GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted
Building Science

Beyond Water Heaters

Water heating is a system, not just an appliance

Water heating is much more than just a water heater.

What’s the first image that pops into your head when you see the term “water heating”? It’s probably a water heater tank. Depending on your interests or what you have in your home, perhaps it’s a tankless or solar water heater. For most people, hot water is all about the device that heats water. Water heating is a system, however, and the water heater is merely one component.

Water heating inputs

The water heater has two inputs: water and energy. The hot water that comes out of the water heater has to travel through a distribution system to reach the various fixtures—faucets, showers, dishwasher, etc.—where the hot water is used. The fixtures have controls. And then you have the people, an important part of any water heating system.

On the input side, the water varies from place to place and from season to season. The mineral content or water softening chemicals can affect the life span and maintenance of a water heater. The temperature of the water supply affects the energy use. In cold climates, the water coming into a water heater is colder than the water being heated in warm climates. The incoming water temperature is close to the average air temperature for a location, but it varies with location of the pipes and with the seasons. In winter, the water coming into a water heater is colder and needs more heating.

The type of energy input is an important choice. It could be electricity, natural gas, propane, wood, solar radiation, or some other type. The fuel you choose narrows your options for a type of water heater. Within types, though, you can choose low, medium, or high efficiency. You also may get to choose from models that store hot water in a tank or those that heat water on demand. I say “may get to choose” because a tankless water heater isn’t always an option. You can’t do solar water heating on demand, for example.

Water heating outputs

The output side of the water heater is the part that’s gotten short shrift. When you choose combustion of a fuel to heat your water, there are two outputs: exhaust gases and hot water. Removing the exhaust gases is a critical health and safety issue. Sealed combustion or direct-vent water heaters are the safest, but the majority of gas water heaters use a natural draft exhaust gas flue, making it the appliance most likely to put carbon monoxide into your home’s air. Consider carefully when buying a water heater. And remember that the exhaust gases are part of your water heating system. Choosing an electric water heater avoids the exhaust gas problem and uses a fuel that keeps getting cleaner.

Pay attention to the hot water distribution

For electric and solar thermal water heaters, the sole output is hot water. Sadly, the process of delivering that hot water is greatly flawed in most homes. The hot water distribution systems still being installed today are based on out-of-date ideas and technology. Since the 1992 Energy Policy Act lowered the upper limits on water-flow rates for different types of plumbing fixtures, hot and cold water lines have been oversized. The result is longer wait times for hot water, which wastes water and time. It also wastes energy because it strands more hot water in the pipes when the tap is turned off than would be stranded with right-sized pipes. Gary Klein is the guy who opened my eyes on this subject, and I encourage you to read his articles and watch his videos.

Also of great importance for the hot water distribution is the location of the water heater and fixtures. The greater the distance between the source of hot water and where it gets used, the greater will be the wasted water and energy. The most efficient hot water distribution system will be the one with the shortest runs in addition to having right-sized pipes. That setup in the vintage Rheem water heater ad [main image] shows an efficient placement of the water heater for laundry room hot water.

The convenience problem, of course, can be solved with a recirculating pump. A continuous recirculating pump that keeps hot water close to every fixture 24/7 will waste a lot of energy. Putting it on a timer so that it runs only during the high-use times can reduce the extra energy usage. A demand-type recirculating system is better. Just push a button, and your shower will be hot in a few minutes without wasting water or energy.

Down the drain doesn’t have to mean lost forever

Drainwater heat recovery improves water heating efficiency
Drainwater heat recovery improves water heating efficiency.

But wait! There’s more. The water heating system also includes the drains at each fixture. When you heat water, use that heat briefly while showering or washing dishes, and then send it on its merry way down the drain, you’re sending energy that you paid for down the drain, too. It’s possible to recover some of that heat in a clever device called—even more cleverly—a drain water heater recovery system. It’s simply a copper pipe wrapped around a drain, absorbing heat from the drain water. Cold water runs through the copper pipe and gets pre-heated on its way to the water heater. Showers are the hot water fixture that can best take advantage of this energy-saving device because the hot water runs down the drain for longer periods. Do your homework, though. You’ll need space beneath the bathroom and enough showering to make it cost effective.

So there’s your 10,000-meter look at hot water. Water heating is a system, not simply an appliance that heats water.


Allison Bailes of Atlanta, Georgia, is a speaker, writer, building science consultant, and founder of Energy Vanguard. He has a PhD in physics and writes the Energy Vanguard Blog. He is also writing a book on building science. You can follow him on Twitter at @EnergyVanguard.


  1. luke_p | | #1

    Has anyone done a life cycle analysis on a copper drain water recovery system to determine how much drain water energy has to be recovered before the energy cost of fabricating the device itself are equal? I mean it takes a lot of energy to mine ore, turn it into usable copper, then fabricate the actual device in a shop, not to mention transport it to the final destination. For some reason my gut tells me it will never be able to last long enough to account for all that energy but maybe my gut is off.

    1. GBA Editor
      Kiley Jacques | | #2

      Fascinating question, Luke. Life cycle assessments don't get near the attention they should in terms of weighing the impact of processes.

    2. tommay | | #3

      Good question. If you look at the picture in the article, it shows a vertical pipe. Hot water flowing out that pipe doesn't completely fill that pipe so heat transfer is minimal. People are better off learning how to use the volume control on their showers to save hot water. Reducing cold water flow that mixes with the hot water gives longer, hotter showers while saving water and energy.

      1. GBA Editor
        Allison A. Bailes III, PhD | | #5

        Tom, hot water doesn't have to fill the pipe. The way water travels down a drain like that is mostly by clinging to surface of the pipe.

        1. tommay | | #7

          Allison, that was my point. Only a fraction of the surface area would be "wet". Even on a horizontal drain, the water only covers a small fraction of the surface area .....not to mention the contact spacing between the copper coil and the conductive properties of the drain material.
          The best way to save money on energy is to use less.

          1. entropic | | #11

            Great Article!

            Tom, this is more of a reply to your first conjecture that “Hot water flowing … so heat transfer is minimal.” The University of Waterloo has pretty compelling evidence that these products are effective (the longest models have over 70% heat recovery):


            Their data show that effectiveness goes up with increasing diameter, which indicates to me, a lay person, that maybe the draining water really does cling to the sides of the pipe as Allison, a physicist, suggests. Increase the heat exchange surface area by either diameter or vertical length of pipe = increased effectiveness.

            Is your interpretation of their data different?

          2. tommay | | #20

            Matt, I read their report. Pretty bogus as far as real world scenarios go. They use a closed loop system, removed all the air from the system, so their piping was completely full of water and temperatures were kept constant. Of course in that scenario the heat exchanger worked okay.
            Any company that does a "test" isn't going to show how bad a product is and will do whatever they have to do to show their product works as they say, hence their experiment setup. To say that it is more effective with larger diameters makes no sense at all since you are creating less contact area with the water inside the pipe thus less area for conduction, unless you completely fill the pipe and maintain a constant temperature as they have done in their tests.
            If you believe that draining water clings to all sides of a pipe perhaps you should do a little experiment yourself and pour some water through a pipe and see for yourself how the water flows within.

          3. charlie_sullivan | | #21

            Tom, that report doesn't have all the details, but it does have one key thing you might have missed: "A shower was simulated by allowing the water to flow into a small catchment basin." The drain water flow is just by gravity from the drain of that basin. That part of the system is not pressurized and is open to the air. The drain pipe is not filled.

            Here's a thesis that explains that setup in much more detail:


            In particular I refer you to the discussion from the bottom of p. 35 to around p. 40, and the photographs, which demonstrate the water clinging to the surface just like Matt described.

            If you want independent confirmation that this stuff works in the field, not just in the lab, here a report from actual installations:


          4. entropic | | #22


            I did not know that you were a professional in this area. I am not. I’m still trying to figure out what I believe in terms of water’s behavior in a vertical drain pipe. Thanks for helping me refine that belief. I think you’ll find this video interesting unless I’m just naively falling prey to manufacturer propaganda:

            Or maybe this one from a homeowner in the real world:

            If there’s another explanation for that infrared video above, I’d like to hear it.

            Maybe this excerpt and Figure 2 from the attached peer reviewed manuscript would also be of interest:
            “The ideal performance of DWHR systems relies on the formation of a falling film of water within the drainpipe. In other words, the water sticks to the walls and falls as an annular film, covering the entire inner surface of the pipe. An example of this is shown in Figure 2, where a camera was mounted vertically inside a DWHR system. The falling film maximizes the area for heat transfer, while minimizing the thickness of water through which the heat must be transferred to the walls.”

            Sounds compelling to me…I think I do believe water clings to the side of a pipe as it falls, until better evidence to the contrary is present.

          5. tommay | | #24

            Matt, the comment by mr smith on that second video link was spot on. Why isn't this thing insulated is a great point. As I mentioned pour some water down a pipe from the inlet of a t-wye, you will see the water does not tend to flow around the entire surface of a vertical pipe. Even on a horizontal pipe it is usually less than half full. Even if it did how much energy do you think can be stored and transferred from a thin film of flowing water?
            Funny on that first video link they show a hot water heater that isn't installed correctly. (I know it's just a mock up but they should at least get it right)
            Charlie, do you really think it would take 179 pages to explain T in and T out? And did you read the conclusion? Or see who benefited from testing.
            As always, buyer beware and do your own research.

          6. maine_tyler | | #27

            "how much energy do you think can be stored and transferred from a thin film of flowing water?"

            Thin film would be ideal for transfer (not sure what is meant by storage). Highest surface area to volume ratio. A full pipe would be bad for transfer. Same general reason air cooled engines have cooling fins.

          7. tommay | | #36

            Tyler, you say a full pipe of hot water would be bad for I guess all hydronic heating systems are working incorrectly or not up to par when all the air is purged out.

          8. entropic | | #45

            I didn’t have to do as Tom suggested (“pour some water down a pipe…water does not tend to flow around the entire surface of a vertical pipe”). The local children’s museum had a great demonstration that I thought I’d share. Saw this today and it reminded me of this post’s reply discussion.


            This demonstration/toy has several runs of horizontal (slightly sloped) clear pipe and two vertical drops that you can just stand and observe. The horizontal pipe going into this vertical segment was a little less than half full. I observed this for quite a while before filming it and would estimate that the vertical pipe has a circumferential falling film of water inside it by about 12-18” after the 90 elbow diverts flow from the horizontal pipe. Not that this will change anyone’s mind that is already made up regarding the existence of falling films of water…never to old to learn something at a children’s museum. So maybe the first (highest) foot or two has < 100% of the pipe acting as a heat exchanger, but the rest should be more likely to have a circumferential falling film of water?

    3. GBA Editor
      Allison A. Bailes III, PhD | | #4

      Luke, I don't know if such an analysis exists but would be interested in seeing the results, too. One thing that would make it better than you indicate, though, is that about 30% of copper is recycled. Still, drain water heat recovery won't pass the test for many homes because they don't get enough hot water through the drain.

      1. luke_p | | #6

        I could see drain water heat recovery being useful in some commercial or industrial applications, maybe even larger multi-family buildings (high-rise apartments) but can't see it being worthwhile in a single family residence. Better off focusing on aspects with high gains in efficiency like air tightness or a better air exchanger.

        1. GBA Editor
          Allison A. Bailes III, PhD | | #8

          I agree. I thought I had revised the article to say that but apparently I didn't.

          1. thierry19 | | #9

            Paybacks range between 2.5 and 7 years


            If so, definitely worth the investment

      2. charlie_sullivan | | #10

        I'm more optimistic than Luke about drain water heat recovery. For a household with electric water heating, if you consider upgrading to a heat pump water heater, or adding a drain water heat recovery system, both entail a bunch of raw materials, fabrication, and shipping. The drain water heat recovery system doesn't save as much energy, but it also weighs about 1/10th what a HPWH weighs, so even if some of that materials in the HPWH entail lower emissions per unit weight, it's almost certain that the overall lifecycle emissions are considerably higher for the HPWH, even if you don't leak any of the HFC refrigerant (a potent greenhouse gas).

