Results change after update to 7.8.20

[[Pa...]]

So I've got these square-ish holes.  Requirement is a surface profile of .007" for form only (no DRF).  Additionally there is a position callout.  It specifies a .010" CYLINDRICAL tolerance zone relative to a fully defined datum set constraining all 6 degrees, and it invokes MMC on the FOS as well as MMB on the secondary and teriary datums.

I scan a curve at the top of the hole and another at the bottom. 

I recall these curves into freeforms for the surface profiles.

I then create a maximum inscribed cylinder recalling points from the freeforms.  This cylinder is the feature in my position characteristic.  We've been making these for like 17 years and until I wrote this we were using a customer supplied hard gage which uses best fit pins through the holes.  Makes sense to me.

I've been running this for a couple years and it always gives results which seem plausible.  We did correlate with a hard gage at FAI and all deltas.  I just checked and we still can.  I do understand that the hard gage gives attribute results (Pass/fail only) while the CMM gives variable results, and that correlation is implied but cannot be evaluated for magnitude.

Just got 7.8.20 installed.  I loaded measured points from a piece run on 7.4 the day before.  I ran plan/subsequent evaluation.  The first 19 pages of the report (which don't include the profile or position tolerances above) match within .00001" in every case.  I've got positions, diameters, symmetry plane distances, fillet rad evaluations... all kinds of stuff.  The update affected none of this at all.  Those positions and profiles generated from curves though?  I'm seeing way too much variation.  The worst hole got .0026" worse on position, the best got .0024" better.  Two did not change at all.  The other 28 holes had variation between these magnitudes when evaluated using 7.4 vs 7.8.  

Anyone got any ideas what I'm up against here?  Did I do a poor job creating those curves?  Am I just seeing the results of an evaluation requiring multiple calculations?  Was this discussed already?

In the screenshot, "show masked actuals" is turned on.

[[Pa...]]

I solved my problem.

The measured points used in my 7.4 runoff were not the same points used in my 7.8 so...

There was literally nothing wrong except for me.  Thanks for reading!

[[Ro...]]

You were able to download it? I keep getting the Terms and Conditions popup that won't go away no matter how many times I hit accept. 

[[Je...]]

Please sign in to view this quote.

This is a known issue and is being addressed by the ZEISS Portal team