Friggin Cones

[[Da...]]

I had all this figured out. I knew that roundness on a cone , just didn't work very well. We have 5 or 6 air table roundness testers, and 2 air table cylindricity testers. They match amongst themselves but never came close to the Micuras. I have a pretty strong theory about that but, that's for a different thread.

So, I ran a small cone. Just 2 circle paths at a diameter around .050", cone angle is 60 degrees. Recalled feature points (one of the circle paths from the cone) in to a circle. Filtered and outlier as usual and asked for roundness. It checked pretty good !!

BUT , when I looked at the diameter size of the recalled circle , it was around .030" larger than any part of the cone. I'm using a 1.0MM probe. Why so far off ? It doesn't look like a ball center issue. Any Ideas?

[[Ju...]]

If you're scanning partial arcs on the cone(s), try evaluating them as L1 features. L1 evaluation is less sensitive to outliers.

[[Da...]]

Full circles, Outlier and filters are working fine !

[[Jo...]]

Have you tried "circle on a cone" ? Maybe it's a vector thing

[[Da...]]

circle on a cone produced a huge roundness reading and run out that was over .002 ! It checks less than .0002 on the concentricity gage.

I don't use circle on a cone.

[[Ju...]]

Is it off by the same amount as the sigma value of the probe measuring it?

[[Da...]]

It's off by .030" roughly .

[[Je...]]

Please sign in to view this quote.

.

Dave, I'm not sure if this will help, but I tend to evaluate roundness on a cone using the Form characteristic. I evaluate per section and also a cumulative result for the entire cone.

.

.

[[Da...]]

Jeff! I have been doing that for a while now, It's not bad, but still not quite right. Luckily, I have alternate equipment here that CAN check roundness on a cone.

[[Je...]]

Have you used "Cone Calculation" to evaluate the cone at a specific location/diameter?

[[Ri...]]

[[Ri...]]

And in Calypso 2023, with the new GD&T engine you will no longer need to do a recall feature points to do a segment form characteristic - you can do the segment analysis directly inside of the characteristic. 🙂

[[Je...]]

Please sign in to view this quote.

Now, that is pretty sweet! Thanks for mentioning it.

[[Da...]]

Please sign in to view this quote.

Using cone calculation gives a size at a location, or a location at a size. Since we cannot project form, the roundness will return as 0.00000. You need actual points to report roundness.

[[Je...]]

Please sign in to view this quote.

Absolutely correct. Your question however was how to report size accurately as you had already achieved a roundness report. Using actual measurement data for your roundess (recalling points from a segment from the feature) and cone calculation for size at that location achieves this.

Also using CY23 with the GDT Beta (as Richard referenced) may give you the results you are seeking.

[[Da...]]

Please sign in to view this quote.

You are correct, Sorry, this post and my other post about roundness are both about the same problem. They are blending together 🤣