Testing Methods to Qualify Operators (Measuring Devices)

[Mikey324]

Hello,

I was hoping to get some ideas from the various professionals out there. My company tests to qualify operators on a yearly basis. We use only digital mic's and digital calipers, so it's pretty basic.

We currently use 6 known standards (3 for mic, 3 for calipers) and have all operators measure these. We then create a control chart for each sample with this data. Any out of control points basically show us if an operator is heavy handed with the device, or just the opposite.

Our R&R's are always very good, and so is the data from our measurement testing. I am just wondering how this is done at other companies.

thanks,

 

[Bill McNeese]

Your approach is definitely one way. I would ask about why yearly. Long time to wait to find out if one operator is heavy-handed.

Another approach would be to have each operator run one standard on each device on a regular basis (weekly for example). Plot the results on an individuals control for the device (so all operators are plotted on the same chart for the same device). This will let you know if the device needs re-calibrating or if there is an issue with an operator.

I am a big fan of running standards/controls for tests methods on a regular basis. But if your R&R is very good, you may not need to do this at all - at least very frequently.

 

[Mikey324]

we do this yearly because calibration is done monthly on all these devices. we also track every measurement that is made in the plant, doing a cap study every month. this lets us know pretty much where we are. the annual test was more to satisfy a TS auditor. They understood that all the data was good, R&R's were good, and the customer was happy. Just wanted to see more of a qualification.
By the way, the control limits found during these tests are far tighter than the customer tolerances we are working to.

 

[bobdoering]

If you are so concerned about bias (operator influence - often, but not always unintentional) that you need to test operators that frequently, I suspect that gaging may not be the best for your process. The bias should be statistically insignificant, or it will easily crop up on the shop floor beyond the actual test. It also will generally cause disputes, slowing down process decisions.

 

[Mikey324]

we only test them yearly (or new to the job) to qualify. As far as matching the gage to the process, our option is pretty limited. We measure thickness and width only.