Fix suspension force calculation in VehicleBody3D to accurately follow Hooke's Law

[FuWan722]

Fix issue 112492

Modify VehicleBody3D suspension force to follow Hooke's Law.

Before, the suspension force was calculated incorrectly, where:

F(N) = stiffness(N/mm) * offset(m) * vehicle_mass(kg).
Meaning, as the vehicle gets heavier, the spring does not compress. This is because under the hood, the force is also multiplied by vehicle_mass.

before_vehicle_suspension.mp4

After, the suspension force is calculated as:

F(N) = stiffness (N/mm) * 1000.0 * offset(m).
This can be found in the Hooke's Law for spring. Multiplication of 1000.0 to convert N/mm to N/m

after_vehicle_suspension.mp4

[Zireael07]

Have you considered making a copy of this PR to Bullet Physics? The Vehicle classes are a 1:1 port of Bullet physics

[skorpnok]

You'll need to squash those commits into one to follow the contribution guideline and the unused line of code should probably just be removed not commented out, git handles the history anyways. Otherwise, nice find!

[shakesoda]

Have you considered making a copy of this PR to Bullet Physics? The Vehicle classes are a 1:1 port of Bullet physics

i'm a bit surprised this issue has persisted in bullet since the class was added in 2006

i have very much run into this and the suspension length issue myself, although the fix might need some extra care/warnings since every project using vehicle bodies would necessarily already be working around this bad behavior

[FuWan722]

Have you considered making a copy of this PR to Bullet Physics? The Vehicle classes are a 1:1 port of Bullet physics

I can try. Im not too familiar with the Bullet physic engine itself, and the stiffness and damping value in Bullet is a bit vague as I do not know what unit it use (could be either in mm or m). I will probably make a PR removing the mass multiplication over there