Root Locus for Positive Feedback Systems
Root Locus for Positive Feedback Systems
If feedback is positive then some of the regular rules of root locus
must be modified. The closed-loop transfer function becomes,
KG (s)H(s)
T (s) =
Unit 7: Part 3: Positive Feedback 1 − KG (s)H(s)
The RL is now defined wherever KG (s)H(s) = 1∠k360o where k is
any integer. This affects the following rules:
Engineering 5821:
3. Real-axis segments
Control Systems I
Faculty of Engineering & Applied Science
Memorial University of Newfoundland
March 17, 2010 Poles and zeros off the real-axis still contribute nothing to
∠KG (s)H(s). Also, poles or zeros to the left contribute nothing.
The RL exists for positive-feedback systems to the left of
an even number of real-axis finite open-loop poles or
ENGI 5821 Unit 7: Root Locus Techniques zeros.
.
e.g. A positive feedback system can be obtained from a negative
5. Behaviour at infinity: The formula for the real-axis intercept feedback system if the gain becomes negative. Assume we have
of an asymptote is unchanged, but the angle is different: the following system with negative gain K 0 :
k2π
θa =
#finite poles − #finite zeros
Other rules for the refinement of the RL only need to be modified We can convert this into a positive feedback system with positive
in that the angle criterion for lying on the RL is changed (e.g. for gain K = −K 0 ,
searches of crossings of the jω axis).
We first determine that the RL exists on the real-axis between -4
and -3 and between -2 and -1. It also exists to the right of the pole Thus we can complete the RL:
at 0. What does this mean?
Continuing, we identify asymptotes,
P P
finite poles − finite zeros
σa =
#finite poles − #finite zeros
(0 − 1 − 2 − 4) − (−3)
= = −4/3
4−1
k2π
θa =
#finite poles − #finite zeros
= 0 for k = 0
= 2π/3 for k = 1
= 4π/3 for k = 2
