Posted: 08 October 2008 at 8:22pm | IP Logged

> Unigraphics

Eng-Tips has a fairly active UG/NX forum. 
NX 4 is pretty old.  Look there is you're curious.

> Pro/E’s default accuracy itself is 0.0012 microns
> when other packages playing with 30 microns default.

Pro/E's default (Relative) accuracy is a unitless value of 0.0012.
_ That value is a factor in an equation used to determine an 'effective'
  accuracy in model units used for feature creation / regeneration.
_ It is "default" only by virtue of being what 'out of the box'
  templates are set to and being recommended for 'average use' though I
  understand that factory recommendations are changing / have changed.
_ That value does not affect the accuracy of some calculations, i.e.
  explicit point positional tolerances and where certain analytic
  geometry types are involved, which are always done to floating point
  accuracy minus a few decimal digits rounding and noise buffer.  
_ If you are using Relative accuracy and want to see what the real value
  is look at the minimum resolution on an IGES or STEP export or an IGES
  export log.

Bottom line: there's little or no difference in the inherent accuracy of
any of the 'brand name' geometry engines. What you will find is that one
or the other tends to create 'cleaner' geometry at a given tolerance under
some or other specific conditions. That's more a function of specific
proprietary geometry creation algorithms.

There may also be some differences in extreme units (like microns) if
internal calculations are done using a specific unit rather than model
units.  Autodesk Inventor, for instance and if I understand correctly,
does all internal calculations in cm rather than model units.  Anyone
know if that's true of UG?  I should probably ask in a UG forum, huh? ;^)

Posted: 09 October 2008 at 4:39am | IP Logged

> I dont want ask UG forums because they don"t have experience in ProE

Uh huh.  Actually there are users posting in this group that also post
in that group and I think a couple of Pro/E users posting there that
don't post here.  The idea, though, was ("look there") to spend a little
time reading instead of asking.  You'll probably learn more.

I thought "ProE vs Unigraphics .....the winner is??????" was the joke.

Posted: 09 October 2008 at 9:07am | IP Logged

I've not used it personally, only heard from co-workers.  We have a seat here for our work with one client.

What I understand is there are two very different modeling philosophies at work here.  Pro|E's is parametric, with one feature driving later ones.  The relationships created are persistent, meaning if you sketch A tangent to existing geometry B, if B changes A will change with it and remain tangent.  UG is what's called Boolean and those relationships aren't necessarily persistent.  You can create A tangent to B, but if B changes you need to know the ramifications and go change A yourself to maintain the tangency.

That's my overly simplistic, and likely incomplete, understanding of the difference.