Hi, I’m trying to do this kind of twist / folding just like in this ring, but neither do twist, sweep2 do the trick, anyone can help me??
Hello - please post a file with what you have so far.
-Pascal
What a painful process to fumble through rigid curvilinear CAD tools to chase after an organic-looking shape. If only Rhino 7.0 offered a Sub-Division toolset as T-Splines once did…
We could then start by extruding & pivot-rotating a flattened shank…
…then begin gumball rotating the successive ends until reaching the apex
(Disclosure, Gumball’s auto-reorienting sucks. Users currently have to brain-boost into clever-monkey mode to compensate so that extrusions properly orbit the exact center. Leave my gumball where I last set it! While we’re at it, do we need to grovel for Gumball extrusion dots for the rotation handles? The Move handles have them. Extruding while rotating was a main T-Splines workflow staple.)
At this shank-half’s completion, world-center the gumball, Alt-rotate a copy to face 180º of the original…
Stitch the mating top vertices. Stitch the mating bottom vertices…
and the tab-toggle gives us a proof-of-concept.
But the peanut gallery is screaming that it’s not the same style… and that’s mainly because the twist in question has an apex cross-section located laterally to the finger. Study the travel direction, adjust accordingly, deploy the “Insert SubD Edge Loop” and “Slide SubD Edges” tools judiciously…
THIS is where a properly working Z-axis Radial Symmetry would’ve been sweet, but I’ve already used up all my genie wishes on unicorns & cupcakes, so we’ll repeat step 3’s copy/rotate.
And stitch vertices…
…to end up here.
5 Likes
Hi @violetadegenciana ,
Sweep2 works fine for me. The curves setup is pretty simple, actually it can be made from a single rail curve which I’ve rotated 180° with history. With the cross sections you can control the appearance of the twist. In this case I first created a simple surface and then used FlowAlongSurface for the profile.
Attached is the Rhino File with construction geometry on hidden layers.
RingSweep2_jM.3dm (1.4 MB)
1 Like
I see there are lots of knots on the rail curves leading to a complex base surface.
It can be naturally simplified, but I am curious, may I ask you which technique has been used to get those rail curves ?
Hi Rodolfo,
I started with a Circle Deformable with 24 Points and made a rotated copy with history. Then I set the initial picture as wallpaper and roughly adjusted the view. I switched the _DragMode to _View and moved the points in a different perspective viewport to fit to the wallpaper in the other viewport.
24 points is really not too much for such a shape and I think I’ve edited 6-8 points. Especially when using _Fair frequently it is better to have more points to get a nice flow. In general I prefer uniform spacing on such organic shapes.
Jess
Thank you for your answer.
I am not used to work with so many points but I understand your logic.
I usually work with less points and add extra data when needed.
Regards
Rodolfo Santos
Well, 8 Control Points is certainly not enough here!
The shape is not planar, please post this curve and/or the resulting ring.
Maybe 24 points was generous and sure it could be made from less points. But as said: when using _Fair it is better to have “enough” points. Another problem you’ll get with uv-texture distortions when using varying spacing…
I like 24 points for a degree 5 deformable circle, if the circle part matters - it is very close to a true circle and has a good constant curvature graph. Degree 3 curvature will wobble a bit more at the same point count.
-Pascal
You are right, I was just mentioning that I usually work with less points, and tune up progressively.
I find interesting to see how other are approaching things and which strategies they use.
Your strategy is perfecly valid and as Pascal mentioned there are advantages in working with more points at a higher degree.
The curve I created is for sure less accurate but is probably more easy to reshape which in jewellery is a recurrent need.
RingSweep2_jM.3dm (55.4 KB)
Yeah, and we should not show off as rivet counters 
The Sweep2 command will vanish the curve simplicity anyways and bring back complexity…
I’d appreciate to see your solution to model that ring!
Regards, Jess
Thank you very much!
Thank you very much Jess!
You’re welcome @violetadegenciana
Now the real question is: What is the purpose of the model?
Do you want to make the ring manually by “folding” metal or eventually make a wax model and cast it?
I would not choose another modeliing strategy ( Sweep 2 ) to approximate the input surface, only try to avoid unneeded complexity for flexibility of use and/or reshaping needs.
Regards
Rodolfo Santos
Set gumball to object alignment, should be much better that way.
A small complementary question
Is there a way to ask a relocated Gumball to stick to the previous location set by the user instead of being reset ?
This is something I see here when working on sub objects VS parent objects.
Rhino for Mac Version 7 (7.6.21117.09002, 2021-04-27)
Regards
Rodolfo Santos
Loft would be an option which could preserve the input curve structure. I made one based on a degree 5 curve with 24 points. I was not able to go below 24 and having control over the shape and the ruling lines. Can you really do it with less points?
Another advantage of loft is to get a developable shape which you can unroll to get the raw sheet metal contour. Just in case it should be made that way…
1 Like
Using Loft it might be possible to use less points and approximate an acceptable more flexible surface.
Then I would use MatchSrf to finalise on top when reshaped.
RingLoft.3dm (109.6 KB)
Regards
Rodolfo Santos
Hm, that does not look any cleaner or easier. Also you messed up the curvature of the surface.
Did you notice that I made this ring from a single curve which I’ve just rotated?
BTW: The most simple surface to make this ring as a solid would be degree 5x5 with 6x7 control points → 42 Control points in total… but I would not recommend that technique for beginners 
1 Like