Ppk (process performance index) incorporates an estimate of the ‘long-term process standard deviation’ which is standard deviation as we all know it. This includes both the within and between estimates of variation which is usually greater so it reduces the estimate of capability. Ppk is a measurement of how close the center (the mean) is to either the upper spec. or the lower spec. It is generally accepted that you want to have at least six standard deviations between the mean and the nearest specification limit. Ppk however is not measured in single standard deviation units. Instead, it is measured in three sigma sets. For example, if Ppk = 2, there would be six standard deviations between the center of the process (the mean) and the nearest specification limit. Remember the difference between population and sample. If we are measuring a population, we would be looking at sigma units. If we are using Ppk to evaluate a sample set, instead of ‘sigma’, we would be looking at ‘standard deviations.’
Use: Ppk is a ‘look forward’ metric for monitoring one of the two ways of making a bad part, or making a customer unhappy is when the centering has shifted toward one of the specification limits. It is for this reason that Ppk is used along with Pp. In contrast, Cpk is more of a ‘where we’ve been’ look at process capability.
