Frequency response: Bode plots

Frequency response by the substitution s = jω, Bode plots of gains, lags, leads, integrators, second-order elements, dead time and PI/PD controllers, and how series elements combine.

Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.

Why it matters

Frequency response tells you how a process or loop treats disturbances of different speeds: slow ones pass through, fast ones are filtered, and every element adds phase lag. Bode plots make this visible, handle dead time exactly (which Routh cannot), and are the basis of the Bode and Nyquist stability criteria and of frequency-based controller tuning.

Key ideas

Frequency response. If a stable linear system receives u = A·sin(ωt), then after the transient dies out the output is a sinusoid of the same frequency, with

  • amplitude ratio AR = |G(jω)| (output amplitude/input amplitude; often normalised by the steady-state gain K), and
  • phase angle φ = ∠G(jω) (negative means the output lags). Both are obtained simply by substituting s = jω in the transfer function. This is the "substitution rule".

Bode diagram. Two plots against log ω:

  • log AR (or 20·log₁₀AR in decibels) on a log scale, and
  • φ in degrees on a linear scale. For elements in series, ARs multiply (so log ARs, or dB, add) and phase angles add. A complex loop is therefore drawn by adding the plots of simple elements.

Building blocks.

  • Gain K: AR = K at all ω; φ = 0 (or −180° if K < 0).
  • First-order lag 1/(τs + 1): AR = 1/√(1 + ω²τ²); φ = −tan⁻¹(ωτ). Asymptotes: AR = 1 for ω < 1/τ and slope −1 (−20 dB/decade) above the corner frequency ω_c = 1/τ. At the corner, AR = 0.707 (−3 dB) and φ = −45°; φ goes from 0 to −90°.
  • First-order lead (τs + 1): the mirror image, slope +1 and φ from 0 to +90°.
  • Integrator 1/s: AR = 1/ω (slope −1 everywhere), φ = −90°.
  • Second-order 1/(τ²s² + 2ζτs + 1): high-frequency slope −2 (−40 dB/decade), φ from 0 to −180° (−90° at ωτ = 1). For ζ < 0.707 the AR curve shows a resonant peak above 1.
  • Dead time e^(−θs): AR = 1 at all ω; φ = −ωθ (rad), unbounded. It does not appear on the AR plot at all, but it dominates the phase plot at high frequency.
  • PI controller K_c(1 + 1/(τ_I s)): AR = K_c·√(1 + 1/(ωτ_I)²), φ = −tan⁻¹(1/(ωτ_I)), from −90° at low ω to 0 at high ω.
  • PD controller K_c(1 + τ_D s): AR rises with slope +1 above 1/τ_D; φ adds lead up to +90°.

General patterns. A transfer function with n poles and m zeros (no dead time) has a high-frequency AR slope of −(n − m) on log–log axes and a limiting phase of −90°·(n − m). Only systems whose phase can reach −180° (third order and higher, or any dead time) can be made unstable by proportional gain alone, which links directly to the next topic.

Formulas

AR = |G(j·ω)|, φ = ∠G(j·ω)

  • ω angular frequency (rad/s); ω = 2π/P for a period P.

AR_dB = 20·log₁₀(AR)

AR = K/√(1 + ω²·τ²), φ = −tan⁻¹(ω·τ)

  • First-order lag; corner frequency ω_c = 1/τ.

AR = 1/√((1 − ω²τ²)² + (2ζωτ)²), φ = −tan⁻¹(2ζωτ/(1 − ω²τ²))

  • Second-order lag (φ taken between 0 and −180°).

AR = 1, φ = −ω·θ (rad) = −57.3·ω·θ (degrees)

  • Dead time θ (s).

AR_total = AR₁·AR₂·…, φ_total = φ₁ + φ₂ + …

  • Elements in series.

Worked examples

Example 1 (standard): two lags. G(s) = 2/((s + 1)(0.5s + 1)), time in s. Find AR (also in dB) and φ at ω = 1 rad/s.

  1. Lag 1 (τ = 1 s): AR₁ = 1/√(1 + 1) = 0.7071; φ₁ = −tan⁻¹(1) = −45.0°.
  2. Lag 2 (τ = 0.5 s): AR₂ = 1/√(1 + 0.25) = 0.8944; φ₂ = −tan⁻¹(0.5) = −26.57°.
  3. AR = 2·0.7071·0.8944 = 1.265; in dB: 20·log₁₀(1.265) = 2.04 dB.
  4. φ = −45.0 − 26.57 = −71.57°.

AR ≈ 1.265 (2.04 dB), φ ≈ −71.6° at ω = 1 rad/s

Example 2 (GATE level): first-order lag plus dead time. G(s) = 5·e^(−0.5s)/(2s + 1), time in min. Find the frequency at which the phase is −180°, and the AR there.

  1. φ(ω) = −0.5ω (rad) − tan⁻¹(2ω). Set φ = −π: 0.5ω + tan⁻¹(2ω) = π.
  2. Solve by trial: at ω = 3.4, 1.700 + 1.4248 = 3.125 (too low); at ω = 3.45, 1.725 + 1.4269 = 3.152 (too high). Interpolating, ω ≈ 3.431 rad/min.
  3. Check: 0.5·3.431 = 1.7155 rad (98.3°) and tan⁻¹(6.862) = 1.4261 rad (81.7°); sum 180.0°.
  4. AR = 5/√(1 + (2·3.431)²) = 5/√(1 + 47.09) = 5/√48.09 = 5/6.934 = 0.721.

