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

Single minisplit for AC?

jeffesonm | Posted in Mechanicals on

Looking for feedback on adding AC to a ~1600 sq ft ranch in central NJ. Current setup is a window unit in the bedroom and a 1980s 220v beast of a wall unit in the dining room. Together they do all right in those worst weeks but for the most part we’ve just been toughing it out. I’m on a well shaded lot and the house stays reasonably cool during the summer, but also a bit damp. So hoping to dehumidify with the AC and of course cool the place down during the hottest weeks of the year.

1960s construction, plaster walls with foil-faced fiberglass batts of indeterminate insulation value. Similar on the ceiling, plus R-30 insulation unfaced batts I added on top, running perpendicular to joists. Original single pane double hung aluminum frame windows, with storm windows. Solid doors. Garage and bedroom ends of ranch are gable ends with no overhang, front and back of house have 18″ or so roof overhang. Front of house (kitchen in front) faces roughly NE.

Heat is primarily from a Jotul F55 wood stove, supplemented by hot water baseboard and Burnham MPO-IQ65 oil boiler when it gets under 20 or we’re not home to feed the wood stove. Not really looking to heat with the mini split but might be a nice option to have during shoulder seasons.

I did a room by room heat loss calc using and came up with 53k btu heating and 21.5k btu cooling. Floor plan, square footage and cooling btu numbers are attached. I did the one bath together with the adjacent hall and the other together with adjacent den just to keep things simple. Also of note is the high living room btus due to 80% of that wall being glass.

That turquoise box on the drawing is where a 2’x2′ whole-house fan is located. I was thinking a simple solution would be to replace it with a ceiling cassette mini spit in the neighborhood of 18k btu/hr. Seems like it would possibly be a little undersized, but only on the worst of days. Given how well we’re getting by now, maybe close the doors to the unused bedrooms and it’d be fine. My better half and I occupy the home and I expect it to remain that way for the foreseeable future. I understand this would require open bedroom doors for air circulation but that’s how we do it with the wood stove and it works fine for us.

I have a full basement so plenty of room down there to run ducts, but I’d really rather not. Don’t want to cut a bunch of holes in my beautiful wood floors or lovely plaster walls, and I hate banging my head on basement ductwork. 4:12 roof so not much room up there. I did consider a ducted mini split centrally located in the attic, with short runs to the bedrooms. Not sure it’s worth the hassle of ductwork for marginally better air circulation, and then I’d have to figure out returns, ducts in uninsulated space, etc. Again I’m designing for the two of us and it’s predominantly the kitchen, living room and bedroom 1.

Thoughts about this plan, or other options I should consider? Thanks in advance.

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

    I don't see a downside. Try it and see if it works.

    You can always add another minisplit in the dining room if you decide you need to.

  2. Expert Member
    Dana Dorsett | | #2

    As a general rule coolcalc overestimates loads, often by 35%. A cooling load of 21.5K for a 1600' house is about a ton per 1000', which isn't unheard of, but somewhat bigger than average. A heating load of 53K for a 1600' is 33 BTU/hr per square foot, which is about twice what I would expect for a reasonably tight 2x4/R11 house with single-panes + clear glass storms. When using that tool be aggressive on all assumptions, particularly on air-tightness.

    Since humidity is a primary issue, be sure to look at the minimum modulated cooling output as well as the max capacity- the lower the better, since that will give you the longest cycle times.

    eg, The SLZ-KA15NA & SUZ-KA15NA combination can throttle back to 3800 BTU/hr in cooling mode whereas the Fujitsu AUU18RLF + AOU18RLFC can back off down to 3100 BTU/hr, about 18% lower. That may not seem like much, but it makes a difference.

    1. Jon_R | | #15

      > Since humidity is a primary issue, be sure to look at the minimum modulated cooling output - the lower the better

      In this case, it's worse. Why - because, for example, the AUU18RLF fan varies only from 400 to 241 CFM. Which means that at 3100 BTU/hr, it's blowing 933+ CFM/ton. Which means NO latent removal (at this load), which is bad. Run a dehumidifier to control humidity and your net efficiency numbers are worse than if the AC cycled. Plus it probably has distribution issues.

