[ODE] Not A Number problems

[j]

Hi Gopi,

I did notice that and your solution works like a charm !
In my case, strict realism is not an issue and by clamping
rotations between -pi and pi I keep reasonable speed
even with 200+ objects without them popping into #NaN.

Big Thanks,

J
----- Original Message ----- 
From: "Gopi Prashanth" <gprashanth at Heavy-Iron.com>
To: "'j'" <ode at dynamica.org>; <ode at q12.org>
Sent: Monday, April 25, 2005 10:44 PM
Subject: RE: [ODE] Not A Number problems


> Here is what I did with a similar problem, if you put a conditional
> breakpoint in the simulation step and check for nan you will see that
always
> rotational speed will go to nan long before linear velocity. This was the
> cause of my woes. To fix it I damped rotational speed and capped the
> rotational speed as well. This fixed my problem. Hope it helps,
>
> Cheers,
> Gopi
>