Hi, on your example, you are starting from a D shape with two sharp corners… no matter how smooth the transition is, at the very beginning there will be a sharp crease. You can smooth it out along the transition, but at the very beginning there will be a crease :
Yes, there has to be a radius starting at zero, but you want to crease to transition to curved very quickly.
Frames 195-1/2 and154-3/4 (about 1/3 and 2/3 of the way down) show no sign of creasing in the plans. When I copy the plan frames, I get ripples when I do sweeps.
@bigjimslade Can you post a .3dm file with the rippled surface? What Rhino commands are you using to create that surface? You mention “sweeps” but the file you posted above does not have sufficient curves for Sweep1 or Sweep2.
Third solution. Surface appeared to have wiggles with the render mesh settings in @bigjimslade original file. Refined render mesh settings and apparent wiggles disappeared. Problem Prop DC2.3dm (175.6 KB)
I probably should have included this in the original I have added in my reference frames that I build from the plans. They consist of lines and tangent circles.
I drew a line through the centers and did a sweep1 and this is the [predictable] result. There is nothing to force the corner into the curve, which is the basic problem.
I have spit and rebuilt the reference curves in various ways but not found a satisfactory solution. I suspect a solution would require multiple surfaces but the trick to knock off the corner.
I’d handle this with multiple surfaces. Start with using Curve interpolate points. To set the split points on the Degree 2 circle. Then use loft making sure to use the to point option. Then it’s just a matter of matching, point editing and remove knots to get a clean polysurface.
How come lofting the frames in red (attached) creates this bizarre surface? I know the frames are not great but it would help to refine them if the loft came out somewhat rationally.
My guess is the unexpected Loft result was due to the order you selected the cross-sections and/or the locations where you selected the cross-sections.
This is typical with lofting polycurves to single curves. Rhino can’t figure out a solution with varying point and knot counts. You could split the single curve to match the knot and point counts of the polycurves. Or just run rebuild option within the Loft command. Other cad packages will just run their own rebuild option without you knowing.
If you move the curves closer to the world axis, or orient the world axis to the curves. The loft command will work correctly. You’ll still get a wonky result.
I understood that. With strange problems it’s best to rule out model tolerance and the world axis orientation first. I guess Rhino 6 doesn’t have this issue.