Why can't I shell this body?

[lemelman]

It was almost certainly the wrong way to create this Y-connector, but it appeared to be working OK until I tried to shell it. I just can't shell it. It gives me the attached error message.

What am I doing wrong?

[TrippyLighting · Accepted solution]

And that would not be obvious at first sight ?

What geometry would you expect for these features to create when shelled ?

'

These threads present better ways to create Y-Branches:

https://forums.autodesk.com/t5/fusion-360-design-validate/y-branches-when-not-using-pipe-command/m-p...

https://forums.autodesk.com/t5/fusion-360-design-validate/creating-a-y-piece-with-tapered-ends/m-p/7...

[chrisplyler]

Sloppy modeling.

I'm attaching my file at the bottom of this post. Open it. Roll the timeline back to the beginning and examine each step in order until you've reached the end. Let me know if you don't understand any particular step.

[chrisplyler]

---

@TrippyLighting wrote:

And that would not be obvious at first sight ?

What geometry would you expect for these features to create when shelled ?

---

I would expect it to create SOME inner wall, perhaps with options on how to deal with unfriendly angles and stuff. You know as well as I do that the Shell functionality is weak.

[chrisplyler]

And here is a version that is shelled.

You might have noticed in my first file that I used the Hollow option on the three pipes, along with outside fillets and inside fillets.

In this second version, I just used solid pipes, outside fillets only, and then Shell at the end.

[lemelman]

Thanks for identifying the problem. I didn't notice the surface irregularities,  but even if I had it would never have occurred to me that it would prevent the shelling function. Actually, now that I know, I'm rather surprised that it did. I would have assumed that the irregularities would have propagated, normal to the surface, to create a corresponding  interior surface. 

In the end I used a completely different method and have produced the required Y-connector. 

[TrippyLighting]

As @chrisplyler has already said, the shell function is pretty weak in Fusion 360. It creates an error, but then does not tell you hat the problematic area is. it also does not offer any options that would allow you to make a decision on how to deal with this.

If the method you used is not covered in any the the other threads please share!

[chrisplyler]

---

@lemelman wrote:

I didn't notice the surface irregularities,  but even if I had it would never have occurred to me that it would prevent the shelling function. Actually, now that I know, I'm rather surprised that it did. I would have assumed that the irregularities would have propagated, normal to the surface, to create a corresponding  interior surface.

---

Yeah, that's kind of what it does. The problem occurs when the shell thickness you specify results in that propagation folding in on itself.

Your original wye WILL shell as long as you make the thickness low enough. But beyond a certain thickness, those little vertical edges cause trouble. Imagine offsetting a V shape inwards smaller and smaller. Well, at some point your V is going to close up. Fusion doesn't know what to do with the offset lines that have become zero or negative length. Same sort of thing happens with the Shell function.

[lemelman]

Apart from the shelling not working, there was another problem with my first attempt. Although I defined parameters to control the sizes of the legs I also defined the angle between them, but it didn't work.

In the new method I drew 3 construction lines, intersecting at the origin, that defined the centrelines of each leg. For one leg I created a sketch of the profile that, when revolved about it's construction line, would create one hollow leg. This profile was then copied to one of the other construction lines and then mirrored to the third. Each profile was then revolved to create 3 bodies which were then combined to create the basic Y-connector. The unwanted interior overlaps were then removed and the required fillets applied.

The resulting model looks OK to me, and all the parameters work.

The timeline in the attached f3d shows the actual steps.