GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Audio Play Icon Headphones Icon Plus Icon Minus Icon Check Icon Print Icon Picture icon Single Arrow Icon Double Arrow Icon Hamburger Icon TV Icon Close Icon Sorted Hamburger/Search Icon
Q&A Spotlight

Is a Ground-Source Heat Pump the Right Choice?

An owner-builder hoping for net-zero performance wonders if a radiant-floor distribution system makes sense

An elevation of the house that Andy Zbojniewicz plans to build in Michigan next year.

In Grand Rapids, Michigan, Andy Zbojniewicz is planning a large, high-performance home that ultimately will be powered by a 15-kW solar array. The house will have more than 3,000 square feet above grade, and another 1,633 square feet in the basement.

His immediate concern is how to heat and cool the house.

“I was told the home was too large to heat and cool with air-source heat pumps and be comfortable,” Zbojniewicz writes in a Q&A post,  “so [I] was primarily looking at geothermal.”

To that end, Zbojniewicz sought quotes from two HVAC contractors for ground-source heat pumps. One of the contractors said the heating load would be 50,797 Btu per hour at an outside temperature (design temperature) of 7°F, and recommended a 5-ton (60,000 Btu/h) heat pump. The second company didn’t provide its load calculations, but suggested that Zbojniewicz would need two 3-ton heat pumps in order to stay comfortable.

“I have a friend who is in HVAC and he feels strongly that I should have radiant tubing run in the basement to help with heating in winter,” Zbojniewicz adds, “and while I was at it I was going to run it in the mudroom, master bath, and three-season porch.”

After doing some reading at GBA, Zbojniewicz realizes there are plenty of critics of both ground-source heat pumps and radiant-floor distribution systems.

“Is there a better alternative to geothermal if I want to have the potential for net zero?” he asks. “If I stick with geothermal, are two 3- ton heat pumps overkill (they’re certainly more expensive)? Should I abort the radiant altogether?”

That’s the topic for this Q&A Spotlight.

There is nothing wrong with air-source heat pumps

The outside design temperature for Grand Rapids, Michigan, is 5° or 6°F, GBA Editor Martin Holladay says, so Zbojniewicz can certainly consider air-source heat pumps. A heat pump from a U.S. manufacturer paired with a conventional forced-air duct system would be fine. Holladay suggests that Zbojniewicz read a GBA article, “Ducted Air-Source Heat Pumps from American Manufacturers.”

Dana Dorsett adds, “I’m aware of a house in Vermont slightly larger than yours and slightly less insulated, in a location with a 99% outside design temp of -12°F that is heated with four ductless mini-splits. The notion that your house is ‘too large’ to heat with a heat pump is just silly. It’s the size of the heat load, not the house, that matters.”

There are some pluses to a ground-source system

Jon R contends that radiant heat paired with a ground-source heat pump is great — as long as Zbojniewicz doesn’t care about how much the system costs. Such a system not only provides space heating but also provides efficient water heating.

“If you are looking at air-to-water geo-exchange, then you can add a water tank (thermal mass), resolving any heat pump over-sizing issue other than cost,” Jon R adds. “If very large, it can enable some use of solar PV that doesn’t involve net metering. Multi-stage or multiple heat pumps also mitigate over-sizing.”

Those load calculations are probably wrong

The capacity of any system that Zbojniewicz installs should be based on two factors: the climate, and how well the house is air sealed and insulated. And Dorsett doesn’t have much confidence in the heat load estimates that one of Zbojniewicz’s contractors has provided.

“Don’t believe the 50K heat loss number unless it was done by a qualified third party,” Dorsett writes, “an engineer, RESNET rater etc.,  somebody who makes their living and reputation on the accuracy of their numbers rather than installing and maintaining HVAC equipment.”

Even a code-minimum house of that size with a measured air tightness of 3 ach50 would have a heat load of less than 40,000 Btu per hour, he says, assuming an outdoor temperature of 0°.

“The only way to stretch it to 50K would be to have excessive air leakage or excessive expanses of window area,” Dorsett says. “Your house is likely to come in closer 30K @ 0F, 36K tops and even less at +7F.”

With that in mind, a heat pump with a capacity of 60,000 Btu/hour is about twice the size that Zbojniewicz really needs. The place to start, Dorsett says, is a Manual J heat load calculation, and it’s his bet that two modulating Fujitsu minisplits or possibly a 3-ton Carrier Infinity Greenspeed air-source heat pump with heating strips would keep the family comfortable.

Steve Theinalienable’s experience bears that out.

“My house insulation, climate, windows, and foundation insulation are almost exactly the same as yours, just a bit smaller (2200 square feet),” he says. “My design heat load at -18°C [roughly 0°F] is 18K Btu, but we had a day like that this week, and my air-source heat pump didn’t have to work that hard to keep up, so, I think it’s probably a bit conservative.”

Get your house rated by a pro, Theinalienable says.

Consider the all-electric option

With a tight building envelope and a relative large photovoltaic system, why not consider an all-electric house? asks Armando Cobo.

“I would highly advise you to consider going all-electric with an air-to-air heat pump, and hopefully installing [Energy Star] appliances and lighting, electric fireplaces and an induction cooktop,” he says. “That’s what I spec on all my houses, which are [zero energy ready homes]. Just a thought!”

In fact, Zbojniewicz is trying to convince his wife that an electric fireplace and induction cooktop are good choices.

Zephyr7 suggests he invest $50 or $60 in a portable, one-burner induction cooktop and experiment with it.

“Have your wife play around with it,” he says. “It’s an inexpensive way to test out induction cooking. Note that not all pots are compatible with induction cooking.”

Insulating the rim joist

One footnote to Zbojniewicz’s efforts to obtain more accurate heat load estimates involves insulation. He’s found a company that will do Manual J calculations, but he’s not sure how to specify the insulation he will use at the rim joist.

What’s the best approach?

“Closed-cell [spray foam] is pretty standard for use on rim joists,” replies Zephyr7. “Lots of people use kits like those from Foam it Green and the Dow Froth-Pak. What you want to avoid is keeping the wood wet. Your assembly sounds like it won’t be a problem, but make sure you have a capillary break between the rim joist and the top of the foundation wall.”

Cobo says open-cell foam is best at the rim joist when rigid foam will be applied to the outside of the wall. Because Zbojniewicz is planning to use Rockwool on the outside of the house, either open- or closed-cell foam can be used, he adds.

Holladay suggests he choose closed-cell foam, not open-cell foam. “In a cold climate, closed-cell spray foam would be a better choice for the interior of a rim joist,” he says. “You’re right that either type of spray foam is an air barrier. The reason that closed-cell spray foam is better is because it is a vapor barrier, unlike open-cell spray foam. The closed-cell spray foam prevents outward diffusion of water vapor, and therefore keeps the rim joist dryer.”

Our expert weighs in

Peter Yost, GBA’s technical director, adds this:

Cold-climate heat pumps: It is interesting to hear the misgivings that air-source heat pumps will be unable to manage either larger homes and/or cold climates. I was just in Wisconsin recently and heard the same litany. Dana Dorsett, as usual, is right on: Heating loads are assessed using U-factors and building enclosure surface areas (and other metrics, including airtightness), not square feet of living or conditioned space.

In the Northeast, heat pumps designed to wring out BTUs at air temperatures well below 0°F have proven efficient and reliable. Marc Rosenbaum, a leading and inherently skeptical mechanical engineer, has helped spread the word by recounting his own great success with cold-climate air-source heat pumps. We did have problematic cold-climate heat pumps in the past (both Nyle and the Hallowell cold climate heat pumps had problems and went out of business some 10 years ago) and that news seems to persist. The technology of today’s cold-climate heat pumps is very different and has proven to be robust. (See numerous GBA resources on the topic, including this one.)

Radiant-floor heating. We just can’t seem to get our arms around the pros and cons of this method of heat distribution.

It is not more efficient. There is nothing magical about radiant-floor heat. Its efficiency is driven by the same physics that governs all other distribution systems. And part of the physics is the mass to which the distribution system is connected. It’s very difficult if not impossible to gain efficiency by way of thermostat setback.

It is more comfortable. It’s hard to argue with this, either empirically or based on heat transfer. In my experience, well-designed and installed radiant floor heating systems deliver superior thermal comfort.

Radiant-floor distribution can be used for space cooling. As Robert Bean has said more times than he cares to count: We keep blaming condensation during cooling on the radiant system when, if the latent load were properly managed, radiant cooling can work just fine (for more, see this).

Insulating the rim joist: In my experience, rim joists get wet from backsplash at grade from the exterior or, less commonly, wicking because there is no capillary break between the mudsill (or sill beam) and the foundation wall.

In cold climates, I don’t see rim joists getting wet from air leakage because the stack effect is pulling cold dry air in during the winter at the rim joist. So, the main reason for me to choose open- or closed-cell spray foam at the rim joist — since both can provide a good air seal at this location — is based on how much drying potential I need to the interior (open-cell spray foam being more vapor-permeable than closed-cell).

Manual J: It’s really heartbreaking to hear how often heat-load calculations are not done. Can you imagine a car company designing a transmission system without knowing the horsepower and torque the engine can generate? On the other hand, it’s heartening to see how often the general public ends up on GBA’s Q&A pages to work with leading building professionals on home performance.


  1. Alan B | | #1

    Theinalienable is an awesome last name. Almost makes we want to change mine to match!

    I don't disagree with the solutions presented, though i would like to hear how it all turns out, even if its a while from now.

    I have wondered how cold air source heat pumps can go, in my case someday i might want to retrofit it. I live near Toronto Ontario, 99% design temp of -20C but we have had -30C and colder in very recent memory

    1. User avater
      Dana Dorsett | | #2

      >"...someday i might want to retrofit it."

      That "someday" could be "today".

      Gree has cold climate mini-splits with capacity tables that go as low as -30C (wet-bulb). Cold climate Fujitsus & Mitsubishis are all specified down to at least -25C (dry bulb). Many minisplits have a specified capacity at -20C (dry-bulb).

      Toronto isn't really all that cold- despite the occasional dip below -25C the 99th percentile temperature bin is still only -17C per ACCA tables ( ). The -20C would be a 99.6% design temp (ASHRAE tables peg the 99.6% design temp at -19.4C at the airport: ).

      Go ahead and use -20C if you like.

      If a mini-split is sized with even 1.2x margin on the load at -20C it will almost always cover your load to some temperature below -25C before it begins to lose ground, even if the capacity tables for that don't go that low.

      1. Alan B | | #5

        Today i have other priorities. I was thinking about starting a Q&A thread about my insulation situation since i have a government inventive that will cover it but i have century old home concerns. They did the energy audit a few weeks ago and i got the e-mail that i am eligible on Friday but have not looked it over yet.
        I also have an almost new furnace and other renos and problems that take priority but someday.
        Can you elaborate on what wet bulb and dry bulb mean in this context?

        1. User avater
          Dana Dorsett | | #7

          At cold temperatures the difference between WB & DB temps isn't very much, but it's historically common to specify the capacity using WB, since the wet bulb temp determines how much time it will spend in defrost. The AHRI protocol for testing heat pump efficiency specifies both WB & DB, but the capacities are specified using the DB temps. For instance:

          Note the ARHI DB + WB test conditions listed in the table on page 4.

    2. User avater
      Stephen Sheehy | | #3

      Our 99% design temperature is -18C and entering our fourth winter, our cold climate heat pumps (Fujitsu) have never had a problem putting out heat, even when much colder than the design temperature.

      1. Aaron Eugene | | #22

        Been reading many of your old posts in regards to your build in Maine back in 2015. I am in the process of new construction myself and (set to begin moving dirt in next month) and am also a fellow Mainer. I had a few questions regarding the ductless Mini-pumps I would love to get your opinion on. Would you mind sending along your email? Mine is, [email protected]
        Hope to hear from you

  2. User avater
    Jon R | | #4

    > Radiant-floor heating...It is not more efficient.

    With a heat pump or condensing boiler, I disagree. Large radiator = lower supply temp = more efficient.

    A radiant floor may (or may not) have enough radiator mass to prevent effective setback. Both the mass and the heat source matter.

    > “If you are looking at air-to-water geo-exchange, then you can add a water tank (thermal mass), resolving any heat pump over-sizing issue other than cost,”

    A typo. Should be "at to-water heat pumps, then".

    1. Charlie Sullivan | | #21

      The potential to realize the benefit of increased efficiency with a heat pump supplying a radiant floor is higher than it used to be, now that we have variable speed air to water (chiltrix) and water-to-water (hydron module) heat pumps. Especially set up with an outdoor reset, they can run at very high COP much of the time. The benefit they get from using low-temperature water is even more than what you get with a condensing boiler.

  3. Roger Berry | | #6

    Air to air heat pumps are pretty clearly able to provide sufficient capacity for homes with your (approximate) heat load. That said, it should be more of a concern for you to look at distribution of that heat. Martin is quite wise in saying he would go with a single head in a central air distribution system. This will make a lot of decisions commonplace to the HVAC people. I also believe it will make the overall comfort of the home better than discrete heads.

    I also support the idea of induction cooktops most vigorously. We are at high altitude and only have propane as a gas option. We do miss having a good burner for the wok, but now that we have induction compatible pans, gas is the past.

    As for heating. My own home is quite similar in size, with an equally large basement that is 80% below grade. Your's seems to be half exposed, so loads may be very different. We have a large percentage of glass area and a design load at -10F of 19,000 BTU/Hr. For a number of reasons, we went with cove heaters in all rooms and avoided the clearly more energy efficient mini-splits. The primary reason is the room layout and three levels to deal with. Some people have referred to our floor plan as a rabbit warren, which is a bit far, but we definitely do not have an open floor plan. Secondly, outrageously high installed costs and sketchy support options during the years we committed to design choices.

    With now three winters experience I can say that discrete heads and mini-splits would have been a very poor choice for us. Many recommendations would have had us place heads in maybe two locations per floor, which for our floor plan would result in strong temperature gradients from room to room. The air distribution we installed was intended for fresh air not heat distribution so a "return" path for each room is not provided. A central core with pickup at the third level is our choice for balancing air quality. For us it works well and we don't have the sound of a central blower to live with. ( a personal phobia of mine)

    The relevance to any home design is how the heat input will be distributed and how it is lost. I do believe that more than a few postings on GBA have cautioned about difficulties with hallway located mini-split heads getting sufficient heat to bedrooms, particularly when doors are closed at night. Martin has mentioned the failure of the Magic Air Arrows to live up to diagramatic expectations in several contexts which bear re-reading. It also supports why he suggests a central head with standard air handling layout.

    Radiant floor heat, though scorned by many, does have the virtue of putting heat into the rooms evenly and without reliance on air movement. It will also place warmth in the very rooms with the highest losses through walls and windows - bedrooms with closed doors. A long cold winter's night with a closed bedroom door will not faze radiant floor heat (or cove heaters). It is, however, a very costly and mechanically complicated way to get even distribution of the heat created by whatever method. By the estimates I did in preliminary design work, quite a bit more costly. It also adds a subtle threat of leaks to deal with over time as well as complicate flooring choices.

    I think rather than going on about whether mini-splits can heat in cold climates, it would be of greater value to have discussion of how to be successful with various floor plans. Many people chiming in are happy with the choice of a mini-split, so lets get more details of how they succeeded in keeping all the rooms comfortable. If they had a failure, cover that, too. Keeping more mistakes from happening will only increase the chances of designs adopting them.

  4. Steve Grinwis | | #8

    Guys guys! That's me! I made the highlight reel!

    Also, I changed my name to be my actual name...

  5. User avater
    Ryan Lewis - Zone 4A | | #9

    Radiant floor heat does have the advantage that it is run at lower water temperatures. This fact should make it stand out. If you are looking to use an air source heat pump, most of them can’t make water at high temps. The highest I have seen is the one from sanden and even that is 150 deg water temps.

    Also radiant off floors/walls/ceilings does not use up floor space. This is clearly an advantage over a large cast iron rad.

    1. User avater
      Dana Dorsett | | #10

      >"The highest I have seen is the one from sanden and even that is 150 deg water temps."

      Sanden's CO2 refrigerant heat pump's compressors are capable of making 180F water, but efficiency (and more importantly, capacity) takes a hit at those temps. But if you really REALLY need 180F, it can do it.

      1. User avater
        Ryan Lewis - Zone 4A | | #11

        Regardless it seems the BTU output is only really appropriate for a passive house or other airtight home.

        Also in terms of performance: for my area (westchester county by) the price of electricity is 3-4x that of natural gas. I wouldn’t be surprised if this ratio is similar for most of the country. Anyways, if the worst case cop isn’t at least 3ish, it’s not obviously dollar efficient to operate.

        1. Steve Grinwis | | #14

          This isn't always about lowest cost.

  6. Andy Bakke | | #12

    I’ve read enough GBA articles to know a proper manual J calculation needs to be made but how does one actually get this done? It seems that most residential HVAC contractors will either say it’s not necessary or lie to your face and make one up. Is this practice unheard of by your average local HVAC company?

    1. Steve Grinwis | | #13

      Find a third party rater, someone who does this for a living, and stakes their reputation on being right.

      I used Building Knowledge.

    2. User avater
      Jon R | | #15

      You actually need a Manual J, S, T & D.

    3. User avater GBA Editor
      Martin Holladay | | #16

      Here is a link to an article that directly addresses your question: "Who Can Perform My Load Calculations?"

      1. Andy Bakke | | #17

        Thank Martin, I must have missed that article but it was right on. I have now found a local RESNET HERS rater that should do the trick for me.

  7. Tom May | | #18

    Doesn't look like you investigated or are going to install any passive solar techniques for heating, cooling and light. How about solar hot water for that radiant zone and some free hot water? Wood stove with heat exchanger or hot water loop to assist. Dish half the PV you are going to purchase and invest in these free sources of heat and hot water. Install as much DC low voltage stuff as you can, then the money you save can buy you more PV in the future if needed.

    1. User avater GBA Editor
      Martin Holladay | | #19

      You seem to be under the misapprehension that a solar hot water system provides free hot water. It doesn't. It provides hot water that is much more expensive than hot water created by other types of equipment. Moreover, choosing equipment that operates on low-voltage DC power rather than AC power makes no sense for anyone except someone who lives off the grid.

      For more information, see "Solar Thermal Is Really, Really Dead."

      1. Tom May | | #20

        Well all my shw customers love their systems, especially those who've had them for 30 years or so and keep them operational for the benefit of free hot water as well those who I have added heating to the system. Not to mention the number of solar showers and pool heating systems I have installed over the years.
        As far as DC, low voltage lighting can be used by anyone, on or off grid, not to mention all the electronics that run off of DC converted from AC. Next time the power goes out, don't go to your car to charge your cell phone.

  8. user-7384703 | | #23

    Personally the best thing one can do is invest in the air tightness and insulation of the house. Shrink your demand, shrink your demand and shrink your demand one more time. Then if you have the money left over, by all means invest in ground source but as mentioned by others its becoming completely unnecessary with modern air source heat pumps that are available today. Something else that I see that breaks my mechanical inclined heart is building these big extravagant houses on a 10 cent millionaire budget. For some in site in NW Iowa (-12*/96* design temps) when I run houses through wright soft, my average cooling demand is 1500 sqft per ton of cooling on the main floor. My worst case scenario house, all open concept, stretched out across half a football field, 20' ceilings and windows to no end and ducts in the attic is 800 sqft per ton on the top floor and 976 on the main with 1259 in the basement. Theses are designed around R19 insulated walls, r38 in the attic (every seam of duct work sealed and wrapped in R8 and berried deep in the blown in) and semi tight construction. pretty basic and stamped out designs....more than enough room for improvement in construction and then there for the mechanicals...

Log in or create an account to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |