Patch.3dm (369.1 KB)
This seems to be a straight-forward task that shouldn’ require any particualar concerns yet it apparently doesn’t work - that is to say it doesnt’ patch the hole but merely creates a surface that’s in the general shape.
What is unsatisfactory about the results?
Patch is very powerful, but frequently requires some experimentation to find the best setting for a particular situation.
Depending on the size of the hole a “Sample point spacing” of 3 may be reasonable, or too large or too small. You probably need at least 100 points around the perimeter of the opening.
The number of Surface U and V spans may also need some experimentation. Too few spans and the resulting surface may not conform close enough to the input cuves and points. Too many spans and wiggles can occur.
Also experiment with Stiffness. A smaller values generally improves how close the resulting surface conforms to the input, but can also result in wiggles.
I[quote=“davidcockey, post:2, topic:34207”]
What is unsatisfactory about the results?
[/quote]
I cant join the surfaces.
I can’t match the surfaces because the patch is a trim.
I think the patch tool could be much more useful if the inputs would respond in real time and there was some way to maintain a joinable continuity. Perhaps the tool could take info for points spacing and UV count from the surfaces being used to create it to at least get close to what’s required rather than leaving it all up to trial and error.
I think if a tool requires so much experimentation in order to get an useable result - a result that is most likely desired by the overwheliming majority of users then it should do a better job.
Here’s another attempt;
I don’t see how this could possibly warrant experimentation-
If your patch is not precise enough , common tips is to temporarily change absolute tolerance to be able to join.
I don’t know for Mac version, it seems you can’t select the patch for joining, strange.
Can you preselect both part and run join with absolute tolerance to 0.1 for example ?
No need to change the absolute tolerance to join two surfaces which with edges too far apart. Use JoinEdge which over rides the tolerance. Using JoinEdge or loosening the tolerance just for joining can cause problems with other modeling so should usually be avoided.
I used Patch on @JKayten file with point spacing of 0.5 and 10 spans in each direction to create a surface which would join without altering the tolerance or using JoinEdge.
Curves could also be created across the opening to use in building the Patch. Use BlendCrv with the Edge option and curvature/G2 continuity, and adjust the handles to the obtain the desired shapes. Patch using those curves. With enough curves Patch should produce a surface with curvature close to the outer surface. Then use MatchSrf to match the outer surface to the patch surface.
JoinEdge will work only one egde at time, in case of polysurfaces you will consume many time to select edges one by one, where changing tolerance do in one shot.
Anyway, i agree, it’s not the best solution, but can be useful in some cases, and need to be know, IMO.
Apart from all the above, sometimes it’s cleaner to just close up the surface directly by adding a couple of knots in the top span and crunching the edge down to a singularity.
NoPatch.3dm (138.4 KB)
-Pascal
Thanks David, that did the trick. I wouldn’t have guessed I needed such a small increment for spaceing. I wonder if there’s an aspect of the math formula that could take into the account the size if the surface used and return a suitable spacing number to insure a joinable surface.
I ended up with something like that by creating a few more closed curves to the loft and simply capped the hole. That was fine since a portion of the form was to be cut away.
One of the tediums of this method is after duplicating edges to create curves for lofting, I must rebuild them to keep the paras down to a minimum. The curves needed for lofting don’t require the precision that generates so many extra points. Working through this particular form I was put off by how heavy the surfaces would become using the curve network, or sweep and match surface tools.
The solution you offered is very clean. It’s not clear exactly how or what you mean by adding a couple of knots. If you find the time, would you make a short vid for that? That technique is probably a very useful one for many of us.
Thanks to all for your kind attention and help!