Hot Water
While water heating is typically only 10-20% of a home's total energy budget, the exact amount varies widely due to differences in how energy efficient the house is in all the other respects and how much hot water the household uses. An house that isn't very energy efficient, but doesn't use a lot of hot water will see little gain from improving the hot water efficiency, while a very efficient house that uses quite a bit of hot water (for example, due to a high level of occupancy), can see sizable gains.
There are 4 things that affect hot water energy use:
Which factors are more relevant depends on the particular situation, but often personal use habits are the largest factor, followed by appliance efficiency and flow rate.
The focus of this document is on equipment efficiency.
Unfortunately there is more hype over how to save energy on hot water than there is solid facts, and alas what strategy makes sense depends on your situation--both your climate and your typical hot water usage. Before embarking on any strategy, it is best to estimate what the actual energy savings might be and then determine if the cost is worth it.
The following list summarizes the strategies:
Heat losses from the hot water system are a bigger deal in hot climates because the escaped heat subsequently needs to be removed by the air conditioner. When its cold, the heat lost costs nothing because it just supplements the space heat. Since most climates have a significant warm period, reducing heat loss is worth doing. In Seattle, because summer is typically short and cool, there is often little problem with letting the standby loss leak into the house--the energy is only wasted 4 months a year (or so).
If the average outdoor temperature is warmer than 70F, the tank is best located outside, but otherwise its generally better inside the heated space.
Typical homes waste multiple gallons of hot water a day simply due to the amount of hot water left in the pipe when the water is turned off. Standard plumbing techniques, which require larger diameter pipes make the problem worse. Much of the problem can be eliminated by centrally locating the hot water heater and locating fixtures as close to the hot water heater as possible, although other design constraints often are in conflict with this.
Pipe loss is easy to measure: get a pot or bottle marked in quart (or even gallon) increments, turn on the faucet and fill the container until the water turns hot. Typical amounts are in the 1/2 to 1.5 gallon range. The total loss in the day is a product of number of times hot water is turned on and how must is lost each time. This value could be as low as 2 gallons in a frugal household with a centrally located tank, up to 15 gallons or more in a larger house.
A simple solution is to use "home run" PEX plumbing. In this system, each fixture (faucet, shower etc) gets its own pipe, and that pipe is of small diameter, typically nominal 3/8" pipe. In a standard plumbing setup, there is usually a 3/4" pipe run for some length, and then it splits into 1/2" and then maybe down to 3/8" to go to the faucet. The issue is that 1/2" pipe holds about 70% more water per foot than 3/8" and 3/4" holds yet another 130% more.4 By using the smaller diameter pipe, hot water gets to the fixture faster and so less water is left in the line when the water is turned off. Note that this waste heat doesn't necessarily occur every time the water is used, only when the time between uses is long enough that the water in the pipe cools off.
Size | Cross section area | Volume/ft | Feet/gallon |
3/8" | .11in2 | 1.32in3/ft | 175 |
1/2" | .196in2 | 2.35in3/ft | 98 |
3/4" | .44in2 | 5.28in3/ft | 44 |
If the pipe runs are very short the issue isn't as big, but it if often the case that pipe runs are 40-50 feet, in which case you could be saving up to 1/2 gallon water on each use. In general, the homerun system costs about the same as the standard one.5
Whether a home run system actually saves energy depends on the usage pattern of the household, and the plumbing layout. If, for example, there are multiple plumbing fixtures near each other that would normally be supplied by a 1/2" line, and those fixtures are typically used at about the same time (ie a shower and sink both used back to back, or even two showers and two sinks on the same plumbing run, all used at about the same time), then running four 3/8" lines could use more hot water. However, if the main line is 3/4" and that branches to two 1/2" lines, then the difference between the two isn't that different, other than in the standard system, the water is run more often (ie by all 4 fixtures), so its more likely to be hot, which would save energy and water. On the contrary if the four fixtures are rarely used at the same time, home run will save energy. The one twist here is that codes were written for branched lines, and many fixtures are traditionally plumbed with 1/2" lines: the issue here is that you won't likely save energy with a 1/2" home run system, because the savings are in the small pipe diameter.
In summary, a home run system doesn't always save energy--it depends on specifics, but certainly when the usage pattern is sporadic, it saves energy.
Because tanks hold hot water (typically 120F), and because they often are manufactured with low insulation values (R5 for example), the use of an insulated blanket is a cheap fix. Pipe insulating, at least for the first few feet of pipe coming out of the tank is also a cheap effective fix. Further reductions can be had by buying a better insulated tank (R10 for example), and by super-insulating the utility closet in addition to the above strategies. An R-10 tank can be put under an R-10 blanket, and then enclosed in an R11 (2x4) insulated room.
On Demand heaters: these units have no tank, so have no tank loss. They are almost a drop in replacement for tank heaters, but are generally quite a bit more expensive and don't operate exactly the same.
Most on-demand heaters run only on natural gas or propane, because by comparison, electric is a low density energy source. Whole house electric demand heaters exist, but because they typically draw up to 20kw (ie about 100amps), most electrical panels can't accommodate them unless the other electrical loads are very low. More typically, electric demand heaters are used for low volume point of source locations. The natural gas units draw gas at 3-4 times the rate of a standard gas hot water heater, and as a result a gas supply line that was adequate for a tank heater may be too small for a tankless unit.
On-demand heaters use energy at a much faster rate than tank heaters, because they must heat the water instantly. The actual energy to heat the water is the same in both cases. Tank heaters deliver hot water from the top of the tank, where the hottest water has risen to, and bring cold water into the bottom. This allows them to heat the water much slower than a tankless unit. Since tankless heaters burn gas at a higher rate, they also draw air at a much higher rate, so they must be installed in a location where ample air for combustion is available, often taken from outside via a 3-4" pipe.
Unlike tank heaters, on demand units have a small delay to create hot water, so you end up waiting a bit longer for hot water to come (but you might not notice it). Additionally, there is a minimum flow rate for the unit to turn on--typically 1/2 to 3/4 a gallon per minute. If you turn the sink on at a low flow rate, the water will never get hot because the unit will not turn on.
Pilot lights typically eat up all the potential savings from avoiding tank loss, so best to choose a unit without a pilot.
Tankless heaters are much more expensive than storage tank units--typically at least four times the cost, plus they tend to have higher maintenance costs.
For anyone trying to reduce their reliance on fossil fuels (aren't we all?), the fact that most tankless units are gas fired causes most super efficient house designer to look at other options for hot water efficiency.
Tankless & Solar: On-demand units do not deal with solar pre-heated water very well. Most units specifically state they are not compatible with pre-heated water, and the ones that do only partially do: they do adjust the heat output, but even when they operate on low, they will overheat pre-heated water that is only slightly too cool (say 110F). What you temperature comes out depends on the flow rate: at the lower end of the flow rate, 110F input will yield 140-160F output. (Calculation: a minimum burn rate of 20k Btu/hr is typical. That's 333Btu/min. At a flow rate of 3/4 gpm, that amount of heat will yield a 55F temperature rise--that's 6lbs of water soaking up 333btu . At 1.5gpm, you'd get only a 27F rise.)
The standard solution for solar hot water is to either use a tempering value to reduce the input temperature of the solar pre-heated water, or one on the output to reduce the output temperature. There is little acknowledgement out there that you need a tempering valve and no consensus on where to put the valve. At issue is the fact that tempering valves lower the temperature of the hot water by typically a minimum of 5F no matter what1 (this is apparently because they never quite shut the cold side off).
All solutions with tempering valves would appear to use more energy than necessary: you're either not taking full advantage of the solar pre-heated water (even if it were fully hot, the tempering valve would lower the temperature), or you're having to heat the the water too much to compensate for the tempering valve cooling it back down (also: since some of the water at the faucet is now cold, you need a higher flow rate to turn on the tankless unit). With tempering valves with only a 5F loss, the problem is probably no big deal.
The simple solution is a to use a small tank and live with the small standby losses, and there are now a couple of products on the market that do just that using approximately 2gal tanks.
On-demand & a small tank ("Hybrid on demand"): Storage tanks deliver heat with no delay and don't have a minimum flow rate, while tankless units reduce standby loss. By using a small tank (5 gallons or less) and super insulating it, the standby loss is nearly as low as a tankless unit by itself. The tankless unit then heats the water in the small tank indirectly (via a circulation pump) making it seem that the tank is actually much larger than it is. This configuration is also likely to eliminate the problem of tankless units overheating solar pre-heated water because the overheated water is now mixed with the existing water in the small tank, thereby limiting the output temperature (note: this is just a theory, I know no one who has tried this). Since circulation pumps draw about 50w, even if it ran for 1 hr a day, the total energy use is 171 BTU, with another 500 BTU standby loss. However, these units will have all the other problems with on-demand heaters: they usually need a large gas line, a lot of fresh air, and have higher maintenance costs.
This isn't as good as finding a gizmo that tempered the output water without always dropping the temperature any at all.
Savings due to tankless heaters: it is my claim that the tankless industry overstates the savings that are likely to be obtained. See the discussion below on savings comparisions
These are passive versions of heat recovery ventilators, and work on similar principles. Also key is that water tends to stick to the surface of pipes, so that all the water going down the drain is going in a thin sheet along the pipe's surface. Like HRVs, the large amount of heat transfer is gained by transferring heat in increments. At the input to the device, the coldest water takes heat from the coolest drain water, and as the cold water moves up the pipe toward the tank it encountered warmer and warmer drain water. The heat recovery is further limited because hot water is mixed with cold at the shower down to typically 100-102F, and then loses heat in the air, exiting the shower at 90-95F.
