How often is Attribute Gage R&R required to be performed?

[hemix81]

I am quite new here at Elsmar Cove. I think I have read all the information about attribute gage R&R I can find on internet. I have got some questions about it.

I am working for a electronics industry and planning to do an attribute gage R&R for a specified product. It is quite clear that in the study there is going to be 26 products, 3 or more appraisers (testers) and 2 trials. Do you think two trials is enough? If we increase amount of trials to 3, testing time will lenghten --> number of products needs to be increased --> confidence interval becomes too wide.

Now I would like to know how often it is recommend to perform attribute gage R&R? Every six months or more often?

Thanks,
Heidi

[Atul Khandekar]

Welcome to the Cove, Heidi!

I believe 26 samples, 3 appraisers and 2 trials should be good enough. Also search the forums for recommendations on selecting the parts for attribute MSA study.

I am not aware of any specific guidelines on the frequency of these studies.

[qualeety]

In Reply to Parent Post by hemix81

I am quite new here at Elsmar Cove. I think I have read all the information about attribute gage R&R I can find on internet. I have got some questions about it.

I am working for a electronics industry and planning to do an attribute gage R&R for a specified product. It is quite clear that in the study there is going to be 26 products, 3 or more appraisers (testers) and 2 trials. Do you think two trials is enough? If we increase amount of trials to 3, testing time will lenghten --> number of products needs to be increased --> confidence interval becomes too wide.

Now I would like to know how often it is recommend to perform attribute gage R&R? Every six months or more often?

Thanks,
Heidi

Unless there is change to the process, why would want to repeat gage GRR?

Once you accept the current measuring system is adequate to measure your product (GRR% < 10%), there is no need to repeat GRR unless

1. specification changes
2. personnel changes
3. new measuring devices
4. procedural changes

note: natural deviation of the measuring devices should be corrected by calibration procedure, not GRR.

[hemix81]

Thank you for your replies.

[Jim Wynne]

In Reply to Parent Post by qualeety

Unless there is change to the process, why would want to repeat gage GRR?

Once you accept the current measuring system is adequate to measure your product (GRR% < 10%), there is no need to repeat GRR unless

1. specification changes
2. personnel changes
3. new measuring devices
4. procedural changes

note: natural deviation of the measuring devices should be corrected by calibration procedure, not GRR.

I agree with the idea that there's no need to repeat GR&R unless circumstances change, and that calibration is the correct remedial method, but the idea that the sole criterion for acceptance of a measurement system is <10% GR&R is a persistent fallacy. For some systems, results >10% may be perfectly acceptable, and for others, 9% may not be adequate. There are also other factors to consider. If the results of an average-and-range study are <10% but there are other statistical issues such as UCLr violations, the system should not be considered unconditionally acceptable.

[qualeety]

In Reply to Parent Post by JSW05

I agree with the idea that there's no need to repeat GR&R unless circumstances change, and that calibration is the correct remedial method, but the idea that the sole criterion for acceptance of a measurement system is <10% GR&R is a persistent fallacy. For some systems, results >10% may be perfectly acceptable, and for others, 9% may not be adequate. There are also other factors to consider. If the results of an average-and-range study are <10% but there are other statistical issues such as UCLr violations, the system should not be considered unconditionally acceptable.

of course, GRR% of 50% is acceptable in exceptional cases...eg. testing hardness of foam on Instron (about 50% GRR) and testing spring on Instron (less than 3% GRR).....so, each and every case is different...having said that, if you show GRR% of less than 10%, most customers will accept the measurement system to be acceptable...unless there is a "pain in the ***" engineer (hate to say this but am one of them ) who reviews suppliers' results and asks questions.

[Jim Wynne]

In Reply to Parent Post by qualeety

of course, GRR% of 50% is acceptable in exceptional cases...eg. testing hardness of foam on Instron (about 50% GRR) and testing spring on Instron (less than 3% GRR).....so, each and every case is different...having said that, if you show GRR% of less than 10%, most customers will accept the measurement system to be acceptable...unless there is a "pain in the ***" engineer (hate to say this but am one of them ) who reviews suppliers' results and asks questions.

If the object is to satisfy a need for paperwork, there's no need to even do the GR&R. To accept study results based on a single (and potentially insignificant) criterion is asking for trouble.

[qualeety]

In Reply to Parent Post by JSW05

If the object is to satisfy a need for paperwork, there's no need to even do the GR&R. To accept study results based on a single (and potentially insignificant) criterion is asking for trouble.

Unfortunately, i would say 90% of all suppliers who prepare ppap have to satisfy a need for paperwork and thus you have to do GRR, unless you want to create a bogus one...and am pretty sure some do!!!!