Gage R&R acceptance impasse

Welshwizard

Involved In Discussions
Why not just use the Intraclass Correlation Coefficient (ICC) to properly and empirically prove the ability of the measurement process to pick up signals in the production process? You already have the data to hand (estimates of PV and TV) although you might like to get a better estimate of part variation than you have from the 10 parts or so that you have already measured if you can.

For more information please see Dr Wheelers paper :https://www.spcpress.com/pdf/DJW222.pdf, extensively covered in his book EMP-III too. Incidentally, you will find an interesting few sentences on NDC in this book too, also reinforced in https://www.spcpress.com/pdf/DJW223.pdf.

The fact that this simple approach is not presented in Minitab should not deter you from using it. When you use this approach, assuming your measurement process is consistent, you will completely and properly arrive at a much better estimate and classification of the ability of the measurement process to pick up signals on a process behaviour chart.

Of course, whether the measurement process can correctly sentence parts to a tolerance is another question which I am sure has been addressed here many times.

Hope this is helpful, good luck with your process.
 
Top Bottom