PDA

View Full Version : Sphere physics

mightyMo
08-16-2001, 04:20 PM
Althought I first thought it would be simple I've since discovered that doing correct sphere to sphere elastic collisions in 3d space isn't as easy as I thought. A lot more thought needed than the ray to triangle collisions I did for a game a while back.

I did a quick search but didn't find a good site. Can anyone point me in the right direction?

DFrey
08-16-2001, 05:08 PM
Actually, I would consider elastic sphere-sphere collisions to be among the simplest collsion cases. First to see if a collision has occured you obviously only have to compare the distance separating the centers to the sum of their radii. Now once you know a collision has occured, it is a simple matter to calculate the plane of collision. That is simply the plane that is tangent to both spheres and perpendicular to the line separating their centers. Now using that plane as a reference, just use the fact that the system's center of mass continues moving with the exact same linear motion after the collision as it had before the collision (assuming the spheres are in free motion before the collsion), and conserve energy and momentum and the final momentums will fall right out after doing a little (ok, a lot of simple) algebra.

[This message has been edited by DFrey (edited 08-16-2001).]

Michael Steinberg
08-16-2001, 06:10 PM
Ahhhh... DFrey, another question. The physics degree you got, did you get it studying computer science or what? I'm nearing my studying time, and I'm also interrested in that stuff. I've got 2 years left, but it's never bad to know it earlier.
Sorry for bothering you!

DFrey
08-16-2001, 06:39 PM
Actually my physics curriculum included only an introductory computer science course (basically a course in Pascal). I'm actually kicking myself; now wishing I had gone into computer science. As the job hunting has been very depressing with what I have. I'm actually considering not telling potential employers I have that physics degree. http://www.opengl.org/discussion_boards/ubb/tongue.gif

mightyMo
08-17-2001, 02:55 AM
I think I've had that grade school algebra http://www.opengl.org/discussion_boards/ubb/smile.gif(but even this algebra took me awhile being that I'm so rusty) figured out but I'm confused with rotations of the spheres. I guess my original question didn't really point to that but here in lies the problem. I want these colisions to realistically affect rotational velocities of the spheres.

And don't go regretting your physics degrees I went for computer engineering and my lack of physics often kicks me in the pants. You should so beable to get a job in scientific visulization if you know opengl and physics. I had an internship at Los Alamos National Labs last summer and we were looking for folks like you in on the viz team.