Create a round cap with continuous surface geometry

Suppose I wanted to cap the following geometry with a curved surface. Using the Network surface command leaves me with completely uneven geometry which doesn’t match either guide curves. Using the sweep command yields similar results. What is the secret to capping such a surface? Is there something I’m missing here? Is there an easier way to model this surface more effectively?

I’ve attached the test model for the inquiry and am using Rhino 5.

Test Model.3dm (1.9 MB)

Hello - see if the attached file has something you like.
Two simple edgeSrfs from matching curves, then some MatchSrf all around to clean up - mostly just for tangency. The edges at the main ‘back’ surface are matched for curvature, the centerline edge for tangency with ‘Average surfaces’ set.
It looks to me like you could make all the surfaces equally simple -

Test Model_Maybe.3dm (1.4 MB)


You could also make this shape by using Sweep2, and adding slashes near the interior corner where the shape gets distorted

Wow that’s really clean. So you used special MatchSrf commands afterwards? What is the edgeSrfs command? Is it like Sweep or Network?


What do you mean by “slashes”?

Hello - I made new (red) curves - looking at the centerline curve (with the curvature graph) , it has a ton of points but looks like it is a straight line in the middle, sloped, and two s-shaped blends out to more level lines, so I made it that way - a line and two simple curves:

I made new cross section curves - these look like they are tangent across the centerline and tangent to the vertical sides - not curvature continuous, again, looking at the existing curve’s graphs for confirmation.


I used slightly edited copies of the three centerline curves to finish off three rectangles:

EdgeSrf makes all three surfaces.
Then ExtendSrf the ‘front’ one(s) and trim like so:


Then create another edgeSrf in the same way - four clean, simple and matching-on-opposite-sides boundary curves. Note they match here even though the point count differs because they are all single span curves.

Then MatchSrf on all edges - for Curvature back to the trimmed edges, for tangency to the lower surface and for tangency with Average across the centerline. I found I needed to InsertKnot > Automatic a couple of times to get a good final match to the vertical wall,


Thank you so much Pascal. I’ve never heard of the edgeSrf command before. I’ll see how it works and try and apply it more often.

Hello - EdgeSrf is useful mostly when you have simple matching curves as in my example and you want to make a four-sided, super clean and simple surface. The surfaces inherit the curve structure, so with complex and especially non-matching curves, it can make a mess, as the curve structures must be combined and this is always more complex than either input.


See the Add Slash section on here