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

Community and Q&A

Drain Water Heat Recover best layout

orange_cat | Posted in General Questions on

I have read what I could find about DWHR (including past discussions). I am curious about the current experiences/thoughts.

I have 2 showers on the second floor, and 1 shower (used least frequently in theory) on the ground floor. Mechanical room is 2 and a half feet below the ground level. Verything is stacked on the same side of the house, but there are two separate stacks (1 main shower, over kitchen, over ground level shower) and 1 other second floor shower, over half bath, over mechanical room.

It seems to me that gravity-fed DWHR for the two upstairs showers plus gravity fed DWHR from all drains for the incoming cold water supply to the domestic hot water tank is the best solution (a total of 3 units).

I also have hydronic heat, but likely with a separate tank from the domestic hot water, so not sure if it also can be fed through the “all drains” DWHR.

Does this make sense? Should I just install 1 DWHR (all drains to the hot water tank cold supply)? None?

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. Expert Member
    BILL WICHERS | | #1

    I'm not so sure a bunch of DWHRs spread over multiple drains and presumbably with the cold water line running series through all of them is going to be very effective. With all that extra supply pipe, you are going to have more pressure loss (unless you upsize the pipe a fair bit), and also more heat loss along the additional length of pipe. While I have no data to prove it, my gut feel here is that the downside to the significant increase in supply pipe length is likely to offset a good chunk of the benefit you would otherwise have from the multiple DWHRs.

    You can't use a DWHR for a shower directly, at least not the way I think you're thinking. If you were thinking to run a DWHR on the drain from the two showers, and use that on the hot water line going to those two showers, that won't work -- DWHRs are passive heat exhangers, so they can only transfer some of the heat from a warmer to a colder fluid. That means you can recover some heat by pre-heating the COLD water line going to a water heater, but if you try to use a DWHR on a HOT water line, it will work in reverse, robbing heat from the hot water to warm up the water going down the drain.

    If you have only one hot water heater, you probably will be best off with one DWHR placed so as to have the majority of the warm waste water pass through it, and use it to preheat the cold water supply to that water heater. If you have multiple drain lines, you're probably best off placing one DWHR in whichever drain you expect will take the highest amount of warm waste water so that you can maximize the thermal energy it can recover over time.


    1. orange_cat | | #2

      Thank you - that is just the type of feedback I am looking for.
      We have not laid out any pipes yet, but are about to. I was envisioning using DWHR on the cold supply lines to the shower with the shower drains - thinking that this lowers somewhat the need for hot water for the showers.

      And then the waste from bathtubs, dishwasher, etc is to use water temp that is at least room temp or higher to preheat incoming water supply into the hot water heater (which much of the year is colder than room temperature in 5A). I am cognizant of the pressure loss, which is why I am hesitating. I also cannot find any reliable reports of true benefit (rather than lab test results in idealized conditions that are not easy to replicate). I am aware of NRCAN website, but I see also that Ecodrain for example seems to have ceased selling their units at the moment.

      1. Expert Member
        DCcontrarian | | #3

        Theoretically that works -- the warmer the incoming cold water, the less hot water you have to put into the mix to get the shower water the same temperature.

        I don't know why heat recovery drains haven't been widely accepted. There must be a reason.

        1. Trevor_Lambert | | #14

          It's required by code here in Ontario. I wonder what they do if you build a single story, slab on grade, or if your shower is on the main floor of a two story.

          1. Expert Member
            MALCOLM TAYLOR | | #16


            From the Ontario code: "Drain water heat recovery is not required where there are no stories beneath the shower in the home.

          2. Trevor_Lambert | | #24

            Ah, much simpler than I thought.

        2. Expert Member
          BILL WICHERS | | #17

          >"I don't know why heat recovery drains haven't been widely accepted. There must be a reason."

          It's probably at least in part economic. I have my doubts DWHR units will ever pay for themselves, especially with today's much higher copper prices. The problem is that there is literally pennies worth of heat being recovered with these units, so if you assume $800 to purchase the DWHR, and 3 cents worth of energy recovered per shower, you need to take about 26,700 showers to pay for the unit. If you assume one shower per day, that's a 73 year payoff, without allowing for interest, which would make the payoff even longer.

          DWHRs DO work, but they are far into the "diminishing returns" area in terms of increased energy efficiency.


          1. Trevor_Lambert | | #25

            Where did you get the 3 cents figure from? If we assume a 15 minute shower, $0.10/kWh typical low flow shower head, that's $0.34 per shower. A good DWHR unit is about 57% efficient, so that is saving 19 cents per shower. That changes the payoff to 12 years.

            Why one shower per day? Most people shower daily, and most homes house multiple people. I'd say two showers per day would be very conservative. That reduces the payback to 6 years.

