Reacting to Gage R&R Findings

[mlowden]

Reaction to R&R findings

Question regarding R&R findings.
We have recently conducted an R&R on a digital balance that is used in our lab for weighting samples for dissolution to run on XRF. The procedure was developed in the early 80’s and specifies a sample size of 2g +/-.0005g. This sample is then digested and diluted to 100ml. When this procedure was developed, it was concluded that a variance in the sample of .002g was insignificant, in other words the XRF could not repeatedly discriminate. However, in an effort to reduce the amount of variation, the author of the procedure chose to limit the total sample variation to .001g.

The results of the R&R indicate the total R&R is 39% of the total tolerance, (.001g), which is above the allowable 30%.

The question is: Should the procedure be changed to allow a tolerance of +/-.001g for the sample, thus increasing the amount of total variation, but would cut the R&R findings to 19% of the total tolerance?

We are just wondering what direction to take in this matter.

Just for information, our plant is certified QS9000 and the lab is A2LA accredited. Thanks for any replies.

 

[mlowden]

Gee - this must be a tough one!
Any thoughts would be appreciated.

 

[tomvehoski]

Who authored the specification? If it was your company I don't see any problem with changing it. You may have to get customer approval though if it is subject to PPAP or other requirements. I would make sure you have other data (experiments) showing the change does not affect the test in case you are challenged.

If it is a ANSI, ASTM or other industry standard, you probably cannot since it would affect 17025 requirements.

Tom

 

[Benoit]

How do you performed the r&r study on scale?

I want to know how do you perform your r&r study on scale?

Regards

 

[mam]

If you have a product tolerance try that in your calculations; we do our GRR's based on product tolerance.
If the tolerance is internal, change it as long as you can meet the product tolerance. Be careful how much.

I'm not sure what the physical procedure is that you are using, but the below may give you some possible directions.

If the EV is high, review the equipment. It may have been able to hold the smaller tolerance fresh out of the box, but may have lost some of its discrimination over time. This would require re-hab at the manufacturer to fix. Avoid contamination of the samples used, from being exposed to heat or cold which may cause condensation. Use the exact same container(s) to avoid container variance. Add controls to be used in the study.

If the AV is high, then you have to fine tune the techniques used by all operators to insure that the item is placed at the same place on the scale(?), that an equal amount of time is spent for stabilization on the scale, that extra care is taken when placing and removing the item from the scale. This is also a good time to train others in 'best' procedures.