ANOVA Gage R&R % Tolerance and Preference

  • Thread starter Thread starter barneyjoe
  • Start date Start date
B

barneyjoe

Is the entire tolerance used in the ANOVA method or is 1/6 of full tolerance? Also, is the preferred method in the ANOVA %tolerance or %variation?


:bonk:
 
Last edited by a moderator:
Elsmar Forum Sponsor
barneyjoe,

Welcome to the Cove!

I'm not sure I understand your question completely. Range Average or ANOVA methods are used to compute the amount of error in your measurement system based on study data. Further analysis method (use Tolerance or Process Variation) is entirely upto you- I'd suggest you use both. Please browse/search the MSA forums for more discussions on this topic.

As for GRR as % of Tolerance, I would calculate it as (6* GRR StDev)/Total Tolerance or as you said divide the GRR StDev by 1/6th of tolerance.
 
Anova versus X bar-R method for gage R & R

Anova method is considered superior over Xbar-R method due to its ability to give interaction of parts*operator which Xbar-R cannot give.
Anova can be used only for crossed study and not for nested study.

Crossed-If you have 15 samples and 3 operators each measuring 3 times, you need to all 15 samples to be checked by each of the operators.

Nested-If you have one operator who can check samples 1 to 5 in first shift multiple times and another operator in second shift who will measure sample # 101 to 105 multiple times and third operator in third shift measuring sample # 201 to 205 multiple times.

None of the current software has the ability to give confidence intervals on gage R & R, hence wherever possible , try to use Anova method.

Arvind
 
Last edited by a moderator:
Back
Top Bottom