Large Assys - SW Not Proven at All

Kman,

No, this customer doesn't typically open the entire 60,000 part top level assembly. They work smarter, not harder.

The engineering manager for this group came from a Pro/E / UG background. He educated his SW users in the proper techniques to handle large assemblies (skeleton parts, configurations, simplified representations, etc.).

They use top down modeling techniques where required, then break the external relations when the design has stabilized.

Regarding you comment about sales people, while I agree that some fall into that category, I have been doing this for 18 years, and am a certified instructor and support technician. I was recently recertified for SW2005.

Some of us have a conscience, pride, and respect for our customers.

CG

Reply to
John Picinich
Loading thread data ...

The problem with the part Template, was with cosmetic threads, not knowing if they wanted to show or not, lol The feature (texture), was disappearing, when the part was rebuilt or shaved.

The problem with the Drawing templates, was that adding dimensions, was pulling up the over constrained feature manager! and telling me that the part was over constrained. When opening the part window, the part had changed and needed updating, lol

The problem with the assembly Template, is the time it takes to open, even a

5 part assembly!

I have since re-made the templates in 2005 sp 2.0 and everything seems to be much better, but I will wait until I have rebuilt the whole assembly on this service pack, before confirming everything is ok again.

Reply to
pete

Of course you are correct about there being good sales persons. I do have respect for many of the SW support as well as direct. Can't say so much for past AutoDesk support and service (just thinking about it raises my blood pressure, take a deep breath, count to ten, better). The Cad Guy's statement was vague and could easily of been interpreted as SW has no problem opening a 60,000 piece assembly in its entirety. Then there is assembly drawing performance and that is another topic to its own.

I also use some of these techniques to mentioned to some degree or less. The artistic in me doesn't like modeling a spring to look like a solid cylinder. Or removing the detail from commercial parts to a point they loose their identity. My simplified SHCS look like they could be a headed rivet, headed punch, trailer hitch pin or maybe a tooling button. Too me models and drawings should look like what they are and be easily interpreted by others not so trained in the field of design or engineering. It takes considerable effort and time (non value added) to create configurations to suppress fillets, fasteners, chamfers, holes etc...

Unfortunately a necessity given the software limitations.

In our business, we design custom assembly, weld and metal removal equipment for automotive front and rear axle assemblies and various interior sub-assemblies. These non-native assemblies cause all sorts of performance issues in SW and SolidEdge for that matter. We have suppressed parts, translated into parasolids and still the performance is generally poor. Add to this fixtures, imported commercial components and fabricated parts and you get the picture. I would be interested in learning other ways to improve performance when importing non-native models.

Kman

Reply to
Kman

Check out Solid Edge from UGS. The newest release (V17) has some new features specifically geared towards massively large assemblies.

Reply to
Ken

Any empirical data to support those big adjectives? You need to coax them into publishing something. If nothing else they can at least add to inventor's trouncing unless all that Functional Design stuff helps 'em (darned sure don't want to overengineer all those cubes). 8~)

Reply to
Jeff Howard

I haven't seen any numbers, but does that really mean anything? They can tell you anything, but untill you try it yourself...

Reply to
Ken

"They" as in the numbers, or as in the sales reps?

There has to be some potential for ROI apparent before I'll spend the time. Haven't seen anything yet to indicate there is any. Sooner or later something will pop up. I'm sure there's a lot of this "testing" going on behind all the various curtains. Of course no one's gonna publish anything that indicates the other guy wins and there has to be justification for buying the ad space...

Reply to
Jeff Howard

In my experience SolidWorks is a huge resource hog.

My performance reference is Pro/E.

SolidWorks slows to a crawl displaying hidden line or wireframe views, even on relatively simple parts. Drawings are slow to manipulate. Assemblies over about 500 parts slow down to a crawl. File sizes are huge.

There are tricks to help but these are basic problems.

I use SW every day and I like it, but it is not the kind of program you would do anything too serious with (no cars, space shuttles, planes, or nuclear submarines, etc.).

Part of this is due to the Windows-PC platform to which SW is married.

I'd suggest looking into a very fast 64-bit UNIX workstation(s) (like a Silicon Graphics Onyx :) and a matching UNIX server and running one of the high end CAD packages like UG or IDEAS or Catia. As always benchmark first!

Regards,

Reply to
Anonymous

Miscrosoft will release a new 64-bit XP operating system fairly soon. It will certainly have its problems at first, but eventually that should improve performance significantly by increasing the amount of available RAM space.

That might be a better route than adding a ton of UNIX hardware and having a hybrid OS/net environment.

James

wrote:

Reply to
James

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.