The following is a plot of the defocus as a function of time for a fixed proportionality gain of 150, and different listed sampling times.

The discrepancy between the behaviors is enough to warrant a proper mathematical explanation; the max. overshoot, for example, is inversely proportional to the sampling rate.