Quick test I made based on difficulties in the past with filleting in rhino
Steps:
Create a 20x20x20 mm box,
Remove Bottom Face,
Duplicate it,
FilletEdge of Radius 10 works, trim, join and everything remains together
FilletSrf of Radius 10 Fails, doesn’t trim and doesn’t join
Cap Bottom Face
Repeat Fillets on Bottom face
Is this expected behavior on the FilletsSrf?
Yes, FilletSrf only trims the surface that are filleted. Pls note that this enhancement to FilletSrf is not meant as a replacement for FilletEdge. FilletSrf and FilletEdge are different tools for different use cases.
you’d fillet edge 1st, then you can fillet srf to get this.
(to be fair fillet edge would do that as well in this example if you wanted a rolling ball fillet)
I guess my example ended up being too simple to represent what I wished to say.
Ok, I get that the side walls would not be trimmed, but why does the fillet fails to form the top round face, forcing to use 9.9 as radius to be able to make it similar to the fillet edge command?
Note, the red line showing up is the surface that was generated but the original surfaces were not trimmed to form the filled.
That is not at all true. Just make the four 10mm fillets and then select them and use them to trim the box. This is nothing new, you could do that with FilletSrf in Rhino1.
For modeling work that consists of making boxes and adding fillets the new development in FilletSrf isn’t going to help much.
The box was just an example of some situations where the fillet fails to auto trim and manually trimming doesn’t work, usually when one of the surfaces is being entirely consumed.
And yes, I know you can trim and join afterwards, however, with complex shapes that becomes very time demanding very quickly and its counter productive to have a tool that does only half of the job, that is part of the frustration with filleting in rhino, having a tool, trying to use it, it not working, and having to just do it manually anyway.
The title of this thread says it all, the purpose of the new functionality is making fillets across tangent faces. You present a model that has no tangent faces and ask why it does not work any better than it always has.
Again your statement is completely wrong.
If you present this tool with a model that has two sets of tangent faces that you want to connect with a long string of fillets it can save you an enormous amount of time. The trimming and joining part is not where the time savings is to be found. Its in making the dozens or even hundreds of fillets each one, one at a time, telling Rhino which two surfaces need to be filleted next. That is not a trivial exercise in itself, but when you’ve worked for half an hour at it and then you discover the fillet size you chose at the beginning of the exercise is not suited to the topology and you have to delete (or undo) it all and start over, you may appreciate the usefulness of a feature that will do it in seconds (or less).
Fair point.
I was looking into the original post geometry with the wrong perspective and was just venting some frustation of having to work the past couple of days filleting mechanical parts.
The improvement is indeed marvelous.
FilletSrf with Continue across faces does not extend the fillet and trim for this shape.
Work In Progress (9.0.25161.12305, 2025-06-10)
A previously reported different failure with an object with the same shape but no back surface was fixed in the new WIP of 2025-06-10. Surface Filleting across Tangent Faces - #39 by davidcockey
However FilletSrf fails with the same shape but with a back surface.
FilletSrfFail3.3dm (84.6 KB)
Yes, that type of trim is currently not supported. It needs the extended fillet to extend through the back surface, get trimmed by the back surface, then trim off the bits above the fillet, includeing parts of the back surface.
We’re not there yet, but I have created a new ticket for this so I won’t forget.
It is probably easier to get to this, where the back face is not trimmed, but the faces that straddle the fillet are.
That’s not the way trimming works in Rhino. In order to get the trim command to work after the fillets are made you have to remove the back surface. So why add that back surface in the first place? It just makes it designed to fail.
People who use this tool would benefit from understanding the tool is designed to make a string a fillets that connect two sets of tangent surfaces. Those two sets do not have to be joined together or even touch each other. People who set things up as if they were going to use filletedge are going to miss out on what this tool can do
Throwing non-tangent surfaces into the mix is just going to throw a wrench in the works.
This file shows how you should be thinking of this to get the most from the “continue across faces” option in Filletsrf. After you make the first set of fillets you again have another two sets of tangent surfaces that can be connected by fillets.
designed_to_succeed.3dm (248.1 KB)
If you make the second set of fillets the same size as the first set of fillets with the current implementation of “continue across faces” it fails because it can’t handle same size fillets properly. Thats is a bug.
Its not supported in Rhino’s Trim command also. If I recall correctly, years ago the developers tried to allow Trim command to do this and it caused problems in some cases.
This would be good. That result now occurs if the back surface is not joined. ExtractSrf the otherwise uninvolved back surface and the FilletSrf works as expected. That difference is what I consider to be the failure.
