Minimum Arc Length (or degrees) for measuring a radius

[[Ja...]]

So we have a part with an .093" radius called out, but the arc length is .016",which is angle of roughly 10°. What it is the minimum arc length or angular dimension that will report a correct, and repeatable, value?

Area in question is the red face below. We are to blend it with a R.020 to the between points A and B (green).

[[Ma...]]

I would say min. 180° but i may be wrong 🤣

[[No...]]

Old and in german only, but still helpful.
The diagrams speak for themselves.

http://www.dr-hernla.de/Hernla%20Info%2 ... chnitt.pdf

[[Cl...]]

Please sign in to view this quote.

With an arc segment that small, you won't get accurate repeatable results by scanning.
I recommend using single points and constraining the axis.

[[Da...]]

I've had good results using curve for tiny features like that. The model provides the nominal shape, and a tolerance can be determined from that nominal. You can create a profile plot and "show" that the radius is in the correct area, and at the correct size. You won't get a numerical size but will have pretty good evidence that the radius is "good". And use single points as Clarke recommended.

[[DW...]]

Please sign in to view this quote.

Agree. We had an issue with a small section of a cone. Calypso would identify it as a cone, and when we would evaluate this small partial conical feature with a high density scan, we would get crazy results. A call to Zeiss, and they recommended when you have a small portion of a cone (may also apply to other normal geometric elements) to use free form surface or in your case curve. Someone with more knowledge can chime in, but I believe they said it had to do with the way the software processes free form shapes versus normal geometric elements.

[[Cl...]]

IMO these types of small radii should be evaluated as a profile. Making life much easier.

[[Da...]]

This was for a spherical radius, but it illustrates the process. The sphere is 2.2 inches and we are using about a .300 section. The sphere section radius is not the rad reported from the "sphere" feature, but instead it is the radius of a circle section applied to the sphere (as single points) . It's pretty cryptic but, it works quite well.