            You also get the benefit of larger effective supply of hot water. Hard to put a dollar value on that, but it's not zero.

            If you compare the economics of the DWHR to that of a heat pump water heater, I think it favours the DWHR. The water heater is probably at least $2000 installed, will almost certainly be inoperable and unrepairable once the 10 year warranty runs out, and you get the fun of dealing with the extra noise and service calls when it fails. The exact savings are hard to estimate, since it depends a lot on your climate, where it can be installed, etc. It likely saves more than the DWHR, but not by a really wide margin.

          2. Expert Member
            BILL WICHERS | | #26

            Reply to post #25:

            I made up the 3 cents number, it's not what I would call "real data", and certainly not based on any actual measurements, but it's probably not that far off from reality.

            You mention a 10 cent/kWh low flow shower head. I'm not sure exactly what that means, but I thought I'd look up some real numbers. 2.5GPM is a typical maximum flow rate for a shower head. To keep things easy, I'll convert that to metric units, so that's 9.46 liters per minute, or 567.6 liters per hour of flow.

            Let's assume the cold water comes in at 50*F, which is 10*C, and we have our hot water heater set to 140*F, which is 60*C. That means we raise the temperature of the water 50*C, and we do that to 567.6 liters per hour to handle the flow rate of that shower head. Water weighs 1kg per liter, and it takes 4,190 joules of energy to heat 1kg of water 1*C. That means we need to heat 567.6 liters of water 50*C every hour, and it takes 118,912,200 joules to do that (567.6kg/L * 50*C * 4,190J/kg). A joule is one watt*second, and 1 watt*hour is 3,600 watt seconds, so 1kWh is 3,600,000 joules (1w*s * 3,600 seconds/hour * 1,000 w/kW)

            That means we need 33 kWh to heat up enough water to keep that shower head running for one hour. Since a typical shower might be 15 minutes as you say, that means we need about 8.25 kWh to supply the hot water for that shower. Google says the average US electric rate is 13.72 cents per kWh, so we spent about $ 1.13 to heat the water we used for that 15 minute shower.

            I'm assuming an electric resistance water heater here, just because it makes the math easier (1kWh of energy into an electric resistance water heater pretty much entirely goes into heating the water). With a heat pump water heater, the energy cost is even less, since the heat pump uses less energy to heat the same volume of water. Natural gas fired water heaters will also be cheaper to run. So the absolute best case scenario is recovering ALL $ 1.13 worth of energy from that shower. But....

            A lot of heat was lost IN the shower, and IN the supply pipes and IN the drain lines BEFORE it got the DWHR. Most of it is probably lost in the shower itself, steaming up your bathroom through the "heat of vaporization" that made some of that hot water go off as steam (similar to how an evaporative cooler works). Let's assume we lose 20% of the energy there, and let's assume your 57% efficiency number for the DWHR. Let's assume another 10% is lost in the pipes. What's left to recover? about 41% of the total input energy, which means we recovered about 46 cents worth of energy with the DWHR.

            That means for an $800 DWHR, you need to save It will take about 1,739 showers to break even, or a bit under 5 years if one person showers every day in the house.

            So the short of it is there is NO question DWHRs DO work. The issue is that they are NOT particularly cost effective in many situations. This is unfortunate, since I love the idea of recovering energy that we already paid for that will be "wasted", but it's surprisingly difficult to recover waste heat like this in a useful and cost effective way.

            BTW, I think my math is accurate, but it's surprisingly difficult to type all that stuff in the little box provided on GBA without losing track and making a mistake. If I made a mistake, please let me know and I'll fix it while I still have time to edit my post.

            Edit: this_page_left_blank noticed that I mistakenly forgot to covert Gallons Per Minute (GPM) to Gallons Per Hour, which is what is needed for my calculations to really work out. That did make a pretty big difference! Thanks for catching that! Now the numbers look a little high though...


          3. Trevor_Lambert | | #27

            I left out a comma, and I guess a couple of words. $0.10/kWh energy rate, and a typical low flow shower head. I offloaded my math to an online calculator, where I just plugged in 15 minutes for shower length, and left defaults for flow rate (7.9l/min) and energy cost. That spat out the $0.19 per show figure, and I used your numbers of 73 years and 3 cents per shower to prorate the payback periods.

            I do think you misplaced a decimal point somewhere, though I didn't go through it with a fine toothed comb. 0.14kWh for a shower doesn't seem right, simply based on energy monitor readings I've looked at. I think that would mean that a 15 minute shower would only be equivalent to a couple of hours of standby losses in the tank, and wouldn't even trigger the heating elements to turn on. On second glance, didn't you forget to convert l/min to l/hr? You said "That means we need to heat 9.46 liters of water 50*C every hour" That should be 570 liters, no?

          4. Expert Member
            BILL WICHERS | | #28

            OK, I think I see the issue. Google says the average shower temperature people like to take (it's amazing what you can look up!) is 105*F. With our 140*F "hot" water, and 50*F "cold" water, it's about a 60%/40% split between hot and cold water to get to 105*F water. That means we are only heating about 60% as much water as I'd allowed in my original calculation, so that $1.13 per shower ends up at about 67 cents per shower. 41% of that means we can recover about 27 cents per shower, so we have a 2,963 shower payback period, or a little over 8 years with one shower per day.

            That seems like a bit spendy for a shower, but I don't see any math errors. If anyone else does, please let me know.


          5. Trevor_Lambert | | #34

            Your 41% recovery is on the low side of efficiency for units out there. I did price these out, and the smaller, 42% variety is closer to $400. Even the 57% one isn't $800, but if you allow for an hour to install, it's pretty close. Having said that, real world recovery is going to be less than rated efficiency, since we have to account for heat lost before the drain water gets to the unit. I doubt there's much in the pipe, unless it's a long run, but going through the air (any steam represents lost energy) might be significant. I'm not sure about the shower floor, since it will equalize to the water temp pretty quickly.

            I've been tempted in the past to test my DWHR, but never got around to it. 8 year payback is in the ballpark I figured when I did back-of-napkin calculations during the house build. That seems pretty good to me for an install-and-forget item.

        3. Expert Member
          MALCOLM TAYLOR | | #30


          I think the code simply conceded that in those situations there were no good options.

  2. plumb_bob | | #4

    These systems consist of a coil that wraps around the drain pipe, with a compressor for the heat exchange? I have never seen one in the wild.

    1. Expert Member
      DCcontrarian | | #6

      No compressor. It's all passive. There is a section of copper drainpipe that is wrapped with a coil of copper tubing. Water flowing down the drain warms the copper drainpipe, which warms the tubing and the cold water flowing through it.

  3. jollygreenshortguy | | #5

    I wonder if there is any hard data on the actual benefits of attempting to harvest heat from drain water. I have some doubts but no real data to support my doubts either.
    At best hot water is draining for a perhaps a few minutes (shower) in relatively small quantities, or very briefly, perhaps a minute, in larger quantities (draining a full bathtub). During that very brief time frame the heat has to be transferred to an incoming water supply and then that heat has to be stored for possibly a considerable time before it is needed again.
    I've seen one installation and it looked like a rather expensive bit of kit. Before I'd buy into something like that I'd want to see some hard numbers of the real efficiency of the system.

    1. Expert Member
      DCcontrarian | | #7

      They're really meant for showers, where there is always warm drainwater running whenever the hot water is running. And the cold water is running too. The idea is that cold water is being warmed at the same rate it's being used.

    2. Expert Member
      Akos | | #8

      Google is your friend for this. They definitely work and are surprisingly effective for showers. Showers are the biggest hot water uses in most households, so good way to reduce your overall energy use.

      DWHR is a also great way to stretch the capacity of a water tank. DWHR is usually cheaper and more efficient than a larger capacity tank. Since it is a passive device it is pretty much maintaince free except maybe in some areas with very hard water.

      As for the OP, you want a single DWHR unit, this should be mounted vertically, as big as possible and connected to the stack with the most used showers. You get some extra efficiency by running both the cold water to the tank as well as the cold water to the shower through it but this can be a challenge in some layouts. Running only the cold water feed for the water tank through it also works well enough.

      1. AdamPNW | | #9

        Are there any drawbacks to connecting the DWHR to a SANCO2 heat-pump combi-system, where the storage tank works relies on stratification to run most efficiently? I’m wondering if warming the cold water return would actually reduce overall energy savings?

        1. charliepark | | #10

          This is something I'm interested in as well. I think I've seen people note that you'll be reducing the COP if you feed warmer water to the cold water inlet of the tank, but that argument's never made sense to me … if what you're going for is overall reduced energy, the COP isn't really relevant at this point, and it seems like the lower "lift" of the warmed water to the heat pump would reduce the energy needed to get it to the target temperature.

          As Small Planet Supply recommends the return of a radiant setup to feed in to the cold water supply, and that water will be coming back warmer than whatever the original supply temp would be, I think a similar effect of warming water via a hot water recovery system would be fine? Screenshot below is from this PDF:

        2. Expert Member
          Akos | | #11

          Look at Fig 7:

          The outlet temperature of a DHWR unit is not that high, about 20F above cold water supply. Yes it does reduce COP a bit, but not enough to matter so it is still a net energy saver.

        3. charliepark | | #12

          Thinking about it a bit more, I think with a SANCO2 combi system, it might be hard to justify the economics of a DWHR. Going with the SANCO2 EnergyStar sticker (, the average cost is $161/year (presumably that's domestic hot water only, not combi; not sure how much more that would add). And with DWHR costing $850, plus installation costs, even if you could reduce your hot water electric costs by a third, you're only saving ~$50/year, which would give you a 17-year payback for the copper DWHR (not even including the installation costs). It might be better to put that $1,000+ towards some other upgrade (solar, insulation, outdoor shading, etc.). (My math might be off on any number of these points; I welcome correction!)

          I love the idea of passively recovering heat, but I suspect that this might be the other bookend to Martin's claim that Solar Thermal Is Really, Really Dead, at least when dealing with a high-efficiency HPWH — the money might be better spent on other optimizations.

          Some more good DWHR comments are in this thread (, especially from Dana Dorsett, who is generally pro-DWHR.

          1. paul_wiedefeld | | #18

            I think the appeal of the DWHR would be that you could avoid installing the Sanden, instead going with a cheaper resistance tank. If you recover 50% of the shower heat, that's similar to improving the resistance COP to 2, which makes the many thousands it'll cost to install the Sanden less appealing.

      2. orange_cat | | #15

        Thank you.

        I have two stacks, both with showers and likely both showers used similarly. One is above mechanical room, so possibly the best candidate. But that leaves the other stack just draining without the recovery?

        The two stacks are approximately 15 feet apart along the same wall - one is above mechanical room, one is 15 feet away from the mechanical room.

        First consists (from the bottom to the top) of mechanical room, 1/2 bath, and a fairly heavily used shower. The second consists (again from the bottom) infrequently used shower, kitchen, and then a fairly heavily used shower.

        Should I ask to have a separate DWHR for each stack then?

      3. Expert Member
        BILL WICHERS | | #19

        I would probably try to get the dishwasher on the DWHR too. I think between showers and the dishwasher, you're probably getting 75-80+ % of the total hot water use in the home going through the DWHR on it's way down the drain. Next after those would probably be the clothes washer. The more of the waste hot water you can run down that DWHR, the more effective it will be.


        1. Expert Member
          DCcontrarian | | #20

          It only works if you're drawing hot water at the same time you're draining it. That's the use case for showers, but not for washing machines or dishwashers, which take in a batch of water and then release it a while later.

          1. Expert Member
            BILL WICHERS | | #23

            That’s a good point, and something I missed here — something would need to be running hot water while those things were draining to make things work. Showers and sinks then are the only things guaranteed to work here, and sinks are pretty minimal in terms of hot water use if they’re just being used for basic hand washing.


    3. Trevor_Lambert | | #13

      You think a full bathtub drains in a minute? Not any one I've even seen. Even half full I'd estimate at least 5 minutes.

      1. jollygreenshortguy | | #21

        Would you like me to time it the next time I drain my tub?
        I think this qualifies as nit picking.

        edit - I just checked. It took 2m44s. Satisfied?

    4. jollygreenshortguy | | #22

      Thanks DC, Akos and the rest. I understand it better now and have a better understanding of when it would be worth considering.

  4. refereedjournal | | #29

    If you have an on-demand water heater, either gas or electric, you may also find that you can go down a size / use a cheaper/smaller model since you won't need the same temperature rise/min in the winter.

  5. StuSid | | #31

    Great discussion and insights:

    My question about DWHR is will the water temperature change during the shower as the incoming temperature of the cold water is heated in the DWHR? I imagine the DWHR lags behind a bit and the user would need to adjust the shower valve a bit colder once the cold water warms up from the DWHR. Not a deal breaker for me because these units seem so cool, just more curious than anything. Thanks.

    1. Expert Member
      DCcontrarian | | #32

      Yes. Unless the shower has a thermostatic valve.

      Although I've never found them to work that well.

      1. gusfhb | | #33

        I have had hansgrohe thermostatic valves in my various showers for almost 20 years
        they work perfectly and I will never install anything else.
        I set my shower temperature in 2010 and have not touched it since
        Maybe other companies are not as good at it.
        Pretty sure they save water as the water temp comes up much faster[they are effectively on full hot until the thermostat is satisfied]
        I would think they would be a really good fit with DWHR

        The other problem is that they only use hot water, so the plumbing would be problematic

        Wonder if laundry would work...

  6. orange_cat | | #35

    Thank you for the lively discussion.
    Yes, the appeal is to not have to worry about lack of hot water for consecutive showers for family members who think they are fish. The DWHR seems weel suited given the average shower is easily 15 minutes.
    And we have hasngrohe thermostatic valves, so I was hoping the temp fluctuation would be minor.

    But nobody answered whether I need one or two units. I included the diagram of two stacks here:

    this is in the context of laying out hot water supply lines, but the drains are in the same places.

    Any suggestions?

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |