BlendEdge Behavior

In constructing the outer surface of a helmet, I’m attempting to use BlendEdge to combine the 3 main surfaces making up the shell with at least G1 continuity. The result is an erratic, failed surface which I haven’t been able to remedy, even after rebuilding the geometry. Is this an unsuitable function for this sort of geometry (is the ‘curvature-continuous’ dependency making this fail in some way)?

I’ve already tried combining these three surfaces by offsetting them, and then using either BlendSrf or 2-rail sweep, and the results aren’t quite clean enough to justify using as a final product. Any alternate suggestions on how to smoothly combine these surfaces are appreciated! Thanks.!

FOAM BLANK RCE 016.3dm (15.3 MB)

Hello - it looks like the angle between the surfaces flips from concave to convex is that the intent? - that is hard to get right, I guess. What are you shooting for here, tangency or better between the surfaces or some kind of transition surface that makes up for the discrepancy?

FOAM BLANK RCE 016_Maybe.3dm (166.3 KB)

-Pascal

Pascal,
I see what you mean – that isn’t the intent, but rather a result of reverse engineering these surfaces based upon an existing STL mesh. I’m shooting for tangency or better, but I’m not opposed to using a transition surface, if that’s what it takes. I tried using OffsetCrvOnSurface (distance=5) on each edge of each surface in order to trim them away from each other, but when I used BlendSrf to transition the surfaces back together, I could not manage to produce a seamless transition (even when using G2 or higher continuity). Thanks.

OK, I added a file where I did some of each - matched up the side surface with the other two for tangency, then added a transition surface across the harder edge at the back.

-Pascal

That’s very close to what I’m looking for. Did you use an EdgeBlend for the transition surface at the back? I think I should be able to take it from here. Thanks for your help Pascal!

Hello - I used MatchSrf, Tangency at both ends, Average set and match by closest points set for matching up side-to-top and side-to-back surfaces. That cleans up those edges some. Then:

ExtractIsoCurve from the top and back surfaces.

At this point you could use these to trim the surfaces and then BlendSrf between. What I did was Loft the isocurves, Rebuild option in Loft and then MatchSrf to the trimmed edges, for curvature I think.

-Pascal

Pascal,
I have one final question regarding smoothing the surface of a helmet. I’ve taken your previous advice to successfully blend surface edges, but the final patch is proving difficult. I’ve tried a variety of surfacing methods (BlendSrf, Sweeps, Loft+MatchSrf), with many combinations of parameters, but the result always has observable irregularities under an environment map. Are there any shortcomings in the way this is modeled which prevent Rhino from creating a cleanly blended set of surfaces? Thank you for your time.
FOAM BLANK 052.3dm (15.4 MB)

It looks pretty good if I sweep 2 rails. Left, right, top edge, bottom edge
with the default position.

Hello - given what you have, I might just move the back blend away from the corner at the top a little do that the top of your gap is a little wider and takes up more of that corner curve. I made an EdgeSrf and then some matchSrf to clean up. I also retrimmed the top surface corner with the top edge of my new surface rather than try to force the serface to the previous edge.

All that said, I can’t help but think that there must be somewhat cleaner, simpler surfaces that would do the job - I see some accents and flats and Zebra wobbles in the surfaces that I’m not sure are part of the design intent…

FOAM BLANK 052_with filler.3dm (201.4 KB)

@fluxguys - are you working off of a mesh? If so, can I get that?

-Pascal

1 Like

Thank you Pascal. All of your points are well taken, and I’m working off of your advice right now (aside from your use of EdgeSrf, which creates a messy polysurf for me). I’m basing this loosely on a mesh, but the shape of the surface I’ve designed here is what’s desired, so the mesh wouldn’t be helpful. Some of the ‘accents’ are intended – I’m just trying to avoid any abrupt, visible surface imperfections, as this may actually be produced at some point.

Hello - OK - one thing is, I neglected to mention, I made the EdgeSrf from rebuilt edge curves, degree 5, 6 points. But, if you are designing this, then I really do think you can simplify the main surfaces - that will allow much more adjustability, so to speak, as well as cleaner curvature. If I have time I’ll see if I can make an example. One of the usual surfacing-suspects on the forum may also jump in with prime advice.

@fluxguys - I know this does not march your design, but just by way of indicating a possible simple-surface layout - FOAM BLANK 052_PG.3dm (121.9 KB)

-Pascal

Aha, ok that’s very helpful. I noticed in one of your past years’ comments you emphasized that the degree of each of the surfs being blended should be consistent, so I’ll focus on getting that right, as well as reducing the UV count on the main surfs. Will let you know if I get this to a ‘finished’ state. Really appreciate your insight on this.

@pascal - That does indeed look much simpler and cleaner. I’ve reduced the UV on the main surfs to about the minimum I need.
Here’s my latest – I matched the UV count on the surfaces to be blended in each direction, re-matched edges on those surfaces, used EdgeSrf to create the final patch, and matched edges on all 4 sides of that patch (unsure if the ‘multiple matches’ option helps). Although this is pretty clean, I’m playing whack-a-mole with MatchSrf, and can’t quite get rid of all of the blemishes on the final patch.
FOAM BLANK 059.3dm (13.0 MB)

OK - looks good - I still think you can simplify - a had a little more of a play with my version… but it looks like you’ve got it handled. Remember - fewer points means fewer opportunities for the surface to wobble. Ancient proverb.

FOAM BLANK 052_PG.3dm (257.2 KB)

-Pascal

1 Like