Gage RR contradiction

Six Sigma – iSixSigma Forums Old Forums General Gage RR contradiction

Viewing 2 posts - 1 through 2 (of 2 total)
  • Author
  • #39855


    I’ve a question about GRR. When we do a GRR-study with 3 products, 3 operators and 10 measurements, we see that our operators find almost the same values for each product and for the products mutual. Great would you say. But, the results of Minitab are not great (%Study var = 71.1, Number of Distinct Catogories = 1). When we change the results of one product with a couple hundredth for each operator we find a very good GRR (%Study var = 10.69, Number of Distinct Catogories = 13). So when we would make one product a couple of hundredth to big we would have a good GRR. How is this possible? Please help.



    Hi Rob,
    I may have read this wrong, but it should be (in theory) 3 people measuring, 3 times each, 10 different associated items, not 3 products (apologies if i misread)
    Anyhow, the gage you have is not picking up the variation in part to part for at least a couple of reasons, 1) the gage may not be sensitive enough 2) there is actually no variation in the parts (thus giving a poor result)
    the reson why you get a good gage when making a bigger part is the fact that there is part to part variation which the gage can pick up – thus giving low % study var. & higher DC’s.
    you may need to conduct a gage RnR using 10 diffirent parts that have been produced under normal production environment, ie some a little small, some perfect, some a little big (varying parts within your process)
    This should give you a better feel for your gage, but you need varying parts. you need part to part variation for a MSA to be succesful (if you do not have variation, maybe there is not a problem!!)

Viewing 2 posts - 1 through 2 (of 2 total)

The forum ‘General’ is closed to new topics and replies.