General Faults after shelling?

[kellne...@my-deja.com]

Anybody had a general fault crop up after a shelling operation? If so
what did you do to find it and fix it?

SW2000 sp1
NT4.0 sp6a
AMD K7 900
500 Mb
ASUS V6800 GeForce 32Mb

Sent via Deja.com http://www.deja.com/
Before you buy.

[wro...@my-deja.com]

Did you try cutting away the model to see where it was failing?

[joel...@my-deja.com]

I've seen this too,

It could be the slightest little sliver or cornered edge that won't
allow the shell to work properly. Sometimes I go through each feature
using the scroll bar to insure a clean model before shelling. This can
be quite time consuming for models with large feature trees. Good luck.

JL

In article <8gn65i$2h7$1...@nnrp1.deja.com>,

[matt]

also try removing/adding fillets, particularly removing small fillets and
adding large fillets. Small / large in this case means relative to the
shell thickness dimension. The "cutting away" method mentioned by
wronsky is a good one, checking the shell for failure every inch or so to
narrow the problem down. Also, don't forget the Tools/Check to check for
short edges and minimum radius. This is a good troubleshooting aid for
questionable geomtery. If it's not good to start with, shelling it isn't
going to help. As another tip, try turning on the "verification on
rebuild" switch (tools/system options/performance). Don't forget to turn
it off when you're done, it will slow you down to a crawl.

matt.

kellne...@my-deja.com wrote:

> Anybody had a general fault crop up after a shelling operation? If so
> what did you do to find it and fix it?
>
> SW2000 sp1
> NT4.0 sp6a
> AMD K7 900
> 500 Mb
> ASUS V6800 GeForce 32Mb
>

[kellne...@my-deja.com]

Thanks for both the replies.

I am still not sure what a general fault means.

The suggestions both of you had were good ones and I recommend them to
others. My practice is to shell after every feature or group of features
as the case may be. I have a hot key setup for Check and I use
verification on rebuild alot. Many times verification on rebuild will
cause the model to fail after shelling. Another little trick that will
pick up subtle errors is to write a step file. Step files are more
persnickety than SolidWorks is about model integrity and if you can't
write one there is a problem.

By cutting away I narrowed the general fault down to a section of fillet
that the shell operation created on the inside of the model. This wasn't
a narrow sliver, just what appeared to be edges of faces that SolidWorks
couldn't quite bring together.

My solution was to redo the method I used to create the loft that was
being filleted eliminating the partial ellipse tangent to line construct
in the loft curves. The bad fillet was crossing one of these lofted
tangencies. I have had vary bad luck with lofts that have tangent
ellipses in them.

You are right about this being time consuming. In CPU time alone I have
racked up about 6 hours in 16 hours at the tube. Note what I am running
on too. I estimate my "old" PIII 500 would have racked up 24 to 48 cpu
hours to do the same work.

SW2000 sp1
NT4.0 sp6a
AMD K7 900
500 Mb
ASUS V6800 GeForce 32Mb

In article <39343DCA...@cadimensions.com>,