Reporting measurement uncertainty for custom items

[greif]

We typically use microscopes to locate the edge of images to determine size. These might be moving microscopes on a CMM, or fixed short distance like filar type eyepieces. Most of the time we measure chrome on glass images (very thin evaporated chromium on a glass substrate) and have well characterized uncertainties for measuring to this type of edge.

My question is: when a customer sends in another type of material (such as images on photo film, or photo paper) , we do not have an uncertainty for that specific material. What is normally done in such a case? Would it be reasonable to have one operator (the tech doing the measurement) measure a feature 10 times and use the std dev from that in the uncertainty calc to determine the k=2 uncertainty for that specific measurement? Or does it need to be more extensive?

 

[dwperron]

It would be similar to the uncertainty budget you are already using. Yes, finding the standard deviation would give you a Type A contributor for repeatability. You will still need to include the usual suspects: your measuring device's accuracy, its calibration uncertainty, resolution, environmental contributors, etc.

 

[greif]

Oh good- simpler than I thought it might be. Thought maybe an R&R and multiple users might be standard for this sort of thing.