[Solved] Major Stability Issues (Coordinate Frame Error)

Game_Ender

15-04-2007 18:11:01

I am developing a simulator of a submersible. So far I have a very simple environment with a buoyancy plane at (0,0,0). A box of dimensions (0.2032, 0.2032, 0.6858) floating at the surface with mass 24.6. The addBuoyancForce function is passed 1000,0.03,0.03 for the first three arguments.

Now I have virtual thrusters (just points where I apply a local force to the box) on either side of the box. I have tried locating these anywhere from 0.16 to 1.0 from the center of the box, it does not affect behavior. When I apply exactly the same force (in the same direction) to them the box will move forward for 5-10 seconds and then start to build a side to side oscillation which eventually builds till the object is spinning abouts it center of mass.

Here is quick diagram (arrows are the forces):
__
^ | | ^
| | | |
--


You will notice, I am applying no torque to the object, but it still starts spinning. Any ideas?

Game_Ender

16-04-2007 00:44:22

So, I solved this, turns I was apply a "local force" with global coordinates. This explains the other behavior I saw (instability far away from the origin).