In this .blend http://regress.homelinux.org/~grahame/files/rails_balls.blend the ball rolls down the rails pretty smoothly until it reaches the point where the rails become wider than the ball. At this point the ball either sticks to the rails, and doen’s drop, or rolls a short way up the underside of the rails depending on the mass of the ball.
I’m using the latest cvs version of blender (21/06/06)
Alright, I took a look at your file in 2.42 RC1 and it would kinda work… kinda. It still had the problem where the sphere would get stuck under the rails. So what I did was change your rails from the convex polyotype to cylinders, mess with the geomerty so it looked right, and widen the rails divergance a bit. Now works alot better. You can get the edited file here:
j-space.homedns.org/blender/rails_balls_worksnow.blend
If I were to make it again I would start off with the cylinders set up as proper cylinders, then play around with the size and positioning.
Hope it helps you, good luck.
Thanks, but that link is pointing to the wrong location. I cut ‘n’ pasted it to get the file though so no biggie.
Can you texture this ball? Texturing with the marble_256.jpg texture would make it easier to debug. (this texture is part of the physics demos).
It would show the the rotation of the ball:
http://www.continuousphysics.com/ftp/pub/test/index.php?dir=blender/&file=rails_balls_worksnow.blend
Doesn’t look too bad, the ball gets stuck between the rails due to rotation, cylinder is best, convex_hull second best, and triangle mesh still works (although sticks longer).
I wouldn’t be suprised you can create such a situation in real-life and get the ball stuck completely!
Thanks,
Erwin
Oh lol, I guess I’m going to have to put the http://www stuff in there for future links or else it thinks to look under the current directory.
Yea I can see a ball getting stuck in a simmilar setup in real life especialy if it were made of rubber or something.