After staying with DIY yeast-generated CO2 for 4 years, I finally got tired of having to remix the yeast and sugar solution every 3 or 4 weeks. So I started looking around for a cheap way to upgrade to a pressurized CO2 system. Here is what I ended up with, for around $100.
I bought a cylinder and double-gauge regulator by mail from Rapids Wholesale Equipment Co., a home brewing supply company. A new steel 5-lb (CO2 weight) cylinder (shipped empty) and double-gauge regulator costs $93. Aluminum cylinders, which are lighter, cost about $10 to $15 more. 2.5 lb cylinders are about $5 less.
Although the 5-lb size is more standard, I bought a 2.5 lb cylinder because I do not have room for a 5-lb cylinder under my aquarium. I also got the double-guage regulator. The single-gauge regulator would be a few dollars less but would show the regulated pressure only. On a double-gauge regulator, one gauge shows the pressure inside the pressurized cylinder, while the other gauge shows the regulated output pressure which you can set by turning a screw. However, since compressed CO2 exists mostly in liquid form, the gas pressure in the cylinder will stay constant (at about 950 psi) until it is nearly empty and all the liquid has evaporated. Then the pressure inside the bottle will start to drop, and you will have only a few days to a few weeks to refill before it is completely empty (depending on your cylinder size and flow rate).
I got a Parker needle valve from a friend who bought a bunch at a swap meet for $2 each. If you aren't so lucky, you can try industrial supply stores (such as Kaman Technologies in San Diego, where a comparable needle valve can be found for about $9). Make sure any needle valve you buy is of the type that can be shut off completely, and has at least 5 turns from completely closed to full open--the more the better, since it would give you finer control. (It took some footwork for me to find the appropriate connecting hardware pieces, since my Parker valve has metric dimensions while the regulator has English specifications, but I finally found what I needed at Hose Pros in San Diego.)
The last piece of brass hardware needed was the reducing barb which connects the needle valve to the flexible air hose. This I found--ready for this?--at Napa Auto Parts, for about $1.50.
I got the BioPlast bubble counter and reactor (diffuser) by trading in plant cuttings at my local fish store. They normally would cost about $50 for a good set. However, if you do not have access to a cheap source for these, you can always make a bubble counter from a soda bottle as shown in the accompanying figure (note: if you use this bubble counter for a yeast-CO2 setup, make sure to change the water inside the bubble counter occasionally to prevent fouling). As for the reactor, you can use the CO2-bell bottle shown in my DIY Yeast-Generated CO2 setup, or use your canister filter if you have one.
I filled the CO2 bottle at a local fire-extinguisher refilling store for about $10. 5-lb bottles are universal, and you can usually bring your empty bottle in to any welding supply shop and exchange for a full one. 2.5-lb bottles have to be refilled at places where they refill the bottle on-site.
I made a bracket to secure the CO2 cylinder to prevent it from falling over. If the CO2 cylinder is horizontal, liquid CO2 could be forced out into the regulator, damaging it and could create very dangerous situations.
I set the regulator for a 5-psi output pressure, then adjust the needle valve to achieve a bubble rate of about 15 to 20 bubbles per minute. This CO2 flow keeps my aquarium pH at around 7.0 to 7.4.* At this rate, my 2.5 lb CO2 cylinder lasts about a year.
The pH will tend to be lower in the morning, and higher in the evening, as the plants use CO2 during the lighted period. Ideally, pH can be kept constant by installing a shutoff valve in the CO2 line that is controlled by an electronic pH sensor. This usually runs about $200. Other people have tried cheaper alternatives, such as installing shut-off valves or air pumps controlled by timers. What some have found is that this usually results in even wider pH swings (in the opposite direction). A possible explanation is that the amount of CO2 consumed by the plants is small compared to the amount injected into the aquarium and lost via normal diffusion into the air. I think the little pH change during the day is just fine and is not worth trying to eliminate.