Nope absolutely not. In any object above X complexity and curviness, there’s a giant hassle with “What do I make my ‘base’ model and what do I add as ‘fillets?’” What is the “solid” I need to make and in what order do I FilletEdge so that everything turns out right? I listened to the other Jim yammer on here about this for ages and had a project where I could really test it out–a boat with hatches with lots of lips for sealing edges and mold details that all have to follow contours of the hull or deck–and it was like the scales fell from my eyes, only the most boring simple shapes get FilletEdged anymore. It’s a far more flexible workflow and produces higher quality results, speeding up the tedious bits of that workflow would blow the doors off of being able to identify 2 or 3 more of the half-dozen fillet situations FilletEdge can’t handle vs SolidWhatever.
Me too. I’ve been trying to get the devs to clearly and concisely explicate the specifics, but all we get are things like “oh it’s not a solid modeler”, or “oh it’s not parametric” – all while obviously there’s nothing Rhino can’t do. It’s just a matter of time, macros, scripts, GH, etc…
My opinion is it’s probably something nurbs volumes could solve, but that’s just me theorizing without a dev saying “no it’s this specific reason, not that one, etc”
Maybe I was a bit facetious. It’s actually more complex, but post op finishing is usually something alot of manufacturers fail to include in their processes.
Something alot of ppl don’t realize is the infinite permutations of how to go about manufacturing a particular ‘inner round’ or ‘outer round’ geometric entity.
Alot of keyboard/mouse button pushers like to think ‘click the fillet button’ and put ‘fillet here’ on the orthographic GD’s&T’s, but the machinist or CAM/CNC programmer/tooling-configurator just needs to know what the ‘radius or diameter’ is intended to be and what the tolerances are.
Most ‘fillet’ operations are temporary graphical aesthetics, cause the real intent is what? Round a 2.5D object’s edges? Why?
The laptop I’m typing on here, has a puny fillet on the edge adjacent to the keyboard which is awful. The fillet should have be 10 times bigger. They jumped the gun willynilly and just went with it like it was good enough for production. They should have just left it sharp 
Maybe their intent wasn’t to make the user have a nice round edge, maybe they were like “oh the part will come out of the mold now, so why make the fillet bigger”.
Or “why use blend edge or any other cool sub-D’s or something, the part releases fine, just leave it square and boxy with puny fillets, good enough”.
Imo, when designers run out of talent they click the ‘fillet button’ and expect it to work automatically. Which is fine, cause why doesn’t Rhino extent srfs or trim edges for fillet commands automatically? 
If it did I would fillet more often lol.
Eventually I plan to test out the ‘blend edge’ more often to learn the differences their, it seems that Rhino needs to learn to have more tolerance for nudging or smudging some parameters to succeed more at filieting things automatically like the other CAD’s do, but we all know even the other CAD’s fail more often at fillets than users admit.
I’m pretty sure all the CAD’s are getting slightly better over time. They all used to be bad at fillets 20 yrs ago. It’s all a matter of formats, macros, scripts, algos etc.
Maybe I was a bit facetious. It’s actually more complex, but post op finishing is usually something alot of manufacturers fail to include in their processes.
What industry are we talking about here?
If I’m designing a plastic housing to be injection molded, the design needs to be 100% as it’s intended to look before handing over to the toolmakers. They don’t want CAD that’s 90% finished with instructions on how to do the other 10% (unless we’re talking about surface finishes) - the responsibility for the design is in my hands, their job is to machine the mold using the cad I give them.
Then there’s 3d printing where especially if it’s a jig or a quick prototype - I want the part coming off the printer without needing to do post work.
If a fillet gets the job done - I’ll use it. I don’t compartmentalise my cad tools into ‘right’ and ‘wrong’ tools - everything gets used to get the design where I need it to go. Fillets in solid modellers have a lot of flexibility and let me focus on design rather than manually building a surface when a fillet achieves the same thing.
Edit: Also in the concept phase of a design, fillets are very useful to make quick changes to geometry.
This is a rectangle with two ‘fillets’. 1st fillet creates a full round on one side of the block, 2nd fillet runs along the curve to create a variable fillet. With Rhino I can’t (or don’t know how?) to use the fillet tool to provide this level of design flexibility and iterating.
Imo, it’s the lack of “deep” typology.
Rhino breps have topology, but each element doesn’t “see” past the first neighbor.
Often trying to walk through the topology you get lost without having idea if you are inside or outside a trim, if a fragment should be kept or deleted, etc etc.
Edges should be connected each-other more (like half-edges in meshes).
Imo, creating a better topology is the first step, then making good fillets will be easier.
Everything I’ve said is a guess.
IMHO:
This is the point with Rhino.
I’ve been using it for 20 years for models, prototypes, moulds … things that usually require a so called Mechanical CAD.
Rhino is not meant for this.
But for artistic work, maybe architecture, sure for aesthetic objects, free-form shapes.
But it’s also very flexible.
You can draw anything with Rhino, but for ‘MCAD’ stuff it often requires quite some time to get the job done, compared to more specialized softwares.
This is correct. Technically you can detect adjacent surfaces and extend them. You then need to ensure that the extension does not break the other continuities.
This is a pure algorithmic problem. However there is another problem…
…if you analyse the shapes coming from parametric modelling software’s you will likely detect that they added surface complexity to the affected surfaces. Something which is not a big deal for technical models usually created by Creo, Catia or Fusion360 and for the majority of their use-cases. However if you decide to create aesthetic surface models you rather choose a direct modelling application. Especially if you want or if its required to control the surface complexity. So it is a problem if during a post-process (which applying fillets usually is) something breaks your clean surfaces into garbage. Clean modelling is not only about continuity but also about representing a shape as simple as possible, and as complex as needed.
So other CAD, like ICEM Surf is choosing a hybrid approach, just as Rhino is doing. Only that the Rhino version is not as good yet. But still I would rather see something which only optionally breaks the surface layout. In the end, a good surface model needs manual intervention from time to time. But with the right tooling its not a problem.
In mass produced manufactured parts, that is seldom how it works. The designer of the part usually doesn’t have sufficient knowledge of tool design to make that happen.
I didn’t mean to imply the part designer is involved with tool design. Was just saying that this idea that the toolmaker solves a fillet issue (if that’s partly what @lander was implying) isn’t how I work.
The toolmaker has zero input on the design of the part - their only concern is tool design (in our manufacturing workflow at least). It’s my responsibility as a designer to ensure the part cad I give them is 100% complete and how I want the part to look - they don’t have any leeway to make adjustments or decisions about fillets or any other aspect of the design.
All industries that make parts that have sharp edges that can cut the customer’s skin open, causing injury – sheet metal, injection molded parts, etc.
It’s way too common.
Most manufacturers are guilty of skipping the steps to mitigate this issue. Which is why the debate over fillets will always be such a silly one, until manufacturers actually start doing their job and having actual deburring departments address the sharp edges they keep allowing their products to have.
Look at any product laying around, and you’ll see sharp edges on it due to avoidance of cost of removing said sharp edge.
True unless they’re in the business of doing that part. It’s funny when ‘sticker makers’ don’t want to have to do any CAD and they just want to make stickers by pushing one button or so.
I agree with Jim’s point. The people creating the real geometry from the digital geometry will have to deal with many things that the designer rarely will ever consider.
Unless the designer is also the person creating the whole design in reality, then they will face every problem associated with that.
But how often does someone have every tool necessary to do all phases of CAD/CAM/CNC etc. …
That’s the dream. So much so, companies often just don’t do post op work, and pretend it’s fine with burrs and sharp edges everywhere.
You might even put a fillet in the design, and the toolmaker will bypass it and pretend it doesn’t have to be there. There might be a burr, and they’ll just say ‘well it’s too expensive to do post op work soo’.
Fillet’s technically get in the way and aren’t very useful. They’re just symantical cosmetic aesthetic accents is all. And if they’re being used very elaborately into the design, then they’re technically copouts for the geometry that should actually be there.
At some point the design should realize that the curvature probably should be alot more organic than some basic fillet radius transition.
Designers often limit themselves to think, ‘oh I’m a really good designer if I figure out how to get these fillets to work out’.
I used to play around with fillets alot in 2004-2008, then I discovered freeform surface modeling.
When you have the freedom to create any compound curvature organic geometry you can dream up, why would you bother to think of things as ‘oh that’s a fillet’ or ‘that needs to be perfect rolling ball true fillet’?
The question should be, does that edge need to be sharp or round? If it needs to be round or smooth then is it sharp? No burrs? K good no burrs. Cause the customer probably wont like burrs cutting their skin during normal use.
As long a the design intent calls for a 2.5D rectangular prism with 1FNR fillet and 1variable fillet then Rhino can do that easy not a problem.
On second thought, I’ve seen Rhino fail to create FNR’s 
poor Rhino 
I hope R8 or 9 won’t have this problem.
Well, that’s why GD’s&T’s are imperative, otherwise the tool maker will be left with too many infinitessimal interpretations of the design intent. And you might end up with different fillets or lack thereof.
Well, yes and no. You definitely want to remove anything up for erroneous interpretation, and the toolmaker needs every GD&T possible, otherwise you’ll end up with lots of scrap and lots of wasted money, etc.
A fair amount of tolerance is important. And please tell them “no burrs please” 
I can’t say that I agree with the sentiment that fillets are not “technically useful” - they can be a great tool with functionality beyond aesthetic purposes.
When I used to design press tooling for sheet metal forming, the process would typically involve modelling a die and punch, then applying a 10mm fillet to the edge of the die so that the sheet metal would slide in smoothly without catching sharp edges.
As another example, machinists would often advise me that the smallest ball-nose cutter they could use for a particular part would be 5mm. In that case, on the part in question, I would apply fillets no smaller than r2.5 - knowing they would be able to machine them. I also know for a fact that if I had sent my machine shop manager a STEP file with no fillets and told him to program it himself, he would have thrown the model back to me and told me to do it properly.
I understand that there are ways to work around Rhino’s FilletEdge command, which cannot compete with what the Parasolid CAD packages offer - but a simple fillet operation that would take literal seconds in another package can take tedious minutes in Rhino. I do think it would be time well spent to improve fillets in Rhino if possible - a fillet command that is as robust as what other packages offer is never going to be unwelcome
Fillet’s technically get in the way and aren’t very useful. They’re just symantical cosmetic aesthetic accents is all. And if they’re being used very elaborately in to the design, then they’re technically copouts for the geometry that should actually be there.
At some point the design should realize that the curvature probably should be alot more organic than some basic fillet radius transition.
I guess I don’t see the fillet tool any differently to a sweep or a blend or anything else - it’s a tool that gives me geometry I desire.
I don’t usually have the luxury of months of design - we typically move through product design & development quite rapidly. Spending hours getting a round detail when a fillet gets me there in a few minutes is often the case. It’s the very few exceptions where a product requires a high level of surface continuity & detail that I’ll build it by hand.
A recent plastic part (housing for some electronics) the design took a week (2 days for initial design, 3d print, another 2 days to fine-tune design). First-off injection molded samples 2 weeks later. So for me there’s times when a client doesn’t need fancy surfaces - they just need a functional part that looks good and fillets are part of the toolbox that gets the job done.
Use of fillet tool highlighted…
(I don’t really understand this idea of fillets getting in the way but maybe we work in very different industries or we’re speaking about the fillet tool very differently.)
@ftzuk can you send me that file?
STEP attached.
radius.stp (55.9 KB)
thanks,
I was trying to see what Rhino could do with this and ran into edge continuity issues between the fillet surfaces. But I’m suspecting it’s impossible to get tangency with rolling ball fillets, as I see the discontinuities in this file as well, up to 1.333 degrees off:
Someone was asking the other day why Rhino isn’t seen more often as a recommendation for 3d printing hobbyists and I think part of the reason is the same reason I don’t use Rhino for a lot of design work: In cases like this I don’t really care about edge / surface continuity: I just want to create something that solves a problem (either for myself or a client) and move on to production or 3d printing. I understand that many industries require a high degree of surface accuracy, but there’s also a lot of industries where surface finish and edge continuity etc. is secondary to the task of solving a problem.
Maybe my case and the hobbyist designing stuff for 3d printing isn’t who Rhino is aimed at, but it just seems like there’s a whole audience who’d jump on Rhino (it’s super affordable compared to other cad packages) if it was more robust / relaxed with things like fillets.
No it’s that Rhino would have to move to subscription pricing to appeal to that market, even though everyone claims to hate it. $995 software is anathema to the crowd that thinks software should be “free” just because they’re “non-commercial users,” as if that matters for anything else people spend money on.
of course, I wasn’t posting that to say it is useless, it would be welcome if Rhino can handle fillet cases like that. In a way the technology to accomplish it is already there, using FilletSrfToRail, but requires quite a bit of hand work. If you know what you want to make, that’s fine, but if this is part of a design iteration, it becomes cumbersome.
… is very buggy.
I had a lot of hopes in this new feature - but I have to say, it is really disappointing.
FilletSrfToRail is failing as soon as at least one surface is circular / degree 2 + rational, non-uniform. (…a revolve or another fillet)
it would be a great feature…
It seems to be working in this case.
The design intent for Ftzuk’s part seems to be to have the bottom edge of the fillet appear as a straight line when viewed from the top. So projecting that line onto the surfaces and then using the projected curves for FilletSrfToRail gives a fairly decent result. Doesn’t take much time either.
radiusx.3dm (172.7 KB)