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

Question about ductless minisplit heat pumps

ctpoponak | Posted in General Questions on


I live in a 1275 sq. ft. Cape Cod in southern New Jersey (zone 4A). My house was built in 1980, is two levels, is somewhat well insulated, has an unconditioned but insulated crawl space, has dual-pane Andersen windows (manufactured in the early 1980s), and it has asbestos shingle siding.

Prior to closing on the house this past January, I had a friend who is an HVAC professional assess the heat and cooling systems of the house. Heat is a 100,000 BTU propane-fired forced air system, while the air is a 1-ton Bryant central air system. My understanding is that a 100,000 BTU heater can heat a house three times the size of mine, but this heater was manufactured in 1996 and I have no idea what the efficiency rating is.

My friend the HVAC professional told me that the heater was on its last legs and he recommended a 5-head Mitsubishi ductless heat pump for $11K – of course I balked and I got a second opinion. Another friend who does plumbing and heating just replaced a control valve for a few hundred dollars and the propane heater is fine.

However, propane is expensive at $2.75 – $3.75 a gallon depending on current demand, and with the return ducts in my small house it often sounds like a 747 is landing in my living room. Plus, the heater itself is massive and it takes up over 1/3 of the crawlspace, so it’s a huge pain in the neck to service it and to replace filters, etc.

I really like the idea of the DMHP, but I really don’t think that I need five heads to the tune of $11K for a house my size as my friend recommended. ALL of the rooms in my house are less than 400 sq. ft. and most are less than 200 sq. ft. It’s my understanding that the smallest heads at 9000 BTU are too large for most of the rooms in my house.

Then I come to this site (GREAT FIND) and I read about homeowners who are heating homes of my size with a one-head system and some have a second head for cooling. I’m certain that my house isn’t as well insulated as homes using one head (envelope? noob here) but the house is well insulated in general. So I’m thinking that I could have the two-head system work in my house — now I’m talking $5K as opposed to $11K. Now I can get rid of my propane heater and the project would pay for itself in three or four years.

Could anyone out there offer any thoughts or suggestions?

Much appreciated-
Chris in NJ

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. GBA Editor
    Martin Holladay | | #1

    The short answer to your question about whether two heads will serve your needs is, "it depends." The most important factor in answering your question is homeowner expectations.

    This issue has been discussed at length many times on GBA. To learn more, I suggest you read the following articles:

    New Englanders Love Heat Pumps

    Are Seven Heads Better Than Three?

    Putting the Duct Back in Ductless

    Two Years With a Minisplit Heat Pump

    Report on Our Ductless Minisplit Heat Pump

    Just Two Minisplits Heat and Cool the Whole House

    Practical Design Advice for Zero-Net-Energy Homes

    Minisplit Heat Pumps and Zero-Net-Energy Homes

  2. mackstann | | #2

    A big limitation is your cold climate. You need a cold climate unit, and that generally means Mitsubishi Hyper Heat or Fujitsu RLS2H. This really limits the configuration options.

    The Mitsubishi Hyper Heat models usually used are single head only and go up to 1.5 ton. So if you want one for each floor, you need two outdoor units. Then there's the P-series multi-head system (so you could have just one outdoor unit with two or more indoor units), but they start at 2.5 ton and I'm guessing they're going to be closer to that 11K price.

    The Fujitsu models are only available in single head models up to 1.5 ton as far as I know.

  3. Expert Member
    Dana Dorsett | | #3

    Putting five ductless heads in a 1275' house is insane. Unless the zone heated/cooled by the head doesn't have a heat load of at least 5000 BTU/hr it will just cycle on/off most of the time and rarely modulate, and you won't hit it's rated efficiency.

    At typical 1300' circa 1980 2x4 framed house with R13 and clear-glass double-panes. R15 batts in cathedralized ceilngs (typcial for dormered-out capes) or R19s between ceiling joists, and a poured concrete crawlspace with no foundation insulation, with maybe some R19s between the floor joists will come in at a whole-house heat load of ~25-28,000 BTU/hr @ +15F (Trenton's 99% outside design temp: ) if you take some pains to air seal the house reasonably.

    Since you have a sealed & insulated the foundation & band joist, and maybe better than typical circa 1980 insulation you're likely in the ~19-22,00 BTU/hr range, which is within the heating output of 1.5 tons of pretty-good ductless The +15F output of a pretty-good single-head 1-ton is about 15,000BTU/hr. You may be looking at a pair of 3/4 tonners, or a 1-ton + 3/4ton. It's often about the same up front money but more efficient to go with two mini-splits rather than a 2-head multi-split.

    For a crude but effective heat load calc:

    Assume a 70F indoor temp, +15F outdoor temp, for a temperature difference of 55F.

    Assuming a framing factor of 25% (typical for 16' o.c. framing) you're looking at a wall U-factor of about 0.1 BTU per degree F per square foot.

    Assume all exterior doors are U0.5 unless they are insulated doors.

    Assume the clear-glass double-panes have a U-factor of 0.5.

    Assume ceiling a U-factor of 0.075 (if it's more than R19s it'll probably be lower.)

    Then on a room-by-room basis, measure the square feet of exterior wall, the square feet of window, and the square feet of upper floor ceiling/cathedral ceiling, and multiply out the U-factors by the 55F delta-T an example, take say a 12 x 13' upstairs corner bedroom with 9' ceilings that has a pair of doulbe-hungs for 20 square feet of window. Being a corner bedroom it has (12' + 13" =) 25' of running exterior wall, for a gross wall area of (25 x 9=) 225' of gross wall area, less 20' of window is 205'. The ceiling is (12 x 13 =) 156'. The heat losses are then:

    Windows: 20' x U0.5 x 55F= 550 BTU/hr

    Walls: 205' x U0.1 x 55F = 1128 BTU/hr

    Ceiling: 156' x U0.075 x 55F = 644 BTU/hr.

    Add it up and you're at 2322 BTU/hr.

    Give it a 15% fudge-factor for air infiltration and your at 2670BTU/hr. Subtract 250 BTU/hr for one sleeping human, and you're at 2420 BTU/hr.

    The smallest ductless heads have an output more than 3x that amount, so fuggedaboudit. You could find a mini-ducted cassette and split the output between two adjacent rooms, but it's easier to just put a ductless head somewhere near the top of the stairs in reasonable proximity to the adjacent rooms and set it to 74F, let it heat the rooms convectivly through open doors most of the time.

    At 2420 BTU/hr that room could be heated fully with one electric cove heater mounted above one of the windows. At 3.412 BTU/watt-hour it only takes (2420 / 3.412 =) 700 watts of radiant cove to get there. If you control it with both a ~$50 wall-type occupancy sensor switch rated for at least 800 watts of incandescent lighting (eg: Leviton ODS10 series ) AND a line voltage thermostat you could even let the room run cold with the door closed when nobody was using it. (Wiring in a bypass switch to the occupancy sensor for sleeping may be a good idea for sleeping quarters.) That way the ductless does the bulk of the heating, and the comfort stays high, since radiant cove heaters come on fast and heat the objects/humans in the room first, unlike baseboards that heat via air convection. That's maybe $250 worth of hardware for one room, with heat output appropriate for the room load, and it sips electricity compared to other resistance heaters due to the control scheme.

    For open floor plan lower floors add up the heat loads of all rooms that have reasonable convective paths for sizing the ductless. Most capes are pretty open except for the laundry rooms, etc, which can also get occupancy sensor limited cove heater treatment. If you have a propane water heater in there, it may not even need ANY- the standby loss of a water heater is more than 500 BTU/hr, and laundry room windows are small. For subtracting plug loads, figure the refrigerator a about 200BTU/hr, 300 BTU/hr if it's a large side-by-side model instead of an over/under version, a digital video recorder is similar, but be aggressive. For sizing the ductless, add maybe 50-75% of the load of doored off adjacent rooms with a common wall to the ductless head. When you're done with this exercise you'll have a pretty good idea of the floor-by-floor loads, and the sizing you'd be looking at for doing it with two ductless heads plus 2-4 cove heaters.

    In your location better-efficiency mini-splits with an HSPF of 11 or better will deliver a seasonal coefficient of performance of about 3.5 or better, meaning they deliver more than 3.5x the amount of heat per kwh of electricity used as the cove heaters. That being the case, it's actually cheaper to overheat the adjacent spaces slightly to minimize power use by the cove heaters. Most people tolerate 72-73F pretty well when it's 15-20F outside, even if a doored-off room is only 65-68F when they first enter it. :-)

    If your R values or construction are dramatically different tell me what it is and I'll estimate a reasonable U-factor if you want to dial it in any finer than that. When you've run the numbers on each room separately, post them here (by floor), so people can suggest actual model numbers for mini-splits & auxilliary heaters that are likely to work for you.

  4. ctpoponak | | #4

    Wow - thanks for all of the info and advice.

    @ Dana- I will post more particulars about my house later on when I get home from work. I really appreciate all of the detailed advice. I'm just concerned that my contractor won't buy it.

    Also, I saw a case study about the Tjernlund aireshare vents and thought that they'd be an ingenious solution for my application but again I don't think that my contractor will buy it.

    @ Nick- didnt think my climate (4A) was cold; my research says there no need for the hyper heat models.


  5. STEPHEN SHEEHY | | #5

    Chris: if your contractor has a detailed response to Dana, you should certainly listen. But if he just disagrees, or says he doesn't do things that way, maybe you need to remind him who will be paying the heating bills. Dana' s analysis appears to be based on considerably more than just a hunch.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |