add_force? Why is this implemented this way? Am I missing something?

wombatstampede

Ok, to be honest, when working with forces and RigidBodies I always use apply_impulse.

But I just had a closer look at add_force ( https://docs.godotengine.org/en/3.1/classes/class_rigidbody.html#class-rigidbody-method-add-force ) and I really don't fully get the concept.

Ok the docs is lacking at this point but I went on to "body_sw.h" and found this:

	_FORCE_INLINE_ void add_force(const Vector3 &p_force, const Vector3 &p_pos) {

		applied_force += p_force;
		applied_torque += p_pos.cross(p_force);
	}

So it is much simpler as I first thought. (I thought Godot kept some list of all positions and applied forces), instead that force is immediately added to the central force and the torque. So it behaves differently than the same force split up into continuous impulses always acting on the same global offset.

But I really don't get the concept of added up forces and torque (3D). I don't see any methods for querying the current total force (applied_force) and torque (applied_torque) of a body. So how do you reset these?

Megalomaniak

Pretty sure the idea of force integration is that you don't reset it and let the physics system resolve it. But I have certainly not put much time into dealing with physics so this is largely a guess on my part.

wombatstampede

To sum this up: add_force() seems faily useless in 3D except for some specific use cases. This use case is when you want to add a constant force to some body (like gravity) and ideally never change it. If you want to change it then have to keep track of your applied forces (and bet the you didn't offset those forces) and apply the difference. Often, it might be just easier to regularly apply impulses per physics frame.

In 2D however you have access to applied_force and applied_torque so you may use these to unset or scale already applied forces.

Ace_Dragon

If you need a constant force, you can declare a variable like this.

var constant_force : Vector3 = Vector3(10,0-10)

and you would place that under the _integrate_forces(state) function as an argument for add_force() or add_central_force()

For applying temporary forces then you can just define other Vector3 variables and do something like... add_central_force(constant_force+my_temp_force).

wombatstampede

Yes, I recently realized after reading the Bullet (physics engine) forums that force values are actually reset after every frame.

So, you don't need to keep track about applied forces as they're reset anyway.

This isn't very intuitive IMHO. I would expect forces to be constant unless changed compared to the temporary nature of impulses. Sadly, the docs don't mention that. Maybe something for a pull request. :-)

Personally, I do exactly as you said but with apply_impulse.