Statistics

Statistical Analysis of Variance. Using standard deviations six times higher and six times lower than your target, determine whether your process is “in control” or “out of control”. Good stuff.

It’s also part of the framework for determining the reproducibility and repeatability of gaging. I’ve discovered that Minitab is one of the best software suites for achieving what you need in calculating ANOVA for your GR&R. It’s a great deal faster than laying out your own Excel spreadsheet, and it allows you to completely automate the process. Doing capability studies, GR&R’s, and X̅&R charts becomes a task that requires seconds, rather than having to deal with irritating Excel spreadsheets or plotting by hand. I’ve learned a ton about statistical process control. I’ve also learned a ton about what works, and what doesn’t.

One of my initial failures (and subsequently, great successes), was discovering that using ten pieces with too similar dimensions will wreak havoc on your GR&R. You’ll wind up with study variance of 40% or worse, which will affect every other aspect of your results (unless you have a huge amount of tolerance).

I think I’m going to write a tutorial on how to conduct a Gage Repeatability & Reproducibility study. Hmm…

27 Oct 2013 07:38 | work, math, hobbies, gr&r, x̅&r, anova, spc
blog comments powered by Disqus