Stability via the Nyquist diagram

Range of gain for stability

Problem: For the unity feedback system be-

low, where
K
G(s) = ,
s(s + 3)(s + 5)
find the range of gain, K, for stability, insta-
bility and the value of K for marginal stability.
For marginal stability, also find the frequency
of oscillation. Use the Nyquist criterion.

Figure above; Closed-loop unity feedback sys-

tem.
1
Solution:
K
G(jω) = |s→jω
s(s + 3)(s + 5)

−8Kω − j · K(15 − ω 2)
=
64ω 3 + ω(15 − ω 2)2
When K = 1,
−8ω − j · (15 − ω 2)
G(jω) =
64ω 3 + ω(15 − ω 2)2

Important points:

Starting point: ω = 0, G(jω) = −0.0356 − j∞

Ending point: ω = ∞, G(jω) = 06 − 270◦

Real axis crossing: found by setting the imag-

inary part of G(jω) as zero,
√ K
ω= 15, {− , j0}
120
2
When K = 1, P = 0, from the Nyquist plot,
N is zero, so the system is stable. The real
K does not encircle [−1, j0)]
axis crossing − 120
until K = 120. At that point, the system is
marginally stable, and the frequency of oscilla-
√
tion is ω = 15 rad/s.

Nyquist Diagram
Nyquist Diagram
0.05
2
0.04
1.5
0.03

0.02 System: G 1
Real: −0.00824
Imaginary Axis

Imaginary Axis
0.01 Imag: 1e−005
0.5
Frequency (rad/sec): −3.91
0
0
−0.01
−0.5
−0.02
−1
−0.03

−0.04 −1.5

−0.05 −2
−0.1 −0.08 −0.06 −0.04 −0.02 0 0.02 0.04 0.06 0.08 0.1 −3 −2.5 −2 −1.5 −1 −0.5 0
Real Axis Real Axis

(a) (b)

K
Figure above; Nyquist plots of G(s) = s(s+3)(s+5) ;
(a) K = 1; (b)K = 120.

3
Gain/phase margin via the Nyquist diagram

We use the Nyquist diagram to define two

quantitative measures of how stable a system
is. These are called gain margin and phase
margin. Systems with greater gain margin and
phase margins can withstand greater changes
in system parameters before becoming unsta-
ble.

Gain margin, GM , The gain margin is the change

in open-loop gain, expressed in decibels (dB),
required at 180◦ of phase shift to make the
closed-loop system unstable.

Phase margin, ΦM , The phase margin is the

change in open-loop phase shift, required at
unity gain to make the closed-loop system un-
stable.

4
Figure above; Nyquist diagram showing gain
and phase margins

Problem: Find the gain and phase margin for

the unity feedback system with
6
G(s) = 2
.
(s + 2s + 2)(s + 2)

5
Solution: From G(s), we see P = 0. The
Nyquist diagram shows that N is zero, so the
closed-loop system is stable.
6
G(jω) = 2
|s→jω
(s + 2s + 2)(s + 2)

6[4(1 − ω 2) − jω(6 − ω 2)]

=
16(1 − ω 2)2 + ω 2(6 − ω 2)2
The Nyquist diagram
√ crosses the real axis at a
frequency of ω = 6. The real part is found
to be −0.3.
Nyquist Diagram

1.5

0.5
Imaginary Axis

−0.5

−1

−1.5
−1 −0.5 0 0.5 1 1.5
Real Axis

Figure above; Nyquist diagram for

6
G(s) = 2
.
(s + 2s + 2)(s + 2)
6
The gain margin is

GM = 20 log(1/0.3) = 10.45dB.
To find the phase margin find the frequency
for which G(jω) has a unit gain. Using a com-
puting tool, we can find G(jω) has a unit gain
at a frequency of 1.251, at this frequency the
phase angle is −112.3◦. The difference of this
angle with −180◦ is 67.7◦, which is the phase
margin ΦM .

Gain/phase margin via the Bode plots

Figure above; Gain and phase margins on the

Bode plots.
7
Problem: Let a unit feedback system have
K
G(s) = .
(s + 2)(s + 4)(s + 5)
Use Bode plots to determine the range of gain
within which the system is stable. If K = 200
find the gain margin and the phase margin.

The low frequency gain is found by setting s to

zero. Thus the Bode magnitude plots starts
at K/40. For convenience set K = 40, so
that the log-magnitude plots starts at 0dB. At
each break frequency, 2, 4 and 5, a slope of
-20dB/decade is added.

The phase diagram starts at 0◦ until 0.2rad/s

(a decade below the break frequency of 2),
the curve decreases at a slope of 45◦/decade
at each subsequent frequency at 0.4rad/s and
0.5rad/s (a decade below the break frequency
of 4 and 5 respectively). Finally at 20rad/s,
40rad/s and 50rad/s (a decade above the break
frequencies of 2,4,5), a slope of +20dB/decade
is added, until the curve levels out.
8
Figure above; Bode log-magnitude and phase
40
diagram for G(s) = (s+2)(s+4)(s+5) .

The Nyquist criterion tells us that we want zero

encirclement of {−1, j0}. Thus the Bode log-
magnitude plot must be less than unity when
the Bode phase plot is −180◦. Accordingly we
see that at frequency 7 rad/s, when the phase
plot is −180◦. The magnitude is -20dB.
9
Thus an increase of 20dB is possible before the
system becomes unstable, which is a gain of
10, so the gain for instability is K > 10 × 40 =
400.

If K = 200 (five times greater than K = 40),

the magnitude plot would be 20 log 5 = 13.98dB
higher, as the Bodes plots was scaled to a gain
of 40.

At −180◦, the gain is −20 + 13.98 = −6.02dB,

so GM = 6.02dB.

To find phase margin, we look on the mag-

nitude plot for the frequency where the gain
is 0dB. As the plot should be 13.98dB higher,
so we look at −13.98dB crossing to find the
frequency is 5.5rad/s. At this frequency, the
phase angle is −165◦. Thus

ΦM = −165◦ − (−180)◦ = 15◦.

10