Gravity hot-water heating

The Old House Web

By Dan Holohan

Q: How long has gravity hot water heating been around?
A: Gravity hot-water heating began quietly in the United States between 1875 and 1885. It was a Canadian import, a safe substitute for steam heat, which had been earning a notorious reputation throughout the world as being a pretty dangerous way to heat a building.

heating.jpg (5566 bytes)
This story is an excerpt of Dan Holohan's book, "How Come? Hydronic Heating Questions We've Been Asking For More Than 100 Years (with straight answers!)"

Dan is the author of a dozen books -- and is a freelance monthly columnist in five trade journals.

You can buy his books -- and find a wealth of other helpful information on heating systems -- at his web site,

Q: What was wrong with steam?
A: The trouble with steam in the early days was that it ran under pressure and frequently exploded with disastrous results. Hot-water systems, on the other hand, were open to the atmosphere, and relatively safe because the old-timers usually limited them to a high temperature of 180 degrees F. In those days, you could liken the difference between a gravity how-water system and a steam system to that of an open simmering pot of water, and a pressure cooker gone berserk!

Q: So gravity hot-water became popular because it was safe?
A: Yes, and because these systems were also easy to maintain and, most of the time, operated with little or no trouble. They had a lot going for them, and they quickly became the preferred way to heat large American homes just before the turn of the century.

Q: Is it a simple system?
A: It theory, it is. The only moving part is the water itself, but to get that water to go where he wanted, a pipe fitter had to meld the knowledge and experience of Mr. Goodwrench and Mr. Wizard. If he did his job well, the system worked beautifully. If he didn't, it became a balancing nightmare.

Q: What did a typical gravity hot-water system look like?
A: Here's a diagram of an "upfeed" system.


Q: Why did they call it upfeed?
A: Because the water fed up from the bottom (the boiler) to the top (the highest radiator).

Q: Where's the circulator?
A: There isn't one! Circulating pumps, which we use on modern hot-water systems, hadn't been invented yet so to move water from the boiler to the radiators, the old-timers depended on a basic law of physics: Hot water rises, cold water sinks.

Q: Why is that?
A: Because of the difference in density between hot and cold water. A cubic foot of water at 180 degrees F takes up about five percent space than a cubic foot of water at 40 degrees F. It also weighs about two pounds less.

Q: Is that where the term "gravity" comes in?
A: Yes! When you heat water in a boiler, it will rise up into the pipes because it's lighter that the relatively cold water in the system piping. That colder water, in turn, falls back down into the boiler (by gravity), and before long, you have a Ferris wheel flow of warm water moving freely from the boiler to the radiators.

Q: What determines how fast the water moves?
A: Several things. First, there's the height of the system. The taller the building, the quicker the flow. Within reason, of course, because if the building is too tall, the water will cool and slow circulation to the upper floors. A three-story house is the practical limit for gravity hot-water heating.

And then there's the size of the pipes. The larger the pipes, the faster the water will flow. This is because large pipes offer less resistance to flow than small pipes. It's also the reason why the old-timers used two supply and two return tappings on their boilers.

Ultimately, the size of the pipes was also the reason steam replaced gravity hot-water heat in American homes. As the years went by, steam heat became safer, but the large-diameter pipes the gravity systems required continued to be expensive.

The third factor that determines how quickly the water circulates is the condition of the pipes. When the pipes are new, they're smooth on the inside. They offer very little resistance to the slow-moving water. However, as they age, the pipes develop little nooks and crannies because of oxygen corrosion. These tiny internal burrs increase frictional resistance, and that, in turn, slows the flow and the movement of heat to the radiators. Nowadays, we usually overcome this problem by adding a circulator to the system.

Finally, there's the difference in temperature between the supply and return water. The hotter the water, the faster it circulates. However, the old-timers always kept the maximum temperature at 180 degrees F to make sure the water never approached the boiling point.

