A candidate solution (to the problem of how the B units can enter the base of the primary tank without water loss) is this.

An airlock at the base of the primary tank could be pressurised using a float activated air compressor to maintain a pressure of 500,000 Pascals inside the air lock. The compressor would only activate if water entered the air lock.

There would be two flaps in the airlock. An outer and inner gate.

The B units would enter the airlock from outside the primary tank. The outer flap (like a cat flap) would close when the B unit entered to keep air pressure as high as possible.

Then the inner flap of the airlock would open allowing the B unit to enter the high pressure water at the base of the primary cylinder.

The water would be unable to flow out into the airlock chamber because the pressure in the airlock would be maintained to 500,000 Pa by the air compressor.

The Abac Genesis air compressor consumes 11kW, but produces 800,000 Pa at a rate of about 0.02 m3/s.

0.02 m3/s is equivalent to 20 cubic litres per second, so the right configuration of B units just might work (though I have not done any buoyancy maths yet).

So the B units should be able to enter the primary tank through an airlock without water escaping.

The water at pressure of just under 400,000 Pa would not be able to escape from the primary tank into an airlock pressurised to 500,000 Pa. Simply not possible.

The output nozzle of the air compressor might also be directed to inject air into the B units if needed to give them extra buoyancy. Again I have not looked at buoyancy maths yet, but this is a candidate solution to the water loss problem.

I agree. It makes no difference whether you use an air lock or pressurise the smaller tank. There should be no water loss if this is done correctly. Both ideas amount to the same thing with different labels.

However, you probably have to leave the primary tank open to atmospheric pressure. If you seal it, the first problem is that the chain keeping the B units together will have to leave the tank through some sort of seal (I think this would be difficult if not impossible).

But my main objection to sealing the primary tank is that there is no reason to seal it, and pumping pressure into sealed tanks is a bad idea if there is no reason for it.

Leaving the primary tank open to atmospheric pressure would be fine. Lower pressure in this tank could not be a bad thing.

You mention pumping water up to the top of the primary tank from the base. This will consume franchise sized amounts of energy. The general rule with pumping water upwards is that you will spend 5 times as much energy pumping it uphill than you will ever get out of it at the bottom of the hill. So anything and I mean anything is better than pumping water uphill in an energy generator.

But this is all detail. My main concern is the speed in m/s at which the buoyant B units will ascend the primary tank.

The speed, when we do the maths, could be as low as 0.5 m/s, in which event the RPM will be low, and therefore the power output will be low.

How would you propose to try and increase the upward velocity of the B units to prevent low RPM?

The circuit of B units is a circuit, and thus what might have been high downward velocity due to gravity will be retarded by low upward velocity due to buoyancy and resistance. It will be upward velocity due to buoyancy and low velocity due to resistance.

This is the key to whether or not the system is viable. Upward speed of the B units against the resistance of the water.

It really is all about buoyancy velocity in terms of power output.

While, knowing the mass-displacement of the bouyancy chamber, while in the working fluid, at a known pressure CAN be calculated.

In practice, the bouyant-force is measured, then the calculations are performed in reverse, to obtain the "effective displacement" volume.
This is because of the unknown pressure expansion curve of the low-pressure chamber seal. At different inner and outer pressures, rate and amplitude of expansion, the inner seal expands differently.
 [note, this is a convex expansion function which can vary greatly with the use of different materials]

The B-unit must be first placed in the fluid, and the automated Zero-B state is iniated. This creates a "weightless" condition, and from there the B-unit scales the Bouyant state: positive (up force) and negative (down force gravity). We can measure bouyant force in either verticle direction at any desired state. With the measured verticle force, we can establish the displacement with respect to the mass of the B-unit, and its initial volume displacement.  It was mentioned before, the energy requirements to run the systems of the B-unit itself. To answer that question, the circuitry and logic chips used in the B-unit control systems consume less energy than a digital wristwatch.
The two power consuming functions are the wireless transmitter (optional) and the hydraulics used to alter the bouyant-state.
The latter accounting for 99% of the energy input, and also presents an engineering hurdle that must be overcome.
the test units are not large enough to carry their own hydraulics in air.
The mass per unit volume of a denser fluid such as water make this problem dissapear, even at test unit size, however, it would be more beneficial to displace additional fluid-mass to lift the additional mass and still provide usable bouyant force. that will get rid of the hose that tethers the B-unit's caliper to the ground-based main cylinder.
There is much room for design improvements, such as the use of plastic cylinders, or a lightweight aluminum tubing that has been investigated. low-density hydraulic fluids, ect.

The worm-gear actuator has been recently improved, it now uses a 5v motor, and is much smaller / lighter

the original test unit used a 12v  car window motor and actuator, to depress and retract, the hydraulic lever using a 6v util bat.
Now this is done using a wormgear on a much smaller compound-lever, lowering the size of the motor/actuator
and its energy use/ battery size.

-------------------------------------------------------------------


The actual power generation should take place with each invididual unit. This eliminates a majority of the timing concerns, and cooridinating all parts of the system.  Link the interface to a large flywheel, and a 1-way "freewheel" greaing to a much smaller driveshaft. simlar to a bicycle.... the moving B-unit gets the flyhweel moving, and this power is translated to the generator shaft at much faster RPM. additional flywheel mass can be added to the shaft itself, which keep the generator in operation constantly, although the B-units are adding energy periodically.

Think of those toys, where you pull a zip-string, and the driveshaft spins, then a freewheel gear and spring combination wind the zip-cord back up.
something kind of like that, but more cyclical, probably two-stage.
up and down.

the up-gear would be larger, to make use of the slower speeds and greater upward (bouyancy dependent) torque,
the down-gear would be smaller than the up, to make use of gravity's constant speed, and mass-dependent torque.

Using that type of generation interface, then you can control the number of units rising and falling independently, as well as the unload and reload phases. That may be one benefit of an "air-lock", that you can separate x-number of B-Units, to reload, without disrupting the main track with upper tank seals. Just pump the water back to the main tank, as necessary.

If the generation interface were to be a continious loop it could be geared like a bicycle-chain, with a clutch that releases during the unload-reload phases, and changes to a larger gear during the "up" phase.

so at any given time, you can have whatever number of up-units, and down-units adding energy to the driveshaft. in theory, you could just store whatever ammount of energy into the mass of the fly-wheel, and generate power at a constant rate via the drive shaft.

in the example with the scuba diver, and his emergency lifevest
That little CO2 cartridge is packing approx ~  881.3 Joules of energy

take your boat, out into the ocean, 80-90 feet or so.
strap a bunch of heavy gear to your body, and swim down the the bottom,
Now give your friend your favorite wenching system, loaded with 881.3 Joules of energy.
and you will understand why scuba gear now includes a BCD (bouyancy control device).
 dont forget to pull the chord on your CO2 cartridge lifevest, before you drown.

Are you building your machine smoky?

I am waiting for a mathematician's feedback in relation to my own machine before a build attempt. Makes sense to do the maths first don't you think?

the B-units i built several years ago flew in air, carrying ther own hydraulic cylinder (not the pump), they work.

Underwater, the bouyancy mathematics are the same, and the logic-chip that controls the B-Unit works in water or air (or any fluid/gas).

there was problems with tensil-strength and outside pressure that collapsed the water unit in on itself, so the unit's casing has to be redesigned to make it stronger. Creating a low-pressure chamber in air is a lot easier than doing it underwater, but it can be done. I have seen floating concrete blocks what were "evacuated" hollow centers, plugged with rubber. the weight of the B-unit is not the problem, and the pressure needed for the hydraulics to expand it is still in range of the pumps we used.
 its the casing that needs to withstand the waterpressure, and inner low-pressure.
the math with the energy involved is straight forward.

the math with pressure inside/out and tensil strength of the casing goes far beyond my bachelor's engineering education... thats going to require some professional assistance to do the math on that one. Probably the same guys that design Submarines.
  i found a work-around solution that can withstand far greater pressures than are required, but as for the math on that i dont know enough to even try..

i can do water pressure at depth-X, but when you add in a low pressure chamber it gets far more complicated. because there is not outward pressure on the inner surface of the chamber, like there is with a pressurized-gas chamber.
----------------------------------------------------------------
In air, we work in terms of Bars, 1 bar for a collapsed B-unit, 2 Bars when it is expanded (maximum bouyant state).
A 0-B state is approx. 1.3 Bars
or in terms of the hydraulic system 130,000 Pa. which is roughly 1/10th of the pressure applied to the brakes on your car when you stop. This is for a low-pressure chamber volume of 0.07 m^3 .
This creates a weightless state, and is (roughly) the starting point of the logic-chip controller when it boots up, or the command is given to return to the 0-B state.

Larger B-units will require more hydraulic pressure to achieve bouyancy. The technology was originally designed for use in a flying craft, but there wasnt enough interest in the device to continue development.

Relating the pressure to actual input energy, in terms of electricity from the battery, is better measured than calculated. because of all the components involved in the electric motor, lever, hydraulic actuator, master-cylinder compression ratio, ect., It can be calculated using standard conversion for pressure, but in practice, the system experiences great losses, and thus the battery consumption is far more than the energy to create said pressure. This, as i noted previously, does not relate directly to bouyant force, created by the device.
 
the logic system is 5v at very low current, and its energy usage has been completely ignored in terms of "input energy", because the value is somewhat negligible.
-------------------------------------------------------------------

My interest in using this as an energy generation system was derrived from a simple equation:   mgh

once this balances out with the energy of the hydraulic system:

leveraged force x distance of lever depression
                             = (mass of the B-unit) x (9.8 m/s) x (height)

everything else is essentially "free energy".
Bouyancy continues indefinitely, within the limitations of our atmospheric pressure, which is theoretically over a mile high.

hey smoke, thought this may be of interest, was watching myth busters and saw and interesting fact that actually changes the perception of many vintage bouancy machines, you know the one where hollow balls are fed into the bottom of a tank, weel they were all debunked because it was said you could not get the balls into the bottom of the tank for the same reason they float up in the tank, mythbusters proved this false, not intentionally regarding the machines, but the pumped ping pong balls down through the water into a sunken boat to see if it would float like the 70 year old donald duck cartoon and it worked refloating the boat, but they pumped the balls down with water and used the viscosity and cohesion of the water holding onto the balls, you see because they were in a tube they were only subject to the pressure within the tube not the whole ocean, and as it was mosttly filled with air (the balls) the pump was quite small and managed easilly. interesting side information from their film that may be of some use in designing such machines, hope you can use it

Archer

ps the viscosity,cohesion and pressue variance regarding the tube is just my analisys of what was happening it iis not in the film, so it may not be exactly the reason, but my physics and logic says that appears most likely why it works