Sum of torque

[EOW]

I prefer a liquid it must be more efficiency. Let the door open and like the pressure moves all the time this will give a torque on the device, like that ? Yes, but I'm not sure of:

1/ water will move because it is compressible, even a small volume of water, this amount can't be ignored, but the energy seems very low
2/ an object rotates => no torque, so this is the delay of pressure to be stabilized that can give a torque, here with the door always open, the pressure change all the time. Look at positions: it's necessary to close door and reopened because the torque is counterclockwise and after clockwise, are you agree ?

I drawn the device, and I think there is one wave inside the upper circle and another wave in the lower circle. In the upper the pressure increases, in the lower the pressure decreases, a part of time, and this must give forces F1 to F4. Length of arm can be higher. The material must absord the wave for prevent round trips of waves.

have you tested on a software ? Maybe if you draw something for understand why you want to recover energy from this device I could understand.

[lumen]

Ok, you are saying that the release of the pressure into the second chamber causes movement and then the centrifugal force would cause the next compression cycle and the release would be delayed until a time when the release would again act to increase the speed. The delay of the gas may cause a shift in weight that works against the rotation like shifting your weight on a swing to slow yourself down.

Would this be easier if the centrifugal force was used instead to rotate a generator and some of the power was used to increase the speed of the disk.
At some point there would be excess power that could be directly used for other purposes and the design would be achieving the same goal using the same forces but would be easier to design and could operate at lower speeds.

[EOW]

There is no gravity, so no weight. For recover energy, maybe turn the device at the max angular velocity possible to do, and recover energy at constant velocity is better. With a constant velocity the device must be more stable.

Maybe it's possible to use a piezoelectric vibrator for create the waves and test with only one circle. Step2: the valve at left, move out the valve quickly. Step3: return of 180° the device around itself, now the valve is at right. Step4: move in the valve. Step1: return the device around itself of 180°, now the valve is at left. Repeat the cycle.

And you, how do you want to recover energy from this device ? I don't understood your Roberval system

[EOW]

I don't use gravity here. All volumes are constant. I use only the red shape, I drawn black circle for show the center where I put the springs.

The pressure come from small balls inside the red 'S' shape. The balls are attrack to the center or repuls from it. This gives a pressure p, 2p, 3p on the side where there are the forces F1 and F2. The balls are vey small.

I give the force F1, this need the work W1 and I recover the work W2 from F2. The move like a snake. But for have a continious mouvement, it's necessary to add another part of circle and give the pressure. Put a pressure inside a shape don't need energy in theory.

[sm0ky2]

the balls do not behave exactly as a fluid, but much slower. pressure does not have to be constant through the length of the tube.
When pressure varies in one part from the springs, more balls move out of the "higher pressure" area quickly.
this lowers the pressure by a proportionate amount in the part of the tube just before the springs.
meaning, to maintain the original pressure in the rest of the tube, you have to increase the input of balls
to bring the pressure back up by an amount of balls squeezed out by the springs.
the torque comes at a cost.