The output of a photodiode is used as the feedback signal of a laser intensity control loop. The controller currently implements PI control, and the laser together with the optical path forms the plant transfer function.
When a step change is applied to the system, the photodiode output shows the response in the attached waveform. The rising edge exhibits a moderate overshoot, which is expected due to loop dynamics. However, at the falling edge there is a sharp positive spike just before the signal drops to the lower level.
The spike is very narrow compared to the rest of the transient response and appears consistently during the falling transition.
What mechanisms in a laser intensity control loop could produce a positive overshoot spike at the falling edge of the photodiode response?

When a step change is applied to the system, the photodiode output shows the response in the attached waveform. The rising edge exhibits a moderate overshoot, which is expected due to loop dynamics. However, at the falling edge there is a sharp positive spike just before the signal drops to the lower level.
The spike is very narrow compared to the rest of the transient response and appears consistently during the falling transition.
What mechanisms in a laser intensity control loop could produce a positive overshoot spike at the falling edge of the photodiode response?

