Process capability

Dear all,

for process validation we are currently using the process capability (Cpk) as a criterion for process robustness. For tablet compressing, the parameters we use are the individual weight, the content uniformity and the hardness. For that purpose this parameters are determined on two sets of 25 samples each, drawn at maximum and minimum tabletting speed.

The issue arose during revalidation, when the process was transferred to a new tablet press with control loops for weight and hardness, and the objection is that the process capability has to be calculated only for parameters which are not controlled, that is, for the machine “free running” with the control loops disabled.

Wonder if there is a statistician out there to clarify whether the calculation of the Cpk makes sense for loop controlled processes or not.

Thanks in advance