The Monkey House

< Writing : Other : Heat Pump Water Heater >

Our water heater, a natural gas-fired non-condensing naturally-vented tank storage heater installed 16 years ago, is finally exhibiting unmistakable signs of age. It's undoubtedly repairable: either the temp/pressure relief valve is getting weak and is leaking, or the thermostat is dying which causes it to overheat the tank and the relief valve is venting like it's supposed to. But it's 16 years old. And it consumes methane with a lousy 59% efficiency since it's a non-condensing design, spewing unnecessary CO2 into the air. And it vents combustion byproducts into our basement for a full three minutes every time it starts up. Additionally, removing a source of combustion will let me cap off the 4" vent pipe that is effectively a giant hole in our house. All of which is to say I have little desire to keep the gas burning heater around.

So nine years after buying our house, after two big half-steps (basically no longer using our gas range, and doing most space heating with the heat pump - more about the big heat pump in a later post!) we will finally be taking the first full step toward electrification: completely removing a gas-burning appliance from the house! Though it would be nice to use something like the Sanden / Eco2systems heat pump which would serve as a redundant system to move ambient heat into the house during winter, local plumbers are unfamiliar with them and there is no local distribution or support. So I've ended up with a traditional style system from Ruud / Rheem. If you're in AO Smith territory, they appear to be equivalent.

Since I'll be 'opening the hood' on our home's plumbing, I'm doing some additional small improvements at the same time. Our home's design is not conducive to a proper hot water supply loop, and also not to an end-of-line recirculation loop. So instead I'll be doing a supply loop just in the basement, with stringers up to the first floor and also up to the second-floor bathrooms. This should dramatically reduce the time-to-hot for the first floor kitchen and laundry, and cut the upstairs time-to-hot by just over half. The water savings from this improvement will be slight: maybe 1-2% of our water use. And the energy use from keeping approximately 70' of insulated pipe hot will also be slight; I've not tried very hard to quantify it since I'm justifying this change as a quality of life improvement rather than an efficiency improvement. But I will be taking the opportunity to add additional sensors - pressure, temperature, and flow rate - to the plumbing system to let me quantify heat loss in the hot loop and to measure our total hot water usage.

Equipment Selection

Our outgoing heater had a 40 gallon storage tank with 73 gallon first-hour rating and 38 gallon per hour recovery rating. We have a large tub which draws enough hot water to empty the tank, so a slightly larger reservoir would provide some advantage. But our other hot water uses are sporadic enough and small enough to be of no concern: we do not have back-to-back demand that would be sensitive to first-hour or recovery ratings on a replacement heater. Plumbers are incentivized to maximize profit and minimize dissatisfied customers' call-backs if they run out of hot water, so they universally recommended a 65-gallon (or even an 80!) replacement when moving to a heat pump heater; this preserves or increases the first-hour rating as compared to our gas heater. But we will be saving $500 and installing a 50-gallon replacement instead, trading a lower first-hour rating for a higher reservoir capacity, as compared with our existing equipment.

The water heater we bought is a 'hybrid' heat pump. That means it is intended to operate mostly in high efficiency heat pump mode, but it also includes two 4500W resistive electric heat elements, like a traditional electric water heater. When hot water levels get low - or the user sets it to high demand mode - it will turn on the heating elements to supplement the heat gathered by the heat pump, for maximum recovery speed. There are models that will run from a 15A circuit, or even from standard 120V power, but they are slightly less efficient and have lower recovery ratings because they lack the additional heating elements. And presumably if there is a failure of the heat pump component, the resistive heaters will step in to provide hot water until the heat pump can be repaired.

The water heater is located less than ten feet from the main breaker panel in our house, allowing us to easily and cheaply run the 30A 240V circuit that the water heater requires. While they're out here I'm also having the electrician install a 50A circuit for an induction range and a 50A circuit to the garage for an EV charger (though in the short term I'll be using a splitter to break out to two 20A 120V circuits instead). More on that in a later post.


As is typical for these posts, I'll be setting down some expectations of how the equipment will perform, to be reviewed in a few years when I have collected data. I do not currently have any metering of our hot water use - just total water use from the city utility - nor of the existing water heater's natural gas use. So I have only rough estimates of energy demand for water heating. That said, I conservatively expect to be displacing about 300kWh (or 10 therms) per month, roughly 1/3 of our current household natural gas use. The increase in electricity use is harder to estimate. If I pretend the basement is a free source of infinite 60°F air, then I expect to add roughly 50kWh/mo of electricity in order to heat the same amount of water. Ten therms, or ~300kWh of heat energy in natural gas works out to about 180kWh of heat added to the water (0.59 fuel usage efficiency); at a COP of 4, that would take 45kWh of electricity for the heat pump to perform the same heating task. I've rounded that up to 50kWh.

But there's seasonal variation to take into account. In the summer, any heat I pull out of the air and dump into hot water (that goes down the drain after a shower) is heat that I don't have to pull out of the air and send out of the house using the air conditioner. In other words, that electricity use is "free" since I would be paying for it anyway, just using a different appliance. Conversely, in winter time, heat from the 'warm' basement air that I am stealing to put into hot water is heat that has to be replaced by the air-source heat pump that heats our house, so I have to pay its electricity cost twice. Let's pretend the 'free' summer power and the 'double cost' winter power cancel each other out: I expect -300kWh/mo in gas use and +50kWh/mo in electricity. That's about 55kg per month of CO2 not coming from the old gas water heater; with Wisconsin's 2023 electric grid carbon intensity of ~540g/kWh, the 50kWh of electricity adds 27kg of CO2, for a savings of just over 50%. As the grid cleans up throughout the life of the heater the savings will increase, to a reasonable near-term limit of about 90% carbon-free electricity. If the new water heater lasts 16 years like our last one did, that's a lifetime savings of a bit over seven tons of CO2. For comparison, that works out to about 10% of our current annual direct CO2 emissions (electricity + natural gas + gasoline) so it seems like a meaningful contribution. Or looked at another way, seven tons is about five times the embodied carbon from the construction of the heat pump water heater itself.

Burning 'good' air?

There's one last effect of switching from combustion to a heat pump water heater to consider. Any time the old gas heater ran during the summer, it also took chilled air out of the house and sent it up the flue. That air got replaced by warm humid ambient air leaking back inside, which requires additional energy to cool and dehumidify. And in winter time, it took warmed humidified air out of the house and sent it up the flue, to be replaced by cold dry air leaking back inside which - you guessed it! - requires additional energy to heat and humidify. So in addition to the increased efficiency at the appliance level, the switch to a heat pump water heater will decrease energy used by the rest of our home's systems as well.

But by how much? One therm of natural gas is about 100 cubic feet. And you nominally want about 10 cubic feet of air per cubic foot of gas for good combustion. But typically furnaces are designed for +50% combustion air flow to have a safety margin, or 15 cubic feet of air per cubic foot of methane. I have no idea how much air goes up the flue of a naturally vented water heater, but let's start there as an approximation, knowing we might be off by a factor of two or more. Ten therms per month for water heating - the estimate I used above - means 1,000 cubic feet of gas, or 15,000 cubic feet of air. That's about one air change in a 1500 square foot home. Home air leakage rates are given in air changes per hour, not air changes per month (~700 hours). In other words, the additional efficiency gain from not burning indoor air and exhausting it outside the house is likely to be quite minor: less than 1% of our home's heating or cooling demand is due to the water heater exhaust, even if we've significantly under-estimated the flue leakage rates by totally ignoring the fact that the flue is a 4" hole in the house that leaks air even when the water heater is not running.

Version 1.0     |     Originally written: May 18, 2024     |     Latest revision: June 8, 2024     |     Page last generated: 2024-06-08 08:09 CDT