The Nyquist Stability Criterion
Born in 1889 in Sweden
Died in 1976, USA
Yale PhD, 1917
Career at Bell Labs
138 patents
Nyquist diagram,
criterion, sampling
theorem
Laid the foundation for
information theory, data
transmission and
negative feedback theory
Frequency Response Plots
| G ( ω ) |= Re 2 ( ω ) + Im 2 ( ω )
Im( ω )
φ = tan −1
Re( ω )
(Review Appendix G in textbook)
Frequency Response Plots
1st Order
system
Im( ω ) − ω / ω1 ω
φ (ω ) = tan −1 = tan −1 = − tan −1 ( )
Re( ω ) 1 ω1
Polar Plot or Nyquist Diagram
K
G (s) = 2rd system
s ( s τ + 1)
Im(ω) −1 − Kω −1 1
φ (ω) = tan
−1
= tan = tan ( )
Re(ω) − Kω τ
2
ωτ
Polar Plots and the Nyquist Criterion
• We have seen how to generate polar
plots of the frequency response.
• The Nyquist criterion is a way of
interpreting the polar plot to determine
stability
• The frequency response in the entire
right-half s- plane with s=-j∞ to s=+j∞
is the critical section of the Nyquist contour.
Mapping Contours in the s-Plane
• Nyquist stability criterion based on
Cauchy theorem on functions of a
complex variable
• Mapping contours
1+L(s)=0
in the s-plane.
Mapping Contours in the s-Plane
Closed contours
F(s)=2s+1
Conformal mapping (angles are preserved)
Mapping Contours in the s-Plane
S-plane F(s)-plane
Cauchy’s Theorem: Principle of the
Argument
If a contour in the s-plane encircles Z zeros and P
poles of F(s ) and does not pass through any poles or
zeros of F(s ) and the traversal is in the clockwise
direction along the contour, the corresponding contour
in the F(s)-plane encircles the origin of the F(s)-plane
N=Z-P times in the clockwise direction
?
N=1-1=0
S-plane F(s)-plane
Illustration of Cauchy’s Theorem
N=Z-P=3-1=2
Illustration of Cauchy’s Theorem
N=Z-P=0-1=-1
The Nyquist Criterion
The Nyquist contour encircles the entire right-half s- plane
For the system to be stable, all
zeros (or roots) of F(s ) must lie in
the left-hand s-plane.
N=Z-P
The number of unstable zeros of F(s)
is thus
Z=N+P
Where N is the number encirclements
of the origin.
When the system is open- loop stable
P=0 and then Z=N
The Nyquist Criterion
In practice it is more convenient to
consider the mapping
in which case becomes the number
of clockwise encirclements of the -1
point.
The Nyquist contour encircles the entire right-half s- plane
Z is unstable zero and P is unstable pole if they are lied within
Nyquist contour.
The Nyquist Criterion
Z=N+P For stable system, the closed loop Z=0.
1) Open-loop stable plant Z=0 => N=0
A feedback system is stable if and only if the contour
in the -plane does not encircle the (-1,0) point.
2) Open-loop unstable plant Z=0 => N=-P
A feedback system is stable if and only if, for the
contour in the -plane, the number of
counterclockwise of the (-1,0) point is equal to the
number P of poles of with positive real parts.
System with Two Real Poles
K 100
GH ( s ) = =
(τ 1s + 1)(τ 2 s + 1) ( s + 1)(0.1s + 1)
Stable
hence for stability we require Z=N=0,
i.e. the contour must not encircle the -1 point in the GH(s)-plane.
System with a Pole at the Origin
R=ε >0
The Cauchy theorem requires
Modified or indented
that the contour cannot pass
Nyquist contour
through the pole
System with a Pole at the Origin
Small circular detour represented by where
varies from
for this portion is an infinite semi-circle
going from
and passing through
System with a Pole at the Origin
Portion from
Magnitude is zero when r is infinite
System with a Pole at the Origin
R=ε >0
Stable
The Cauchy theorem requires
Modified or indented
that the contour cannot pass
Nyquist contour
through the pole
System with Three Poles
System with Three Poles
The system is stable or not depending on the k value.
System with Three Poles
System with Two Poles at the Origin
Angle of
Unstable for all K
N=2 hence Z=2, i.e. two roots in the RHP
System with a Pole in the RHP
P=1, hence we require N=-P=-1
for stability.
With
Infinite semicircle in the LHP
With
Two roots in the RHP
System with a Pole in the RHP
Adding derivative action
When
Stable
Crosses real axis when
System with a Zero in the RHP
-2K
Unstable for K>1/2
Relative Stability
Ultimate point
crosses the real axis at
u=-1 when
Relative Stability and Nyquist Criterion
• Clearly, from the Nyquist criterion the
distance between the (-1,0) point and
the polar plot of the loop transfer
function GH(jω) is a measure of the
relative stability of the system
• Note that the (-1,0) point on a Nyquist
diagram corresponds to the 0 dB and -
180 degrees on the Bode magnitude
and phase diagrams
Relative Stability
• The gain margin is the increase in the
0
system gain when phase =- 180 that
will result in a marginally stable
system with intersection of the -1+j0
on the Nyquist diagram
• The phase margin is the amount of
phase shift of the GH Nyquist plot at
unity magnitude that will result in a
marginally stable system with
intersection of the -1+j0 point on the
Nyquist diagram
Gain Margin
The factor by which the gain at ultimate point can be
increased to reach marginal stability
Reasonable values are
gain margins in the range
Expressed in dB: [2 5], or in dB [6 14]dB.
Phase Margin
The phase margin is the difference between and
the phase of the loop at the crossover frequency
Reasonable values are phase
margins in the range [30 60]
degrees
Gain and Phase Margins on the Bode Diagram
Sensitivity and Nyquist Plot
The radius r of the circle centered at -1
and tangen to L is the reciprocal of the
nominal sensitivity peak
Critical point
Effect of Time Delay
• The Nyquist criterion is applicable to
non-rational transfer functions
• A time delay is described by
• It does not introduce additional poles
or zeros within the contour
• However delay introduces phase lag
and consequently has a destabilizing
effect.
Effect of Time Delay
adds a phase shift
At the crossover frequency
might be sufficient to destabilize the system or
to significantly reduce the phase margin
Effect of Time Delay
Phase margin
for delay-free
system
Delayed
system is
unstable!
System with Delay
Consider the following delayed system:
With a second-order Pade approximation, the delay
can be written as
Hence
System with Delay
K = 2.41
System with Delay
K=3
K=2
K=1