GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter X 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

Minisplit design/Manual J question

Bdgray | Posted in Mechanicals on

Hello all,

I am working with an HVAC contractor to design and install a multisplit system for our 1950s ranch renovation. As an FYI, we’ve done a substantial amount of upgrades to improve the tightness and insulation of the home although at 60 years old the house still isn’t perfect. I’m now hoping to replace our aging ducted air handler with a new ductless system. Motivation to go ductless is driven by many factors that I won’t get into in this post (both efficiency and a more practical issue of headroom in our finished basement).

Our contractor has run a detailed Manual J and has provided me the report using a program written by Wrightsoft. I’m happy to attach the report if this would help. The open question is how to handle three guest bedrooms that are all connected via a single hallway. The rooms are occupied (our three kids) so I do want a solution that prioritizes comfort (i.e. I don’t want to place a single unit in the hallway and hope the rooms are comfortable when the doors are shut).

The contractor has presented me with two options (all Mitsubishi):
1) installation of small 6K wall units in each room (MSZ-GE06NA-8)
2) installation of a single 12K ducted unit for all bedrooms (SEZ-KD12NA4R1.TH)

Below are the calculated loads for heating and cooling in each room and in total:
ROOM NAME; Area (sqft); Htg load (Btuh); Clg load (Btuh)
Bedroom 1; 225; 3372; 2392
Bedroom 2; 192; 3127; 2127
Bedroom 3; 156; 1586; 1551
Bedrooms Total; 573; 8085; 6070

I’m hoping for advice/thoughts on which approach is recommended when balancing performance, efficiency and comfort. It seems like either way I will loose efficiency – either oversizing the units or using ductwork.

As a final question, given I’m in Chicago I also wonder whether there would be any merit in going with a hyperheat model in each BR. While I like the ability to heat to a lower temperature I think these units would create an even larger oversizing issue (I believe these start at 9K BTU).

Lastly, I do plan on adding supplemental heat in any scenario. I’ve recently installed high-efficiency fireplace inserts and I plan to also add some electric radiant to have a set and forget option. To be clear though, I really want to find a multisplit design that heats our home (comfortably and efficiently) for 90%+ of the winter. If that’s not realistic then I’ll grudgingly keep a forced air system.

Thank you,

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. Expert Member
    Dana Dorsett | | #1

    The loads of each of those rooms are small enough to would a 6K head per room ridiculously oversized. If the rooms are all reasonably close to one another so that the duct runs would be relatively short a mini-duct solution would be more efficient and more comfortable. A 3/4 ton HyperHeat in each room would be even more ridiculous- they'd be cycling on/off even when it's pretty cold out!

    But the SEZ-KD12NA4R1 may be oversized for your 90% temperature bin load. The SEZ-KD09NA4 might be a better fit:

    But lets check that: Take a look at the output tables on p.25 of this document (

    The KD12 has a capacity of 8400 BTU/hr @+15F, and 6500 BTU/hr @ +5F (in a 70F room, a bit more in a 65F room.)

    The KD09 has a capacity of 6700 BTU/hr @ +15F, and 5200 BTU/hr @ +5F. (in a 70F room, a bit more in a 65F room.)

    Your heat load was probably calculated at an outside design temp of 0F (your approximate 99th percentile bin: ) and in inside design temp of 70F, or a total delta of 70F. Your 90th percentile temperature bin in Chicago is about +10F, so you have to do a bit of interpolation on the table numbers, but since 10F is only a 60F delta from the indoor design temp, the heat load at 10F is going to be about 60/70 of that 8085 BTU/hr number, which is 6990 BTU/hr.

    The output of the KD12 @ +10F will be something like 7500 BTU/hr- that definitely covers it with margin

    The output of the KD09 @ +10F will be something like 6000 BTU/hr, which is a bit shy of 6990 BTU/hr

    Mind you, if you backstop the room temp at 65F with the auxilliary heating the -KD09 will be delivering about 6200 BTU/hr, and instead of a heat load of 60/70 x 8085, with the temperature difference shrinking to 55F from 60F the heat load with a 65F room temp is 55/70 x 8085= 6350 BTU/hr.

    That's only a 250 BTU/hr shortfall, which is a close enough match- a sleeping adult human puts out about 250 BTU/hr, after all. If you set & forget the KDxx for 70F, the thermal mass of everything in the room will carry you for hours before it hits 65F at an outdoor temp of +10F.

    The smaller unit will modulate more during rest of the time too, leading to stabler room temps. The KD12 can deliver about 12,000 BTU/hr @ 35F (per the table on p.25), but the minimum modulated output of these things is only about 1/3 of the max capacity, so you're looking at min-output of about 4K @ 35F. At 35F your heat load is half what it is at 0F, so that's your whole heat load, and above 35F it'll be cycling. The KD09 has a capacity of 9600 BTU/hr @ 35F, so the min-mod output will be about 3200 BTU/hr- it'll still modulate at outdoor temps up into 40s.

    The rated COP of the KD09 unit at 47F is 4.2, compared to 3.6 for the KD12, which is a pretty big difference. But at +17F the KD12 is marginally more efficient than the KD09, at 2.86 compared to 2.76. Since your mean winter temp is above 25F, the KD09 will probably deliver slightly higher annual efficiency.

    SOooo....'s really your choice- take the KD12 for more low-temp margin with a 70F indoor design temp, or sleep at 65F on the colder nights and go with the KD09. Since you have auxilliary heating to cover the capacity issues at much lower outdoor temps, it's going to be just fine to go with the smaller more efficient unit. Either would cover you at least 90% of the time if you're comfortable sleeping at 65F.

  2. Bdgray | | #2


    Thank you very much for the well thought out response. I have to admit that I needed read this several times to make sense of it all but I think I understand the gist. Considering I plan to backstop the multisplit, the ducted KD09 seems to make sense. Also, the Manual J was designed at -5F which I think only strengthens your argument. If you don't mind I have a few follow-up questions.

    1) We have a large open floor plan kitchen & living room (~1225 sqft). Total heating load at -5F is ~20K BTU. The contractor has split the load between two 12K units on different sides of the room. Cost aside, is there logic in splitting the load in a large open space between two units versus one 20K unit in the center?
    2) Not sure whether your knowledge extends to the electric heating appliances that will backstop the multisplit. I assume the units sold at the big box stores are garbage (although this is pure conjecture). Are there recommended suppliers in this space who manufacture high efficiency units?

    Thank you again,

  3. GBA Editor
    Martin Holladay | | #3

    Q. "Concerning the electric heating appliances that will backstop the multisplit: I assume the units sold at the big box stores are garbage ... Are there recommended suppliers in this space who manufacture high efficiency units?"

    A. All electric-resistance heaters have exactly the same efficiency (100%). Any manufacturer of electric-resistance heaters who claims that their units are "high efficiency" compared to other electric-resistance heaters is lying.

    The only way to get more than 100% efficiency out of an electric heater is to use a heat pump. If the heater has a red, glowing element, it's a resistance heater.

  4. Expert Member
    Dana Dorsett | | #4

    A design temp of -5F is below the 99% temperature bin for Chicago-proper, and probably below the 99% bin for any location in Cook County IL. It's pretty standard (but lousy) practice to tweak the inputs on Manual-J tools to give a bit of margin "just in case" (the contractor doesn't want the call-back from an irate & cold customer at 5AM on the coldest day of the year), but with resistance heating backstopping the temperature extremes beyond the 99% point designing right up to the line has zero comfort risk.

    There is no point to using multiple heads in a large open space- convection is powerful, and if the different areas are divided only by large open archways rather than 25 square foot door openings the end-to end temperature delta won't be more than a couple of degrees.

    While the efficiency of all resistance heating is the same, the comfort levels are not. (The reliability of resistance heaters is also pretty good, but the fit & finish varies all over the place.) Finned-element baseboard convectors raise the air temperature, and the air keeps the humans warm, whereas radiators heat the objects (including the humans) directly. Small electric panel radiators or radiant cove heaters on the walls with windows will average out the average radiation temperature in the room by countering the low radiation temperature coming from the direction of the window. Radiant floors are very cushy under bare foot, but quite a bit more expensive than panel radiators. Radiant ceilings may have a bigger comfort effect while you're in bed, since the coldest hours when the mini-split might not be keeping up typically occur in the pre-dawn period when most people are sleeping. Radiant ceilings are also not cheap, but is a bit faster acting, heating the room a bit faster than radiant floors, since it isn't being blocked by furniture or rugs.

    When designing resistance heating as supplemental to mini-splits, you have do decide whether to treat it as auxilliary heating (making up the difference between what the mini-split can't deliver) or as back-up (has enough output to carry the full load.) With a heat load calculation it's dead-easy to size it for the whole load- simply divide the BTU/hr figures by 3.412 BTU/watt to come up with watts, eg: Your lossiest bedroom has a heat load of 3372 BTU/hr, which is 3372/3.412= 988 watts, so a 1000 watt resistance heater would cover it even at -5F. But like any other heating system, if it's only making up the difference between an estimate/WAG of 1500 BTU/hr of mini-duct output @ -5F and the full load of 3372 BTU/hr , the resistance heat requirement is about half that. The 500 watt auxilliary heater will cycle a lot less, and be more comfortable than 1000 watts, but it's not insane to size it for the full load.

    You also have to consider how much it's worth spending for something that will only be operating less than 10% of the time, primarily during times when you're sleeping. It may well be that pretty-good oil-filled baseboard radiators or even finned baseboards are the "right" way to go, rather than springing for radiant floor/ceiling or panel radiators.

    The ~2' square ~400-500 watt wall panel radiators that can fit under most windows can be under $100/pop if you shop around, and would be a reasonable way to go too. Most are designed with on-board controls- you'll want to hard-wire it and use a line-voltage wall thermostat rather than a off-low-medium-high switch (though keeping the switch allows you to use it at low power for aux heat even if you size it for full backup.)

  5. jackofalltrades777 | | #5

    Your contractor left out option #3:
    3. A ductless multi-zone unit (3) that will run off of one single outdoor compressor unit.

    They are more efficient than a ducted mini-split and when each zone calls for additional heat/cool the wall or ceiling unit ramps up just for that room. Mitsubishi makes multi-zone units.

    Don't forget, most codes require EACH BEDROOM to have a some type of heat source. Having one mini cover an open kitchen or dining room is OK but having one mini set in a hallway doing 3 rooms via natural airflow through the hallway will NOT pass code in most places.

  6. Expert Member
    Dana Dorsett | | #6

    Petee=r: His option #1 IS the 3-head multi-split option, making #3 redundant.

    The smallest ductless heads for multi-splits are about 6000 BTU/hr, and ridiculously oversized for the calculated loads. That would be a less efficient than the mini-duct cassette option, since both the heads AND the compressor would be even more oversized for the load. (The smallest M-series 3-zone would be a an 2-ton compressor MXZ-3B24NA-1, instead of the 3/4 tonE-KA09NA.TH or 1 ton SE-KA12NA.TH compressors designed for single mini-ducts.) A 24,000 compressor for an 8,000 BTU/hr design heating load is never going to work efficiently.

    Codes do not require a heat source for every room. They only require that the heat source in the house be capable of keeping all rooms above 68F at the 99% outside design temperature.

    A high-R low-load house can do that without a heat emitter in each room, but that's not really an option with 2000 BTU/hr design heat loads on those rooms.

    The mini-duct cassette approach with output split three ways works here, provided the duct runs can be made ulltra-short.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |