Introduction to Nyquist Plots and Lead-Lag Compensation

In control system design, engineers routinely face the challenge of improving both stability and transient response. Lead-lag compensation is a classic technique that addresses these goals by modifying the open-loop frequency response. However, choosing the right compensator parameters and verifying their effectiveness require careful frequency-domain analysis. Among the most insightful tools for this task is the Nyquist plot, which provides a complete graphical map of a system's gain and phase behavior across all frequencies. By comparing Nyquist plots before and after compensation, engineers can directly observe changes in stability margins, robustness to uncertainties, and overall performance improvements. This article explains how to use Nyquist plots to evaluate lead-lag compensation, offering practical steps and interpretation guidelines that apply to real-world control loops.

Fundamentals of Nyquist Plots

A Nyquist plot is a polar plot of the open-loop transfer function G(s)H(s) evaluated along the Nyquist contour in the complex plane. It maps the real and imaginary parts of the frequency response as the angular frequency ω varies from –∞ to +∞. The resulting curve reveals key stability information, particularly the proximity to the critical point (–1, 0j). The Nyquist stability criterion states that the number of closed-loop unstable poles equals the number of anticlockwise encirclements of –1 plus the number of open-loop unstable poles. Therefore, the plot serves as a direct visual test for closed-loop stability.

Key Features of the Nyquist Plot

  • Magnitude and Phase: The distance from the origin at a given ω indicates the gain, while the angle indicates the phase shift.
  • Critical Point: The point (–1, 0j) represents the threshold where the open-loop gain is unity and the phase is –180°. Encircling this point implies potential instability.
  • Stability Margins: The gain margin is the reciprocal of the magnitude at the frequency where phase is –180°. The phase margin is the additional phase lag required to reach –180° at the crossover frequency where magnitude equals 1.

Nyquist plots are especially valuable for systems with time delays or non-minimum phase elements, where Bode or root locus methods alone may be insufficient. Because the Nyquist plot considers the entire frequency range, it captures interactions between multiple crossover frequencies and provides a comprehensive view of robustness.

Lead-Lag Compensation Essentials

Lead-lag compensation blends the advantages of lead and lag networks into a single transfer function. A lead compensator adds positive phase shift near the crossover frequency, increasing phase margin and improving transient response (faster rise time, lower overshoot). A lag compensator increases low-frequency gain, reducing steady-state error while often decreasing phase margin. By combining them, engineers can simultaneously achieve higher low-frequency gain for accuracy and adequate phase margin for stability.

Typical Transfer Function

A lead-lag compensator has the form:

D(s) = Kc · (s + z1)(s + z2) / (s + p1)(s + p2)

where zeros and poles are placed such that one zero-pole pair provides lead (zero at lower frequency than pole) and the other provides lag (pole at lower frequency than zero). The design aims to shape the open-loop Bode plot: the lag portion raises the low-frequency magnitude, while the lead portion boosts phase at the desired crossover region.

Impact on Frequency Response

Because Nyquist plots directly display the compensated open-loop frequency response, they reveal how the lead-lag network alters the locus around the critical point. A well-designed compensator pulls the Nyquist curve away from –1, increasing the gain and phase margins while maintaining a smooth trajectory. Conversely, poor compensation can cause the curve to come dangerously close to –1 at multiple frequencies, indicating fragile stability or excessive peaking in the closed-loop response.

Evaluating Compensation with Nyquist Plots

To assess whether a lead-lag compensator achieves its objectives, follow a structured comparison of uncompensated and compensated Nyquist plots. The steps below outline a practical workflow suitable for both analytical models and measured frequency response data.

Step-by-Step Analysis Procedure

  1. Obtain Open-Loop Transfer Functions: Write the uncompensated open-loop transfer function G(s)H(s) and the compensated version Gc(s) = D(s)G(s)H(s).
  2. Select Frequency Range: Choose a range from near zero frequency to well beyond the expected crossover (typically 0.01 to 1000 rad/s for many electromechanical systems).
  3. Compute the Nyquist Data: Evaluate the complex frequency response for both systems across the chosen frequencies. Use a high enough resolution to capture sharp changes – 100-200 points per decade is often sufficient.
  4. Plot Both Curves: Display the uncompensated and compensated Nyquist plots on the same axes (or side-by-side). Ensure the critical point (–1, 0j) is clearly visible.
  5. Mark Stability Margins: Identify the frequency where the uncompensated phase crosses –180° and note the gain margin. Then locate the frequency where magnitude crosses 0 dB (unit circle) and measure the phase margin. Repeat for the compensated plot.
  6. Check Encirclements: Count the net encirclements of –1 for each plot, accounting for any open-loop poles in the right half-plane. Zero net encirclements indicate stable closed-loop if the open-loop is stable.

Example Comparison

Consider a unity-feedback system with G(s) = 1 / [s(s+1)(s+2)] (uncompensated). The Nyquist plot crosses the negative real axis near –0.75, giving a gain margin of about 1.33 (2.5 dB) and a phase margin of 15° – borderline stability. After adding a lead-lag compensator D(s) = (s+0.1)(s+5) / (s+0.01)(s+10), the compensated Nyquist plot shows the –180° crossing shifted to a much lower magnitude (gain margin > 10 dB) and the 0 dB crossover occurs at a phase of –140° (phase margin > 40°). The curve remains well away from –1. These improvements are directly visible on the Nyquist diagram, confirming the compensator’s effectiveness.

Interpreting Nyquist Plot Changes

Once both plots are generated, interpret the variations to determine whether the compensation meets specifications. The three most important indicators are phase margin, gain margin, and the shape of the curve near the critical point.

Phase Margin Improvement

In an uncompensated system with low phase margin (often below 30°), the Nyquist curve passes close to –1 at the crossover frequency. A successful lead-lag compensator rotates the curve counterclockwise near that region, increasing the distance from –1 and pushing the phase crossing to a higher frequency. The result is a larger phase margin – typically 45° to 60° for well-damped systems. If the compensated plot shows only a marginal increase (less than 10°), the compensator’s zero and pole locations may need adjustment.

Gain Margin Considerations

Gain margin indicates how much the gain can increase before instability. A lead-lag compensator often reduces the peak magnitude of the Nyquist plot near the negative real axis. If the compensated curve intersects the negative real axis at a magnitude of 0.3 or less, the gain margin is at least 10 dB – generally acceptable. However, if the compensator introduces a resonance peak (a bulge in the Nyquist curve), the gain margin may actually decrease, indicating that the lag or lead portion has been poorly tuned.

Encirclements and Relative Stability

For systems with open-loop unstable poles, the number of encirclements must match the required condition for closed-loop stability. After compensation, the Nyquist plot may encircle –1 fewer times (if the compensator stabilizes the system) or the same number of times but at a safer distance. Always verify that the compensated system satisfies the Nyquist criterion; a plot that does not encircle –1 at all for a stable open-loop process indicates robust stability.

Additional Shape Features

Beyond margins, examine how the Nyquist curve approaches the origin at high frequencies. Ideally, the curve should spiral into the origin quickly without oscillations. A compensated plot that oscillates near the origin may indicate high-frequency noise amplification or unmodeled dynamics. Also check for any sharp bends near the –1 point, as these suggest that the compensator is creating a near-cancellation of poles and zeros, which can lead to sensitivity issues in practice.

Practical Considerations and Limitations

While Nyquist plots offer deep insight, engineers must be aware of practical challenges when using them for lead-lag evaluation.

Accuracy of the Model

The Nyquist plot is only as accurate as the system model. In real systems, unmodeled high-frequency dynamics or nonlinearities can cause the actual Nyquist curve to differ significantly from the theoretical plot. Always validate the compensated design with experimental frequency response data if possible. A common practice is to perform a sine-sweep test on the physical system and construct the Nyquist plot from measurements.

Numerical Resolution

Lead-lag compensators with widely separated poles and zeros (e.g., lag zero at 0.1 rad/s and lead pole at 100 rad/s) require careful selection of frequency points. Standard logarithmically-spaced vectors with 200 points may miss sharp transitions, leading to incorrect margin calculations. Increase resolution near the critical region – the frequencies around crossover and phase inversion – to avoid misinterpretation.

Multiple Crossings

In some systems, the Nyquist plot crosses the negative real axis more than once. Each crossing affects stability margins differently. The smallest gain margin among all crossings dominates the system’s robustness. When evaluating compensation, check all crossings; a lead-lag network might improve one crossing while worsening another. The overall evaluation should consider the worst-case margin.

Importance of Nyquist vs Bode

Bode plots are often simpler for initial design, but Nyquist plots provide a more complete stability picture, especially for non-minimum phase systems or those with delays. When evaluating lead-lag compensation, combine both tools: use Bode for gain and phase adjustments during design, then use Nyquist for final verification. For further reading, see the classic text Franklin, Powell, and Emami-Naeini's "Feedback Control of Dynamic Systems" and the comprehensive guide on Nyquist stability criterion on Wikipedia.

Conclusion

Nyquist plots remain one of the most powerful tools for evaluating the effectiveness of lead-lag compensation in control systems. By comparing the uncompensated and compensated open-loop frequency responses on the complex plane, engineers can directly observe changes in stability margins, encirclements, and overall robustness. The visual nature of the Nyquist diagram makes it easy to identify whether a compensator achieves the desired improvements in phase margin and gain margin, and to detect potential issues such as multiple phase crossovers or high-frequency resonances. While careful modeling, numerical resolution, and experimental validation are necessary, the Nyquist-based evaluation provides a definitive check that complements Bode and root locus methods. Mastering this approach enables engineers to design lead-lag compensators with confidence, ensuring that real-world control loops meet performance and stability specifications.