Introduction to Nyquist Plots in Control Engineering

Nyquist plots are a cornerstone of frequency-domain analysis in control engineering, offering a graphical method to evaluate the stability and performance of feedback systems. Developed by Harry Nyquist in the 1930s, these plots map the frequency response of a system’s open-loop transfer function G(s)H(s) onto the complex plane. Unlike Bode plots, which separate magnitude and phase into two graphs, Nyquist plots combine both into a single polar plot, making it easier to visualize how gain and phase interact at each frequency. For engineers designing controllers for aircraft, robotics, or process control, mastering Nyquist plots is essential to ensure robust stability and avoid catastrophic oscillations.

In this expanded guide, we will explore the interpretation of magnitude and phase in Nyquist plots in depth. We will cover the underlying mathematics, practical stability criteria such as the Nyquist stability criterion, and how to extract gain margin and phase margin directly from the plot. Additionally, we will discuss common misinterpretations, the relationship between Nyquist and Bode plots, and provide practical tips to apply this knowledge in real-world engineering scenarios. By the end, you will have a thorough understanding of how to read and analyze Nyquist plots to design reliable control systems.

The Basics of Nyquist Plots: Polar Representation of Frequency Response

A Nyquist plot is essentially a polar plot of the complex function G(jω)H(jω) as the frequency ω varies from 0 to ∞. In many textbooks, the Nyquist plot also includes the symmetric portion for negative frequencies, making a closed contour. The plot is drawn in the complex plane where the x-axis (real axis) represents the real part of the transfer function and the y-axis (imaginary axis) represents the imaginary part. Each point on the curve corresponds to a specific frequency, and the vector from the origin to that point encodes both magnitude and phase.

Mathematically, for a given frequency ω, the open-loop transfer function evaluated at s = jω yields a complex number:

G(jω)H(jω) = |G(jω)H(jω)| e^{jφ(ω)}

Here, |G(jω)H(jω)| is the magnitude (gain) at that frequency, and φ(ω) is the phase angle. In the Nyquist plot, the magnitude is the distance from the origin to the point, and the phase is the angle measured counterclockwise from the positive real axis. Typically, the curve starts at ω = 0 (which corresponds to the dc gain, often on the positive real axis) and ends at ω = ∞ (which tends to the origin if the system is strictly proper).

One key distinction: Nyquist plots are often confused with polar plots or quot;Nyquist diagramsquot; in other fields. In control engineering, the Nyquist plot specifically refers to the frequency response of the open-loop transfer function used for stability analysis via the Nyquist criterion. It is not simply a plot of impedance in electrical engineering (though similar concepts apply). For a solid foundation, refer to classic texts like this introduction to Nyquist plots from Control Systems Lab, which covers the fundamentals.

Interpreting Magnitude in Nyquist Plots

The magnitude at a given frequency indicates how much the system amplifies or attenuates a sinusoidal input at that frequency. In the Nyquist plot, the magnitude is the radial distance from the origin. For a point at coordinates (x, y), the magnitude is sqrt(x² + y²). This quantity is also known as the gain of the open-loop system.

A magnitude greater than 1 means the system provides amplification at that frequency; a magnitude less than 1 means attenuation. The critical point for stability analysis is the magnitude at the frequency where the phase crosses -180° (the phase crossover frequency). At that frequency, the magnitude is equal to the reciprocal of the gain margin. Specifically, if the magnitude at -180° phase is less than 1, the system is conditionally stable; if it is greater than 1, the system may be unstable.

Gain Margin from Nyquist Plot

Gain margin (GM) is a key stability metric. It is defined as the factor by which the gain can be increased before the closed-loop system becomes unstable. In a Nyquist plot, the gain margin is inversely proportional to the magnitude at the phase crossover frequency (ω_pc), where the phase is -180°. Mathematically:

GM = 1 / |G(jω_pc)H(jω_pc)|

On the plot, this corresponds to the distance from the point where the curve crosses the negative real axis to the critical point (-1, 0). If the curve crosses the negative real axis at -0.5 (magnitude 0.5), the gain margin is 2 (or 6 dB). If it crosses at -2 (magnitude 2), the gain margin is 0.5, indicating the system is already unstable. Engineers often prefer a gain margin of at least 6 dB (factor of 2) for robust stability.

To visualize, imagine the Nyquist plot approaching the -1 point. The closer the curve gets to -1, the smaller the gain margin. If the curve encircles or touches -1, the closed-loop system is marginally stable or unstable.

Magnitude at Different Frequencies

Beyond the phase crossover, the magnitude at low frequencies (dc gain) is important for steady-state accuracy. A high dc gain reduces steady-state error for step inputs, but it may also affect stability margins if not compensated. The magnitude at high frequencies rolls off due to poles, which is why strictly proper systems tend to the origin as ω→∞. The shape of the magnitude in Nyquist plot is not as directly readable as in Bode plots, but you can trace radial distances along the curve to gauge how gain changes with frequency.

In practice, engineers often overlay a unit circle (magnitude = 1) on the Nyquist plot. The intersection of the Nyquist curve with this unit circle gives the gain crossover frequency (ω_gc) — the frequency at which the magnitude equals 1. This frequency is used to compute phase margin.

Interpreting Phase in Nyquist Plots

Phase in a Nyquist plot is represented by the angle of the vector from the origin to a point on the curve, measured from the positive real axis. A positive phase means the output leads the input (counterclockwise rotation), while negative phase means the output lags (clockwise rotation). For most physical systems, phase lag dominates, especially at high frequencies due to time delays and poles.

The most critical phase value is -180° (or -π radians). When the phase reaches -180°, the feedback becomes positive if the magnitude is also sufficiently large, leading to instability. The frequency at which the phase is -180° is called the phase crossover frequency (ω_pc). At this frequency, the Nyquist curve crosses the negative real axis. If this crossing occurs to the left of -1 on the real axis (i.e., magnitude > 1), the system is unstable; if to the right (magnitude < 1), the system is stable.

Phase Margin from Nyquist Plot

Phase margin (PM) is the additional phase lag required to bring the system to the verge of instability (i.e., to reach -180° phase at the gain crossover frequency). It is measured at the frequency where the magnitude equals 1 (gain crossover frequency, ω_gc). On the Nyquist plot, the phase margin is the angle between the negative real axis and the line from the origin to the point where the curve intersects the unit circle. More precisely, if the Nyquist curve meets the unit circle at an angle φ (in degrees), then the phase margin is:

PM = φ + 180°

For example, if the curve intersects the unit circle at a phase of -135°, the phase margin is 45°. A phase margin less than 45° indicates a system with significant overshoot in the step response; a phase margin close to 0° means near-instability. Good engineering practice typically aims for a phase margin between 45° and 75°.

To find phase margin from a Nyquist plot directly, draw a circle of radius 1 centered at the origin. Where the Nyquist curve crosses this circle (for the first time, typically), measure the angle from the negative real axis to that point. This is the phase margin. Many control design tools automatically compute it.

The Nyquist Stability Criterion: From Encirclements to Stability

The Nyquist stability criterion is the most powerful feature of the Nyquist plot. It relates the stability of the closed-loop system to the encirclements of the critical point (-1, 0) by the Nyquist plot of the open-loop transfer function G(s)H(s). The criterion accounts for open-loop poles in the right-half plane (RHP). The formal statement:

Let N be the number of clockwise encirclements of the point -1 by the Nyquist plot of G(s)H(s) as s traverses the Nyquist contour (including the imaginary axis and a large semicircle in the right half-plane). Then the number of unstable closed-loop poles Z is given by:

Z = N + P

where P is the number of open-loop poles in the right-half plane (including on the imaginary axis). For stability, Z = 0, so the Nyquist plot must encircle -1 in a counterclockwise direction as many times as there are open-loop RHP poles.

In practice, if the open-loop system is stable (P=0), the system is stable if the Nyquist plot does not encircle -1. If the plot encircles -1 clockwise, the system is unstable. If it encircles -1 counterclockwise, it is stable only if there are an equal number of RHP poles.

It is important to use the correct direction: the Nyquist plot for positive frequencies is only half of the contour; engineers typically complete the plot by mirroring for negative frequencies to form a closed loop. The encirclements are counted for the complete closed contour. For most practical systems (no RHP poles), simply checking whether -1 lies to the left or right of the Nyquist curve as frequency increases is enough. However, for systems with RHP poles or zeros, the full criterion is necessary. Detailed examples can be found in MathWorks documentation on Nyquist plots.

Reading Encirclements from the Plot

To determine encirclements, draw a line from the point -1 in any direction; count how many times the Nyquist curve crosses this line. The direction (clockwise or counterclockwise) determines the sign. Alternatively, you can use a mental quot;winding numberquot; approach. For a simple system with no RHP poles, the stability condition is that the Nyquist plot should not have -1 on its right side when traversed in the direction of increasing frequency. In other words, the point -1 must be outside the critical region.

Practical Analysis: Combining Magnitude and Phase Using Nyquist Plots

While Bode plots separate magnitude and phase, Nyquist plots show both simultaneously, which is useful for visualizing how gain and phase interact. For example, when designing a compensator, you can see how adding a lead network shifts the Nyquist curve toward the right (improving phase margin) while increasing magnitude. Conversely, lag compensation pulls the curve toward the origin at high frequencies but may reduce phase margin at low frequencies.

Gain and Phase Margins from a Single Plot

One of the greatest practical values of the Nyquist plot is that you can read both gain margin and phase margin from it, provided you know the unit circle and the negative real axis. Steps:

  1. Identify the intersection of the Nyquist curve with the negative real axis (phase = -180°). The magnitude at that point gives the gain margin as described earlier.
  2. Identify the intersection with the unit circle (magnitude = 1). The angle of that point (measured from the positive real axis) gives the phase margin: PM = angle + 180° (if angle is negative).
  3. If the Nyquist curve does not intersect the unit circle (meaning magnitude never equals 1), then the system has infinite phase margin, but you still need to check gain margin.

Important: For systems with multiple crossings, the most critical crossing determines the margins. Always check the crossing closest to the -1 point.

Common Pitfalls in Interpreting Nyquist Plots

  • Confusing magnitude and phase vectors: Remember that the vector length is magnitude; the angle is phase. Novices often misread the distance to the real axis as magnitude. Use a grid or scale.
  • Using only positive frequency branch: The stability criterion requires the entire contour, including the mirror image for negative frequencies. Many software tools only show the positive frequency branch; you must mentally complete the closed loop. If the plot is for negative frequencies, it is symmetric about the real axis.
  • Ignoring open-loop poles on the imaginary axis: Systems with integrators (poles at s=0) cause the Nyquist plot to go to infinity at low frequencies. The Nyquist contour must indent around these poles, leading to arcs at infinity that must be accounted for in encirclement counting.
  • Overlooking time delays: Time delays add additional phase lag without affecting magnitude. This can cause the Nyquist plot to spiral inward, potentially causing instability at high frequencies. The Nyquist plot is still valid, but the infinite frequency behavior must be handled carefully.

Using Software Tools for Nyquist Analysis

Modern control design is performed using computational tools like MATLAB, Python (control library), or Scilab. These tools generate Nyquist plots automatically and can compute gain and phase margins with a single command. However, understanding the underlying interpretation is vital for verifying results and troubleshooting unexpected instability. For instance, MATLAB’s nyquist() function plots the positive frequency branch; the command margin() returns gain and phase margins for the open-loop system. Python’s control library provides similar functionality. When using these tools, always double-check the direction of encirclements and remember that the plot might only show frequencies from 0 to ∞; to see the full contour, you may need to manually add the mirror image. A comprehensive tutorial on using Python for Nyquist analysis can be found at Python Control Systems Library documentation.

Relating Nyquist Plots to Bode Plots

Nyquist and Bode plots are complementary. Bode plots provide clear information about magnitude and phase separately across a wide frequency range, making it easier to identify resonance peaks, roll-off rates, and crossover frequencies. Nyquist plots compress all frequency information into a single curve, emphasizing the critical point -1 and the overall shape. Engineers often use both: start with Bode to design a compensator, then validate stability with Nyquist. The two are related through the concept of the complex plane mapping. For example, the gain crossover frequency identified from the Bode magnitude plot corresponds exactly to the point where the Nyquist plot intersects the unit circle. Similarly, the phase crossover frequency from Bode matches the Nyquist plot’s intersection with the negative real axis. For more on this relationship, see this university resource on the Nyquist criterion.

Advanced Topics: Sensitivity, Complementary Sensitivity, and Disk Margins

While gain and phase margins are traditional metrics, modern robust control theory uses additional measures like sensitivity peaks. The Nyquist plot can also be used to derive the sensitivity function S(s) = 1/(1 + G(s)H(s)). The maximum sensitivity Ms is the maximum distance from -1 to any point on the Nyquist curve. A smaller maximum distance implies better stability robustness. The complementary sensitivity function T(s) relates to performance. Disk margins, which simultaneously consider gain and phase variations, can also be visualized on the Nyquist plot as a region around the -1 point. Engineers in aerospace and automotive fields frequently use these advanced metrics to ensure system reliability under parameter variations.

To compute disk margins from a Nyquist plot, imagine a disk centered at -1 with a certain radius; if the Nyquist curve stays outside this disk, the system can tolerate a certain amount of simultaneous gain and phase uncertainty without becoming unstable. This is a more comprehensive measure than separate gain and phase margins.

Practical Worked Example: Second-Order System

Consider a simple open-loop transfer function: G(s)H(s) = 1 / (s(s+1)). This is a Type 1 system (one integrator). The Nyquist plot for positive frequencies starts at infinity when ω→0 (due to the pole at s=0), then spirals toward the origin at high frequencies. The phase starts at -90° (due to the integrator) and approaches -180° as ω→∞. The plot crosses the negative real axis at some finite frequency. For this system, the crossover occurs at a magnitude less than 1, so the closed-loop system is stable. The gain margin is infinite (since the plot never actually reaches -1 in magnitude because it crosses the negative real axis to the right of -1). The phase margin can be found where the magnitude is 1. By solving |G(jω)H(jω)| = 1, we find ω ≈ 0.618 rad/s, and the phase at that frequency is about -128°, giving a phase margin of 52°, which is acceptable.

Now modify the system: add a gain of 5: G(s)H(s) = 5 / (s(s+1)). The Nyquist plot will be scaled by 5. The intersection with the negative real axis now occurs at magnitude 5 × (something greater than 1) — actually the crossing point magnitude becomes 5 times the previous crossing magnitude. If that product exceeds 1, the system becomes unstable. In this case, the crossing magnitude for the original system was 0.5, so with gain 5 it becomes 2.5, meaning the Nyquist plot encircles -1, and the closed-loop system is unstable. This illustrates the direct relation between gain and stability.

Conclusion: Mastering Nyquist Plot Interpretation

Interpreting phase and magnitude in Nyquist plots is a critical skill for every control engineer. By understanding that magnitude is the radial distance from the origin and phase is the angle from the positive real axis, engineers can extract both gain and phase margins directly from the plot. The Nyquist stability criterion provides a robust method to assess closed-loop stability, even for systems with open-loop instability. Combining Nyquist analysis with Bode plots and modern tools allows for comprehensive design and troubleshooting. As you practice interpreting more complex Nyquist plots — especially those with time delays, non-minimum phase zeros, or multiple frequency-dependent elements — your intuition for frequency-domain behavior will improve. Always remember the critical point (-1,0) and use the unit circle as your guide. With these skills, you can design control systems that are both stable and performant across a wide range of operating conditions.

For further reading, explore the original paper by H. Nyquist, quot;Regeneration Theoryquot; (1932), or modern textbooks like Modern Control Engineering by Ogata and Feedback Control of Dynamic Systems by Franklin, Powell, and Emami-Naeini.