What is derivative kick and how can it be mitigated in PID implementations?

• A. A sudden change in the derivative term caused by a step change in setpoint; mitigated by using derivative on measurement (D on M) or a derivative filter.
• B. Derivative kick refers to noise in the measurement due to sensor heating.
• C. Derivative kick can be mitigated by increasing proportional gain.
• D. Derivative kick is a problem of integral action causing windup.

Answer: (A)

Derivative kick shows up when a setpoint change happens: the derivative term reacts to how fast the error is changing, so a sudden jump in the setpoint makes the error slope change abruptly and the D term spikes. Since the derivative is effectively listening to the rate of error change, a step in the reference translates into a large, momentary derivative contribution that can drive the actuator hard and produce a transient upset.

The best ways to mitigate this are to modify how the derivative is computed. Using derivative on measurement means the derivative term reacts to the rate of change of the process variable rather than the error. If the process variable changes smoothly, the derivative contribution stays tame even when the setpoint steps. This decouples the derivative action from the setpoint jump and reduces the kick.

Another effective approach is to filter the derivative term with a low-pass filter. By smoothing the high-frequency content of the derivative, rapid spikes caused by setpoint steps (and noise) are attenuated, preventing large transient actions while still providing the stabilizing effect of the derivative on the overall response.

Increasing proportional gain does not address the spike in the derivative path and can alter loop behavior in undesired ways. The derivative kick is not related to integral action or windup, so those aspects aren’t the fixes here.