Theoretical Features Vs. Non-Theoretical Features

[[An...]]

Hi all,
I recently started at a new position and have inherited a program that I am hoping for some thoughts on. I believe it's a relatively simple question. I unfortunately do not have access to the person who wrote the program.

The program checks an ID and does this by scanning two circles, creating a theoretical cylinder using Recall Feature Points with these two circles, and reporting the diameter of the theoretical cylinder as the ID.

Were I to have created this program, I would create a (non-theoretical) cylinder with 3 circle paths and report the diameter of this cylinder as the ID.

My question is this, then. Is the approach that the original program uses appropriate? What are the differences in calculation between defining a theoretical cylinder in this way and the way that I would have done it?

If more info is needed, I will do what I can to provide it.

[[To...]]

Please sign in to view this quote.

One thing about the diameter of your cylinder is, it's the average diameter of the 3 paths. Some people like to know what the diameter is at the local area measured. I generally create the cylinder like you are doing, then create 3 circles, recalling the feature points from the 3 paths. Half a dozen of one, 6 of another.

[[Ma...]]

I would point out, that using cylinder in base alignment is not trustworthy with manual alignment. It can easily flip axis direction.
To avoid that you use circles - so maybe this can be previous colleague explanation.

[[Ca...]]

Hey Andy!

I am wondering about the depth of the cylinder. It is often more stable to measure circles then recall into a 3D line for a base alignment, as Martin was mentioning as well, if the depth of the cylinder is less than the diameter of the cylinder. This is because the circles are constrained to a plane, then you can get the center points, then recall into a line. Are those circles being used elsewhere in the program, maybe in a 3D Line? If so, then it would be just as easy to recall those circles back into a cylinder rather than measuring the cylinder again.

[[Ow...]]

Cylinder features can mathematically be tricky and right or wrong, bad results have left many programmers to use circles instead.
The ratio of Diameter to Length (as previously mentioned), as well as the measurement strategy and cylinder form can cause cylinder calculation challenges.

Solving a circle is not very challenging, given the orientation of the circle is controlled by the alignment and nominal A1/A2. So, circles are relatively bulletproof, and creating a 3D line from the fit centers works very well to create a reliable axis.
For best results on cylinders, always scan at least 3 equally spaced circles throughout the length of the cylinder. The Zeiss strategy cookbook (supposedly now given out for free at class, I had to pay for mine☹) does a good job at defining the length vs amount of scan paths needing to be used and datum constraints for those not deep enough.

The Stepped cylinder feature is nice for reporting the diameter and X,Y,Z positions of ALL the scanned circles used in the cylinder feature but, won’t report the roundness of each scanned path, maybe one day it will 🙄

[[Ri...]]

Please sign in to view this quote.

I have used this method before and do not recall any issues.
I do think that "Recall", "Recall Feature Points" and "Theoretical" are three sperate entities.
In particular, theoretical. A theoretical feature will not have any of the form characteristics. Recall and Recall Feature Points will.

I also agree with Martin Jánský, adding 2d Lines.
It is one of the reasons I believe the Base Alignment should be no more than describing to the CMM where something exists in its volumetric area.