I know this is a common issue, but I'm having problems getting the object in the attached model to fillet. It's a sphere with a letter boolean'd out of it so that the text appears punched into the object (or at least that's the idea). I want to fillet the edges to give it a softer, more organic look.

However, either fillet gives no feedback at all and just fails, or it sits there calculating. Is this just the wrong tool for the job? And if so, what should I use?

TIA
Dan

Hi Dan. If you scale you model up by 10, then fillet and scale it back down again, you can get a fillet. I think it's just some tiny numbers in that curved part of the Inner J.

Hi Dan, also what radius are you attempting to fillet your original model with?

That model is already at a small size, so only a pretty darn small fillet radius will actually fit in the space that is available there. You may be trying to use too large of a radius than will physically fit in that area.

Here is what it looks like over here loading your model, selecting the entire object and then running Fillet with a radius of 0.01 units:



Sometimes you may be surprised at how much space fillets will actually take up, when you run into filleting problems one of the first things you can try is to try a much smaller radius, like 1/10 or even 1/100 of the value you were previously trying with and see what that does.

- Michael

I achieved the fillet by changing the units from cm to mm, I find this a strange behavior, I wouldn't expect changing units would effect anything in MoI's calculations, if a sphere is 0.06cm in MoI and I change the units to mm isn't the sphere still the same size ?
I would think the fillet should work in any unit type, be it inch, ft, mm etc, I understand if you model something tiny weeny, but this particular example should have no problem no matter what unit type is used.

-

But Danny, do you have "Scale on unit system change" selected? under unit options...This would change the behaviour you describe.

Hi Danny,

> I wouldn't expect changing units would effect anything in MoI's
> calculations, if a sphere is 0.06cm in MoI and I change the units
> to mm isn't the sphere still the same size ?

Well, generally algorithms work on just plain numbers...

So for example if you have a sphere that is 20 units in size (whether current units are cm, mm, miles, whatever), that could potentially give a different result than a sphere of 0.0001 units in size even if the units were set to give that some equivalent physical size.

Mostly tolerances are based at some particular fractional value, like 0.001 units regardless of what the current unit system is.

That's pretty much the regular way that CAD programs are set up. Stuff works on numeric values, and a unit system is just a separate tag to mark what those numbers are supposed to be referenced as physically.

But if you try to model a 0.06cm sphere with your units set to Miles, you will be using some really small numbers for your sphere and run into various issues, precision stuff, running close to the tolerance value, etc...


Anyway, I'm able to make this example work by just loading the file and using radius = 0.01, it should not actually be necessary to change the units or scale...

- Michael

Hi Danny,

> I understand if you model something tiny weeny

But what "tiny weeny" means to the program is if the numeric value are small, like if you are using numbers like radius = 0.0000000001 or things like that.

For example if you have units = Miles, and then have a sphere with radius = 0.00001 miles, that is a nice reasonable size in cm, however it will still be seen from the software's view as being tiny weeny because the numbers being used are tiny.

- Michael

""""Anyway, I'm able to make this example work by just loading the file and using radius = 0.01, it should not actually be necessary to change the units or scale...""""''''

I originally tried that on my P4 with an old Radeom ati card and it just sat there like the original poster described. Could be this resources issue.

Hi Burr, over here I'm actually using an old P4 and old Radeon card as well, both about 7 years old.

Is this on your machine with the Radeon 7000 in it? That's even beyond old, getting more into "antique" land as far as video card stuff goes... :)

- Michael

Hi Michael,
Hahaha. Actually i did get rid of the older 7000 and replaced it with a 2400 hd. Since your last post, I revisited the file. I mispoke.

The only time the fillet doesnt work at that value is picking only a single edge here:



It just sits there. I think I was going around looking for a problem edge and just jumped straight to the scale post I made.

Indeed it fillets at .01 from the original file with a normal selection.

Hi Burr, yeah actually believe it or not filleting just 1 edge like that where it has sharp corners is a lot more difficult than the whole boundary.

It's because the fillet itself won't naturally connect to other fillet pieces, it will want to open up a kind of hole in the model sort of like this:



A couple of different things get difficult there - the fillet does not all by itself cut the top piece, so some additional connections need to be traced on the top to cut the top surface, and then the "side wall" piece has to try to be extended to fill in the gap that is opened up after that.

It's a pretty different kind of scenario than when everything is being filleted - when everything is filleted the fillets and corners all by themselves define a complete trimming boundary.

- Michael

I had seen that if I selected the "Straight corners" option, then it actually would produce that "Fillet-but-leave-the-original-surface-alone" result, which would allow me to do the lower level surface work if I really needed just that edge filleted... :o

Thanks to Michael, Danny and BurrMan for their help with which I was able to resolve the issues. I have realised that some edges just can't be filleted but I managed to get a useful result nonetheless. Cheers, Dan