Hi,
I have one rail curve and one profile curve.
Rail:
degree = 3
control points = 4
Profile:
degree = 3
control points = 4
But after Sweep1, the surface becomes:
U degree = 3, CV count = 19
V degree = 3, CV count = 4
Why does the U direction become 19 CVs if the rail has only 4 control points?
Is this normal for Sweep1, or should I rebuild / fit the surface after sweeping?
Thanks.
From my basic understand this happens when the section doesnât match an knot of the rail.
The section should also be perpendicular to the rail
I tried creating the profile perpendicular to the rail and sweeping it, but it still creates many isocurves.
When I try it as a half profile, the result looks cleaner, but then I have to create two separate surfaces.
I still do not fully understand why this happens.
I also attached the CAD file.
soru.3dm (241.4 KB)
Have you tried cutting the rail in two where it touches the section?
I tried it 2-3 times, but Sweep1 still adds extra isocurves / CVs.
I also recorded a short video showing the steps:
https://drive.google.com/file/d/1seTK6an42J7uAZoO0QthYuSo2zPpqjnk/view?usp=sharing
I still cannot understand why it creates such a dense surface. 
what you re after was called simple sweep in older rhino versions.
but simple sweep was never nicely supported - for example with error-messages etcâŚ
and there was not much feedback in this community when i tried to push this featureâŚ
even worth - it became a âhiddenâ feature.
from an old help file:
Simple sweep
Simple sweep works in cases where the shape curves intersect the rail at rail edit points. This option generates simpler surfaces in cases when the curves are perfectly set up. The surface inherits the structure of the input curves.
the most basic setup that will work looks like this:
red (one or two rails)
green (shape)
rail and shape (at start/end) need to have G0 (position) - test with _gcon command.
if you want inner shapes they must be positioned at edit Points and all shapes need the same nurbs structure.
_editPtOn (for the rail)
_pointsOn (for the shape(s))
i like _setPt best to move the shapes end points to the rails edit points.
⌠feel free to push (by answering) some of the old topics ⌠search for simplesweep and simple sweep in this forum
my 2 cents - cheers - tom
Why do you use the âRefit cross sections withâ option? Usually it destroys the original structure of the input curves and will try to generate a refit surface with a quite different structure.
Have you tried the âExtrude along curveâ command?
! _ExtrudeCrvAlongCrv _Pause _Solid=_No _SubCurve=_No
It preserves the original structure of the input curves, but with one caveat: the extruded ends are parallel along the path.
An alternative is to use a trick with merging the two halves and then removing the unwanted intermediate knots:
The Refit cross sections with option was not used. Please ignore that in the screen recording.
I also tried ExtrudeCrvAlongCrv, but it does not give the same result as Sweep1 in my case. I added an image for comparison.
I attached the CAD file too. If anyone can make a clean/simple Sweep1 result from that file, it would help a lot.
Where?
Sorry, the CAD file was in an earlier message above. I am attaching it again here.
soru.3dm (241.4 KB)
Have you tried the thick that I described in my previous post? It works flawlessly with your curves. I also added a 3rd video above where I showed the solution with your 3dm file.
Thanks, I understand the half-profile + mirror workflow.
I can already make it that way.
My main question was why the full profile creates extra CVs / isocurves with Sweep1, while the half-profile gives a cleaner result.
Thank you, this is exactly what I was trying to understand.
So it seems that when the rail passes through the middle of the profile, Sweep1 does something extra internally and adds more CVs / spans.
But when the sweep is made from the profile end point, it preserves the control point structure much better.
Now it makes sense.
The mirroring itself does not give you a clean result. Itâs the merging and subsequent removal of the extra knot in the middle that give you a clean result which matches the input curves.
Thanks, I understand your workflow.
I was mainly trying to understand why the full profile behaves differently. @Lagom answer clarified it for me: when the rail goes through the middle of the profile, Sweep1 adds extra CVs, but when it touches the profile endpoint, it stays cleaner.
Thanks again for your help.
As I mentioned in the video description under two of my 3 videos, the Rhino developers must fix the big related to the unwanted rebuilding of the âSweep 1 railâ surface, in order to achieve a clean result similar to the âExtrude along curveâ command. This bug have been reported many years ago by multiple people, yet itâs still here in year 2026. If we are lucky, Rhino 10 may finally offer a better âSweep 1 railâ tool.
In many cases itâs not possible to use an end rail, because the latter is located in the middle, thus the user is left against the challenge to look for an unusual solution.
An alternative is to orient a copy of the profile curve to the opposite end (with the âOnCurveâ option to maintain the orientation relative to the rail curve), then build a 3d scaled copy of the rail along the end of the profile curves.
Use the following commands to achieve these two steps:
! _OrientOncrv _Pause OnCurve Copy
! _Orient _Pause Copy=Yes Scale=3D
If you build two scaled copies at either end of the profile curves, you will be able to use either ! _EdgeSrf or ! _Sweep2.
Yes, this is why I showed 3 alternatives to choose from. I didnât include âLoftâ that could be built through the 3 rail curves, because it generates too much of a deviation at the top and bottom.
