Gage R&R - Is my sheet calculation wrong?

kedarg6500

Quite Involved in Discussions
Dear friends

I took a GRR study. But I could not get OK results. The GRR sheet I am using is as per MSA 4 edition example.

I have attached the GRR study I conducted recently. I don’t know whether sheet calculation is wrong, or study conducted is wrong. I may be wrong, but I could see consistency in data.

Please help

Kedar
 

Attachments

  • GRR8.xlsx
    23.3 KB · Views: 252
Last edited:

Jen Kirley

Quality and Auditing Expert
Leader
Admin
A quick bump - can anyone answer this? I don't trust my memory of MSA formulas.
 

howste

Thaumaturge
Trusted Information Resource
I took a GRR study. But I could not get OK results. The GRR sheet I am using is as per MSA 4 edition example.

I have attached the GRR study I conducted recently. I don’t know whether sheet calculation is wrong, or study conducted is wrong. I may be wrong, but I could see consistency in data.
The "Conclusion" cell is evaluating cell B54, which is % of total variation (30.08%). This has nothing to do with the tolerance of the part. I believe it should be looking at cell C54, which is % of tolerance (2.76%).
 
N

ncwalker

Your sheet calculation is wrong. The numbers for % Total Variation are correct, but the numbers for % of Tolerance are NOT correct. For all four percentages you are dividing the different variations by the total tolerance 0.007 - 0.003 = 0.004. You need to divide this result by 6 to get what is called Tolerance Variation. 0.004 / 6 = .00067 and use this as the denominator in all your % of Tolerance calculations.

Source: NOT the MSA manuals, I copy pasted your data into a sheet that I know works. Double check with the MSA.
 

normzone

Trusted Information Resource
Also, you are working in the Quality field.

You were either mad to take that step, or have been driven mad by the pressures.

Or both.

:bigwave:
 
N

notadog

This is to "ncwalker" could you please provide a copy of the GRR that you know that works or tell us where it can be found? I have tried a few of the GRR forms (using excel) and all I could do is repeat what "kedarg6500" had calculated?
 

kedarg6500

Quite Involved in Discussions
Your sheet calculation is wrong. The numbers for % Total Variation are correct, but the numbers for % of Tolerance are NOT correct. For all four percentages you are dividing the different variations by the total tolerance 0.007 - 0.003 = 0.004. You need to divide this result by 6 to get what is called Tolerance Variation. 0.004 / 6 = .00067 and use this as the denominator in all your % of Tolerance calculations.

Source: NOT the MSA manuals, I copy pasted your data into a sheet that I know works. Double check with the MSA.

Attached is MSA example from MSA 4 handbook. I am using the same excel sheet for calculation then how my calculations are wrong?

please help
 

Attachments

  • GRR MSA4 EXAMPLE .xlsx
    32.2 KB · Views: 749
Last edited:
N

ncwalker

Try this one.


This one is clearly dividing the tolerance by 6. But - it is ONLY reporting the % of Tolerance. Which means it is a suitable Gage R&R (I haven't really double checked everything, so I'm not certifying it is error free ....) for answering the question - "Is this gage suitable for checking these parts against this tolerance?" Which is the PPAP question - are you going to catch bad parts?

It does not report % Study Variation, which is the six sigma or DOE question. "Given these parts, is my gage good enough to detect the difference?" Which is the question you ask when doing a study. Remember - you're tweaking inputs looking for some optimization. You have to demonstrate you can detect the changes, which has nothing to do with the tolerance.
 

Attachments

  • GRR Tol Only.png
    GRR Tol Only.png
    27.1 KB · Views: 237
N

ncwalker

Attached is MSA example from MSA 4 handbook. I am using the same excel sheet for calculation then how my calculations are wrong?

please help

Your calculations are wrong because people aren't careful. Everyone goes and hacks together an Excel spreadsheet because due to computers today, it's "easy." Then, nobody goes back and TESTS their spreadsheets against known working models. "It's in a computer, therefore it must be correct."

Absolutely false.

You should question ANY spreadsheet you are given, ESPECIALLY if it isn't protected. Copies float around and people trample through them thinking they understand and the click and copy and try to do things quickly - next thing you know, broken formulas start appearing.

There's LOTS of reasons your calculations could be wrong.

That's, in fact, how I learned so much about Gage R&Rs. Suppliers were giving them to me, everyone had their "own" spreadsheet. One day, I had some free time, and started copy pasting the input data between sheets. And found that in MANY cases the spreadsheets did not agree. And I found that not all spreadsheets are created equal. Heck, I found calculation errors in the professional gage management software my own company bought.

So I made a gage R&R spreadsheet and locked it down. That way, I KNEW nobody had accidentally stepped on a formula. And I don't mean to say people were giving me intentional errors to make their gages look better, they were just careless spreadsheet errors, that propagated copy by copy. At that point, we no longer accepted Gage R&Rs other than those done in Minitab or those done on my Excel sheet. We found too many instances where the underlying math was wrong.

Same thing with capability studies.

Now, like I said, *I* may be wrong, but I don't think I am.

PS - It would be nice if the Cove created a library of TESTED sheets. I see a lot of people flinging up sheets. And they are OK (NOT protected). USE WITH CAUTION. So if, say, like this in these comments someone flings up a sheet, great. BUT ... it would be nice if we could submit an "official" sheet to the Cove, have it peer tested and certified, and then have a "Cove approved" tool section. I don't know how that would fit with their mandate or not.
 

Attachments

  • GRR Excerpt.jpg
    GRR Excerpt.jpg
    184.8 KB · Views: 231
Top Bottom