram will intersect the real axis at 100=480. Thus,
G
M
20 log
480
100
13:62 dB. FromSkill-Assessment Exercise Solution 10.4, the 180
frequency is 6.63 rad/s.

10.6
a.
1 10 100 1000
180
160
140
120
100
80
Frequency (rad/s)
2
0

l
o
g

M
60
Frequency (rad/s)
P
h
a
s
e

(
d
e
g
r
e
e
s
)
1 10 100 1000
300
250
200
150
100
50
0
b. The phase angle is 180
at a frequency of 36.74 rad/s. At this frequency the gain is

99:67 dB. Therefore, 20 logK 99:67, or K 96; 270. We conclude that the
system is stable for K < 96; 270.
c. For K 10; 000, the magnitude plot is moved up 20log10; 000 80 dB. Therefore,
the gain margin is 99:67 80 19:67 dB. The 180
frequency is 36.7 rad/s. The

gain curve crosses 0 dBat v 7:74 rad=s, where the phase is 87:1
. We calculate the
phase margin to be 180
87:1
92:9
.
10.7
Using z
ln%=100
p
2
ln
2
%100
_ , we nd z 0:456, which corresponds to 20% over-
shoot. Using T
s
2; v
BW

4
T
s
z
1 2z
2

4z
4
4
_
z
2
2
_
5:79 rad=s.
10.8
For both parts nd that G jv
160
27

6750000 101250v
2
j1350v
2
1350v
v
6
2925v
4
1072500v
2
25000000
.
For a range of values for v, superimpose G jv on the a. M and N circles, and on the
b. Nichols chart.
a.
M = 1.3
1.4
1.5
1.6
1.8
2.0
M = 0.7
0.6
0.5
0.4
M = 1.0
2
1
1
2
3
Im
Re
1 2 1 2 3
3
= 20
25
30
20
40
50
30
70
40
50
70
25
G-plane
4
b.
Open-Loop Phase (deg)
O
p
e
n
-
L
o
o
p

G
a
i
n

(
d
B
)
Nichols Charts
350 300 250 200 150 100 50 0
200
150
100
50
0
6 dB
3 dB
1 dB
0.5 dB
0.25 dB
0 dB
1 dB
3 dB
6 dB
12 dB
20 dB
40 dB
60 dB
80 dB
100 dB
120 dB
140 dB
160 dB
180 dB
200 dB
220 dB
240 dB
Plotting the closed-loop frequency response from a. or b. yields the following plot:
120
100
80
60
40
20
0
Frequency (rad/s)
2
0

l
o
g

M
1 10 100 1000
Frequency (rad/s)
1 10 100 1000
300
250
200
150
100
50
0
P
h
a
s
e

(
d
e
g
r
e
e
s
)
10.9
The open-loop frequency response is shown in the following gure:
Frequency (rad/sec)
P
h
a
s
e

(
d
e
g
)
;

M
a
g
n
i
t
u
d
e

(
d
B
)
Bode Diagrams
40
20
0
20
40
10
1
10
0
10
1
10
2
160
140
120
100
The open-loop frequency response is 7 at v 14:5 rad=s. Thus, the estimated
bandwidth is v
WB
14:5 rad=s. The open-loop frequency response plot goes through
zero dBat a frequency of 9.4 rad/s, where the phase is 151:98
. Hence, the phase margin

is 180
151:98
28:02
. This phase margin corresponds to

z 0:25: Therefore; %OS e
zp=
1z
2
p _ _
x100 44:4%;
T
s

4
v
BW
z
1 2z
2

4z
4
4z
2
2
_
_
1:64 s and
T
p

p
v
BW
1 z
2
_
1 2z
2

4z
4
4z
2
2
_
_
0:33 s
10.10
The initial slope is 40 dB/dec. Therefore, the system is Type 2. The initial slope
intersects 0 dB at v 9:5 rad=s. Thus, K
a
9:5
2
90:25 and K
p
K
v
1.
10.11
a. Without delay, G jv
10
jv jv 1
10
vv j
, from which the zero dB fre-
quency is found as follows: M
10
v
v
2
1
p 1. Solving for v; v
v
2
1
p
10, or
after squaring both sides and rearranging, v
4
v
2
100 0. Solving for the roots,
v
2
10:51; 9:51. Taking the square root of the positive root, we nd the 0 dB
frequency to be 3.08 rad/s. At this frequency, the phase angle, f v j
3:08 j 162
. Therefore the phase margin is 180
162
18
.
b. With a delay of 0.1 s,
f v j vT 3:08 j 3:080:1180=pi 162 17:65
179:65
:
Therefore the phase margin is 180
179:65
0:35
. Thus, the system is stable.

c. With a delay of 3 s,
f v j vT 3:08 j 3:083180=pi 162
529:41
691:41
28:59 deg:
Therefore the phase margin is 28:59 180 151:41 deg. Thus, the system is
unstable.
10.12
Drawing judicially selected slopes on the magnitude and phase plot as shown below
yields a rst estimate.
Experimental
Frequency (rad/sec)
40
30
20
10
0
10
20
1 2 3 4 5 6 7 8 10 20 30 40 50 70 100 200 300 500 1000
95
90
85
80
75
70
65
60
55
50
45
P
h
a
s
e
(
d
e
g
)
G
a
i
n
(
d
B
)
We see an initial slope on the magnitude plot of 20 dB/dec. We also see a nal 20 dB/
dec slope with a break frequency around 21 rad/s. Thus, an initial estimate is
G
1
s
1
ss 21
. Subtracting G
1
(s) from the original frequency response yields
the frequency response shown below.
40
50
60
70
80
90
1 2 3 4 5 6 7 8 10 20 30 40 50 70 100 200 300 500 1000
0
20
40
60
80
100
Experimental Minus 1/s(s+21)
Frequency (rad/sec)
G
a
i
n
(
d
B
)
P
h
a
s
e
(
d
e
g
)
Drawing judicially selected slopes on the magnitude and phase plot as shown yields a
nal estimate. We see rst-order zero behavior on the magnitude and phase plots with a
break frequency of about 5.7 rad/s and a dc gain of about 44 dB 20 log5:7K, or
K 27:8. Thus, we estimate G
2
s 27:8s 7. Thus, Gs G
1
sG
2
s
27:8s 5:7
ss 21
. It is interesting to note that the original problem was developed from
Gs
30s 5
ss 20
.
CHAPTER 11
11.1
The Bode plot for K 1 is shown below.
Frequency (rad/sec)
P
h
a
s
e

(
d
e
g
)
;

M
a
g
n
i
t
u
d
e

(
d
B
)
Bode Diagrams
180
160
140
120
100
80
60
10
1
10
0
10
1
10
2
10
3
250
200
150
100
A20%overshoot requires z
log
%
100
_ _
p
2
log
2 %
100
_ _
_
0:456. This damping ratio implies a
phase margin of 48.10, which is obtained when the phase angle 1800 48:10
131:9
. This phase angle occurs at v 27:6 rad=s. The magnitude at this frequency is
5:15 10
6
. Since the magnitude must be unity K
1
5:1510
6
194;200.
11.2
To meet the steady-state error requirement, K 1;942;000. The Bode plot for this gain
is shown below.
Frequency (rad/sec)
P
h
a
s
e

(
d
e
g
)
;

M
a
g
n
i
t
u
d
e

(
d
B
)
Bode Diagrams
40
20
0
20
40
60
10
1
10
0
10
1
10
2
10
3
250
200
150
100
A 20% overshoot requires z
log
%
100
_ _
p
2
log
2
%
100
_ _
0:456: This damping ratio
implies a phase margin of 48:1
. Adding 10
to compensate for the phase angle

contribution of the lag, we use 58:1
. Thus, we look for a phase angle of 180
58:1
129:9
. The frequency at which this phase occurs is 20.4 rad/s. At this

frequency the magnitude plot must go through zero dB. Presently, the magnitude plot is
23.2 dB. Therefore drawthe high frequency asymptote of the lag compensator at 23:2
dB. Insert a break at 0:120:4 2:04 rad=s. At this frequency, draw 23:2 dB/dec
slope until it intersects 0 dB. The frequency of intersection will be the low frequency
break or 0.141 rad/s. Hence the compensator is G
c
s K
c
s 2:04
s 0:141
, where the gain
is chosen to yield 0 dB at low frequencies, or K
c
0:141=2:04 0:0691. In summary,
G
c
s 0:0691
s 2:04
s 0:141
and Gs
1;942;000
ss 50s 120
11.3
log
%
100
_ _
p
2
log
2
%
100
_ _
0:456. The required bandwidth is
then calculated as v
BW

4
T
s
z
1 2z
2

4z
4
4z
2
2
_
_
57:9 rad/s. In order to
meet the steady-state error requirement of K
v
50
K
50120
, we calculate
K 300;000. The uncompensated Bode plot for this gain is shown below.
Frequency (rad/sec)
P
h
a
s
e

(
d
e
g
)
;

M
a
g
n
i
t
u
d
e

(
d
B
)
Bode Plot for K = 300000
60
40
20
0
20
40
10
1
10
0
10
1
10
2
10
3
250
200
150
100
The uncompensated systems phase margin measurement is taken where the magnitude
plot crosses 0 dB. We nd that when the magnitude plot crosses 0 dB, the phase
angle is 144:8
. Therefore, the uncompensated systems phase margin is 180
144:8
35:2
. The required phase margin based on the required damping ratio is

F
M
tan
1
2z
2z
2
1 4z
4
_
_ 48:1
. Adding a 10
correction factor, the

required phase margin is 58:1
. Hence, the compensator must contribute

f
max
58:1
35:2
22:9
. Using f
max
sin
1
1 b
1 b
, b
1 sin f
max
1 sin f
max
0:44.
The compensators peak magnitude is calculated as M
max

1
b
p 1:51. Now nd the
frequency at which the uncompensated system has a magnitude 1=M
max
, or 3:58 dB.
Fromthe Bode plot, this magnitude occurs at v
max
50 rad=s. The compensators zero
is at z
c

1
T
. v
max

1
T
b
p : Therefore, z
c
33:2.
The compensators pole is at p
c

1
bT

z
c
b
75:4. The compensator gain is chosen to
yield unity gain at dc.
Hence, K
c
75:4=33:2 2:27. Summarizing, G
c
s2:27
s 33:2
s 75:4
, and
Gs
300;000
ss 50s 120
.
11.4
log
%
100
_ _
p
2
log
2
%
100
_ _
0:591. The required bandwidth is
then calculated as v
BW

p
T
p
1 z
2
_
1 2z
2

4z
4
4z
2
2
_
_
7:53 rad=s.
In order to meet the steady-state error requirement of K
v
10
K
830
, we calculate
K 2400. The uncompensated Bode plot for this gain is shown below.
Frequency (rad/sec)
P
h
a
s
e

(
d
e
g
)
;

M
a
g
n
i
t
u
d
e

(
d
B
)
Bode Diagrams
100
80
60
40
20
0
20
40
10
1
10
0
10
1
10
2
10
3
250
200
150
100
Let us select a new phase-margin frequency at 0:8v
BW
6:02 rad=s. The required
phase margin based on the required damping ratio is F
M
tan
1
2z
2z
2
1 4z
4
_
_ 58:6
. Adding a 5
correction factor, the required phase margin

is 63:6
. At 6.02 rad/s, the new phase-margin frequency, the phase angle is

which represents a phase margin of 180
138:3
41:7
. Thus, the lead comp-

ensator must contribute f
max
63:6
41:7
21:9
.
Using f
max
sin
1
1 b
1 b
, b
1 sinf
max
1 sinf
max
0:456.
We now design the lag compensator by rst choosing its higher break frequency one
decade below the new phase-margin frequency, that is, z
lag
0:602 rad=s. The lag
compensators pole is p
lag
bz
lag
0:275. Finally, the lag compensators gain is
K
lag
b 0:456.
Now we design the lead compensator. The lead zero is the product of the new phase
margin frequency and
b
p
, or z
lead
0:8 v
BW
b
p
4:07. Also, p
lead

z
lead
b
8:93.
Finally, K
lead

1
b
2:19. Summarizing,
G
lag
s 0:456
s 0:602
s 0:275
; G
lead
s 2:19
s 4:07
s 8:93
; and K 2400:
CHAPTER 12
12.1
We rst nd the desired characteristic equation. A 5% overshoot requires
z
log
%
100
_ _
p
2
log
2
%
100
_ _
0:69. Also, v
n

p
T
p
1 z
2
_ 14:47 rad=s. Thus, the
characteristic equation is s
2
2zv
n
s v
2
n
s
2
19:97s 209:4. Adding a pole at
10 to cancel the zero at 10 yields the desired characteristic equa-
tion,s
2
19:97s 209:4s 10 s
3
29:97s
2
409:1s 2094. The compen-
sated system matrix in phase-variable form is
A BK
0 1 0
0 0 1
k
1
36 k
2
15 k
3
_
_
_
_
. The characteristic equation for this
system is jsI A BKj s
3
15 k
3
s
2
36 k
2
s k
1
. Equating coef-
cients of this equation with the coefcients of the desired characteristic equation yields
the gains as
K k
1
k
2
k
3
2094 373:1 14:97 :
12.2
The controllability matrix is C
M
B AB A
2
B
_
2 1 1
1 4 9
1 1 16
_
_
_
_
. Since
jC
M
j 80, C
M
is full rank, that is, rank 3. We conclude that the system is controllable.
12.3
First check controllability. The controllability matrix is C
Mz
B AB A
2
B
_
0 0 1
0 1 17
1 9 81
_
_
_
_
. Since jC
Mz
j 1, C
Mz
is full rank, that is, rank 3. We conclude that
the system is controllable. We now nd the desired characteristic equation. A 20%
overshoot requires z
log
%
100
_ _
p
2
log
2
%
100
_ _
0:456. Also, v
n

4
zT
s
4:386 rad=s.
Thus, the characteristic equation is s
2
2zv
n
s v
2
n
s
2
4s 19:24. Adding a pole
at 6 to cancel the zero at 6 yields the resulting desired characteristic equation,
s
2
4s 19:24s 6 s
3
10s
2
43:24s 115:45:
Since Gs
s 6
s 7s 8s 9
s 6
s
3
24s
2
191s 504
, we can write the
phase-variable representation as A
p

0 1 0
0 0 1
504 191 24
_
_
_
_
; B
p

0
0
1
_
_
_
_
; C
p

6 1 0 . The compensated system matrix in phase-variable form is A
p
B
p
K
p

0 1 0
0 0 1
504 k
1
191 k
2
24 k
3
_
_
_
_
. The characteristic equation for this
system is jsI A
p
B
p
K
p
j s
3
24 k
3
s
2
191 k
2
s504 k
1
. Equat-
ing coefcients of this equation with the coefcients of the desired characteristic
equation yields the gains as K
p
k
1
k
2
k
3
388:55 147:76 14 . We
now develop the transformation matrix to transform back to the z-system.
C
Mz
B
z
A
z
B
z
A
2
z
B
z
_ _
0 0 1
0 1 17
1 9 81
_
_
_
_
and
C
Mp
B
p
A
p
B
p
A
2
p
B
p
_ _
0 0 1
0 1 24
1 24 385
_
_
_
_
:
Therefore,
P C
Mz
C
1
Mx

0 0 1
0 1 17
1 9 81
_
_
_
_
191 24 1
24 1 0
1 0 0
_
_
_
_
1 0 0
7 1 0
56 15 1
_
_
_
_
Hence; K
z
K
p
P
1
388:55 147:76 14
1 0 0
7 1 0
49 15 1
_
_
_
_
40:23 62:24 14 :
12.4
For the given system e
_ x
A LCe
x
24 l
1
1 0
191 l
2
0 1
504 l
3
0 0
_
_
_
_
e
x
. The characteristic
polynomial is given by jsI A LCj s
3
24 l
1
s
2
191l
2
s504 l
3
.
Now we nd the desired characteristic equation. The dominant poles from Skill-Assess-
ment Exercise 12.3 come from s
2
4s 19:24. Factoring yields 2 j3:9 and
2 j3:9. Increasing these poles by a factor of 10 and adding a third pole 10 times
the real part of the dominant second-order poles yields the desired characteristic poly-
nomial, s 20 j39s 20 j39s 200 s
3
240s
2
9921s 384200. Equ-
ating coefcients of the desired characteristic equation to the systems characteristic
equation yields L
216
9730
383696
_
_
_
_
.
12.5
The observability matrix is O
M

C
CA
CA
2
_
_
_
_
4 6 8
64 80 78
674 848 814
_
_
_
_
, where
A
2
25 28 32
7 4 11
77 95 94
_
_
_
_
. The matrix is of full rank, that is, rank 3, since
jO
M
j 1576. Therefore the system is observable.
12.6
The system is represented in cascade form by the following state and output equations:
_ z
7 1 0
0 8 1
0 0 9
_
_
_
_z
0
0
1
_
_
_
_u
y 1 0 0 z
The observability matrix is O
Mz

C
z
C
z
A
z
C
z
A
2
z
_
_
_
_
1 0 0
7 1 0
49 15 1
_
_
_
_
, where A
2
z

49 15 1
0 64 17
0 0 81
_
_
_
_
. Since Gs
1
s 7s 8s 9
1
s
3
24s
2
191s 504
,
we can write the observable canonical form as
_ x
24 1 0
191 0 1
504 0 0
_
_
_
_
x
0
0
1
_
_
_
_
u
y 1 0 0 x
The observability matrix for this form is O
Mx

C
x
C
x
A
x
C
x
A
2
x
_
_
_
_
1 0 0
24 1 0
385 24 1
_
_
_
_
,
where A
2
x

385 24 1
4080 191 0
12096 504 0
_
_
_
_
:
We next nd the desired characteristic equation. A 10% overshoot requires
z
log
%
100
_ _
p
2
log
2
%
100
_ _
0:591. Also, v
n

4
zT
s
67:66 rad=s. Thus, the character-
istic equation is s
2
2zv
n
s v
2
n
s
2
80s 4578:42. Adding a pole at 400, or 10
times the real part of the dominant second-order poles, yields the resulting desired
characteristic equation, s
2
80s 4578:42s 400 s
3
480s
2
36580s
1:831x10
6
. For the system represented in observable canonical form e
_ x

A
x
L
x
C
x
e
x

24 l
1
1 0
191 l
2
0 1
504 l
3
0 0
_
_
_
_
e
x
. The characteristic polynomial is given
by jsI A
x
L
x
C
x
j s
3
24 l
1
s
2
191 l
2
s 504 l
3
. Equating
coefcients of the desired characteristic equation to the systems characteristic equation
yields L
x

456
36; 389
1; 830; 496
_
_
_
_
:
Now, develop the transformation matrix between the observer canonical and cascade
forms.
P O
1
Mz
O
Mx

1 0 0
7 1 0
49 15 1
_
_
_
_
1
1 0 0
24 1 0
385 24 1
_
_
_
1 0 0
7 1 0
56 15 1
_
_
_
_
1 0 0
24 1 0
385 24 1
_
_
_
_
1 0 0
17 1 0
81 9 1
_
_
_
_:
Finally,
L
z
PL
x

1 0 0
17 1 0
81 9 1
_
_
_
_
456
36; 389
1; 830; 496
_
_
_
_
456
28; 637
1; 539; 931
_
_
_
_
456
28; 640
1; 540; 000
_
_
_
_
.
12.7
We rst nd the desired characteristic equation. A 10% overshoot requires
z
log
%
100
_ _
p
2
log
2
%
100
_ _
0:591
Also, v
n

p
T
p
1 z
2
_ 1:948 rad=s. Thus, the characteristic equation is s
2
2zv
n
s v
2
n
s
2
2:3s 3:79. Adding a pole at 4, which corresponds to the
original systems zero location, yields the resulting desired characteristic equation,
s
2
2:3s 3:79s 4 s
3
6:3s
2
13s 15:16.
Now,
_ x
_ x
N
_ _
A BK BK
e
C 0
_ _
x
x
N
_ _
0
1
_ _
r; and y C 0
x
x
N
_ _
,
where
A BK
0 1
7 9
_ _
0
1
_ _
k
1
k
2

0 1
7 9
_ _
0 0
k
1
k
2
_ _
0 1
7 k
1
9 k
2
_ _
C 4 1
Bk
e

0
1
_ _
k
e

0
k
e
_ _
Thus,
_ x
1
_ x
2
_ x
N
_
_
_
_
0 1 0
7 k
1
9 k
2
k
e
4 1 0
_
_
_
_
x
1
x
2
x
N
_
_
_
_
0
1
_ _
r; y 4 1 0
x
1
x
2
x
N
_
_
_
_.
Finding the characteristic equation of this system yields
sI
A BK BK
e
C 0
_ _
s 0 0
0 s 0
0 0 s
_
_
_
_
0 1 0
7 k
1
9 k
2
k
e
4 1 0
_
_
_
_
s 1 0
7 k
1
s 9 k
2
k
e
4 1 s
_
_
_
_
s
3
9 k
2
s
2
7 k
1
k
e
s 4k
e
Equating this polynomial to the desired characteristic equation,
s
3
6:3s
2
13s 15:16 s
3
9 k
2
s
2
7 k
1
k
e
s 4k
e
Solving for the ks,
K 2:21 2:7 and k
e
3:79:
CHAPTER 13
13.1
f t sinvkT; f
t
1
k0
sinvkTdt kT;
F
s
1
k0
sinvkTe
kTs

1
k0
e
jvkT
e
jvkT
_ _
e
kTs
2 j
1
2 j

1
k0
_
e
Tsjv
_
k
_
e
Tsjv
_
k
But,
1
k0
x
k
1
1 x
1
Thus,
F
s
1
2 j
1
1 e
Tsjv
1
1 e
Tsjv
_ _
1
2 j
e
Ts
e
jvT
e
Ts
e
jvT
1 e
Ts
e
jvT
e
Ts
e
jvT
e
2Ts
_ _
e
Ts
sinvT
1 e
Ts
2cosvT e
2Ts
_ _
z
1
sinvT
1 2z
1
cosvT z
2
13.2
Fz
zz 1z 2
z 0:5z 0:7z 0:9
Fz
z

z 1z 2
z 0:5z 0:7z 0:9
46:875
z
z 0:5
114:75
z
z 0:7
68:875
z
z 0:9
Fz 46:875
z
z 0:5
114:75
z
z 0:7
68:875
z
z 0:9
;
f kT 46:8750:5
k
114:750:7
k
68:8750:9
k
13.3
Since Gs 1 e
Ts
_ _
8
ss 4
,
Gz 1 z
1
_ _
z
8
ss 4
_ _
z 1
z
z
A
s

B
s 4
_ _
z 1
z
z
2
s
2
s 4
_ _
:
Let G
2
s
2
s
2
s 4
. Therefore, g
2
t 2 2e
4t
, or g
2
kT 2 2e
4kT
.
Hence, G
2
z
2z
z 1
2z
z e
4T

2z1 e
4T
z 1z e
4T
.
Therefore, Gz
z 1
z
G
2
z
21 e
4T
z e
4T
.
For T
1
4
s, Gz
1:264
z 0:3679
.
13.4
Add phantom samplers to the input, feedback after H(s), and to the output. Push
G
1
(s)G
2
(s), along with its input sampler, to the right past the pickoff point and
obtain the block diagram shown below.
H(s)G
1
(s)G
2
(s)
G
1
(s)G
2
(s)
R(s)
C(s)
+
Hence, Tz
G
1
G
2
z
1 HG
1
G
2
z
.
13.5
Let Gs
20
s 5
. Let G
2
s
Gs
s

20
ss 5
4
s
4
s 5
. Taking the inverse
Laplace transform and letting t kT, g
2
kT 4 4e
5kT
. Taking the z-transform
yields G
2
z
4z
z 1
4z
z e
5T

4z1 e
5T
z 1z e
5T
.
Now, Gz
z 1
z
G
2
z
41 e
5T
z e
5T
. Finally, Tz
Gz
1 Gz
41 e
5T
z 5e
5T
4
.
The pole of the closed-loop systemis at 5e
5T
4. Substituting values of T, we nd that
the pole is greater than 1 if T >0:1022 s. Hence, the system is stable for
0 < T < 0:1022 s.
13.6
Substituting z
s 1
s 1
into Dz z
3
z
2
0:5z 0:3, we obtain Ds
s
3
8s
2
27s 6. The Routh table for this polynomial is shown below.
Since there is one sign change, we conclude that the system has one pole outside the unit
circle andtwopoles inside the unit circle. The table didnot produce a rowof zeros andthus,
there are no jv poles. The system is unstable because of the pole outside the unit circle.
13.7
Dening G(s) as G
1
(s) in cascade with a zero-order-hold,
Gs 201 e
Ts
s 3
ss 4s 5
_ _
201 e
Ts
3=20
s

1=4
s 4
2=5
s 5
_ _
:
Taking the z-transform yields
Gz 201 z
1
3=20z
z 1

1=4z
z e
4T

2=5z
z e
5T
_ _
3
5z 1
z e
4T

8z 1
z e
5T
:
Hence for T 0:1 second, K
p
lim
z!1
Gz 3, and K
v

1
T
lim
z!1
z 1Gz 0, and
K
a

1
T
2
lim
z!1
z 1
2
Gz 0. Checking for stability, we nd that the system is stable
for T 0:1 second, since Tz
Gz
1 Gz
1:5z 1:109
z
2
0:222z 0:703
has poles inside the
unit circle at 0:957 and 0:735. Again, checking for stability, we nd that the system
is unstable for T 0:5 second, since Tz
Gz
1 Gz
3:02z 0:6383
z
2
2:802z 0:6272
has
poles inside and outside the unit circle at 0:208 and 3:01, respectively.
13.8
Drawthe root locus superimposed over the z 0:5 curve shown below. Searching along
a 54:3
line, which intersects the root locus and the z 0:5 curve, we nd the point
0:58754:3
0:348 j0:468 and K 0:31.

3 2.5 2 1.5 1 0.5 0 0.5 1 1.5 2
1.5
1
0.5
0
0.5
1
1.5
Real Axis
I
m
a
g

A
x
i
s
z-Plane Root Locus
54.3
(0.348 + j 0.468)
K = 0.31
s
3
1 27
s
2
8 6
s
1
27:75 0
s
0
6 0
13.9
Let
G
e
s GsG
c
s
100K
ss 36s 100
2:38s 25:3
s 60:2
342720s 25:3
ss 36s 100s 60:2
:
The following shows the frequency response of G
e
jv.
Frequency (rad/sec)
P
h
a
s
e

(
d
e
g
)
;

M
a
g
n
i
t
u
d
e

(
d
B
)
Bode Diagrams
60
40
20
0
20
40
10
1
10
0
10
1
10
2
10
3
250
200
150
100
We nd that the zero dB frequency, v
F
M
, for G
e
jv is 39 rad/s. Using Astroms
guideline the value of T should be in the range, 0:15=v
F
M
0:0038 second to
0:5=v
F
M
0:0128 second. Let us use T 0:001 second. Now nd the Tustin
transformation for the compensator. Substituting s
2z 1
Tz 1
into G
c
s
2:38s 25:3
s 60:2
with T 0:001 second yields
G
c
z 2:34
z 0:975
z 0:9416
:
13.10
G
c
z
Xz
Ez
1899z
2
3761z 1861
z
2
1:908z 0:9075
. Cross-multiply and obtain z
2
1:908z
0:9075Xz 1899z
2
3761z 1861Ez. Solve for the highest power of z oper-
ating on the output, X(z), and obtain z
2
Xz 1899z
2
3761z 1861
Ez 1:908z 0:9075Xz. Solving for X(z) on the left-hand side yields
Xz 1899 3761z
1
1861z
2
Ez 1:908z
1
0:9075z
2
Xz. Finally,
we implement this last equation with the following ow chart:
e
*
(t)
x
*
(t)
e
*
(t-0.1) x
*
(t-0.1)
x
*
(t-0.2)
e
*
(t-0.2)
Delay
0.1 second
1899
3761
1861 0.9075
1.9.08
+
+
+
+
Delay
0.1 second
Delay
0.1 second
Delay
0.1 second

Solutions To Skill-Assessment Exercises

Caricato da

Informazioni sul documento

Titolo originale

Copyright

Formati disponibili

Condividi questo documento

Condividi o incorpora il documento

Opzioni di condivisione

Hai trovato utile questo documento?

Questo contenuto è inappropriato?

Copyright:

Formati disponibili

Solutions To Skill-Assessment Exercises

Caricato da

Copyright:

Formati disponibili

SOLUTIONS TO

Taking the inverse Laplace transform yields,

From which the eigenvalues are 2 and 3.

Chapter 8 Solutions to Skill-Assessment Exercises 21

. Solving for the angle of departure, u 233:1.

. Thus, the compensator pole must contribute 180

. Using the following

in going from v 0 to v 1. Thus, the resultant rotates 180

while its magnitude

in going from v 0 to v 1. Thus, the resultant rotates 270

frequency is 6.63 rad/s.

at a frequency of 36.74 rad/s. At this frequency the gain is

frequency is 36.7 rad/s. The

. Hence, the phase margin

. This phase margin corresponds to

. Therefore the phase margin is 180

. Thus, the system is stable.

to compensate for the phase angle

. Thus, we look for a phase angle of 180

. The frequency at which this phase occurs is 20.4 rad/s. At this

. Therefore, the uncompensated systems phase margin is 180

. The required phase margin based on the required damping ratio is

correction factor, the

. Hence, the compensator must contribute

correction factor, the required phase margin

. At 6.02 rad/s, the new phase-margin frequency, the phase angle is

. Thus, the lead comp-

0:348 j0:468 and K 0:31.

Potrebbero piacerti anche