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

DWHR and HPWH vs Electric, Total Efficiency?

lance_p | Posted in Mechanicals on

So it’s well known that a Heat Pump Water Heater (HPWH) is far more efficient at heating water than a standard electric water heater, but what happens when we look at the entire system top to bottom, including Drain Water Heat Recovery (DWHR) as well as the building’s main heat source?

A HPWH is most efficient with low water temperatures; the colder the incoming water and the lower the hot water temperature set point on the tank, the higher the overall Coefficient Of Performance (COP) will be. Using Rheem’s hybrid model as an example, they claim a COP of 3.5, or 350% efficiency, under standard test conditions.

A standard electric water heater is 100% efficient regardless of the water temperatures. Only standby losses are a factor.

So what happens to this situation when we throw DWHR into the mix? All of a sudden the incoming water temperature is much higher because DWHR captures 40-60% of the heat from the outgoing drain water. This, in theory, will raise the average water temperature in the tank and decrease the COP of the HPWH. A standard electric tank keeps on going at 100%.

Now consider the heat source for the building. A HPWH is taking heat from the building and putting it into the water. If the building has a 95% efficient gas furnace, the heat the HPWH uses is starting off with a 5% efficiency hit. If the building is heated with an Air Source Heat Pump that’s operating at a COP of 2.0 in cold weather what happens to system efficiency then?

A standard water heater might last 10-12 years. A HPWH will likely be the same as the tanks are built the same, assuming of course the heat pump is still working trouble free. So the cost to replace the HPWH is about twice as high every 10-12 years.

Now, just to make things extra confusing, throw in a standard water heater like the Rheem Marathon which has a lifetime guarantee against leaks (no replacement costs) as well as greatly reduced standby losses (3″ of foam instead of 2″). How does that measure up over time?

Maybe this has been studied before and I just haven’t found the paper? I’ve looked and all I seem to find are reviews of the water tank only, not the math to figure out the whole picture.

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. GBA Editor
    Martin Holladay | | #1

    You are overstating the case when you say that with a drainwater heat recovery (DWHR) unit, "the incoming water temperature [at the water heater] is much higher because DWHR captures 40-60% of the heat from the outgoing drain water."

    It's not "much higher." At least 50% of the warmed water is going directly to the cold line supplying the shower, and the other 50% of the warmed water is going to the bottom of the water heater tank. While warmer incoming water slightly lowers the COP of the heat-pump water heater, it reduces the amount of energy required to heat it.

    Remember, too, that the effect of a drainwater heat recovery unit depends on occupant behavior. It's useful for families that take a lot of showers, but fairly useless if family members prefer baths.

  2. Expert Member
    Dana Dorsett | | #2

    The EF testing is run at 56-60F incoming water temperature, at a room temperature of 65-70F, and an indoor relative humidity of 50% (+/- 2%). The efficiency of the heat pump is lower at lower RH and lower room temperature, and higher RH & room temp. Efficiency is also lower at 140F delivery temp (required setting by code in many areas) than it is at the tested 125F.

    The NRCAN test protocol for DWHR units calls out 8C/46F incoming water temperature and 10 litres per minute (2.6gpm) flow, and 46C/114F at the shower mixer. Efficiency is higher at lower flow rates, lower at higher flow. Efficiency is also lower at higher incoming water temps, higher at lower temps. This

    So the nameplate efficiency of the DWHR could be either higher or lower than reality, and the same is true of the HPWH. There are lots of variables here, and a home is not a test laboratory with those variables controlled to test lab conditions.

    But the raw physics isn't defeated by variations in efficiency- heat that is recovered rather than flowing down the drain is in no way defeated or even significantly altered by the relatively minor changes in heat pump efficiency that occur during the 20-40% of the total hot water use that occurs during showers. It's still a significant net reduction in showering energy use, even if the model isn't a simple-math problem.

    The efficiency of the space heating heat pump is only relevant to the fraction of the heat the HPWH is drawing from the room air, which will vary with conditions, but typically 1/2-2/3 of the total heat going into the water. The remainder is the power used by the compressor & fan at an effective COP of 1. Clearly a COP of 2 is better than a COP of 1, and even if the warmer incoming water reduces the fraction drawn from the room air to 1/2 of the total instead of 2/3 it's still a net win if it means it only needs half as much total heat.

    There is no way a Marathon will come anywhere NEAR the efficiency of a HPWH, with or without DWHR (on either.) Whether it's more financially effective or greener depends a lot on the local utility rates and the local grid energy resources.

    Comparing gas furnace efficiency to a heat pump efficiently isn't legitimate, since the net source-fuel efficiency of the generator and delivery to the heat pump, which is well under 100%. But the grid source mix and grid efficiency varies a lot by location, and also over the lifecycle of the heating equipment. The grid is trending toward higher fuel to load efficiency for the fossil burning generators but the fossil burners are also a shrinking share of the overall energy mix. A heat pump averaging a COP of 2 in the ISO New England grid region is substantially lower carbon than a condensing gas furnace, and will be even more so over the next 15-25 years. A heat pump averaging a COP of 2 in coal-heavy West Virginia or Ohio can't make that same claim.

  3. lance_p | | #3


    I have to disagree that 50% of the water from the DWHR is going to the cold side of the shower. If the water heater temperature is set at 120F and the shower water temperature is 105F, very little of the shower's flow will be coming from the cold water side. But I'm not sure it matters since 100% of the water entering the water heater has gone through the DWHR.

    Assuming the shower is the only thing using water and the drain water is at 90F entering the DWHR unit (after losing heat in the shower and pipes) and the DWHR is 40% efficient @ 2.5 gal/min, incoming water at 58F will be heated to roughly (90-58=32 x0.4=13 +58=71) 71F.

    The COP of a HPWH will drop when operating from 71-120F compared to 58-120F. By how much? Probably not a lot, maybe down to 3.0 from 3.5? Who knows, but it's worth considering the specified efficiency will be lower when doing the payback math. DWHR is mandatory for new construction in Ontario, the minimum efficiency required is 42%.


    Thanks for your thoughts. I tend to agree with you that the HPWH would still be far ahead of even a good standard tank like a Marathon. I also didn't realize the incoming water for EF testing was ~58F, higher than I would have thought and 46F in Canada, closer to where I was thinking. In which case the efficiency of a HPWH would be higher in Ontario where the incoming water is colder.

  4. user-6908581 | | #4

    In another thread I recently asked about going with a Marathon and the DWHR instead of the HPWH; while I don't expect it to be near the HPWH in efficiency as Dana pointed out, I think it will be good enough. I bought a Marathon for my previous home as I disliked that day every 8 to 12 years when you would find your water heater leaking and need to replace it. Lance, I think you bring up a valid point, in that if you go the Marathon/DWHR route and don't have to replace the HPWH in 10 years, is your long term cost closer to even? Of course there is no guarantee the Marathon won't fail, I just know I flushed mine every 3 years, with an Iron Curtain in front of it on a well, and it was always clean and running as good as when it was installed 10 years earlier. Neighbors even with an Iron filter and the typical tank and sacrificial anode rod had failures quicker than you would expect. Perhaps if you are on a well and have less than ideal water quality, that may make the Marathon and it's non-metallic tank a more cost effective long term choice. The new build is in the same area so I'm going to stay with what worked with the addition of the DWHR. The HPWH is tempting though, as while being cheaper to operate, could potentially also eliminate the need to run a stand alone dehumidifier in the basement.

  5. Expert Member
    Dana Dorsett | | #5

    Lance: The drain water is usually a bit over 100F, not 90F, unless you like really cold showers. There's not a lot of thermal mass in the drain pipe or much heat loss do the rooms (even cool basements) for the water to give up heat to. It's more realistic to assume 100F at the top of the heat exchanger.

    This is probably still too much information, but a brief description of the characterization of DWHR performance testing from which the NRCAN test protocol was eventually derived can be found here:

    The additional HPWH efficiency of the lower incoming water isn't as significant as the difference in final storage temperature and the room temperature % humidity. Even nearly freezing 35F incoming water will be 58F (a 23F difference) or higher for about 3/4 the heating cycle on it's way to a 125F storage temp.

