Problem with positioning of components when creating a static joint

[davecorsello]

When I create a static joint between two joint origins, one on an imported component and the other on an existing component in the current design, things usually behave as expected--the existing component (and the entire existing design) remains stationary, and the imported component is repositioned to mate with it.  But sometimes, the imported component remains stationary and the entire existing design is repositioned to mate with it.  Grounding the existing design has no effect.  This seems to happen most often (perhaps always) when the imported component is an uploaded step file.  Is there anything I can do to ensure that the expected behavior occurs?

[sutherland-]

Hi @davecorsello 

Can you share the files that are acting weird?

Side note: please remember that the order of how you select the joining components matters, so the first component that you select will always mate to the second one.

[davecorsello]

Hi @sutherland- thanks for your response. Yes, I'm aware that the order of selection is important.  The component I'm inserting is comprised of sub-components, four of which are uploaded manufacturers' step files.  I tried creating a new, empty, local component in the current design and importing the sub-components, and this caused no unexpected behavior.  So it seems the problem is not with the uploaded sub-components in this case, but in the sub-assembly I'm importing.  But this problem happens randomly, and this case doesn't necessarily represent all the conditions under which it occurs.  I'll try uploading the files.

[jeff_strater]

"Grounding the existing design has no effect"

What, exactly, are you grounding?  In Fusion, grounding a sub-assembly will not ground all the child components of that sub-assembly.  Instead, assuming that the sub-assembly is fully connected with joints, try grounding one of the components at the bottom of the hierarchy (e.g. by selecting in the graphics area, not the browser).  That should force the existing design to stay put.

"Yes, I'm aware that the order of selection is important"

That is only partially true.  Order of selection is more a hint to the solver than a hard requirement.  If you have components A and B, and both are ungrounded, and you select A first in a rigid joint, then B, yes, A will mostly move to B.  However, if A is grounded, B will always move to A, regardless of the order you pick them.

[davecorsello]

Hi @jeff_strater,

Thanks for your suggestions.  I rarely ground components and I rarely have this issue.  In this particular design, all was working fine with no grounding until I inserted the sub-assembly in question.  Grounding the housing to which all sub-components are joined made no difference.

While preparing a pared-down example to upload to this thread, I encountered positioning issues while  building the sub-assembly.  I inserted the main component of the sub-assembly and then inserted a manufacturer's .step file.  While joining the mfr.'s component to the main component, the main component moved.  In this case, grounding the main component, inserting the sub-component, creating the joint and ungrounding the main component solved the issue.  I was then able to insert the sub-assembly into a simplified main design and join it to the ungrounded housing with no positioning difficulties.

I worked around the problem in my working design by building the sub-assembly locally within the main design.  This worked without grounding the housing despite using the same mfr.'s .step file that caused the problem I described in the previous paragraph..

In summary, many third-party .step files join to other components without causing positioning problems.  But some seem to be grounded by default with no way to turn off grounding.  Grounding the component to which a problematic third-party .step files is joined fixes the issue sometimes, but not always.  And it seems that a mfr.'s step file behaves differently with respect to this issue in different main designs.  This would take some time to document.

Dave

[jeff_strater]

Thanks for the explanation.  If you have a simplified example that shows a problem, I would be happy to take a look.

One point, though, that surprises me:  "But some (third-party STEP files) seem to be grounded by default with no way to turn off grounding".  If you run across such a STEP file, I would be interested.  My understanding of the Fusion translation process is that no joints or grounding is ever applied when translating a neutral file format such as STEP.  And, ground, especially, should not be happening if you are inserting any sub-assembly.  Ground in Fusion is stripped when inserting a design into another design.

My usual approach for imported sub-assemblies is:  Import them, open them in a separate tab, and fully "Joint" that sub-assembly.  If the entire thing is rigid, I would just use Rigid Group, and select all components.  I usually test this by dragging in that sub-assembly.  If the entire design moves, and no stray components are left around, you are good.  Then, save that design, open the target design, insert the sub-assembly, and just create whatever joints are needed - a single Rigid Joint from one component in the sub-assembly to a component in the target, if the sub-assembly is rigid, and wants to be rigidly connected to the target.

[davecorsello]

@jeff_strater,

Thanks again for your latest explanation of how this is designed to work.

I said, "But some [third-party .step files] seem to be grounded by default with no way to turn off grounding."  It would have been more accurate to say that some behave as if they are grounded under a set of conditions that I haven't yet identified.

Very strange.  I just tried three times to create a simplified design to demonstrate the problem, but now I can't reproduce the problem.  The best I can do is send you the third-party step file that seems to have originally caused a glitch.

Dave

[davebYYPCU]

The top level of the Fusion Browser is always grounded,

if you Insert with top level, you can’t move it.

If it inserts as a new Component, it is never locked unless the user does it, as mentioned.

Might help....

[jeff_strater · Accepted solution]

Thanks for the STEP file, @davecorsello .  I did not see a problem with this STEP file.  Note that the file itself contains a sub-component, so it defines a small sub-assembly.  The video below shows:

1. the translated STEP file, and dragging the switch model within this sub-component (I turn on the origin to show that the move is moving the component with respect to the root origin).
2. inserting this component into a top-level assembly.  Note that this results in 3 levels of hierarchy - the root of the top level, the sub-assembly, and a component (the switch) within that sub-assembly.
3. Note that nothing is grounded at this point (I can freely drag the switch, which actually moves within the sub-component)
4. Grounding the sub-assembly root does not prevent the child (switch) component from moving.
5. However, grounding the switch component itself does prevent that component from moving when dragged