Step by step GR&R with Minitab

I’m going to go through an incredibly quick rundown of how to conduct your own Gage Repeatability and Reproducibility study using Minitab. The first thing you should keep in mind is that there are ways you “should” conduct a GR&R and there are ways you “shouldn’t”. You “should” be doing a blind, randomized study that doesn’t allow the “operator” to influence the study with what he expects. This isn’t always the easiest way to begin your first study or introduce yourself to these types of studies. For your first study, just do your trials in order.

The first thing you need: three participants. Your three participants do three trials. You can do more or less, but you need at least two participants with at least two trials. Next you need ten pieces to measure. CLOCK YOUR PIECES. What do I mean? If you’re measuring a circle, mark your circle with a marker as to where it’s being measured on all ten pieces. There are no perfect circles, and if three people measure ten circles in different ways three times a piece, you’re not going to have collected anything but entropy. Another really, really important tip: You need to have samples to measure at all extremes of your tolerance. If you have ten pieces that are all within a tenth of a thousandth of an inch of one another and you have a twenty thousandth of an inch for tolerance, you’re going to have an awful GR&R. Your study variance is going to be enormous and you’re not going to have a good time.

Here is the format for your GR&R worksheet:

Once you’ve collected your data, paste it into Minitab. You’ll have to stack it with this command:

(Enter commands by pressing Ctrl+L.)

1

Stack C6-C15 C3.

That takes stacked data from columns C6 to C15 and places it in column C3.

Next, we’ll automagically propagate our data from the table.

1
2
3
4
5
6

Set C2
  1( 1 : 10 / 1 )9
  End.
TSet C1
  10(  "Operator 1" "Operator 2" "Operator 3" )3
  End.

Those two commands insert part numbers and operator names / trials into the table as necessary.

Now that we have our data in a format that Minitab likes, we can continue on to our Nested ANOVA GR&R. Easy command, but you’ll have to edit it to your needs. Change the USL and LSL to the dimension tolerances of your part.

1
2
3
4
5
6
7

GageRR;
  Parts C2;
  Opers C1;
  Response C3;
  Studyvar 6;
  LSL 8.0;
  USL 11.0.

There you have it. The absolute bare necessities to analyze graphically and statistically the capability of any gage you study. It’s incredibly handy. One of these days I’ll write a followup post on how to analyze the graphs and data output by Minitab. There are other resources out there, but this is the legwork for getting you going. Nearly anyone can be a metrologist with this information. Now go, analyze your measurement equipment!