Gage R&R for Vision System

[QualityEngBrady]

Hello,

At my plant we use vision systems to measure dimensions of capacitors (length, height, width, terms). I assume the gage itself is deemed an automatic gage because it takes a overhead shot of all the capacitors & automatically picks the measurement points from a pre-programmed measurement profile looking for min/max edge detection points.

Since a calibration has been done it is good for bias and linearity with standards not the actual product. To make sure it is good (repeatability) at measuring our product which has intra product variation (aka width varies by location on the part) I plan on testing out 30 randomly selected capacitors, all in the same dimensional familiy and run that 3 times. I plan on resetting and moving around the capacitors (while keeping track of pc idenity location). With the data I plan on running an ANOVA single operator analysis in Minitab and assess the %Variance Contribution of Gage and Tol% as this is used as product acceptance and not SPC.

Is this a kosher way performing this type of analysis. Please let me know.

 

[Miner]

Assuming that you correctly analyze the data, the design appears to be sound. Keep in mind that the study only holds true for that type of capacitor. To generalize it for other components, you will have to demonstrate with additional studies.

 

[QualityEngBrady]

Assuming that you correctly analyze the data, the design appears to be sound. Keep in mind that the study only holds true for that type of capacitor. To generalize it for other components, you will have to demonstrate with additional studies.

Hello Miner, yes I would be doing it for each different capacitor case size - aka the capacitors that are designed to have the same dimensions. Repeat study for different case size families.

And I would also like to add - I am not comparing it to a different vision system - we only use one vision system for this specific test.