Step response calculator
Enter a transfer function and get its unit-step response with rise time, peak overshoot and settling time. Choose K·G(s) itself or the unity-feedback closed loop K·G/(1 + K·G).
- Poles of G(s)
- −2−10
- Zeros of G(s)
- none
- DC gain G(0)
- ∞
- Closed-loop poles at this K (1)
- −2.325−0.3376 ± 0.5623j
- Stability
- Closed loop stable: every pole is in the left half-plane.
Simulated system: T(s) =1s³ + 3s² + 2s + 1
- Final value
- 1
- Rise time (10–90%)
- 2.677 s
- Peak overshoot
- 14.46%
- Peak time
- 6.078 sPeak value 1.145
- Settling time (2%)
- 12.36 s
Same result in code
% G(s) = (1) / (s^3 + 3s^2 + 2s)
G = tf([1], [1 3 2 0]);
K = 1;
T = feedback(K*G, 1); % unity negative feedback
step(T); grid on
S = stepinfo(T) % 10–90% rise time, 2% settling band by defaultimport control as ct
G = ct.tf([1], [1, 3, 2, 0])
K = 1
T = ct.feedback(K*G, 1) # unity negative feedback
resp = ct.step_response(T)
print(ct.step_info(T)) # RiseTime 10–90%, SettlingTime 2% by defaultMATLAB uses the Control System Toolbox; Python uses python-control 0.10 or later (pip install control).
What this calculator computes
The response y(t) to a unit step applied at t = 0 with zero initial conditions, and the standard transient specifications measured from it.
For the default example 4/(s² + 2s + 4), a standard second-order system with ζ = 0.5 and ωn = 2 rad/s, the calculator gives 16.3% overshoot at 1.814 s, a 10–90% rise time of 0.819 s and a 2% settling time of 4.04 s. The overshoot and peak time agree with the closed-form values e^(−πζ/√(1−ζ²)) and π/ωd.
How to read the result
- Final value: the DC gain T(0), reached only when every pole is in the left half-plane. All other metrics are measured against it.
- Rise time: the time from 10% to 90% of the final value. Peak overshoot: the largest excursion beyond the final value, as a percentage of it.
- Settling time: the last time the response leaves the ±2% band around the final value. A dip in the wrong direction first (a right-half-plane zero) is reported as undershoot.
Method and conventions
- The response is exact at every sample, not integrated step by step: a state-space realization is extended with the constant input, and each sample is advanced by the matrix exponential e^(Mh). Stiff systems with very fast and very slow poles need no special treatment.
- Crossing times are then located between samples by bisection on the same exact solution, so the metrics do not depend on the plot resolution. The definitions match MATLAB stepinfo and python-control step_info defaults.
If any pole is on the imaginary axis or in the right half-plane, there is no final value, and the calculator plots the growing or oscillating response without metrics. Initial conditions, input limits and delay are not modeled.