Hello
The network surface and the 1-rail sweep will create an edge between the two created by the network surface and the 1-rail sweep.
Because of this, the shape will not be straight.
What could be causing this and what is the solution?
best
1 Like
If you could share your file it would be much easier to answer
Include the input geometry as well as the surfaces.
As the others mentioned above, the best way to receive free help and proper advice is to upload your 3dm file.
Here is the simplest solution, though it depends on the shape you want to achieve:
- Create two extruded vertical surfaces from the middle profile (one upwards, the other downwards).
- Then match the ring shapes to their corresponding extrusion using tangency continuity.
- Delete the two extrusions.
Tip: Avoid using “Network surface”. This is the worst NURBS surfacing tool in Rhino along with “Patch”. If it was mine project, I would prefer to use “Sweep 2 rails”, “Sweep 1 rail”, “Loft” or “Edge surface”. Depending on the case.
1 Like
I created the green surfaces with network srf and the red surfaces with Sweep1.
I’m a perfectionist so I’m a little concerned about the joint edges.
I just i share the 3dm file.
sample1.3dm (9.0 MB)
1 Like
Hi @davidcockey
Edge was handled more nicely than I expected.
Thank you so much.
I’d prefer to scale the control point after MatchSrf where it’s a little snaky, or cut the section and do Swepp1 again, right?
Here is another approach with single-span surfaces. I created entirely new input curves with degree 5 and degree 3. Assuming that the lower half of the ring must have the same thickness along its entire length, I made it with the “Revolve” command to keep the control points at their minimum.
It’s not fully optimized, though. There is about 0,18 degrees deviation between the two upper surfaces.
To achieve better tangency continuity (0,05 degrees), you can apply “Match surface” (with the “Refine match” option) to the adjacent surface edges on the upper half of the ring. But that will also add more control points, making it more difficult to modify the shape of the current simple surfaces.
An alternative approach is to apply “Match surface” to the upper surface (not to be confused with the top cap) and then convert it into two Bezier surfaces, in order to preserve the single-span structure of the ring. If you then follow that with another run of “Match surface” to each of the two halves, you can achieve 0,01 degrees tangency.
In my opinion, a gradually widened lower half of the ring helps to achieve a more natural transition towards the top surface. You can clearly see that from the light lines.
This one has 0,13 degrees deviation between the upper surfaces.
2 Likes
Is this method correct for the process of adjusting from a single surface using revolve?
If you don’t mind, could you share the procedure on your Youtube channel?
See also some alternative modeling methods in this thread:
Thank you
Although the use of revolve is a traditional approach, we are considering a variety of approaches.
I am a trial-and-error person because I sometimes change the shape and length of the Right and Bottom quarter points of the Ring in a resourceful manner.
The modeling process is quite slow and tedious, because Rhino’s “Match surface” tool does unwanted extreme chaotic arrangement of the control points of single-span degree 5 surfaces while I point edit them with active History. Alias, on the other hand, does that in a more predictable way, thus the process is multiple times easier there. For some reason, Rhino has a really poor implementation of surfacing with degree 5 surfaces. Maybe because more effort was put into degree 3 surfaces (many tools will convert the surfaces to degree 3 by default). Also, there is no “Explicit control” in Rhino (Alias has it), which further makes it harder to do a proper NURBS surfacing with single-span Degree 5 surfaces.
You can see the same method of control point modeling in one of my old videos, where I showed the process of creating single-span surfaces in a simple Y-branch shape:
Couple of years ago I already posted two alternative methods to create rings with other tools.
One of these uses the “Loft” tool and its “Loose” option:
The second method uses basic loft surfaces, which are then modified with the “MoveUVN” tool:
1 Like
@Rhino_Bulgaria
Signet rings are an ordinary form of jewelry, but the jewelry industry has long advocated various methods of making them, and they are surprisingly difficult to make.
It is easy to shape, but it takes time to accurately reproduce the cross section, length, and shape desired by the creator.
After all, it is impossible to completely remove the surface edges created by network surfaces and Sweep1…
Yeah, even the signet ring tool in RhinoGold (for example) is pretty bad. I usually end up using SubD now to make signet rings.
RhinoGold is too old…
MatrxiGold should still be able to make something better.
No doubt…but then I’d have to dish out at least $3500 for it (that’s the such-a-deal price for trading in my RhinoGold license too). If my dayjob will pay for it, then I’ll certainly use it.
The alternative is to learn how to create your own rings. 
That saves you huge money.