Odd fillet behavior from one model to another

[sportbikeryder]

Hello,

     I have two separate parts that have similar (nearly exact) geometry. One part (plate) will accept a fillet while the other (body) will not. The fillet is sort of odd in that it becomes a partial sphere and the depth of the cylindrical bore is smaller than that of the fillet radius. I understand why this can create a burden on fillet calculation, however I am not sure why it works in one part but not the other. 

While not relevant to the issue, it is sometimes neat to know what the parts are...these are gear pump components, with the feature in question being a local pressure relief cutout. The depression created by the fillet is machined via a plunge from a .250" ball end mill.

John

Fillet selection and size

Successful fillet

Failed fillet

[chrisplyler]

I don't know why you would expect a 0.125 fillet to work around the bottom of a hole that has a 0.1135 radius and is 0.070 deep.

If you want the shape that plunging a 1/4" ball end mill will produce, either set the hole up right such that a 0.125 fillet will work properly, or just plunge a 1/4" ball end mill into what you've already got like this....

[sportbikeryder]

 It isn't unusual to be able to fillet items that aren't extremely easy to calculate in many CAD packages. 

The first model posted worked fine, while the second did not. I understand there are a variety of methods that can create the same end geometry, what I don't understand is why the fillet worked in one model and not the other. I was surprised it functioned in Fusion on the first part as Fusion is not particularly "advanced" with offering a great deal of control on geometry creation. 

John

[sportbikeryder]

I added some additional geometry to the part in which the fillet "worked" and it will solve, however I was unable to achieve similar results in the second model. 

A small flat actually exists in the bottom of the "divot" created with the fillet. I hadn't noticed this was there, but makes it even more odd as to how the solver interprets the geometry and where it is placing the "center" for the fillet calculation.

The first model was originally created prior to a few of the recent updates that were pushed. I don't know why that would matter, just some additional info. 

[TrippyLighting]

I think any answer would require someone with intimate familiarity with how fillets are created by the software.

Filleting in itself is not exactly difficult, but it is the myriad of special and corner cases that a software like Fusion 360 would have to cover to make this an easily usable feature that works flawlessly in 95%+ of use cases.

[chrisplyler]

If there is a small flat in the center of the divot after filleting with 0.125, then that circle's diameter is greater than 0.250 of course. In your second part, I inspected/measured the diameter and it is LESS than 0.250, so of course the fillet would intersect itself around the center.

[sportbikeryder]

That's one of the odd things...the circle is not greater than .250"...both parts have the same sized divot. 

Its just weird that one part solves and the other doesn't. I thought maybe I was just missing something obvious. 

[chrisplyler]

Well, they aren't 0.250 anyway.

The second part has a lower wall height around a portion of the divot. Perhaps that wall height is not enough to support a 0.125r fillet at the bottom of it. Do the round divot, make the fillet, and THEN do the larger recess on the side of it. Worth a try anyway.