solution 7.9.1.3
(a)
Wehave
G
p
=
10
(s+3)(s+30)
and G
c
=
K
c
(s+0:5)
s+ b
Figure 1(a) shows the root locus without the lag compensator, while g-
ure 1(b) shows the root locus with the lag compensator, assuming we place
the pole of the lag comensator near the origin. At the pointwherethetwo
Re(s)
Im(s)
Re(s)
Im(s)
(a)
(b)
-30
-3 -30 -3
Figure 1: Cascade Compensation with UnityFeedback
root loci cross the line of constant damping =0:8, the loci are very simi-
lar. For the locus of part(a), wecannd the intersection analytically. The
break-out pointisats = ;16:5. To nd the imaginary part of the proposed
closed loop poles, we note that
=cos
;1
(0:8) = 36:87
:
Then
tan =
x
16:5
;;
or
x =16:5tan =16:5tan36:87
=12:375:
1
Thus the closed loop poles that meet the specications are
s = ;16:5j12:375
The gain required to place the closed loop poles at the desired locations is
10K
c
=
p
13:5
2
+12:375
2
p
13:5
2
+12:375
2
=335:4
Then the closed loop transfer function is
T
c
(s)=
335:4
(s +16:5;j12:375)(s+16:5+j12:375)
;;
and the step response is
C(s) =
335:4
s(s+16:5;j12:375)(s+16:5+j12:375)
=
A
s
+
M
s+16:5;j12:375
+
M
s+16:5+j12:375
:
A = sC(s) j
s=0
= T
c
(s) j
s=0
=
335:4
(16:5;j12:375)(16:5+j12:375)
=
335:4
16:5
2
+12:375
2
= 0:7884
M = (s+16:5;j12:375)C(s)j
s=;16:5+j12:375
= T
c
(s) j
s=;16:5+j12:375
=
335:4(s+16:5;j12:375)
s(16:5;j12:375)(16:5+j12:375)
s=;16:5+j12:375
=
335:4
(16:5+12:375)(j24:75)
= ;0:3942+j0:526 = 0:6570
6
;2:2143:
The time response is then
c(t)=
h
0:7884+ 1:3140e
;16:5t
cos(12:375t;2:2143)
i
1(t):
We see that the steady state error to a step input is
e
ss
= 1;0:7884
= 0:2116;;
which exceeds our specication.
2
(b)
We knowthat
e
ss
=
1
1+K
p
;;
or
K
p
=
1;e
ss
e
ss
:
Then for
e
ss
0:1;;
wehave
K
p
9:
Next note that
K
p
= lim
s!0
335:4(s+0:5)
(s+ b)(s+3)(s+30)
=
1:8633
b
:
Thus for K>9, wemust have
b <
1:8633
9
;;
or
b<0:207:
Wechoose b =0:1.
Wenow nd the angle of
G
c
G
p
(s)=
335:4(s+0:5)
(s+0:1)(s+3)(s+30)
at s = ;16:5+j12:375.
6
G
c
G
p
(s = ;16:5+j12:375) = [180
;tan
;1
(12:375=16)];[180
;tan
;1
(12:375=16:4)]
;[180;tan
;1
(12:375=13:5)];tan
;1
(12:375=13:5)
= tan
;1
(12:375)=16:4);tan
;1
(12:375=16)
= ;179:3
:
Thus, wesee that the root locus with the lag compensator added passes very
close to s = ;16:5+j12:375, so close, in fact, that wecanassume it does
pass through this point, and choose
G
c
(s)=
335:4(s+0:5)
(s+0:1)
:
3
Then
T
c
(s)=
335:4(s+0:5)
(s+0:426)(s+16:34;j12:16)(s+16:34+j12:16)
;;
and the step response is
C(s) =
335:4(s+0:5)
s(s+0:426)(s+16:34;j12:16)(s+16:34+ j12:16)
=
A
s
+
B
s+0:426
+
M
s+16:34;j12:16
+
M
s+16:34+ j12:16
:
A = sC(s) j
s=0
= T
c
(s) j
s=0
=
335:4(0:5)
(0:426)(16:34;j12:16)(16:34+ j12:16)
= 0:949
B = (s+0:426)C(s)j
s=;0:426
=
335:4(s+0:5)
s(s+16:34;j12:16)(s+16:34+ j12:16)
s=;0:426
= ;0:1452
M = (s+16:34;j12:16)C(s)j
s=;16:34+j12:16
=
335:4(s+0:5)(s+16:34;j12:16)
s(s+0:426)(s+16:34;j12:16)(s+16:34+j12:16)
s=;16:34+j12:16
= ;0:401905+j0:542472
= 0:6751
6
;2:2083
The closed loop response is
c(t)=
h
0:949;0:1452e
;0:426t
+1:3502e
;16:34t
cos(12:16t;2:2083)
i
:
The steady state error to a step input is
e
ss
= 1;0:949
= 0:051;;
whichmore than meets the specication.
4
0
0.2
0.4
0.6
0.8
1
0 0.5 1 1.5 2 2.5
Time in Seconds
Response with Lag Compensation
Respone with Gain Compensation
Figure 2: Time Response with Gain and Lag Compensation
The time responses of the two systems are shown in Figure 2. Note
that the response with gain compensation reaches steady state rapidly, but
unfortunately the steady state is o bytwentypercentfrom the desired
value. The lag compensated system comes within vepercentofthe desired
steady state but exhibits the long slowcreep to nal value characteristic of
this kind of compensator.
5