I've been looking at this for two days and I don't understand what's up here. I've installed the TSM rig numerous times with no problems but for some reason the right elbow won't deform correctly this time. I've checked the CP assignments, weighting, even deleted the fan bones I'd put in the elbows, but the problem persists.

I'm posting the model and a screen shot of both elbows bent at 60 degrees for comparison.

Without saying I've solved it, here are some clues. I'm guessing this got run through TSMRigger twice.

Click to view attachment


As an aside, remember to save your model after you run TSMrigger, then close A:M and reopen the model to get rid of most of those "String" errors.

As Robert said, you had duplicate relationships. I went through all TSM2 relationships and deleted all the "error loading string" relationships. Here' a fixed model.

Wow, thank you gents. I can assure you it didn't get run through TSM rigger twice! But I did see some strange behavior when I ran Flipper, now that I think about it. I should have documented it at the time, but it made me wonder at the time if v15.0j had some compatibility probs with TSM.

As I recall when I ran Flipper some bones came in with their sizes zeroed out (ie, occupying a point in space) and a text string in their size boxes, like NNaN, something like that with several N's in it. I typed in values corresponding to the original side bone and everything seemed fine, so I moved on. Rigger ran fine. I can't explain the constraint to the eyelid!!

But by coincidence Robcat I only today started playing with the new filtering feature but I didn't know what I was doing so let it go. Your movie is a big help on that front too.

And Mark, can I ask how many duplicate sets of constraints you found?

All TSM2 relationships. I don't know exactly how many constraints, I deleted the "error loading string" relationships, which hold the constraints.

Are you doing this on Windows or Mac?

Gerry, now that I read your post, I can say I ran into the exact same problem. Unfortunately I can't go back to recreate it. But I had to go in and manually clean up the illegal values. I also ran into nulls getting scaled to huge sizes and a a whole host of issues, that at the time I chalked up to user error.

Was this with 15.0j?

P.S. Robcat, this was on a pc.

I haven't had this problem happen to me, but this is another reason to do your fan-boning and CP weighting and save that before you run Rigger, then if Rigger got an odd result you could easily go back to the pre-Rigger state and look for clues.