Stability Analysis Interview Questions

Q1. What does it mean for a control system to be stable?

A control system is stable if every bounded input produces a bounded output — this is called BIBO stability — meaning the output does not grow without limit when a finite input is applied. A correctly tuned industrial PID temperature controller returns to setpoint after a disturbance without oscillating indefinitely or saturating the heater relay. In terms of poles, stability requires all closed-loop poles to lie strictly in the left-half s-plane; a pole exactly on the imaginary axis gives marginal stability, which is usually unacceptable in practice.

Follow-up: What is the difference between BIBO stability and Lyapunov stability?

Q2. How do you apply the Routh-Hurwitz criterion to test stability?

You form the Routh array from the coefficients of the characteristic polynomial and check that all elements in the first column are positive — any sign change indicates a right-half-plane pole and therefore instability. For the polynomial s³ + 6s² + 11s + 6, the Routh array has all positive first-column entries, confirming all three poles are in the left-half plane and the system is stable. The number of sign changes in the first column equals the exact number of RHP poles, giving you more information than just a pass/fail result.

Follow-up: What does a row of zeros in the Routh array indicate?

Q3. What is gain margin and how do you find it from a Bode plot?

Gain margin is the factor by which you can multiply the open-loop gain before the closed-loop system becomes unstable, measured at the phase crossover frequency where the phase is exactly -180°. If at the phase crossover frequency the open-loop magnitude is -12 dB, the gain margin is 12 dB, meaning you have a factor of 4 in gain before instability. A gain margin below 6 dB is generally considered insufficient for robust industrial designs because component tolerances and aging can easily reduce it further.

Follow-up: Why is gain margin alone insufficient to guarantee robust stability?

Q4. What is phase margin and why is it important?

Phase margin is the additional phase lag that would bring the open-loop phase to -180° at the gain crossover frequency where the magnitude is 0 dB. An under-damped servo with a phase margin of 30° will exhibit significant overshoot and oscillation, while a phase margin of 45°–60° gives the well-damped transient response preferred in industrial motion controllers. Phase margin directly predicts the closed-loop damping ratio: ζ ≈ PM/100 is a quick engineering approximation valid for typical second-order systems.

Follow-up: What is the typical recommended phase margin range for industrial control systems?

Q5. Explain the Nyquist stability criterion.

The Nyquist criterion states that the number of unstable closed-loop poles equals the number of clockwise encirclements of the -1+j0 point by the Nyquist plot of the open-loop transfer function G(jω)H(jω). For a stable open-loop system like most industrial plants, you simply check that the Nyquist plot does not encircle -1 — if it does, the closed-loop is unstable. The criterion handles systems with open-loop RHP poles and time delays more rigorously than Bode analysis, making it the preferred tool for analyzing sampled-data and delay-dominant systems.

Follow-up: How does the Nyquist plot relate to the gain margin and phase margin?

Q6. What is the difference between absolute stability and relative stability?

Absolute stability is a binary condition — the system either is or is not stable — while relative stability quantifies how far the system is from the stability boundary, which matters for robustness. A closed-loop motor drive with all poles at s = -0.1 is absolutely stable but has very poor relative stability because small parameter changes push it toward instability, whereas poles at s = -10 are far more robust. Gain margin and phase margin are the standard engineering measures of relative stability used in Bode-based design.

Follow-up: Which is more useful in real controller design — absolute or relative stability? Why?

Q7. What happens to stability when you increase the proportional gain of a controller?

Increasing proportional gain moves the closed-loop poles toward the right-half plane, reducing phase margin and damping until the system eventually becomes unstable at the ultimate gain. In a flow control loop for a process plant, increasing proportional gain beyond the ultimate gain Ku causes sustained oscillations at the natural frequency, which is exactly the Ziegler-Nichols oscillation test. Each plant has a specific gain crossover frequency that determines how fast oscillation begins, so practical tuning always stays well below Ku.

Follow-up: How does the Ziegler-Nichols method use the ultimate gain to tune a PID controller?

Q8. What is a conditionally stable system?

A conditionally stable system is stable only within a specific range of gain — it goes unstable if the gain is either too low or too high, which is the opposite of the typical behavior. Some high-order process control systems with three or more integrators exhibit conditional stability, where reducing controller gain actually drives the system unstable. This is dangerous in practice because a sensor failure or power-down that effectively reduces gain can cause instability, so such systems require careful analysis and are avoided where possible.

