Evaluate Different Spec. Limits in Capability Analyses
Mar 23, 2012 8:36 AM(1316 views)
I spend a lot of time on the development phase of new components we make (we're a job shop). Many times a week I present initial capability data after a first manufacturing run of parts. I recognize that the measurement and process both affect a capability score, but often a part meets the functional requirements yet shows poor capability because of the tolerance, or because of the lack of a better measurement system. For this reason, I tend to always recalculate (and recalculate again...) new and alternate specification limits because that is the cheapest, fastest solution to ensure good capability down the road given the existing process and measurement system. Of course, not always is a new tolerance the way to go, but sometimes it's a fine solution. I'd like to present or inquire as to the possibility that the Capability Analysis function in JMP include a provision for evaluating different specification limits easily. For example, in my work, I ask, 1) what is the minimum tolerance *width* that yields a 1.3 Cpk (Ppk) or greater keeping the existing specification target. 2) What is the minimum tolerance width that yields a 1.3 Cpk (Ppk) or greater if the spec. target is moved to or near the process mean. In both cases, I have to keep in mind the number of decimal places a reasonable specification should include (i.e. 1, 2, 3, or 4 plcs for inches).
For this need, and because I present my data in a Word document, I have a function to do the above in Visual Basic. However, my code is still clumsy since I end up having to tweak the two outputs for an appropriate number of decimal places to present, and I'm sure that the programming crew for JMP could do a far better job with this task than me with my vb code.