Nyquist Stability Interactive Calculator

Nyquist Stability Criterion

Z = N + P

Z = Number of closed-loop right-half-plane (RHP) poles (dimensionless)
N = Number of clockwise encirclements of -1+j0 point by Nyquist plot (dimensionless)
P = Number of open-loop RHP poles of G(s)H(s) (dimensionless)
System is stable if and only if Z = 0

Phase Margin

PM = ∠G(jω_gc)H(jω_gc) + 180°

PM = Phase margin (degrees)
∠G(jω_gc)H(jω_gc) = Phase of open-loop transfer function at gain crossover (degrees)
ω_gc = Gain crossover frequency where |G(jω)H(jω)| = 1 (rad/s)
PM > 0° required for stability; PM = 45-60° optimal for most systems

Gain Margin

GM = -20 log₁₀|G(jω_pc)H(jω_pc)|

GM = Gain margin (dB)
|G(jω_pc)H(jω_pc)| = Magnitude of open-loop transfer function at phase crossover (dimensionless)
ω_pc = Phase crossover frequency where ∠G(jω)H(jω) = -180° (rad/s)
GM > 0 dB required for stability; GM = 6-12 dB typical design target

Critical Gain (Stability Limit)

K_cr = 1 / |G(jω_pc)H(jω_pc)|

K_cr = Critical gain at stability boundary (dimensionless)
|G(jω_pc)H(jω_pc)| = Open-loop magnitude at phase crossover (dimensionless)
Maximum gain before system becomes unstable; used in Ziegler-Nichols tuning

Gain Margin from Linear Magnitude

GM_dB = 20 log₁₀(GM_linear)

GM_dB = Gain margin in decibels (dB)
GM_linear = Gain margin as linear ratio (dimensionless)
Conversion between linear gain ratio and logarithmic representation

Scenario: Quadcopter Flight Controller Stabilization

Marcus, an aerospace engineering graduate student, is designing the pitch attitude controller for a research quadcopter drone weighing 2.8 kg intended for agricultural inspection flights. After system identification flights, he's obtained the open-loop transfer function from servo input to pitch angle: G(s) = 3850/[s(s² + 14.2s + 287)]. Initial test flights with a proportional gain of K = 0.8 showed acceptable hover performance but exhibited concerning oscillations during forward flight transitions. Using the Nyquist calculator, Marcus inputs the system parameters and discovers his phase margin is only 28.3° with a gain margin of 4.7 dB—both below the 45° and 6 dB minimums recommended for aerial vehicles. He calculates that reducing gain to K = 0.52 achieves PM = 48° and GM = 8.1 dB. Subsequent flight tests confirm dramatically improved handling with smooth transitions and excellent disturbance rejection in 15 mph crosswinds, validating that adequate stability margins are non-negotiable for safe autonomous operation.

Scenario: Industrial Temperature Control Retrofit

Jennifer, a controls engineer at a pharmaceutical manufacturing facility, is troubleshooting persistent temperature oscillations (±3.2°C) in a jacketed reactor vessel that's causing batch quality issues and yield losses. The original PID controller was tuned 15 years ago, and process modifications have altered system dynamics. She conducts frequency response testing by injecting sinusoidal setpoint changes and measuring amplitude ratio and phase shift across 20 logarithmically-spaced frequencies from 0.0001 to 0.1 rad/s. Plotting the Nyquist diagram reveals the root cause: the critical -1 point lies dangerously close to the frequency response curve, with measured phase margin of only 18° and gain margin of 2.3 dB. Using the Nyquist calculator in criticalGain mode, Jennifer determines the system is operating at 78% of the critical gain. She reduces the proportional gain by 40% and increases integral time constant by a factor of 1.8, achieving PM = 52° and GM = 9.7 dB. The improved tuning eliminates oscillations, reducing temperature variance to ±0.4°C and increasing batch yield by 6.2%, saving the facility $340,000 annually while demonstrating the tangible economic value of proper stability analysis.

Scenario: Automotive Active Suspension Development

David, a vehicle dynamics engineer at an automotive OEM, is developing the control algorithms for an active suspension system that adjusts damping in real-time based on road conditions and driving maneuvers. During laboratory rig testing of the quarter-car suspension model with electromechanical actuator, he observes a concerning 7 Hz oscillation when transitioning from smooth road to pothole simulation. His colleague suggests simply reducing controller gain, but David wants to understand the underlying stability issue systematically. He uses the Nyquist calculator to analyze the measured frequency response data from the hydraulic actuator, suspension linkage, and vehicle body dynamics. The analysis reveals two critical insights: first, the system has a lightly-damped structural resonance at 6.8 Hz producing significant phase lag, and second, the current controller design provides only 22° phase margin at the 4.2 Hz gain crossover frequency. David implements a notch filter centered at 6.8 Hz to add 25° of phase lead at that frequency, combined with slight gain reduction. The modified controller achieves PM = 51° and GM = 10.2 dB, completely eliminating the oscillations while maintaining excellent ride comfort metrics. The Nyquist plot visualization helps him explain the design changes to management, showing how the modified frequency response curve maintains safer distance from the critical -1 point across all frequencies, ultimately leading to production approval for the new suspension technology.

What is the difference between Nyquist stability analysis and Bode plot analysis? +

While both Nyquist and Bode analysis examine frequency response to determine stability, they present information differently and offer complementary insights. Bode plots display magnitude and phase separately as functions of frequency on semi-log axes, making it straightforward to read gain and phase margins directly and to visualize how these margins change with frequency. The Nyquist plot, conversely, combines magnitude and phase into a single polar representation in the complex plane, mapping the open-loop frequency response from ω = 0 to ω = ∞. The critical advantage of Nyquist analysis is its rigorous handling of systems with right-half-plane poles through the encirclement criterion—Bode analysis assumes open-loop stability and can give misleading results for unstable plants that are stabilized by feedback. Additionally, Nyquist plots reveal behavior at very high frequencies (approaching infinity) which affects contour closure, information that's compressed or lost in Bode plots. For routine controller tuning on stable plants, Bode analysis is often more convenient; for formal stability proofs, systems with unusual dynamics, or when open-loop instability exists, Nyquist analysis provides the theoretically complete picture based on Cauchy's argument principle.

Why does the Nyquist plot include negative frequencies when physical systems only respond to positive frequencies? +

The complete Nyquist contour in the s-plane must be a closed path to apply Cauchy's argument principle, requiring evaluation along the entire imaginary axis from -j∞ to +j∞, not just positive frequencies. However, for systems with rational transfer functions (ratios of polynomials with real coefficients), the frequency response exhibits conjugate symmetry: G(-jω) = G*(jω) where * denotes complex conjugate. This means the Nyquist plot for negative frequencies is simply the mirror image (reflected about the real axis) of the positive frequency portion. In practice, engineers typically plot only the positive frequency segment and mirror it to obtain the complete plot, or use software that automatically generates the full contour. The negative frequency portion doesn't represent physically meaningful excitation but is mathematically essential for closure of the contour in the complex plane. Some subtle effects do require consideration of both portions—for instance, when counting encirclements, you must track how the complete closed contour (positive and negative frequencies plus the infinite semicircle) winds around the -1 point. For most stable minimum-phase systems, the infinite semicircle maps to the origin and contributes no encirclements, but for systems with more poles than zeros or RHP poles, careful contour completion becomes critical for correct stability determination.

How do I interpret the Nyquist plot when my system has poles on the imaginary axis like integrators? +

Systems with integrators (poles at s = 0) or marginally stable oscillators (poles at s = ±jω₀) require a modified Nyquist contour that deviates around these poles using small semicircular indentations to the right. For a single integrator K/s, as you traverse the small semicircle of radius ε around the origin from -jε to +jε, the transfer function traces a large semicircular arc from -j∞ to +j∞ in the G-plane, contributing -180° to the net phase. This appears as the characteristic infinite discontinuity in standard Nyquist plots of type-1 or higher systems. When counting encirclements, you must include contributions from these indentations: each pole on the imaginary axis contributes a semicircular arc in the G-plane. For stability analysis, what matters is the net encirclements of -1+j0 by the complete modified contour. A common mistake is forgetting to account for these indentation contributions when systems have multiple imaginary-axis poles or when poles are slightly offset from the axis due to numerical precision. Practically, most software handles this automatically by evaluating the frequency response starting at a small positive frequency (like 0.001 rad/s) rather than exactly zero, effectively implementing the indentation. The key engineering insight is that integrators don't violate the Nyquist criterion—they simply require careful contour construction, and the resulting -90°×(number of integrators) phase contribution at low frequencies is what limits achievable phase margin in type-1 and type-2 systems.

Can a system be stable with negative gain margin or negative phase margin? +

Surprisingly, yes—under certain circumstances a system can be stable despite negative classical stability margins, though this represents unusual and generally undesirable design. This occurs in conditionally stable systems where the Nyquist plot crosses the negative real axis multiple times, encircling the -1 point in a way that satisfies Z = N + P = 0 despite negative margins at one crossing. A classic example involves systems with three or more crossings of the 180° line: the plot might pass outside the unit circle at the first crossing (positive GM), inside at the second (negative GM), and outside again at the third. If the net encirclements satisfy the criterion, the system is technically stable but exhibits dangerous characteristics—small gain increases can cause instability (when negative GM is present) or small gain decreases can cause instability (conditional stability). Such systems are highly undesirable in practice because they're sensitive to parameter variations and may become unstable with component aging that changes gain. Another scenario involves systems where phase margin is measured at a local gain crossover that isn't the dominant one—multiple crossings of the 0 dB line can occur in complex systems. The critical engineering lesson is that while the Nyquist criterion (Z = 0) is the definitive stability test, negative margins indicate problematic design even if technically stable. Redesign is virtually always warranted to achieve positive margins at all critical frequencies, ensuring robust stability across the full range of operating conditions, component tolerances, and environmental variations that real systems encounter.

How do time delays affect the Nyquist plot and what are practical implications for remote control systems? +

Pure time delays e^(-sτ) introduce phase lag that increases linearly with frequency (-ωτ radians) while maintaining constant unity magnitude, causing the Nyquist plot to spiral inward toward the origin as frequency increases. Unlike typical system dynamics where phase asymptotically approaches finite limits, delays contribute unbounded phase lag—the higher the frequency, the more negative the phase becomes. This creates severe practical constraints: for a given delay τ, there exists a maximum achievable bandwidth beyond which adequate phase margin becomes impossible regardless of controller sophistication. A useful guideline is that gain crossover frequency should satisfy ω_gc·τ much less than π/2 radians (about 90°) to maintain reasonable phase margin. For networked control systems, this translates directly to performance limits: a 100 millisecond round-trip communication delay (τ = 0.1 s) constrains crossover frequency to roughly 3-5 rad/s, implying settling times no faster than 1-2 seconds. This explains why teleoperated surgical robots over intercontinental links feel sluggish, why cloud-based control of physical processes is challenging, and why 5G networks emphasize low latency (1-5 ms) as enabling technology for remote control applications. Compensation strategies like Smith predictors can help by creating a delay-free model path in parallel with the delayed real path, but these require accurate delay knowledge and plant models. The Nyquist plot makes the delay impact visually obvious—the spiral shape shows how delay progressively erodes stability margins as frequency increases, providing intuitive understanding of why high-bandwidth control over communication links with even modest delays becomes fundamentally problematic.

What stability margins should I target for different types of engineering applications? +

Stability margin targets vary significantly across applications based on performance requirements, safety criticality, and environmental uncertainty. For general industrial servo applications like CNC machine tools or industrial robots, industry practice recommends phase margin of 45-60° and gain margin of 6-12 dB, providing good balance between performance and robustness. Aerospace applications are more demanding: commercial aircraft flight control systems typically require PM ≥ 60° and GM ≥ 10 dB to ensure safe handling across the full flight envelope including turbulence, icing, and off-nominal configurations. Military aircraft with advanced control laws might accept PM as low as 40° in high-performance regimes where agility is prioritized, but only after extensive flight testing validates safety. Chemical process control generally accepts PM of 30-45° since slow dynamics and overdamped response are preferred over speed. High-precision positioning systems for semiconductor manufacturing or optical alignment often target PM ≥ 65° to minimize overshoot and settling oscillations. Medical robotics, particularly systems in contact with patients, require conservative margins (PM ≥ 70°, GM ≥ 12 dB) to ensure gentle, predictable behavior. The underlying principle is that higher margins trade bandwidth for robustness—tighter margins allow faster response but increase sensitivity to modeling errors, parameter variations, and unmodeled dynamics. When designing to margins, also consider that PM and GM can vary with operating point in nonlinear systems, so margins should be verified across the full operating range, not just at a single nominal condition. Additionally, modern robust control theory suggests looking beyond classical margins to structured singular value (μ) analysis for systems with significant parametric uncertainty, but Nyquist margins remain the most widely used metrics due to their intuitive physical interpretation and experimental measurability.

Free Engineering Calculators

Explore our complete library of free engineering and physics calculators.

🔗 Explore More Free Engineering Calculators

• Bode Plot Transfer Function Interactive Calculator
• Joint Torque From Payload Interactive Calculator
• Harmonic Drive Ratio Interactive Calculator

Browse All Engineering Calculators →

About the Author

Robbie Dickson — Chief Engineer & Founder, FIRGELLI Automations

Robbie Dickson brings over two decades of engineering expertise to FIRGELLI Automations. With a distinguished career at Rolls-Royce, BMW, and Ford, he has deep expertise in mechanical systems, actuator technology, and precision engineering.

Wikipedia · Full Bio