Solution: 7.9.4.4
R
+
C
G
c
G
p
Σ
-
Figure 1: Unityfeedback
For the system 0f Figure 1 wehave
G
p
(s)=
1
(s +0:1)(s+ 1)(s +20)
(a)
Figure 2 shows the poles of the plantand the zero of the PD compensator
and the desired closed loop poles.
XXX
O
Re(s)
Im(s)
-20 -a -1 -0.1
-4
3
Figure 2: Angle calculation
The zero is at
;a = ;
K
d
K
pd
:
1
The gain is K
pd
.Thus, for the root locus to pass through s = ;4+j3we
must satisfy the angle condition at the point, that is, wemust have
;
1
;
2
;
3
= ;180
;;
or
=
1
+
2
+
3
;180
= [180
;tan
;1
(3=3:9)]+ [180
;tan
;1
(3=3)]+ tan
;1
(3=16);180
= 142:43
+135
+10:62
;180
= 108:05
:
Thus, the zero lies to the rightofs = ;4, so by simple trigonometry we
have
a = 4;
3
tan(180
;108:05
)
= 4;
3
tan(71:95
)
= 3:022:
Then
K
pd
=
p
3
2
+3:9
2
p
3
2
+3
2
p
3
2
+16
2
p
3
2
+0:978
2
=
4:924:2416:28
3:155
= 107:7:
Finally,
K
d
= 3:022K
pd
= 3:022108
= 325:5:
b
The closed loop transfer function is of the form
T
c
(s)=
108(s+3:022)
(s +4;j3)(s+4+j3)(s+
)
2
The simplest waytond
is just to searchalong the real axis to the right
of s = ;20 until wendagain of 108. Try s = ;16.
Then
K =
(4)(15:9)(15)
12:978
= 73:51:
Too small try s = ;14
K =
(6)(13:9)(13)
10:978
= 98:76
Close, try s = ;13
K =
(7)(12:9)(12)
9:978
= 108:6
Just a little too large, try s = ;13:1
K =
(6:9)(13)(12:1)
10:078
= 107:7
So wesettle for third pole at s = ;13:1and
T
c
(s)=
108(s+3:022)
(s +4;j3)(s+4+j3)(s+13:1)
The following MATLAB dialogue locates the third pole exactly and then
nds the unit step response shown in Figure
EDU>g= zpk([-3.022],[-0.1 -1 -20],108)
Zero/pole/gain:
108 (s+3.022)
--------------------
(s+0.1) (s+1) (s+20)
EDU>h = 1
h=
3
Time (sec.)
Amplitude
Step Response
0 0.5 1 1.5 2 2.5 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1
EDU>tc = feedback(g,h)
Zero/pole/gain:
108 (s+3.022)
--------------------------------
(s+13.07) (s^2 + 8.033s + 25.13)
EDU>step(tc,4)
EDU>step(tc,3)
EDU>print -deps pd4step.eps
EDU>
4