Profile Text ON deviations don't match CAD of same point

[Aaronious1]

I've been coming to this forum for some time. Only now I have a question I haven't found an answer for by searching.

Hopefully I can make this question short. I should start off with saying that I'm not a CMM programmer nor play one on TV. I can do minor edits and understand (I think) enough to discuss things with the CMM programmers.

My problem simplified: Profile output "TEXT ON" when laid out in CAD does not match deviation showed by PC-DMIS, it can be off as much as 20% or .0015", sometimes more.

Basically a lathe produces a conical part with the OD being measured with a bar gage at a set height from the datum surface and the gauged diameter is conforming, angle is within expect range. Then the part is then rejected by the CMM which led to my investigation. Our profile data is output with "TEXT=ON" so for curiosity sake I entered all the points from the report into CAD (which was verified correct).

In CAD, the points are within the profile allowed. I've verified that the theo numbers for the points used in the profile set (x, y, z, i, j, k) are correct. Then I used the formula given by dph51 in this thread: https://www.pcdmisforum.com/forum/pc...urface-profile in Excel and the values given there match to the 6th decimal place what I get in CAD. Also, SNAP is set to YES on all the profile points. Also, we're using legacy profile.

Lastly, I verified the vector directions are not transposed, reversed or otherwise.

Any ideas where we've potentially gone wrong here or what I'm misunderstanding?

Aaronious1

[RIDER]

Are you dimensioning surface profile or line profile? Is it form only or form and location? Are you using ASME or ISO as the standard you are working to?

[Schrocknroll]

You say "the part is then rejected by the CMM" but you don't say how it's being checked. As a profile? As a gage diameter on a cone?

[JEFMAN]

Snap is on... I believe it's a bad idea, ecause you are assuming that the cmm goes exactly to the theo coordinates, except along the vector.
In adidition, you don't describe how you realized the alignment, so you can't compare to the CAD directly.
The CAD is perfect, the part isn't.
The CAD doesn't use any algorithm, the CMM uses one (this that you choose).

If you compare some T_values of a perfect cone which has a axis deviation, all the t_values will be off, just because of the alignment, but the form should be ok.
Create a cone with those points, then align on it and look at the form, it should match beteen text and CAD.