BuoyancyEffector2D - should this be happening?

A negative magnitude makes it go in the opposite (180-deg) direction to the flow angle so yes.

BTW: You should use Rigidbody2D.useAutoMass to set to true and set the Collider2D.Density to be relative to the BuoyancyEffector2D.density to make it sink/float. With auto-mass off, you'll get a collider density fixed to 1.

 

It use density only. The object won't sink unless it enters the effector with some velocity i.e. falling via gravity. From a standstill, it would rise if the density is less than the effector.

 

No idea, that's not what happens when I test it. There was an issue a while back where Rigidbody2D.gravityScale wasn't taken into account but as I say, that was fixed a while ago.

Try using the latest version or set gravity scale to 1 as a test.

From above you mention that mass is 3. As I stated already; if you turn-off Rigidbody2D.useAutoMass and set the mass explicitly then the density on colliders is fixed at 1. The mass is irrelevant for the effector.

Beyond that I have no idea why that would happen.

If you are still seeing issues then I would suggest creating a bug case and posting the case number here so I can help you further.

 

So my previous statement wasn't entirely accurate. The actual relative density is what is used which is what I was trying to get across however the total force applied (just like any other force) is relative to the mass. In other words, a larger mass will react less per unit force than a smaller mass.

With a fixed density of 1 then to get more buoyancy you need to either increase the area of the collider and/or reduce the mass to increase the forces effect.

This is why useAutoMass should be used for all objects in the effector. This way, their size and density ensure they all react relative to each other. When you select this in your project, the mass is changed from 3 (explicit) to 1 (auto-calculated) i.e. the force has three times the effect which is why the upwards force can then counteract gravity.

The buoyancy force is calculated as:

Code (CSharp):

1. var buoyancyForce = -EffectorDensity * SubmergedCollider2DArea * WorldGravity * Rigidbody2DGravityScale

 

By relative density I mean the density difference between the effector density and the collider density. The bigger the difference, the bigger the force. Also the bigger the submerged area of the collider the bigger the force. Finally, the lower the mass of body, the more effect the force will have.

In simple terms; use auto-mass. When you've got multiple objects floating, the mass will be calculated automatically meaning that they'll all act similarly according to the density you give them as the mass will be calculated as the density * collider-area. If you don't use it then the density is fixed at 1 and the only variable you have is the mass of the object (which changes how the force acts on it) and the area of the collider which is used to calculated the submerged area and the size of the force.

In your case, you assumed that because the density was 1 that it would float in an effector set with a density of 2. This wasn't the case because the mass of the object reduced the effect of the buoyancy force. Reducing the mass to 1 (for instance) meant it floated. Alternately, you can increase the buoyancy force by increasing the area of the collider.

 

I presume you mean not using auto-mass on Rigidbody2D components as it isn't limited to whole projects using it or not. Sure, as I've stated, with it off on the Rigidbody2D, the Collider2D have their density fixed at 1 (because you are explicitly setting the mass) so it's down to the area of the colliders and the mass of the body only.

In your example you don't give a reason why you would want auto-mass off; you use it and give the lead a high density and the wood a low density (use real values if you like).

 

The density is specified on the colliders when the Rigidbody2D they are attached to is using auto-mass. With it off, it is not shown.