Bug when reorienting Z' in ball joint to LCS z?

[Anonymous]

Creating a ball joint for an articulated humanoid is trickier than with regular solids. Ball joints in Fusion are mathematically modelled in a non-intuitive way, with an Euler-relative arrangement, but apparently for good reasons as explained here.

In reference to the essential information provided in the above thread "ball joint mysteries”, when following the workflows recommended there I am *sometimes* experiencing what could be a bug that also destroys some data, namely the orientation of the local coordinate system (LCS) for a component to a ball joint.

Goal: I’m attempting to define a ball joint orientation by defining a joint origin to align normal to z from joint component 1's LCS (local coordinate system).

This workflow fails in a couple of ways in the case illustrated in the attached video:

1/ Fusion's joint command fails to accept as an argument a joint origin I’ve created where I have "reoriented" the Z' angle of that joint origin.

2/ Fusion erroneously re-aligns the implied joint face from the joint command of joint component 1 as I build the joint, losing its original alignment entirely.

3/ Fusion moves joint compoent 1 as the joint is built to an uncommanded location.

Result: Creating a joint with the desired Z'XZ" orientations is impossible, and while there are some kludgy workarounds, none saisifies with a ball joint aligned precisely as desired.

Steps to repeat:
Create joint origin in joint component 2
Reorient Z' for this joint origin to become normal to z from joint component 1's LCS. Works OK!
Show joint component 1, press j key to start create joint command
Hide joint component 2
Select location for joint in joint component 1, works OK!
Show joint component 2

Try to select joint origin in joint component 2 as second argument for joint command. Fail! Can't select it.

Instead fusion offers me a joint origin aligned normal to joint component 2's LCS z. Not what I want.

Click to create joint anyway. Joint is created but its Z' is located normal to joint component 2's z, not joint component 1's z.

Attempt to use "edit joint->motion->pitch->z axis command to reorient joint. Fail 2!

See video 1

Discover that somehow a component orientation and joint orientation joint component 1 have been shifted to

...and other downsteram problems... (See video pt2)

I can share the subject file with AD staff but not publicly as it is copyrighted.

[Anonymous]

This is the second screen cast

[Anonymous]

Video from initial post cut off, here's another...

[Anonymous]

same error this time with as-built

[kgrunawalt]

Hi,

I'm sorry you've had such a frustrating time with this. Thanks for the great documentation. If you could share the assembly with me (share the project if possible), I'll take a look a the ball joint issue. Is this an imported model? If so, was the left leg mirrored from the right? I'm just taking a guess that there might be a mirrored transformation matrix that could cause Fusion issues.

I think I see three issues in your videos:

1. Difficult selecting joint origin? I'm not certain that you created two aligned joint origins and selected them in first joint attempt. If they weren't aligned, a move is expected. If they are aligned but you could not select both, perhaps geometry was interfering (try click-hold to get list of selectables).
2. Setting home position didn't work
3. as-built axes jumped on OK

The as-built joint should be the way to go, but they should not behave as you observed.

K

[Anonymous]

@kgrunawalt What's your email address please?

[kgrunawalt]

katrin.grunawalt@autodesk.com

[kgrunawalt · Accepted solution]

I have found two issues to log and have a workaround. Issue 1: you cannot select a joint origin when defining as-built joint position. The command should allow this. Issue 2: the as-built joint orientation changes on OK in some cases (duplicated in simpler example). I'll log these.

The workaround is to create two aligned joint origins, one in the left thigh and one in the left pelvis sphere. The orientation should reflect the thigh origin orientation. This requires reorienting the joint origin in the pelvis to use the thigh z and x axes. It may require rotating the joint origin in the thigh by 180. Below is an image showing these joint origins. Once these are created. Create a ball joint between them (not as-built). There should be no movement in the leg and the joint should have the right orientation. Note that the thigh origins are not perfectly symmetrical.

I think this results in what you want -- or close enough that tweaks to the joint or joint origins can get there. Sorry for the hassle again. Once those issues are addressed, it should be much simpler. Remember to delete the joint or joints that already exist between the left thigh and pelvis.

Let me know if this helps.