Thermal Expansion Tanks

The shortcomings of previous designs lead to new flow-through expansions.

As part of my consulting business, I often perform inspections of domestic hot water systems associated with various personal injury or property damage incidents where there has been scalding, Legionnaires’ disease outbreaks and property damage from floods. I have found that on a significant number of newer water heater installations, there is no thermal expansion tank or other means to relieve the pressure build-up from thermal expansion as required in the model plumbing codes. The International Plumbing Code and the Uniform Plumbing Code address requirements for thermal expansion. 

When there is no thermal expansion tank and the hot water usage during peak periods is significant, the temperature swings can be drastic. During the heating cycle, if there is no water usage, there needs to be a way to relieve the build-up of pressure from thermal expansion. 

I often find the temperature and pressure relief valve is spitting a small amount of water as the pressure builds higher than 150 psi to relieve the pressure. Constant, or intermittent, relief valve discharges can lead to calcification or scaling-up of the valve seat, and in the future the valve will not operate at the designed pressure, if it operates at all.  

There must be a problem with your water heater; it’s leaking!

Many years ago, a water heater sales rep in the Michigan area was called to an automotive plant, where the plant operators told him there was a defective relief valve on the new water heater that he had sold. The water heater was serving a restroom in the plant that was heavily used several times a day when the assembly line was idled for workers to take a 15-minute break. During the break, hundreds of workers headed for the restroom to wash their hands. When the break was over, they all had to return to their stations. There was virtually no restroom usage during the heating cycle when the water heater burner was turned on. 

There were no thermal expansion tanks installed, but there was a four-inch dial face pressure gauge installed on the piping near the water heater, which allowed us to see the pressure in the system. The sales rep came to look at the installation, and while he was there, he witnessed the pressure gauge on the water heater moving like a drag racing car speedometer when the car leaves the starting line. The pressure started at about 60 psi and steadily rose up to about 150 psi, where the relief valve would spit about a cup of water onto the floor, and the pressure gauge needle would drop down to a lower pressure. 

The maintenance man said, “See, something is wrong! It’s not supposed to leak like that.” The sales rep noticed there was a swing check valve on the cold water branch to the water heater, as well as a small circulator pump, and they put a check valve between the cold water supply and the hot water return connection to the cold water pipe entering the water heater. This is common, but there was no thermal expansion tank. The pressure showing on the gauge would build to 150 psi, discharge, and then the pressure would drop back down to about to about 60 to 80 psi. This rise in pressure and discharge of water would continue as long as the burner was on. It lasted about 20 to 30 minutes and cycled every few minutes. 

This clearly illustrated why a properly-sized thermal expansion tank was needed to allow a cushion of air in a thermal expansion tank (bladder tank) to absorb the thermal growth. The sales rep recorded what was going on for the purpose of training programs. 

Most thermal expansion tank manufacturers have sizing guides to select a tank that will take into consideration thermal expansion based on the temperature differential of the system and volume of the system. Select a volume that will allow the system pressure to stay under 80 psi. When the system pressure is closer to 80 psi, it will require a larger tank to keep the pressure rise below 80 psi. 

The solution was to add a thermal expansion tank and replace the temperature and pressure relief valve that had been cycling so often for a couple of months.  

A thermal expansion pressure problem 

A few years ago, I was asked by a product manufacturer to assist with investigating a plumbing system in a building where there appeared to be a pressure problem that caused some plastic components in a product it had manufactured to rupture. When the component ruptured over a holiday weekend, it led to significant flooding and several million dollars of building and product damages. The insurance company paid for the damages and then sued the manufacturer of the component to recover money under a subrogation case. 

The mystery was why and how there was a pressure build-up that was enough to rupture the component. I visited the building for an inspection and found that there was a backflow preventer on a process water supply to a portion of the building where the flood occurred. There were also five 250-gallon electric water heaters in the process water system and no thermal expansion tanks. I noted many other code violations and other problems in the system that were not directly related to the flood incident, but there were many signs of poor installation and maintenance practices. 

One of the signs of a long-term problem was white calcified stains on the floor under every temperature and pressure relief valve and stalactites hanging from the relief valve pipe. There were also piping connections for a water softener that had been removed. 

At some point there was a water softener. The hard water in the area and constant cycling of the water heaters caused pressure to build in the system, and the relief valves were discharging constantly. Eventually each one of the relief valves scaled-up to a point where they ceased operation until all five no longer worked. The event caused a significant pressure build-up. 

Piping failure occurred when a hurricane hit the area and caused a major power loss of all five electric water heaters for a couple of days. When the power was restored, the water heaters experienced a cold start, and the thermal expansion was so great with seized-up pressure relief valves, the pressure built up to the point that the weakest point in the system failed. There was no thermal expansion tank, and the relief valves could not relieve the pressure. The many hundreds of pounds of built-up pressure exceeded the rating of the piping system and plastic component in the piping system. 

A materials expert was hired to testify for the insurance company that was suing the component manufacturer in the subrogation case. The expert conducted tests on a similar component and found that it failed at 1,100 psi. He was later able to get the component to fail at 440 psi, with 180-degree hot water and water hammer spikes introduced at the same time. The manufacturer had a published pressure rating of 320 psi in the catalog, so he was still good. 

In the courtroom, the expert testified that the component should have been made of metal, not plastic, because metal is stronger. The materials expert, a PhD with multiple degrees, taught material science at a major university. The judge read my report, which stated the code has a maximum allowable pressure in the plumbing system of 80 psi; that the code required thermal expansion tanks to relieve the pressure, and there weren’t any. There also weren’t any relief devices that could keep the pressure below 80 psi installed in the system. All of the relief valves on the five water heaters, which were rated at 150 psi, were seized up because they had removed the water softener to save on maintenance costs, and the relief valves were not checked and maintained properly. 

The judge asked the expert if he had any experience with plumbing. He said he installed plumbing in his house many years ago. The judge then asked him for a technical reason for why he thought the component should be plastic instead of metal. He said because in his opinion metal is stronger and would have held a higher pressure. The judge then asked if the he knew what the maximum pressure was supposed to be in the plumbing system. The expert said about a 100 psi; he did not know the maximum allowable pressure in the plumbing code. 

At that point the judge asked the attorneys to come to his bench for a conference. He already knew the answer because he had read my report. During the sidebar, he asked the insurance company's attorney if he had any more witnesses. The attorney said no. The judge said he was going to give his expert one more chance to come up with an answer. 

The judge had the bailiff call the expert back into the courtroom with no jury. Once again, the expert could not give the maximum allowable pressure in a plumbing system, and he couldn't provide a technical reason why metal is stronger than plastic. The judge dismissed the expert and thanked him for his testimony. 

Then he asked the insurance company's (plaintiff) attorney to call his next expert. He had no other expert, so the judge dismissed the case. There are many lessons here, but the most important one is thermal expansion tanks are very important to prevent pressure build-ups and water damage.  

Thermal expansion tanks and Legionella bacteria 

Thermal expansion tanks have been installed with a way to drain downward, and add an air charge to the top half of the tank. In years past, the hydro-pneumatic tanks were only used on heating hot water systems that were closed loops for when the system heats up from a cold start. Thermal expansion can increase pressure and cause problems with fixtures, pump seals, and flanges and joints. Years ago, there weren't any thermal expansion tanks on domestic hot water systems because there weren't backflow preventers on building water service entrances to create a closed system. Back then, water would heat-up and expand back into the city water main through the building service pipe connection. 

With the introduction of backflow preventers, domestic hot water systems became closed systems. No one was using water, so the need for thermal expansion tanks grew. The original thermal expansion tanks were steel- or glass-lined tanks with an air charge on top and water on the bottom, and a sight glass to see the water level. There were many problems with those types of tanks because there was no separation of the air and water. The air would often dissipate into the water over time and then there was no thermal expansion capability, as water does not compress like air. 

