Design suddenly having severe issues when Computing All

[fuzzyepictnt]

Hello,

I've been working on a design for a custom vex robot for my team. At the bottom of this post is a link to the file. Please excuse the excess of parameters, they're for easy access since a lot of aspects of this design need to be changed over time the more we develop. This is also why some components have been removed.

To preface, on Monday, I wasn't able to mirror the right side of the drive as it would repeatedly crash when trying to do so, due to an overload of some sorts. In case this is relevant, I attached what would happen when trying to mirror, multiple crash reports were sent at the time. The same thing happened when trying to create a simulation model, but this issue isn't a problem to me anymore. I ended up mirroring each subcomponent individually, and although it wasn't the prettiest workflow in the timeline, compute all was still working.

The issue: Today, I noticed three errors in the middle of the timeline that weren't there before, saying "Position calculation failed due to initialization error". I tried doing compute all and suddenly this design, which Wednesday and before computed fine, has the "Position calculation failed due to initialization error" as warnings and errors all throughout the timeline. The errors start at Rigid4.

When I edit a joint with this error/warning, it appears to be a perfectly valid action:

Editing Rigid4

When I try reselecting the edges in the joint however, it says that there is already a relationship between the two components with an existing rigid group, in this case saying RigidGroup1. RigidGroup1 is the next Rigid Group past the history marker in this screenshot: 

Joint conflict warning when reselecting the edges

RigidGroup1 has not been created at this point and is now conflicting. Deleting RigidGroup1 and moving the history marker back to where it is in the screenshot ends up causing further errors even earlier in the timeline. Why are these joints conflicting with something that does not yet exist? This was not an issue before. 

Furthermore, when the history marker is moved earlier in the timeline, future joints are still showing in the window and in the browser. For example, in the following screenshot I have the history marker placed before Rigid4 was made, however even joint Rigid45 is showing, both in the component tree and in the window. (far before Rigid45 is created in the timeline)

Future joints still showing in this screenshot

Other joints under components that also do not exist at that point in the timeline are shown in the window, such as the three rotate joints toward the left of the window.

I think that something may have happened past a recent update. A school computer I recently used was saying that I needed to install a later version of Fusion 360 to open the file, however Fusion 360 wasn't prompting to install an update as it usually would if an update was pending. I had to download the installer and install a new copy of F360 to be able to open the file, I am unsure if that is relevant but I have a feeling that these issues have stemmed from a recent update, since those school computers update every Sunday. 

I am very concerned as we have a deadline for this in two weeks. What happened? How can I go about resolving this without restarting the design? 

Thank you,

Mae

The design file itself is too large to attach (85mb), I uploaded it here to Google Drive: 15.mk2 v52.f3z - Google Drive 

[sunil.v.b]

Thanks @fuzzyepictnt for reporting this issue.

Created an internal ticket FUS-104250 to track this issue and also assinged this to respective development team to take a look.

Thanks

Sunil

Fusion360 dev team

[euclidmandd]

Has there been any progress on this issue?  I also have a large file that's been very unstable lately with the same seeming random "position calculation" errors.  I could share it if that might help root out the cause.

[sunil.v.b]

Hi @euclidmandd Unfortunately there is no update to share on this issue. We haven't been able to figure out the root cause yet. Thank you for willing to help us here.

Definitely, you may share the model with us, we can take a look and see. In addition, could you also mention the steps you are trying out, is it compute all or something else? Also do you recollect the state when things were working fine for you ? 

Thanks

Sunil

[TrippyLighting]

Yes, please share the design!

[euclidmandd]

Hi Sunil and Peter.  It's a somewhat proprietary design, so I'll send a link through a private message.  Thanks in advance for your help.

[TrippyLighting]

I suggest you place this into a new thread.

So far I've only looked at the last design in your PM.

1. Joints are fine.
2. Compute all does not return any warnings or errors, but some of the half-U-profiles turn into solid block. That should not be happening and is definitely a bug! 

Edit1:

I am looking through your second design at the moment. Here's a little tip:

If you want to de-feature geometry, for example all the hardware/screws and want to remove the threads, instead of adding that feature to the timeline, edit the base feature of that component and simplify there. That will keep you from polluting your timeline with all these little delete face features.

Edit2:

@sunil.v.b I would take a look at these designs. To me there's buggy behavior with the as-built joints (and not only). For example, I just rolled the timeline back just after this joint and it turned from red to OK. 

[sunil.v.b]

Thanks @TrippyLighting for looking into the designs shared by @euclidmandd

There are 2 issues here

1. the compute ALL showing by the extra boxes and edit boundary fill bringing it back to desired result. I agree this is a bug and I have created internal ticket FUS-114764 to track this issue.

2. the position calculation errors. I have created internal ticket FUS-114765 to track this. 

[euclidmandd]

Thanks for looking into this.

In case it helps someone else having the same issues, I'm starting to suspect that the issues in both files are related to copy-pasting bodies into a new component (or using boundary fill to do the same thing).  I think I only used that workflow once in the big problem file, and after removing it, the errors all cleared.  When I get time I'm going do a more permanent re-work that doesn't involve copy-past bodies and see if it holds up long term.

@TrippyLighting,

Thanks for the tip.  I do directly edit the imported components sometimes.  I think in this case I went the editable route just in case I wanted the features back in the future.  But yeah, I agree it make the timeline a little messy.