V2016 Parallelism error

[BKulpa]

Working offline using a CAD model and dimensioning a plane to plane parallelism and it's showing 0.074mm error.
Tried exact measure and legacy, both show the he same thing.
Dimensioned flatness to be sure I'm not hitting another surface by accident.
All planes measure 0 flatness error. Vectors are all 0,0,1 or 0,0,-1.
I have this problem on 2 different plane to plane dimensions. Just makes me wonder if my actual results are going to be valid when I can go back to plant.

[JEFMAN]

You can try
ASSIGN/V1=MAX(DOT(PL1.HIT[1.PL1.NUMHITS].XYZ,PL2.XYZ))-MIN(DOT(PL1.HIT[1.PL1.NUMHITS].XYZ,PL2.XYZ))
ASSIGN/V2=MAX(DOT(PL2.HIT[1.PL2.NUMHITS].XYZ,PL1.XYZ))-MIN(DOT(PL2.HIT[1.PL2.NUMHITS].XYZ,PL1.XYZ))
It will give you 2 arrays of values.
If one of them is not an array of zero, then the defect is on the plane which uses hits.

Are you "measuring" with a 0.5 mm ball ?

[BKulpa]

Nope measuring with a 2mm ball, all offline. Should be perfect.

[JEFMAN]

Yep, that's why I used quotation marks on "measuring"...
I was thinking about a bad compensation of the ball radius on one hit... But it's stupid, flatness wouln't be zero in this case...

[Schlag]

If you take individual vector points on those surfaces are you still seeing perfect vectors ? Have someone with access to CAD look at the model. These would be my 1st starting points.

[BKulpa]

I checked the theoretical values of the points I measured in the routine, all seems well. My only thought is one is scanned and the others are ttp. The 2 TTP planes dimension parallel to each other (0).
Maybe it's the filtering in the plane composed of 3 rotary scans. But the flatness of that is zero and the vector is 0,0,1.