Linearity Analysis - Measurement System Analysis - AIAG MSA Manual 3rd Edition

[jscholz]

Thanks for your quick response. That will help me a lot.

Hello again,

1) Comparing to the spreadsheet you poiunted me to, it seems that the formula on page 93 of MSA 3rd Edition for s is incorrect. The book states that s = Sqrt (Sigma yi2 - b Sigma yi - a Sigma xi yi / gm-2); the spreadsheet calculates s = Sqrt (Sigma yi2 - b Sigma yi - a Sigma xi yi2 / gm-2). That also gives one the correct graph shown on page 95. Would you agree?

2) I'm trying also the best fit line approach, which is not part of the spreadsheet I believe. In the formula for slope and intercept, what are x and y? x as in bias and y as in measurement values?

Thanks again in adavance.

 

[Alison]

still have question



ok, I wonder if "xi" is the value given in table 5 in page 94, for example, x1=2.70, while "yi" is the value given in table 6 in page 95,for example, y1=0.7? But if I calculate these "a","b","s","ta","tb" using the "xi","yi" into the equations given in page 93,94, I can't get the exact value as given in the manual. I've calculate them in excel, the result is a=-0.14114,b=0.785969,s=0.2746,ta=-9.823,tb=8.496. I have sent you an e-mail on how can I make out these values which is different from these given in 3rd edition MSA. Please check it! The title of the e-mail is "linearity analysis" .
I know that the "a" and "b" is exactly the same in both the 2nd edition and 3rd edition MSA, but the equaltions are different,how could they make out the same values? I'm really puzzled about it, and hope you can help me! Thank you very much!

Hi Datony, I am currently doing the MSA linearity test and get the result that exactly same with you. I am confused, have you find the way to solve already? Can you help me on this?

 

[Marc]

CAUTION: This is a very old discussion thread.

Also see: Search results for query: linearity

 

[optomist1]

good point, and AIAG MSA is now in 4th Edition....

 

[Miner]

This is a link to an updated MSA 4th edition file with Linearity.