Hi!
I’ve tried numerous ways to fillet this but couldn’t find a clean solution. I’d be forever grateful if anyone can solve it in the blink of an eye! (File attached)
Thanks,
Rizon
FilletHelp.3dm (2.5 MB)
Here is a pure rolling ball solution made with FilletSrf:
FilletHelpx.3dm (656.2 KB)
2 Likes
2 Likes
I think rhino 9 WIP has a new command for this
fillsrf
1 Like
vanilla rolling ball rh8
also pure / “vanilla rolling ball” with another focus, other order of which fillet “runs” first.
fillSrf rh9 / Wip
… i did not try with current Rh8 patch though…
@menno - great work, thinking of skipping some lectures for my students on the workflows in above solution.
FilletHelp_tp01.3dm (393.3 KB)
cheers - tom
4 Likes
just playing around.
2 surfaces
i guess it should be possible with a typology like this - but i gave up, after it got late ans _matchSrf did not do, what I thought (hoped?) it should …
3 surfaces
I also played with this typology
but both attempts above are only nice screenshot-concepts… not worth sharing the mess of CVs.
EDIT:
3 surfaces - accessible
i think this is a nice accessible vanilla Rh8 approach:
build 2 simple surfaces left and right
(“left” is connecting the big Arc / Cylinder with the planar Srf - a simple loft)
(“right” is a modified initial loft between both R0.8 fillets.
build the inner “stripe” that finally has more dense Isocurves.
this stripe was also single span at the start.
matchSrf this stripe with history and pull option, keep isocurve, no refine
this setting above will allow to modifiy the left and right surface until the shape and flow is nice.
stepwise add Knots to the “stripe”
optionally do some _matchSrf with refine at the end.
FilletHelp_tp04_srf6conly.3dm (251.8 KB)
cheers - tom
3 Likes
Hi @rizon Well adding to this discussion… I keep experimenting… please find 3 different approaches… In each of the files youll find organized in layers accordingly to the order of the gometry creation (layer 01 was dont first… layer 02… second. and so on)…
I think I like V3… but V2 is also nice… and easy to make…
V1 was my attempt where I was trying to use single span surfaces… I was not able to have a nice result…
Best regards…
FilletHelp_V1.3dm (530.5 KB)
FilletHelp_V2.3dm (494.1 KB)
FilletHelp_V3.3dm (519.3 KB)
1 Like
i like your V3.
here is my go with this surface-setup
FilletHelp_tp03_v3.3dm (1.1 MB)
i started with a simple _loft (green) between 2 edges (tangent) and one additional crv (blue).
_split at the curve position / knot
_matchSrf to get the smaller surface to the edges
_matchSrf to refix the inner connection
_setPt y=yes to get the C-shaped triangle nice
_extrudeCrv _changeDegree _matchSrf for the remaining extension of the R0.8 Surface
… i feel like there will even be an easier solution…
2 Likes
i gave it a shot yesterday and ended up with something looking virtually the same to this solution
but i used up more than an hour experimenting and in the end i still would have to match more and just deleted the entire thing completely again.
it seems coming up with a solution here without matching or using some weird filet surface kung fu ala @jim is impossible or what… i mean we have some hope on the horizon with FillSrf, still should be easier than that. i wonder how many lost souls are out there that gave up on a “simple” fillet 
2 Likes
yes - i already thought of doing a topic “_matchSrf Graveyard” in the gallery category
4 Likes
WOW! Some beautiful stuff here, didn’t see this coming! Thanks a lot all of you, this is perfect!
Through this process I discovered Plasticity, intriguing package… just sharing FYI what came out of it.
1 Like
This first (plasticity) solution is especially ugly (to me)!
A perfect collection of what not to do .
When a parametric SW uses rolling ball to solve complex situation the result is always weird.
Hi again.
I kept experimenting… while my previous v3 “worked”… i thought that v2 was more "elegant… so I kept experimenting with that and came with these two other solutions… (in both of them the “sweeped” profile is cutted in a diagonal fashion using a simple diagonal line using “flow along surface”…
V5 has a almost impossible to detect tangency discontinuity… but V6 is I think the simplest and easier to do…
Looking at the solutions you post from plasticity… I see too that it wants to add an additional "flat triangle on horizontal surface… like in my V2…
Haven’t ever looked into plasticity… I guess that it will get tricky too when solving some stuff…
My only regret in this exercise is that I really thought I was going to be able to solve it with single span surfaces. but I failed… … but matchsrf is not very intuitive… and gets even more tricky when the patch layout gets some “triangles”
FilletHelp_V5.3dm (265.1 KB)
FilletHelp_V6.3dm (822.5 KB)
4 Likes
This single-span exercise can be proposed in this topic; someone from McNeel might find a better solution, perhaps using the new commands in Rhino 9.
1 Like