Shell type offsets circular face's snap point

[shtanko_andrey]

Dear Sirs,

I've encountered an issue when shell type seems to offset circular face's pivot point used for move/align. Namely, problem occurs in case of sharp+outside/inside shell (inside/outside affects which face exactly gets this issue).

Sample model and screens attached. I've also managed to reproduce it in new file, steps:

1. XY sketch, center diameter circle, aligned to origin;

2. Offset plane from XY,

2a. Sketch on offset plane, center diameter circle, aligned to origin;

3. XZ/YZ sketch, project sketches from p. 1 and 2, make vertical line from midpoint to midpoint of projections and spline on one side, connecting ends of projections;

4. Revolve profile drawn in p.3.

Either:

5. Shell top surface, sharp+outside shell.

6. Trying to select to move or align bottom face (from bottom view) shows that pivot point is off origin point. Aligning that pivot point to origin point causes whole body to move off origin axis.

OR

5. Shell top surface, sharp+inside shell.

6. Trying to select to move or align body's bottom upper face (from top view) shows that pivot point is off origin point. Aligning that pivot point to origin point causes whole body to move off origin axis.

* problem seems not to occur with round shell;

** problem seems to go away if I align side spline's tangent handle from p. 3 with bottom sketch projection from p. 1, smoothing transition between spline and bottom line; changing spline to straight line or 3-point arc also sees to be producing correct result, so problem seems to be in surface generation when splines are used.

[CGBenner]

Fusion gurus, do we not have any takers on this challenging post?  😉

[laughingcreek]

I can't reproduce this with your file.  can you attach a version where the align is incorrect?  

also-

using align in this way is generally not suggested.  you end up moving the body away from the sketch that created it.  ideally when you make changes of this nature you want to go as far back in the time line as possible instead of layering additional steps onto what you;ve done.  in this case that means just drawg the revolve profile where it was suppose to be to begin with.

also-

since this is a revolve, sketch bottom and sketch top, along with the offset plane are unnecessary.  you only need a dimensioned profile.

also-

the primitives such as sphere are not very well developed in fusion, so from a parametric viewpoint, it would be better to just revolve an arc

see attached

[shtanko_andrey]

---

@laughingcreek wrote:

also-

using align in this way is generally not suggested.  you end up moving the body away from the sketch that created it.  ideally when you make changes of this nature you want to go as far back in the time line as possible instead of layering additional steps onto what you;ve done.  in this case that means just drawg the revolve profile where it was suppose to be to begin with.

---

Body initially was where it was supposed to be (as initial sketch was started from origin point). External shell made body's outline larger and I decided that it's not neat and tried to move it to align with initial plane, that's when I noticed above mentioned strange stuff. Drawing profile with shelling in mind is possible, of course, but amended it every time when I decide to change wall thickness would be strange 🙂 Usually I don't use move/align at all, therefore I was quite surprised to see that things went off origin and central axis downstream after aligning.

---

also-

since this is a revolve, sketch bottom and sketch top, along with the offset plane are unnecessary.  you only need a dimensioned profile.

---

I know, those 2 sketches were not used. 

---

also-

the primitives such as sphere are not very well developed in fusion, so from a parametric viewpoint, it would be better to just revolve an arc

---

Err yes, it was not related to the case, actually, but thanks)

---

@laughingcreek wrote:

I can't reproduce this with your file.  can you attach a version where the align is incorrect?  

---

In file you've attached - cuckoo nest LC.f3d - it's already reproduced: just switch to bottom view zoom in to origin, and select move or align command, then aim cursor at body4 (the "pot) and you'll see that handle which will be used to move/align body is not aligned to origin on XY plane (about 0.5 mm to the side); at the same time from top view (i.e. aiming at bottom's upper surface) handle is at origin, as expected.

 

 

If you go back and switch shell type of body4 to inside (which will keep it perfectly aligned with sketch), then bottom surface handle will be at origin point, as expected, but from top view handle will be off origin, i.e. situation reverses.

If you switch shell type to "both" - both handles will be unaligned.

At the same time, rounded shell produced expected results (both handles aligned with origin point) both with inside and outside shell (which is why I think that it's not a design issue).

I've attached initial file's version with all distractions like sphere, etc, removed and a new one where it is reproduced from scratch.

I could only reproduce it when spline was used to draw outside wall, btw. And it seems that aligning spline's  handle in point where it connects to straight line fixes the issue. 

  

P.S. I've also noticed that I can click on surface itself, not on that handle, to select it from moving/aligning, which seems to work as expected, but in that case I can't understand what's that handle is there for and in any case why it's not in the center of surface. Seems like a bug, even though minor one.

[laughingcreek]

I agree.  align isn't working right here.  maybe someone from AD can take a look.  I've attached file from above with a sketch showing the misalignment after using align function. @Phil.E ?

[laughingcreek]

---

@shtanko_andrey wrote:...Drawing profile with shelling in mind is possible, of course, but amended it every time when I decide to change wall thickness would be strange 🙂

...

---

Stranger than bodies visually disconnected from their sketches?  six of one I suppose...

a vid for your viewing pleasure.

[Phil.E]

I don't see anything unexpected in your file. Can you tell me what to do/look for?

[laughingcreek]

before the align feature, the body is centered on the origin.  after the align (using the "center" snap) the body is not centered on the origin.  the last sketch shows how much the body is shifted.  

[Phil.E · Accepted solution]

Okay I'll send this to the team that does shell. It might have to do with the new shell types, and we just didn't see this yet. Thanks for all the information everyone.

For now, if you are working with shelled objects, perhaps try using a vector Move to keep the objects on axis. Align shouldn't shift like this but it doesn't offer a "vector" like the Move command does, therefore the move command is by default more accurate,  and it's parametric.

[shtanko_andrey · Accepted solution]

Selecting face itself instead of a handle seems to be working as intended despite shifted handle; perhaps it can be used as a workaround.