Solution 7.9.6.4
For the system of Figure 1 wehave
Σ
G
C
G
P
+
R
C
Figure 1: Cascade Compensation with UnityFeedback
G
p
=
1
(s +0:5)
2
:
Part a.
Wehave
G
c
G
p
(s)=
K
c
(s +0:5)
2
s(s + a)
1
(s +0:5)
2
=
K
c
s(s + a)
The root locus is shown in Figure 2. For the poles to be at s = ;4 j2we
must have a =8.Then, as shown in the gure,
K
c
= jsjjs +8jj
s=;4+j2
=
p
2
2
+4
2
p
2
2
+4
2
=20:
Thus
T
c
(s)=
20
s
2
+4s +20
This is T
N2
(s)for!
n
=
p
20 and
=cos
tan
;1
(0:5)
=cos(26:57
)=0:89
As a result the system will be essentially critically damped. The peak occurs
at
t
p
=
2
=1:57 s.
but will be barely discernible since the overshoot is less than one per cent.
K
v
= lim
s!0
s
20
s(s +8)
=
20
8
The steady state error to a ramp input is
e
ss
=
1
K
v
=
8
20
=0:4;;
and the steady state error to a ramp input is 40%
1
Im(s)
-a = -8
2
- 2
s + 8
s
-4
Re(s)
Figure 2: Root Locus for First Compensator
Part b.
G
c
(s)G
p
(s)=
K
c
(s +0:5)(s+0:1)
s(s + a)
1
(s +0:5)
2
=
K
c
(s +0:1)
s(s + a)(s +0:5)
:
The root locus and angle calculation to place closed loop poles at s = ;4j2
is shown in Figure 3. The angle equation that must be solved at s = ;4+j2
is
;
1
;
2
;
3
= ;180
Im(s)
2
s + a
s
-0.1
-0.5
-a
s + 0.1
s + 0.5
θ
1
θ
2
θ
3
α
Re(s)
Figure 3: Root locus and angle calculation for second compensator
2
or
3
= +180
;
1
;
2
= [180
;tan
;1
(2=3:9)]+ 180
;[180
;tan
;1
(2=4)];[180
;tan
1;
(2=3:5)]
= 180
+ 152:85
;153:43
;150:255
= 29:17
:
Then
a =4+
2
tan(29:17
)
= ;7:584
Then, asshown in the gure,
K
c
=
jsjjs+7:584jjs+0:5j
js +0:1j
s=;4+j2
=
p
20
p
16:25
p
16:85
p
19:21
= 16:88
Thus
T
c
(s)=
16:88(s+0:1)
(s + )(s +2; j4)(s+2+j4)
:
The third closed loop pole can be found bysearching the interval [0;;;0:5]
for a closed loop pole location that will yield a gain of 16.88. Wedonot
need the third pole of the closed loop compensator to compute the steady
state error to a ramp input, because with unityfeedbackwecanuse the
forward loop transfer function to compute
K
v
=lim
s!0
s
16:88(s+0:1)
s(s +0:5)(s+7:854)
=0:445:
However, wecannd the the third pole witht the MATLAB statements
gcgp = zpk([-0.1],[0 -0.5 -7.584],16.88)
tc = feedback(gcgp,1)
whose output is
Zero/pole/gain:
16.88 (s+0.1)
-------------------
s (s+0.5) (s+7.584)
3
Zero/pole/gain:
16.88 (s+0.1)
---------------------------
(s+0.08441) (s^2 + 8s + 20)
Thus, the steady state error to a ramp input is
e
ss
=
1
K
v
=
1
0:43
=2:25:
Thus the steady state error to a ramp input is 225%, far worse than the rst
design, primarily because of the zero at s = ;0:1whichcauses T
c
to havea
pole close to the origin.
The following MATLAB dialogue generates the closed loop transfer func-
tion and the step response shown in Figure 4
EDU>gcgp = zpk([-0.1],[0 -7.584],16.88)
Zero/pole/gain:
16.88 (s+0.1)
-------------
s (s+7.584)
EDU>tc = feedback(gcgp,1)
Zero/pole/gain:
16.88 (s+0.1)
--------------------
(s+24.39) (s+0.0692)
EDU>gcgp = zpk([-0.1],[0 -0.5 -7.584],16.88)
Zero/pole/gain:
16.88 (s+0.1)
-------------------
s (s+0.5) (s+7.584)
4
EDU>tc = feedback(gcgp,1)
Zero/pole/gain:
16.88 (s+0.1)
---------------------------
(s+0.08441) (s^2 + 8s + 20)
EDU>step(tc,30)
EDU>print -deps sr7964b.eps
EDU>
Time (sec.)
Step Response
5 10 15 20 25 30
0.2
0.4
0.6
0.8
1
1.2
1.4
Figure 4: Step response of compensated syteme, part (b)
By partial fractions we nd the step response to be
c(t)=[1;0:1612e
;0:0844t
+1:8816e
;4t
cos(2t;2:0328)]1(t)
5