G
Guest
this is awsome, thanks a lot I have sent it to several customers and they think is a great form,, again thanks
Intro to Measurement System Analysis (MSA) of Continuous Data – Part 5b: R&R
- Thread starter Miner
- Start date
G
Guest
Thanks! I needed something to verify that the form I created worked and this was perfect! Your numbers match my numbers!
G
Guest
I have a problem when I do GR&R that bothers me. The molded plastic parts that are used in the study vary very little since all 100 cavities are made with the same electrodes and the processing of the molding machine is very consistent. When choosing 10 parts to measure, the variation is so small, the ndc values are always 3 or less. I use high quality digital verniers and optical comparators which can easily measure the parts accurately and well trained inspectors. How can I get a GR&R to satisfy customer Automotive Engineers?
M
Mr.Happy
Hi Xavier,
If you don't get more than 3 NDC's because of to Little variation in the parts try to measure them with a higher resolution Gage (micrometer or CMM).
Good luck,
Mr.Happy
If you don't get more than 3 NDC's because of to Little variation in the parts try to measure them with a higher resolution Gage (micrometer or CMM).
Good luck,
Mr.Happy
B
BatteryHen
Hello Minor.
Am new to MSA and got lost in the MSA manual! This helps alot and clears my head for trying to work out things. Many thanks!
Am new to MSA and got lost in the MSA manual! This helps alot and clears my head for trying to work out things. Many thanks!
QualityEngBrady
Registered
Hello @Miner,
Where does the % Contribution acceptance criterion derive from? I do not see it in AQS MSA book and it seems to be an elusive topic for me.
@QualityEngBrady This comes from Six Sigma, not from AIAG. I do not recall the specific reference.
Huh. I don’t remember the original source either. I don’t think it was strictly six sigma. I htink Mikel HArry made the limits ‘popular’ or commonly known beyond automotive?
In the eighties I used the Western Electric control chart approach then switched to the Youden approach in the 90s.
AIAG manuals in the early 90s do mention them and that they are GUIDELINES, not hard and fast rules.
Larry Berrentine probably has the first publicly documented publication in his Concepts for R&R studies published in ‘91.
The first public article was “The rubber ruler” by Robert Traver, published in the 1962 ASQC Convention Publication. It was reprinted in Volume 5, issue 1 1992 of Quality Engineering. He discusses ‘acceptability criteria’ but doesn’t mention any specific guidelines. He only says that engineering and economic judgement must be applied. (This paper is a summery of what Robert learned from GM’s use of R&R). Robert was a consultant who worked for Dorian Shainin for a time.
In any account the 10%/30% rule is 1) arbitrary. There is NO mathematical or statistical derivation or proof of these limits. 2) the %coverage formula is a blatant violation of mathematical rules, so any such derivation would be based on a fallacy. As @Miner has pointed out the limits are fairly conservative due to their violation of mathematical rules…
In the eighties I used the Western Electric control chart approach then switched to the Youden approach in the 90s.
AIAG manuals in the early 90s do mention them and that they are GUIDELINES, not hard and fast rules.
Larry Berrentine probably has the first publicly documented publication in his Concepts for R&R studies published in ‘91.
The first public article was “The rubber ruler” by Robert Traver, published in the 1962 ASQC Convention Publication. It was reprinted in Volume 5, issue 1 1992 of Quality Engineering. He discusses ‘acceptability criteria’ but doesn’t mention any specific guidelines. He only says that engineering and economic judgement must be applied. (This paper is a summery of what Robert learned from GM’s use of R&R). Robert was a consultant who worked for Dorian Shainin for a time.
In any account the 10%/30% rule is 1) arbitrary. There is NO mathematical or statistical derivation or proof of these limits. 2) the %coverage formula is a blatant violation of mathematical rules, so any such derivation would be based on a fallacy. As @Miner has pointed out the limits are fairly conservative due to their violation of mathematical rules…
I think you may have misunderstood the OP's question. They were asking about % Contribution, which is based on the variances, not the standard deviations. AIAG applied the 10/30 criteria to % Tolerance and to % Study Variation but did not apply any such criteria to % Contribution. They just kind of threw % Contribution in there as a metric and neglected to say much about it. I first saw guidelines applied to % Contribution in Six Sigma training materials. I won't say that that is the original source, but it wasn't AIAG. The GM precursor to the AIAG manuals did not even show % Contribution.
Similar threads