    It also would more than half the heating cycle from 75F incoming water (a realistic output temp of a ~50% NRCAN tested DWHR unit at 2.5gpm with 35F incoming water) going to 125F storage temp. It's not going to have a huge impact in average COP efficiency. I doubt the hit is anywhere near 0.5 COP, but its probably more than a COP drop of 0.1.

    But whatever the initial incoming water temp is, it still has to cross through 75F, so giving it 75F water means it's just running a shorter cycle, at the same efficiency that it would be with the cooler intake water situation would be for that part of the cycle.

    The delta-T between a 65F room at 125F storage temp is 60F. Raising the storage temp from 125F to 140F would become a 75F delta from room to water, a 25% higher hill to climb. That might result in a COP reduction of 0.5, but it's probably not quite that bad.

  6. charlie_sullivan | | #6

    I'm realizing there's something I don't know about how a modern HPWH condensor is configured, and scanning a few pictures online I see different configurations. The more popular version seems to be that the condensor coils wrap around the tank liner, inside the insulation, and thus the whole inside surface of the tank acts as a heat exchanger between the hot refrigerant gas and the bulk stored water. The other arrangement has the condensor in the top with the rest of the heat-pump hardware, and the water from the tank is pumped through the condensor to heat it.

    In the former, the incoming water temperature doesn't directly affect the COP--all that matters is the temperature of the water in the tank. If it drops from 130 to 110 over the course of the shower, the heat pump is delivering to a 120 F average water temperature.

    In the latter, it could be configured so that the incoming water flows in through the condensor. It might get warmed from 50 F to 80 F on that first pass through, and the heat pump would have a very high COP during that process. The when the shower is done, the mixed tank water gets circulated through and the COP is about the same as in the first design.

    So the incoming water temperature affects the COP more in the second design than the first. But still, most of the heat pump run time happens after the shower is over, so the effect isn't all that great.

    Anyone know the actual configuration used?

  7. GBA Editor
    Martin Holladay | | #7

    Online pricing for the 50-gallon Marathon water heater is about $1,045 to $1,114.

    Online pricing for the 85-gallon Marathon water heater is about $1,400.

  8. Trevor_Lambert | | #8

    Here are a couple of tidbits to keep in mind. When comparing the Marathon and its lifetime tank warranty, don't forget that thing is probably going to cost an arm and a leg, it's not just going to be 10% more than a standard electric. The DWHR unit is code requirement in Ontario now, so you have no choice, you have to put it in there. I really can't believe it's ever going to justify either its initial cost, or embodied energy, at least in my case. My wife has baths, I only shower at home a couple of times a week. I don't know how many showers it will take to pay it off, but I'm guessing I won't live anywhere near that long. I actually have mine hooked up only to the shower cold inlet, not the water heater inlet. The instructions said this configuration would be "slightly" less efficient than splitting it (it also described directing it only to the water heater inlet as "slightly" less efficient). Having it split to the water heater would have have made the plumbing to integrate into my open direct hydronic heating exponentially more complicated, so I nixed that idea. As for the heat pump water heater, you won't find one locally for any kind of reasonable price. They don't sell them at big box stores, and plumbers/plumbing supply stores mark them up 100% over retail. So unless you are planning a trip to the states with a pick up truck or a trailer, probably just forget about that for now. I made the mistake of buying a used Geosprings hybrid for a "great" price. Not surprisingly, the heat pump is completely non-functional, so I just bought a big, used electric water heater for the price of a new, more compact one.
    I just looked up the price of a Marathon, it's pretty much the same cost in US$ as an equal sized hybrid heat pump. So at least C$1500, maybe as much as $3000. Westinghouse makes a much cheaper version of the same concept.

  9. Trevor_Lambert | | #9

    I wish I saw that Westinghouse one before I bought a conventional one a few weeks ago, probably would have driven over the border and got it. 5500W and max temp of 170F... put a mixing valve on that and it would deliver a lot of hot water. Oh, the cheaper one on sale only goes up to 150F.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |