Nyquist

The Nyquist criterion determines closed-loop stability from the open-loop frequency response — even for open-loop unstable plants and systems with time delay. Learn the polar plot, encirclements of −1, and margins.

The Nyquist plot draws the open-loop frequency response L(jω) as a curve in the complex plane: at each frequency, a point with that magnitude and phase. Whether and how that curve circles the critical point −1 tells you, exactly, how many closed-loop poles are unstable.

Unlike Routh or root locus, it works directly with measured frequency-response data and handles time delay exactly.

An analogy

Imagine walking around the point −1 on a map, following the loop's frequency response. If your path winds around −1 (and the plant was stable), the closed loop will be unstable — the path 'traps' the dangerous point.

The polar plot

For each ω from 0 to ∞, plot the complex number L(jω). A simple lag starts on the positive real axis and curls clockwise toward the origin. The mirror image (ω < 0) completes the closed contour.

Encirclements and the criterion

Count how many times the full contour encircles −1 clockwise (N). With P open-loop RHP poles, Z = N + P closed-loop RHP poles. For a stable open loop (P = 0), stability simply requires no encirclement of −1.

Margins on the Nyquist plot

The gain margin is how far you could scale the plot before it reaches −1 (1/|L| where it crosses the negative real axis). The phase margin is the angle between the unit-circle crossing and −1. The closest distance from the curve to −1 equals 1/M_s — a direct robustness measure.

Open-loop unstable plants

If the plant itself has RHP poles (P > 0), the closed loop is stable only if the plot encircles −1 counter-clockwise P times (N = −P). Magnetic levitation, inverted pendulums and some aircraft rely on this, and the Nyquist criterion is the tool that proves stability.

Time delay

A delay rotates each point by −ωT without changing magnitude, so the plot spirals into the origin, crossing the negative real axis infinitely many times. The Nyquist criterion handles this exactly, while pole-based methods need approximations.

Formulas

Worked example: Critical gain of K/(s + 1)³

Given: L(s) = K/(s + 1)³

Find: Phase-crossover frequency and the K for marginal stability.

Answer: Stable for K < 8; oscillation at 1.73 rad/s

Worked example: Open-loop unstable plant

Given: L(s) = K/(s − 1) (P = 1)

Find: K range for stability.

Answer: K > 1 (feedback stabilises the plant)

Common mistakes

Applications

Key takeaways

Going further