Bode and Nyquist stability criteria, gain and phase margins

The Bode and Nyquist stability criteria, phase- and gain-crossover frequencies, ultimate gain and period from frequency response, and gain and phase margins, including loops with dead time.

Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

Routh tells you whether a loop is stable; gain and phase margins tell you how far it is from instability, which is what matters when process gain drifts, valves stick and models are wrong. Frequency-response stability tests also handle dead time exactly, so for most real process loops they are the practical way to find the ultimate gain and set safe controller gains.

Key ideas

Open-loop frequency response. All tests use the open-loop transfer function G_OL(s) = G_c·G_v·G_p·G_m with s = jω, i.e. the loop broken just before the comparator. Its amplitude ratio AR_OL and phase φ_OL come from the Bode rules of the previous topic.

Bode stability criterion. For open-loop stable systems whose phase crosses −180° only once:

  • Find the phase-crossover (critical) frequency ω_co, where φ_OL = −180°.
  • If AR_OL(ω_co) < 1 the closed loop is stable; if AR_OL(ω_co) > 1 it is unstable; AR_OL(ω_co) = 1 means sustained oscillation at ω_co.
  • Physical idea: a signal going around the loop at ω_co is inverted by the −180° lag and inverted again by the comparator, so it returns in phase; if it also returns larger, it grows.

Ultimate gain and period. For proportional control, AR_OL = K_c·AR_p, so the ultimate gain is K_cu = 1/AR_p(ω_co) (with AR_p including valve and sensor gains) and the ultimate period is P_u = 2π/ω_co. These are the same values Routh gives for rational transfer functions, but the Bode method also works with dead time.

Gain margin (GM). GM = 1/AR_OL(ω_co). It is the factor by which the loop gain could increase before instability; in dB, GM_dB = −20·log₁₀AR_OL(ω_co). Typical design: GM = 1.7–2.0 (about 4.6–6 dB).

Phase margin (PM). Find the gain-crossover frequency ω_g, where AR_OL = 1. PM = 180° + φ_OL(ω_g). It is the extra phase lag (for example, extra dead time) the loop could tolerate. Typical design: PM = 30–45°. Extra dead time Δθ that would destabilise the loop is about PM (in rad)/ω_g.

Nyquist criterion. The Nyquist (polar) plot traces G_OL(jω) in the complex plane as ω goes from −∞ to +∞. With P open-loop poles in the right half-plane and N clockwise encirclements of the point (−1, 0), the number of unstable closed-loop poles is Z = N + P. For the usual open-loop-stable process (P = 0), the closed loop is stable if the plot does not encircle −1. Nyquist is more general than Bode: it handles open-loop unstable processes and phase curves crossing −180° several times. On the polar plot, GM is 1/(distance from origin to where the curve crosses the negative real axis), and PM is the angle between the negative real axis and the point where the curve crosses the unit circle.

Practical reading. Increasing K_c lifts the AR curve without changing phase, so it lowers GM and PM. Integral action adds phase lag at low frequency (lowers margins); derivative action adds lead (raises PM). Dead time lowers ω_co sharply and therefore K_cu.

Formulas

φ_OL(ω_co) = −180°, GM = 1/AR_OL(ω_co), GM_dB = −20·log₁₀ AR_OL(ω_co)

AR_OL(ω_g) = 1, PM = 180° + φ_OL(ω_g)

K_cu = 1/AR_p(ω_co), P_u = 2π/ω_co

  • P control; AR_p is the AR of G_v·G_p·G_m; ω_co in rad per unit time.

Z = N + P

  • Nyquist: Z unstable closed-loop poles, N clockwise encirclements of (−1, 0), P open-loop right-half-plane poles.

φ = −ω·θ − tan⁻¹(ω·τ), AR = K/√(1 + ω²·τ²)

  • FOPDT process K·e^(−θs)/(τs + 1).

