Part Accuracy/Tolerance 2

[mloew]

I am trying to change the part accuracy/tolerance in SolidWorks 2007 SP3.1. I did a search in the help files, on this site and with Google and did not find what I was looking for. I found settings for the rendering quality, but not the part quality. In Pro/ENGINEER what I am trying to do is called accuracy (Edit, Set-up, Accuracy). There are then options for relative or absolute accuracy.

What is the SolidWorks equivalent and how do I set it?

Thanks in advance.

Best regards,

Matthew Ian Loew

http://www.daxcon.com

 

[TheTick]

I know what you're looking for. It's not there.

What exactly is the problem?

 

[CorBlimeyLimey]

The Tolerance Strategies section explains briefly;

http://www.iti-oh.com/Education/Articles/BadThingsCad.htm

 

[TheTick]

The article misses some very important points.

[ul][li]It completely ignores problems that can be caused by the Pro/E scheme. Especially true for very large models that have some features with very small faces (think large molded part with a few tiny ribs)[/li]
[li]It also completely ignores the fact that UG (and SE, I think) do, in fact, have adjustable tolerancing. UG's can even be adjusted on a face-by-face basis.[/li][/ul]

Anyhoo, I don't want to further detract from mloew's needs. It would be helpful to know the exact nature of his difficulty.

 

[ewh]

UG also allows model checking for many of the geometry flaws the author mentions. I wonder how much stock in CADIQ he owns...

Believe it if you need it or leave it if you dare. - [small]Robert Hunter[/small]

 

[mloew]

Tick,

You hit the nail on the head: I have a physically large model (major feature element is an 85ft diameter weldment) with ribs that have features with geometric details on the order of inches. The main circular feature is rendered with visible faceting in the model and shows gaps in the model where a cylindrical surface intersects the planar circle. This is a surface model that is being used for generating a FE mesh (in HyperMesh via, unfortunately, STEP). The sketched geometry displays well, the surfaces, however, do not. If this was just a cosmetic issue, I would not complain (as much). But it seems to effect the quality of the exported file that is then meshed.



Best regards,

Matthew Ian Loew

http://www.daxcon.com

 

[CorBlimeyLimey]

This article is a couple of years old, but should still be valid;

http://www.solidworks.com/swexpress/dec02/200212_techtip_03.cfm

 

[mloew]

I'm going the other way: the SolidWorks data is native and I am exporting out in a STEP format for HyperMesh to read.

The part just looks awful inn SolidWorks with gaps and protruding surfaces...

Best regards,

Matthew Ian Loew

http://www.daxcon.com

 

[TheTick]

Are they two separate bodies, or are the surfaces sewn together?

 

[mloew]

Independent surface features; i.e. not stitched. Would that help? There are no solid bodies.

Best regards,

Matthew Ian Loew

http://www.daxcon.com

 

[TheTick]

SW renders surfaces with less resolution than solid bodies.

Any reason not to use IGES? I think SW's IGES export might be better than its STEP. The IGES export options also have an option for high trim curve accuracy.

 

[Heckler]

I wonder if Relative Accuracy & Absolute Accuracy is something SWx plans to make a user setting?

Heckler
Sr. Mechanical Engineer
o
_`\(,_
(_)/ (_)

This post contains no political overtones or undertones for that matter and in no way represents the poster's political agenda.

 

[TheTick]

Things you could try:

1. Stitching the bodies together

2. Using the bodies to trim one another. You may need to extend the surfaces first.

3. Try other formats (IGES, ACIS)

Compounding the problem will be the fact that FEA programs don't always have the best geometry drivers. This was true with MSC NASTRAN when I used it.

 

[CorBlimeyLimey]

You may have to 'round trip' the file to fix the surfaces, which is where the tips linked to in my second post might be useful.