Some other failled fillets ... (or booleans)

Next
 From:  PaQ
1698.1 
Hello,

Just a little report on a filleting failure. I know it will be fixed later, but I still take the time to report it, might be usefully ... or not :)



So there are 6 cloned spheres, booleans added to the bottle ... 4 fillets successfully created with 0.3 cm radius ... 2 of them don't what to be created higher than 0.05 cm.
I suspect it's a boolans problem in fact, but I can't isolate it.

Btw I have created this nice object to test the shelling function (in viacad), actually moi gives a strange result here (shell at 0.3 cm) (the exploded surface is on a sphere where
fillet was succesfully created).



PaQ

EDITED: 4 Dec 2015 by PAQ

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
1698.2 In reply to 1698.1 
Hi PaQ, yup that is probably currently the most annoying fillet bug, the problem is that it won't cross an edge that starts a small distance away from the fillet edge.



See how the seam edge of the sphere comes very close to the edge you want to fillet, but does not quite touch it? You can do a fillet radius up to the point where the fillet starts to touch that edge and then it will fail.

This is definitely the #1 biggest problem currently in fillets, when this one is fixed up it should be a pretty big boost.

If you rotate that sphere around so that the seam edge is something like this:



Then the filleting should actually work there.

The shelling looks like an example of the #1 shelling bug which is a problem sometimes when offsetting surfaces with singularities in them - that's with a surface that has "poles" in it where the surface is collapsed down to a single point. Once that one is fixed up for shelling that should be a good step forward in that area as well.

- Michael

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  PaQ
1698.3 
Thanks a lot for the explanation, and for the tip ;)
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
1698.4 In reply to 1698.3 
Hi PaQ - also one other workaround method is to use surface/surface filleting instead of edge-based filleting, the surface/surface fillet will be constructed in this type of situation.

To use surface/surface filleting, the objects involved must be individual single surfaces instead of joined-together objects. To get that you could select the 2 surfaces (sphere and the main body piece), then push Edit/Separate once to break that 2-surface chunk off from the main object, then push it again to break it out into 2 completely individual surfaces. Then you must deselect those objects and select each surface with a click - surface/surface filleting needs a click selection point on it to know which fillet to build in cases like 2 planes crossing over each other like an X.

That will build the fillet, then you can join those pieces back together.

Surface/surface filleting can succeed in places where the edge-based filleting will fail, the edge based calculations are more complex (but also more convenient because they know how to build corner pieces where different fillets meet).

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
 From:  PaQ
1698.5 In reply to 1698.4 
Excellent, I didn't know fillet can work in that way, with individual surfaces :)

Thanks a lot Michael !
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged
 

Reply to All Reply to All