I am implementing a collision system which will involve a sphere hitting a surface, from that friction will occur and subsequently causes the sphere to change its angular velocity. I have looked for ways of calculating frictions given the collision point and all necessary info of the 2 object. It seems from the tutorial I looked that friction is proportional to the normal impulse to the impact surafce, I am not convinced by this.I beleive friction should be proprotional to the tangential(with respect to the collision surface) impulse, but I am not well versed in physics to tell if I am right or not. Can anyone me if am right or not?

I don''t know what tangential impulse is, but friction (both the force needed to get an object to start moving and the force needed to keep it moving) is definitely directly proportionate to the normal force.
BTW, you''d probably get more information if this was in the Math forum.

The official zorx website

They say friction is a girl''s best friend.

Friction, okay.
First. air resistance, dampen all motion by a few percent per second

Secondly. friction, dampen the motion further every time you colide with the ground by a few more percent per second, i use about twice as mutch as air resistance.
Here is the tricky part, you need to multiply the friction dampning with the angle (between the normal and your motion vector) so that when the angle is at 0 degrees you multiply by one and when it is at 90 degrees and more you multiply by 0.
Atleast i think this is the way.

[edited by - lc_overlord on December 1, 2003 2:56:32 PM]

Friction is almost exclusively calculated from the normal force of the surface that is assumed to be at rest (the motion is relative, but easier to calculate if you assume one surface is at rest).

Friction can be calculted from the incidental angle of motion, but I can''t remember the equations from my statics class, and my book isn''t here. In general, though, it''s easier to calculate the normal force (which will be affected by the angle), then multiply by the coeffecient of friction.

tangential impulse is the impulse vector running parallel to the colision surface.

correct me if i am wrong but it is not the case that more friction should be produced if the sphere hit the surface at great angle from the normal than it would from a lesser angle from the normal. (e.g a ball hitting the side of a pool table head on(or almost head on) will not spin as much as one hit at an angle.

actuarly it's the oposite, when a ball strikes the pool table at a great angle it starts to spin because one side of it get more friction than the other, that spinn will infact result in less friction.
But a ball that are rolling across a pool table does not have a motion vector that is aimed at 90 degrees from the normal.
Because of gravity it is aimed downwards, thus you wont get 0 friction.
And because of gravity you can only get 0 friction when the ball is falling straight down or if it moves sideways along a steap curved ramp.

Because of gravity you allso get higher friction when you go up a ramp since speed and gravity press it down into the ramp.



[edited by - lc_overlord on December 1, 2003 3:48:19 PM]

Think about it - paperweights (which increase the normal force by a lot compared to the tiny force of gravity which the paper has by itself) prevent papers from moving sideways - all due to friction.

The official zorx website

I think he is talking about friction as it relates to angular momentum, not to transitional momentum. I think the hypothesis you suggested... was right on track... the more acute the angle between the ball''s vector & the surface''s normal, the more change in angular momentum. What everybody else is saying... is that will be less change in transitional momentum. There are two calculations that take place... angular & transitional momentums. You also have to account for the change in transitional momentum because of the angular effects... say for instance your ball entity has a lot of friction (like a bouncy ball on pavement) & its spinning quickly & hits the surface... you know from experience that this angular momentum will GREATLY effect transitional momentum, but now you need to figure out what equations govern this fact.

quick notes:
since angular momentum is taken into acount here, you ball entity should have mass (duh, for momentum equations), & a coefficient of fricion (maybe a static one & a kinetic one)

For the angular momentum OR angular velocity calculation:
The frictional force (orthagonal to the normal of the surface) that influences the angular velocity or angular momentum of your sphere is directly proportional to the normal impulse. The normal impulse has to do with how fast the sphere was travelling, its mass, & its angle to the surface. The frictional force''s applied to the sphere has a vector that is oriented tangental to the sphere''s surphace & opposite in direction to the sphere''s vector projected onto the surface. I think you can consider the frictional force that is applied to sphere to rotate as a TORQUE (angular force) & you can use this along with the sphere''s mass to determine its angular velocity. Keep in mind... whatever force that is applied as torque to spin the sphere also has to be subtracted from the force that affects the sphere''s transitional velocity. Since we don''t want to get TOO complicated about this aspect... you can assume that it will be subtracted from only the X-axis of the sphere''s velocity (sphere was traveling from left to right & bounced back up). How much do you subtract? well I don''t have the equation... but rest assured that the energy of the sphere will remain a constant, assuming no wind resistance & no energy lost from the collisions. So kinetic energy from angular velocity + kinetic energy from transitional velocity + potential energy from the sphere''s position (above ground) = CONSTANT. You won''t have to calculate all these equations for every collision, but write them out to see how an increase in angular velocity would decrease the transitional velocity (in one direction probably).

good luck

ps - if the sphere already has an angular momentum when it hits the surface... things could get tricky, because if the sphere is already rotating (say... like a motorcycle wheel spinning in the air when you jump a ramp) faster than it is travelling in the X direction, than you need to calculate speed of the point that will contact the surface... & that includes transitional & rotational speeds, always apply the frictional force vector opposite to the vector of the speed of teh point on the sphere that collides. So in the case of the spinning motorcycle wheel... angular speed of the tire is decreased & the transitional speed of the bike is increased along the X-axis when the wheel "collides" with the ground. I hope this is clear as a pile of HOG SLOP, e-mail & I''ll explain better.