bappM_epdf 12/18/2018 21:27:38 Page 1
Appendix M
Root Locus Rules: Derivations
M.1 Derivation of the Behavior of the Root Locus
at Infinity (Kuo, 1987)
Let the open-loop transfer function be represented as follows:
K sm a1 s m 1
am
KG s H s M1
sm n b1 s m n 1 bm n
or
K
KG s H s m n m n 1
M2
s b1 s bm n
sm a1 s m 1 am
Performing the indicated division in the denominator, we obtain
K
KG s H s M3
sn b1 a1 s n 1
In order for a pole of the closed-loop transfer function to exist,
KG s H s 1 M4
Assuming large values of s that would exist as the locus moves toward infinity,
Eq. (M.3) becomes
sn b1 a1 s n 1
K M5
Factoring out sn, Eq. (M.5) becomes
b1 a1
sn 1 K M6
s
Taking the nth root of both sides, we have
1 n
b1 a1
s 1 K1 n
M7
s
A-172
bappM_epdf 12/18/2018 21:27:39 Page 2
M.1 Derivation of the Behavior of the Root Locus at Infinity (Kuo, 1987) A-173
If the term
1 n
b1 a1
1 M8
s
is expanded into an infinite series where only the first two terms are significant,1 we obtain
b1 a1 1 n
s 1 K M9
ns
Distributing the factor S on the left-hand side yields
b1 a1 1 n
s K M 10
n
1 n 1π n
Now, letting s = σ + jω and K K1 n
e j 2k , where
1 n 1π n 2k 1π 2k 1π
1 e j 2k cos j sin M 11
n n
Eq. (M.10) becomes
b1 a1 2k 1π 2k 1π
σ jω K1 n
cos j sin M 12
n n n
where k = 0, ± 1, ± 2, ± 3, . . . Setting the real and imaginary parts of both sides equal to
each other, we obtain
b1 a1 2k 1π
σ K1 n
cos M 13a
n n
2k 1π
ω k1 n
sin M 13b
n
Dividing the two equations to eliminate |K1/n|, we obtain
b1 a1 2k 1π
σ cos
n n
M 14
ω 2k 1π
sin
n
Finally, solving for ω, we find
2k 1π b1 a1
ω tan σ M 15
n n
The form of this equation is that of a straight line,
ω M σ σ0 M 16
where the slope of the line, M, is
2k 1π
M tan M 17
n
1
This is a good approximation since s is approaching infinity for the region applicable to the derivation.
bappM_epdf 12/18/2018 21:27:39 Page 3
A-174 Appendix M Root Locus Rules: Derivations
Thus, the angle of the line in radians with respect to the positive extension of the real axis is
2k 1π
θ M 18
n
and the σ intercept is
b1 a1
σ0 M 19
n
From the theory of equations,2
b1 finite poles M 20a
a1 finite zeros M 20b
Also, from Eq. (M.1),
n number of finite poles number of finite zeros
M 21
#finite poles #finite zeros
By examining Eq. (M.16), we conclude that the root locus approaches a straight line
as the locus approaches infinity. Further, this straight line intersects the σ axis at
finite poles finite zeros
σ0 M 22
#finite poles #finite zeros
which is obtained by substituting Eqs. (M.20)
Let us summarize the results: The root locus approaches straight lines as asymptotes
as the locus approaches infinity. Further, the equation of the asymptotes is given by the real-
axis intercept and the angle with respect to the real axis as follows:
finite poles finite zeros
σ0 M 23
#finite poles #finite zeros
2k 1π
θ M 24
#finite poles #finite zeros
where k = 0, ± 1, ± 2, ± 3, . . . Notice that the running index, k, in Eq. (M.24) yields a
multiplicity of lines that account for the many branches of a root locus that approach infinity.
M.2 Derivation of Transition Method
for Breakaway and Break-in Points
The transition method for finding real-breakaway and break-in points without differentiat-
ing can be derived by showing that the natural log of 1/[G(σ)H(σ)] has a zero derivative at
the same value of σ as 1/[G(σ)H(σ)] (Franklin, 1991).
We now show that if we work with the natural log we can eliminate the step of
differentiation.
2
Given an nth-order polynomial of the form sn + an–1sn–1 + ∙ ∙ ∙ , the coefficient, an–1, is the negative sum of the
roots of the polynomial.
bappM_epdf 12/18/2018 21:27:39 Page 4
Bibliography A-175
First find the derivative of the natural log of 1 [G(σ)H(σ)] and set it equal to zero.
Thus,
d 1 d 1
ln Gσ H σ 0 M 25
dσ Gσ H σ dσ G σ H σ
Since G(σ)H(σ) is not zero at the breakaway or break-in points, letting
d 1
ln 0 M 26
dσ Gσ H σ
will thus yield the same value of σ as letting
d 1
0 M 27
dσ G σ H σ
Hence,
d 1 d σ p1 σ p2 σ pn
ln ln
dσ Gσ H σ dσ σ z1 σ z2 σ zm
d
ln σ p1 ln σ p2 ln σ pn
dσ
ln σ z1 ln σ z2 ln σ zm
1 1 1 1 1
σ p1 σ p2 σ pn σ z1 σ z2
1
0
σ zm M 28
or
n 1 m 1
M 29
i 1σ pi i 1σ zi
where zi and pi are the negatives of the zero and pole values of G(s)H(s), respectively.
Equation (M.29) can be solved for σ, the real axis values that minimize or maximize K,
yielding the breakaway and break-in points without differentiating.
Bibliography
Franklin, G. F., Powell, J. D., and Emami-Naeini, A. Feedback Control of Dynamic Systems, 2d ed.
Addison-Wesley, Reading MA, 1991.
Kuo, B. Automatic Control Systems, 5th ed. Prentice-Hall, Englewood Cliffs, NJ, 1987.