You are at The Krib ->Plants ->CO2 [E-mail]

DIY CO2 Controller Project for Planted Tanks

Jim Hurley (, July 20 1993

[Editor's note: Here are some comments by others who've built controllers based on this design.]

A posting I made. I have a scanned GIF file of the schematics for this project here.

I started fishkeeping a little over 3 years ago and didn't know very much. I bought some books, followed the Internet aquaria newsgroups and made some friends with knowledgeable fish-folk. My first tank was a 35g overcrowded community tank, then a year later my wife bought a 27g community tank for our kitchen. A few months later I bought a 100g tank for our family room that was going to be a display tank with lots of plants. A friend who knew I was a techie thought I should get involved with CO2 fertilization, so I studied up on the subject and communicated with a few folk who had set these things up. I latched on to the idea of building a CO2 controller and spent a weekend at the library studying pH probe literature and checking the Thomas Registry for manufacturers. I wrote them and got a lot of free literature that was very helpful. I also called the companies and asked for technical assistance. The tech guys were very helpful - it seems that as long as you do a little homework and get a basic understanding, they are very talkative and give you a lot of useful practical advice. By the way, these companies; Markson, Hach, Hanna, Cole-Palmer, all have excellent, large, free catalogs full of chemical and scientific goodies. Prior to this I knew almost nothing about aquatic plants, let alone terrestrial plants.

So one weekend I bought the pH probe from a local scientific supply house along with some calibration pH buffers and cleaning solutions. After a little experimentation, I felt ready to design a unit.

I had forgotten all the high school chemistry I ever knew and I am basically self-taught in electronics, but what the heck, I'm ambitious.

I looked in the Yellow Pages for some Welding Gas suppliers and found one nearby who sold tanks of CO2. At first I bought a 10 pound tank, but that lasted 6 months, so I later traded it in and bought a 20 pound tank of CO2. This lasts a long time, almost a year.

I also bought the tank regulator at that place.

I next looked through the Yellow Pages for metal plumbing fitting suppliers and found a local supplier who was very helpful. I had done some homework by writing to companies listed in the Thomas Registry and once again I had a lot of free literature that proved very useful. A call here and there brought the necessary advice.

So, I now planned out the layout. A 20 pound tank of CO2 is at 1200 pounds of pressure, so you need to add a pressure reducer to the tank to lower this pressure down to something manageable. That is the job of a pressure regulator that screws onto the tank output and has a tank pressure gauge that shows you when the tank gets empty. The regulator reduces the tank pressure so that the gas output is at something like 50 psi. Some regulators have two pressure gauges that will show tank pressure and output pressure.

Next you need a way to control the flow, because you can't just send 50 psi of CO2 into an aquarium. CO2 dissolves very readily in water, you only need a tiny trickle - about a bubble or 2 per second.

There are two common approaches at this point - manual control or electronic control of the gas delivery. With manual control, you get a fine control valve (called a metering valve or needle valve) that will further reduce the 50 psi output pressure into a tiny stream of bubbles. Needle valves have a tiny hole with a small needle that screws into that hole. The gas comes through that hole and is restricted by the needle. The needle has a very fine turning mechanism that lets you make extremely precise adjustments in the output flow. The valve I use has about 50 turns from off to full on, and I can turn it one revolution to get from one bubble out per second to two bubbles out per second. With such a fine control of the gas flow, you can experimentally find an equilibrium point where the CO2 is dissolved at the same rate it is utilized by the tank and released into the atmosphere.

With electronic control, you put a pH probe into the tank that measures the pH of the water and produces a control signal that can be used to shut off the gas flow via an electronic air solenoid valve. As the CO2 dissolves in the tank, the water becomes more acidic. When the acidity reaches a set point, you can shut off the solenoid and stop the gas flow. Later the CO2 will be absorbed by plants or released by osmosis into the atmosphere and the pH will start to rise again. At another pH setpoint, the pH controller will turn the CO2 back on again. You will still want a needle valve with electronic control, because you don't want to change the pH of the tank too abruptly. When CO2 bubbles into the tank at about 3 or 4 bubbles per second, the pH will drop about .01 ph every ten minutes. This actually depends on many factors, the most important is the carbonate hardness of your water. More on that later.

So now I had this idea. Add a CO2 air-valve after the tank regulator and follow that with the needle valve. A control box I would design and build would actuate the air-valve at the proper pH setpoints by sensing the tank pH via a pH probe.

The next problem was to decide how to mix the CO2 bubbles efficiently into the tank. CO2 dissolves quite rapidly, but if I just bubbled it into the tank, most would dissipate into the air and I would waste a lot. In commercial systems, the thing that does this is called a reactor and is composed of a long tube with things like bio-balls inside. Water is pumped into the top and comes out the bottom. CO2 is bubbled into the bottom and mixes in the column with the water.

What I did was get a plastic T-valve at an aquarium store and cut the inlet into my 100g tank's Magnum 220 filter. I put the CO2 gas into the one section of the T and it mixed in with the incoming tank water inside the Magnum. By the time the water left the Magnum, the CO2 was practically 100% dissolved. With experimentation of the needle valve flow rate, you can get this to happen quite easily.

In fact, for most people, just that part mentioned so far would be enough to get a practical CO2 system running. There really isn't a need to get so involved in the electronic control. I know of several people who just bubble a few drops of CO2 per second into their cannister filter-reactor all the time. By measuring the tank pH over time, they are able to find a bubble rate that keeps things fairly constant with very little hassle. As long as the pH changes slowly and kept within reasonable bounds, the fish and plants don't seem to mind at all.

I won't go into detail on the pH controller, but will highlight the essential features. A pH probe is basically a tiny battery, like those lemon-zinc batteries we all made in chemistry class. It puts out a small signal that is very linear with respect to pH - about -53 millivolts for every +1 pH change. The output is almost exactly 0 at a pH of 7.00. The current is infinitesimal, so you need an extremely high impedance op-amp input circuit to amplify that signal. I used the LF356 op amp, but I later found that an LF411 would have been better. I made the gain of this stage 16.8 so that I would see -1 volt per delta pH instead of the standard -.053v/delta pH. After that I added in a reference voltage of -7 volts to get a true pH signal, although it was negative, not positive (since the output of the probe is 0 when the pH is 7). This signal is then sent to an adjustable comparator with adjustable hysteresis. The comparator setting I put at -6.9 and the hystersis is -.1, so I get the comparator output ON when the pH is above 6.9 and OFF when the pH drops to 6.8. This gives the tank a .1 pH swing which is very tiny. In practice, the gas is on for about an hour and off for about an hour and the gas tank lasts almost a year. The comparator output drives a solenoid driver transistor which actually does the actuation of the air-valve. Air valves can be bought to be activated by almost any voltage: 120VAC down to 5VDC - I used a +/-12 volt circuit, so I use a 24VDC solenoid. Most people might get an AC line level solenoid switched with a timer circuit - controlled by the same circuit that turns the lights on and off.

I have had this circuit in operation for about 18 months now and the plant growth is incredible. I need to prune things every two weeks or the plants take over the whole tank.

I have 4 40W Tritons providing the lighting.

Our local water is 3.5 KH which is a little low, but OK. There is a table in the book 'The Optimum Aquarium' that shows the relationship between pH, KH and dissolved CO2. If the KH is very high, you need to dissolve a large amount of CO2 to get the pH down near 7.0 where most plants thrive. If the KH is too low, the pH will swing wildly with very little CO2 added - not good for the fish. Optimum CO2 levels are about 20-24ppm. More than that will interfere with the respiration of the fish (and the plant growth will be astronomical). You can change KH by adding a little baking soda to increase it, if necessary, or distilled water to decrease if that's your problem. Probably most people won't have any trouble. I use the LaMotte CO2 and pH test kits and Tetra KH test kits to make these measurements, as well as having calibrated pH buffers for adjusting the pH controller.

Don't forget that plants need to inspire O2 at night, so some people with manual controllers, use an air-valve connected to a 12 hour timer, so the CO2 bubbles only in that day-time. If the KH is too low, this may cause dramatic pH swings night-to-day, so be careful.

So here's a rundown of the parts and costs of as typical CO2 system:

PS: This weekend I saw the Tetra CO2 system which uses a small tank to bubble CO2 under an inverted jar in a tank. You manually turn on the CO2 to fill the jar and then turn it off. The CO2 dissolves slowly in the tank.

This is a pain to do every day. I originally tried this idea by connecting a hose to the CO2 tank and bubbling CO2 under inverted Mason jars that were suspended in the sump of my 100g. I did this for a few weeks while I was building the pH Controller. It's a drag but it works. The CO2 dissoves slowly in a few hours and the jars fall slowly into the sump.

Another disadvantage is that the Tetra CO2 tank looks very small and may not be refillable, so you may need to buy a new tank every few months.

Addendum (Long-term)

From: Jim Hurley <>
Date: Sun, 12 Jan 1997 19:45:52 -0800
Subject: pH probes

A few months ago, a pH controller I built in Feb 1992 broke. It was stuck in the ON position, so I just throttled back the CO2 via the needle valve until I could get some time to work on the controller.

This weekend I had that opportunity and I discovered that the controller was fine, in fact it was still very well calibrated, benefitting from the low-drift op amps I chose in the design.

What had happened was the the pH probe was no longer working. It kept reporting a very high pH all the time.

I guess that's not too bad considering I had done zero maintenance on it for the last few years.

Anyway I bought another one, this time the one had an exposed glass bulb, whereas the original one had a plastic cover around the tube. (The original was a Hannah 1910B, the new one a Hannah 1912B, perhaps the old one isn't made anymore. except for the protective plastic they look the same).

Unfortunately, after one day, the pH probe bumped into something while I was doing water changes and the fragile exposed bulb broke.

Moral - don't get that type of probe - or protect it with a porous substance like a piece of sponge. The probe tip is very delicate, you're not supposed to even touch it, so maybe pushing it into a soft sponge isn't so good an idea. From what I recall when I was researching the probes back in '92, the glass is a specially made type and scratches will affetct the pH response.

Up to CO2 <- Plants <- The Krib This page was last updated 23 June 1999