Q: Did the old-timers work with a certain temperature difference between supply and return?
A: Yes, and to get the best efficiency, they limited the maximum temperature difference between supply and return to 20 degrees F. This was a function of pipe sizing (the smaller the pipes, the greater the temperature drop, and vice versa). So on the coldest day of the year, if water left the boiler at a maximum of 180 degrees F, it would return at a minimum of 160 degrees F. This assumes, of course, that the pipe fitter followed the accepted piping practices of the day.

Q: Did the hot water take up more space than the cold water?
A: It sure did! As I said before, when you heat water from 40- to 180-degrees F, you wind up with about five percent more water than you started with. You have to have a place to put that "extra" water.

Q: How did they deal with the "extra" water?
A: They used expansion tanks.

Q: What does an expansion tank look like?
A: A typical one looked like this.


Q: Where did the expansion tank go?
A: Typically, at the high point of the system. You'll usually find them in the attic. The tank gives the expanding and contracting water a place to rise and fall.

Q: Suppose I put too much water into the system when I first fill it up. What will happen?
A: It will overflow from the tank through its vent and wind up on the roof.

Q: Can this do any harm?
A: Not to the system. It might leave some rust stains on the roof if the system is old, but that's about it.

Q: How much water should I put into the tank when I'm first filling the system?
A: Normally, you should maintain the tank at one-third full when the water is cold (there's often a gauge glass on the side of the tank so you can see what you're doing). As the water heats and expands, it will rise into the upper two-thirds of the tank and stop before spilling over onto the roof.

Q: How did they fill these tanks?
A: Some tanks had an automatic fill valve, which is very similar to the ballcock in a toilet tank. Others, the old-timers filled by hand with a valve that was either down in the basement or up in the attic.

Q: Wait a minute, if you're in the basement, how can you tell how much water is in the attic tank?
A: Good question! Chances are the boiler had an "altitude" gauge that showed the height of the water in the system. The gauge registered feet of altitude as well as static pressure.

Q: What's static pressure?
A: It's the pressure created inside the boiler by the water as it stacks up in the system piping. The gauge records static pressure in "pounds per square inch" (psi). One psi will lift water 2.31 feet (that's 28 inches) straight up, and that's where the "altitude" comes in.

Q: Do you have to take any special precautions with the upfeed gravity system?
A: Yes, if you have to drain the system, be careful how you refill it. Start off with all the radiator vents open. Then, slowly fill the system, one floor at a time. When water flows from the vents on the first floor, quickly close them all. Then, continue filling until water rises to the second floor. Shut all the air vents and move up to the third floor. Once you have all the radiators filled, fill the system to the one-third-full point in the expansion tank.

Q: Why is this method important?
A: Because there's so much air in those large pipes and radiators. If you try to fill the system all at once, and then go back and bleed each radiator, the escaping air from one radiator will cause the water to drop out of the expansion tank and the nearby radiators. This can pull more air into the system piping.

Q: What happens if I don't follow this fill procedure?
A: Usually, you'll wind up with "phantom" air problems. The air appears in this radiator today. You vent it. Tomorrow, it's in that radiator over there. You vent it. The next day, the problem appears somewhere else. It can be maddening.

Q: How does the air from the heated water get out of the system after the initial purge?
A: It vents out through that overflow pipe that sticks out through the roof. Usually, the tank sits atop the main system riser at a high point. The tank vents most of the air that the heated boiler water releases. Should some of this air wind up in the radiators instead of in the tank, it can slow the flow of heat to the rooms. Ideally, with this type of system, someone should bleed the radiators at the start of each heating season.

Q: Is there a danger of the attic tank freezing if the attic isn't properly insulated.
A: Yes, there is. And if that happens, the expanding system water will have nowhere to go. To avoid this potentially dangerous situation, many old-timers piped their tanks like this.


That second pipe, connected into the side of the tank, allows hot system water to circulate through the tank. Since the water is hot and in motion, it's much less likely to freeze.

Q: Why didn't they just go ahead and pipe all their tanks this way?
A: Because by circulating the water through the open tank in this way, the rate at which water will evaporate from the system increases. That means someone has to add more fresh water. Fresh water increases the rate of corrosion in the system and, over time, slows circulation.

Next Page (How to modernize a system, and more) -->

About the Author
The Old House Web

Search Improvement Project