ω ≈ 3.43 rad/min, AR ≈ 0.721. Without dead time, the phase of a single lag never reaches −180°; the dead time supplies the extra 98°.

Example 3 (short): PI controller. For K_c = 2, τ_I = 1 min at ω = 1 rad/min: AR = 2·√(1 + 1) = 2.83, φ = −tan⁻¹(1) = −45°.

Common mistakes

  • Using frequency in Hz (cycles per time) where ω in rad per time is needed; ω = 2πf.
  • Mixing radians and degrees when adding the dead-time phase −ωθ to the lag phase in degrees.
  • Expecting dead time to change the AR plot; it changes only the phase.
  • Adding ARs instead of multiplying them (or adding dB values but multiplying phases).
  • Using the asymptote value instead of the true value at the corner: the true AR there is 0.707, not 1.

For GATE CH

Expect: AR and phase of a given transfer function at a given frequency, the corner frequency, slopes of asymptotes, and the frequency at which the phase reaches −180° for systems with dead time (by trial). Questions also ask for the amplitude of a sensor reading for a sinusoidal input. Practise trial-and-error solution of phase equations with consistent units.

Quick check

  1. What is the corner frequency of 3/(4s + 1) (time in min)?
  2. What is the high-frequency slope of 1/((s + 1)(2s + 1)(3s + 1)) on log–log axes?
  3. What AR and phase does e^(−2s) give at ω = 0.5 rad/s?
  4. Convert an AR of 0.1 to dB.

Answers: 1. 0.25 rad/min. 2. −3 (−60 dB/decade). 3. AR = 1, φ = −1 rad = −57.3°. 4. −20 dB.

Try answering each one aloud before you open it.

  1. 1.What is a Bode plot and why is it used in process control?Concept

    A Bode plot is a graphical representation of a system's frequency response. It consists of two plots: one showing the magnitude (gain) versus frequency and the other showing the phase shift versus frequency. Bode plots are used in process control to analyze the stability and performance of control systems, helping engineers design controllers that ensure desired system behavior.

  2. 2.Explain the significance of the gain margin and phase margin in a Bode plot.Concept

    The gain margin is the amount by which the gain can increase before the system becomes unstable, measured at the phase crossover frequency where the phase angle is -180 degrees. The phase margin is the additional phase lag required to bring the system to the verge of instability, measured at the gain crossover frequency where the gain is 1 (0 dB). Both margins are indicators of system stability; larger margins generally imply a more stable system.

  3. 3.How do you interpret the slope of the magnitude plot in a Bode plot?Concept

    On log–log axes each first-order lag adds a slope of −1 (−20 dB/decade) above its corner frequency 1/τ, and each lead adds +1. So at high frequency the slope is −(n − m), where n and m are the numbers of poles and zeros; −40 dB/decade means two more poles than zeros, whether from two lags or one second-order element. A pure integrator gives a slope of −1 at all frequencies. The slopes and corner frequencies let you read the time constants off an experimental Bode plot.

  4. 4.Why are logarithmic scales used in Bode plots?Application

    A log frequency axis covers several decades compactly, which matters because process time constants range from seconds to hours. Plotting AR on a log (or dB) scale turns the multiplication of amplitude ratios of elements in series into addition, so the plot of a whole loop is the sum of simple element plots, and each lag becomes straight-line asymptotes that meet at its corner frequency. Phase angles add directly, so the phase is plotted on a linear scale against log ω.

  5. 5.What happens to the Bode plot if a pole is added to the system?Application

    Adding a pole to the system introduces a -20 dB/decade slope to the magnitude plot starting at the pole's frequency. It also introduces a phase lag that starts at one-tenth of the pole frequency and reaches -90 degrees at ten times the pole frequency. This affects the system's stability and transient response, often requiring adjustments in the control strategy.

  6. 6.How does a zero affect the Bode plot of a system?Application

    A zero introduces a +20 dB/decade slope to the magnitude plot starting at the zero's frequency. It also introduces a phase lead that starts at one-tenth of the zero frequency and reaches +90 degrees at ten times the zero frequency. This can improve the system's transient response and stability, depending on the location of the zero relative to the poles.

  7. 7.Why is it important to consider both magnitude and phase plots in a Bode plot?Application

    Considering both magnitude and phase plots is crucial because they provide complementary information about the system's frequency response. The magnitude plot shows how the system amplifies or attenuates signals at different frequencies, while the phase plot indicates the timing or phase shift of the output signal relative to the input. Together, they help assess system stability and performance.

  8. 8.What is the effect of time delay on a Bode plot?Application

    Time delay in a system introduces a phase lag that increases linearly with frequency, without affecting the magnitude plot. This phase lag can reduce the phase margin, potentially leading to instability if not properly compensated. Time delay is often represented as an exponential term in the transfer function, and its impact is more pronounced at higher frequencies.

Finished this topic? Mark it so your progress, study plan and readiness keep up.

Stuck on something here?