Surfaces are not connected well

hello, im still at the beginning of learing jewelry modeling. I made this earring, visualy it came as I needed. but I think I used a bad technique, because surfaces are not connected well. at first I drawn curves, then I duplicated them and made surface by using loft and patch. and I think it was a wrong way. Maybe any of you have suggestions? EARRING.3dm (431.0 KB)

The use of Patch has a limitation in that it may not Join into the other surfaces if it doesn’t have enough points. You may already know, but you can use ShowEdges to display the naked edges or gaps. If Join doesn’t fix these, you need to remodel those surfaces. In your model I would use ExtractSrf and delete the patches. Then use the EdgeSrf command with the resulting four sided boundaries. You could alternately use NetworkSrf with positional continuity or Sweep2. In this way you can Join the result into a closed solid.

Hey @BrianJ,

With regards to tangency of Patch, does the rule also follow that more UV points specified=increased likelihood of tangency? I have this file below where it’s still getting only positional towards the corners with about 25X25 points.

PatchExampleTangency.3dm (379.9 KB)

Any thoughts on a different patch layout would be interesting too so as to avoid patching!

You can use the JoinEdges command, but you risk to be stuck at corners.In any case it’s better to extract curves from the sides of the surface you like and then reuse as rail to rebuild the surfaces or use the sides as rails directly.

Hi all - don’t use this command unless you know what it does… See Help.


Hi Jonathan,

The surfaces surrounding the patch also are G0 to their adjacent surfaces which will add to the difficulty here. Remodeling those can help some in getting the patch corners smoother. In general, I only add as many points as needed for the patch to Join without naked edges.

Hello - given what you’re starting with, I’d try to arrange the surfaces something like in the attached file - but that area is a little ambiguous - it looks to me like there are transition surfaces that are driving the shape there, not a primary, and it gets a little messy where they converge.

PatchExampleTangency_PG.3dm (216.6 KB)


Thanks for this Pascal - I would have been tentative forming the surfaces in that way for fear of the pull command inadvertently creating any nasty intersection or really thin surfaces. But what you’ve done seems more than clean enough with the BlendCrvs going into the SplitEdge. Like below, I ended up getting with the surface not going all the way.

I think some of the G0-ness was caused by not realising I had a used the refit option on the sweeps - so that I could have the Tangent option for the surface edge as rail.

What refitting options did you use? I can’t seem to replicate what you managed.

Hi Jonathan - the surface I had as two is probably just fine as one-

If things are nice and aligned you can split edges on adjacent surfaces and MatchSrf using ChainEdges

PatchExampleTangency_PG.3dm (222.5 KB)


Thanks Pascal.

Apologies OP!