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

Two choices for minisplit sizing

bac478 | Posted in Mechanicals on

Trying to size a Mitsubishi Hyper Heat Mini Split:

Climate Zone 5A, Zip 01301, Heating Design Temp -2F (approx), second story space, 2 bedrooms and a hallway, 450 sqft floor area, R19 walls, ~R60 ceiling, ~30sqft window area in N and S walls, none on W or E. Don’t know what to use for airtightness. 1890 farmhouse with dense pack in the cavities, membrain VB, thin polyiso backer for the vinyl siding that was installed over wood clapboards. I paid attention to airtightness details when I gutted this part of the house. I was estimating “average” tightness.

Depending on how I input airtightness I get anywhere from just over 9k Btu to just under 11k Btu heat load. I am sizing the unit for the house now, I have plans to improve airtightness and add exterior insulation when I install new siding.

I am trying to decide between a 9k unit and a 12k unit. When I look at the mitsubishi engineering manual it looks like heating capacity at 5F outdoor temp is only about 5200 Btu for the 9k and 6500 Btu for the 12k. These are both below the heat load. Clearly the unit can’t be sized for the worst case load or it would be oversize the other 99% of the time. What is the most sensible way to decide between the 9k or 12k unit. For reference it looks like the $ difference is only about $200 between the two and the 12k is slightly less efficient.


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

    It's seems unlikely that 450' of conditioned space with R19 walls and only 30 square feet of window ( or is that 60 feet total, 30 on each wall?) has a true heat load anywhere near 9000 BTU/hr (20 BTU/hr per square foot, which is higher than a lot of 2x4 construction would be) let alone 11,000 BTU/hr @ -2F as long as it has glass in the windows. It's conceivable the load could be as high 6500 BTU/hr, but it could also be as low as 5500 BTU/hr if it's very air tight.

    If you were paying attention to air sealing details during the rehab it's going to be far tighter than "average". What was the BTU/hr number associated with air infiltration in your calculation?

    With the air tightness presumtion of "tight" what is the calculated heat load at 20F, 30F and 45F?

    The mean temp in Greenfield in January is about 22F, the December mean temp is about 30F, and the March mean temp is about 35F, eyeballing it from Weatherspark datasets, so don't want it to be maxed out at 20F, but modulating rather than cycling when it's in the 30s. The minimum modulated output of the FH09 at +47 F is 1700 BTU/hr, but for the FH12 it's 3700 BTU/hr. Ideally you'd want the thing to be modulating rather than cycling whenever it's actually cold out:

    The HSPF doesn't really tell you much about it's operational efficiency in a zone 5 climate, nor does it tell you the efficiency when the unit is oversized or undersized for the load. From a raw efficiency point of view it'll do somewhat better if oversized by 25-50%, but from a comfort point of view you'll be better off if it's modulating when it's below 40F (or at least by 30F) rather than cycling. If it turns out the FH09 is undersized for the load you'll get better as-used efficiency out of the FH12, since it'll be modulating down to a much lower outdoor temp than the FH09. If running flat-out full speed at +15F it won't have a COP better than about 2, but if modulating it'll be in the mid-2s or higher.

    When cycling on/off they're somewhat less efficient than at minimum-modulation, but the room temperatures also take much bigger swings. The temperature swings are sometimes solvable by retrofitting a wall thermostat, but that's also a significant cost-adder, since they are manufacturer-specific items, and produced in fairly low volumes.

  2. bac478 | | #2

    Dana let me try to answer your questions:
    1. 3x5 window, qty 2 per wall; qty 2 walls (N and S), = 30sqft/wall or 60 sqft total window area
    2. 1 ACH air change assumed using guideline here
    3. If I set infiltration to tight I get 5566 Btu/hr @20F, 4453 Btu/hr @ 30F, and 2783 Btu/Hr @ 45F. Used calculator here

    Let me explain 2 and 3 above. I used two calculators as a sanity check. One takes in ACH as a number and gives guidelines of what to set it at with 1 equated to "leaky". The other calculator just has a subjective drop down ranging from very tight to very poor with 5 options. They also operate slightly differently, one bases results on heating degree days and the other lets you input an indoor and outdoor temp and calculates the heat req for those parameters only. I used outdoor @ -3 so the results are a worst case number. Results as follows:
    Calculator 1 (-3F outdoor, 70F indoor):
    1. At very tight I get 8126
    2. At average I get 9925
    3. At very poor I get 12983

    Calculator 2 (-3F outdoor, 7200 heating degree days)
    1. At .33 ACH I get 6782
    2. At .5 ACH I get 7533
    3. At 1 ACH I get 9740
    4. At 1.7 ACH I get 12831

    Sounds like the goal is to keep the load above the min modulation limit for as much of the time as possible without going so small that the unit can't keep up on an "average" winter day. Clearly in fall spring there will be cycling because outdoor temp is close to desired indoor temp.

    At 20F the FH09 maxes out at about 7500 Btu/hr
    At 20 F the FH12 maxes out at about 9300 Btu/hr

    One of the calculators allows me to change the outdoor temp and adjusts load accordingly. I can use this to figure out around what outdoor temp the unit would start to cycle
    1. The FH09 would hit min modulation at about 55 deg outdoor
    2. The FH12 would hit min modulation at about 36 deg outdoor

    Looks like the FH09 will keep up almost all year and will avoid cycling maybe 10 months out of 12. Are there things I am failing to consider?


  3. Reid Baldwin | | #3

    If the ACH on the second floor is from stack effect, the air that is leaving the 2nd floor will be replaced by air coming upstairs from the 1st floor, which is presumably already heated.

  4. bac478 | | #4

    That is a good point and I hadn't thought of it. Thinking FH09....

  5. bac478 | | #5

    Just looked into the fujitsu units. Looks like the 9RLS3H is comparable. I estimate I am going to need about 35' of lineset and the Fujitsu can handle up to 49' with the factory refrigerant charge in the unit. The Mitsu is only good for 25'. I would have to find a tech with a refrigerant scale to meter out the additional charge. I can handle the vacuum/nitro purge myself.

    Looks like the Fujitsu has a min load rating of 3100 Btu/hr so it would cycle more than the Mitsu in the transition seasons and hurt efficiency a little. The "h" version (Fujitsu) looks like it has a built in pan heater, additional drain holes, and some structural enhancements for cold weather. Advertised down to -15F heating.

    That said I think the mitsu is a better fit based on load but the Fuji would be easier to install because it wouldn't require additional refrigerant.

  6. Expert Member
    Dana Dorsett | | #6

    The 9RLS3H is more comparable to the FH12 than the FH09, in heating applications, including it's minimum modulation level @ 47F. At -5F the max capacity of the 9RLS3H running at max speed puts out about 14,000 BTU/hr the same as the max capacity of the FH12 at +5F.

    The "nominal" heating output of 12,000 BTU/hr of the 9RLS3H is about what it delivers at -15F(!). So from a heating capacity point of view the 3/4 ton Fujitsu is a somewhat bigger unit than the 1-ton Mitsubishi, and it'll have you covered even if the worst case numbers in your calculation runs prove true (not likely.)

    Note, that as the temperatures drop, so does the minimum modulated output. The min & max output at 47F is just the data set for 47F. I'm sure at 30F the

    If you were building with air-tightness in mind you'll be tighter than the BuildItSolar guidelines, at least for the part that you were rehabbing. This is why mechanical ventilation is almost always necessary in new tight construction. But the heat loss related to that is exaggerated, even under Manual-J methods, since the air isn't all entering or leaving via a single straight & round hole. There is significant "heat exchanger effect" in real-world random leakage- the air never enters the house at the outdoor temperature, nor does it leave the exterior surface of the house at the interior temperature. It's not like a 90% efficient HRV, but it's not a straight pipe- it's something in between. The higher-R you make the house, the bigger the infiltration loss numbers appear compared to the other load factors, but you have to understand that they're large relative to typical in-situ reality, often by a factor of 2 or 3.

    Even with blower door test numbers you don't know where the leaks are or how big they are, just the aggregate size. But the location matters in terms of how much heat exchanger effect you get out of it.

  7. bac478 | | #7

    Dana - clearly you have a lot of experience with sizing these units. Based on the information provided which seems most appropriate; the 9k Btu Mitsubishi, the 12k Btu Mitsubishi, or the 9k Fujitsu?

  8. Dana1 | | #8

    For a fairly open & connected 450' space with reasonable convective connection:

    If it's a couple of doored off rooms with little convective distribution between them:

    (Oversize the ducts, keep them short.)

  9. bac478 | | #9

    There are two bedrooms and a small 1/2 bath upstairs. I was going to mount the head in the hallway above the stairs to one side if I went with the ductless. See floorplan attached with location highlighted in yellow. This is pretty similar to the Carter Scott arrangement on this site. This is what I was planning on doing so I designed around it.

    If I move a light fixture I could probably install a drop ceiling above part of the stairs with a return grille right underneath the ducted unit. Then I could have a plenum off the supply side with about 10" of retangular duct off of each side (like a giant tee) that went into each room. Terrible sketch attached: blue is the unit, red is the ductwork. This would provide distribution to each bedroom with very minimal ductwork. It is more work than the ductless.

    Currently it is me and my fiancee, the second bedroom won't be used. Open doors during the day shouldnt be an issue. I was thinking heat would distribute pretty well. I was worried about cool air falling down the stairs but there is a door at the bottom. Is the ducted worth the extra work & materials, lower efficiency (HSPF 12 vs 14), and higher min heating temp (-5F vs -15F) for better distribution?

  10. Dana1 | | #10

    If the heat loads of the individual rooms were 1500 BTU/hr a single head would work fine, but they're roughly twice that. A ducted unit guarantees that it all just works.

    The efficiency number tests are for some arbitrary zone 4 set of conditions with an outside design temp of +17F, sized exactly for the load. The test doesn't reflect the true efficiency of modulating units, or the beneficial t effects of modest oversizing. The larger coils of the larger units give them a higher efficiency at some intermediate load when running at 1/3 modulation vs. a smaller unit running at 2/3 of it max speed. Since the ducted Fujistu units all modulate down to the same 3100 BTU/hr @ 47F, there isn't a comfort hit for oversizing. The nominal capacity of the 9RLFCD is 12,000 BTU/hr and will already have some oversizing factor built in for the likely load. But if you wanted to be sure, you could bump that to the 12RLFCD, and have guaranteed margin, and despite the lower HSPF numbers, it'll be more efficient than the 9RLFCD as-used:

    The down side other than the modest uptick in upfront cost is the higher air volumes- it might feel pretty breezy even at mid-speed. My gut tells me based on the range of numbers presented is that the 9RLFCD is already approaching the optimal oversizing factor, and the improvement of going to the 12 would be pretty small. If the true load is at the high end of the range presented the 12 might be a better choice.

    For ease of setting up the condensate drain, it might be better to install it in the top of the bathroom ceiling, due to the proximity of the plumbing chases, but that's not the only factor. Routing the refrigerant lines or placement of the compressor unit may trump that.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |