"Too little" variation in gage R&R

H

Hamlet

#1
Dear Sirs;

I was starting out to perform a simple gage R&R study on a digital vernier gauge, used in our incoming inspection to control some metal parts from a new supplier.
Now, since I have only received one batch so far, of app. 2.000 pcs., I really do not have any ?Historical process data? to rely on in terms of determining the STDV to be used for calculating the performance in terms of ?assessing the process performance? - and more so: The characteristic that is critical to our production, (- to be measured with the vernier gauge in question), is the width of the parts, stamped out from standard sheet metal - which obviously results in an extremely low STDV if I let MiniTab estimate the value from the study alone.

Better do the example to clarify:

Nominal part width: 3,35 mm.
LSL: 3,42 mm
USL: 3,30 mm
Study: 10 items, 3 appraisers, measurements x 2.
Result: 60 measurements in all; approximately 30 x 3.35 mm and 30 X 3,36 mm, almost perfectly distributed.
Tolerance variance: Perfect ! (Around 8-9%)
Study variance in terms of the ability to "assess process performance": Totally through the roof? (70% or more?) which is probably due to the low STDV calculated by Minitab, (0.002), which is low alright, but nevertheless plausible hence the stamping process if you were able to obtain data from a long term production.

Obviously the problem is how to demonstrate the R&R as well as the ability to monitor process variance - but how the heck do I do this with a digital vernier gauge ? - which by the way, is far within the 0.2 acceptance criteria, and the actual measurements only vary one single increment?
Would it be OK to simply set a higher STDV for the study ?
Or should I just explain - if anyone asks that is - that I really do not care for the ability to assess the process performance, since I am only interested in demonstrating the reproducibility and repeatability of the measuring equipment in order to ensure the conformity of parts in our incoming inspection ?

Any good advise would highly appreciated?
 

Miner

Forum Moderator
Staff member
Super Moderator
#2
Since this gauge is used in receiving inspection, will it only be used to assess product to specification? That is, you do not intend to use it for SPC or capability studies?

If the gauge is only used for inspection, you need only worry about the %Tolerance metric. %Study Variation and ndc do not matter unless you intend to use the gauge for SPC or capability studies.

See my blog on MSA, particularly parts 5a and 5b for more detail.
 
H

Hamlet

#3
Hi Miner;

Thank you so much for clarifying ! I had a hunch that the answer was that simple, but I was not sure.
As recommended, I just took a glance at your fine blog, where a another thread also confirmed the fact that %tolerancerVar has to do for incoming inspection.
Still, I know that I will probably face another challenge soon, when I have to perform a R&R on a Zeiss CMM running SPC on an extreeeemely stable process? - but that?s another story.

Thanks again..!
 

Top