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

Separate ducted mini-splits for each floor or a multi-split for the whole house?

whitenack | Posted in Green Products and Materials on

Hi all,

Building a house in Central KY (zone 4A) and trying to research what the appropriate strategy for zoning the HVAC. The traditional approach would be a conventional heat pump for each floor, allowing you to adjust the use for each floor when not occupied.

However, I am building a house with extra insulation, and the heat and cooling loads are not going to be “traditional”. I have done the room by room manual J calculations, and I am getting a total house heating/cooling load of 29k/27k. Breaking that down per floor gets me 18k/12.5 for the 1st floor and 11.4k/14.2k for the 2nd.

So, the question is whether it is better to keep the 1st and 2nd floors on separate, smaller systems (outdoor units) or combine the entire house into one system with multiple heads? I know that you lose some efficiency when you have an outdoor unit sized for two or three different indoor heads and only one or two are operating, or worse, the indoor unit is calling for a load less than what the outdoor unit requires to even kick on. Also, I guess having separate systems creates some redundancy if one unit goes out in the middle of peak demand you at least have one floor that is conditioned. Sounds like I am talking myself out of the multi-split idea.

Also, is there a rule of thumb for maximum length the ducts can be run? My first floor is fairly long from side to side and I’m worried about it being too long for a single ducted air handler. Wondering if a multi-split, at least for the first floor, would perform better.

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

    Unless you are willing to post a floor plan, it's going to be hard to provide full answers to your questions.

    Here are some basic principles:

    1. You're going to want to have your first floor and your second floor on separate zones (because most of the heating load needs to be delivered to the first floor, and most of the cooling load needs to be delivered to the second floor). This can be accomplished with either a multi-split unit or two ductless minisplit units.

    2. Whether or not your family will be satisfied with point-source heating on each floor (instead of heat delivered to each bedroom) depends on the heating load and cooling load of each bedroom, as well as the flexibility of family members.

    Here are links to two articles which you may be interested in reading:

    Rules of Thumb for Ductless Minisplits

    How To Buy a Ductless Minisplit

  2. Expert Member
    Dana Dorsett | | #2

    It's almost always better to zone multi-story houses by floor, even if it's operating on a single compressor.

    Efficiency-wise it's usually better to run each zone on it's own compressor than with a multi-split approach, since the minimum modulation levels are lower. Multi-split compressors typically have a minimum modulation of 6000-7000 BTU/hr, whereas single zone compressors are half that or less. So when just one of the zones has a load of 4000 BTU/hr and the other nearly none, it would force the compressor and zone cassette/head into a higher less efficiency mode than needed, cycling on/off rather than modulating, cutting even further into efficiency.

    Duct lengths are never determined by rules of thumb- it has to be designed for the flows required for meeting the heating/cooling requirements, as well as the specifications of the air handlers. Not all mini-duct cassettes can push the same amount of air at a given static pressure.

    A heating load of 29,000 BTU/hr in a zone 4 location with better than code insulation indicates that it's a pretty big house!

  3. whitenack | | #3

    Sorry, Martin, for not being clear. I'm for sure going with different zones on the 1st and 2nd floors. What I was trying to ask was whether I should have separate outdoor units for each floor, or run the whole house off of one outdoor unit with multiple heads. Thanks for the links. I'll do some studying.

    Dana, thanks for the info. I'm not sure my manual J calculations are exact...I've never done this before. I used an online program at that seemed pretty thorough, but there were some variables that I had to guess at. My local utility company ran some numbers for me and came up with 38k BTU/hr which didn't seem right at all. I'm looking at 3600-3800 square feet of living space, so I'm not sure if that is in line with what you were guessing or not. I've got quite a bit of glazing on the northwest wall.

    My floor plan is below, for those that are interested.

  4. whitenack | | #4

    By the way, I'm not promoting that website. Just referencing it in case someone wanted to check it out. Most online manual J calcs are pretty basic, and I didn't want someone to think I had just done a quick, back of the envelope calculation from some simple website.

  5. iLikeDirt | | #5

    I remember seeing this floorpan a while back, Clay. If I might be so bold, I think there are a lot of problems with it:

    1. There is a ton of wasted space. Your entry foyer is the size of a bedroom. The upstairs hall is the size of two bedrooms. You have four bathrooms, and three of them are gigantic. The upstairs laundry room is nearly the size of a small bedroom. There's room for a lot of improvement and consolidation without any loss of functionality, yielding significant cost savings. You could probably chop 200 square feet of hallways, foyers, and excess bathroom and laundry room space out of a house containing these rooms without even noticing it.
    2. The office, formal dining room, living room, and two of the upstairs bedrooms have windows on only one wall, and the office only has a single window, period. Those rooms will be be gloomy and depressing much of the day without artificial lighting.
    3. The kitchen work triangle is inefficient. Put the fridge where the cooktop or wall oven is, and put a freestanding slide-in range on the island.
    4. There appears to be a formal dining room as well as a dining area in the kitchen. Seems redundant, and the formal dining room will probably be rarely used. The public living room/private great room paradigm still makes a certain amount of sense but IMHO the formal dining room is an anachronism, especially with a dedicated dining area elsewhere.
    5. The master suite is way too big; you'll never use all that space. It will feel huge and intimidating, not warm and cozy the way it should.
    6. Having to walk through the master closet to get into the master bathroom is gonna get real old, real fast, and the clothes probably face an elevated risk of mildewing and smelling bad due to close proximity to the bathroom. That layout would be a negative feature for me if I were looking at this house as a buyer.
    7. The place for the laundry facilities upstairs should be as close as possible to the most-used closets, which in your case look to be the master closets. The washer/dryer should be within or just off of the master suite.
    8. Is that a mother-in-law suite off the kitchen? If it's not, seems weird that it would have a bathroom and laundry facilities. But it seems too small for that. What is it?

    And finally, I hesitate to include this part, but feel I must: my god, those walk-in closets! Each bedroom has its own large walk-in closet, and the master suite has TWO of them, one size of a bedroom! Walk-in closets comprise 233 square feet of this house! Nobody sane has that many clothes, it's crazy. Make your kids and your wife give some of their clothes to goodwill or something. That many walk-in closets of such enormous size just invites unhealthy hoarding, and if you resist that urge, they will look awkwardly bare. It's a lose-lose situation.

    To answer the original question, just spitballing I'm going to predict that there is no way a non-ducted mini-split per floor is going to be sufficient for this house in its current state unless it's insulated and air-sealed to above-passivehaus levels. It's too big, with too many interior walls and doors, and too much glass that's not crazy German superwindows. In the great room alone, you have 126 square feet of patio door space that is R-3.5. Those doors alone are losing you 35 BTUs of heat per degree of delta T. A ducted mini-split per floor seems eminently sane. But where are the mechanical rooms to hold them?

  6. Expert Member
    Dana Dorsett | | #6

    Most tight 3800' code-min houses would come in with heat loads at about 9-10 BTU/hr per square foot or 34-38K @ +10F (a typical 99% outside design temp in KY: )

    Your utility company probably just did a pro-forma quickie with all code-min assumptions, including code-max air leakage.

    But without knowing your actual wall stackups and insulation values, window U-factors there's no way to assess whether 29K is the right number or not.

    The average new house in the US runs 2400-2600 square feet- yours is on the large-ish side probably at the 1-sigma tick on the bell curve.

    For yuks I tried to run the Cool Calc tool on a house I already had good data on as a sanity check. The selections of wall stack ups were not even possible for this not-so-exotic house, not even the 24" roof overhang s were a selectable option in that tool. It's hard for me say how accurate it is, since garbage-in == garbage out, and I didn't follow it through to the end.

  7. whitenack | | #7

    Nate G, yes we have had this conversation before. You don't like the design, and that's fine, but it is what it is. There are reasons for all of our decisions, some of which are functional and some of which are personal preferences. I won't go into defending all of your critiques, as that isn't the point of this discussion. You didn't even answer the original question...I wasn't asking anything about non-ducted mini-splits vs ducted mini-splits. No hard feelings, I appreciate your opinions. In response to your question, no that is not an in-law suite, it is a mud room. This room connects to an attached garage (not pictured).

    Dana, thanks again. I agree with garbage in = garbage out. The cool calc tool didn't have my wall construction (2x6 with cellulose and 3" of polyiso), and didn't go up to R60 attic insulation, so I maxed it at R56 and figured that was close enough and if anything my figures would be a little high. The biggest variable I couldn't figure out what to use was the Effective Leak Area. I had no idea what to put in and this had a huge effect on the end results.

  8. iLikeDirt | | #8

    To each his own! I understand the grumbling about size is kind of a curmudgeonly thing, and I know the reason why you clustered the rooms with plumbing close together, but I do hope you think about the lighting situation a bit more. Windows on two sides of a room--especially on perpendicular walls--makes it a much nicer room. I worry that the office will be a really depressing room with only one window and a large flat wall, just begging for a window to be in it.

  9. Expert Member
    Dana Dorsett | | #9

    Nate: Please don't construe my observations as "...grumbling about size...", it's only meant to possibly explain the magnitude of the calculated heat load for a presumably high-R house.

    Clay: A 2x6/R20 + R18 polyiso wall has a U-factor of about U0.028 - U0.030 BTU per hour per square foot per degree-F delta (depending on actual framing fraction) which is well below any "standard" construction. The lowest U-factor wall in Cool Calc is U0.049, so that's a significant error. If you're using advanced framing throughout, use U0.028. If it's 16" o.c. standard framing, use U0.030. I believe the tool allows you to just enter a U-factor rather than the canned stackup selections (?).

    The difference in load between an R56 roof and an R60 roof is "in the noise", within the likely construction differences between modeled and actual framing fractions, not a huge error.

    I didn't dig into the specifics on air infiltration in that tool, but if you're religious about air sealing, assume 25 square inches for the " Effective Leak Area" (reality will likely be smaller.)

  10. whitenack | | #10

    Dana, Thanks for this info! I was assuming the differences in R-value for the walls was also inside the "noise" factor. This is great news. Cool Calc says I can change the U-factor of the walls, but then it doesn't let me edit that box. Not sure what the deal is there. I'll play around with it some more and see if I can find out how to change it. I guess I could also see if I could switch the construction type to ICF or something to see if I can get to that U0.030 level.

    I'm not sure what I plugged in for ELA. At the time I remember trying a couple of different values and it makes a huge difference in the total load. I think I used 50. That makes a huge difference.

  11. Expert Member
    Dana Dorsett | | #11

    The static U-factor of ICF construction isn't necessarily the best indication of what they will do for reducing peak loads. I'm not sure if Cool Calc correctly factors in the dynamic thermal mass issues with ICF, which adds a bit of complexity (they might- some load calculation tools do an OK with that, others don't.) The thermal delay of the internal mass reduces but broadens the peak.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |