HELP. Issue with fillet edge command on model


I’m having some issues with the fillet edge command on my model (jug), please see images attached. This is a new version of the jug where the issue has arose. An earlier iteration sees the fillet edge doing what it should (see image). Where in the new one it’s breaking up and not forming as should. I’ve tried a couple of things to get around it but to no solution. If someone could kindly advise that would be greatly appreciated!!



Screenshot 2021-10-05 at 12.01.33|562x500

dear @ld.power

_filletEdge is not able to handle the inner transiton.
this transistion / blend is made by several surfaces (Polysrf) i selected the most critical in the screenshot.
solution 1
you can still build single fillets using _filletSrf and then trim / join everything together.
the blue surfaces are _filletSrf R2.5

imagine fillet as a ball / sphere is rolling along the edge - if this ball is touching surfaces, that are not connected to the initial edge - the command fails.
you see this in the screenshot: the s-shaped surfaces - as part of the polysurface, does not belong to the edge…

solution 2

you can avoid those issues if you redesign the trimming and do a transition / _blendSrf or fillet or whatever that is a single surface … (instead of the s-shaped …)
then filletEdge should work

hope this helps . kind regards -tom

Hello - I’d greatly simplify that whole area, for example -


Hi Pascal,

Simplifying it sounds great. Could you please show how you have done that or what commands?


Hello - in this case I cleaned up the sharp corner in the edges there and made a smoothly curved trim on the spout surface and the blend in there - then made a loft surface to bridge across, having rebuilt the curves at the edges to match each other (degree 5, 6 points) Then MatchSrf.

But. That is only patching up one area - I would (if it were me, and it’s not) step back and rethink the structure a bit - I don’t want to mess with your design but just as an example, something like this:

might be cleaner.

(as an aiside, a spout that does this:

is pretty much guaranteed to send milk all over the table - the edge needs to be rolled over - I don’t know if that matters here but just in case :slightly_smiling_face:)


Pascal this is the modelling I need to learn! Looks far cleaner and you obviously know how to build not just what the commands do - very impressive. I don’t suppose you can make a video of what you did exactly? It’s hard to piece it together from what you wrote to actually modelling it. After getting this project out the way I’m certainly going to be looking at improving my surface work.

Also good tip on the milk :+1:t3:

not talking care about the milk tip … sorry, but i appreciate.

ok and if we discuss about modelling in general - a very simple and clean approach for the basic body would be the following workflow:
_untrimAll (the outer surface)
_changeDegree to 3-3 (not deformable)
_inserKnot (2 horizontal)
_inserKnot symmetrical=yes (2 knots, close to the spout)

move the control point s with gumball

kind regards -tom

1 Like

Hi Tom,

Thanks for your assistance! that surface is looking very nice, also like the use of the control points to change the form! I am having trouble with the initial command if you wouldn’t mind helping - Untrim All. Please see attached screen recording for reference.

just use the main surface - untrim all does not work on polySrf… you need a single surface to start above workflow:
_extractSrf (copy yes) - the main Surface “oval-truncated-body…”
… see above

Why Deformable=No in ChangeDegree? That puts multi-knots in the surface if the input surface is multi-span and the new degree is greater than the old degree. Multi-knots can become an annoyance or worse.

Deformable=Yes does not introduce multi-knots but the shape may change if the input surface is multi-span. When the surface will be changed the difference in input shape may not matter.

The Deformable option does not affect the results if the input is single span or if the new degree is less than than old degree.

in this specific case the input surface is based on a ellipse. (degree 2 weight, … special knot structure)
rebuild with deformable option will change this setting and will change the shape.
rebuild without deformable will result in the same shape and offers a CV-Structure suitable for this small modification.

ok ? or i am wrong ?


Makes sense. I overlooked the special cases such as ellipses.

Hi Tom,

Thank you so much for these tips, this is going to help massively with the development of the product!

MANY thanks