Why did Rhino add extra UV's to this corner blend?


I’m building a custom corner blend. The 4 surfaces surrounding the corner blend are all 6 point 5 degree surfaces. Rhino seems to have overridden these input UV’s when generating the corner blend, adding a ton of CVs. I used the ‘Surface from 2,3 or 4 Edge Curves’ tool to make the blend surface. Any idea why this has happened?

Thank You,

Hi Justin - hard to say without a file but if the edges are trimmed, you might get that - try Untrim on the edges above and below.


Here’s the file:

Corner Blend.3dm (1.1 MB)

I used split to cut through all of the surfaces, deleted the theoretical corners and shrunk the remaining surfaces. Then I tried to build blends using adjustable curve blend and ‘Surface from 2,3 or 4 Edge Curves’ tool. The end most blend worked fine, the one on the corner added a bunch of CVs. If I untrim the surfaces it creates irregular surfaces edges to blend between.

The reason I shrink the surfaces after splitting/deleting is so that I can use the MatchSrf command to match the blends to the primary surfaces. MatchSrf doesn’t work on trimmed surfaces that still have the pre-split CV layout.

Hi Justin - Untrim is your friend here - (DupEdge, Untrim, then Dupedge again and look at the points on each curve)

BlendSrf with ‘Interior shapes’ unchecked, and Refine set.


I see the upper trimmed edge created more CV’s. Why does Rhino add CVs to a 5 degree surface that is split by a 1 degree line? The lower open edge was created with the same split/delete process and kept the correct 5 degree edge intact.

When I untrim the upper surface I get this deviation from the line I originally split to:

Hi Justin, Ah, I see - a new friend is RefitTrim in this case - don’t Untrim but RefitTrim. If a surface is trimmed by a single curve across the entire U or V domain, then RefitTrim will attempt to make a new, untrimmed edge that matches the trim. As to why the trim needs to be worked on - trimming an arbitrary surface other than with an isocurve is always a tolerance-sensitive, so to speak, operation - generally there is not an exact result on the surface that keeps the surface structure. (that is what refitTrim is for - it is not exact but can be very close to the original trim curve, while maintaining the surface structure.)

(btw, another way out would be to Untrim and then matchSrf the long edges of that surface to the side surfaces)


RefitTrim worked! I’m guessing the lower edge split more accurately because it has less complex geometry i.e. less curvature and more perpendicular to the line used for split. Thanks Pascal!

The lower edge’s split falls exactly on an isocurve. That is a special case…