Max Concave, Max Convex Strategy

[[Er...]]

I have this callout on this flatness. What strategy would you use measure the max concave, and max convex requirement? As of now, I am leaning towards a manual inspection where the indicate the high and low spot and record the result.

[[Da...]]

Regarding the question uniaxial or biaxial, a Fourier analysis springs to mind. Don’t really know if it would also allow to mathematically determine the direction of the wave surface, but I would guess that’s the area you should look into.

Alternatively, you could always add graphical views of the surface for anyone reviewing the report to determine the concaveness/convexness for themselves. If you project the plane profile into two different line elements, you could evaluate two different straightness tolerances for each direction. If the deviations are very continuous, you might evaluate whether your extreme points are on the outer edges of your profile or more in the center.

P.S.

I would strongly suggest to make prototype workpieces with concave and convex deviations both in one and two axes and one that definitely exceeds the allowed deviations (a warped surface), just to test the evaluation. Doesn’t have to look like your final part at all, it’s just about the right strategy.

[[Ri...]]

Please sign in to view this quote.

Wouldn't you just be able to use a Profile with unequal distribution for this type of analysis?

Not a fan of surfaces that have a flatness spec, and then allowable concavity/convexity. Had a customer that wanted flatness held to 0.020mm across the entire surface, but then wanted the concavity set to 0.015mm nominally. Made my head hurt.

[[Ma...]]

It is not uncommon at all for us to have concave only flatness specs but ours only occur on circular shaped planes. In that case we can simply compare the height value between inner and outer circular scans to determine if the surface is convex or concave. This is a much more difficult analysis problem to attempt with Calypso.
If the values can be determined by the CMM operator visually from a flatness plot, perhaps the most expedient solution would be to input the results via result element at the end of the run.