Nested or crossed gage R&R?

[JohannesD]

Hi all

I am currently setting up a measurement plan for a gage R&R. My measurement system consists out of 13 gages which are located on 13 positions across a plane.
To assess the reproducibility of the gages, I would rotate each of the gages on every position and repeat all measurements (so that every gage has measurements on each position).

However, this would require too much measurements and therefore I would only rotate 3 times.
So for example,
position one will be measured by gage 1, 2 and 3
position two will be measured by gage 2, 3 and 4
position three will be measured by gage 3, 4 and 5

This situation is neither a 100% crossed one, nor a 100% nested one.

What would be correct: crossed or nested gage R&R? Or neither and should I redesign my test?
Thanks!

 

[Miner]

You need to provide more context. Are these gages identical? In normal usage are these gages randomly chosen for use in any given location? Are they measuring the same type of feature in all locations?

 

[JohannesD]

Hello Miner

Thanks for the reply.
Yes they are identical. They will be randomly chosen and once installed they remain fixed on that location. The gages are measuring brightness of a lightsource, I have 13 spread across a plane since I am also interested in the uniformity of the light source (so they are measuring the same type of feature: brightness).

 

[Miner]

This does appear to be a crossed study as the same locations can be replicated by multiple gages. Just because you are not replicating it with 100% of the gages doesn't matter. You typically do not include 100% of all the operators that could use a gage in a study.

You will have to be very careful in the design and the analysis of your study to ensure that you are truly capturing the repeatability within a gage and the reproducibility between gages, but the concept is sound.

 

[JohannesD]

Ok thanks.
I will repeat the measurements for one gage in a very short timeframe (thinking of programming 3 measurements within 5 seconds).
For reproducibility, I will run the calibration sequence whenever I switch the gage, so I am sure it is not influenced by handling,...

Any pitfalls you see in this way of working?