Glossary term
Stability Margin
A quantitative measure of how far a feedback control system is from the boundary of instability.
Definition
metricA stability margin is a quantitative measure of the distance between the operating point of a feedback control system and the boundary of instability, expressed in terms of allowable gain increase (gain margin) or allowable phase lag (phase margin).
A feedback loop that is nominally stable can become unstable if the loop gain increases or additional phase lag accumulates — due to model uncertainty, component ageing, or operating condition changes. Stability margins quantify how much variation can be tolerated before instability occurs. They are the primary robustness specifications in classical control design and are computed from the open-loop frequency response using Bode plots or Nyquist diagrams.
A feedback control system is designed to be stable: all closed-loop poles lie in the left half of the complex plane and all transient modes decay. However, the model used for design is always an approximation of the real plant. Parameter variations, unmodelled dynamics, and changing operating conditions all perturb the actual loop behaviour relative to the design model. Stability margins quantify how much perturbation the loop can absorb before becoming unstable, providing a practical measure of robustness.
The gain margin GM is defined at the phase crossover frequency \omega_{pc} — the frequency at which the open-loop phase response \angle L(j\omega) reaches -180°. At this frequency, the open-loop transfer function L(j\omega_{pc}) is real and negative; a loop operating at unity gain at this frequency would be on the boundary of instability (Nyquist criterion). The gain margin is the factor by which the loop gain can be multiplied before instability occurs:
Expressed in decibels: GM_\text{dB} = -20 \log_{10} |L(j\omega_{pc})|. A gain margin of 6 dB means the gain can double before the system becomes unstable. Typical design requirements specify GM \geq 6 dB, though demanding applications may require 10–12 dB or more.
The phase margin PM is defined at the gain crossover frequency \omega_{gc} — the frequency at which the open-loop magnitude |L(j\omega)| equals unity (0 dB). It is the additional phase lag that would bring the system to the boundary of instability:
A phase margin of 45° means the open-loop phase at the gain crossover frequency is -135°; an additional 45° of lag would make it -180° and drive the system to the verge of instability. Typical design requirements specify PM \geq 45°. Low phase margin produces a closed-loop response with a pronounced resonance peak near the gain crossover frequency and large overshoot in the step response.
Reading stability margins from the Bode plot
On a Bode plot of the open-loop transfer function L(j\omega):
- The gain margin is read at \omega_{pc}: it is the distance in dB from the magnitude curve to the 0 dB line.
- The phase margin is read at \omega_{gc}: it is the distance in degrees from the phase curve to the -180° line.
For stable open-loop, minimum-phase loops with a single relevant crossover, positive gain and phase margins are expected for closed-loop stability. Multiple crossovers, right-half-plane poles or zeros, delays, sampled-data effects, or unusual loop definitions require a Nyquist or robust-stability check rather than relying on one pair of Bode margins alone.
Relationship to closed-loop performance
Stability margins are not only robustness indicators — they also correlate with time-domain performance. The phase margin is directly related to the damping of the dominant closed-loop poles: a phase margin of 45° corresponds approximately to a damping ratio of \zeta \approx 0.45, and a phase margin of 60° to \zeta \approx 0.59. Higher phase margin means more damping, less overshoot, and a more sluggish response. There is an inherent trade-off: increasing stability margins typically reduces bandwidth and slows the response, while pushing for higher bandwidth reduces stability margins. Controller design is in large part the management of this trade-off.
Limitations
Gain and phase margins characterise robustness to gain and phase perturbations separately. They do not capture robustness to simultaneous gain and phase variations or to structured uncertainty. The H_\infty robust stability framework and the concept of the stability radius provide more comprehensive robustness measures, but gain and phase margins remain the standard specifications in most industrial and aerospace control design.
Common mistakes
A common mistake is accepting a controller because nominal gain and phase margins are positive while ignoring delay, actuator saturation, rate limits, sensor filtering, sample time, and unmodelled high-frequency modes. Another is assuming larger margins are always better; excessive margin can indicate low bandwidth and poor disturbance rejection. A strong stability-margin review states the open-loop definition, feedback sign, crossover frequencies, gain margin, phase margin, delay margin where relevant, operating point, and model uncertainty.