[ODE] question about euler angles and angular motor

[jon klein]

Hi all,
I'm having a strange problem with an angular motor.  I have
the motor set up to use euler angles, and I set the "twist"
(axis #2) to be a constant velocity.  The problem is that
although axis #2 is supposed to twist at a constant rate, it
turns out that its rate is dependent on the rotation of axis
#1 (which is the automatically computed axis of the motor).

So when the joint is relaxed on axis #1, the rate of twist
is just as expected -- but when the joint is bent on axis #2
(to an angle of about 1.4), the body twists over twice as
fast as it should.

I hope that explanation made sense -- has anybody encountered
this problem before?  Am I misunderstanding how this joint is
supposed to be working?

Thanks,

- jon klein