Disturbing issue with interference check in ST6.

I have a troubling issue with the way ST6 evaluates interference checks.  I have an assembly that consists of a number of heavy steel plates weleded together.  Some of these plates are curved and have various cutouts where oither curved plates are welded in.  As such, these are modelled as sheetmetal parts and the cutouts are modelled in using the normal cutout feature using copied sketches driving the mating parts.  So there should be absolutely NO WAY that anything should intefere.


Now, when I do an interference check by selecting all the parts I get an interference being reported where there should be none.  The interference volume is shown as a tiny sliver of a line segment and it makes no sense at all.  I believe this is a bug.  If I do the interference check again and select just the two parts that were reported as having an interference then NO interference is reported.  This is quite disturbing and greatly diminishes one's confidence that a model in really error-free.  It also generates many unnecessary extra hours of work to make sure there really are no clashes in the model.


The way I see it there might be a number of things going on here:

  1. There is a problem with how SE calculates interferences? (seemingly already evident above)
  2. There is a problem with the normal cutout command?
  3. Maybe the precision of the geometry has been degraded in SE? (Accuracy = 1.0 in the parts concerned, so that is as good as it gets)


Now I realise we are chugging along in the 19th century with ST6 MP14 while most of you are probably on ST7 or even ST8 by now and this may no longer be an issue with those more modern versions of SE, but has anybody run into any issues with inconsistent or buggy interference checks?