Eventually someone designed a tank with a rubber diaphragm, and since then many designs have used newer synthetic rubber bladder materials. The older tanks were not designed for use in drinking water systems; they were adapted from hydronic heating hot water systems. Now, the tanks have NSF-approved surfaces and bladder materials suitable for drinking water systems. 

The material and bladder membrane issues were good improvements to the design of hydro-pneumatic tanks used for thermal expansion, but they did not address the dead-end tank designs, which has become apparent in recent years with many Legionella outbreak investigations. Water sits in the thermal expansion tank. After a few days, the water treatment chemicals, such as chlorine, mono-chlorine, or chlorine dioxide, dissipate down to levels that are ineffective at controlling Legionella bacteria growth. When there is a sudden drop in system pressure, the outflow of water from a dead-end thermal expansion or hydro-pneumatic tank used for a domestic water pressure booster system can discharge a very high concentration or dose of Legionella bacteria into the domestic water system, which could significantly increase the chances of an outbreak. 

A few manufacturers have solved this issue with flow-through thermal expansion tank and hydro-pneumatic tank designs that should be the new design standard for safe, clean drinking water in domestic hot water systems. Flow-through thermal expansion tanks assure a constant supply of fresh water with water treatment chemicals from the water treatment facility that can help control Legionella bacteria growth.  

Many of the thermal expansion tanks that I have inspected have been located close to the water heaters, but the piping connections needed to be connected to the cold water main ahead of the hot water return piping connection, which in some cases creates a long piping connection to the tank which is a dead leg full of stagnant water. I often find where the thermal expansion tank is connected downstream of the hot water recirculation connection allowing hot or warm water to circulate by the thermal expansion tank piping connection. This creates a long dead leg in the piping system with warm stagnant water, which is an ideal location for Legionella to form in the thermal expansion tank. Dead legs are a common source of Legionella and other organic pathogen growth in the plumbing system because of the stagnant water. 

Design professionals should be aware of the piping layout and route piping to minimize dead legs. I try to keep dead legs to a maximum of five pipe diameters from the flowing main to a shut-off valve. And I try to locate drains in long dead legs or branches after the isolation valve. If there are fixtures that are seldom used, consideration should be given to a flushing program, or consider a motorized valve for automated flushing as needed.

Make sure thermal expansion tanks are connected to the cold water line to the water heater ahead of the connection for the recirculated hot water piping, which can cause warm water that is in a temperature range ideal for Legionella bacteria growth to flow past the tee for the thermal expansion tank connection. The thermal conduction along the pipe walls and through the fluid will create an area just off the tee which is stagnant water at a temperature in the Legionella growth range. Therefore, I suggest making expansion tank tie-ins in the cold water main serving the water heater, upstream of the hot water return circulation tee and downstream of the cold water system check valve.  

I would also suggest considering replacing any dead-end style bladder-type thermal expansion tanks and install a properly sized flow-through type thermal expansion tank. This will improve the water quality in most buildings and minimize liability related to Legionella outbreaks. The illustration in Figure 1 shows that if the pre-charge pressure is too low, it allows stagnant water to reside in the tank. If the pre-charge is too high, the water will not expand into the tank until the thermal expansion pressure exceeds the pre-charge pressure. Check with the manufacturer of the tank for the correct size tank based on the system pressure and acceptance factor.

 If the pre-charge air pressure in an expansion tank is below the system operating pressure, the expansion tank will always have stagnant water held in the tank. I checked the air pressure on the tank on the day of my inspection and found that the air charge was approximately the same as or within a pound of the system pressure. Ideally, the air charge should be about two or three pounds higher than the system pressure to be assured there is never stagnant water in the expansion tank, except when there is a start-up period or heating cycle after a large draw, followed by a period with no flow and where there can be significant thermal expansion.

There will soon be many flow-through designs for thermal expansion tanks. I am aware of a few. Consider contacting one of these manufacturers for a flow-through style of thermal expansion tank. See or for more information on flow-through expansion tanks.

Ron George, CPD, is president of Plumb-Tech Design and Consulting Services LLC. Visit

Content Type: