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
Musings of an Energy Nerd

How To Buy a Ductless Minisplit

This cheat sheet is intended for designers and builders who are confused by available equipment options

Mitsubishi and Fujitsu make air-source heat pumps that can heat a home in a cold climate, even when outdoor temperatures drop to -15°F or -20°F.
Image Credit: Alex Wilson

Green builders usually specify high-performance windows and above-code levels of insulation, while striving to reduce air leaks in their homes. As a result of these efforts, most green homes have relatively low heating and cooling loads.

Increasingly, these low-load buildings are being heated and cooled by ductless minisplits or ducted minisplits. Many of these air-source heat pumps have ratings in the 9,000 to 15,000 Btu/h range — an appropriate range for low-load houses. Because they are fueled by electricity, these systems are a good match for a home equipped with a roof-mounted photovoltaic system.

Of course, there are still plenty of builders who have their doubts about minisplits. Some wonder whether it’s really possible to heat and cool a house with just one or two ductless minisplits; others wonder whether every bedroom needs a separate heater or forced-air register. (GBA has published a quite a few articles on those topics; for links to these articles, see the “Related Articles” sidebar below.)

In this article, I won’t be addressing questions about cold bedrooms or providing advice on bedroom door operation. Instead, I’ll assume that readers know how many ductless or ducted units they want to install — but just want some guidance on equipment selection.

To keep things simple, this article will focus on just two manufacturers: Mitsubishi and Fujitsu. These two brands have captured a strong percentage of the U.S. market for minisplits, and both companies manufacture equipment that works well in cold climates.

Start with a load calculation

When designing a heating and cooling system, the first step is always to perform a heating and cooling load calculation — ideally one using the Manual J method.

Be careful: many HVAC contractors don’t know how to perform these calculations. The vast majority of HVAC contractors use rules of thumb that…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial


  1. STEPHEN SHEEHY | | #1

    minisplit cost
    Great article by Martin. It may be useful for readers to get an idea of what these things cost.
    I just had two Fujitsu units, 9RLS3H and 12RLS3H, installed in my new house in Maine. Total cost was about $6500.

  2. skidmorebay | | #2

    Mini-split noise
    I haven't yet found any real-world information on the noise produced by a mini-split heater. I'm actually concerned about the outdoor noise level in my quiet wooded neighborhood. Could anyone give me an idea of how loud the condenser is when running, and how loud the indoor unit is?

  3. GBA Editor
    Martin Holladay | | #3

    Response to Justin Smith
    Most homeowners report that Mitsubishi and Fujitsu's outdoor units are quieter than split-system air conditioners from U.S. manufacturers.

    Indoor units are generally quieter than a refrigerator. Again, most homeowners report that they are surprisingly quiet -- in many cases, quieter that the whooshing noise coming from the registers of a typical forced-air system.

    Forced-air systems are often oversized, so they suffer from on-again-off-again syndrome -- a type of operation that sounds noisy. Inverter-driven minisplit systems, when properly sized, tend to run continuously, and are therefore less obtrusive.

  4. GBA Editor
    Allison A. Bailes III, PhD | | #4

    It's not always a condenser
    Martin, you wrote:

    "The outdoor unit of a minisplit system is also known as the condenser. Because the condensing unit includes a compressor, some people call the outdoor unit the compressor — but 'condenser' is more accurate. (The condenser includes a compressor, a condensing coil, and a fan, among other components.)"

    If you were talking about air conditioners only, you'd be correct. With heat pumps, however, the outdoor coil is the evaporator coil when it's in heating mode, the condensing coil when it's in cooling mode. It's most accurate just to call it the outdoor unit.

  5. STEPHEN SHEEHY | | #5

    Reply to Justin
    The minisplits are very quiet. At the low fan setting, the indoor unit is inaudible. Even on high, it's pretty quiet. Both Mitsubishi and Fujitsu provide sound levels for all their indoor and outdoor units with their specifications.

    I've only had mine for several weeks, but I've never heard the outdoor unit running, even though it isn't far from the entry door.

  6. GBA Editor
    Martin Holladay | | #6

    Response to Allison Bailes
    Good point! Thanks for the correction.

  7. Expert Member
    Dana Dorsett | | #7

    What he said!
    It's only a condensing coil when in air conditioning mode. Since as a rule the unit also contains the compressor, I'll stick with calling it the "compressor" &/or "outdoor unit". never "condenser" if it's a heat pump.

    Costs and installer competence are all over the place- it's important to shop around, and it's important to look closely at the equipment specs.

    I recently talked/walked a co-worker through the process of getting mini-splits installed in his suburban Boston 1920s antique, starting with room by room heat load calcs. The upstairs rooms (2 bedrooms, and office and a bath, about 800-850' total) had a total heat load just shy of 8KBTU/hr @ +10F, the approximate 99% temperature bin for Arlington, MA. The first floor had a total heat load of about 15K. Since the first floor kitchen was doored off from the rest of the rooms on the first floor (which had open archways between them) he wanted to do the first floor with a mini-duct cassette and a few ducts in the basement, but couldn't find a bidder willing to go there.

    A highly recommended Fujitsu installer (who said "I don't do ducts- period"), suggested two -9RLS3s, one upstairs one for the kitchen, and a 12RLS3 for the rest of the space. The min-mod output of the 9RLS3 @ +47F is 3100 BTU/hr which is above the +47F load for both the upstairs zone and WAY above the +47F load of the kitchen zone. I don't remember the exact quote, but it was in the $10K range. It would work, but less than ideal, since two of the three would be cycling rather than modulating most of the heating season.

    He had a handful of Mitsubishi installers in to look at it. The sales-droid from one of the larger installers opened the conversation with "So, how much do you want to spend?" (never a good sign...), to which my co-worker responded "I want what makes the most sense." The first suggestion the guy came up with was to install a ductless head in every room. The co-worker hauled out the heat load calc and pointed out that even the smallest heads were way oversized for the room loads. They went back & forth a bit discussing various options, and the salesman went away, and submitted a bid the next day for three GE12NA heads and a single 2.5 ton multi-split compressor. (Sigh...)

    The min-mod output of the compressor @ 47F was a bit over 6K, and while not ridiculously above the whole house load @ 47F, the minimum output of the 1 ton heads for the kitchen & upstairs was well above the loads. Not a great fit. The quoted price was on the order of $16-17K too (seemed a bit steep) and the HSPF of that equipment combination was about 9.8, not exactly rocking the efficiency world. I pointed out in conversation that the house a pair of FH09s had sufficient capacity handle his first floor load and a third could handle the upstairs, at an HSPF of 13+. He called the contractor and asked him do bid that, which came back less than an hour later at $15K (REALLY?).

    A second Mitsubishi installer (from a 5 person company, 6 if you count the dog) suggested that a 1-ton would be more appropriate for the larger zone on the first floor (true), and that the higher cfm would be useful if the upstairs unit for heating the rooms via open doors (probably), and probably better sized from a cooling point of view (maybe- no cooling load calcs were done.) That bid came in at $12.6K for a pair of FH12s and an FH09.

    Another Mitsubishi installer submitted a proposal for a pair of FH09s and an FH12 for similar money.

    Yet another Mitsubishi installer (one of the largest in the region) quoted ~$21K for a pair of FH12s and one FH09(!). Don't THINK so! They can't be very hungry- this is a dead-easy installation to pull off, nothing complicated about it.

    The $12.6K for two FH12s and one FH09 was the winning bidder. It's substantially more expensive than the Fujitsu installer, but more appropriate modulating ranges for the zones they serve, and there is quite a bit more Mitsubishi support than Fujitsu support in this area. They'll be installing them next week. With the locally available efficiency rebate incentives for the higher efficiency units the net cost will be about $11k.

    Even the winning quote was a bit high compared to quotes I had been seeing over the past few years, (I've seen quotes for three 1.5 ton units on a single building for that kind of money) but not as insanely high as the outliers. The scheduling time between signed contract and install-dates were also a bit on the longer side, which may be an indication that the market is a bit super-heated locally.

    It's pretty clear that some (most?) of the people selling this stuff have absolutely no clue about minimum output issues (with either multi- or single zone systems), the HSPF efficiency, or the specified heating capacity at a given temp, and none of them offered to do load calculations. It's "Caveat emptor". Without having done the load calculations on his own and looking online for the efficiency test submittal pages for the proposed equipment my co-worker would not have had the means to assess which proposals made sense, and which were completely out to lunch at any price.

  8. STEPHEN SHEEHY | | #8

    what Dana said: "no clue."
    A few minutes ago . I received an e-mail from my insurance agent. The carrier recently send someone out to inspect the (almost) finished house. The only issue is the lack of central heat, because the insurer doesn't consider heat pumps effective at low temperatures. In 23 years here, we've seen -20F twice. Typical coldest winter temps are -15 for one or two nights each winter. 99% design temp is around zero F. I guess I'll just install a 100,000 btuh coal-fired boiler and they'll be happy.

  9. GBA Editor
    Martin Holladay | | #9

    Response to Dana and Stephen
    Dana and Stephen,
    Of all the subcontractors working in residential construction, none are more frustrating, or more likely to be ill-informed, than HVAC contractors. This has been a problem for at least 30 or 40 years. (Needless to say, there are some excellent, well-informed HVAC contractors out there; it just seems that they are in the minority.)

    If Mitsubishi and Fujitsu want to take the residential market seriously, they need to come up with some method of solving the ill-informed HVAC contractor problem. I don't have a magic wand, but the problem is real, and not getting any better.

  10. Expert Member
    Dana Dorsett | | #10

    It's surprising... or maybe not.
    Code inspectors will often insist on "backup" for mini-splits, even for low-load houses, even when the extended temp capacity tables clearly indicate that the mini-split would cover it.

    I pointed out to one building owner who was getting some push-back about backup heat that a couple 1500 watt plug-in space heaters could handle the zone loads at the 99% outside design temp. Thankfully nobody insisted on hard-wired resistance heating backup complete with wall thermostats on that one.

    Codes usually only demand that all rooms be able to be heated to 68F / 20C at the 99% outside design temp, not the all time record low, or the 99.7th percentile temperature bin (which is still often used as a design temp.)

    Martin: It's not just the mini-split vendors, it's literally the whole industry. I doubt one out of 10 have ever done a load calculation that was any better than garbage-in = garbage-out. But of course most home owners don't even know what it is, let alone how to ask for a load calc, or properly assess a calculation if one is submitted. Contractors can install 3x oversized equipment but if they are prompt & friendly, and clean up after themselves they get top reviews on AngiesList etc. . It takes a lot to wade through the wealth of ignorance on subjects like this, but it's hard to blame the manufacturers.

    California Title 24 prescribes hard oversizing limits and requires a Manual-J. The CA inspectors there DO look at both the Manual-J and the equipment specs. There is still probably some thumb-weight in some of the contractor-submitted load calculations, but it's hard to oversize by 3x (or even 2x) and get it by the inspector. It's a start.

  11. kyeser | | #11

    Don't overlook the fact these units will make noise, especially in heating mode. I have a single head low temp fujitsu, can't remember the model. The condenser is about 5' from my bedroom window and it can really start to hum during the winter. I am going to move it about 20' down the house so it isn't quite as annoying.

    Very satisfied with the performance, but do not put the confessor any where near a bedroom where late at night during the perfectly quiet winter, you will hear all the cycle changes and loud humming that can accompany these machines.

  12. Expert Member
    Dana Dorsett | | #12

    BTW: There's a new big brother to the FH series Mitsubishis
    Seems the FH18NA has been added to the FH09 / FH12 / FH15 lineup:

  13. Tim C | | #13

    Garbage in and out indeed
    One contractor I had out for a GSHP estimate wouldn't believe I had better than '90s code minimum insulation and refused to even give me a quote for anything less than a full ton over the right size. I was almost proud of another who wanted to go off of my propane usage rather than guess at insulation... until he came back with a heat loss 50% higher than my own calculation from the same numbers (which turned out to be within 5% of later observational data).

  14. Svig | | #14

    Building in a small town in zone 7
    The only HVAC contractor sells Mitsubishi, so I am pleased with that. I won't be moving there until August and I am glad I will be armed with the useful knowledge and experiences found here at GBA. I have done a preliminary Manual J on the home I plan to build, and I will do another on the home I actually build. So I am looking forward to see how the load calc/sizing/bid process goes; hoping for a good experience, but ready that it may be something less.

  15. GBA Editor
    Martin Holladay | | #15

    Response to Dana Dorsett (Comment #12)
    Thanks again for your sharp eye. I have added information about the FH18NA to the article. Your input is always appreciated.

  16. christopherw | | #16

    Are the individual heads on a multi-split still variable?
    Looking through the specs on Mitsubishi's site, I notice the minimum for a multi-split is the same as the capacity for the smallest single head. That in itself isn't surprising. However, when I dig in and look at the specs on linkdrive for the MSZ-FH09NA, it shows two submittals, one for pairing with a mini split system, and another for configuring as part of a multi-split system.

    The mini-split configuration shows a rated capacity of 9,000 Btu/h and a capacity range of 1,700 - 12,000 Btu/h.

    The multi-split configuration only shows a single cooling capacity of 9,000 Btu/h.

    Looking across all the internal units, I see the same thing. The multi-split configuration shows a single capacity rather than a range. It seems to imply that, in a multi-split configuration, the internal units are not actually variable, but simply modulate on or off. Rather than carefully regulating the flow to the individual heads based upon efficiency, need, and humidity requirements, it the system simply turns them on or off.

    This would mean that a multi-split system is not at all a VRF system in the same sense that the mini-splits are. It would be more like a series of single-speed systems, with some rather significant implications for the potential comfort and efficiency of such a system.

    This seems so shocking that I'm probably missing something here.

  17. GBA Editor
    Martin Holladay | | #17

    Response to Christopher Welles
    I don't have an answer to your question, but I hope to dig deeper next week, with inquiries to Mitsubishi, unless a knowledgeable reader posts the answer to your question before then.

    Here's what I know:

    1. The compressor of a multi-split system definitely has a variable capacity. According to information on a Mitsubishi web page, "Inverter compressors ramp up quickly, providing the energy necessary to achieve the cooling or heating demand of the zone. Then, working in tandem with system controls and sensors, the inverter compressor varies its speed to maintain the desired comfort level. Thus, the system performs at only the minimum energy levels necessary and does not waste electricity when partial-load conditions are present, which is 97% of the time in most locations."

    2. A single-zone minisplit system (one that pairs a single indoor head with a single outdoor unit) is definitely more energy-efficient than a multi-split system, due in part to the fact that the multi-split compressor has to provide enough capacity to handle the total load, including the load of the hungriest zone. This approach is inevitably less efficient that using a separate outdoor unit for each zone, because the outdoor unit's compressor is working harder on a multi-split system than it might be if there were two compressors for two zones.

    3. I'm not sure whether the indoor units on a multi-split system have ramping capabilities. But whether they do or don't, multi-split systems are less efficient than single-zone minisplit systems.

  18. user-593033 | | #18

    Multi head or multiple head
    So say you have a single room load of 24000 BTU's. Would it be more efficient to use a single head unit with that capacity or to use two smaller individual units (each with there own outdoor unit)? It seems that you could use only one unit in part load seasons and run the second unit only when the seasonal loads require more BTU's.

  19. GBA Editor
    Martin Holladay | | #19

    Response to Chaz Steffen
    1. That must be a very large room, or a very poorly insulated room, if it has a design heat loss of 24,000 Btu/h. That's a lot.

    2. The most efficient way to go is always a single-zone approach with a ductless indoor head paired with its own outdoor unit. Neither Mitsubishi nor Fujitsu make indoor heads with such a high capacity (24,000 Btu/h). So, if you are sure that your heat load was calculated correctly, you should choose two units, each rated for 12,000 Btu/h.

  20. user-593033 | | #20

    big load
    The room is a laundry with a couple of commercial dryers. The load is only when the dryers are running. It is about 1000 square feet of floor space. The walls are From the outside in: double layer of plank,one vertical and one horizontal, about 3" thick, Plaster, 2.5" gap then 2x4 framing all filled with dense pack cellulose, 1/2" drywall. The floor is uninsulated over an insulated basement. The ceiling is a ten inch joist with dense pack cellulose with an apartment above. The only windows are on the North side. This is in zone 5. So you can see that the biggest load is the cooling load with the dryers running. I usually deal with residential so having larger equipment loads is new to me. Thanks for your input.

  21. DarkNova | | #21

    Fujitsu Heat Output
    You show the heat output for the Fujitsu AOU9RLS3H as:

    At -15°F, it has a heat output of 11,100 Btu/h.
    At -5°F, it has a heat output of 14,000 Btu/h.
    At 5°F, it has a heat output of 15,400 Btu/h.

    I'm wondering where you got this information? I'm looking for the heat output for AOU12RLS3H and AOU15RLS3H.

  22. GBA Editor
    Martin Holladay | | #22

    Response to Nick Hall
    I have so many documents on my computer, I should be able to find it. But I haven't found it yet.

    Nor have I found it with a web search, which is evidence of how bad the Fujitsu web site is, and how apparently reluctant minisplit manufacturers are to share heating capacity specifications.

    Fujitsu has published this information -- I just can't find it right now. I'll post a link to the document when I find it.

  23. DarkNova | | #23

    Fujitsu Document

    I searched around myself after I posted and think I found the document: looks like where Fujitsu hides the technical data, just like Mitsubishi does. Now I just need to find the COP data. Those ranges given at 5F are pretty wide and I'm wondering how the COP does at other temperatures.

  24. DarkNova | | #24

    Fujitsu Heating Capacity
    So the heating capacities from this document are:

    ASU9RLS3/AOU9RLS3H: 15,400 BTU/hr at 5F, 14,000 BTU/hr at -5F, and 11,100 BTU/hr at -15F.
    ASU12RLS3/AOU12RLS3H: 16,600 BTU/hr at 5F, 15,000 BTU/hr at -5F, and 11,800 BTU/hr at -15F.
    ASU15RLS3/AOU15RLS3H: 21,000 BTU/hr at 5F, 18,600 BTU/hr at -5F, and 16,300 BTU/hr at -15F.

  25. GBA Editor
    Martin Holladay | | #25

    Minisplit COPs
    Thanks very much for those links.

    I'm posting an Excel spreadsheet (it appears at the bottom of this comment) -- one I modified from the Excel document posted at

    The one posted on Google docs is hard to read. Click on the link below to see my modified spreadsheet. (I eliminated some of the columns to focus on what's important.)


  26. DarkNova | | #26

    Minisplit COPs
    Thanks Martin, that's very helpful.

  27. DavidAlex | | #27

    HVAC contractors
    This thread makes me want to go get an HVAC degree (certificate?) and go into business. I'd probably use my physics degree more often in HVAC than I do in my current job.

  28. GBA Editor
    Martin Holladay | | #28

    Response to David Hicks
    Do it! We need more HVAC contractors with a background in physics.

  29. hallie17 | | #29

    location of interior unit
    In zone 5, would you locate the unit near the ceiling or closer to the floor? It seems most intallations expect locating high on the wall, but I read one artcile that suggested putting it 18" above the floor in cold climates. We do use some air conditioning here, but the main need is heating.

  30. GBA Editor
    Martin Holladay | | #30

    Response to Hallie Bowie
    The wall-mounted units are designed to be installed high up on a wall; their fans and vanes are engineered to do a good job distributing warm air or cool air throughout the room.

    If you want a floor-mounted unit, most minisplit manufacturers sell specially designed indoor units for that location. (See photo below.)


  31. Adam Emter - Zone 7a | | #31

    Hi all,
    I am building a new

    Hi all,

    I am building a new home in North Dakota, zone 7a. Main heat source will be radiant hydronic heat, but I am considering a ductless split for shoulder-season heating and also for summer cooling. We get very cold here, but on those days the radiant would be the main heat source, so I don't see the low-efficiency issue with air-source heat pump to be an issue in this case. I was initially just planning to use two separate window air-conditioners with through-the-wall sleeves, but these mini-splits are much more efficient and would also give me heating capabilities. My questions: would this be a smart use of a mini-split only being a backup heater and summer AC? Can any of these Mitsubishi or Fujitsu units be purchased and installed by myself or do I need to go through an HVAC contractor?

  32. GBA Editor
    Martin Holladay | | #32

    Response to Adam Emter
    Q. "Would this be a smart use of a mini-split only being a backup heater and summer AC?"

    A. I'm not sure what you mean by "smart." These units are excellent air conditioners, but they cost more than window-mounted units. I'm always a little leery of the idea of installing two heating systems when one will do, but if you can afford it, why not? You may even find that you end up using the ductless minisplits for most of the winter.

    Q. "Can any of these Mitsubishi or Fujitsu units be purchased and installed by myself or do I need to go through an HVAC contractor?"

    A. That depends. Do you have a vacuum pump, and are you good with a flaring tool? For more information on DIY minisplit installations, see Dare to DIY a minisplit install?

  33. user-5356986 | | #33

    ducted systems
    I was a little confused by the following paragraph/section in the article. Near the end was this section:

    "On ducted systems. “If you’re going with a ducted minisplit in U.S. climate zone 6, the only realistic option is the Fujitsu xxRLFCD series. They at least have a rated output down to -5°F. The Mitsubishi SUZ/SEZ units have rated output down to 5°F, but the output and efficiency is comparatively poorer. But in a U.S. climate zone 4 or warmer they’d do just fine, and there is a lot more distributor support for Mitsubishi than Fujitsu in many markets.”

    However, this seems to contradict or at least ignore the earlier discussion in the article of the Mitsubishi MXZ- xxxNAHZ product lines which have full hyperheat temp range. They are listed in the NEEP list as multizone ductless units, but this article said that they could also be used to support multizone ducted delivery systems if the right inside units were purchased. Doesn't that mean they are a good alternative to the Fujitsu? What am i missing?

  34. user-5356986 | | #34

    minimum ouput penalties and multizone ductless
    I'm not sure how active this comment list is anymore here in October, but I have another question to add in case Martin is still checking it. Something I had never heard before was the danger of having too high a minimum stable heat output in the shoulders of the heating season when the actual demand might be less than the minimum stable output and therefore require cycling on and off. That is very interesting. If COP values suffer, how important is that likely to be in the full season average COP? Won't it, like the very lowest temperatures, contribute only a small amount of heating to the total and so have only a small impact on seasonal averages? I would expect this to be more days than the very low minimums in winter, but by definition very little energy consumption per day so a small impact on seasonal COP. Any comments on this reasoning?

    I see what you mean about the Mitsubishi multizone ductless minisplits having exceptionally low. Most heatpumps have minimums at 47 outside temps that are around 15-20% of their max at 47, these Mitsubishi multizone models have a minimum that is all of 50% of their maximum. If any heating systems were going to suffer this 'oversize penalty' during the shoulder heating seasons it would be them. Fujitsu has a couple cold-climate mujltizone ductless models (AOU18RLXFZH and AOU24RLXFZH) that are rated as functional down to -15 (full hyperheat) but if you look at the specs they lose capacity at an unusual high rate between +5 and -15, meaning you would have to oversize dramatically to meet the demand. It seems like there are drawbacks to all the multizone ductless systems if singlezone systems will do, but I guess installation costs of several single zones would be higher. I'd appreciate any further commentary you have on these subjects.

  35. GBA Editor
    Martin Holladay | | #35

    Response to Jonathan Comstock
    Thanks for your comments. Yes, there is increasing evidence that the COP of a minisplit suffers during the shoulder seasons (fall and spring) because of cycling. If the heating demand is lower than a unit's minimum heat output, the unit will shut off and start up again frequently, hurting the COP.

    I'm now writing a blog on the topic, so stay tuned.

  36. charlie_sullivan | | #36

    Cycling vs. part load operation
    There are two effects--the cycling means you get start-up/shutdown losses, mostly associated with pressurizing the refrigerant in the condensor on startup and then letting the pressures equalize on shut down. But the other effect is that you gain efficiency by running the compressor and fans at lower speed. Consider operation at a load that is 50% of full capacity. If you compare meeting that requirement by cycling on and off with 50% duty cycle, to steady operation at 50% capacity, the steady operation is better not only because you avoid those startup/shutdown losses, but also because the COP is higher during steady operation at lower speed.

    Both of those effects are in addition to the fact that the COP is higher with a higher outdoor temperature.

    I would think that shoulder-season operation could still have decent overall COP, given that the operation can be at low speed, and is with a relatively high outdoor temperature, even if it has to cycle. But perhaps Martin with have data to contradict that in his upcoming blog.

  37. user-5356986 | | #37

    correct system sizing to avoid excess cycling
    It will be primarily a shoulder season issue, unless the system is oversized. I have a friend who has been arguing that the heatload at the minimum F design temperature should be matched to the 'rated' capacity rather than the 'maximum' capacity, I guess for a safety factor for extreme cold days and perhaps insurance against uncertainty as to the precise accuracy of heatload calculations. I've been arguing that that degree of oversizing is a bad idea because it will result in 'cycling' not just in the shoulder season but most of the entire heating season. What do you think?

  38. user-5356986 | | #38

    refined question on sizing
    Relevant to comment 37, please give a brief clarification this paragraph of the article (see below) for me. Specifically, when you talk about oversizing as much as 1.5 times, at what temperature are you making the 'oversized calculation? In my area the 'design temp' for heating is 0F. Do you mean I could oversize the capacity at 0F by 1.5 the expected demand (for safety etc) and still have efficient operation throughout the heating season (even though the 'overcapacity at 40F outdoor temps would be much more than 1.5x), or do you mean that I can match more exactly match the heating load needed at 0F and as long as the overcapacity at higher outdoor temperatures doesn't exceed 1.5 times the capacity at those higher temperatures the system will still operate with high efficiency ?

    (From this article above): On the advantages and disadvantages of oversizing. “Oversizing by as much as 1.5 times would usually leave the equipment in the modulation zone most of the season and provide some margin for lower temperatures, and run slightly higher efficiency than if exactly sized for the load. Above that [in other words, oversizing by more than 50%] and it’s pretty much downhill on both comfort and efficiency.”

  39. Expert Member
    Dana Dorsett | | #39

    The ~1.5x oversizing thing...
    Oversizing by 1.5x for it's output capacity at the 99% temperature bin is something of an upper bound, not an optimum, and even 1.5x isn't always a good idea, depending on the minimum modulated output of the unit in question.

    The modulation range of minisplits vary by manufacturer & model, but it's not infinite. Oversizing by more than 1.5x usually results in a minimum modulation higher than the load at temperatures in the 20s or 30s F, which puts the mini-split in the mode of cycling on/off with significant room temperature swings.

    Assuming a 60F heating/cooling balance point for the house and an outside design temp of 0F, the heat load at 30F is half the design load, and the heat load at +45F is a quarter of the design load. At 1.5x oversizing at 0F the load at 30F is less than 1/3 of the output capacity at 30F, since the capacity increases with outdoor temps. At 45F the load is less than 1/6th of of the output capacity. Many mini-splits will not be able to modulate low enough to be running continuously at 45F with a 1.5x oversizing factor @ 0F, but some will. Almost all would be in a modulating mode at 30F at a 1.5x oversizing factor @ 0F, but virtually none would be modulating if the oversizing factor were 2x.

    It's useful to consider the minimum modulated output at +47F (part of the HSPF test- min & max capacity at +47F is in the submittal sheets as well as the modulation level @ 47F at which the efficiency was tested, the "rated" capacity). Compare the minimum output @ 47F to your heat load @ 47F to get a sense of the outdoor temp at which it begins cycling rather than modulating. If it's cycling at too low an outdoor temp it can directly affect comfort.

    Example case, an exception that proves the rule :-) :

    I recently specified a 1.5 ton Fujitsu with a capacity of 20,000 BTU/hr @ +17F. The load at local design temp of +12F was ~11,800 BTU/hr, which means it was roughly a 1.7x oversizing factor- more than ideal. But since that unit modulates down to 3100 BTU/hr @ 47F it's not a problem: Assuming a 60F heating/cooling balance point, a heat load of 11.8K @ +12F implies a heat load of about 5900 BTU/hr @ 36F, and a heat load of about 2950 BTU/hr @ 48F, so it'll still be capable of running in a modulating mode well into the 40s F but not into the 50s. If the heating/cooling balance point turns out to be closer to 65F it'll be modulating into the very low 50sF.

    The next size down in that series was a 1 ton, with an output of 13,500 BTU/hr @ +17F, which means it would still cover the load at +12F (I didn't bother to look up the extended temp capacity tables, but it'll be at least a 1.1x oversizing factor @ +12F). But it is a ducted unit, with the ducts being routed in an as-yet-unfinished and not directly heated basement, the potential duct losses meant it would be hanging it on the hairy edge for capacity without a modulation advantage at warmer temps. (A 1.25 tonner would have theoretically been ideal for this situation, but it doesn't exist in that series.)

    The larger coils of the oversized unit will give it bit of an efficiency advantage over most of the heating range, enough of an advantage that it will beat the smaller more exactly sized unit in as-used efficiency, despite a somewhat lower tested HSPF. Had the min-modulation of the 1.5 tonner at 47F had been 5000 BTU/hr or higher it would have been better to go with the smaller unit, since it would then be at the very low end of it's modulating range at the average winter temp at that location, and cycling on/off constantly during the shoulder seasons. It might have slightly higher efficiency or not due to running only in low-modulation range, but not enough to be worth the shoulder season discomfort.

  40. user-5356986 | | #40

    seasonal COP in cold winters
    I appreciate these informative answers!

    I further question I have is related to expected seasonal heating COP, again with a design temperature near 0 F. I'm aware of several review articles citing seasonal COP values of 2.5-3 even in cold winter situations. Again, I am trying to reach agreement with a friend who does not actually deny these data but doesn't fully accept them because a lot of the reviews are not clear on methodologies used. Are the COP referred to actually the entire energy consumption by the compressor and all pump units etc, or sometimes just the compressor? Those kinds of questions come up. He tends to rely on a laboratory result that comes up with an estimate of only 2.2. Its not a true field number, but the methodologies are transparent. I think its a miserly estimate. Can you point me towards the best documentation for cold climates that would have compelling descriptions of the measurement methodologies? We are formulating some guidelines for a community program encouraging adoption and I think this 2.2 is too low a figure to be citing.

  41. GBA Editor
    Martin Holladay | | #41

    Response to Jonathan Comstock
    I hate to hold my cards close to my chest, but I have written an article that will answer many of your questions. The article is scheduled to be published on Friday. I hope you can wait until then for the answers.

    [Later edit: Here is a link to the article: Ductless Minisplits May Not Be As Efficient As We Thought.]

  42. user-5356986 | | #42

    Great Stuff Here
    Martin- Thanks! I will be watching for it!

  43. T_T | | #43

    Mitsubishi MSZ-FH09NA

    Mitsubishi MSZ-FH09NA
    Min-max range of 3,100 - 12,000 BTU/h at 47F

    a] Is a certain part of that range most efficient? (for example halfway at 5,900)
    b] Is the *lower* bound of the modulation range affected by temperature?
    b] Is the unit always modulating when the output is in that range?
    So for example when outputting 9,000 BTU/h at 5F it still modulates.

  44. srivenkat | | #44

    Single MVZ connection possible
    Per the below manual:
    "Single unit connection is possible only with MVZ model. Connect 2 or more units for models other than MVZ."
    I heard the single-MVZ/MXZ combo is not AHRI-listed but is possible.

  45. user-5994916 | | #45

    Hello, Does anyone had experience with the MSZ-EF version (comes in 9, 12,15, and 18)? I like the appearance of these a lot more. I am having trouble finding information on the EF (or Zen) series that is geared for the US, and am trying to figure out if BTU and Seer ratings are comparable to the FH range.

    Thank you!

  46. GBA Editor
    Martin Holladay | | #46

    Response to Whitney Scurlock
    I don't have any experience to share, but here is a suggestion: if you post your question on GBA's Q&A page, you are more likely to get an answer (because more readers will see your question). Here is the link: Q&A page.

  47. srivenkat | | #47

    It's mentioned: "At an outdoor temperature of 5°F, the COP of this equipment varies from 2.09 to 2.86."

    I am trying to arrive at this using the Capacity tables on PDF page 16:

    For indoor temperature of 75F, TC=14.6 kBtu/hr, IP=2.34kW

    Using the formula 1kW = 3412.142 BTU/hr, 2.34kW translates to 7984.41 Btu/h?

    COP should be 14,600 Btu/h / 7984.41 = 1.83, right?


    Please ignore my above question. After studying Dana's excellent explanation in comment 11 on the below thread, it has become aparent to me that I was not considering the "nominal/rated" output of this unit at 5F, rather I was looking at the "Total Capacity" output when the unit is running flat-out.

  48. Jon_R | | #48

    > At an outdoor temperature
    > At an outdoor temperature of 5°F, the COP of this equipment varies from 2.18 to 3.76.

    Would be interesting to see an explanation of how to replicate these numbers from Fujitsu data.

  49. ethant | | #49

    It is difficult to read through this post and not see any reference to the fact that refrigerants are a major contributor to global warming and climate change (sea level rise, etc.)

    According to "Drawdown,"

    HFCs, the primary replacement, spare the ozone layer, but have 1,000 to 9,000 times greater capacity to warm the atmosphere than carbon dioxide.

    I am eagerly awaiting alternatives to HFCs, including, perhaps CO2, though my understanding is that CO2 as a "refrigerant" is mainly good for heating, which is why it is found in the Sanden and Ecocute water heaters. Perhaps we should be discussing eliminating cooling loads vea sufficient insulation and shading and then applying non-HFC technologies to our heating needs. Or am I overthinking this?

  50. iconoclast2222 | | #50

    What precisely constitutes a zone? Sure, I generally know what it means but in some contexts in this series of articles, I am left scratching my head. Notably, when stated that a ducted system is used, there must be at least 2 zones. I’m probably missing something fairly straightforward, as sometimes happens. 🤷‍♂️

    1. GBA Editor
      Martin Holladay | | #51

      A zone is a room or rooms controlled by a single thermostat. Most ducted systems are single-zone systems, although zoning a ducted system is possible.

  51. rpratt905 | | #52

    Do you provide consulting services? I am building a "near" Passive House in the mountains of Colorado and want a second opinion on my 3 zone mini-split heating strategy. Framing is nearing completion and I want to give the HVAC contractor (which I am still looking for) first crack at the MEP portion of the up coming work. [email protected]

    1. GBA Editor
      Martin Holladay | | #53

      R. Pratt,
      I'm a journalist, not a consultant. But if you post questions on GBA's Q&A page (here is the link), you'll get excellent advice and feedback -- and the advice will be free.

  52. tsd2626 | | #54

    2021 and I can confirm, Mitsubishi's website is still horrendous.

Log in or become a member to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |