[ODE] [Newbie] Problems with torque and inertia, please help me

[Umberto]

Hi all,

I'm realizing the simulator of a robotic arm, with links connected by hinge 
joints.
To move the arm I apply a torque to each joint.
The problem is that torques don't move the arm like they should, unless 
they are extremely high (about 10^10 Nm)
So I simulated a simple inverse pendulum, but I still had the same problems.

Suppose I want to rotate a body around its x axis, and the body is linked 
to the world with a hinge joint, with joint axis along body x direction.
The body has inertia Jxx around x axis, and it's initially still.
If I apply a torque T to the hinge joint, for a time = TIMESTEP, then, 
after the timestep, the body's acceleration alpha_x is:

alpha_x = T / Jxx

and its angular velocity omega_x is:

omega_x = alpha * TIMESTEP

But if I apply that to the inverse pendulum (or whatever), it doesn't work!!!

The mass of the pendulum is 0.7853981634 Kg

and its inertia matrix is
0.0659407208        0.0000000000        0.0000000000
0.0000000000        0.0659407208        0.0000000000
0.0000000000        0.0000000000        0.0009817477

If I apply a torque of 1.0 Nm (Newton per metre) around its x axis, after 1 
TIMESTEP, the angular velocity should be 0.0345 rad/s.
But the angular velocity I get is 10^ -10 !!!! And I get the same result 
even if I change the mass of the pendulum !!
I made several tests, and I always got this:
angular speed = 10^ -10 * torque !!!!! Why????

Thanks in advance


Umberto Battista
University of Genova, Italy
DIMEC - Laboratory of Design and Measurement for Automation and Robotics