Re: CAD software?

From: Michael Brindley, Chuck Kirschman (Clemson University)
Date: Friday, January 27, 1995

From: Michael Brindley, Chuck Kirschman (Clemson  University)
To: RP-ML
Date: Friday, January 27, 1995
Subject: Re: CAD software?
> From: brindley@ECE.ORST.EDU
> >      AutoCad makes a pretty decent file, considering the low cost of the 
> >      system.  Curved surfaces, however may be a bit course for top of the 
> You didn't state which version of AutoCad you were talking about.
> R12 had an 'unsupported feature' for making STL files (an AME 2.1
> Lisp program).  The new version, R13, completely changed it's solid
> modelling software (it uses the ACIS engine).
> 
Ah yes, but not perfectly.  Through my experiences at Clemson and other
places, I've dealt with a lot of CAD packages, and I've often been 
responsible for declaring _exactly_ what is wrong with the file.

I can say that all venders seem to be improving, but several
have not yet perfected it.  The level of commitment varies
from company to company, depending on who's pushing them.
Your milage may vary.

Here are some of my observations (on more known CAD packages):

Autocad leaves holes.  V12 is worse than V 13.  Meshing is
sub-optimal.

Unigraphics and Intergraph do not always trim surfaces correctly.

I have received only 1 bad ProE file.  I also like the meshing that they
do.  Plus the "fine detail" meshing option.  

I've never received a bad Aries file, though I've had some difficulties
in creating them on occasion.

CV had about the worst I've ever seen, so if you must go the CV route
I recommend a call to Brock Rooney.  I've also had very good luck
with his IGES translator STLs, though I've never used the package.

I-DEAS did not follow the format very closely.  I hope they've fixed that.

MicroStation is one that the poor man should look at.  Bently Systems is
now doing their own marketing rather than hanging with Intergraph.  The
mesher is good, and the STL's were perfect from the Beta I was using.
Much slicker than ACAD, and cheaper too.

Most of the rest of the systems I deal with are somewhat limited in
distribution, and I don't always know where the files come from, only
that they are bad.

Another point is that various RP (HF?) systems have different tolerances
for the errors.  So, a bad STL is sufficient for some systems, but
will not work at all for others.

Again, all opinions expressed are my own from the many files I've
diagnosed.  I don't have detailed notes, but certain things stand
out over time.  Everything in any package may have been fixed yesterday.

If you have specific problems/questions, feel free to ask.

chuck


Previous message | Next message
Back to 1995 index