in many cases - also for varaible fillets to cheat to get better results - avoid crease-split / tangent-split surfaces. this is a much nicer workaround than RXX.999
common commands:
_useExtrusions → Polysurface
_creasSplitting
_mergeSrf (smooth = no)
_DivideAlongCreases
_shrinktrimmedSrf
_shrinktrimmedSrfToEdge
… and a few commands support “split At Tangents = No”
_revolve
_extrudeCrv (useextrusion = no)
…
for the posted geometry:
_extactSrf copy=no
_mergeSrf (smooth = no)
_join
_filletEdge … wil do the job
did not test original data as it is a rh9 file that i can no longer acess on mac intel… sadly.
As a clue for the developers, this appears to fail because the errant surface and the half-cylinder are both trimmed from taller surfaces. Untrim and scale the original surfaces to fit and the resulting polysurface will fillet along the desired edge ok.
Probably explains why people recreating the form don’t see the bug - no reason for them to have used trimmed surfaces.
I have been using Rhino for a while now, and am still fascinated by some intricacies of the software. Some issues are not obvious at all, and I would have absolutely no clue how to resolve them without this community.
Thanks for giving insight as to why this bug occurs, it helps a lot.
This has evolved into a nice discussion. However, no matter how many explanations, reasons, topologies, user habit, surface aspect ratios etc… the fact remains that Rhino should be able to construct such a simple fillet.
Let’s look at fillet edge / fillet surface as some mentioned. What is stopping Rhino, under the hood, after feeding it the edge in the example to run a fillet surface command after fillet edge fails?! Where the input is the two surfaces shared by the edge or something along that logic. In reality this topic shouldn’t even exist -there are plenty of automations, even stupid ones like the one exemplified just now, that can be built and run in 0.1 seconds on today’s hardware… it really is bewildering to me that I had to start a topic like this one.
if you often run into situations where simple fillets dont work you might be doing something wrong. a simple fillet quickly turns impossible with the right amount of mess. should be “fairly easy” to understand.
Yes, of course it should. But the world isn’t perfect and sometimes (always?) software has bugs, omissions, oversights, false aims etc and these - maybe - get ironed out over time as people report them. Software is, except in trivial cases, too complex for mere humans to make defect free so this correction process is inevitable. Thanks for reporting more filleting defects and helping make Rhino better.
I think the point is clear and we all agree that these cases should just work. If you fillet all day long using FilletEdge then Rhino currently is not your best tool. This situation will improve over time, since Menno started working on it for Rhino 9.
no problem here, you can either do both at the same time or just one first then the other, in both instances it works. again you must be doing something wrong, it also looks pretty weird looking at your screen shot maybe you explain what exactly you are trying to achieve there? i did not have even the slightest of issues doing these indeed simple filets whatsoever. maybe its a mac vs windows thing?
unless you are indeed talking about making one side of the fillets bigger than the residual dimension, in which case you can use FilletSrf, it is slightly more involving since you have to trim off the adjoining surfaces but nothing that would worry me honestly.
This case does not fall in the categories you listed. It is simply a tool left under or un- developed for too long. Devs need to improve or adapt this tool to current standards, that’s all. Issues with filleting in Rhino have been reported for years, and because Rhino is not parametric it is a very big time sink for everyone to have to manually address these trivial cases.
Yes, I mean 4.5 or bigger. I’m sorry my dimensions didn’t properly reflect this in the screenshot, but in the step by step I posted I mentioned 4.5.
Now, to your argument: of course, one can indeed build these surfaces control point to control point, or even write his own nurbs kernel, or a million other ways. But that still shouldn’t excuse Rhino’s fillet srf or fillet edge tool failing. Which they do.
V8 FilletEdge works as I would expect. I do not agree that FilletEdge should be able to fillet the combination of 6 and 4.5 radius fillets on a box 10 wide. FilletEdgeExample.3dm (2.3 MB)
A radius of 6 followed by a radius of 4 works as expected.
A radius of 6 followed by a radius of 4.5 does not trim the second fillet. There is not sufficient space for the second fillet to be tangent to both surfaces. If the user wants a partial fillet then another method needs to be used.
What would you expect as the result? There is not sufficient space for a complete tangent fillet of radius 4.5. There are several alternatives for a partial fillet of that radius. How should FilletEdge decide which alternative is “correct”?
(4.5 fillets in image below were not created by FilletEdge)
if you cant stick to one essential dimension in 2 short descriptions it definitely clears my doubts about your modelling methods. also actually modelling one type that fails and explaining a completely different topic in the worded description is as weird as it can get.
also now your topic became a feature request not a fillet fail bug report.
again just use FilletSrf and be done in 1 minute versus 30 seconds. just imagine how many fillets you could have made during the time complaining here about theoretical issues.
Just imagine if no one ever questioned paradigms ever how you’d still be using Autocad or a pen and paper. Can you stop flying off on tangents (pun intended) every time you fail to grasp the point of something?
You ranted how I fail at descriptions yet you failed miserably at reading the title of this thread. Hint: it’s not a bug report, nor necessarily a feature request.
