A funny question from a customer 2 weeks ago. “We want to avoid to calculate too much and too often capability indices!”.
An outcry from most colleagues, especially the younger ones... "But we have this software, world leader in calculating capabilities... why?"
There was only a gentle smile on the face of Mr Gastgeb (me, the author), because this customer, although only using our software for a few weeks, understood.
Of course, the calculation of capabilities, following the distribution time processes, is the core of the software. Of course, we have tools to automatically and cyclically inform users about the capabilities.
But this is all downstream. If I look at a report every morning with yesterday's data to see which characteristics would now no longer be capable, then perhaps a defined amount of rejects were produced for a few hours. Using the SPC system, it wouldn't have happened. Or not as much. Because with consistent data recording under SPC rules, the following happens
- A process is calculated. Whether it is capable or not, that is its status.
- A quality control chart is defined for the status (taking all points into account in advance)
- If this quality control chart is now adhered to, it can be assumed that the capability will not deteriorate, but rather improve.
Ergo: You don't have to constantly calculate capabilities.
And when I'm explaining this, there's always this one question: Can I have a C-value alarm during data recording? An alarm if my capability would no longer exist?
No!!
Why?
I have already created a documentation for this:
Link to "Why Cp/Cpk makes no sence in O-QIS"
Nevertheless, the main part of this documentation should be repeated here for reading: