Improved filleting?

    
  Subscribe Topic

Rank: 1

Paul

Newbie

posts: 0

Registered: 2011-7-12

Message 1 of 8

 Improved filleting?
11-08-2005 01:44 . pm   |   View his/her posts only
From time to time I run into a problem (such as just now) where I try to apply a solid fillet, the fillet actually previews, but the command errors out when you try to place the fillet. In this case it does not even allow for a partial result. The intelligent blend will place a fillet, but only in several segments. On top of this it is necessary to trim all the edges, and sew everything up.

In this Spring's MUG meeting I heard that some big advances were under developement in release 12 for solid filleting. I checked the release notes for version 12 and didn't see anything regarding solid filleting, can we expect any big filleting advances in version 12?

It sure would be nice to be able to preview everything and if it has problems with a specific area be able to tweak these areas.

Rank: 1

Robert

Newbie

posts: 0

Registered: 2011-11-22

Message 2 of 8

12-08-2005 11:45 . am   |   View his/her posts only
Paul, you heard correctly... lots of new functionality coming soon in Version 12.

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 3 of 8

13-08-2005 01:19 . pm   |   View his/her posts only
Hi Paul

Nine times out of ten, the fillet failure occurs when the command is unable to auto-trim all of the model faces. There will always be cases where this is so. However, instead of a partial result, what we hope to offer is an intact fillet surface (i.e. a surface as per the preview), leaving you to handle the trimming manually. It's not just solid filleting, it is hybrid, you can use the command on "open" faces along sewn edges.

There are things you can do right now. First of all, study that preview carefully! Zoom in tight and you will probably be able to see where the fillet is going to fail and can adjust the model or the fillet (e.g., use a variable fillet or reduce/increase the fillet rad). Watch out for regions where the fillet overlaps itself, or swaps from concave to convex - you might have a feature that presents an unintentional undercut for example. If there are no apparent glitches in the preview, then a computational tolerance change may tip the balance in your favour. Typically, your default tol might be 0.01mm for the model. If you change this to 0.1mm (Edit/Preferences), then try the fillet, it is likely to solve. Remember to change the tol back again for the rest of your modelling. If you are working on a tiny model, then the default tol wants to be 0.001mm. At any time, you can try changing the tolerance of all the features (entire model), via Edit/History Operation/Modify Tolerance.

Rank: 1

Kevin

Newbie

posts: 0

Registered: 2004-4-26

Message 4 of 8

15-08-2005 05:54 . am   |   View his/her posts only
I noticed on the planned changes in v12 (I think) that the minimum tolerance will now be 0.0001mm. How will this impact on filleting and export to other systems (if at all)? Also one thing I would like to see is a curvature continuous filleting option (as can be found in SolidWorks for example). The designer arcs in 2D are good - how about a 3D filleting equivalent?

Rank: 1

Paul

Newbie

posts: 0

Registered: 2011-7-12

Message 5 of 8

15-08-2005 08:39 . am   |   View his/her posts only
I do look forward to version 12, and hope that the new code will be able to handle some real world geometry. I find that bad filleting situations can add days to a project.

I'll have to try some of the tweaks that you suggested Chris, I am quite mindful that you can really mess things up if you get the tolerance too loose, but if it works on one continious fillet, it would be much better than stitched in geometry.

3d Monkey; I'm not sure what you are referring to as far as curvature continuous filleting and how it works.

Perhaps if it was easier to tweak the surfaces which cause the filleting failure it would be a big help, it might also help if the highlited preview showed the failure area in red.

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 6 of 8

15-08-2005 11:46 . am   |   View his/her posts only


Well, in addition to arc fillets, VX offers conic and elliptical fillets - give them a try!




The advantage of a very fine computational tolerance is realised on "micro" parts, which will have a correspondingly small manufacturing tolerance. However, as you may have observed in other CAD systems, very fine computational tolerance often presents problems - this is because your 32bit system can't handle the decimal places. Will be great on 64bit though.......

Rank: 1

ChrisWard2k2

Newbie

posts: 2

Registered: 2011-11-22

Message 7 of 8

15-08-2005 11:51 . am   |   View his/her posts only
One trick in exposing the problem areas is to increment up the size of the fillet radius until the first problem area reveals itself.

Rank: 1

Kevin

Newbie

posts: 0

Registered: 2004-4-26

Message 8 of 8

16-08-2005 06:09 . am   |   View his/her posts only
Quote

Originally posted by: ChrisW
Well, in addition to arc fillets, VX offers conic and elliptical fillets - give them a try!



I have Chris, but they never seem to work that well. I get the same problems on applications that use the ACIS kernel using their Elliptical fillets. The Parasolid based ones seem to be more robust.



The advantage of a very fine computational tolerance is realised on "micro" parts, which will have a correspondingly small manufacturing tolerance. However, as you may have observed in other CAD systems, very fine computational tolerance often presents problems - this is because your 32bit system can't handle the decimal places. Will be great on 64bit though.......




The other problem is on solid kernels like Parasolid and ACIS is that the tolerance is something like 0.0000001. This makes for very robust watertight models BUT also creates all sorts of issues with some shelling situations, and stitching surfaces into solids. As soon as the dreaded "tolerant edge" is used in these apps the stability and robustness of the file seems to deteriorate quickly.
See also
X