Recalled 3D line evaluation

[[Jo...]]

Hi all,

Thank you all for the wisdom and indirect help, I am still a bit of a novice with Calypso but have found direction here frequently.

Please help me understand the behavior I am experiencing. I recalled 2 circle paths on the ID of a cylinder into a 3d line. Both circles are evaluated as max inscribed features with standard filtering and outlier elimination. The 3d line showed it deviated by .015" from nominal (for this part is huge), top circle center is within a few 10ths of nominal, and the bottom circle was off .002" in X from nom. How is the line between the 2 centers so much farther off? When I change the 3d line evaluation method form LSQ to Min feature, it is then 10ths from nom.

background:
Calypso 18, XXT RDS head, 3mm x100mm probe
(Overall, I am comparing the CL of features cut into the OD of this cylinder to the CL of the cyl. I have been getting numerous funny results, so was trying different methods looking for one that made the most sense.)
The base alignment is the cylinder ID, top plane, and circle in an alignment hole for clocking. I was comparing the 3d line to the cyl CL to ensure other CL measurements were sound. In this case, reporting the x value of deviation from CL.
Thanks,
John

[[An...]]

See attached.

Contribution_13_03_2020_.pdf

[[SH...]]

Please sign in to view this quote.

He told it is a circular path,so the point density should be same, how does the point density change??

[[An...]]

The point density only changes in this very unrealistic simulation.
Just wanted to show the functionality of "3D line" and "Recall Feature Points".

[[Ja...]]

"Recall feature points" of a cylinder will consider all points. "Recall feature"(not recall feature points) and select two circles (not a cylinder) will give you a line from the center points of the circles.

[[Jo...]]

Thank you all for the info! I ended up finding it was the difference between recall feature and recall feature points.
Thanks,
John