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

Heating a small house with a PTHP

Martin Holladay | Posted in PassivHaus on

In an upcoming article in Fine Homebuilding, I’ll be writing about heating options for small, well-insulated houses. One option we’re looking at: packaged terminal heat pumps (PTHPs). These are the through-the-wall units you see in hotel rooms. They’re inexpensive, somewhat inefficient, and occasionally noisy, but they represent a simple way to heat a small house.

Has anyone out there done it? I’d love to hear your stories.


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. jwyman | | #1

    My 620 square foot office in Western Massachusetts is heated by a Rinnai EnergySaver 1004F 38,000BTU direct vent wall furnace unit. It's not a PTHP, but a simple way to efficiently heat my space. Cost for the unit was $980 through a plumbing wholesaler and cost to operate is in the $400 yearly range. Construction of my office includes 8" roof SIP panels and cellulose filled wall trusses that average R-30. There is a fairly large window to wall area (for daylighting) and the Rinnai unit is quiet and comfortable.

    You can see the building here:

    and construction images here:

  2. user-659915 | | #2

    Interesting question. We've used them for years for single-volume spaces where the somewhat elevated noise levels are not an issue such as detached home offices and studios, including my wife's 450 s.f. painting studio. Perfect for intermittent occupancy with a low thermal mass enclosure which heats up or cools down quickly, and we have found it very economical to run though we don't have separate figures for it as it's metered with the house.

    I'd not be happy living with the compressor noise full-time though. I guess with a super-insulated house you wouldn't have to run it constantly, you could put it on a timer so that you could have a regular quiet period every day, or every night. Ductless minisplits are probably a better bet: more expensive but much quieter and, at least so far, vastly more efficient: 22 SEER and up for some of the Mitsubishi units which is way better than any of the through-the-wall units I've seen.


    I agree with James that they are too noisy to rely on for heating but are much better than a dehumidifier for dehumidification and cooling of an energy efficient envelope in a hot-humid envelope.

    The vertical, ducted versions can be placed in an insulated closet with return air supplied through a short section of 16" flex with reasonable sound control but are more commonly installed in a closet with a return grille cut through the wall or in the bottom of the door. This is what you see in more upscale hotels such as the Hotel Del Coronado in San Diego. They can achieve a heating COP of 3.1 but still be listed as 10.2 SEER.

  4. user-869687 | | #4

    I have come across a number of complaints about packaged terminal units installed in rental apartments, reading comments made on apartment rating websites. These units are relatively common in the NW because they cost 1/3 as much as ducted central air, a big motivator for multi-family developers. The typical complaints are that the units are noisy and not energy efficient, or simply ineffective, and that they don't distribute conditioned air into all the rooms. I would also point out that the units are installed in a hole in the wall, which interrupts the thermal enclosure.

  5. Kevin Dickson, MSME | | #5

    I'm planning on doing this. I only found one other guy who has done it, but he's happy:

    "Because the house is so compact and highly insulated, we eliminated the central heating system. The air source heat pump, which is like one found in a hotel room both heats and cools. Heat in the lower level is provided by small baseboards in the bedrooms and an electric heater with a fan in the bathroom.

    We tracked the performance of one unit from February 1st 2009 to February 1st 2010. The cost to heat the unit came to $185.24 which averages to $30.87/month or $1.03/day. The square footage of the unit is 1150. A separate meter which tracks electric used by the heaters showed 2795.5 kilowatts used during the six month heating season of mid-October to mid-April."

    Mini-splits will cost $2500 and up installed, but a PTHP is only about $700. That amount will take a long time to recoup with just a 30-50% better COP.

    Yes, it is a "hole in the wall", but so is a window. So, if installed properly, the added heat loss is about like a 26" x 14" window.

    A small, superinsulated house is a perfect application since the output of a PTHP is 12,000 btuh, and that should match the house pretty well.

    What about when the outside temperature is below 25F? PTHPs have backup resistance heat if necessary, but we ran a yearly simulation that showed that resistance heat was needed for only 4 hours for the entire year in a 1200sq.ft. house.

    A PTHP also has a "fresh air" option, that just might save you another $500-$3000 on an HRV

    More discussion at

  6. Kevin Dickson, MSME | | #6


    A good review of small, low energy home heating systems is included in the NREL report:

    After the dust settled, a better choice would have been a PTHP instead of a direct vent gas wall heater. The main reason for that is the "service fee" for natural gas, which is now $144/yr and going up.

  7. user-659915 | | #7

    Is there a case for a new class of heat/cool unit specifically formulated for highly insulated small homes? I'm imagining a split PTHP with the two components (air handler and compressor) installed back to back saddlebag-style on either side of an insulated wall with only the refrigerant loop passing between the two, with simplified coupling/commissioning. Compressor unit built to Mitsubishi standards. Quieter and more energy-efficient than a standard PTHP, cheaper and more easily installed than a mini-split.

  8. Kevin Dickson, MSME | | #8


    Great idea!
    A design that may be a little simpler than your proposal would have the insulated "wall" built into the unit. The $80 metal sleeve that is used for roughing in would also need a good thermal break.

    As long as we're redefining how hotel rooms will be conditioned in the future, they should also throw in a tiny air-to-air recouperator for the fresh air mode.

  9. John Donley | | #9

    I apologize if this thread is dead, but I thought you might still benefit from our experience. Two years ago, we installed a 18000 btu GE vertical zoneline in our poorly insulated, 1200 sq foot house. After some discussion with GE technical folks, I put the unit in an insulated closet dormer in the attic. I then ran flexible duct to four adjacent rooms (a bathroom, living room, kitchen, and sun room). The return duct is in the ceiling below the unit. All bedrooms are on the opposite side of the house. The unit does a fine job of keeping the entire house cool, even though it works pretty hard when the temps reach 100. But it struggles with heating, mostly because the ceiling is wood planks with nothing to keep the heated air from rising up through the joints into the attic. I have no doubt that the unit will be satisfactory year round once we add attic insulation.

    Incidentally, before we installed the GE, we tried an 18,000 btu Soleus mini split which worked about as hard. We were generally pleased with the performance, but not the reliability. Within one year, the fan motor on the indoor unit failed, the defrost feature on the outdoor unit quit working causing the fan on that unit to explode when it encountered ice buildup, and the compressor failed. The unit is still installed but not working (I can't bring myself to pay someone to drain the refrigerant after spending so much on repairs).

    I'm currently planning to install a PTHP in a well insulated shop. I only wish the horizontal versions were available in 18,000 btu (the max seems to be about 14,000).

  10. Eric | | #10

    We recently renovated and expanded a 1250 sq ft bungalow to 1800 sq ft, superinsulated the shell with 4 inches of rigid foam and replaced our 1950s oil furnace with a Mitsubishi Mr. Slim heat pump. With one outdoor compressor, we have three zones, 2 of them ductless and the 2nd floor as a ducted systems. Although we are still working out the kinks we are pretty pleased with the system although it clearly works hard when the temps approach zero, which is not all that common (we live near Boston). In the first year in the house, which is now all electric, we used about 9200 kWh and spent about $1500 on all energy combined. That's about half of what we spent previously. Although more expensive to install than the hotel-like units, I think we should be moving away from inefficient electric heat wherever possible. The Japanese mini-splits are definitely moving in that direction and seem to be getting more efficient all the time.

  11. Hoping for Heaven | | #11

    Please forgive me, if because the answers here are so valuable, that I spend time mentioning ideas (as a futurist ) than solving problems. But I'm also a
    materials scientist degreed in electronics. This is my first post.

    I'm curious and a bit confused about heat pumps because my experience, as most of you, is with room window air conditioners OR my refrigerator.
    Both of these cool the area, of interest, while a heat pump for heat is supposed to warm an area. My impression after shopping is that heating
    'conditioners', if you will, don't seem to work as efficiently as cooling air conditioners and their prices appear to be 2x more than a corresponding
    cold heat pump (window ac). If you've notice this same thing then I'd like to offer a scientific suggestion. As an experiments (1) buy a $100 freezer
    and route a way to move ground water (that you normally use anyway for the sink or toilet) through the freezer. My idea is to put the freezer in an
    area you want to heat by pumping heat from ground water and discharging the heat into the room. The water is moving and comes out the other side
    as cold as you choose depending on 1.capacity of the freezer 2. the amount of water flow 3. time in the freezer chamber.

    But for my 'money and time' this kind of experiment demonstration woudl teach me more about what I might get out of a heat pump than all of the
    reading and shopping that I've done looing at them.

    However. One heat pump I have not looked into which will become important to us, all, is the fresh air exchange heat recovery units. Because the
    amount of heat they recover is , I think, relatively small they'll be easier, for me, to see how they work and whether they work. I feel like I can
    put my hand in 'there' and see how cold the intake and how warm / cold the discharge is. Right now I don't feel like I can do that with any heat
    pump, heaters, that I've seen

    How about you? Did what I say have any merit to your experiences about the heat pump market?

    I appreciate having a forum where, I believe, people reading understand what I said. Thank you!.
    Andre in CA USA

  12. GBA Editor
    Martin Holladay | | #12

    It sounds like you want to build your own ground-source heat pump. I'm sure you can do it, although you won't get much heat output if you're using the compressor from a $100 freezer.

    Home-made appliances are rarely as dependable as the equipment available from large manufacturers. The device you are talking about -- a ground-source heat pump -- already exists. If you want one, you can buy one from Water Furnace or Econar.

  13. Hoping for Heaven | | #13

    I posted before I read your posts. A distinct fail I know. Sorry. And I realize the size of a 'freezer'
    is not 18,000 btu.. which by the way is 1000 lb of water heated or cooled 18 degrees F. per hour.
    This means if you cool 1000 / 7 gallons = 142 gallons of water / hour from 65F (my winter water
    temp) to 45 F you'll have more than 18,000 btu of energy transferred out of the water. (I hope the math is right) With inefficiency of the heat pump dumping heat into the room I'd guess that would help the efficiency.

    For some reason the heat pump industry is no where close to the window a/c industry. It could be
    due to the lack of the right refrigerant to pull heat from water at 60F or air at 30F. However we have
    , in our labs, refrigerators pumping down to -80F from 70F room temp. The heat from that would be
    the simple BTU calculation of difference of tempeature and the mass (i.e. 8lb water = 8lb of frozen tissues) changing temperature from 70F to -80 F . If it's 200lb and 150F delta that's a btu of 30,000 on the initial temperature change. IF it's water at 70F moving 200lb every hour then that's 30,000 btu per hour .. which it won't be because there's no way the 'water' will get to -80F .. but the reason I mention is to help to understand.. in my mind at least, how very simple it is to figure .. hey.. these heat pumps should be working better than they do.. and why aren't we using ground water to pull heat instead of air.. or why not glass box in my exchanger and in daytime heat the
    condenser coil (in a heating pump) with some solar heat. I see a/c evaporators on the s.west side of
    roofs and my heat pump has the condenser on the East side (bad for winter) and shaded by trees
    besides. The minimal winter sun will never touch it.. .. and why not? It's defective thinking the way I look at it. -- sorry to drone. I seem to have the last post in an old thread. I'll back off. : ) - Andre in CA . USA

  14. GBA Editor
    Martin Holladay | | #14

    I'm not sure why you wrote, "Why aren't we using ground water to pull heat instead of air?" In fact, heat pumps that use ground water as a heat source are extremely common; they are called ground-source heat pumps.

    Since you seem to be unfamiliar with the fact that these appliances exist, and since you seem interested in the concept, perhaps you would like to read about them:

    Ground-Source Heat Pumps

    Ground Source Heat Pumps, Part 1: The Basics

    GBA Encyclopedia: Heat Pumps

  15. user-622614 | | #15


    I have never seen a through-the-wall unit that was not so noisy that it would drive you bonkers (unless you were deaf already), and the seal between the unit and the sleeve seems to be impossible to seal, in spite of modern sealants and gaskets. The trend for more upscale hotels, "extended stay" facilities and condos is to use small split systems, where only the fan coil is in the space. My vote is for a mini-split or a small conventional split system with ductwork.

  16. Kevin Dickson, MSME, P.E. | | #16


    Have you published the article yet?

  17. GBA Editor
    Martin Holladay | | #17

    Not yet; look for it in an upcoming issue of Fine Homebuilding.

    My conclusions are similar to those expressed here:
    Heating a Tight, Well-Insulated House.

    I think that for most homes, a ductless minisplit system makes more sense than a PTHP, because of efficiency and noise issues.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |