Thicken cut of a helical surface makes faces segmented

[mroek]

As a learning exercise for myself, I tried to recreate something I did in Onshape before I moving to Fusion, and that was to do a helical trisection of a cube (as shown here: http://www.thingiverse.com/thing:629887 (by another user)).

I was able to get it done in Fusion, but since the goal was to be able to 3D-print it, the final step(s) was to do a thicken cut of the helical surfaces to remove material to get some tolerance. However, Fusion then splits the cut faces into segments, as shown here:

Before the thicken cut (right after the split), the faces of the three bodies are single, non-segmented faces. While this wouldn't really be an issue if I were to print these, I don't understand why Fusion can't create proper faces after the cut? Is this some limitation of the modeling kernel in Fusion? My Onshape model has proper, non-segmented faces after the thicken cut.

I've attached the design file in case anyone wants to look at it. Try rolling the history back two steps (to before the thicken cuts), and you'll see what I mean.

[michallach81 · Accepted solution]

Hi Øyvind, you should try another method. Your sweep on helix creates a very bad inner edge (should be a straight line). If you would provide line first, that could solve the issue:

[mroek]

@michallach81

I had seen the lofting method previously, but I find it really difficult to imagine the result. Visualizing a loft between a straight line and a helix isn't easy to wrap my head around.

The method of sweeping a line is much more intuitive, but in Fusion it doesn't work the same as in Onshape, because in Fusion it also needs a guide rail to create the surface (without it, it creates a very different surface). And as you showed, that result may not be all that good either. I've never really used the curvature analysis tool before, but it was clearly very useful, so thanks for showcasing it.  🙂

I reworked my example using the loft method (lofted one helical surface, then made two copies of it at 120 degrees apart, and then stitched two of them to be able to make the first split), and I got a proper end result.

So thanks for looking at it, I'm happy that I learned something!  🙂

[michallach81]

One more little improvement to my method. Include 3D do a bit of a mess with a helix. Instead of sketch feature, you can pick an edge of the coil. Since it can't be an edge of solid, you have to remove some faces while in Patch workspace. This time result is clean:

[mroek]

Very nice, good trick!

However, I'm a bit disappointed that "Include 3D" doesn't give the exact same result. Using tricks like this shouldn't really be necessary.

[TrippyLighting]

---

@mroek wrote:

Very nice, good trick!

However, I'm a bit disappointed that "Include 3D" doesn't give the exact same result. Using tricks like this shouldn't really be necessary.

---

Disappointed ? No, it's a disgrace for a modern parametric CAD  that we have to go through these convoluted workflows to create a helical structures. It would not even have occurred to me to check the result of "include 3d" with a curvature comb.

Projecting a 3D sketch is the method Iv'e seen mostly used here on the Forum and have suggested myself a time or two. 

Why would we first have to create 3D geometry and then have to delete 2/3 of it to create the base for the helix just to delete the rest of that geometry later.

I really appreciate @michallach81 (yet again) for his contribution and for figuring this one out, however, this should not even be necessary. 

[michallach81]

Something like two years ago I've said that first thing that needs to be rebuilt is sketching. The problem is that everything else depends on sketches. I don't want to say that sketches are useless. There are just a couple of problems, but if we would like to see more advanced surfacing we need more tools in sketching. Again I don't want to say we can't create complex elaborated surfaces with current set, just daydreaming, class A surfacing. Because I understand (and share) reasons for the current course of development, I've learned to "debug" modeling and solving problems immediately after they occur. What I don't like is that learning resources don't mention limitations. Users especially new one, can't predict tools behavior.