Rigid Body collision problems

[FloydBishop]

Just out of curiosity, how are you calculating the collision? Verts, bounding box, etc? Also, what kind of forces are you using? Any gravity or drag?

[FloydBishop]

OK. I re-read your post and you've already answered some of my questions :)

What scale are you building in? Sometimes an odd size can cause some undesired results.

[colt]

The tracks are 0.1 units in diameter, the car about 1 unit wide. I build in meter, but I tried changing to cm, and the car jumped at the same position. The car weights about 50 units.
The only force I use is a standard gravity, without modifications.
I tried all three apply force settings: Bound., verts, center of mass ... no change in behaviour.

I'm really running out of options. Do you think re-building it bigger or smaller would help? What would be more promising, bigger or smaller? Or as a last resort, trying to switch to 6.0 ... did Alias bugfix the dynamics?

Anyway, thanks for your thoughts. :)


Edit: Ok, several hours later. I tried rebuilding in 100x scale, same problems. And I did the rebuild in Maya 6.0. Now I'm out of options. :cry:

Scenefile_288k (http://www.uni-weimar.de/~rehberg/challenge/rollerco.mb)
Heres the rebuild in Maya 6.0 format. I'd be happy, if someone would get something better than this. mov_650k (http://www.uni-weimar.de/~rehberg/challenge/crash04.mov)

[king21]

Hi!
Just one thing , be carefull of normals. For me works, that normals of one object ar in other direction as the rail


cheers

[Mr Majestic]

Ive done something similar before, maybe yours is ser up different or maybe im just really tired and not reading or thinking of this correctly. But on the track did u just make the rails by having a curve and making a cylinder extrude to that curve? You said its polys, maybe is catching on divisions of that rail, i wonder if a high smooth wouldnt help you out, then again I could be totally wrong :) just tossing out ideas.

[alexgk]

I tried smoothing my track, but this hasn't worked so far. I had better success (though not great) with lofting curves and making that the passive rigid body. Surprised it worked better than sliding on a pair of simple poly cylinders.

Are nurbs better to work with for collision surfaces?

[caulfield]

I'm having some issues with this at the moment aswell - I think rigid body solvers are great up to a point, but the longer your shot, the more likely it is that your shit will hit the fan.

I'm animating a giant version of Spike Jonze sliding face first down a glacier, and yes the same thing happens. Sadly, rigid body solving is indespensible to my solution, but why yours? Couldn't you achieve the same look with animated points on the curve, and four nice little springs? Rig the thing like you would a car on four wheels.

You could add some noise to the Y offset of yours points to simulate the vibrations and this would cause the springs to bounce around.

You could always recreate the hobbit genome, but Peter Jackson was smart. He just got actors to talk like geezers and wear big feet and stupid clothes.

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.