[ODE] Failed capsule-box collision

Jon Watte (ODE) hplus-ode at mindcontrol.org
Tue Jan 9 09:43:11 MST 2007

This sounds like you're testing something different from what you think 
you're testing (or rendering something different from what you think 
you're rendering). Box/Capsule works very well and is fairly robust in 
ODE (and has been for a long time).

Assuming that you've gotten ODE to work well previously, and that, say, 
sphere/box works fine, then I would look at the orientation of the 
created capsule (as the default axis might be different from what you 
expect). I would also look at how you extract the data for rendering 
(remeber: ODE quaternions put W first), and make sure the correct 
contact joints are created (connecting the box with the capsule) and 
cleared each frame.


/ h+

Tim Sorrells wrote:
> I’m trying to track down an oddity between a capsule and a box I’m 
> colliding in ODE. The box is a stationary geom with no body, and I 
> spawn a capsule above it and let gravity do it’s thing. When the 
> capsule collides it reports a collision took place, however visually 
> nothing seems to change with the falling capsule. It continues to fall 
> right through the box. The values I’m getting in the dContactGeom are 
> strange as well. The values are always identical all the way through 
> the box: depth 0.001684, normal (1.0 0.0 0.0). This is odd since the 
> capsule falls through the box so you’d expect the depth to be 
> increasing. It’s also hitting the box from above, so you’d expect the 
> normal to be positive Z instead of X.
> I tried swapping the box out with a capsule and the collision works 
> correctly with the falling capsule bouncing down the side. Any ideas?
> ------------------------------------------------------------------------
> _______________________________________________
> ODE mailing list
> ODE at ode.org
> http://mooshika.org/mailman/listinfo/ode

More information about the ODE mailing list