I’m a university student from South Korea, and I’ve hit a major roadblock in my Rhino modeling process. After countless attempts, I still cannot achieve the desired continuity and smoothness in my surface using Blend Surface (G3 continuity). I’ve spent hours experimenting with different workflows, but the results are not working as intended—Zebra analysis shows broken continuity, and attempts like using Patch create unattractive surfaces.
Here’s my current workflow:
I started with three polysurfaces and performed a boolean operation to combine them.
Then, I used Pipe with varying diameters to cut sections of the geometry.
For the top part, which was originally sharply cut, I used Blend Curve to connect the ends with G3 continuity and then trimmed the surfaces.
Finally, I attempted to connect the trimmed surfaces using Blend Surface (G3 continuity), but the result was not satisfactory. Zebra stripes are broken, and the surface is far from smooth.
My Issues:
I cannot get a clean and smooth result for the blended surface.
The Patch command creates very uneven and unattractive geometry.
I’m unsure if my workflow is flawed or if there’s a better approach to this kind of modeling.
Request:
If anyone could take a look at my Rhino file (attached) and suggest:
A better workflow for achieving smooth G3 continuity in this situation.
Advice on how to handle these challenging transitions or recommend alternative commands/settings.
Any tips for optimizing my process overall.
I would be incredibly grateful for your help, as I’ve been stuck on this issue for quite some time. Thank you for reading, and I deeply appreciate any guidance you can offer.
a quick start could be to go back to the state of boolean, then use BlendEdge, select all edges you want to fillet at the same time, maybe that gives you a good start since it will try to create a smooth fillet across all edges, alternatively experiment with FilletSrf these will create g2 max though, specifically the filletsrf, the blend edge maybe not even but give it a try.
Why are you looking for G3 continuity?
You better start over with a good tangency workflow and then raise the level.
Sincerely for that kind of shapes even G2 look too much.
Extrusion has zero curvature and revolve has it constant = none of them are ideal for matching at higher grade.
As usual start by the amazing thirtysixverts YouTube channel
Understand that higher continuity isn’t a command but a way of life and you will on the right way… Young Padawan.
I used to think that a higher G level always meant better results. Through this issue, I’ve learned that this isn’t always the case. Thank you for helping me realize this!
@jim i would be curious how you use with FilletSrf here, i never use it because it throws me off each time i attempt it, i never seem to get the order right. if you have time for a short tutorial
The existence of Gs beyond 2 is the thesis that when milling a mold for a car body panel using the highest-speed equipment on the planet, an abrupt change in the rate of change of the rate of change of the curvature of the toolpath(which of course to most of us is just a bunch of straight lines and maybe arcs) might cause a flaw or break the tool or something.
What you’re really looking for is a “nice” surface transition, which has not really a lot to do with the level of continuity that technically exists right at a joined edge. The parameter space of options for G2 and beyond is so infinite that trying to use some tool to achieve it(BlendSrf just puts points in a row, which is crude) is really limiting.
use other tools like xnurbs (as Plugin or via plaxticity)
use parasolids (kernel) fill function from software that is based on it (siemens nx, plasticity, onshape, solidworks, …)
do it with subD in rhino
poor rhino backup: use _netWorkSrf or try _patch
check out rhino WIP / Rhino 9 _npatch
vanilla rhino
my vanilla rhino ( rhino only ) solution would be to find a surface / patch layout, where each surface stepwise reduces the complexity of the transition.
And all surfaces are (or at least try to be) either convex or concave in one direction (U,V).
untrimmed edges are better then trimmed edges
avoid rational surface to match to (use deformable option in revolve)
_matchSrf is painful but the only tool that will make it happen, use the pull option for edges that trim other surfaces later (like the edge of the green surfaces, that trims the blue one)
if you want to get really nerdy: you can combine _matchSrf with history. (refine = no) and massage the CV’s of the surfaces.
First I modified the main handle(?) surface to remove a couple weird glitches at the quadrants. Was the section made by modifying a circle? Don’t. Then I really cheated by point-editing it a bit to make it curve a bit towards the round shape instead of being totally straight.
Anyway after that I did the largest fillet first between the handle and round parts, then used that result as a guide to draw curves to trim them back for a blendsrf. Then I could do the same thing for the remaining edge but left it at the FilletSrf result. Point being most of these “how do I fill in this corner that’s literally impossible to make continuous to all the neighboring edges,” problems go away if you look at them as fillets/blends rolling over top of other fillets/blends.
Thanks for taking time giving your solution.
I’m kindly asking if you can add a couple of screenshots of your solution.
I’m mainly looking through my tablet and here I can’t view the file.
From my oint of you it’s more appealing if the thread has more visuals.
I won’t to be critic, I’m just so curious to see the different approaches that I can’t wait until Monday to check your file.
In my opinion the solution for this depends on the design intent. For me the two pipes are related so I’d like to see that back in the flow. Therefore I’d approach it in this direction:
I would rebuild everything to single-span surfaces with degree 5. The blends will be much cleaner and easy to edit then. Most likely they would stay relatively simple even when using “Match surface” with the “OnSurface” option.
From a design point of view this indeed seems to be the better or even the “correct” approach.
Sometimes (and when I say “sometimes” I mean “always”) a little bit of conceptual thinking/planning beforehand can avoid problems downstream.
I disagree. check what the old masters did before cad:
I think exactly this nice mixture between a geometrical approach and a virtuous and free interpretation on the details give does designs their strong character.
With the round looking object did you split with a pipe then blend srf? I’m not sure of the angle but upper right side could use some slashes and perp snap to straighten those lines out I always get nice smooth transitions that way
While I love many of the classic 70ies products in that thread, I wouldn’t consider them the last word on surfacing.
Design intend is of course a matter subjective opinion, that’s why I put “correct” in inverted commas.
For my design sensibility @Gijs 's approach seems to create the more logic surface flow and also seems to create simpler and more pleasent fillet-/blend-surfaces.
