I wanted to get such on the long item in that file, but could see an issue, so try on a cube, and as filletSrf makes a surface at a time, and I have Jims settings of extend = No and trim = No as advised, I dont get the rounded corners. I have no other options in FilletSrf in V5.
Is that ok or is my simple logic flawed. All it does is take the edge of the fillet and sweep it round the corner.
Unsure about the straight bit though, but cannot see how to avoid it.
In the case of the cube-ish part, Iâd turn on Int snap and use Split Surface with Isocurve (right-click on split.) Split each fillet where it hits another fillet, so that all fillets are trimmed back to each other. Then you could use networksrf to make the corner.
The other blend is more annoying and there really arenât great solutions to that one. But hereâs what I did to do it anyway. Split with Isocurve the lower two fillets where they intersect the fillet thatâs coming down at an angle (you may need to extend that angled fillet first because itâs a little short.) Then blendcrv the bottom edges of those fillets. Then draw a line between the other two corners of those fillets and pull that line to the angled fillet. Trim the angled fillet with that funky curve. Do a sweep2 to fill the rest in. (Youâll need to do some trimming so that you can make your sweep 2 tangent to the other surfaces.)
Just do the fillets all in one single use of the command (remembering to set the radii to 0.02 or 0.01 as needed). Everything closes and the corners are all done for you.
Hi,
I have just spent an hour struggling with it, using FilletSrf, and now find that FilletEdge is the correct way.
I feel like a ping pong ball , I am out of time here and cannot get over the l;ast fence in this race, rounded corners and filletSrf and now filletEdge is the way here.
What are the check points to do to establish which should be used ?
dont boolean I am told, so follow that, but then that item elsewhere, (I cropped it for simplicity here), needs filletSrf elsewhere and that needs no booleaned surfaces, so how can one have filletEdge and FillerSrf on the same object ? Does one boolean and filletEdge, then explode and FilletSrf ?
My attitude is that FilletEdge is quick and convenient so I will often try it first. If it works then Iâm done. If it doesnât work then I move on to FilletSrf. And quite often a failed FilletEdge will leave some useful surfaces that I can trim and use.
The more you use these tools in different scenarios, the more you develop an instinct for where they will work and where they wonât.
And yes, I mix boolean operations on closed polys with explosions (or extractions) all the time.
The biggest problem I have with filletEdge is that it mostly fails. That means you set things up to use filletedge and 9 time out of 10 that turns out to be a waste of time. Not only that it most often turns turns out you have to tear things apart again to fix the problem or just start over and organize things differently. `
But the main problem I have with filletedge is that it produces sloppy inaccurate geometry (as do many of Rhinoâs other solid modeling tools)
Personally, in my work, its impossible for me to use filletedge. The models I create I have to be exported for CNC machining and for approval by the companies that contract for the work. When i use filletedge I get complaints that the fillets are not recognized as fillets.
Bottom line is â if use filletedge I would have been out of work long ago.
If your work process doesnât have to be accurate and you donât mind playing Russian roulette with your modeling strategies go ahead use FilletEdge.
You have to admire Steve. He knows what he is tackling involves a lot of fillets and he knows almost all of those fillets will not work with whatever method he decides to use to make them, but he persists.
The check point is âwhatever worksâ, works. Rhino is a surgical tool; worst case scenario you make everything manually âfreeformâ.
Theyâre just two constituents of Rhinoâs hundreds or thousands of tools.
What? I mean sure? A well distilled âwork flowâ will definitely help clarify any routine process. You might even be able to script it or create an Eto Framwork to antiquate Rhinoâs âfilletedgeâ conundrum
The first thing to notice here is that the the fillet going toward the right of the image is bigger than the other two. First rule of fileting; biggest first.
So, as a first step, fillet the top and the right surfaces. Second step, fillet the top surface, the right surface and the fillet you created on step one, with the left surface.
Youâl get exactly the result you see in this image.
Also note in that image that at least two out of the 4 edges of the corner fillet are cut planar. The plane is achieved by creating a tangent blendcurve which matches positional to the corresponding end of the small fillet curve . This makes 3 controls points and this forms a plane. It is very important to reduce the tension of all edges, to minimize the deviation. If you know how to construct such a fillet, it is easy to do. But depending on the corner fillet, you sometimes need to adjust the surroundings as well. Its all about creating perfect conditions. This is also why automated tools often fail, even in Catia etc.
from biggest to smallest Radius
_filletSrf (Extend = no)
same radius
will result in a part of a sphere in the corner.
in V7 there is
_SphereTangentToThreeSurfaces
in V5
(1) _filletSrf
(2) remove unwanted parts of the fillet
_split Isocurve shrink = yes (first object snap is intersection, next endpoint)
check if the 3 edges meet
_gcon (g0 expected)
(3) draw a sphere with your Radius
_sphere centerSnap for the center
(4)rotate the sphere, so the surface-Seam is of the corner
_rotate3d or Gumball
(5)
_trim
_shrinkTrimmedSrf
the idea is to not have a pointy pole - which is more robust but maybe other call it nit-pickingâŚ
There is a test command with a lame interface that I wrote for testing spherical corner creation in FilletEdge. testSphericalPatch. Pick the surfaces that were filleted, not the fillets themselves. The order doesnât matter if there are 3 of them. For more than 3, you have to pick them in order around the patch.
âfreeform surface modelingâ 
