Howdy,

I've stumbled upon an issue I've spent way more time than I should probably have before asking for help.
The first attached image shows a simple setup - a three bones chain with a simple IK solver. A control null is placed at the “base” of the last bone. THis last bone is constrained (position only) to the said ctrl null.

I've reproduced this setup in Softimage as well and the following gifs shows the fundamental difference in the way the bone chain behaves, with some big implications in the way the rig of a quadruped leg I'm working on, should be created.
The fact that in XSI, parented bones will not part ways no matter what, has been very helpful in various circumstances. Maybe detrimental in others, but I wouldn't know about those.

In the attached hip file all you have to do is to grab the tip ctrl and rotate it on its X axis you'll notice the chain break. No matter what I try, like keeping the bones in the same hierarchy, something gives: if the chain will not brake, some bone jitter upon rotation will occur; or I'll have more bones than I want affected by a twist affector.

I'm hopeful there's a solution that doesn't involve a compromise of sorts.

Any kind of help is appreciated.

Here's the issue that as I described above as some sort of bone jitter.

The gif actually shows miccetro's leg from here [www.sidefx.com] in order to dispel any suspicion that it might be something that I'm doing wrong/differently.

I have zoomed in on the leg_toe_bone and slowly increased the Smart Footroll value. The twitching of the yellow bone is not due to the gif's low fram_rate. That shouldn't matter for stationary objects and besides, anyone can download that file and check it out.

What is going on here?


edit:
Hoping to save someone's time in the future: apparently increasing the Tracking Threshold Factor accuracy (by lowering value) will get rid of the twitching.
p.s. Changed the title to fare better in searches.

And the leg rig itself for anyone interested.
Ctrls not zeroed out.