Offsets need to be determined again after restart

[TimWindecker]

Hi,

sometimes we have to determine the offsets again after the Gello was turned off and on again, because otherwise some of the joint values are completely off. Our theory is that this happens when the Gello arm is turned on at a different position than the time before. This could be caused by the Dynamixel motors since when they are turned on at a value of about 359° and are rotated a couple degrees clockwise (~ 5°), they output values > 360°. While if the Gello is restarted at this position, it returns a joint value of about 4°.

Is this the normal behavior or do you have an idea how it can be mitigated?