How can I obtain single span degree 2/5 surfaces here that (because they have few CVs) are easy to align G2 with Match (or NUV) and then fillet G2 (like in the Alias equivalent)? As I have it now, the ultra high density surface on the left can be joined in without a naked edge, but makes FilletEdge fail (all prior surfaces disappear). When I rebuild the COS obtained from projecting, I can build far simpler surfaces - but then the whole shebang can’t be joined, and thus not filleted. What is the common procedure for such things?
Iterative rebuilding of the surfaces helped to reduce the number of spans while still being able to join. FilletEdge always fails with chordal fillets above 0.2 mm. What’s the procedure to get this right?
I use a different approach to building the primary surfaces. Basically overbuild/extend and then trim to create the opening rather than try to build exactly to the opening.
Build the the “large body” without the opening.
Build the lower “duct” surfaces.
Create a curve with the shape of the opening. Do not try to have this curve be exactly on the large surface. Make as simple as possible with the desired shape.
Build the “funnel” surfaces from the duct to the opening curve.
Extend the funnel surfaces so that they protrude through the large body.
Trim the large body and funnel surfaces to create the opening.
Fillet the edge of the opening.
@davidcockey Yeah, overbuilding and intersecting (unlike trim, it works with history) to find a pleasant theoretical edge would be better, but in this case here, I have to follow the 2D curves and sections exactly.
@Gijs Manual filleting, with trim and G2 blend, provides no chordal fillet, making it look “pinched”, and the trimming ain’t working automatically; please see screenshot below.
It’s always the rebuilt (still very dense) inner left surface that kills FilletEdge with a too large value; without that particular surface, the rest can be filleted 1 mm chordal no problem. Maybe more CV massaging necessary and a better G2 match, so the FilletEdge can “find its way around”.
Example.3dm (889.0 KB)
Hi Lagom, WIth the Pipe trick and about 15 minutes of manual curve building and edge splitting , 2 rail sweep , I got this . Not sure if it really means anything. —Mark
Cheers, that’s the best solution for this problem so far. Still, the two surfaces are too dense for direct CV manipulation. I wonder (having to strictly keep to the 2D curves) if there’s some way to simplify them further; down from 91 to 23 CVs is at least something ; )
It depends on how the curves and any other constraints on the surfaces are defined, and the allowable tolerances. Depending on the curves and other constraints 0.001 mm maximum deviation can be a very different situation than 0.1 mm deviation.
I work with modeling boat hulls from “lines and offsets” and from scan data, and have developed methods minimizing the complexity of the surfaces while matching the input data within the needed tolerances.
Yes I put that lower surface in there because of a naked edge in the original. So , maybe I rushed it just to help. Thank you Gijs for providing your Global Continuity checker.—-Mark
I cannot rebuild the curves and surfaces (degrees/spans) further (see screenshot) and still achieve appropriate edge/surface continuity for going down the Join + FilletEdge route.
If the shape of the inner 2D curves would be up to the modeller, then one would go down the overbuilding route.
If FilletEdge fails yet surface continuity is good, try the “pipe + trim” method.
Is this the right way to understand things like these in the “Rhino way”?
