Don’t undertake capability analysis without control charts

Don’t undertake capability analysis without control charts

It is very common, in my experience, to see people in industry undertaking capability analysis, without reference to the corresponding control charts. This is understandable to a degree where there is very little data available, for example, during the early stages of process validation, or during preliminary R&D work. However, when there is adequate data available, it is of vital importance that the control chart is studied in conjunction with the capability analysis.

Many people are aware that there are two sets of capability indices, Cp/Cpk and Pp/Ppk, although they frequently don’t understand how the difference between them arises. This distinction between the indices is of recent origin, and I have seen its arrival on the capability analysis scene in the mid 1990’s. The first published document to distinguish clearly between the two sets of indices was the SPC manual that accompanied the QS 9000 standard produced by the American “Big Three” automotive companies, Ford, General Motors and Chrysler in 1995. Prior to the 1990’s there was only one set of indices in use, namely, Cp/Cpk; the lower Pp/Ppk values were there, but were not distinguished from Cp/Cpk.

In my SPC training courses, I emphasise the importance of understanding the difference between the two sets of indices, and explaining how the differences arise. It is common to find a considerable difference between Cp/Cpk on one hand, and Pp/Ppk on the other hand, with Pp/Ppk at a lower, less attractive, level. To understand this difference, the practitioner must look at the control chart. The instability, should it exist, leading to poorer capability, and the difference between Cp/Cpk and Pp/Ppk, can be readily seen on the control chart.

I am a great fan of control charts. I see them as one of the most powerful of analytical tools, and this useful application is over and above their main function, which is to assist in the control of manufacturing processes. When people send me data, and I have analysed millions of data values over the years, the first thing I will do is to analyse the data on a control chart. I can immediately see important features of the manufacturing operation, in particular, whether is the process is being run in a stable condition.

SPC practitioners can be readily trained in the correct interpretation of control charts, and how they lead to better understanding of capability analysis.