hi !
It seems that i have a litlle problem how i aply suspensions forces on
the car body.
First i explain how i compute the forces.
1) Calculating spring and dampers forces in suspension coordinate.
2) Creating suspension force vector in car coordinate.
3) Now it's more complicated
here's is a little graph to explain:
http://magicfr.multimania.com/images/apllysusp.jpg
a) In red, the suspension, in this exemple rear suspension are longer
and stiffer than front.
In green the suspension force vector.
If i apply the force directly to the body like this ,the car will move
without any engine..waow.. i'm going to be rich ;o) ...well .. maybe not
;o)
b) I decompose the suspensions forces in 2 component, normal to ground
and tangeant to ground.
c) In fact suspensions force are applied in the same amount ( is this
correct ? ) to body and tires. So in the tangeant to the ground ( pink
vectors ), the force apply to the body and to the tires are canceled (
is this correct ? ). So i applied only the blue vectors to the body.( is
this correct ? )
The problem is that it seems to be incorect when there is high angle
between body and ground :o(
Where am i wrong ? I think i forget a force on ground contact impulse
because of the wheel ( in my model ) can only move verticaly to the car
, so there must be a constraint force perpendicular to the wheel motion
in car coordinate ( vertical in my model ).
I'm using a rigid body physics with inverse inertia tensors.
Any help would be very appreciate, and especialy if you have some
solution for a wheel that is constraint with not a particular vector in
the car coordinate. so i could implemente variable geometric suspension.
king regards,
--
Sebastien TIXIER - Game Developer
Dynamics and Car Physics
http://www.eden-studios.fr
GPLRank Normal:-42.98 Monster:-117.44
Hi Sebastien,
the thing you are missing (and you correctly identified it) is that you
aren't taking the wheel contact with the ground properly into account.
Basically, how you should handle things is, when you calculate the
forces acting on the wheels, you have to find the component that acts
along the wheel travel axis (vertical with respect to the car in your
case), and the remainder of the force that is acting perpendicular to
this axis. The force perpendicular to this axis gets transfered directly
to the car body, while the force which points along the wheel travel
axis is used, along with the suspension forces, to accelerate the wheel
along this axis, and only the suspension force is added to the total
force acting on the car body. The wheel is then accelerated according to
the suspension forces and the component of the road force along the
wheel movement axis.
On the other hand, if you are neglecting the wheel mass (and I think you
do, Brutus ;) ), then the suspension force needs not be used for
anything else than to determine the wheel travel (I believe you use the
damper force to determine the relative wheel velocity in this
approximation), and all the external (which excludes the suspension
ones) forces that are acting onto the wheel are directly added to the
total force on the car body. The suspension forces are considered as
internal ones, as the parts (wheels) of the rigid body are effectively
considered as not gaining any momentum in this approximation, and they
do not have anything to do with the body motion directly.
I know the above may sound obscure, so just ask about the details that
aren't clear so I may say more about them.
Greetings,
-Gregor
Sebastien Tixier wrote:
Quoted text (52 lines)
> > hi ! > > It seems that i have a litlle problem how i aply suspensions forces on > the car body. > > First i explain how i compute the forces. > > 1) Calculating spring and dampers forces in suspension coordinate. > 2) Creating suspension force vector in car coordinate. > 3) Now it's more complicated > here's is a little graph to explain: > > http://magicfr.multimania.com/images/apllysusp.jpg > > a) In red, the suspension, in this exemple rear suspension are longer > and stiffer than front. > In green the suspension force vector. > > If i apply the force directly to the body like this ,the car will move > without any engine..waow.. i'm going to be rich ;o) ...well .. maybe not > ;o) > > b) I decompose the suspensions forces in 2 component, normal to ground > and tangeant to ground. > > c) In fact suspensions force are applied in the same amount ( is this > correct ? ) to body and tires. So in the tangeant to the ground ( pink > vectors ), the force apply to the body and to the tires are canceled ( > is this correct ? ). So i applied only the blue vectors to the body.( is > this correct ? ) > > The problem is that it seems to be incorect when there is high angle > between body and ground :o( > > Where am i wrong ? I think i forget a force on ground contact impulse > because of the wheel ( in my model ) can only move verticaly to the car > , so there must be a constraint force perpendicular to the wheel motion > in car coordinate ( vertical in my model ). > > I'm using a rigid body physics with inverse inertia tensors. > > Any help would be very appreciate, and especialy if you have some > solution for a wheel that is constraint with not a particular vector in > the car coordinate. so i could implemente variable geometric suspension. > > king regards, > -- > Sebastien TIXIER - Game Developer > Dynamics and Car Physics > http://www.eden-studios.fr > GPLRank Normal:-42.98 Monster:-117.44hi,
well, it's perfectly clear ;o) while waiting for an answer i came to the same
reasoning.
In fact to simplify the model ( PS2 is not very powerufl in cpu ) , my car
don't really have wheels ;o)
I use a ray/triangle intersection to get the contact patch wheel/ground
position. The ray start frrom the suspension mounting point on the chassis to
the ground. So wheel don't have a weight ... so if the car is not too far away
from the ground the wheels always touch the ground ... no wheel bouncing
effect... BUT actually, i don't have any force acting on the wheel.
I tell how i'll handle this, maybe you'll have a better idea.
On my rigid body engine i have a function that can calculate impulse forces
from a contact point , and a vector that is the normal to the collision. This
impulse force only void the velocity regarding to the normal vector. So this
force is the minimum constraint force that make the wheel not penetrating the
ground .
It's quite complicated formula that use the relative position to CG, inverse
tensor and angular velocity and give a force or a linear force + a torque
So i can consider that at the wheel position the force acting on the wheel is
this impule force.
Then calculate the component of this force perpendicular to suspension travel
, and apply it on the body.
What do you think about this ?
thanks for your help.
--
Sebastien TIXIER - Game Developer
Dynamics and Car Physics
http://www.eden-studios.fr
GPLRank Normal:-42.98 Monster:-117.44
Gregor Veble wrote:
> Hi Sebastien,
>
> the thing you are missing (and you correctly identified it) is that you
> aren't taking the wheel contact with the ground properly into account.
>
> Basically, how you should handle things is, when you calculate the
> forces acting on the wheels, you have to find the component that acts
> along the wheel travel axis (vertical with respect to the car in your
> case), and the remainder of the force that is acting perpendicular to
> this axis. The force perpendicular to this axis gets transfered directly
> to the car body, while the force which points along the wheel travel
> axis is used, along with the suspension forces, to accelerate the wheel
> along this axis, and only the suspension force is added to the total
> force acting on the car body. The wheel is then accelerated according to
> the suspension forces and the component of the road force along the
> wheel movement axis.
>
