For discussion only,im just curios on the possible approaches:

question: if my measurement system applies only sampling due to its destructive characteristics, then
1. Can i do MSA? How?:confused:
2. Can i run a Stability Analysis? How?:confused:
3. What are the other ways to assure the my measurement system is "measuring it right"?:confused:

Any thoughts anyone? :D

question: if my measurement system applies only sampling due to its destructive characteristics, then
1. Can i do MSA? How?:confused:
2. Can i run a Stability Analysis? How?:confused:
3. What are the other ways to assure the my measurement system is "measuring it right"?

1. Yes. You can perform an MSA for a destructive test. The AIAG MSA Manual has a good section on the method. A brief description of the method for a 3-operator, 3-trial, 10-part study is: Select 10 sets of 9 parts/samples (3 operators x 3 trials = 9). Each set must be highly homogeneous such as consecutive samples or samples within a single part. The 9 samples must be randomly assigned amongst the 3-operators for testing. This method depends on the homogeneity of the samples within a set. If this is unlikely, the method may give high Repeatability results.

2. Maybe. You would need homogeneous samples sufficient for the entire course of the study. The samples themselves must be stable over the entire course of the study. That is, the samples must not deteriorate or change in performance over time. A stability study is intended to assess the stability of the gage itself, not the product. If the test has standard that may be used such as a viscosity standard, this would be a viable option.

3. A Linearity Study would also be possible. The same criteria as #2 would apply, but sets of homogeneous samples/standards throughout the measurement range would be required.

2. Maybe. You would need homogeneous samples sufficient for the entire course of the study. The samples themselves must be stable over the entire course of the study. That is, the samples must not deteriorate or change in performance over time. A stability study is intended to assess the stability of the gage itself, not the product. If the test has standard that may be used such as a viscosity standard, this would be a viable option.


Thank you for the response. Just to check if i understand it correctly. The idea is the collect homogeneous samples, then instead of the usual measuring the same sample repeatedly, use the set of homogenous samples and measure 1 by 1. this way i'm measuring the samples with same charateristics, thus minimizing part-to-part variation and highligting only Gr&R. Do i get it right?

Thank you for the response. Just to check if i understand it correctly. The idea is the collect homogeneous samples, then instead of the usual measuring the same sample repeatedly, use the set of homogenous samples and measure 1 by 1. this way i'm measuring the samples with same charateristics, thus minimizing part-to-part variation and highligting only Gr&R. Do i get it right?

Yes. You are correct. One homogeneous set of 9 would allow 3 operators to take 3 repeat measurements.

For discussion only,im just curios on the possible approaches:

question: if my measurement system applies only sampling due to its destructive characteristics, then
1. Can i do MSA? How?:confused:
2. Can i run a Stability Analysis? How?:confused:
3. What are the other ways to assure the my measurement system is "measuring it right"?:confused:

Any thoughts anyone? :D

Why not try simulation technique, for example if it is a force measuring instrument (bond testers)...use a device to duplicate the bond force...

Why not try simulation technique, for example if it is a force measuring instrument (bond testers)...use a device to duplicate the bond force...

Yes that is also possible, and very creative. But if there is a possible interaction between the measuring instruments and the actual products, care must be taken such that this can be captured by the simulated measurement.