Problems with True Position

[MrComment]

I have a program with suspect True position problems.

The attached pdf shows my report, from the default PCDMIS report template "textonly.rpt" just after running the actual part.

The v4.1 dimensions Dim#601 and Dim#597-X.
Versus the old style legacy dimensions Dim#601_Old and Dim#597-(1-4) seem to give conflicting results.

I dont know which one to believe, or what the heck Im doing wrong. Ive tryied a number of different alignment techniques already thinking it was the way I aligned to the feature or something, but Im getting wierd results regardless.

Looking at the CAD model on the screen after running the part CIR13 "Dim#601" doesnt appear to my calibrated eyes to be anywhere close to 4.311mm off from the nominals. I would tend to believe its more like .25mm off maybe.

The 597 dimension might be .45mm off, but I also would have a hard time thinking they are 2mm off. Especially since they pass with the GO Gage as well.

Anyone have any idea why these two dimensions wouldnt match, legacy vs new style? Or even be remotely close?

[ag1979]

If you post or send me the follwing:
1) Nominal x, y, z, i, j, k, size
2) Actual x, y, z, i, j, k, size
optional
3) Tolerances

I'll send you the true position values 2D and 3D.

[Jan d.]

You have an VC on your datum! Since version 4.1, PC-DMIS takes a virtual condition on a datum into effect. Before that, it always took the datums at RFS (you know, Kevin's really fuzzy stuff).

Redo the dimension and make your datum RFS (take the MMC off). Then compare the results.

If you have a VC on a datum, PC-DMIS will start to "wiggle" the simulated hard gauge to get you the very best TP. In the DATUM SHIFT line it tells you how far it wiggled the gauge around (in your case for DIM 601 I see very significant wiggle in X and Y plus a large rotation around your Z-axis). This is EXACTLY what VC's on datums are supposed to do! That is totally correct. Maybe a little surprising, but correct.

The "old" PC-DMIS" (up to even V4.0) does NOT do this wiggly thing. It therefore always gave results that have been too restrictive (in other words, you are failing parts that would really be good!). The idea being that VC's on datums are really hard to explain and therefore it seems as if you pass parts that really should not be passed at all. If your management gives you a hard time over this, go back to RFS! That'd be my advice.



Jan.