        1. paul_wiedefeld | | #17

          Bingo - I'd promote a DWHR over a heat pump water heater any day. Plus manufacturers actually know how to build a DWHR, HPWHs have shaky reliability.

  2. pjpfeiff | | #12

    A somewhat related article I would like to see is, if you currently have an old water heater of type (A), what do you need to do in order to be ready to switch to type (B) when your water heater dies? For instance I have a natural draft gas water heater. If I wanted to switch to a HPWH I understand I would need to ensure enough height in the basement, a place for condensate to drain, a large enough connected air volume, and ideally a 240V outlet (although I think I recall an article about 120V models?). Or to go to a direct-vent I'd need either a separate duct for input or a larger hole in the wall for concentric ducts. Maybe that's half the article...but it'd be nice if someone who knew what they were talking about wrote it. Or maybe that article exists and I haven't found it?

    1. GBA Editor
      Allison A. Bailes III, PhD | | #14

      Paul, that's a great suggestion. I went through the transition from a gas water heater to a HPWH myself and have added this to my list of topics to write about.

      1. irene3 | | #25

        I ended up bailing on a heat pump water heater last time because of the difficulty of coordinating the electrical work and the plumbing work. As I recall, no single company would both sell you a HPWH and do the installation, and it was also not as simple as just getting the electrician in first to prepare the area and then getting a plumber. I am now considering just getting a regular electric water heater the next time we need a new one. There are only two of us, and our usage is quite low. Currently the gas water heater is our only gas appliance, so getting rid of it means getting rid of the base monthly charge to have gas to the house, which is about half the bill. So even if our cost for actual water heating doubles, we'd be out about the same amount of money (and by then I expect we'll have solar panels anyway). But the replacement dance with the electrician and plumber might be just as complicated, dunno.

        1. GBA Editor
          Allison A. Bailes III, PhD | | #26

          Irene3, I don't understand why you think it's so complicated to get the electrician out and then the plumber. That's what I did. I did do the last bit of wiring to connect the wires to the water heater myself. If you weren't doing that, you might have a gap between the plumber finishing and the electrician doing the final wiring and startup. But getting a standard electric water heater is a good alternative plan.

          1. irene3 | | #31

            The regular way to replace a water heater (for us unmechanical types - you really, really don't want me doing your wiring or plumbing, and that goes triple for my husband) is to get one company to sell you a water heater and the installation thereof, and everything gets done in one day. But if you're going from gas to electric, well, you have two and possibly three sets of people involved (depending on who you get the HPWH from), and it all seems super complicated and very unlikely to get done in one day. I was hoping to find a plumber who would either coordinate with an electrician themselves (I suppose that was unrealistic) or who would at least say, yeah, call XYZ Electric, they do all these jobs for us, they know exactly what to do. But if we plan ahead we shouldn't be under nearly as much time pressure, and could get wiring put in way ahead of time. We were thinking of calling in an electrician to do the various small jobs that have been piling up (e.g., we have some areas where we planned to install light fixtures and have not yet done so) and could just add the prep work for water heater installation to the list. The final hookup would then be the only wrinkle, but presumably that wouldn't take much. But new commercial and multifamily construction in Seattle is going to require HPWHs pretty soon, so there should be a much bigger market for them, and with any luck they will get cheaper and more outfits will be prepared to install them.

  3. charlie_sullivan | | #13

    That's a really good idea. It's really hard to argue for anything other than a heat pump water heater in most cases, so I'd focus on that. I think your list of issues is correct:

    1. Electric. Last we had a discussion here of 120-V HPWHs, they were on the drawing board but not available. However, Rheem has a unit that needs only a 15-A 240 V circuit. That can overcome most concerns about available electrical capacity, and is easy to run wire for. It could make sense to have an electrician install that circuit ahead of time, terminated in a box near the intended location. 30 A is better if that's as easy to do from a service capacity and wiring perspective, but 15 A can work.

    2. Height. The shortest Rheem I found is 62" high, but it would make sense to consider how many gallons you need and check the height for that size.

    3. Air volume. If you don't have enough, you can duct it. See this article:

    4. Condensate drain. If you don't have an appropriate drain available,. a simple, inexpensive condensate pump can do the job just fine.

    So of those four, only the wiring is something you might want to work on in advance. I think the rest are more about reassuring yourself that there's not a problem so you will have confidence in making that decision under pressure.

    But I agree, a full article on this would make sense.

    1. GBA Editor
      Allison A. Bailes III, PhD | | #15

      Charlie, I think you're right about the availability of 120 V models. Regarding the current capacity of the circuit, it depends on which model of HPWH you get. I have an 80 gallon model, which required a 30 A circuit. As you suggested, I had the electrician set up that circuit for me, putting a box next to where I installed the HPWH a month later. He left a flexible conduit whip for the wiring, which I connected myself after the plumber installed the water heater.

    2. GBA Editor
      Allison A. Bailes III, PhD | | #16

      One more thing: You do need to put in the required circuit capacity, but you don't have to use it. I've got my HPWH set for heat pump only mode, so it never hits electric resistance heating elements. I don't have monitoring set up on my circuit yet, but the current is much lower in heat pump mode.

    3. this_page_left_blank | | #18

      The biggest issue with HPWH right now, aside from availability and cost, is noise. I had this on my to-do list until I heard the recent reports of excessive noise in the latest generation of Rheem units.

      1. GBA Editor
        Allison A. Bailes III, PhD | | #19

        Trevor, I don't know about the models that have come out since 2019, but I've lived with a Rheem 80 gallon HPWH and can barely hear it. It's in the mechanical room in the basement, not far from where I'm sitting now. Currently, there's no door on the mechanical room, so I do hear it when it runs, but it's far from what I'd call noisy. Once I remodel the basement and have a door on that room, I won't hear it at all.

        1. Robert Opaluch | | #28

          We recently installed a Bradford White HPWH and I was surprised it wasn't louder in operation, after hearing all the complaints about HPWH noise. However, its located in the basement (not located next to a quiet bedroom which might bother people at night). It sounds like a typical fan. We never heard it upstairs, despite it being 5' from the open stairway with a hollow core door at the top of the stairs in the center of the house. Even though we wanted to notice it running, to see how loud it might be. Its quieter than the previous oil-fired boiler heating-HW combo unit. The STC rating on HPWHs isn't high. But obviously louder than a simple electric resistance hot water heating tank.

    4. charles3 | | #32

      Any idea how many CFM of air the HPWH fan moves? Guessing since Martin says 1/10 of a ton it's 40 CFM.

  4. monitor_top_fan | | #23

    Seeing the black-and-white photo reminded me of an advertisement I always liked that was embedded in an episode of The Secret Life Of Machines:

  5. Robert Opaluch | | #29

    Allison this is a great article. Its difficult to get consumers to understand the complexities involved in systems.

    My opinion is that most important part of optimizing the hot water system is keeping the distance of all of the hot water plumbing runs minimized, through close grouping of the bathrooms, kitchen sink, and hot water heater. Not distributing them all around the building. From an engineering perspective, it makes more sense to co-locate everything than to invent, design, build and maintain a complicated hot water recycling loop. At least designers should attempt to co-locate hot water appliances and minimize the hot water delivery time. Or consider a hot water temp boost or small tank at a more remote location.

    The house I designed and built had two bathrooms stacked vertically, next to the hot water heater and washer-dryer, with the kitchen sink and adjoining dishwasher on the other side of the 10' long, 2-story plumbing wall. No hot water pipe run was longer than 10' from the hot water heater output. Almost no wait time. Minimal wasted hot water. This is not difficult to design and build, one (or maybe two) hot water system locations in the home. This should be one design/review step in the home design process.

  6. JonathanBeers | | #30

    In Madison, Wisconsin, incoming water temperature peaks in Oct/Nov, and is lowest in April/May. Not a big deal, but noticeable if you're trying to subtract water heating energy use from use for heating.

    Link to ENERGY STAR HPWH checklist:

    Link to HPWH research in Minnesota as of 2015. Anybody know of more cold-climate field research?

    Drain water heat recovery works a lot better in multifamily and commercial settings than in single-family. Drain water heat recovery study:

  7. charles3 | | #33

    I have a comfort vs. efficiency question: A HPWH will be located between a living room and a kitchen. In the winter, the living room is much warmer than the kitchen because it contains a wood stove. Should the HPWH draw air from the living room or the kitchen? If the living room, the HPWH will run more efficiently, but exacerbate the temperature difference between the kitchen and living room. If the kitchen, the HPWH will run less efficiently but help to reduce the delta t. What say you?

    1. charlie_sullivan | | #34

      There's no one answer to a comfort vs. efficiency question, but I'd lean towards drawing from the kitchen and dumping in the living room. That will suck some living room air into the kitchen, so the air you'll be using will be plenty warm for the HPWH to work well.

      The other option is to draw from and dump to the living room, with the vents spaced as far apart as possible.

      1. charles3 | | #37

        Charlie, thanks for your perspective. Now let's look at summer: It's more important to have cooling and dehumidification in the kitchen than the living room, because the kitchen appliances give off heat, and because people work in the kitchen. So which way do you lean now?

        1. charlie_sullivan | | #38

          Now you are tempting me to recommend a duct monster with all possible configurations and eight different dampers that need to be adjusted four times a day for optimum performance!

          Actually, for the summer scenario you describe, drawing from the kitchen and delivering to the living room sounds good for removing humidity from the kitchen air. As long as it isn't too close to the stove and sucking grease droplets into the system and clogging things up. So that seems like a pretty good option all year.

          1. charles3 | | #39

            Interesting, Charlie. I would have guessed exactly the opposite. It is as if you were specifying an HVAC return for a room with a significant cooling load rather than a supply.

          2. charlie_sullivan | | #41

            My thinking is in part that comfort in the living room is more important than comfort in the kitchen, but maybe if people actually spend more time in the kitchen than the living room. And grease in the system is a real concern.

        2. Jon_R | | #40

          If there is a large opening between the rooms, then my answer is mostly "it doesn't matter".

  8. user-6998645 | | #35

    Very good and timely article Allison. I'm a huge fan of demand-type recirculating systems where a button or motion sensor activates a pump and sends tempered water from the hot water line back down the cold water line (where a dedicated return line is not plausible). I have tried unsuccessfully to get energy efficiency programs in Canada to implement these devices, but the one hurdle I get is from government health departments who state the risk of propagating Legionella is too big a risk. Yet, I am aware these devices are common in many U.S. programs. Can anyone send me any references to show what the risk is (e.g. scientific papers)?

    Also, I was until recently heavily involved from the get-go with the testing and certification process for drain water heat recovery units (CSA B55.1 and B55.2). As a plumber and an energy efficiency geek (proud of both), I helped to ensure that the test process and results made practical real world sense. Nonetheless, whether a person believes how draining water down a vertical pipe can influence energy transfer (the pipes are never filled and there is full air movement for ventilation purposes in the test drains) really doesn't matter as ultimately energy in versus energy out is what counts. By the way, the CSA B55 Standards now include test procedures for horizontally installed DWHR units.

    Conrad Baumgartner

    1. Jon_R | | #42

      Data suggests that people died due to past "turn down your water heater to save energy" advice. IMO, intermittent recirculating systems shouldn't be recommended until it's clear that they don't have similar results. It's entirely possible that the best place for stagnant, previously heated water is down the drain.

      1. charles3 | | #43

        Jon, I assume you are referring to Legionnaire's disease. Would you share the suggestive data? Why doesn't normal chlorination stop Legionnaire's? IIRC, Legionnaire's was first discovered in air conditioner condensate, not potable chlorinated water.

        1. this_page_left_blank | | #44

          From the CDC:

          "Warm temperatures also make it hard to keep disinfectants, such as chlorine, at the levels needed to kill germs like Legionella."

          Note that this statement is within the context of hot tubs, which would have a far higher concentration of chlorine than the domestic water supply. So we can extrapolate from that that chlorine is not an effective counter measure (not to mention there's still a lot of people on private wells with no chlorination at all).

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |