NYQUIST STABILITY CRITERION

Nyquist Stability Criterion(NSC)

Relates the location of the roots of characteristics equation to

open loop frequency response G ( j ) H ( j ) of a system
Unlike root-locus technique, computation of closed loop poles is
not necessary in NSC.
Stability study can be carried out graphically from open loop
frequency response.
Experimentally determined open loop frequency response can
be used directly for the study of stability when the feedback
path is closed(i.e. closed system stability)

Mathematical Preliminaries

Consider a function q( s )
(s 1 )(s 2 )L(s m )
q(s) =
(s 1 )(s 2 )L(s n )

(1)

S=complex variable
q ( s ) = also complex variable


Eq.1 indicates theta for every point s in the s-plane at which

q(s) is analytic (function and all its derivatives exist) we can
find a corresponding point q(s) in the q(s)-plane.
Since any number of points in s-plane can be mapped into
q(s)-plane, for a contour in the s-plane which does not go
through any singular points (points in the s-plane where the
function or its derivatives does not exist), there corresponds a
contour in the q(s)-plane.

Mathematical Preliminaries

Fig.1

The region to the right of a closed-contour is considered enclosed by the

contour when the contour is traversed in clockwise direction. Thus, the shaded
area in Fig.1 is enclosed by the closed contour.
We are not interested in exact shape of the q(s)-plane contour. But, it is
important to know the encirclement of the origin by q(s)-plane contour for NSC

To investigate the encirclement:

Consider an s-plane contour which encloses only one of the zeros of q(s), say s = 1, while
all the poles and remaining zeros are distributed in the s-plane outside the contour.

Mathematical Preliminaries

For any non-singular point s on the s-plane contour, there corresponds a

point q(s) on q(s)-plane contour.
( s 1 ) ( s 2 ) L ( s m )
From (1), the point q(s) is given by q( s) =

( s 1 ) ( s 2 ) L ( s n )
q(s) = (s 1) +(s 2 ) +L(s 1) (s 2 ) L

Fig2

From Fig.2(a) it is found that as the point s follows the prescribed path ( i.e.
clockwise direction on the s-plane contour, eventually returning to the
starting point, the phasor ( s 1 ) generates a net angle of 2 ,while all
other phasors generate net angles.

Mathematical Preliminaries

This implies tip of q(s)-phasor must describe a closed contour

about the origin of the q(s)-plane in clockwise direction.
As said before, exact shape of the closed contour in q(s)-plane
is not important for us rather it is sufficient to observe that this
contour encircles origin once.
If the contour in the s-plane is so chosen that it does not enclose
any zero or pole, the corresponding contour in q(s)-plane then
it will not encircle the origin.

Mathematical Preliminaries

If s-plane contour encloses 2 zeros (say s = 1 , s = 2 ) , q(s)-plane contour

encircles origin twice in clockwise direction.
Fig.3

We can say, that for each zero of q(s), enclosed by s-plane

contour, the corresponding q(s)-plane contour encircles origin once
in CW direction.

Mathematical Preliminaries

Consider now the enclosure of a pole of q(s), say s = 1 by s-plane contour.

Phasor ( s 1 ) generates an angle 2 as s traverses the prescribed path.
Since ( s 1 ) is in denominator of Eq.(1), the q(s)-plane contour experiences
an angle change of 2 , which means one contour CW encirclement of the
origin.

Mathematical Preliminaries

This argument holds for other poles of q(s). Thus, if there are P poles and Z zeros of
q(s) enclosed by s-plane contour, then the corresponding q(s)-plane contour must
encircle the origin Z times in CW direction and P times in CCW direction resulting in
a net-encirclement of the origin (p-z) times in CCW direction.
This relation between enclosure of poles and zeros of q(s) by the s-plane contour
and encirclement of origin by q(s)-plane = principle of argument.
Ex: 1 zero and 3 poles enclosed by the s-plane contour, the net encirclement of the
origin by the q(s)-plane contour is 2 (3 1) = 4 rad i.e. two contour CCW
revolutions.

Fig.4

Nyquist Stability Criterion

Consider Single-loop feedback system

(3)

Ch. Eq: q( s) = 1 + G ( s) H ( s) = 0

Open loop transfer function (pole zero form):

G ( s) H ( s) =

K ( s + z1 )( s + z2 )L( s + zm )
( s + p1 )( s + p2 )L( s + pn )

From (3) and (4)

(4)
q( s ) = 1 + K

( s + z1 )( s + z2 )L( s + zm )
( s + p1 )( s + p2 )L( s + pn )

Nyquist Stability Criterion

q( s ) = 1 + K
=

( s + z1 )( s + z2 )L( s + zm )
( s + p1 )( s + p2 )L( s + pn )

( s + p1 )( s + p2 )L( s + pn ) + K ( s + z1 )( s + z2 )L( s + zm )
( s + p1 )( s + p2 )L( s + pn )

( s + z11 )( s + z12 )L( s + z1n )

=
( s + p1 )( s + p2 )L( s + pn )

(5)

(6)

From (6) Zeros of q(s) at z11 , z12 ,L z1n are the roots of Ch.Eq. Closed loop
poles of q(s) at p1 , p2 ,L pn are the same as the open loop poles of the
system.
For a system to be stable, the roots of Ch.Eq and hence zeros of q(s) must
lie in LHSP
Note: Even if some of OL poles lie in RHSP, all the zeros of q(s) i.e. CL poles
may lie in LHSP meaning there by that an OL unstable system may lead to
a CL stable operation.

Nyquist Stability Criterion

In order to investigate the presence of any zero of q(s)=1+G(s)H(s) in the

RHSP, let us choose a contour which completely encloses the RHSP. Such a
contour C is called Nyquist Contour (Fig.6)
Fig.6

It is directed CW and comprises of an infinie segment C1 along

axis
j
and an arc C2 of infinite radius.

Along C1: s = j with s varying from j to j

Along C2: s = Re j with varying from to 0 to
2

Nyquist Stability Criterion

The Nyquist contour so defined encloses all the right half s-plane zeros and
poles of q( s ) = 1 + G ( s ) H ( s )
Let there be Z zeros and P poles of q(s) in the RHSP.
As S moves along the Nyquist contour in the s-plane, a closed contour is
traversed in the q(s)-plane which encloses the origin N=P-Z times in CCW
direction.
In order the system to be stable, there should be no zeros of
q(s)=1+G(s)H(s) in the RHSP i.e. Z=0.
This condition is met if N=P, i.e. for a CL system to be stable, the no. of
CCW encirclements of the origin of the q(s)-plane by the contour should be
equal the no. of right half s-plane poles of q(s) which are the poles of the
OL TF of G(s)H(s).

Nyquist Stability Criterion

In special case of p=0 (OL stable system), the CL system is stable if N=P=0
which means that net encirclement of the origin of the q(s)-plane by contour
should be zero.
It is easily observer that

G ( s ) H ( s ) = [1 + G ( s ) H ( s ) ] 1

(9)

From 9: Contour GH of G(s)H(s) corresponding to Nyquist contour is s-plane

is the same as contour q of 1+G(s)H(s) drawn from the point (1 + j 0)
Thus, the encirclement of the origin by the contour q is equivalent to the
encirclement of the point (1 + j 0) by the contour GH (Fig.7)

Fig7

Nyquist Stability Criterion

If the contour GH of the OL TF G(s)H(s) corresponding to the Nyquist contour

in the s-plane encircles the point (1 + j 0) in the CCW direction as many times
as the no. of RHSP poles of G(s)H(s), the CL system is stable.
The CL system is stable if the contour GH of G(s)H(s) does not encircle (1 +
point, i.e. the net encirclement is zero.

j 0)

Mapping of Nyquist contour into contour GH is carried out as follows :

Mapping of Imaginary axis: Substitute j axis in G(s)H(s). This converts mapping

function into a frequency function of G ( j ) H ( j )

In physical systems (m n)
limj G ( s) H ( s ) =
s = Re

real constant

Thus the infinite arc of Nyquist contour maps into a point on the real-axis.
Complete contour GH is thus the polar plot of G ( j ) H ( j ) with varying from to
Nyquist plot is symmetrical about the real axis since
G * ( j ) H * ( j ) = G ( j ) H ( j )

Example-1:

Consider a feedback system whose OL TF is

G ( s) H ( s) =

G ( j ) H ( j ) =
G ( j ) H ( j ) =

K
(T1s + 1)(T2 s + 1)

K
(1 + jT1 )(1 + jT2 )

(1)

K
2

2
1

Fig.7
2

(1 + T ) (1 + T2 )

G ( j ) H ( j ) = tan 1 T1 tan 1 T2

lim G ( j ) H ( j ) = K , lim GH = 0

lim GH = 0, lim GH = 1800

Rationalize (1) & separate the real & imaginary parts

K (1 2T1T2 )
(T1 + T2 ) K
K
=

j
(1 + jT1 )(1 + jT2 ) (1 + 2T12 )(1 + 2T22 )
(1 + 2T12 )(1 + 2T22 )

Example-1:
Equating real parts to zero, we get =
GH =

1
T1T2

1
T1T2

KT1T2
T1 + T2

The plot of GH is shown in Figure. The infinite semicircular arc of the Nyquist contour
maps into the origin.
As the point 1 + j 0 is not encircled by the plot, N=0, P=0, Z=0 which implies stable
system.

G ( j ) H ( j ) does not encircle (1 + j 0) point for any +ve values of

Therefore system is stable for all +ve values of

K , T1 & T2

K , T1 & T2

Example-2:
G ( s) H ( s) =

s+2
( s + 1)( s 1)

Determine whether the system is stable when the feedback path is closed.

Fig.8
From OL TF, one pole in RHSP therefore P=1,
G ( j ) H ( j ) locus encircles (1 + j 0) point once on CCW direction therefore N=1=P
Thus Z=0, i.e. there is no zeros of 1 + G ( s ) H ( s ) in RHSP and hence the CL system is stable

OL Poles on the imaginary axis:

If G(s)H(s) and therefore 1+G(s)H(s) has any poles on the imaginary axis, the Nyquist
Contour is defined as

Fig.9

To study the stability in such cases, the Nyquist contour must ne modified so as to bypass
any imaginary axis poles. This is accompanied by indenting the Nyquist contour around
The imaginary axis poles along a semicircle of radius , where 0 . Fig.10

Fig.10