Worked examples

Example 1 (standard): margins of a three-lag loop. G_OL = K_c/((s + 1)(2s + 1)(4s + 1)), time in min. Find K_cu, P_u, and the GM and PM for K_c = 4.

  1. Phase crossover: tan⁻¹ω + tan⁻¹2ω + tan⁻¹4ω = 180°. By trial ω_co = 0.935 rad/min (43.1° + 61.9° + 75.0° = 180.0°).
  2. AR_p(ω_co) = 1/(1.369·2.121·3.873) = 1/11.25, so K_cu = 11.25 and P_u = 2π/0.935 = 6.72 min (the same as the Routh result).
  3. GM at K_c = 4: GM = K_cu/K_c = 11.25/4 = 2.81, i.e. 20·log₁₀2.81 = 8.98 dB.
  4. Gain crossover: 4/(√(1 + ω²)·√(1 + 4ω²)·√(1 + 16ω²)) = 1. By trial ω_g = 0.542 rad/min (1.137·1.474·2.386 = 4.000).
  5. φ_OL(ω_g) = −(28.4° + 47.3° + 65.2°) = −140.9°; PM = 180 − 140.9 = 39.1°.

K_cu = 11.25, P_u ≈ 6.72 min; at K_c = 4: GM ≈ 2.81 (8.98 dB), PM ≈ 39°

Example 2 (GATE level): first-order-plus-dead-time process. G_p = 2·e^(−s)/(5s + 1), time in min, with P control and unity valve and sensor gains. Find ω_co, K_cu, P_u and the K_c giving GM = 2.

  1. Phase condition: ω·1 + tan⁻¹(5ω) = π rad.
  2. Trial: ω = 1.6 gives 1.600 + 1.446 = 3.046; ω = 1.7 gives 1.700 + 1.454 = 3.154. Interpolate: ω_co ≈ 1.689 rad/min (96.8° + 83.2° = 180.0°).
  3. AR_p(ω_co) = 2/√(1 + (5·1.689)²) = 2/8.503 = 0.2352.
  4. K_cu = 1/0.2352 = 4.25; P_u = 2π/1.689 = 3.72 min.
  5. GM = 2: K_c = K_cu/2 = 2.13.

ω_co ≈ 1.69 rad/min; K_cu ≈ 4.25; P_u ≈ 3.72 min; K_c ≈ 2.13 for GM = 2

Common mistakes

  • Using the closed-loop transfer function instead of the open-loop one in Bode or Nyquist tests.
  • Swapping the two crossover frequencies: GM is read where the phase is −180°, PM where AR = 1.
  • Writing PM = −180° − φ instead of 180° + φ (a phase of −150° gives PM = +30°).
  • Forgetting the valve and transmitter gains in AR_p when computing K_cu.
  • Applying the simple Bode criterion to open-loop unstable systems; use Nyquist (Z = N + P).
  • Mixing radians and degrees in the dead-time phase term.

For GATE CH

This topic regularly produces numericals: the crossover frequency of a loop with dead time (by trial), the ultimate gain and period, gain margin (as a ratio or in dB) for a given K_c, the K_c for a specified GM, and phase margin. Conceptual questions test the Nyquist encirclement rule. Practise solving tan⁻¹ sums equal to 180° quickly and carry four significant figures.

Quick check

  1. The open-loop AR at the phase-crossover frequency is 0.25. What is the GM in dB?
  2. At the gain crossover the open-loop phase is −150°. What is the PM?
  3. A loop has K_cu = 6 and runs at K_c = 2. What is the GM?
  4. Can a first-order process with no dead time under P control have a finite gain margin?

Answers: 1. 12.0 dB (GM = 4). 2. 30°. 3. 3. 4. No: its phase never reaches −180°, so GM is infinite.

Try answering each one aloud before you open it.

  1. 1.What is the Bode stability criterion?Concept

    For an open-loop stable system whose phase crosses −180° once, find the phase-crossover frequency ω_co where the open-loop phase is −180°. The closed loop is stable if the open-loop amplitude ratio there is less than 1, unstable if it is greater than 1, and oscillates continuously if it equals 1. The reason is that at ω_co the loop lag plus the comparator's sign reversal return a signal in phase, so it grows if the loop amplifies it. The ultimate gain follows as K_cu = 1/AR_process(ω_co).

  2. 2.Explain the Nyquist stability criterion.Concept

    The Nyquist plot traces the open-loop G_OL(jω) in the complex plane as ω runs over all frequencies. The number of unstable closed-loop poles is Z = N + P, where N is the number of clockwise encirclements of (−1, 0) and P the number of open-loop poles in the right half-plane. For the usual open-loop stable process (P = 0), the closed loop is stable if the plot does not encircle −1. Unlike the Bode test, it handles open-loop unstable processes and phase curves that cross −180° more than once.

  3. 3.What are gain and phase margins, and why are they important?Concept

    Gain margin and phase margin are measures of the stability of a control system. The gain margin is the factor by which the system gain can be increased before the system becomes unstable, while the phase margin is the additional phase lag required to bring the system to the verge of instability. These margins are important because they provide a quantitative measure of how close the system is to instability, allowing engineers to design systems with adequate stability margins to ensure reliable operation.

  4. 4.How do you determine the gain margin from a Bode plot?Application

    To determine the gain margin from a Bode plot, locate the frequency at which the phase angle is -180 degrees. At this frequency, read the gain from the magnitude plot. The gain margin is the reciprocal of this gain, expressed in decibels (dB). A positive gain margin indicates that the system is stable, while a negative gain margin indicates instability.

  5. 5.Why is the Nyquist plot used in control systems analysis?Application

    The Nyquist plot is used in control systems analysis because it provides a comprehensive view of the system's frequency response, including both magnitude and phase information. It allows engineers to assess the stability of the system by examining the encirclement of the critical point (-1,0) in the complex plane. This makes it a powerful tool for analyzing systems with complex dynamics and for designing controllers that ensure stability.

  6. 6.What happens if the phase margin of a system is zero?Application

    If the phase margin of a system is zero, it means that the system is on the verge of instability. At this point, any additional phase lag will cause the system to become unstable, leading to oscillations or even system failure. Therefore, a zero phase margin is undesirable, and systems are typically designed with a positive phase margin to ensure stability and robustness.

  7. 7.How can you improve the gain margin of a control loop?Application

    The simplest way is to reduce the controller gain, since GM = K_cu/K_c for proportional control. Reducing dead time or measurement lag raises the phase-crossover frequency and the ultimate gain, which also raises the margin at the same K_c. Adding derivative (phase-lead) action pushes the −180° crossing to a higher frequency, and using less integral action (a longer τ_I) removes low-frequency lag.

  8. 8.Calculate the gain margin if the gain at the phase crossover frequency is 0.5.Numerical

    The gain margin is calculated as the reciprocal of the gain at the phase crossover frequency. If the gain is 0.5, the gain margin is 1/0.5 = 2. In decibels, this is 20 * log10(2) ≈ 6.02 dB. A positive gain margin indicates that the system is stable.

  9. 9.A system has a phase margin of 30 degrees. What does this imply about the system's stability?Application

    A phase margin of 30 degrees implies that the system has a moderate level of stability. It means that the system can tolerate an additional 30 degrees of phase lag before reaching the verge of instability. While this indicates that the system is stable, it may not be sufficiently robust for all applications, and further analysis may be needed to ensure adequate performance under varying conditions.

  10. 10.Determine the phase margin if the open-loop phase at the gain crossover frequency is −150 degrees.Numerical

    PM = 180° + φ_OL(ω_g) = 180° + (−150°) = 30°. The loop could tolerate 30° of extra phase lag at the gain-crossover frequency, for example from additional dead time, before oscillating continuously. A PM of 30° is at the low end of the usual 30–45° design range.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?