Drain heat recovery only works for water that goes down the drain at the same time hot water is being used, eg a shower. It doesn't work for dishwashers or washing machines since they hold the hot water for quite a while.
To use drain heat recovery, the shower drain must pass near enough to where the cold water pipe comes into the hot water heater. This rules out slab-on-grade homes as well, since the drain is buried under concrete.
Unless you use a lot of water showering, drain heat recovery will not make much of a dent in your energy budget. For the frugal, two minute shower at 2gpm, the max savings is 80% of the difference between the 90F drain water and the 50F (or so) cold water. So you reduce your heating need by about 32F, times 5 gallons/shower is a reduction in heat energy of about 1300Btu. Obviously if your family uses four 20 gallon showers, your savings is 8 times greater.
Electric heaters are nearly 100% efficient, and potentially carbon neutral (more about this here), but since the US average is that 70% of the electric is generated by burning fossil fuel, a process that is typically only 33% efficient, the actual efficiency of an electric hot water tank (based on fossil fuel consumed) is more like 53%. While both natural gas and fuel oil are not renewable resources, it is more efficient to burn them directly in your own hot water heater than to have the electric company burn them. A typical new gas hot water heater will be of the "mid efficiency" type--somewhere around 82% efficient. The next step up, a "high efficiency" or "condensing" unit, will be more like 94% efficient. Buy a unit with a piezoelectric starter rather than a pilot light, as pilot lights consume a significant amount of energy.
In heating dominated climates (ie the northern US), it is best to put the hot water tank somewhere within the heated space so that the heat loss from the tank goes into the house. When installed inside the house, select a direct vent model that brings in outside air for combustion and puts it exhaust outside in order to avoid indoor air quality problems. In cooling climates the opposite is true--keeping the tank out of the heated space avoids having the air conditioner having to remove its standby heat.
Note: As a practical matter, electric hot water tanks heat water much slower than gas units--typically 3-5 times slower. As long as the house uses hot water efficiently and the tank is big enough, this isn't a problem.
Heat pumps: These are electric devices that are effectively air conditioners that produce hot water. They are 2-2.5 times more efficient than regular "resistance" hot water heater because the extract heat from the air2 --at least as long as the air in the room is relatively warm. Current units are actually hybrid heat-pump/resistance units because as room air temperatures drop below somewhere around 60F, the heat pump becomes inefficient and so the resistance element is used in addition to it. As the air temperature gets closer to the minimum operating temperature (typically 40°F), the resistance element takes over essentially all of the work.
Heat pumps need a significant quantity of air to work from (750-1000Ft3, or 100-125SF of floor space), and will cool that air considerably (and hence lower the efficiency) if there is no heat leaking into the space from somewhere. This means that where it is installed is critical.
It would seem as if you just exhausted the cold air outside, the situation would be better, but the problem is that you need make-up air, and if the outside air is colder than the exhaust air, then you've just made things worse.
Because the heat pump is attached to the tank, when the unit is installed in a cool space like an unheated basement, the standby loses will be greater than if the unit was installed inside the envelope. Tank insulation values are typically only about R10. You can add a blanket to the unit, but because of the compressor on top, this will be harder to do than the typical blanket install. If the units worked at temperatures colder than 40F, the standby losses will start to get so great as to nullify the heat pump benefit. When comparing these to air-source space heat units, keep in mind that the output temperature required for hot water is at least 120F--which is much higher than is required for space heat, hence the COP.vs.air temperature curves are not the same.
In spite of their efficiency, their recovery time is quite slow--slower even than normal electric tanks, hence you typically get a slightly larger tank. The Rheem unit, for example delivers 8700Btu/hr, an electric resistance unit about 15k Btu/hr, a gas unit about 40k Btu/hr, and an on demand will do 150k Btu/hr. Using solar preheated water will significantly increase how many gallons of hot water you'll get out of the small output of the heat pump.
Other drawback are that the units are expensive (over $1k), are fairly noisy, often quite tall and are heavier than standard units.
While no such unit exists, a water source heat pump that draws is heat from a tank of solar heated water would allow the tank to be in conditioned space. If the solar tank water gets too cold, the unit just reverts to resistance heating. The interesting side effect is that as the heat pump cools the water in the solar tank, the solar collector could potentially become more efficient, particularly in weak sunlight when the collector temperature is relatively low.
There are no ground source hot water units that I'm aware of, but it is likely someone makes a ground source heat pump that supplies both heat and hot water. Ground source heat pumps can be very expensive because they typically require a digging long trenches or drilling many holes so as to use a large area of earth as the heat sink (heat moves quite slowly in soil).
