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

Community and Q&A

Choosing a Heating System

Gillit62 | Posted in GBA Pro Help on

We’re re-doing our heating system in Zone 7a and have been looking at using air source heat pumps with a buffer tank and in-floor radiant heat.  My contractor also suggested taking a look at solar PV with a combination of electric in-floor heat and radiant Cove heaters.  He’s apparently worked with an engineer that calculated the cost of the solar array would be paid back within 3 years by eliminating the boiler work and taking advantage of tax incentives.  What’s the collective wisdom on these alternatives?  My current cost for electricity is 13 cents/kWh.  Thanks!

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. paul_wiedefeld | | #1

    The radiant heat hasn’t been installed yet? How is the house currently heated?

    1. Gillit62 | | #3

      Previous in floor staple up (upper floor) and QuikTrak (lower floor) has been removed so clean slate on choice of systems to heat going forward.

      1. paul_wiedefeld | | #5

        Got you: there is definitely a lot going on (in an interesting way)
        1. Solar should be a go either way at that electricity rate and with tax credits. The question then becomes how big an array do you need/can you fit.
        2. Boiler work is definitely expensive and air-to-water heat pumps are uncommon. A ducted/ductless heat pump would be more efficient and probably be substantially cheaper/easier to source and repair. If you're willing to entertain air-to-air then:
        3. What's the heat loss of this house? At a low enough heat loss (assuming air conditioning isn't a huge deal in your climate) duct work can also become expensive. If it's super low, ductless might be expensive relative to just using resistance. If you're willing to give up whole house radiant, you can mix and match between ducted, ductless and/or electric resistance based on what makes the most sense. That way you can have electric radiant floors in the bathroom(s), ducted in the largest areas and ductless in a harder to reach room if needed.

  2. Expert Member
    NICK KEENAN | | #2

    You're mixing a lot of apples and oranges there. Solar vs. grid, heat pump vs. resistance heating, in-floor vs. Cove.

    Each one is a separate conversation.

    1. Gillit62 | | #4

      Thanks; agree those are all separate conversations, my original hope was for air source heat pump run off solar PV, battery and grid when necessary with a propane boiler backup for the 112 hours/yr when ambient temp is below zero F, connected to in-floor hydronic radiant. I'll still have solar PV, just wondering if electric radiant (wall) and in-floor resistance is better.

      1. Expert Member
        NICK KEENAN | | #6

        Resistance heating is cheaper to install than a heat pump but uses probably three times as much electricity to get the same heat. If you're aiming to produce all your electricity with solar the question is whether a three times as big solar array costs more than the difference between a heat pump and resistance heating. You need to price it out to see, there's no fixed rule.

        There's a reason air-to-air heat pumps are the most popular, they give a good balance between efficiency and initial outlay.

        The Cove heaters make a lot of claims about producing greater comfort with less heat than other heating systems. Don't believe it, if it were true everyone would do it.

        1. Expert Member
          MALCOLM TAYLOR | | #7


          I specified cove heaters in two h0uses. One owner balked at the price difference from baseboards, the other put them in and the only advantage they can see is they are up on the walls as opposed to being in the way on the floor.

      2. Expert Member
        BILL WICHERS | | #8

        I very much recommend you to not bother with batteries. Batteries are a big cost and maintenance item for solar installations, and are really only a good idea when you're off grid and have no choice. For all other systems, you're better off just using a grid tie system and using grid power when solar is not available. You don't need to have a net metering rate available to be able to do this, either -- you can either run as a "peak shave" system, where solar provides power to as much of your load as possible when it's available, and you use grid power the rest of the time, or with net metering you sell back your excess solar when it's producing.

        With a peak shave system, you'd want to over heat your home during the day to use it as a thermal battery and carry you into the night as much as possible. Exactly how much of a system you'll need to do this depends on a number of things (size of the home, insulation and air tightness, local climate conditions, etc.), so it's not easy to guess at here. I think you'll find it easiest to shoot for a "as much as I can fit in for $x" type of system and then adapt as needed down the road. Heat pumps will let you get the most from your solar system in terms of useful heat, but they'll cost more to install too.

        My guess here is that the extra cost for the heat pumps is likely to be the better overall cost/benefit compared with putting the savings into a larger solar system and using electric resistance heat. The reason for my thinking is that the heat pump allows a smaller solar system during the day, but also requires less grid power at night when the solar system isn't producing. That means the heat pump is saving you money ALL day, where the larger solar system is only helping you during the day when it's producting, and there are limits to how much of that energy can be stored by pre-heating your home (or with actual batteries if you were to go that route).


  3. Expert Member
    NICK KEENAN | | #9

    The cost-effectiveness of solar varies enormously depending on local regulations and subsidies. I installed a solar array on my house in Washington DC and it paid for itself in about 3.5 years. But I estimate that 90% of the benefit was from government subsidies, either tax credits or SREC's, and net metering. DC has I believe the most generous SREC programs in the country as well as one of the best net metering programs. Without government support the payback would have been 35 years. So you need to find out what you local government is doing.

    1. Expert Member
      BILL WICHERS | | #11

      It depends on how you size the system too. For me, here in SE MI, a ~3kw system will pay off in about 3.5-4 years or so with NO subsidies at all. That is with me doing the installation, a grid-tie system with no batteries, no net metering, and a time-of-day utility rate. The system pays for itself by offsetting daytime load during most of the on-peak hours on weekdays, and gives a little bonus on weekends (when power is much cheaper), and it helps out in the winter too, but much less.

      The important thing is that solar can make economic sense on a small scale, even without gov't subsidies, if you size it appropriately, and only try for modest goals (offsetting your daytime load) instead of lofty "net zero" goals.


  4. Expert Member
    NICK KEENAN | | #10

    I've seen descriptions of solar heating systems that run DC straight from the panels into a resistance element -- no inverters, no grid tie, no batteries, just a thermostat. Basically like an old-school solar thermal system but using electrons instead of water as the medium. They claim higher efficiency and lower cost but I don't know how it compares to a system with an inverter and a heat pump.

    I would wire it up so that when the thermostat is off the panels would power inverters and produce useful electricity.

    1. Expert Member
      BILL WICHERS | | #13

      Such a system would work fine, and you would get pretty good efficiency since you have very few energy conversion steps. A heat pump would be better whenever the COP of the heat pump allowed it to deliver more BTUs than the absolute energy available with the resistive system. With the resistive system, your only real losses are in the wiring.

      Note that if you try this, you need a special DC-rated contactor to control the on/off cycling of the heating element(s). A regular thermostat can't really handle any amount of power, and regular contactors that would appear from their amp rating to be OK probably won't be in a DC system, since DC is much more difficult to interrupt. The reason for this is that things that switch AC rely on the zero crossings -- of which there are 120 every second -- to help to turn things off, since those zero crossings will kill any arc that forms across a set of switching contacts. DC doesn't have zero crossings, so the arcing issue is a much bigger deal when switching DC.


      1. robgood | | #15

        Excellent insight on the DC power arcing issue.1

      2. Expert Member
        NICK KEENAN | | #16

        Sounds like an automotive relay on the thermostat would do the job.

  5. onslow | | #12


    You haven't shared envelope insulation levels or a floor plan, so if you have a well insulated, but less than open floor plan, you may find a blending of two systems (cove heaters and mini-split) to be advantageous. I would have done my own home with this combo when building except heat pumps were almost unheard of locally and insanely costly. My envelope insulation is quite high, but my floor plan was once described as a "rabbit warren".

    A great deal of duct work would have been required with tradtional forced air to provide heat and return air runs. Estimates were well north of 25K for a (no AC) heating install, which also would have tied us to needing propane as natural gas is not available to me.

    A potential mini-split installation would have taken two units, each with multiple heads or possibly ducted heads, options frequently noted in GBA Q&A postings as less than optimal due to minimum head sizes or dial down of outputs. The issues of fan noise, line set lengths, condensate lines and aesthetic concerns plus insane installation charges nixed this choice. We did lose the AC bonus of mini-splits, but our climate and home design has made it possible to not really miss it.

    The option of doing baseboard over cove heaters was briefly discussed, but nixed due to furniture placements and pets. Hair is not an issue with cove heaters, nor is furniture placement. Our envelope and triple glaze windows make it quite unnecessary to place heat sources under windows as traditionally done. The aesthetics may leave some cold, but they are certainly no uglier than cassette units. Plus they are silent.

    I investigated solar early on and again, local conditions made the choice sketchy. Being advanced in years, the ROI meant only my heirs would likely start seeing "free" energy. This despite favorable net metering and government rebates. The output of a solar array is of course biased to summer input, so net metering would be perfect for beating down the winter electric charges. Be very cautious of actual output estimates during the winter. Your heat demand is highest, solar output lowest and snowy cloudy days will degrade any size array's output. Also be mistrustful of the way sales people project rising energy costs. The greening of our overall grid is driving costs at the municipal/state levels down sharply. Ask you local coal mining operations how that's going for them.

    If you wish to make heating system cost savings into a solar array just be sure that your load and output needs are balanced by some form of assured energy banking with your grid supplier. Many don't like the solar trends and there are numerous cases of rate structure changes that go very harshly against solar. The fight is still on against individual energy.

    So, if you have hot and/or humid summer temps, I would definitely go with some form of mini-split to gain the AC function. How you get the AC to successfully cool bedrooms is going to be the challenge. Similar issues will be present during heating season, especially when down around zero all night. At least in this situation, cove heaters can help with providing a local (controllable) boost for those not desiring to sleep like a Scotsman.

    Except for perhaps a bathroom, I would avoid under floor electric. Failure of the floor sensors twice in two years along with a erratic thermostat put me off the idea. Per square foot costs of the wiring is very high and affects your flooring choices. Carpeting or an envelope close to PGH levels will mean not experiencing the warm toes everyone desires. You could with effort, but keeping a house at 85 is usually not desirable. Mostly, when it screws up, you have a major problem on your hands.

    So, to provide some actual experience: My wife and I were quite done with forced air systems and wanted to advance to an all electric home free of gas. With mini-splits off the table economically, we went with all cove heaters on advice of our electrician's past experience installing them in other high altitude homes. All 18 heaters, 12 thermostats, wiring AND installation set us back approximately 6K. We pay 15 cents a Kwh and average annual heating costs in CZ6B is $1500. Our resistance water heater is the big pig in our $3100 annual electric usage. I won't explain now why we passed on a HPWH. Having saved over 15K by not going with forced air or a wildly over charged mini-split installation, I feel reasonably good about my choice.

    I essentially have 10 years of free heat at a minimum. Avoided propane charges extends my "free ride" further, as does my lack of service calls or repairs. My earlier experience with a computerized high efficiency furnace was not favorable. Neither has my experience with AC units integrated into forced air systems. Fears of winter freeze-up or snow issues with a mini-split are not a worry for me. Any potential efficiency savings with the mini-splits were eaten up by the front end costs and risk of expensive repairs should a unit go down. Additionally, the higher COPs for units 7 years ago were only achievable in modest tempertures. This, of course, is not now the case, but CZ7 still presents a more challenging overall environment. Review your room layout, individual loads and then review how you get the heat and cool where needed.

    1. Expert Member
      BILL WICHERS | | #14

      >"The greening of our overall grid is driving costs at the municipal/state levels down sharply. Ask you local coal mining operations how that's going for them."

      That's not true. I agree that solar companies are likely to overinflate potential future energy costs, but "greening of the grid" is NOT reducing costs. Costs have been increasing. My main work is in the telecommunications in industry and my typical customer has at least a $10,000/mo electric bill, many are up in the six figures -- every month. Obviously we watch utility rates closely, and we work with the utility companies with cost projections. A lot of my projects are energy efficiency projects, and I often get paid a percentage of the savings I'm able to gain for a customer over a period of years. Energy costs are a BIG deal, and they are NOT going down.

      Most of the new generation that has been replacing coal fired plants has been new natural gas fired plants, and some conversions of coal fired plants to use natural gas instead. You can see this in the generation mix graphs of most of the ISOs in the Eastern US (MISO, PJM, etc.). It's not wind and solar replacing coal, it's natural gas. The main reasons for this is that natural gas prices over the past 15-20 years have been lower per BTU than coal, and natural gas has a lot less emissions per unit energy too which makes the plants cheaper to operate. If those natural gas plants were to be shutdown or retired, energy costs would skyrocket and system reliability would plummet. We may actually see some system stability issues this winter if the projections for a colder than usual winter come true.

      BTW, it is for these reasons that I discourage people from going "all electric" if their goal is reduced emissions, UNLESS they use heat pumps. Heat pumps gain you in terms of overall emissions because they can provide more heating BTUs than the BTU content of their input energy. This is because they move heat, they don't make it. If you use electric resistance heat to replace natural gas fired heat, you are very likely to actually INCREASE your overall emissions because the total system efficiency of the electric system, when using primarily natural gas and coal as fuel as is the case in much of the US, is lower than if you were to just burn the natural gas for heating directly.


    2. Expert Member
      NICK KEENAN | | #18

      "If you wish to make heating system cost savings into a solar array just be sure that your load and output needs are balanced by some form of assured energy banking with your grid supplier. Many don't like the solar trends and there are numerous cases of rate structure changes that go very harshly against solar. "

      In fairness, if you build a net-zero house with net-metering, your power company gets nothing from you, but bears the cost of maintaining the grid that you get the use of for your power banking. I've gone five years without paying an electric bill, and that's not sustainable. We need to work out a new billing model that fairly compensates the distribution companies for the service they provide.

      People in the industry talk about "grid abandonment." What they're worried about is not so much individual residential accounts, but big commercial users -- universities, hospitals, data centers. When one of those decides to put thousands of solar panels on a back lot and go net zero the utility loses hundreds of thousands of dollars of annual revenue -- but still has to provide the same level of grid capacity.

      There's also an equity issue. Without reform, as more people go net-zero, the entire cost of the grid gets shifted onto the remaining customers. Broadly speaking, who buys solar panels? People with money. People who can afford the up-front cost, and who own their own property to put them on. So more and more of the burden gets shifted onto the poorest.

      These are real issues.

      1. Expert Member
        BILL WICHERS | | #20

        >"but big commercial users -- universities, hospitals, data centers."

        Hospitals and data centers don't have enough land to make enough power with solar to run their operations. It's not even an issue there. I would imagine most large commerical operations are similar, although I have most of my experience with the telecom world (data centers), and after that some with hospitals. A typicall smallish data center is the equivalent of around 700-1,000 homes in terms of electrical consumption, large ones can be many times that.

        I agree about the issues with net metering, and it's not just economic in terms of billing, there are very real issues with power flow too. I think a fair model for billing is to bill customers for their power use AND production on the distribution side, since net metering effectively uses the distribution network twice: once in each direction. Have the utilities buy excess solar production at the wholesale market rate, possibly the spot market rate (which can be very much higher than the long-term rate when capacity gets constrained). This way the utility isn't paying a premium for people's excess solar power which the utility cannot count on for planning purposes. I'd even be OK with a discounted rate here, since again, the utility can't count on it, so they can't plan for it. Have the net zero customer buy power at the regular retail rate from the utility, since they're buying power from the plants just like anyone else when they're doing that. Basically net zero customers are regular electric power consumers like anyone else when they're not producing.

        This way the distribution costs are covered, net-zero customer power consumption is paid for at normal rates, so it's not a detriment to the utility, and excess solar production gets purchased as what it is -- excess (surplus) electric power that cannot be planned on and is not dispatchable, so it's of less value to the utility than the commercial producers that can be relied on are.


  6. robgood | | #17

    Great contributions.

    I am in Duluth, MN, where the ambient temperature may vary 125 degrees F over a year. Add wind chill in the winter. The house I bought was well constructed as an All Electric house in 1978; including in-ceiling radiant heat! Successive owners added gas fireplace inserts which I am updating. I have two levels, each of 1500 ft^2 . The NG fireplaces offer very cozy base load heat of 30K btu (lower level) and 40k btu (upper level). I had installed a 48K btu Mitsubishi high efficiency heat pump compressor driving 5 head units. I plan on an additional 36k unit with 3 head units next year. The efficiency tables claim 25% of max output at ambient temps of -15 F. I’m skeptical, and maintained all the wiring and breakers for the radiant ceiling heat in the event the Mitsubishi fails or the ambient temp goes to -40.
    Next year I’m having the insulation contractors in to redo that critical system.
    Part of the rationale for the Mitsubishi is that this summer we had forest fire smoke and high temps in the upper 90’s F. I may not get the full benefit of the AC side, but I am sure that the capability will be attractive to buyers. When I grew up AC was rare in Minneapolis and unheard of in Duluth. Temps have gone up so much in this state that AC is practically a requirement over $400K housing prices.

  7. onslow | | #19

    I perhaps should have been more clear with my reference to production costs going down with increasing solar and wind installations. I did say at the state/municipal level by which I meant wholesale grid sales. I may be also guilty of a certain level of parochialism. Our local grid is very focused on sourcing green energy and our major provider is set to achieve 50% renewable by 2024. We do have a distinct advantage of low industrial demand on our grid, which makes accomodation of demand fluctuations less challenging. Our cost per kWh has been stable for five years. I do note that despite a major shift toward renewable sources, the people on the front range in Colorado are being hit with double digit increase requests. I have also noted that Minnesota is facing the same rate increase requests from their providers.

    Just the same, production costs for solar have dropped by 90% between 2009 and 2020. Current production is roughly 4 cent kWh global average with Saudi desert installations hitting 1 cent. Wind produced electricity ranges from 9 to 3 cents a kWh which may fall further with the newest windmills that produce up to 15 MW per unit. Vestas, GE, and Siemens all have truly huge windmills slated for the next few years. Offshore leasing is targeting a goal of 30 Gigawatts of production by 2030, which admittedly is a bit ambitious. Still, fuel free energy is going to be needed if we are to make any headway with global warming. I am really anxious to see if Helion is finally going to bring fusion to actual grid use.

    The newest coal fired plants production costs are 11 cents kWh and massively detrimental to our future. Locally, our coal mining operations have shut down due to loss of market. The lower costs of kWh production from gas, wind and solar will continue to render coal fired sources uneconomic. Nuclear is said to be 16 cents kWh, which does not likely cover the 60 years of massive subsidies and decommissioning costs to say nothing of the deferred waste storage issues. A major urban provider closed three nuclear plants over the last decade or so. Natural gas at 6 cents is the current go to for numerous reasons, but the environmental impacts are only marginally less than coal fired, just distributed differently. Actual methane leakage associated with production has been found to be at least 30% higher than believed.

    The google says that nationally we have 3300 utilities, 7700 power plants and 200,000 miles of high voltage power lines and 5.5 MILLION miles of local distribution lines. Long term embedded costs, regulatory messes (politics) and deferred upgrading expenses all work against seeing price reductions for delivered energy. The wholesale costs are shifting, but the utilities certainly don't like to share. The issue of grid tied solar being unfair strikes me as suspect or at least a regulatory/political distortion.

    My limited understanding of hook up fees either for gas or electric is that they are supposed to be for grid or pipe network maintenance and captial improvement. When I did have natural gas at a prior home, the source of the gas was separately selectable and charged separately from the delivery of the gas. My fixed costs for accessing the natural gas system was close to $40 a month regardless of gas used.The electric provider also charged delivery fees that ranged around $40 though it fluctuated with total delivered kWh. With a customer base in my city of nearly 4 million and 10 million system wide I have to suspect that grid tied solar was not going to ding them in any real way due to the monthly connection fees.

    Perhaps it would have been more precise to say that the rate of increases SHOULD be tempered by the availability of green energy sources. The solar salesmen I dealt with projected 3-5% per year to support payback periods that don't really hold water. I also agree that heat pumps are far and away better than straight resistance heat with newest units far outperforming what was available when I built. My local conditions distorted the choices I had to work with. If I didn't make that clear, my bad. I would say that an all electric house will be the future and our local electrical provider agrees. They are pushing heat pumps now and three new solar farms are scheduled by 2024-5

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |