16.

06 Principles of Automatic Control

Lecture 12

Root Locus Rules

Rule 1 The n branches of the locus start at the n poles of L(s). m branches end at
the zeros of L(s). n m branches end at s 8.
Rule 2
zeros.

The locus covers the real axis to the left of an odd number of poles and

To the left of the pole, 180

To the left of a zero, 180 .

s to the right of the pole

=0o

s to the left of the pole

=180o
To the right of a pole, 0
To the right of a zero, 0 .
So,
m

=Lpsq

i1

i1

m1 180 n1 180
pm1 ` n1 q180 n1 360
180 ` l 360
if m1 ` n1 is o, where
1

m1 number of zeros to the right of s

n1 number of poles to the right of s
Rule 3 For large k, n m of the loci are asymptotic to the lines emanating from the
point s 8, with angles
l

180 ` 360 pl 1q
,
nm

l 1, ...n m

i zi
where pnm
.

Why? If s 8, k 8,then to highest order the equation

dpsq ` knpsq 0
becomes
sn ` ... ` kpsm ` ...q 0
So the solution satises

sn ksm , pk, s 8q
"asymptotic to"

s nm k
1

s pkq nm
1
180 ` 360 pl 1q
k nm =
nm
To get the point s , do asymptotic analysis with next terms:
Result is that center of pattern is at:

pi zi
nm

A related rule, not in FPE, is:

Rule 3a If n m 2, the centroid of the closed-loop poles is constant, and is at

pi
n

To show this, consider a polynomial with roots z1 , z2 , ... The polynomial is then

ps p1 qps p2 q...ps pn q
sn pp1 ` p2 ` ...pn qsn1 ` ...

Therefore, a1 pi .

Now, the closed loop polynomial is given by

dpsq ` knpsq
sn ` a1 sn1 ` ... ` kpsm ` b1 sm1 ` ...q
sn ` a1 sn1 ` ... ` panm ` kqsm ` ...
That is, the rst term to change in the polynomial is the anm term. If n m 2, the a1 term
is unchanged, and the centroid is a constant.
Note that if m poles go to the m zeros zi , the centroid of the remaining n m poles must
go to

pi zi
nm

in agreement with rule 3.

Rule 4 The angle(s) of departure of a branch of the locus from a pole of multiplicity
q is

i i 180 360 pl 1q
qdep

and the angle(s) of arrival of a branch at a zero of multiplicity q is given by

qarr

i ` 180 ` 360 pl 1q

where the sum

excludes

to poles (or zeros) at the point of interest.

Example:

Lpsq

s
ps ` 1qps2 ` 1q

Im(s)
dep=90o-45o-90o-180o
1=90o
1=45o
2=90o

dep 90 45 90 180
dep 225 135 (mod 360 )

Re(s)

MIT OpenCourseWare
http://ocw.mit.edu

16.06 Principles of Automatic Control

Fall 2012

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.