Checking a cylindrical profile

[Mr_Cris]

Hi folks,
Somebody put a +-.0005 tolerance on the radius of a small arc segment and I am hoping to convince them that using a profile tolerance would be more ideal.

I want to check the profile of an arc without using datums. My part looks kind of like a chunk of tubing that was saw cut into segments. What I need to know is how much that arc deviates from a 3.5" cylinder. So far, I have scanned the top and bottom of that arc, using linear open scans, and constructed a cylinder out of that using the least_sqr type. No problem, but now I just need to figure out how to compare that profile to a 3.5" cylinder. I don't really care about the position of the cylinder, I just need to know about the form.

Any suggestions? If it matters, I am using v4.1 using and analog scanning probe.

[spazus_maximus]

Roundness & Diameter?

[Mr_Cris]

I'm not sure. I know my part is not perfectly round, which makes measuring the radius to the expected precision impossible.

I have tried to use the cylindericity feature, but that only compares the profile of my cylinder to the best fit diameter. I want to check the profile against a perfect 3.5000" cylinder.

[Don Henger]

I'm not sure. I know my part is not perfectly round, which makes measuring the radius to the expected precision impossible.

I have tried to use the cylindericity feature, but that only compares the profile of my cylinder to the best fit diameter. I want to check the profile against a perfect 3.5000" cylinder.

If you want to compare the differance between a perfect 3.5". Create a 3.5" cylinder or circle at the exact location of the one you have measured and the circle lines should show you the differance between the one you created and the one you measured. (just for a visual look) & (just a thought)

[Mr_Cris]

I think I got it

I think I got this licked. What I did was use the Best fit find nominals method for finding the nominals of the scan, that gave me the 3.5000" profile I was looking to compare the arc to (CAD file was needed for this). Then, I reported a surface profile using the form only option with a best fit comparison. This way, although the radius was reported to be oversized (meas: 3.5025). The profile fit a .0010" profile band.

I attached a copy of the graphical analysis.