Automated Measurement Machine MSA - Small Parts

[arthurs]

We're thinking of purchasing a new tool that automatically measures small parts. When I say automatically I really mean that the Operator is taken out of the equation. The only input the operator has is to dump the parts in a bin. The machine does the rest. I can perform R&R study but the operator portion should contribute 0% as it make no odds to the machine who loads it. That said, I expect the R&R to show any run to run variation.

What else can I do wrt consistency trial or is an R&R good enough.

 

[Proud Liberal]

I just bought a $50k vision system that claimed parts could be randomly arranged on the stage and found R&R values far in excess of 30% when they were. The only way I could get the unit to repeat was to fixture the parts in a fixed orientation.

I guess what I'm saying is don't believe the salesman. Make the do the R&R for you before you commit.

 

[Miner]

We're thinking of purchasing a new tool that automatically measures small parts. When I say automatically I really mean that the Operator is taken out of the equation. The only input the operator has is to dump the parts in a bin. The machine does the rest. I can perform R&R study but the operator portion should contribute 0% as it make no odds to the machine who loads it. That said, I expect the R&R to show any run to run variation.

What else can I do wrt consistency trial or is an R&R good enough.

Replace the Operators field with run to run. When you interpret the Reproducibility remember that it represents the run to run instead of operator to operator.

 

[arthurs]

Well, I'm thinking of doing some kind of Bias study and also correlation to a known good measurement. I'm not too clear on the criteria for success on a consistency study or method. I think that we should demonstrate that the tool is accurate to 10% of the process tolerance too.