In cooling climates, specially designed air conditioners can supply hot water for free as a side effect of their operation using a unit called a desuperheater. This is because air conditioners are just heat pumps--the move heat from inside to out, so it is nearly as easy to heat water instead of dumping the waste heat outside.
Combined heat-H/W. These systems offer a small efficiency in that they attempt to use the burner more efficiently. In hydronic heating systems, one hot water tank often supplies both heat and hot water. These are usually tank systems, not tankless.
These units are not cheap (typically a few thousand dollars), but make a much bigger impact than any other strategy. In sunny climates, a solar collector will supply nearly all the hot water you need. In dark, rainy climates like Seattle, solar hot water will supply most of your hot water for about six months a year, and give you at least a little heat the rest of the year.
In the typical configuration, the solar collector heats a separate pre-heat tank, and then water from the pre-heat tank is fed into a standard hot water tank which adds whatever amount of heat is necessary to bring the pre-heated water up to the set-point temperature (typically 120F). Depending on your climate, the pre-heat tank may be considered to be the main tank, which then uses some auxiliary system to supply backup heat. If the solar tank is always quite warm, using electric demand heater might be practical, since the energy demand is much less. In most situation the only practical solution is an electric element, either in the same tank or in a separate tank.
It would seem like on-demand heaters are the ideal backup heat source for solar hot water systems, but there are complications that make it much less ideal than it seems (see the tankless section).
Solar hot water heaters come in a variety of configurations, the most common of which are "indirect" systems using glycol-water running thru the collector and a heat exchanger in the tank. Flat plate collectors are the most common, while evacuated tubes are using in cloudier climates. The major advantage of evacuated tubes is a reduced footprint on the roof.6
How much energy you'll save by using any of these strategies depends on the specific situation, but there is still a range of likely possibilities.
It should go without saying that the more hot water you use, the less standby, or any of the other losses affects your total energy use. If your heater uses a pilot light (considered desirable for people who's power goes out often) that use is likely to be much greater than any standby loss. The following table compares the energy used for the hot water itself and a range of savings related to various strategies.
Energy Use | BTU/day | Notes |
Use: 25gpd@20F | 4165 | low use(25 gallons), solar preheat to avg. 100F |
Use: 25gpd@70F | 14577 | low use, no preheat, avg cold water at 50F |
Use: 50gpd@70F | 29154 | regular use, no preheat |
Pipe loss: 2gpd | 1166 | efficient pluming layout and low usage, assume 70F rise |
Pipe loss: 15gpd | 8746 | bad layout and much use |
Drain loss: 10gpd | 2665 | recoverable heat from frugal showering |
Drain loss: 60gpd | 15993 | recoverable heat from ample showering |
pilot | 16800-26400 | 700-1100Btu/hr from www.energyideas.org |
standby loss R20/120F/70F | 1830 | quality unit, 30.5SF area@R10+R10 blanket set at 120F in 70F room |
standby loss R6/140F/50F | 10980 | std unit, 30.5SF area @ R6, set at 140 in 50F bsmt |
As you can see how much savings you get varies by the situation. If you use little hot water and its solar heated, even small losses look significant, but if you use a lot of hot water, that energy starts dwarfing all other savings.
The appeal of on-demand (tankless) units is great because they have no standby losses, and the idea of using a small, very well insulated storage tank may well alleviate their limitations. It seems likely that neither electric nor heat pumps are likely to power any whole house on-demand unit. This makes the only reasonable carbon neutral choice to be some kind of biogas--although I haven't looked into how practical this is, if it was, its a very appealing solution.
The US DOE, energy efficiency and renewable energy website guide to water heaters at: http://www.energysavers.gov/your_home/water_heating/index.cfm/mytopic=12770
American council on energy efficiency guide to hot water heaters at: http://www.aceee.org/consumerguide/waterheating.htm
AnAnother take on the topic from Home Energy magazine:
http://homeenergy.org/archive/hem.dis.anl.gov/eehem/96/960510.html
1: but there is some mixing valve that lowers the temp to 120, while still outputting 120F if the input is 120F. In 2004, I couldn't find a tempering valve that didn't drop the hot temperature less than 20F, but newer tempering valves reduce it by as little as 5F (examples: Watts 1170-m2 & MMV). Still beware, because most probably still drop the temperature by 20F.
2: Theoretically it could be ground source or use a pond, but since those systems are much more complex to install, only air source heat pumps are readily available.
3: But the cold air is still spiting into the heated space, so these ideas may not really be practical.
4: pipe sizing is very confusing, because they are measured by outer diameter (OD), inner diameter (ID), and copper pipe size (CPS, in which 1/2" pipe isn't either 1/2 ID or OD.) The calculations I did was using ID. I used a CPS table and got similar results.
5: at least that's what I was told by one plumber. Its more pipe, but often less labor.
6: evacuate tubes can collect more heat on cloudy days, but in practice it does seem to result in much additional hot water.