Re: How to select parts for a MSA study ?
Greetings!
I realize this discussion has been "dead" for quite a while. But it relates directly to some questions I have. So hopefully I can resurrect it.
Firstly....
Parts should be selected randomly such that they are representative of the full range and distribution of the process.
While this is the oft quoted advice concerning part selection, am I the only one who finds this to be a frustrating oxymoron? In the first place, if part selection is supposed to be "random" then how can I add a condition about the results of the selection? Secondly, how am I supposed to verify that the parts fit the criterion....oh, hey, I'll just measure them....but wait, I'm supposed to be verifying my measurement system ....so..???
Second.....
What is the actual effect of part variation on the results? I had a GRR that failed R&R at 78%. Just as an experiment, I added 0.02 to all the measurements from all the operators on the 10th part. The GRR results passed with an R&R of 3.08%. Why?
Assistance with understanding is greatly appreciated!
Thanks!!