Best STL output

    
  Subscribe Topic

Rank: 1

Paul

Newbie

posts: 0

Registered: 2011-7-12

Message 1 of 11

 Best STL output
05-03-2007 11:35 . am   |   View his/her posts only
I am having difficulty in exporting die STL'S to use for forging simulation in DEFORM.

Many times there are hundreds of open edges in the STL import even though the VX model is sealed.

There actually is documentation with VX that recommends checking the close gaps with Refine Facets. Unfortunately I haven't had any luck with this.

Are there optimal settings that I should use for export?

Are there any options that might easily clean up tangency or bad edge violations?

Rank: 1

Kevin

Newbie

posts: 0

Registered: 2004-4-26

Message 2 of 11

05-03-2007 12:01 . pm   |   View his/her posts only
I have had issues with VX stl exports in the past. I tend to use the Materialise NextDay for initial quotes if nothing else, but that system has a very precise STL definition and sometimes VX stls show up bad edges etc. I am quite sure the actual STLs are OK but perhaps the simulation software you describe is as fussy with the stl? Try the Materialise Next Day software and runa few VX STLs through it to see if they highlight the same issues. If they do then there is a strong chance it is VX in some way or another.

Rank: 1

Paul

Newbie

posts: 0

Registered: 2011-7-12

Message 3 of 11

05-03-2007 01:42 . pm   |   View his/her posts only
We personally don't use Magix. I have e-mailed DEFORM support to see whether the VX generated STL errors are the same in DEFORM as they are in Magix (because they do own this software).

I IGES'd this to Solidworks in the hope of patching together a more useable STL. There were no open surfaces but a lot of surface defects.

Rank: 1

Mike

Newbie

posts: 0

Registered: 2002-8-28

Message 4 of 11

05-03-2007 04:25 . pm   |   View his/her posts only
Paul,
We have spoken on the phone about this. If there is still an issue I would like to see the file. The developer of our stl translator has viewing software though I don't know off hand what he uses. According to him, of the two numeric fields, you should only change the Max mesh size.

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 5 of 11

05-03-2007 06:20 . pm   |   View his/her posts only
Hello Paul
I have come across this problem with other software. On that occasion, the problem was in fact that the VX output, which is optimised for rapid prototyping solutions (that is after all what the STL format is intended for), was too fine (too accurate). It may very well be the case that DEFORM actually requires a large-sized mesh. Check your DEFORM help for guidance on mesh size. It may also be the case (this is certainly true of other sim programs) that too much fine detail in the model can be a problem for the sim software. If there is any insignificant detail in your model that can be suppressed that might help matters - again, consult the DEFORM help.

Rank: 1

Paul

Newbie

posts: 0

Registered: 2011-7-12

Message 6 of 11

06-03-2007 10:05 . am   |   View his/her posts only
I have been trying a few things to be able to observe the effects.

One STL that would not seal up as an STL from VX sealed up when I exported the model as an IGES into Solidworks, and exported it from Solidworks as an STL.

Yesterday I ended up pretty much hung up when I tried exporting an STL with a mesh modified from 0.0 to 0.5. (I gave up after waiting over an hour) Today on a hunch I changed the mesh size to 500 an exported the STL without an issue. I haven't imported this into DEFORM for trial (as it is presently in use), but it does look clean in Solidworks.

Why does the 0.0 display similar results as 500? What would be the difference? It still seems to show all the detail that the 0.0 setting had. Does this modify the little details? I am crossing my fingers that this will produce better results.

Chris what would you mean by suppressing detail?

I talked with DEFORM in regards to the type of export that VX performs. He believes that there might be one type of export that would be optimal with Rapid Prototyping that wouldn't necessarily be optimal for finite element analysis.

I do have a large file that could be sent for observation. The model is sealed but there are some tangency issues. This will not fit in an e-mail, if a devoloper would like to look at this is there a place that I could send this to?

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 7 of 11

06-03-2007 10:54 . am   |   View his/her posts only
Hi Paul

By suppression I mean that, in your history list, you can temporarily omit features. Typical omissions for FEM are small fillets, but features such as holes and bosses might also be safely omitted, it depends entirely on the aim of the FEM process. The actual maximum mesh size used to export an STL is still determined by the physical size of the objects faces - if you tell VX that the maxium mesh size is 1ft and the largest curved face is 1/2inch, then clearly the max is not going to be reached From what you have told us so far, I don't see that there is an issue with VX's output. It is possible for an STL to appear to have edge stitching flaws when it arrives in another product, but that will normally be to do with how the other product builds an STL file on import. Certainly, most FEM programs will apply some simplification during import where possible.

Rank: 1

Paul

Newbie

posts: 0

Registered: 2011-7-12

Message 8 of 11

07-03-2007 12:21 . pm   |   View his/her posts only
In our application removing detail would not actually work. It's actually in these details that we can observe potential defects. I was just wondering if there might be a simple way to simplify the detail that already was there.

Where I noticed an issue with VX is where I could take a VX IGES into Solidworks and export an STL with workable results where-as the VX STL would not build a mesh which is necessary to achieve die stress calculations.

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 9 of 11

07-03-2007 03:55 . pm   |   View his/her posts only
Hello Paul
If that is the case, it should merely be down to finding the best mesh output from VX. It is an overall advantage to be able to adjust the STL output because you can realise the best possible STL for different requirements. In my experience, FEM is not so good at fine detail - be wary of false positives.

Rank: 1

Kevin

Newbie

posts: 0

Registered: 2004-4-26

Message 10 of 11

08-03-2007 04:49 . am   |   View his/her posts only
It would almost be worthwhile to have a better STL mesher in VX that can distinguish between large areas and small details - perhaps by assigning specific mesh parameters to specific regions or features? I know the problems you are having Paul as the FEA you are doing is for press/stamping (to see if the part can be made by pressing)? As opposed to strength analysis FEA for design? One of my customers does this and it is the details they are interested in as it is those areas that generally cause the problems in that the metal can't be formed with sufficient detail in one hit (hence requiring a multi stage operation). The system they use though uses native CATIA v5 data and they can assign specific mesh parameters to zones on the sheet.

But getting back to the stl issues one of the problems I see in the use of generic stl meshing is that if your part has subtle curved surfaces and lots of fine details as well you need to make a very fine mesh to define the subtle zones but this is total overkill for the detailed areas where a looser mesh would be more than enough for most SLA builds. Does any CAD system do this? If not then it would be a good VX advantage.

BTW if you end up with a monster stl file (for the above reasons) try using Materialise's STL ZIP application (free) as this is fantastic at zipping any stl file to a fraction of the original size - far better than WinZip for example. Typicalling a 20MB stl will WinZip to maybe 10MB. In STL ZIP it is under 1MB.

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 11 of 11

08-03-2007 07:40 . pm   |   View his/her posts only
Hi Paul

DEFORM has it's own mesher - if you are really concerned about fine details on the model, then I would be looking at exporting an IGES model and using DEFORM to produce an optimised mesh from that, which should be more efficient than STL. I expect the SFTC people can guide you to get the best from their product with this regard.
See also
X