Folks,
We are two students who are conducting a Gage R&R study on destructive methods for a glassfactory.
Has anybody found a way to measure gage R&R for these kind of testing mehtods???
Or does anybody know an alternitive method to measure destructive testing with the same output as gage R&R??????

Thanks
Bas&Jasper

According to the rules that apply for a gage R&R Study, the test needs to be "REPEATABLE".
ANY destructive test is not repeatable and therefore is beyond the capability of the R&R test.

Although destructive testing precludes 'repeatability' as defined in classical R&R, that does not negate the capability of determining measurement error. Many guidelines exist in the chemical industry, as in qualifying a lab for EPA testing. Find some of Dr. Wheeler's writings, Evaluating Measurement Systems in particular.

I have had some success with R&R studies using destructive test methods. I use Don Wheeler's methods, and in one of his books he validates the concept of destuctive testing. He states that we have to make an assumption that samples taken near one another or consecutively (depending on your process) are identical.

It's easy for the products we test (adhesive products made in web/roll form). We just test areas a few inches over on a web. It may not be so easy for other processes.

First, it is important to note that even if the gauge is destructive, there is a measurement error and it is needed to evaluate it ! Doing nothing is the worst case.

We have developped a new method to evaluate repeatability on destructive gauge, it will be published this year in Quality Engineering (ASQ). It is a two steps procedure. The variance of the error (repeatability) is confonded with the variances of the locations is the first step, and with the parts in the second step. Using additivity of the variance, solving a simple equation give the variance of the error.

It is then possible to estimate standard %R&R and %P/T.

FR.

Francois,

I do not have any copies of Quality Engineer available to me. Would you have a version available for e-mail transmission? This is a pretty major sticking point for me, as some of our tools are destructive, and it seems that a normal Gage R&R will be confounded by the destruction caused by the measurement. Thanks for any help...

-benwah