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
Guest Blogs

Getting Into Hot Water — Part 2

Installing an add-on heat-pump water heater

Image 1 of 3
The Geyser heat-pump water heater is an add-on unit that sits beside a conventional electric water heater.
Image Credit: All photos: Marc Rosenbaum
The Geyser heat-pump water heater is an add-on unit that sits beside a conventional electric water heater.
Image Credit: All photos: Marc Rosenbaum
This photo shows what's inside a Geyser heat-pump water heater. I removed the lower electric-resistance heating element from the Marathon water heater so I could hook up the Geyser to the tank.

In my last blog, I discussed the basics of our domestic hot water (DHW) load, and looked briefly at adding a solar hot water system to satisfy most of that load. What I decided to do first was to try a heat-pump water heater (HPWH), partly because it was a much simpler and less costly installation, and partly because I was just curious to see how well one would work.

Since I already had the 85-gallon Marathon electric water heater, I didn’t need additional storage, so I looked at the add-on products available and chose the Geyser, which is made in Maine by Nyle Systems. It costs in the range of $1,000, plus installation.

An add-on heat-pump water heater

The Geyser is a small cube (about 1 1/2 feet on a side) which can sit on the floor or on a platform adjacent to the water heater tank. The way it (and other HPWHs) operates is that it removes heat from the air in the space in which it is located and transfers that heat to the water.

The Geyser has a compressor to operate the refrigeration cycle to take the heat out of the air (like an air conditioner or dehumidifier), a fan to move air across its refrigerant coil, and a pump to circulate water to and from the adjacent storage tank. It connects to the tank at or near the bottom, so it heats the coldest water, maximizing the efficiency of the heat pump. Image 2 (below) shows a the interior of the Geyser, including its key components.

As a HPWH extracts heat from the air, it also cools it, which means that it acts as a dehumidifier as well, with one key difference. Conventional dehumidifiers cool the air, condense moisture out, and then reheat the dehumidified air, so they remove moisture and add heat to the air. Since warming the air lowers its relative humidity, the twin effects of moisture removal and heat addition are exactly what we want — lower relative humidity means lower moisture absorption by materials in the space being dehumidified, and therefore less mold.

A HPWH removes moisture from the air, but cools it rather than heats it, because the heat goes into the DHW, so it isn’t as effective a dehumidifier. I put a Hobo four-channel data logger in place to look at the effects of the HPWH on the basement environment.

The normal Geyser installation uses a clever tube-in-tube fitting that is inserted in the drain fitting on a conventional electric water heater (or other water heater tank). This allows the Geyser to both extract water to be heated, and return heated water through the same port on the water heater.

Removing the Marathon’s lower heating element

In the case of the Marathon, the drain port location precludes this. I chose to do something a bit unconventional — I removed the lower electric heating element (residential electric water heaters have an upper element and a lower one, and only one operates at a time) and used an adapter to connect the heated water returning from the Geyser to this location. The water to be heated comes from the drain connection.

This approach loses the functionality of having the lower element as back-up to the HPWH — only the top element is usable. The Geyser uses the existing line voltage thermostat on the water heater to tell it when to operate. In a typical electric water heater, the controls first ensure that the top thermostat, that controls the upper element, is satisfied, then the lower thermostat and the lower element, which properly prioritizes the water at the top of the tank, which is drawn off first.

In my set-up, once the upper thermostat is satisfied, the power is switched to the lower thermostat, which can call for the Geyser to operate if needed. I’ve set the upper thermostat to about 90°F, and the lower thermostat to 120°F, which forces the heat pump to do all of the water heating.

I installed the Geyser on July 5, 2011, and we haven’t used any power for electric resistance heating since then; all water heating has been done by the Geyser. Note that HPWHs are slow heaters relative to typical water heaters. The Geyser’s output is under 2 kW, whereas an electric water heater is typically 4.5 kW, and gas or oil heaters are much more. This is why the tank size is large for a HPWH. We’ve had a house full of guests a number of times this summer and fall, and no problem with enough DHW.

Monitoring performance with water meters and a Kill-a-Watt meter

Since this is a new technology for me, I wanted to understand how well it was working. I installed a DLJ75 water meter on the cold water feed line to the water heater, before it splits to the thermostatic mixing valve, so I would be able to measure how much DHW we were using. This is a simple, time-tested mechanical water meter. We’re installing a couple of others right now that have a pulse output that can be counted and totaled, so they won’t need to be read by a human daily to get a sense of the usage patterns.

The Geyser is a 120-volt machine that simply plugs into an outlet, so I’m tracking its electrical usage with a Kill-A-Watt meter. I also have been using the two exterior channels of my Hobo logger to measure cold water inlet and DHW outlet temperatures. With flow and temperature differential I can calculate the energy demand of the DHW system, and with the kWh I can calculate the system efficiency at satisfying that energy.

Handling the condensate

Like dehumidifiers, HPWHs remove moisture from the air and therefore generate liquid water as that water vapor condenses on the coil. The Geyser has a drain pan to catch this condensate and a tap that can be connected to a plumbing trap or a condensate pump. Mine is simply draining to a 5-gallon pail.

Interestingly enough, in the middle of the summer, when the basement temperature and relative humidity was highest, there was a puddle on the concrete below the unit, which went away as we left the peak cooling season behind. I tried to determine the issue and the folks at Nyle were very helpful, but in the end I think that somehow there is a leak in the drain pan, probably at the corners, that only is operative when the rate of condensate is high. Alternatively, there may be condensation on a component that is not above the drain pan. I may need to wait to next summer to resolve this one.

I have been pleased that the machine is fairly quiet, and so far, completely reliable.

Add-on or integrated unit?

If I had decided at the beginning that I wanted to install a HPWH, I probably would have selected an integrated unit, that has the heat pump built-in on top of a storage tank. These are made by Stiebel Eltron, GE, A.O. Smith, and Hubbell, amongst others.

There is an excellent report on the field performance of HPWHs authored by Steven Winter Associates.

In the next post, I will comment on other aspects of the DHW installation; then in a subsequent post I’ll share some data on the performance of the HPWH and its effect on the temperature and moisture level of the basement air.

[Editor’s note: This series will continue; stay tuned for “Getting into Hot Water — Part 3.”]

Marc Rosenbaum is director of engineering at South Mountain Company on the island of Martha’s Vineyard in Massachusetts. He writes a blog called Thriving on Low Carbon.


  1. user-984364 | | #1

    This looks really interesting
    I have to run a dehumidifier in the basement for much of the summer, and the notion of capturing that heat for DHW is really appealing. But for the rest of the year, when we are actively heating the basement, sucking that heat back out for DHW ... I don't know if that makes sense in my case. I'll be very interested to see your data for basement temp & humidity.

  2. watercop | | #2

    Well, dang!
    Having gone up one side and down the other on Marc's selection of an 85 Gal Marathon for 13 GPD usage back in part one...

    ...Now I get to admit I have a somewhat similar setup.

    I have a 3400 SF ICF house with 4.5 souls onboard (one part time teen) It is heated and cooled (well, mostly cooled, since we're in Florida) by a two stage 3 ton WaterFurnace Envision water source heat pump locked in low stage. It works nifty and has the heat recovery desuperheater option.

    The WF desuper feeds its very own 80 gallon (unfired GE 12 year electric water heater) storage tank which it keeps at 100-115*F May-Sept and Dec-Feb. During shoulder seasons with minimal HVAC the preheat tank is 70-90*F.

    The desuper's 80 feeds a 2nd identical GE 12 year 80, initially fired by conventional 5500 Watt resistance element(s) in the usual fashion, but, beginning in July 2009, de-energized in favor of a Nyle (then North Road) rev 0.0 serial number one (or thereabouts) Geyser.

    The early Geyser is noisy and required almost complete disassembly to adjust temperature setpoint. It also featured a balky Hall effect current sensor that routinely tripped the unit out at amperages well below rated maximum.

    I've since hacked a hole in the top of the cowling so as to adjust temperature and bent the current sensor away from its host relay. I also added a digital timer to interrupt power to the Geyser for ten minutes twice each day at 0400 and 1600. That overcomes the non-adjustable 20* control deadband that otherwise caused tepid hot water delivery since the preheat tank routinely delivers 95-105*F water to the Geyser tank - too warm to trigger a run cycle unless I set the Geyser for at an inefficient (and unsafe for small children) 125*F+ setpoint.

    Having done all that, the Geyser reliably delivers God-knows-how-much, but certainly more than 50 gallons of 110*F hot water daily for all our needs at a cost of about $3 per month all summer and no more than $15 during shoulder (no HVAC) months. I know that to a fair degree of certainty since, like Marc, my Geyser has been monitored using a Kill-a-Watt since Day One, July 15, 2009.

    P.S.: Though I got a sweet deal on the twin GE 12 year 80s back in 2008 when a local Home Depot was discontinuing that model, I do regret not having gone with Marathons - their clearly superior insulation would substantially reduce standby losses.

    P.P.S.: I like Marc's creative thinking on using the lower element port since the Marathon's geometry precludes using the Geyser's tube-in-tube plumbing arrangement. I declined to use the tube-in-tube owing to concern of heat crosstalk between tube walls. Instead I withdraw water from the cold water inlet (whose dip tube extends to near the tank's bottom) and reinject warmed water into the bottom drain. I prefer and advocate this arrangement since it prevents sediment / scale that may accumulate at the very bottom of a tank from being drawn into the Geyser's heat exchanger

  3. Marc Rosenbaum | | #3

    Curt, my Geyser is a later
    Curt, my Geyser is a later version. It's acceptably quiet for the basement. It uses the lower element's thermostat in the tank, which simplifies the installation and makes it more robust. The exterior case is still designed such that if you want to get inside it's much more of a nuisance than it needs to be.

  4. Alex Wilson | | #4

    dB rating on the Geyser?
    Marc, have you actually measured the decibel level from the Geyser, or does the company publish that. Noise seems to be the Achilles heel of HPWHs. Of the integral models, noise ranges from about 55 dB (GE GeoSpring) to 65 (Stiebel Eltron), which is a ten-fold difference. Other manufacturers don't seem to publish that information. I'd be interested to know how the Geyser compares.

  5. kevin_in_denver | | #5

    Split HPWH
    Marc, I disagree with your comment: "If I had decided at the beginning that I wanted to install a HPWH, I probably would have selected an integrated unit, that has the heat pump built-in on top of a storage tank."

    Your Geyser could last 30 years or more. Hot water tanks only last 12 years. It would be a shame to send a perfectly good heat pump (filled with refrigerant) to the land fill because the steel tank rusted through. Replacement cost will be at least double that of a tank only.

    So a Geyser with a Marathon split system will have by far the lowest life cycle cost of any HPWH. And you will never have the hassle of replacement or the potential water damage typical of steel tanks.

    Folks, be careful using Curt's advice above about sucking water out of the dip tube. Some manufacturers install a built in check valve or "heat trap" that prevents the flow of water in that direction.

    BTW - Here's a website that informs us how to make our steel water heaters last 30 years or more....

  6. Marc Rosenbaum | | #6

    I don't have a sound meter,
    I don't have a sound meter, so send one along and I'll measure both the Geyser and the Stiebel in place. Geyser specs say 60 dB at 3 feet. At this point I'm reluctant to recommend the Geyser anyway - not enough improvement over straight resistance, and not Energy Star rated.

    I'd say that the Achilles heel I'm worried about is longevity, not noise, at least in cold climates where these will be located in the basement in most cases.

  7. watercop | | #7

    Kevin makes a good point
    Kevin makes a good point about the heat trap in the cold water dip tube. I normally yank those, but forgot to mention that detail.

    One time I forgot to yank the check valve but found that the circ pump had no problem overcoming it - it resembles a rubber washer with inward radial rubber flaps able to resist the very small delta-P of natural convection but open right up in the face of pump operation.

    The check valve is part of a nipple that also includes a plastic sleeve insert that serves to reduce the cross-sectional area of the water heater inlet and outlet from 3/4" to 1/2", so if a client has need of high (> 7 GPM) hot water flows such as for a Roman tub faucet I replace the nipples with full port 3/4" and compensate by insulating both the hot and cold water lines near the tank.

    I can fill our third floor 80 gallon Roman tub in less than 10 minutes with just 30 psig well water because I have done away with all the nuisance check valves and sleeved nipples in favor of full port 3/4" fittings.

    I'm skeptical of Kevin's hypothesis that a Geyser may last 30 years...what durability or accelerated wear testing supports that premise? Marathons have been around long enough to suggest that they outlast conventional steel tanks, but Geysers are fairly new and comprised of metal....without visiting Kevin's link I suspect it provides good advice to renew the anode in conventional water heater tanks...good advice, but I'm not aware of a replaceable anode protecting Geysers.

  8. user-758514 | | #8

    Geyser with solar
    I recently installed a Geyser RO into a 40 gallon indirect tank off my System 2000 boiler. The cold water comes in from a 80 gallon solar heated tank. I am still tweeking the set point and differential on the Johnson controls A419. The sensor is mounted on the tanks bottom cold feed. I am currently set at 120F with a 12 degree differential. Inlet temperature the last couple days has been 107-110. The Geyser came with a 120 setpoint and 30 differential. This may work fine with a cold inlet but with pre-heated water it just insures that I have a tank of tepid water. Any words of wisdom on determining the best settings?

  9. kevin_in_denver | | #9

    Geyser has no steel to rust
    Response to Curt Kinder:

    The main component of the Geyser is a hermetically sealed compressor unit that looks just like the guts of a refrigerator. Fridges will last 30 years but usually get thrown out before that because of reasons other than compressor system failure.

    In the Geyser, all the wetted components are copper, and there is no tank to protect.

    Heat trap valves come in many different flavors, some of which can stymie a tiny pump.

  10. barkingeater | | #10

    is the Geysor still available

    1. GBA Editor
      Martin Holladay | | #11

      Q. "Is the Geyser still available?"

      A. No. For more information, see "Revisiting Heat-Pump Water Heaters."

      In addition to reading the article, you should check out Comment #2.

    2. Foton | | #12

      Almost 2 years later...
      Well, it looks like Nyle Systems might be readying a new "Geyser" after all: the e8.

      At 8 kBTU/h it may be usable for some (warm-climate) space heating applications too, possibly as an alternative to a Sanden/Eco2 SanCO2?
      Specifications lack details and are only preliminary, and as far as I can tell, no availability date or pricing has been released yet either.
      To be continued, I guess.

      1. charlie_sullivan | | #13

        On the comments of the article linked below, there's a report that it's due in April, 2022.

        For heating use, you'd need to duct it outside, and it says it only works down to 38 F. So the climates where that's useful are quite limited.

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |