MSA on set of parts with low variation

[Welshwizard]

Hi Risca,
A situation many people find themselves in is where your customer insists on a certain ability of the measurement process, in this case we are talking about the fidelity of the measurement process for spc. A lot of

Dear Welshwizard,

Thank you for your input. You wrote: "but crucially in your case the patterns made by joining the dots by lines should be parallel or the points outside the limits should be at least on the same side of the limits!". You mean that the patterns of the lines from operator 1 should be parallel or on the same side of the limits compared to other operators or is it also valid for Xbar charts from data only from one operator? You are more then welcome to elaborate what you mean with parallell and on same side, I would be thankfull for that! I have always worked under the assumption: Points beyond the control limits in the Xbar chart (limits from an R-chart) is good when looking at measurements systems. It means that the measurement system can tell the difference between samples.

Thank you!

You are correct in your understanding that points outside the limits on the X Bar Chart indicate a level of resolution.

Yes, if you join two adjacent points on one X Bar Chart - say operator 1 join the points to make a line and compare the same line in terms of parallelism with the X Bar chart for operator 2 you should get reasonable parallelism. Bear in mind that looking for parallelism for points that lie inside the control limits is pointless due to the fact that points inside the limits represent noise......but any points outside are fair game!
Note that this isn't an exact science, but if lines are 90 or 180 degrees opposed to each other then the process is trying to tell you something.

A very powerful yet simple graphical tool.

Hope this helps