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
Musings of an Energy Nerd

Drainwater Heat Recovery Can Lower Your HERS Score

Although these simple devices have always saved energy, they weren’t recognized in HERS Index calculations until recently

Hidden inside the coil of copper tubing is a vertical length of 3-inch-diameter copper drain pipe. The warm water descending the drain pipe transfers heat to the cold water flowing through the small-diameter tubing.
Image Credit: Image #1: Lyndon Than / PassiveHouseOne

Drainwater heat recovery (DHR) devices have been around for more than twenty years. By now, over 60,000 of the units have been installed in North America. When one of these devices is installed in a typical single-family home, it can reduce the amount of energy used for domestic hot water by 15% to 22%.

Two years ago, changes were made to the rules governing HERS Index calculations to give credit for drainwater heat recovery devices when calculating a HERS rating. For builders who are interested in advertising an impressive HERS Index, this is important news — because installing a drainwater heat recovery device can be a relatively inexpensive way to get a two-point improvement in a HERS number.

Researchers measure energy savings

The first such drainwater heat recovery device to hit the market was called the GFX. In the December 1996 issue of Energy Design Update, editor Ned Nisson wrote, “The GFX is a ‘coil-and-tube’ counterflow heat exchanger that is installed vertically in the home’s plumbing waste line and connected to the cold water main. Warm wastewater from showers and sinks runs down through the central copper pipe while incoming cold supply water runs up through the tightly would coil of copper tubing.”

Over the years, a parade of researchers looked into the potential savings attributable to the use of a drainwater heat recovery device:

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial


  1. Dana1 | | #1

    Other facts to consider.
    It's usually the case that the additional marginal cost of a bigger (= both fatter & longer) drainwater heat recovery unit "pays off" quicker than going a smaller (= skinnier &/or shorter) unit. The installation labor is about the same independent of size, so installing the largest one that fits the available space is usually going to be "worth it", even if it is a few hundred USD more for the unit than a skinny stubby one.

    Natural Resources Canada developed an efficiency testing protocol to enable apples-to-apples comparisons between different models & manufacturers. They maintain a database of third party tested return efficiency (at their test parameters), and an updated spreadsheet can be downloaded from a link on this site:

    The NRCan test protocol is at a full flow shower rate (~2.5gpm), but lower flow rates increase the energy return efficiency, whereas higher flow rates lower the return efficiency.

    Get pre-approval from the building inspectors- there have been instances where local inspectors condemned them on a (flawed IMHO) basis that potable piping in contact with drain piping (even though it's double-walled construction, and it's manufactured in a way that they can't abrade or interfere with one another.) Some models also use thicker tubing than used for typical potable piping, and lack the stripe indicating that it's suitable for potable piping per ASTM specs. The quality of that tubing is actually higher than typical potable piping, but it's thicker walled than the materials spec that falls under that which can be marked. This to has led to problems with inspectors, and condemned heat recovery units.

    This is an inspector or letter-of-code problem, not a safety problem . Those same units have been condemned in some locations for lack of a marking stripe are pre-approved in Massachusetts, which has fairly stringent rules for what may/may-not be connected to potable plumbing, and the state maintains a (searchable, online: ) database of fixtures that are allowed to be hooked up. Anything not listed requires a variance in Massachusetts, and not all drainwater heat recovery units are on the list (though they could be- the manufacturer has to apply.)

    Some states offer rebate subsidies for drainwater heat recovery based on efficiency, though the details vary.

  2. GBA Editor
    Martin Holladay | | #2

    Response to Dana Dorsett
    Thanks for your useful comments. I was unaware that some building inspectors were raising questions about these units; thanks for the warning.

    Battling building inspector ignorance and obstructionism sometimes feels like a full-time job.

  3. Expert Member
    Dana Dorsett | | #3

    Times change, but it can be slow.
    This thread is going 6 years old, but chronicles a particular case where this occured:

    The same product condemned by that inspector is a pre-approved plumbing fixture in Massachusetts:

  4. Jon_R | | #4

    Note that the cost
    Note that the cost effectiveness is highly influenced by the usage and the cost of heat. With the 1 GPM I use, the ROI wasn't at all attractive.

  5. Expert Member
    Dana Dorsett | | #5

    If it saves you the cost of upsizing...
    If it saves you the cost of upsizing from 50 gallon heat pump water heater to an 80 gallon or 100 gallon HPWH it pays for itself up front.

    If the goal is to actually hit your Net Zero Energy goals it can be cheaper than the additional insulation / window /solar it would take to get there.

    It's not always about the net present value of first + financing costs relative to future energy cost savings. The title of this piece is, after all

    "Drainwater Heat Recovery Can Lower Your HERS Score"

    If the goal is to limbo under some HERS target, it can cheaper than some other methods of getting there.

  6. Westhawk39 | | #6

    Mandatory in Manitoba
    I was just reading a revised building code document for the province of Manitoba and it stated DWHR devices are mandatory for all new home construction. It states that they are only to be installed where physically possible (not for mainfloor shower with no basement). But doesn't state if all showers require them or other appliances or fixtures require them.

    Would it make sense to install for a washing machine or dishwasher?

  7. GBA Editor
    Martin Holladay | | #7

    Response to Scott Benson
    If a drainwater heat recovery device only serves a washing machine or dishwasher, it would save so little energy that the payback period would be greatly extended. Most analysts would conclude that this type of installation is a waste of perfectly good copper.

    Thanks for letting us know about the Manitoba requirement. As far as I know, that's a first.

    I can't help noting one unfortunate feature of the PowerPoint presentation you linked to: it includes an illustration that depicts a flawed installation. In the example shown in the illustration, the pre-heated water is sent to the cold inlet of the water heater, instead of to a tee as required for an "equal flow" installation.


  8. charlie_sullivan | | #8

    Doubly flawed
    Not only is that illustration missing the T--it's also connected backwards. Heat exchangers work much better in a "counterflow" configuration, with the supply water coming in at the bottom (near the exit of the drain water) and coming out at the top.

  9. GBA Editor
    Martin Holladay | | #9

    Response to Charlie Sullivan
    Good catch! You're right.

  10. exeric | | #10

    If it isn't already obvious
    If you have marginal requirements for longer hot water shower times than you are currently getting AND you don't currently have a low flow shower head then save some money and time and go for the low flow shower head and see if it works for you. I know, everyone here already has them but I figure its better to be comprehensive than not allow for outliers. There are some pretty good ones now that don't feel like low flow showerheads.

  11. Expert Member
    MALCOLM TAYLOR | | #11

    Different waste sources
    I understand their effectiveness with simultaneous water use and drain situations, but can someone tell me how they work where that doesn't occur? With baths, dishwashers and washing machines, where there is a delay between water use and drainage, how does their efficiency suffer? The pre-heated water in the coil would presumably be available for some period of time, but would be cooled if any other cold water drains into the plumbing stack.

    Are these effectively shower heat recovery devices? And if so would there be any benefit to twinning the drains so that they only served showers?

  12. GBA Editor
    Martin Holladay | | #12

    Response to Malcolm Taylor
    For all intents and purposes, these devices recover the heat that goes down the drain from a shower.

    Of course, there will be some incidental heat recovery from other uses of hot water -- for example, when someone washes dishes in a sink, or on those rare (coincidental) occurrences when some family member is drawing hot water at a sink when water is being drained from a bathtub -- but those occurrences aren't enough to justify the cost of the drainwater heat recovery device.

  13. Expert Member
    MALCOLM TAYLOR | | #13

    Thanks for the reply. I have been thinking about these units through the lens of the last two houses I designed.

    The first was two storey slab on grade, with one shower upstairs and another on the main floor, so the unit could only serve one.The owners have been surprised by their electrical bills, but that appears to be largely due to the demands of washing the clothes of their two infant children. Perhaps this balance will shift over time towards more showers.

    The second is on a crawlspace, and the occupants favour baths over showers. Their total annual electrical consumption represents somewhere in the region of $1000.

    From a design perspective the questions I face are whether they work well enough for a generic family that it is worth including them even if they don't make sense for the present occupants, and whether they work well enough that it is worth making design changes to try and incorporate them, where otherwise they couldn't be installed.

  14. Expert Member
    Dana Dorsett | | #14

    It's probably more than you ever wanted to know but...(@Malcolm)
    The NRCan modeling and test protocols were developed from research summarized here in some detail:

    There is very little thermal mass in gravity film heat exchangers themselves, so unless flow is simultaneous the amount of heat recovered falls below what is easily measurable. The only residental use-case for this technology really boils down to showers- even hand washing & toilet flushing heat recovered is so tiny as to be difficult to measure, though it can be theoretically calculated.

    Focusing that use-case entirely on NPV of future energy cost savings is to miss a piece of the boat. There is some savings on the size/capacity requirements of the hot water heating mechanicals as well. It's true that some families will be shower-bathers who literally never shower, at which point it's value is limited to scrap copper, but those instances are probably more than one standard deviation out. Whether it makes good policy sense to require them by building codes requires a bit more research on hot water use patterns across a large number of households to know the distribution of shower users vs. tub-only bathers in the region. (Hopefully that research was actually done in Manitoba.)

    From a total carbon footprint of the heat exchanger, a few years ago I made a napkin-math estimate that with a standard gas tank type water heater and 20 minutes of 2 gpm showering use per day the "payoff" on atmospheric carbon emissions was less than one year. Most of the carbon emissions are from the smelting of copper from ore, with much lesser amounts attributable to transportation & fabrication energy. With more use data and a survey of water heating sources one could convert that into lifecycle carbon tons avoided to tons emitted ratio, and the relative cost to come up with a $/ton number, which would also be part of the policy discussion.

  15. Expert Member
    MALCOLM TAYLOR | | #15

    "It's probably more than you ever wanted to know but.."

    Summarizing that sort of stuff is kind of what I've come to rely on you for on such topics :)

  16. user-5942299 | | #16

    Chimney pipe heat recovery
    Can this concept be applied to a chimney pipe coming off a wood stove or masonry heater? Are there examples of this and products designed for this application?


  17. GBA Editor
    Martin Holladay | | #17

    Response to ADK Creator
    Any device that lowers the temperature of a wood-stove flue increases the rate at which creosote accumulates and clogs the flue. This raises the frequency of required cleanings and increases the risk of chimney fires. That's why lowering the temperature of a wood-stove flue is a bad idea.

  18. user-5942299 | | #18

    Understood. Thanks

  19. jchas | | #19

    Been working well for me for over 3 years
    Installed a 42x4" PowerPipe in our basement 3 years ago here in Victoria BC. (Zone 4 Coastal)

    In winter, here was the heat recovered while using the shower (it stabilized within 30 seconds of starting the shower):
    40F: temp of water from street
    100F: temp of water from shower head
    90F: temp of drain water by the time it reaches the plumbing stack
    68F: temp of warm water leaving DWHR unit
    This gave us a 28F temp gain "for free" while running the shower (28/60=43% energy use reduction).

    This year I added a second PowerPipe (66x4") in series with the original one. Drain water from the 2nd floor shower runs through both units while water from the ground floor shower runs only through the original one.

    The temp gains from the ground floor shower remained the same. However the heat gains from the 2nd floor shower are now close to the theoretical limit. My measurements show:

    40F: temp of water from street
    100F: temp of water from shower head
    90F: temp of drain water by the time it reaches the plumbing stack
    82F: temp of warm water leaving DWHR unit
    This gave us a 42F temp gain "for free" while running the upstairs shower (42/60=70% energy use reduction).

    Payback for this system will be under 4 years. Performance is better in colder climates (where incoming water temperature is lower). There is no significant heat recovery from dishwasher, toilets, or baths.

    Having the two heat exchangers in series has worked well even when only the lower one is being fed with drain water because its outlet temperature (68F) is room temperature and therefore is energy neutral as it runs through the upper unit.

    Highly recommended for those in cooler climates.


  20. Expert Member
    Dana Dorsett | | #20

    Bigger really IS better, in this instance @ John Charlesworth
    Those numbers correlate reasonably will with NRCan test data. The 4x 42" PowerPipe tested ate 43.7% steady state under the standard test conditoins:

    Have you bucket-tested the actual shower flows?

    The much higher return efficiency of longer units almost always pays back sooner than shorter ones. I only had space for a 4" x 48" running at ~50% average return efficiency (in service since late 2008) at my bucket tested ~2 gpm, but it's been worth it, despite the decline in the retail cost of natural gas that has occurred since it was installed.

  21. jchas | | #21

    Actual efficiency even higher
    I haven't measured the gpm flow yet--I should do that.

    The biggest loss is the cooling that happens as the shower water falls through the air from shower head to bathtub drain--almost 10F.

    Looking at the heat recovery of the heat exchangers themselves based on the actual temperature of the drain water entering them, it's even higher as a percentage.

    For the R4-42 unit on its own, actual measured efficiency is 28/50=56% and for the combination of the two in series, it's 42/50=84%.

    Both of these are better than the NRCan ratings so I must be at the sweet spot regarding flow rates. I do have it plumbed so that the pre-heated water feeds both the water heater tank *and* the "cold" side of the shower controls.

  22. Expert Member
    Dana Dorsett | | #22

    NRCan measures the net efficiency, not the simple efficiency.
    The test protocol measures the temperature at the showerhead, not the temp at the drain for determining the efficiency, since that's the fraction of actual net energy returned of the raw energy input.

    That's a more relevant number than the raw efficiency of the heat exchanger itself in a closed system, without the losses to the room that happen between the shower head and the drain.

    One could argue that in a heating dominated climate a large fraction of the energy lost between the showerhead & drain isn't really lost, since it offsets the heating load, but that gets a lot squishier, and isn't what they are really trying to measure.

  23. Buzz_Burger | | #23

    What about commercial use?
    I can't help but think that in certain commercial applications there could be a tremendous savings/quick ROI. At my local health club, for example there is usually continuous multiple shower use for about 15 hours a day, Is the appropriate hardware available for that situation or can the same hardware work? Often, but not always there is no basement below which could be a problem.

    I've also wondered if the hardware exists to somehow connect the heating of water with refrigeration. Probably totally impractical for a single family home, but what about a large restaurant or other food service facility that requires a lot of refrigeration and a lot of hot water use? Is there a practical way to suck the heat out of a freezer with a heat pump and transfer it to water?

  24. GBA Editor
    Martin Holladay | | #24

    Response to Buzz Burger
    Manufacturers of drainwater heat recovery devices are very familiar with commercial installations at establishments like hotels, laundromats, etc. Owners of commercial establishments should talk to these manufacturers about their needs -- most manufacturers can help design a good installation.

    If you use the right terms in the GBA search box -- maybe "hybrid refrigerator water heater," without the quotes -- you'll read lots of threads on the topic. The topic comes up on GBA like clockwork, twice a year. Makes no sense for a residential installation, as you guessed.

  25. timgodshall | | #25

    Can a drainwater heat recovery unit serve two showers?
    I have two showers located close together and am wondering if one of these units can serve both of them? If so, I suppose the equal flow installation would have to include a pipe serving both showers. Also, how far away from the shower can the drainwater heat recover unit be located and still be effective? Due to the wall layout in the basement under my bathrooms, the vertical drop won't occur till about 15-20 feet away from the shower drain.

  26. GBA Editor
    Martin Holladay | | #26

    Response to Timothy Godshall
    Q. "I have two showers located close together and am wondering if one of these units can serve both of them?"

    A. Yes -- certainly.

    Q. "If so, I suppose the equal flow installation would have to include a pipe serving both showers."

    A. That's correct.

    Q. "How far away from the shower can the drainwater heat recovery unit be located and still be effective? Due to the wall layout in the basement under my bathrooms, the vertical drop won't occur till about 15-20 feet away from the shower drain."

    A. It's still worth installing the drainwater heat recovery device. If you want to boost the performance of the system slightly, you could wrap the drainpipe that extends from the showers to the drainwater unit with insulation.

  27. Dana1 | | #27

    No need to insulate the distributed output of the heat exchanger
    The easiest way to have the heat exchanger's output serve multiple showers is to have it supply the entire cold water distribution for the house. The output of the heat exchanger is typically between 68-75F (depending on incoming water temperature, flow, and the size of the heat exchanger), It would take an unusually large heat exchanger for it's output to be much more than that.

    So if the distribution plumbing is inside of conditioned space (even if it's a unheated 60F basement), the total amount of heat loss from even 50' of distribution plumbing is pretty low, not enough to rationalize insulating the tepid heat exchanger output plumbing to the shower(s) and water heater.

    Losses from the drain distribution are slightly higher due to higher temperature of the drain, but again the losses are pretty small, especially with plastic drain pipe. The losses will be slightly higher on vertical sections of drain due to the gravity-film distribution on the interior surface, but on the horizontal sections the water is flowing in a tight stream along the bottom of the pipe. Only with very long drain plumbing would insulating it be worthwhile.

  28. timgodshall | | #28

    Response to D Dorsett
    Thanks for your response. I have a question about your recommendation: "The easiest way to have the heat exchanger's output serve multiple showers is to have it supply the entire cold water distribution for the house." Wouldn't this dilute the efficiency of the system since the hot water gains are being dissipated in the whole house's cold water lines rather than going directly to the shower that is drawing the water? (I acknowledge that you did say "easiest" not "most efficient"!)

  29. Dana1 | | #29

    Not much (response to Timothy Godshall)
    "Wouldn't this dilute the efficiency of the system since the hot water gains are being dissipated in the whole house's cold water lines rather than going directly to the shower that is drawing the water?"

    The tepid water only flows toward the open taps, and it's at roughly room temperature, with very little heat being lost to the rooms.

    If a sink or toilet is filling while the shower is running there is some loss, since the heat going into the sink or toilet isn't being returned to the hot water heater, shower mixer, or the drain. If free-flowing (cold only) into the sink at high flow the dilution factor will be real enough. But in practical terms those events aren't usually long in duration or very frequent in most homes.

  30. ewoodhouse | | #30

    Why not just send the heated water straight to the shower?
    So, dumb question, perhaps: Assuming the DHR device is attached to a single shower drain, why is the "equal flow" installation better than just sending all the preheated water to the cold side of the shower mixing valve? I was in a class on heat pumps for water heating a couple weeks ago and it was mentioned that the COP of the heat pump drops as the cold supply temperature increases, so that returning preheated water to a HP tank isn't a good idea... but as I think about it, it seems like using the preheated water immediately would be most efficient, regardless of the heat source. Am I missing something here?

  31. charlie_sullivan | | #31

    Response to Eric Woodhouse
    The efficiency of heat recovery goes up when the volume of incoming water going through it increases, and goes down when it's lower. So you simply get the best heat recovery when all the incoming water flowing at that time goes through the heat exchanger. It's true that the heat pump COP goes down when the water temperature goes up, but it's really the mix temperature in the tank, not the incoming water temperature, that matters. If you are significantly dropping the tank temperature, the people showering aren't going to be happy, so there isn't really a good way to make use of that phenomenon.

Log in or become a member to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |