How it works:

The yeast/sugar/water generates CO2, probably at a rate of about 1 bubble per second. If the solenoid valve is open the CO2 goes directly to the bubble counter and to the tank. If the solenoid valve is closed the CO2 goes to the storage bottle, displacing some of the water in that bottle which goes to the pressurizing bottle. After a few hours enough water is displaced into the pressurizing bottle to compress the trapped air there and that pressurizes the storage bottle and the yeast/sugar/water bottle. To calculate how much pressure is there, assume each CO2 bubble is about 1/16 inch in diameter. So the volume of gas in that bubble is 4/3 x pi x the bubble radius cubed, which is .002 ml. If the solenoid is closed for 16 hours, that is 57600 seconds. The total CO2 generated in that time would be 57600 times .002 ml per bubble, or 115 ml. In a 1 liter bottle that would change the water level by less than 10%., generating less than 1 inch of water pressure - a very low pressure.

When the solenoid opens that will dump all of the CO2 into the tank, at a rate determined by the resistance to flow through the diffuser used. The pressure is so low, just barely above atmospheric pressure which is about 360 inches of water, that a metering valve may not be needed. If the bubble rate is 2 bubbles per second the storage bottle will not completely empty in 8 hours. This would cause the storage bottle to accumulate more and more CO2 from day to day, until the generated CO2 drops below 1 bubble per second. Some experimenting will be needed to determine if a metering valve is necessary.

An added benefit of shutting off the CO2 when the lights are off is that the time each charge of the yeast/sugar/water bottle will last should be three times longer assuming an 8 hour photoperiod.

MCSLABS, do you think this will work? And, is it much different from your system.

Illustration changed to add needle valve and change storage set-up