Solution 5.8.1.3
G
H
+
R
CΣ
Figure 1: Standard Closed Loop Conguration
For the system of Figure 1 wehave
GH =
K
(s +3)
3
:
The rst step is to plot the poles and zeros of GH.Thepoles of GH are
not the closed loop pole locations, but they can be used to nd the closed
loop poles. The closed loop zeros can be found immediately: they are the
zeros of G and the poles of H.Thezeros of GH also help in nding the
poles of the closed loop system.
The root locus is shown in Figure 2. The root locus on the real axis
occurs only to the left of the triple pole at s = -3. The pole zero excess is
p =3;0=3:
The number of asymptotes is equal to p and hence there will be three asymp-
totes at
`
=
(1+ 2`)
p
180
` =0;;1;;2
The three asymptotes intersect the real axis at:
i
=
P
real part of poles of GH ;
P
real part ofzeros of GH
p
=
;3;3;3;0
3
= ;3
Since the asymptotes meet at the triple pole, wecan see that along the
entire length of eachofthe asymptotes the sum of the angles of the three
poles must equal ;180
.Thus, the root locus is just the three asymptotes.
1
Re(s)
Im(s)
3
Figure 2: Rootlocus
We can verify this bycomputing
dK
ds
=
d
ds
h
;(s +3)
3
i
= ;3(s +3)
2
Hence, the break-out occurs at s = ;3.
We can nd the gain that places closed loop pole on the imaginary axis
in a number of ways. Analytically,wecould write out the denominator of
the closed loop transfer function in unfactored form to obtain:
1+GH = 1+
K
(s +3)
3
=
s
3
+9s
2
+9s +27
(s +3)
3
Toobtain 1 + GH =0requires that
s
3
+9s
2
+27s +27+K =0
Setting s = j! in this last equation yields
;j!
3
;9!
2
+ j27! +27+K = (27+ K;9!
2
)+j(;!
3
+27!)
2
If this expression is to equal 0+ j0wemust have
27 + K ;9!
2
=0
and
;!
3
+27! =0
The second equation is easily solved, yielding
! =
q
(27)
= 3
q
(3)
Substituting into the rst equation then yields
K = 9!
2
;27
= 216
These calculations are satisfying in the sense that they are precise. As a
general rule, even for more complicated transfer functions, the equations are
usually reasonably easy to solve. However, they obscure the connection with
Nyquist theory.For that reason, weoerthefollowing alternativeanalysis
which students usually nd unappealing because of its imprecise nature.
We know that at every pointonthe root locus wehave
GH =1
6
;180
Thus, at those points where the root locus crosses the imaginary axis, the
angle condition must be satised. In the presentcase it is easy to see that
weautomatically havethe angle condition satised all along the asymptotes,
so that
! = 3tan(60
)
= 3(
p
3)
= 5:196
The resulting gain to place poles at s =0j3
p
(3) is simply the cube of
the vector drawn from s = ;3tos =0+j3
p
(3) or
K =
h
(3
2
+(3
p
3)
2
)
1=2
i
3
= 6
3
= 216
These same basic ideas can be applied to more complicated examples with
only a small increase in computational eort. Further, this approachis
consistentwith understanding the Nyquist analysis.
3