Hi Lang, I would actually recommend that your easiest course of action for filleting problems is to get ViaCAD to help out in some of these specific instances - ViaCAD does have an overall more robust filleting mechanism in it and so it can be easier to use it to get the job done more easily in situations like this.
Also the ViaCAD 2D/3D edition only costs $100 so it's pretty inexpensive:
http://www.punchcad.com/p-9-viacad-2d3d-v8.aspx
Filleting is an area of quite complex calculations and so it's just not a bad idea to have another tool in the toolbox for handling difficult situations, particularly when it's inexpensive like ViaCAD. Or also check out Autodesk 123D which is actually free.
> Yeah if you want to have a go at it for me that would be
> great to see what you can come up with too.
So first I reconstructed your tube sweep so that it had the seam on the outer part like I described earlier.
Then I did the same fillet that you showed earlier that gives the funny looking result - that result is a set of fillet surfaces that are sitting inside the dome. Hide the dome to see them more clearly:
After doing some experiments with using those surfaces to try and cut the main object I was able to determine that it was a small area at the top of the fillet that was malformed and not properly hugging the outer dome surface that was preventing things from working properly. It's this small piece up here:
So I deleted that little piece, it's basically junk. Now the problem is to get some closed boundaries to make it possible to trim the dome surfaces. I took a look at the fillet surfaces to see if they might extend a bit further than their current ends, and they did, so I untrimmed them (use Edit > Separate to break things into individual surfaces, then select all boundary edges of a surface and hit delete to untrim it and reveal the entire underlying surface), and that gives this result:
So there's only a pretty tiny missing area now, I'm not really sure why the fillet engine decided to stop the fillets right there, it looks like it was doing a good job right up until that spot, maybe it has something in it where it gives up if the fillet cross sections get too tiny or something. Anyway I decided that probably putting in some curve blends would be good enough to make a close enough boundary for trimming.
To do that I duplicated these edges to make stand-alone curves out of them:
Then used Construct > Blend to put in these blend curves here:
Then I used Edit > Trim on the dome, using the edges of the fillets and also those little blend curve pieces as the trimming objects, and doing that I was able to cut away the material of the dome to make room for the fillet to be placed in there. That leaves this little hole (which is in the attached 3DM file still), which can be filled in using Network or Sweep:
So anyway, it's a sequence of pretty advanced surface editing and trimming operations to deal with this kind of stuff, it's far easier to just export an SAT file to ViaCAD and give it a try over there and then import it back into MoI. That's usually something that you wouldn't need to do very often, just in cases where MoI's filleter gets confused, but when you do need to do a fillet tuneup that's actually a lot faster than mastering all these different low level surface modeling and trimming tools that you would otherwise need to use.
- Michael
