# Drivers & Relative Bone Rotations?

I've been trying to figure out how to create a driver that relies on relative bone rotations only. Is this possible?

The concept is very simple. A parent bone attempts to rotate itself based on its child's relative rotation. So imagine a hand twisting, and a forearm twisting along with it. The problem I'm running into is that the hand is a child of the forearm, so the forearm (while observing the hand rotation) ends up reading its own rotation changes, which causes feedback. But this doesn't make sense if the rotation I'm reading is relative.

I was under the impression that "local space" rotations would mean relative rotations? This means the parent rotations have no influence on the result, right? But if that was the case, why would the result become unstable when a parent bone attempts to rotate based on a child bone?

Am I misunderstanding the way local space works? If so, is there a simple way to obtain "relative rotations" of bones for drivers? A method to use that would allow a parent to change itself based on a child?

Edit: After some testing, I'm pretty sure the rotations are indeed relative. So I'm guessing there is just some type of glitch or error that occurs when trying to use child transforms in parent drivers when IK is involved. Can anyone confirm this? Any way around it?

Edit 2: Just to clear everything up, I have a hand (child most), a wrist (parent of the hand, and it only twists), an elbow (parent of the wrist), and an upper arm or shoulder (parent of the elbow). I want to make the wrist automatically twist when the hand is twisted (the axis that runs down the bone) so that the wrist twisting runs down the entire forearm.

When I attach a driver (to the wrist) to make this happen, the mesh warps at the wrist area and stretches all over, as if the bone transform is exploding, even while the hand twist value is at zero. If I temporarily disable the IK chain applied to the hand (the little eyeball toggle of the constraint), the problem looks the same. If I delete the IK chain applied to the hand, the problem goes away and looks perfect.

Does anyone know why or how to deal with this? I've tried tweaking the IK rotation limits of related bones, but nothing seems to help. My IK bones are separate from my real bones. The real bones simply follow the rotations of the IK bones. But I've tried applying the IK directly to them and it doesn't seem to change anything.

Thanks for any clues!

• Do you use the IK chain to deform the mesh directly? The wrist bone should be parented to the elbow so that you can read the rotation from there and then use that to drive the individual deforming bones. You can also set this up using a bendy-bone which have this functionality built in. Commented Aug 8, 2019 at 9:44
• Sort of, but not exactly. I have copies of the real bones that I use for IK, and the real bones follow the rotations of the copies when IK is enabled. But I don't think this influences the glitch I'm seeing. The elbow (forearm 1) is still the parent of the wrist (forearm 2). And the wrist can only rotate in one direction - twisting. I wanted to make the wrist automatically twist when the hand twists. That is my goal. But apparently, as long as an IK chain is involved, the skin-mesh warps all over the place when I try to use a driver to do this. If I delete the IK-chain, it works fine. Commented Aug 8, 2019 at 11:52
• I don't get it... care to share? Commented Aug 8, 2019 at 12:09