I'm trying to determinate a machine and process capability and I have those questions. Can I determine a machine capability measuring only one shaft with a plenty of processed surfaces, then each surface has its own fit tolerance, becouse of its cylinderity? If I can should I consider all fit tolerance measuring results for only one shaft? Or should I consider MIN and MAX measured fit tolerance for a plenty of shafts?
Is it better for the process control to use XR ratio or Xs ratio?

Suzi

This isn't my area of expertise, never the less:

Rule of thumb is that you should run 200 pieces if possible but no less than 50 for a machine capability study. One piece will not give you a time component and therefore will not show linear drift of the process. You also need to take subgroup measurements in order to get subgroup averages and then a grand average in order to determine the shape of the distribution.

Repost your question in the Tooling and Equipment Suppliers Forum. I believe you will find someone there with practical experience in machine capability studies.

Rick

This is an oldie, but does anyone have any contemporary thoughts on this?