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

Minisplit heat pump sizing issue

nhsportsman79 | Posted in Mechanicals on

Hi All,

I am hoping someone can point me in the right direction.
We are thinking to add mini split ductless heat pumps to our existing 1550 sf raised ranch in Manchester NH. The house has good insulation for the age with R50 in the attic and 15 year old windows. It has 2×4 walls so not sure about R-value there. Currently have hot water baseboard heating system with 25 year old oil fired boiler and our oil consumption averages about 450 gal / year.

We had contractor do head load/loss last year, but my fear is that he used higher design temperatures that those specified in the Manual J catalog which for Manchester NH are (-3 and 88). He used (-20 and 95) so I am not sure if his designed temperatures would affect size of the equipment. For living room he recommended Mitsu MSZ-FE18NA but I think it’s oversized for our living room. My plan is to use Fujitsu 12RLS3(H) or non H for living room and Fujitsu 12RLS3(H) or non H for open area that includes (dining room, kitchen, entry area). For the 3 small bedrooms I would go with multi (3 zone) 7RLF1. I think this is oversized for the size of our bedroom but these are the smallest Fujitsu units.

Attached is heat load/loss for your reference.

Please let me know if my thought process is right or if you can point me in the right direction.

Looking forward to your feedback.

Alex Frost

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. nhsportsman79 | | #1

    Sorry....forgot to include heat load/loss attachment.

  2. Expert Member
    Dana Dorsett | | #2

    He didn't spell out the U-factors in the calculations, only presented the results, and there must be other thumbs on the scale than the outside design temp. Even -20F I don't quite believe 49,000 BTU/hr of load for a 1550' house that has a fraction of the house below grade. That's a ratio over 30 BTU/ft.- I'd be willing to bet it's under 20 BTU/ft, which would be less than 31,000 BTU/hr of total load @ -20F.

    The difference between the 71F inside design temp and the -20F outside design temp is 91F. If you used -3F that would be a delta of 74F. So even if you used his 49K heat load and scaled it, the load at -3F drops to 49K x 74/91= 40K, and the 31K WAG upper bound would imply a heat load of 31K x 74/91= 25K @ -3F a ratio of 16BTU/ft-hr, which isn't out of the question, but probably higher than reality.

    If you keep the place at 70F and only use 450 gallons of oil over a 7000 heating degree-day winter the implied heat load can be estimated this way:

    At 138,000BTU/gallon and 85% efficiency you're getting 0.85 x 138,000 = 117,300 BTU/gallon out of the oil.

    At 450 gallons/year that's 450 x 117,300= 52.785 MMBTU/year (MM= million)

    52.785MMBTU/7000 HDD= 7541 BTU per degree day.

    7541/24= 314 / BTU per degree-hour.

    Working backward from the 65F degree-day base, at -3F you have 68F heating degrees, for an implied heat load of:

    314 x 68F= 21,352 BTU/hr @ -3F.

    At -20F you'd have 85 heating degrees, and an implied heat load of

    314 x 85F= 26,690 BTU/hr

    That's barely more than half his calculated heat load.

    If you have a mid to late winter fill-up with the exact fill-up dates and oil quantities we can look up the degree-days for Manchester to zoom in on that more precisely, but his total numbers are more than a bit off, unless you were heating with a wood stove part of the time or keep it at 55F indoors or something.

  3. nhsportsman79 | | #3

    Hi Dana,

    Thank you so much for elaborate explanation and confirming my hypothesis that the heat load contractor provided to me is not quite accurate.

    Here is more info on our 450gal oil consumption. We have 2 tanks in our basement and I fill them up (around 350gal) in July or August when the oil is the cheapest. Then in March or April I add 100 gal just to make sure I don’t’ run out and the cycle repeats again every year. Also, we have 3 zones: #1 ( living room), #2 dining room, kitchen, entry way, bathroom, #3 (3 bedrooms). Our bedrooms thermostat stays about at 60-62 during the day (have small kids that don’t spend much time in their bedroom) we keep at 66 at night in bedrooms. Zone #1 and #2 are kept at 68 during the day and drop it at 60 during the night.

    Do you know of any free spreadsheet or online calculator that will allow me to input his measurements and change the parameters that you suggested?

    Lastly, assuming your estimate of 26,690 BTU is close enough what kind of mini split system / configuration would you suggest.

    Many thanks!!

  4. Expert Member
    Dana Dorsett | | #4

    With those thermostat settings it's probably worth re-calculating based on a 55-60F heating degree day base, which is far fewer HDD. But it's also clear that you shouldn't be using 71F for an indoor design temp. Code requires that the heating system be capable of heating every room to 68F at the 99% outside design temp, so projecting the load forward a few degrees to 68F is sufficient.

    Looking at the numbers from, Manchester NH logged 19999 base 65F HDD over the past 36 months, which is 6666 HDD/year, but 12,859 base 55F HDD, or 4286 HDD/year. Using base-55 F will overestimate reality slightly, but let's see how that looks.

    52.785 MMBTU/year over 4286 HDD is 12,316 BTU/HDD or 513 BTU/degree hour.

    At -3F you are 58F below the presumptive 55F base, so the heat load @ -3F is 58F x 513= 29,754 BTU/hr.

    At -20F you are 75F below the presumptive 55F base, so the heat load @ -20F is 70F x 513= 35,910 BTU/hr.

    It's ridiculous to design to -20F to cover the "extra" 6000 BTU/hr of load below the 99th percentile temperature bin of -3F. If it STAYS that cold during a major Polar Vortex event you can almost make that up with a single 1500W electric space heater (!), and even with modest oversizing of the mini-splits you won't actually need even that much auxiliary heat.

    If you add 5 heating degrees to cover the difference between your average thermostat setpoints that adds 2565 BTU/hr to the estimated heat load, but in fact it's less than that, since your true balance point is going to be a bit above 55F in your 60-68F home. Call it 30,000 BTU/hr as the probable worst-case, 26-28K is more likely, but it's probably not less than 20K (yet).

    This is actually pretty typical for a 2x4 framed house that size with no foundation insulation, but would be on the high side if the foundation walls are fully insulated. If yours has a cantilevered second floor you may have some rectifiable air leakage & insulation issues at the cantilever. If there is no below-grade insulation there may be worthwhile retrofit possibilities there too, but it really depends.

    The zone-by-zone loads don't necessarily have the same proportional error as the whole house load, which makes it a bit difficult to make good recommendations. If you run your own room-by-room I=B=R spreadsheet based on the U-factors of the different types of exterior surfaces it would be a lot better. Simply down-sizing proportionally from mysteriously-derived numbers based on the estimated real load would NOT be a good bet.

    Any of the units on the Efficiency Vermont pre-approved list for rebates would be suitable in a Manchester NH climate:

    Of the Fujitsu xxRLS2s or RLS3s the "H" versions are also a much safer bet, since the others can self-damage the outdoor unit's blower during extended cold weather from accumulated defrost ice in the bottom of the pan. Mitsubishi also sells automatic pan heaters for their FExxNA and FHxxNA series to avoid those problems, and somewhere in the fine print of the warranty you won't be covered without them in climates as cold as yours.

  5. nhsportsman79 | | #5

    Thanks once again Dana...

    Can you clarify what is I=B=R spreadsheet or where to get it? I did quick google search and nothing came up..

  6. nhsportsman79 | | #6


    Sorry I spoke (wrote) to fast. I found IBR spreadsheet, but I am more puzzled with results it provided because when I plugged in all the measurement and assumption the heat loss came close to 18K but for living room which is similar to the heat loss contractor provided but my design temperature were different. Could you please look at the attached file and let me know if and what I am doing wrong.


  7. Expert Member
    Dana Dorsett | | #7

    That's a pretty crummy tool where did it come from?

    The U-factor of a 2x4 16"o.c. wall with R13s (just about everything built after 1962) with wood sheathing & siding would come in at about U0.095 BTU/hr per square foot per degree F (some would estimate U0.08, counting the air-films) At a (68F- -3F=) 71F temperature difference and U0.095 that's a heat loss of 6.75 BTU/hr per square foot of wall area. Count ONLY the wall area, subtracting out the window & door area.

    Clear glass windows with clear glass storms run about U0.5, so at a 71F delta that's 36 BTU/hr per square foot of window. If they are low-E replacement windows or low-E storms you're looking at no more than about 25 BTU/hr per square foot of window.

    Solid 2" doors are also about U0.5.

    An R50-ish attic call it U0.02, for 1.42 BTU/hr per square foot of ceiling.

    Subsoil temps are about 50F, but dry soil has both R-value and thermal mass working in your favor, so the modeling used in your downloaded tools are going to exaggerate by a very large amount. Ignore the slab losses unless it's slab-on grade, since the load doesn't change with daily outdoor temperature swings and the magnitude of that load is comparatively small. If the slab is more than 2' below grade call it zero for now, maybe fudge-factor it in later if it's an issue.

    Build a room-by-room spreadsheet on those numbers and see where it comes out. If it's a tight house the infiltration losses won't exceed the interior heat sources from electrical plug loads and mammalian body heat by very much, but we can fudge-factor those in too with a bit more information.

  8. nhsportsman79 | | #8

    I believe I found it on website.

    Ok..I will build my own spreadsheet with the data you provided.

    How do I figure out heat gain now that I am working on this?

  9. nhsportsman79 | | #9

    This is my cut at the heat loss based on your inputs. Can you let me know if I am on the right track before I do other rooms?

    Thanks Dana!!!

  10. Expert Member
    Dana Dorsett | | #10

    That's the way you would do it at a fixed 71F delta-T. If you want to do it at some other delta-T make a cell for delta-T, and use that, inserting a formula into the boxes to come up with the room loads broken down by surface type (wall, window, ceiling), then summing those loads for the room load. If you use a cell to insert the U-factor constants you can then change them globally if you find out something different about the wall stackup or ceiling insulation based on what you find poking around.

    U-factor x area x delta-T = BTU/hr

    If you arrange it nicely on the spreadsheet you can then sum up the total window, wall, and ceiling losses separately.

    Then the question marks become how leaky is the house REALLY, and how cold is the slab. But you'll at least be starting from a good base line. Most infiltration loss estimators grossly overstate reality, since there is a significant amount of "heat exchanger effect" along most leakage paths. The incoming infiltration air isn't really as cold as the outdoors, and the outgoing air isn't at the full room temp. (The Lunos HRV system uses this effect and the thermal mass of it's cores to change the air direction every 70 seconds with every little heat loss even at substantial air volumes per hour. Your air random leak heat exchangers aren't nearly as good as that, but they are better than most heat load calculators would imply.)

    For bedrooms you can subtract 250 BTU/hr per sleeping human. For other rooms look for 24/7 loads like refrigerators, DVRs , hot water heaters, etc. and we can come up with some reasonable estimations.

  11. nhsportsman79 | | #11


    I have finished whole house calculation and the total house loss seems low and I haven’t used any of the “reductions” i.e. human body heat, refrigerator, TV’s..etc as you suggested and by doing so the numbers would drop even more.

    I increased the indoor design temperature to 71 which increased delta-T to 74 because my experience with radiant heat and forced hot air is that hot air heat makes us comfortable at 71-72 while radiant heat makes us comfortable at 68-69.

    Our lover level walls are insulated as well and concrete foundation has 2 inch insulation on it. Our lower level temperature stays around 60F in the winter (due to boiler and hot water piping) being in lower level. Recently I have installed hybrid heat pump water heater and I noticed about 2F decrease in the temperature (down to 58F) which is expected for these type of water heater due to amount of cold air exhaust.

    Let me know what you think.

  12. Expert Member
    Dana Dorsett | | #12

    A raw unadjusted I=B=R heat load of 20K for a 1550' house with an insulated foundation is about what I would expect. If the slab is uninsulated it might add another 2K before subtracting off plug-loads, etc. and the infiltration numbers could add as much as 25-30% but probably not more, if it's a pretty tight house. With all of the adjustments you're still probably under 25K total. If you want some capacity assurance, upsize the equipment capacity at -3F by 25% from your I=B=R numbers and you'll be fine, but not more than 50% if you can help it. You'll have to look at the curves or extended capacity tables to figure that. If the step sizing between the units in the same series oversizes the capacity by more than 50% it won't modulate much in the shoulder seasons, so don't go whole-hog for covering the entire load at -20F.

    Modulating mini-splits are more comfortable than on/off ducted hot air systems since they don't produce much wind-chill and the room temps stay steady. It's not the same comfort level of radiant, but way better than on/off cycling hot air.

    The hybrid HW heater is essentially a net-wash, neither a big heat sink nor a big heat source. If you're using a lot of hot water and sending that heat literally down the drain it's a net heat loss, but unless you're taking an endless shower when it's -3F outside it's not going to affect your load numbers much. Without the standby loss of the boiler it might hit 50F down there in the depths of winter, but it won't freeze.

    The living room zone is well within the output of the 9RLS3H, which can deliver about 14,000 BTU/hr at -3F with a 70F interior, which is a hair over 50% higher than the calculated load, but it can modulate down to about 3K even at warmer temps, so the on/off cycling isn't going to be a problem. If there is free & open air flow between the kitchen/dining/entry and the living room it would balance the combined load almost exactly, and the 1-ton or even 1.25 ton might make more sense, if it's handling both zones. If that's not the case, if the layout is conducive you might combine that smaller zone with the mini-ducted bedroom zone, if that's even possible (it gets awkward if they are on different floors or are at opposite ends of the house.)

    The capacity tables for the xxRLS3H are similar to those of the RLS2-H series. The capacity table for the 9RLS2H can be found at the top of p15 (PDF pagination) here, the 12RLS2H table is at the bottom of the page, and the 15RLS2H is on p.16:

    (The TC numbers in the boxes are "Total Capacity", in KBTU/hr)

    The three bedrooms combined add up to about 5.5K, which is within the output of the 9RLFCD mini-duct cassette. It won't modulate much during the shoulder seasons, but should do just fine in winter, and since you spend most of your bedroom time at night when the loads are higher, it's not much of a comfort issue if it's cycling during the day.

    If you combine the kitchen/dining/entry into the mini-ducted zone into that you'd be at a bit over 10K of load and the bigger units make more sense. The 1.5 ton unit can almost cover the whole house load, but the 1-ton might be more appropriate for the bedrooms + kitchen load.

    I don't have the extended temperature tables for the mini-ducted units, but the -5F total capacity will be more than half the max-capacity @ 47F in the submittals, so the 9RLF will have at least 9K @ -3F, the 12RLF would have more than 10K, and the 18RLF more than 13K. A Fujitsu contractor should have access to those tables if it needs more refinement. Since the minimum modulated output at 47F is 3K for all three in that series, there isn't a comfort penalty for oversizing, and oversizing by 25-50% delivers slightly higher average efficiency than something that has no margin.

    I could walk you through it on Mitsubishi's lineup as well, but I'll leave that as an exercise- you can find most of the necessary information in the tech & design manuals, or other documentation on the web.

    At least you're starting out with better-ballparked load numbers. If anybody tells you that they are wrong, make them prove it with a formal Manual-J calc based on something real, not a WAG. No way is the raw I=B=R calculation undershooting reality by 60%, which it would have to be if your original contractor's load were correct. More often than not the raw I=B=R calc overshoots reality a bit due to the internal heat sources. Three bedrooms/4 people is 1000BTU of internal heat source right there, then there's the plug loads. Methods that presume ridiculous infiltration losses are often the drivers of ridiculous oversizing factors. The infiltration losses are not zero, but they are also nothing remotely like 29,000 BTU/hr (with the windows closed, anyway :-) )

  13. nhsportsman79 | | #13

    I can't believe my true heat load comes out to be under 30k and my boiler is 100k BTU. Last year when I was thinking to replace the boiler the contractor wanted to put a new one with 100K BTU (granted I was going to go with Super Store water heater tank) but that would still be an overkill. I wonder if I had newer tripple pass boiler if my oil consumption would go even lower.

    I am concerned with comfort level with these mini split compared to radiant heat we have (primarily because wife might complain more with mini split as they won't make her comfortable as radiant heat).

    The only problem I see with going with concealed slim ducts is the loss of efficiency compared to RLS units.

    Lastly, just doing quick math on ROI, it does not look like it's a quick payback based on my average seasonal heating bill (oil consumption). It might take 10 years to recoup the investment unless the oil skyrockets over $5 / gal

  14. nhsportsman79 | | #14

    Hi Dana,

    Can you look at the attached floor plan and let me know if I could getaway with 1 unit (9 or 12 ) in the living room (on the western outside wall?). There is opening between living room and dinning room/kitchen continued down the hall toward bedroom. The goal would be to use the unit for heat in the shoulder months or down to 5 degrees or so and for cooling in summer.

    Or do you still think I need another unit in dinning room/kitchen area?

    I am on board for getting 9RLFCD and ducting in it to each bedroom.


Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |