[ODE] Balance of torque

[Gib Bogle]

I think I can now answer my own question (that was quick!).  It seems to 
be a matter of the timestep size.  The period of revolution in the 
current case is about 4 sec.  I thought that dt=0.01 would be small 
enough, but I find that if I reduce it to dt=0.001 the ratio of the 
torques becomes 40:39, and presumably reducing dt still further will 
further reduce the discrepancy.

Gib

Gib Bogle wrote:

> I'm simulating the motion of a heavy rope spinning about a vertical 
> axis.  The rope is represented as a number of rods connected by 
> ball-and-socket joints, and the two ends are connected to the vertical 
> axis by hinge joints.  The motion is driven by external applied forces 
> on the rods.  A constant rotational speed is maintained by suitable 
> application of a resisting torque (applied as a force in each 
> timestep) on one of the rods that is hinged to the axis (the other is 
> free).
>
> The behaviour looks fine, but there is a discrepancy in the torque 
> balance.  Since the rotational speed is held almost constant, the net 
> torque of the applied forces, averaged over one rotation, should match 
> the resisting torque, averaged in the same way.  (There is no air 
> drag.)  In fact these two average torques are roughly in the ratio 43:38.
>
> I'd like to know the explanation for this discrepancy, and also if 
> there is a way to reduce it.
> Thanks
> Gib
> _______________________________________________
> ODE mailing list
> ODE at q12.org
> http://q12.org/mailman/listinfo/ode
>
>