I really enjoy using MoI, but I always seem to run into fillet and chamfer problems. Sometimes they work just right as I would expect and sometimes they either tell me they fail to calculate, or maybe they don't fail, but nothing happens, or they don't technically fail, but create some wild and crazy results.

I could use some guidance on what is required to set up a valid edge to apply a fillet or chamfer to.

Here is a picture of what seems to be a not-so-complex shape, with some edges highlighted that I'd like to chamfer.

And then three pictures showing different ways it can fail.

I just can't get this object to chamfer. What about it is invalid, and what have I done wrong? The chamer/fillet operations in my experience are kind of mysterious when they will work and when they will fail. But it may be that I just don't understand the constraints.

I've attached the model file if anyone wants to take a look.

And here's a picture of a successful model I made in MoI, just to show that fillets and chamfers do indeed work sometimes for me.

Hi LM, the object you posted will Chamfer ok if you select the entire object instead of only some of the edges:





It's a lot more difficult to chamfer or fillet only some of the edges that come to a shared vertex instead of all of them. That's because the fillet surfaces have to be extended and intersected against each other in that situation instead of being able to make a corner juncture patch.

In your last screenshot the issue is probably the distance you're using is too large to fit within the available area.

- Michael

Hi LM, I think there's some kind of other bug coming into play with that particular model though.

I've attached a version that I made using Shell and it can do the partial chamfers ok.

- Michael

Michael, thanks for taking a look. I was able to do a small chamfer on the whole object like you did.

I also tried another object that to my human eyes looks kind of similar.



And I can chamfer just the front edges.



I know that to my eyes the shapes look kind of similar, but they must be quite different mathematically. Are there any suggestions for what to avoid if you want to later chamfer something? "Stick with only straight edges" isn't it, because I've done lots of fillets on curves. "Only planar surfaces that join each other" isn't it either, because this model shows chamfers on curved surfaces and edges.

Weird. I tried your model too and it works for me as well (the partial chamfer).

Then I took my first model, deleted the interior surfaces and recreated them using the Shell command, and then I could chamfer it in the way I first wanted to as well.

Do you know what the difference is between those two? My original model, and one recreated with Shell? Visually they look pretty much like the same thing.

In my original model, I created a sphere, used Shell on it, then boolean subtracted I think three other spheres.

Hi LM,

re:
> Do you know what the difference is between those two? My original model, and one recreated with Shell?
> Visually they look pretty much like the same thing.

Well in the shell model the "side walls" are basically a loft while in the original they were trimmed pieces of spheres.

I'm not exactly sure yet why your original version won't work, but it seems to be failing to trace out the fillet rails properly.

It seems to be a bug when processing these particular surfaces. I tried with making 1.1 times larger sphere cuts and that one worked ok too.

So I don't think it's illustrative of some kind of general case that you would try to avoid.

Also if you run ShrinkTrimmedSrf on your original object it will do the partial chamfer ok too. Select the object, type the Tab key and then type ShrinkTrimmedSrf and push Enter. That makes the underlying surfaces of the side walls to be smaller pieces of a sphere instead of being full spheres.

If you run into fillet troubles it can be worth a try to run both Merge and ShrinkTrimmedSrf on the object and try again.

- Michael