Solution 10.8.7.4
The characteristic equation is
1+GH(s)=0:
If we let
GH(s)=
(s)
(s)
;;
then we can write
1+
(s)
(s)
=0;;
or
(s)+(s)
(s)
=0:
Clearly,another waytowrite the characteristic equation is
(s)+(s)=0:
We next divide both sides of
1+GH(s)=0
by GH(s)toobtain
1
GH(s)
+1=0:
We then rewrite this last expression using
GH(s)=
(s)
(s)
;;
to obtain
1
GH(s)
+1 =
1
(s)
(s)
+1
=
()
(s)
+1
=
(s)+(s)
(s)
:
Thus, we see that, as must be the case,
1+GH(s)=0;;
1
j Im(s)
Re(s)
III
I*
r =
8
Figure 1: Root locus
and
GH(s)
;1
+1=0
are simply dierentways of writing the characteristic equation. If we use
GH(s)
;1
+1=0;;
Then wesimply reverse the roles of the poles and zeros of GH(s), but ev-
erything else remains the same.
Now consider
GH(s)=
K(s+3)
(s;1)(s+1)
Then
GH(s)
;1
=
(s;1)(s;1)
K(s+3)
Wenow draw the Nyquist plot for the contour shown in Figure 1 The Bode
phase and magnitude plots are shown in Figure 2. SInce GH has on pole
in the righthalfplane, GH
;1
will have one zero in the righthalf plane and
as a result the phase goes from 180
to 90
as ! goes from zero to innity.
Since GH
;1
has twozeros and one pole its magnitude will increase without
bound as ! !1. Thus, the Nyquist plot for part I of the
contour is
shown in Figure 3
As weevaluate GH
;1
along the semicircle of ininite radius, the mag-
nitude of GH
;1
remains innite, but the angle rotates through 180
in the
clockwise direction. This is because wehavetwozeros contributing 180
2
10
-1
10
0
10
1
10
2
10
3
-60
-40
-20
0
20
40
10
-1
10
0
10
1
10
2
10
3
80
100
120
140
160
180
200
Figure 2: Bode magnitude and phase plots
3
Re(s)
Im(s)
Figure 3: B GH(I)
4
Re(s)
Im(s)
_
8
a
III
Figure 4: Completed Nyquist plot
eachintheclockwise direction and a pole contributing ;180
in the clock-
wise direction. The evaluation of GH
;1
along I
is just the mirror image of
GH(I). The completed Nquist plot is shown in Figure 4.
Wenowcan complete the analysis. One thing to remember. The gain
K is in the denominator so that as we increase K the Nyquist plot shrinks
and as wedecrease the gain the Nyquist plot expands. The value of gain K
that makes GH
;1
=1can be determined from the Bode plots. The phase
starts at 180
and goes to 90
Thus the only place where the Nyquist plot
touches the real axis is for s =0.Itiseasy enough to see that for s =0
jGH(s)
;1
(0)j =
js;1jjs+1j
js+3j
s=0
= 0:3333:
The Bode magnitude plot was drawn for K =100,and we see from the
5
Bode magnitude plot
jGH
;1
(j0)j;50 dB = 0:003162
That is
js;1jjs+1j
100js+3j
s=0
=0:003162:
Dividing both sides by0.03162 weobtain
js;1jjs+1j
0:3162js+3j
s=0
=1
Thus if If we decrease the gain byafactor of
1
0:003162
=316:23
the point`a' will pass through the point(;1;;0) in the GH plane. The value
for K wehave obtained from the Bode plots, namely 0.3162, is pretty close
to the precise value of one third. Remember we are working with the inverse
of the loop transfer function and K is in the denominator.
Thus, fore 0 <K<0:3333: the point(;1;;0) will be region I. For
0:333K<1,thepoint(;1;;0) will be region II.
Since GH(s)
;1
has no poles inside the righthalf plane, in region I the
Nyquist equation is
Z = N +P
= 1+0
= 1
Thus there is one closed loop pole in the righthalfplane.
In region II the Nyquist equation is
Z = N +P
= 0+0
= 0;;
and there are no closed loop poles in the right half plane.
Thus, the the system is stable for K>0:03333. This can, of course, be
seen from the root locus shown in Figure 5.
6
-8 -7 -6 -5 -4 -3 -2 -1 0 1 2
-3
-2
-1
0
1
2
3
Real Axis
Imag Axis
Figure 5: Root locus
7