I have been planning to conduct an experiment to compare a particle performance on our CVD chamber between a new and old process. But the problem is that the particle event is random and unpredictable, how do we I determine the optimum sample size for this kind of evaluation