Follow-up: Why is conditional stability considered dangerous in industrial applications?

Q9. How does a time delay affect the stability of a control system?

A time delay adds a phase lag of ωT radians at frequency ω, which reduces the phase margin and can destabilize a system that would otherwise be stable without the delay. In a networked control system with a 50 ms feedback delay, the extra phase lag at the gain crossover frequency can eat into the phase margin by 20–30°, requiring a detuned (lower-gain) controller to maintain stability. Unlike a pole, a delay cannot be compensated with a standard lead network — a Smith predictor is used in process control to handle predictable delays.

Follow-up: What is the Smith predictor and how does it deal with time delay?

Q10. What is Lyapunov stability and how does it differ from BIBO stability?

Lyapunov stability is a state-space concept: a system is Lyapunov stable if trajectories starting near an equilibrium point remain near it, and asymptotically stable if they also converge to the equilibrium. BIBO stability focuses on input-output behavior and only guarantees bounded output for bounded input, but does not say anything about unobservable or uncontrollable internal states. A system with a hidden unstable mode that is cancelled in the transfer function can be BIBO stable but Lyapunov unstable, which is why state-space analysis is essential for safety-critical systems.

Follow-up: How do you construct a Lyapunov function to prove stability of a nonlinear system?

Q11. What is the significance of the characteristic equation roots lying on the imaginary axis?

Roots exactly on the imaginary axis indicate marginal stability — the system's natural response is sustained sinusoidal oscillation that neither grows nor decays. A passive LC tank circuit with ideal components oscillates at ω = 1/√(LC) indefinitely because its poles sit at ±jω on the imaginary axis. In a feedback control system, marginal stability is almost always unacceptable because real-world nonlinearities, noise, or parameter drift will push the system toward either growing oscillations or decay.

Follow-up: Is marginal stability ever deliberately used in engineering applications?

Q12. How does adding a zero to a system affect its stability?

Adding a left-half-plane zero increases the phase of the open-loop transfer function, which increases the phase margin and typically improves stability and transient response. A lead compensator, which adds a zero to the left of its pole in the s-plane, is specifically designed to inject this phase boost — a lead network with a zero at s = -2 and pole at s = -20 can add up to 55° of phase margin near the gain crossover. An RHP zero does the opposite, reducing phase margin and making stability harder to achieve.

Follow-up: What is the maximum phase boost achievable with a single lead compensator stage?

Q13. What is the Routh-Hurwitz condition for a second-order system?

For a second-order system with characteristic polynomial as² + bs + c, the Routh-Hurwitz conditions reduce to simply requiring all three coefficients a, b, and c to be positive and non-zero. A DC motor speed controller described by 2s² + 8s + 10 satisfies this immediately — all coefficients positive — so both poles are in the LHP and the system is stable. For second-order systems this coefficient sign check is sufficient, but for third-order and above you must actually fill in the full Routh table.

Follow-up: What is the minimum number of Routh array rows needed for a third-order system?

Q14. How do you determine the number of unstable poles from the Routh array?

The number of sign changes in the first column of the Routh array equals exactly the number of poles in the right-half s-plane. For the polynomial s⁴ + 2s³ - s² + 3s + 5, filling in the Routh table reveals two sign changes in the first column, meaning exactly two poles lie in the RHP and the system is unstable. This count is extremely useful in compensator design because it tells you how many unstable modes you need to stabilize, not just whether stabilization is needed.

Follow-up: If the Routh array has a zero in the first column but not an entire row of zeros, how do you proceed?

Q15. What is sensitivity function and how does it relate to stability robustness?

The sensitivity function S(s) = 1/(1 + L(s)) describes how output errors respond to disturbances, and its peak magnitude |S|max, called the sensitivity peak Ms, is a direct measure of stability robustness. An industrial pressure control loop designed with Ms = 1.4 (about 3 dB) ensures the Nyquist plot stays outside a circle of radius 1/1.4 centered at -1, guaranteeing adequate gain and phase margins simultaneously. Minimizing the sensitivity peak is a more complete robustness criterion than specifying gain and phase margins separately.

Follow-up: What is the relationship between sensitivity peak Ms and gain/phase margins?