Has anybody performed MSA studies on Hardness Measurements (Rockwell/ Brinnel / Knoop / Vickers)?

Has anybody performed MSA studies on Hardness Measurements (Rockwell/ Brinnel / Knoop / Vickers)?
I am at a loss. What is MSA???

MSA stands for " Measurement Systems Analysis" - a requirement referred both in QS 9000 and TS 16949:2002

We do a lot of testing for our semi-solid aluminum parts - almost all with the Rockwell scale. While I've gone through the excersize of a study, I've never submitted the results as they were pretty ugly (even though I used a test standard for the study). One of the reasons (I think) is that you are measuring an average for one point and (at least in our methodology) then measuring two more points and averaging all three of the averages for your "measurement" :eek: . That was a few years ago and I've not attempted another since.

I've not had a customer ask for a study on a hardness tester.

I think you mean non-replicable MSA.
According to MSA 3rd, it requires studies on Discrimination, Accuracy (Bias, Linearity) and Precision (GRR) for a variable gage system to make the system is suitable for a characteristic.
Since, Accuracy and Precision studies are based on repeatable measuring data, it is very difficult to conduct such studies on non-replicable system.

I always ignore Accuracy Study.

I focus on Precision (GRR) and try my best for that. Starting with Homo-genius and Hetero-genius assumption. Then, ANOVA GRR on the data.

We are a commercial heat treating facility registered to TS16949. For many years we have been able to "avoid" performing Gage R&R studies for hardness measurements under the guise that they are destructive tests in nature and it would be difficult to objectively perform such studies. My background is 25 years in the specialty steel business responsible for chemical and mechanical testing labs. Having gone through accreditation experiences to ISO Guide 25/ISO 17025 and NADCAP I have experienced creative ways to deal with the issue relating to Measurement Systems Analysis/Uncertainty of Measaurement and the like.

The primary issue from my experience with respect to Gage R&R specifically regarding hardness measurements is that the challenge is to avoid erroneously determining appraiser and equipment variation when in fact the most significant source of variation is in the material itself (speaking of ferrous materials).

We have started an MSA program where I require our calibration vendor to perform linearity and bias studies using NIST traceable test blocks during one of their quarterly calibration visits annually. We verify hardness testers daily with master test blocks to show stability compliance. I have started a program of Gage R&R studies using both test blocks and homogeneous sample parts. We purchase test blocks specifying that the variability (sigma and repeatability) is as small as possible. We will section off the test block into zones resembling a "test part" for each zone/section. The gage R&R is performed using both these test blocks where a measurement is made in each section by each appraiser over 3 trials. We will in parallel perform a Gage R&R using typical sample parts to differentiate between variation attributable to inhomogenous material and equipment appraiser variability. To date we have had encouraging results. We have performed this procedure for our Vickers and Knoop measurements and will begin a similar study on our Rockwell C and R15N scales.

I'd be interested in any comments, thoughts or suggestions that the group may have. I hope this helps.