Second-order systems: underdamped and overdamped response
Second-order systems K/(τ²s² + 2ζτs + 1): where they arise, how the damping factor sets overdamped, critically damped and underdamped responses, and overshoot, decay ratio, period and settling time.
Drafted with Aria, reviewed by the AiCanCode.org team. Spotted an error? Use Give Feedback at the bottom of the page.
Why it matters
Second-order behaviour appears whenever two lags act in series, in inherently oscillatory elements such as a U-tube manometer or a control valve with its actuator, and, most importantly, in almost every feedback loop under PI control. The terms overshoot, decay ratio and period are how plant engineers describe loop performance, and the classic tuning target of a quarter decay ratio is a second-order idea.
Key ideas
Standard form. In process control a second-order system is written G(s) = K / (τ²s² + 2ζτs + 1), where K is the steady-state gain, τ the natural period of oscillation (time) and ζ the damping factor (dimensionless). The equivalent mechanical form uses the natural frequency ω_n = 1/τ: for a mass–spring–damper, ω_n = √(k/m) and ζ = c/(2√(k·m)).
Sources of second-order dynamics.
- Multicapacity processes: two first-order lags in series give τ² = τ₁τ₂ and 2ζτ = τ₁ + τ₂, always with ζ ≥ 1.
- Inherently second-order elements: a U-tube manometer, a pneumatic valve actuator, a spring-loaded instrument.
- Processes with controllers: a first-order process under PI control gives a closed loop that can be underdamped.
Poles and damping. The poles are s = (−ζ ± √(ζ² − 1))/τ.
- ζ > 1, overdamped: two distinct real poles; sluggish, no overshoot. The system factors into two first-order lags with time constants τ/(ζ − √(ζ² − 1)) and τ/(ζ + √(ζ² − 1)).
- ζ = 1, critically damped: a repeated real pole; the fastest response without overshoot.
- 0 < ζ < 1, underdamped: complex poles; the response overshoots and oscillates with decaying amplitude.
- ζ = 0, undamped: sustained oscillation at the natural frequency 1/τ rad per unit time.
- ζ < 0: growing oscillation (unstable).
Underdamped step response (unit step, K = 1). y(t) = 1 − (1/√(1 − ζ²))·e^(−ζt/τ)·sin(√(1 − ζ²)·t/τ + φ), with φ = tan⁻¹(√(1 − ζ²)/ζ). Its characteristics depend only on ζ (and τ for the time scale):
- Overshoot (fraction of the final change by which the first peak exceeds it).
- Decay ratio (ratio of successive peak heights above the final value), equal to the square of the overshoot.
- Period of oscillation and time to first peak.
- Rise time (first time the final value is reached) and response (settling) time (time to stay within ±5 % or ±2 %). Smaller ζ means faster rise but more overshoot and longer settling.
Process-control targets. A quarter decay ratio (DR = 0.25) corresponds to ζ ≈ 0.215 and an overshoot of 50 %; it was the classic Ziegler–Nichols target. Many modern tunings aim at ζ around 0.5–0.7 for less overshoot and more robustness.
Formulas
G(s) = K / (τ²·s² + 2·ζ·τ·s + 1)
- K gain; τ natural period (s); ζ damping factor (–).
Overshoot OS = exp(−π·ζ / √(1 − ζ²))
- Valid for 0 < ζ < 1, as a fraction of the final change.
Decay ratio DR = exp(−2·π·ζ / √(1 − ζ²)) = OS²
Period T = 2·π·τ / √(1 − ζ²), ω_d = √(1 − ζ²)/τ
- T in s; ω_d damped angular frequency (rad/s).
Time to first peak t_p = π·τ / √(1 − ζ²)
Settling time t_s ≈ 4·τ/ζ (±2 %), ≈ 3·τ/ζ (±5 %)
- Approximation based on the decay envelope.
ζ = −ln(OS) / √(π² + (ln OS)²)
- Damping factor from a measured overshoot.
τ₁, τ₂ = τ / (ζ ∓ √(ζ² − 1))
- Overdamped system split into two first-order time constants.
ω_n = 1/τ = √(k/m), ζ = c / (2·√(k·m))
- Mass–spring–damper analogue: k stiffness (N/m), m mass (kg), c damping coefficient (N·s/m).
Worked examples
Example 1 (standard): characteristics of an underdamped system. G(s) = 2/(s² + 0.8s + 1), time in minutes, unit step input. Find ζ, τ, overshoot, decay ratio, period, time to first peak and the peak output.
- Compare with τ²s² + 2ζτs + 1: τ² = 1, so τ = 1 min; 2ζτ = 0.8, so ζ = 0.4.
- √(1 − ζ²) = √0.84 = 0.9165.
- OS = exp(−π·0.4/0.9165) = exp(−1.371) = 0.254.
- DR = OS² = 0.0644.
- T = 2π·1/0.9165 = 6.86 min; t_p = T/2 = 3.43 min.
- Peak output = K·(1 + OS) = 2·1.254 = 2.51.
- Settling (±2 %) ≈ 4τ/ζ = 4·1/0.4 = 10 min.
OS = 25.4 %, DR = 0.064, T = 6.86 min, t_p = 3.43 min, y_max ≈ 2.51
Example 2 (GATE level): identifying ζ and τ from a test. A closed-loop step test shows an overshoot of 30 % and a period of oscillation of 6 s. Find ζ, τ and the decay ratio.
- ln(OS) = ln(0.30) = −1.204.
- ζ = 1.204/√(π² + 1.204²) = 1.204/√(9.870 + 1.450) = 1.204/3.365 = 0.358.
- √(1 − ζ²) = √(1 − 0.128) = 0.934.
- τ = T·√(1 − ζ²)/(2π) = 6·0.934/6.283 = 0.892 s.
- DR = OS² = 0.30² = 0.09.
ζ ≈ 0.358, τ ≈ 0.89 s, DR = 0.09
Example 3 (short): overdamped factoring. G(s) = 1/(4s² + 5s + 1): τ = 2, ζ = 5/(2·2) = 1.25. Time constants τ/(ζ ∓ √(ζ² − 1)) = 2/(1.25 ∓ 0.75) = 4 and 1, so G = 1/((4s + 1)(s + 1)).
Common mistakes
- Confusing τ (natural period) with a first-order time constant; for an underdamped system the oscillation period is 2πτ/√(1 − ζ²), not τ.
- Taking the decay ratio equal to the overshoot; DR = OS².
- Reading 2ζτ as ζτ when matching coefficients, or forgetting to normalise the denominator so the constant term is 1.
- Using the overshoot formula for ζ ≥ 1 (there is no overshoot).
- Using a mass–spring formula with ω_n in Hz instead of rad/s.
For GATE CH
Expect: extracting K, τ and ζ from a transfer function; computing overshoot, decay ratio, period or time to peak; working backwards from a step test to ζ; deciding whether a response oscillates from the characteristic equation; and finding the closed-loop ζ of a first-order process under PI control or of two lags under P control. Practise normalising the denominator before matching coefficients.
Quick check
- For G = 1/(9s² + 3s + 1), find τ and ζ.
- What decay ratio corresponds to an overshoot of 0.5?
- Can two non-interacting first-order lags in series give ζ < 1?
- For ζ = 0, what does the step response look like?
Answers: 1. τ = 3, ζ = 3/(2·3) = 0.5. 2. 0.25. 3. No, ζ ≥ 1. 4. A sustained oscillation of constant amplitude about the final value.
Interview questions
All Process Instrumentation and Control interview questionsTry answering each one aloud before you open it.
1.What is a second-order system in the context of process control?Concept
It is a system described by a second-order differential equation, written in process control as G(s) = K/(τ²s² + 2ζτs + 1), with gain K, natural period τ and damping factor ζ. Second-order behaviour arises from two first-order lags in series (always ζ ≥ 1), from inherently oscillatory elements such as a U-tube manometer or a valve actuator, and from feedback loops, for example a first-order process under PI control, which can be underdamped. The damping factor decides whether the step response is sluggish, critically damped or oscillatory.
2.Explain the difference between underdamped and overdamped responses in second-order systems.Concept
In a second-order system, an underdamped response occurs when the damping ratio is less than one, leading to oscillations before the system settles at its final value. An overdamped response occurs when the damping ratio is greater than one, resulting in a slower response without oscillations. Critically damped systems, with a damping ratio equal to one, return to equilibrium as quickly as possible without oscillating.
3.Why is the damping ratio important in the analysis of second-order systems?Application
The damping ratio is crucial because it determines the nature of the system's response to disturbances. It affects the speed and stability of the system's return to equilibrium. A low damping ratio can lead to excessive oscillations, while a high damping ratio can cause the system to respond too slowly. Engineers use the damping ratio to design systems that balance speed and stability.
4.What happens to the response of a second-order system if the damping ratio is zero?Application
If the damping ratio is zero, the system is undamped and will exhibit sustained oscillations at its natural frequency. This means the system will continue to oscillate indefinitely without settling to a steady state, which is often undesirable in practical applications as it can lead to instability.
5.How does the natural frequency affect the response of a second-order system?Application
The natural frequency of a second-order system determines the speed of the oscillations in the system's response. A higher natural frequency means the system will oscillate more quickly. It is a key parameter in designing systems to ensure they respond appropriately to changes or disturbances.
6.In what scenarios would an engineer prefer an overdamped system over an underdamped one?Application
An engineer might prefer an overdamped system when stability and avoiding oscillations are more critical than speed. For example, in systems where overshoot could cause damage or safety concerns, such as in certain chemical processes or structural applications, an overdamped response ensures a smooth and steady approach to equilibrium.
7.Calculate the damping ratio for a system with a damping coefficient of 50 Ns/m and a mass of 10 kg, given that the natural frequency is 5 rad/s.Numerical
The damping ratio (ζ) is calculated using the formula ζ = c / (2 * √(m * k)), where c is the damping coefficient, m is the mass, and k is the stiffness. First, calculate the stiffness using the natural frequency: ω_n = √(k/m), so k = ω_n² * m = (5 rad/s)² * 10 kg = 250 N/m. Then, ζ = 50 Ns/m / (2 * √(10 kg * 250 N/m)) = 50 / (2 * √2500) = 50 / 100 = 0.5.
8.An overdamped second-order system has a damping factor of 2 and a natural frequency of 3 rad/s. What are its two effective time constants?Numerical
Here τ = 1/ω_n = 1/3 s. An overdamped system factors into two first-order lags with time constants τ/(ζ − √(ζ² − 1)) and τ/(ζ + √(ζ² − 1)). With √(ζ² − 1) = √3 = 1.732, these are (1/3)/0.268 = 1.244 s and (1/3)/3.732 = 0.089 s. The slower one dominates the response.
9.Explain how feedback control can change the damping of a loop.Application
Closing a loop changes the characteristic equation, so the controller settings move the closed-loop poles. For example, two lags under P control give 1 + Kc·K/((τ₁s + 1)(τ₂s + 1)) = 0, and raising Kc reduces the closed-loop ζ from above 1 to well below 1, so the response becomes faster but oscillatory. Integral action usually lowers damping further, while derivative action adds damping, which is why PID tuning is largely the choice of an acceptable ζ or decay ratio.
10.Is a critically damped response always the best target for a process control loop?Application
No. Critical damping (ζ = 1) is the fastest response without overshoot, but a slightly underdamped response reaches set point sooner and rejects disturbances faster, which is why classic tuning aimed at a quarter decay ratio (ζ ≈ 0.22) and many modern tunings use ζ of about 0.5–0.7. A critically damped design also becomes underdamped as soon as the process gain rises or extra lag appears, so engineers choose damping by trading speed against overshoot and robustness to model error.
Finished this topic? Mark it so your progress, study plan and readiness keep up.
Stuck on something here?