Ah, air! It’s grand for living creatures but often a problem for hydronic heating systems.

Air is basically 80% nitrogen and 20% oxygen. Oxygen likes to corrode iron and steel, and there’s plenty of that in our hydronic systems. The oxygen enters with the cold fill water. It’s dissolved in solution and it will react with the metal to create rust. What’s left after that takes place is nitrogen, which will also dissolve in the water and come out of solution, and only when the water gets hot or the system pressure gets low.

This is why most of us position the air separator very near the boiler’s outlet. That’s where the water is hottest, and all of the water has to flow through that big pipe at some point. We could also pipe the air separator at the top of the system, where the static pressure is at its lowest. That’s a good place to catch nitrogen, but not all the water will flow through that high point, so it will probably take longer to get rid of it.

The oxygen in the fill water will also react with any oil, flux, and other stuff in the system, and that can form a volatile gas, so it pays to flush the system with a good hydronic-system cleaner before you turn it over to your customer. Hold a match above an air vent in a dirty system. You may be surprised by what happens next.

Nitrogen and oxygen are both lighter than water so they want to rise up. That’s why the Dead Men put their expansion tanks at the top of those gravity systems. When they closed the tanks and moved them to the basement to keep them from freezing (around 1915), the expansion tanks became compression tanks. The Dead Men had to put the air in the tank, but it would often leave. It took them a while to figure out why, but they eventually learned that absorption and gravity circulation were the culprits.

When you connect a plain compression tank (that’s one without a diaphragm) to a hydronic system, the water in the tank will absorb the air in the tank because that water is relatively cool. Gravity circulation within the pipe that connects the tank to the system will move the cooler, air-laden water out of the tank. Hot water rises and cold water sinks; you don’t need two pipes to make this happen — one will do. The rising and falling waters pass each other within that single pipe.

When you vent a radiator, the system pressure drops. The feed valve adds cold water, which enters the tank. This problem inspired Amtrol to come up with the diaphragm-type compression tank. They figured that if you could separate the air from the system water there would be fewer problems.

Their first tanks had a pre-charge of 12 psi because that’s how much pressure you need to fill a typical two-story house. They used a piece of dry ice to charge each tank. I thought it was ingenious. So did they.

But, as time went by, they learned that the carbon dioxide from the melted dry ice would move through the rubber diaphragm by osmosis and enter the water. There was no way to recharge those early tanks, so Amtrol switched to the Schraeder valve that we see today.

Today’s rubber diaphragms still allow the air to move by osmosis, and at the rate of about 1 psi of loss per year. It’s the same process that causes air to leave a party balloon. It’s always a good idea to check the air pressure in the tank if you’re having pressure problems with the system. You have to isolate the tank from the system to do this, though, because the water pressure will squeeze the diaphragm and compress the air, giving you a false reading on your tire gauge.

When we size compression tanks, we look at three things, any of which can easily change. They are:

  • The amount of water in the whole system;

  • The difference in pressure between the feed-valve setting and the relief-valve setting; and

  • The average water temperature of the system on the coldest day of the year.

Suppose you have a system with a lot of water in it. Think about those older buildings with the big pipes and high-volume radiators. You’re going to need a compression tank that’s larger then it would be on a modern system for those jobs.

Here’s how the Dead Men sized their old-school compression tanks. First they measured the total system equivalent direct radiation (EDR). Once they had that, they did this:

  • If there was less than 1,000 square feet of radiation on the job, they’d multiply the total EDR by .03 to determine the tank size in gallons;

  • If the total EDR was between 1,000 and 2,000 square feet, they’d use .025 as a multiplier; and

  • If the total EDR load was greater than 2,000 square feet, they’d use .02 as a multiplier.

More water means more expansion — that’s why the tank will be so big. If you decide to use a diaphragm-type compression tank on that job, it’s also going to be bigger than it would be if it were on a modern system. Probably a lot bigger.

Here’s a rule of thumb for the diaphragm tank on a high-volume system: Take the size of the standard steel compression tank in gallons and multiply by 0.55 if the building is two stories tall or 0.44 if the building is three stories tall. The answer will give you the volume of the diaphragm tank.

For example, let’s say we have a two-story house with 1,000 square feet of radiation. We’ll size a standard steel tank first: 1,000 x .03 = 30 gallons. Now, since it’s a two-story house, we have to multiply that by .55 to get the volume of the diaphragm tank: 30 x 0.55 = 16.5 gallons of required volume in the diaphragm tank.

You can find the volume of the diaphragm tank in the tank manufacturer’s specification sheet. Table 1 includes, for instance, the volume ratings of the diaphragm-type tanks Amtrol makes:


TABLE 1: Amtrol diaphragm-typetank volume ratings


Amtrol Model Number
Volume (in gallons)
15 2.0
30 4.4
60 7.6
90 14.0
SX30V 14.0
SX-40V 20.0
SX-60V 32.0
SX-90V 44.0
SX-110V 62.0
SX-160V 86.0


In this case, you’d use an Amtrol SX-40-V, or any combination of smaller tanks that will equal or exceed 16.5 gallons of volume. If you wanted, you could use four Amtrol 30s instead of the one big tank.

Now, about those other two variables — the average water temperature and the difference in pressure between the feed valve and the relief valve. If you add water to the system, you’re going to need a bigger tank. If you increase the system pressure by, say, 50% — from 12 psi to 18 psi — the required tank size is going to double.

This sometimes happens without your assistance. If you were to have your circulator on the return side of the system, pumping at the compression tank, and if your feed valve was connected to the suction side of the circulator, then that feed valve is going to add water to the system, but it will only do it once, and on the very first cycle. It does this because the compression tank is the point of no pressure change.

When the pump comes on, it drops its suction pressure below 12 psi. The feed valve senses this and adds water to the system. The system is now at a higher pressure, so the tank is too small. The pressure creeps up when the water gets hot and the relief valve pops. It’s a common installation mistake, and it leads people who don’t understand all of this to use larger tanks than they would need if the piping was right.

The other variable, water temperature, is easily changed. If you size your compression tank for an average water temperature of 170° F and someone changes that to a higher setting, the compression tank is going to be too small for the job.

So, there you go. Not very complicated, is it? A little chemistry, some simple physics, and human nature, of course. All good things to consider while you’re troubleshooting.

Go get ‘em.