    2. Jon_R | | #16

      > When using that tool be aggressive on all assumptions, particularly on air-tightness.

      Except don't be aggressive on air-tightness when it comes to latent load. This will cause an incorrect SHR and lead you astray when it comes to solving your dampness problem.

  3. Expert Member
    Dana Dorsett | | #3

    With a 1.25-1.5 ton ceiling cassette and through-wall unit in the dining room you probably have more than sufficient total cooling capacity.

    The beastly wall unit in the dining room is probably ridiculously or even ludicrously oversized for the combined cooling load of the dining room + den/bath area, which leads to lousier latent load control. (What are it's nameplate BTU/hr and EER specs?) Since it's 30 years old it probably no longer hits even it's original EER numbers. The load numbers for the dining/den area don't really add up to a mini-split worthy load, since the minimum modulation of most mini-splits will be above the cooling load of the dining room, so it won't really modulate with load (much).

    But replacing the past-it's-prime cooling-only wall unit with a half-ton point terminal heat pump (PTHP) or more appropriately sized half ton through-wall air conditioner might be "worth it". (A half-ton PTHP runs about $200 more than a half-ton AC-only unit. $700-800 vs $500-600) If you get one that has a scroll or rotary compressor it will be a lot quieter than what's there, and by being closer to right-sized for the loads the duty cycle will be higher, with better humidity control. Something like the Amana PBH073G35CB (high wall mount) or PTH073G25AXXX (better as low-mount) would do it, and would probably more than cover the heat load for that zone in heat pump mode too. There are others.

  4. jeffesonm | | #4

    Thanks for the feedback. I was eyeing up that Mitsu but will check out the Fujitsu as well. I will have to dig around for the manual for the beastly wall unit, I don't see the BTUs anywhere visible. Probably give the mini split a go and replace the wall unit down the line, if it's still needed.

  5. Expert Member
    Dana Dorsett | | #5

    The BTUs, EER and model numbers are usually on a plate inside the front grille where you can peek at it with a flashlight, or by the filter access door/slot, sometimes on the side of the unit (making it impossible to read when it's installed.)

    All else being equal, Fujitsu AUU18RLF / AOU18RLFC ceiling cassette is also about 20% more efficient than the Mitsubishi SLZ-KA15NA / SUZ-KA15NA . (SEER 20 vs. SEER 16, HSPF 11.5 vs. HSPF 9.6). That's enough to make measurable difference in the power bills. It's AHRI rated capacity at +17F is literally twice that of the Mitsubishi too- 21,600 BTU/hr vs. 10,500 BTU/hr. Even if it's a bit more expensive up front, it's probably going to be worth the up-charge to go with the 1.5 ton Fujitsu than the 1.25 ton Mitsubishi for the extra capacity alone.

  6. Lita_ | | #6

    Dana, thanks for the great info comparing ceiling cassettes between Fujitsu and Mitsubishi, its just what I was looking for. Where did you find the capacity at +17F for the Fujitsu AUU18RLF / AOU18RLFC ceiling cassette, please? Only the 47F min-max capacities are listed on the submittal sheet I found.

    Do you have any info comparing the Daikin ceiling cassettes to these two brands?

  7. Expert Member
    Dana Dorsett | | #7

    The ARHI "Nominal heating" capacity of 21,600 BTU/hr is the modulation level at which it was tested at +47F for efficiency, not necessarily it's maximum. But to run the test it has to be able to deliver at LEAST that much at +17F.

    If it's consistent with some other Fujitsu cassettes compatible with the AOU18RLFC compressor it can probably still deliver the full 21,600 BTU/hr even at +5F, and something close to 23,000 BTU/hr @ +17F when running at it's maximum speed.

  8. jeffesonm | | #8

    Just over a year later, I am finally ready to pull the trigger on the ceiling cassette mini split. Before I order the Fujitsu combo mentioned above, any other more recent market entries I should consider?

    1. Expert Member
      Dana Dorsett | | #10

      My understanding is that any minute now Mitsubishi is going to be releasing a series of "hyper heating " cold-climate single zone versions of their MLZ ceiling cassettes. They are compatible with their cold-climate multi-zone compressors, but right now the only single zone versions are using non-hyper heating SUZ series compressors, so the heating capacity craps out pretty fast at temps below freezing. Even though they're pretty efficient, unlike the Fujitsu AUU18RLF, it's not going to provide as much heat as you would like when it hits single digits F outside.

      I would expect a hyper-heating version of the -KP18 to have capacity at +17F comparable to the Fujitsu AUU18RLF. When married to one of the MXZ hyper heating compressors it's good for 21,000 BTU/hr at any outdoor temperature, as long as the MXZ has that much capacity.

      The advantage to the MLZ cassettes is that they can be installed between 16" o.c. joist framing, whereas the square AUU18RLF ceiling cassette will in most cases requires modifying existing framing to get it in. Read the installation instructions and measure your whole-house fan cut out very carefully to be sure it fits where the whole-house fan once stood. If I'm reading the diagrams on page 7 correctly it takes an absolute minimum of 22-7/16" of clearance between the framing. With 24" o.c. joists there is a nominal 22.5" which is a 1/16" squeaker fit (but if it's tight you can always make it fit with grease and a big hammer. :-) ) With the "...2’x2′ whole-house fan..." it could make it, but it might not.

      Given your floor plan and heat load, if it fits the AUU18RLF is clearly the better choice, since it blows in four directions, whereas the KP18 blows to one side. If going with the KP18, mounting it to blow primarily toward the Living Room makes the most sense.

      1. jeffesonm | | #11

        Thanks for that info.

        No worries on space, that hallway install would be replacing a 1950s whole house fan so I've got a solid ~25"x25" to work with.

        Not super concerned with the heating either. There's a wood stove about 6' from the future AC position and it does a good job heating the entire house until it gets down to 20s.
        Below that I run the oil boiler for ~1/2 hr in the mornings just to take the chill off and then the wood stove keeps up throughout the day. I do expect I might use the mini split a bit in the shoulder seasons when it's not quite cold enough to justify keeping the wood stove going.

        While shopping for the AUU18RLF I came across some LG setups like a 7+7+7+12 four ceiling cassettes and one compressor for not that much more money (seems like the combo's are way better $/btu value than a single 18k?) It would be more install work, but allow for better control if I had units in each bedroom, especially for sleeping. The other downside I see would be minimum cooling for that ~40k compressor was like 12k, so even at night time when I need just one bedroom cool, it would be basically just wasting electricity, if I understand how that works? And then 40k overall seems very high, and I wonder if a 7k unit would just cycle all the time for a single bedroom? So maybe just stick with the single Fujitsu?

        1. Expert Member
          Dana Dorsett | | #12

          >"While shopping for the AUU18RLF I came across some LG setups like a 7+7+7+12 four ceiling cassettes and one compressor for not that much more money (seems like the combo's are way better $/btu value than a single 18k?) It would be more install work, but allow for better control if I had units in each bedroom, especially for sleeping."

          The efficiency tends to take a serious dive due to gross oversizing when going with the "cassette per bedroom" approach, and can even lead to LOWER comfort, since even when nominally "OFF" there is still refrigerant flowing through each cassette whenever any one zone is active. If a 12K is off the living room and right-sized for the load it'll be running all the time, and if the 7K in the bedroom is 3x oversized for it's room load it can sometimes overcool/overheat the space even without the blower running.

          For low-load rooms it's usually better for both efficiency and comfort to combine a few of them into a single zone and use an appropriately sized ducted cassette.

          To get a better handle on what makes sense, run separate Manual-J calculations on each bedroom. It's conceivable that a 1.5- 2 ton multi-split with ducted zone for the bedrooms and bath could work, running the ducts in the basement. Fujitsu's ARU7RLF is a decent 7K mini-duct cassette compatible with their 1.5 ton AOU18RLXFZ and AOU24RLXFZH multi-splits, if the combined loads are that low. Almost all vendors have 9K (and up) mini-ducted units.

          Online freebie Manual-J tools such as or deliver reasonable numbers if you are sufficiently aggressive on R-value and air tightness assumptions. (With those tools it's usually better to assume the house and ducts are completely air tight to avoid oversizing.) It's hard to get the shading factors dialed in without a better tool though, which are more important for cooling than for heating. If your house gets much shade from trees/hills/buildings it's going to over-state reality.

          1. jeffesonm | | #13

            I pulled up my room-by-room cool calc from last year and double checked everything, only thing that needed adjustment was the 99% outdoor design conditions which I pulled from here: Cooling 99% went from 76 to 89 which moved overall BTUs to ~24k.

            The lot is fairly wooded and the house does get quite a bit of shade so by the sounds of it the software estimates will be higher than reality (~35% stated above?)

            Based on these revised estimates, it seems like a 7 concealed duct (ARU7RLF) + 12 ceiling cassette (AUU12RLF) and 18k compressor (AOU18RLXFZ) would match the load nicely and be pretty close to maxed out on the hottest days, which I understand to be the goal.

            Overall this seems like a pretty plan, and worth the extra $ vs just the one ceiling cassette. It also begs the question, why not do another concealed duct unit instead of the ceiling cassette? I could place it more centrally in the [non-bedroom] living area and maybe even get that cold into the den/dining rooms. More ducting required, but otherwise install would be easier working in an unfinished basement vs crawling around a 3/12 attic in the middle of summer.

            Any good resources offhand you can point me to regarding duct sizing, lengths, considerations etc for concealed duct mini splits?

            (also Dana, as an aside, I really appreciate all your help)

          2. jeffesonm | | #14

            For others who might find this thread, some good ducting design info here:

            and here:

  9. kjmass1 | | #9

    I'd recommend a wired controller placed somewhere centrally if you decide to put a cassette in that hallway space. The hallway will cool quickly and may turn off and not cool the living room well. A thermostat on the stairway wall in the LR may help.

  10. Jon_R | | #17

    > So hoping to dehumidify with the AC

    You are on-course for continuing humidity problems.

    You need to re-evaluate what your SHR need is (.93 that you posted is Albuquerque high, .63 is far more likely), how it increases significantly at lower loads and what equipment will provide you with a low enough SHR (and resulting comfortable humidity) at all of your loads. Or plan on the efficiency hit (and distribution problems if you close interior doors) of additional dehumidifier use .

    1. Expert Member
      Dana Dorsett | | #18

      >"Or plan on the efficiency hit of additional dehumidifier use."

      The better/lesser efficiency hit is to occasionally take advantage of Fujitsu's "DRY" mode during low sensible load conditions. A dehumidifier converts the latent load to a sensible load. The DRY mode puts that latent heat outdoors where it belongs.

      NJ can be sticky, but nothing like the Gulf Coast states. Even with a shaded lot and low sensible loads DRY mode can take it down a lot. There still needs to be some sensible load, but the SHR becomes optimized for dehumidification.

      The only series that can dehumidify without sensible cooling is Daikin's Altherma mini-splits, but they only come in high-wall versions, not ceiling or mini-duct cassettes.

      1. Jon_R | | #19

        Sounds good, but is it? Some Fujitsu specs indicate that compressor output is severely limited in dry mode (eg a max of 22%, the opposite of what I'd expect for maximum latent removal). So you often can't get better drying and adequate cooling - I wouldn't want to get up in the middle of the night to switch modes.

        Can anyone say what the SHR vs output will be in dry mode? I don't advise the "hope+bandaid" method of HVAC design.

        Fujitsu isn't helpful - they write (about one of their model's dry mode) "Allows for moisture removal in low load conditions of up to 50% of rated moisture removal capacity." Vague (to the point of meaningless) and not adequate when you need more than 100% of rated moisture removal.

        Some old, other model data here in Table 7 (low load SHRs are terrible).

        1. kjmass1 | | #37

          A real world example: I have a 12k + 7k Fujitsu hooked up to a 24k multi on my second floor. At 72F indoor and 90s outside with high humidity, I was getting about 1 drip/second on the drain line.

          I have another Fujitsu 24k multi for other parts of the house so I put the first floor 12k on dry mode today as well, I was getting 3-4 drips/second just running that one by itself, with lower loads being on the first floor.

          We are going to get wicked heat this week I may try running them all in dry and see how it works. RH in home all day was 52% and comfortable.

          1. jeffesonm | | #38

            Thanks for the real world feedback... it's easy to get lost in the design theory of all this so nice to hear.

            I also found this report "Phased Retrofits in Existing Homes..." which includes one example where they installed a single ARU12RLF in a 900 sq ft FL home, and it did a good job cooling and removing humidity:

    2. jeffesonm | | #21

      Okay SHR is new to me, one more thing to tirelessly research.

      Some other GBA thread:

      DOE report "Evaluating Moisture Control of Variable-Capacity Heat Pumps..."

      "Assessment of Energy Efficient Methods of Indoor Humidity Control"

      If I'm understanding your concern, it's that the high SHR of the mini splits means that it won't dehumidify the house sufficiently, especially at low loads?

      What are the possible solutions? Do other units offer a significantly lower SHR? Oversize the unit and run at partial loads? Whole house dehumidifier?

      1. Expert Member
        Dana Dorsett | | #22

        >"If I'm understanding your concern, it's that the high SHR of the mini splits means that it won't dehumidify the house sufficiently, especially at low loads?"

        That is indeed the concern, but that's what DRY mode is designed to take care of, and by all reports Fujitsu's version of it works pretty well. It won't be necessary to run it in DRY mode all the time, but even though it runs a lower SEER when in DRY mode it's still more efficient than running a dehumidifier if there is any sensible load AT ALL.

        I doubt your shoulder season sensible loads are any lower than many of John Semmelhack's PassiveHouse & Net Zero houses comfortably cooled & heated with Fujitsu xxRLFCDs. See his comment regarding dehumidification in the fifth paragraph of response #4:

        1. kjmass1 | | #23

          I like to use my 2nd floor unit in cool mode and my first floor unit in dry mode as needed. It always runs on low fan and will knock the humidity from 58% to 50% in quick order on rainy, cooler days like today. I'll use the 2nd floor unit in dry mode more in the shoulder season as temps get cooler.

      2. Jon_R | | #24

        As the links show, there are large differences in humidity control with different units (the Unico did quite well, worth a look). Unfortunately, manufacturers don't provide the specs needed to make good comparisons.

        Other than AC equipment choice, a portable dehumidifier plus open interior doors may be your most viable solution. It will only run when the AC equipment fails to properly control humidity.

        Maybe someday manufacturers will get their act together and provide full independent control of sensible and latent output over the load entire range - without very inconvenient manual mode changes. A dehumidifier shouldn't run until the AC coil approaches freezing (or sensible load is zero).

        1. Expert Member
          Dana Dorsett | | #25

          >"Maybe someday manufacturers will get their act together and provide full independent control of sensible and latent output over the load entire range - without very inconvenient manual mode changes."

          Such as Daikin's Quaternity series mini-splits, with independently settable humidity and temperature setpoints, and the ability to dehumidify (almost) without sensible cooling?

          While the Quaternity also has separate COOLING, DRY, and DRY COOLING modes, in most cases you wouldn't have to switch it up. The difference in those modes is briefly explained on p49 of the manual:

          1. Jon_R | | #31

            > Such as Daikin's Quaternity series mini-splits

            A limited compressor range and multiple heat exchangers to alter cfm/ton sounds good (and expensive), but it's another "who knows". As far as I can tell, they don't publish low load humidity control specs either. Using their capacity tables for an example, I get a terrible SHR = .97.

  11. Expert Member
    Dana Dorsett | | #20

    ... and yet people in humid parts of east Asia and Australia are doing OK with it.

    I don't think it's possible to armchair-engineer it's latent load handling abilities simply by reading the the manual.

    Whatever the SHR in any cooling mode, it will provide better dehumidification that it's getting now, and if some use of a dehumidifier is needed on the worst days (not likely, but maybe) so be it. A whole house dehumidifier would be like swatting flies with a sledge hammer at NJ style shoulder season latent loads.

    It would be silly to try to dry your basement (which in NJ has a negative sensible load most of the time), but for the above grade conditioned space a maximum compressor output of 22% would cover both latent and sensible loads during the shoulder seasons, and much of the day even on design day. With Fujitsus you DO need a sensible load (or drop the set point) to do much drying, since it will still follow the sensible temperature setpoint. Some other mini-splits will run forever when it DRY or DEHUMIDIFY mode, over cooling the place when the sensible loads are low.

    John Semmelhack designs around xxRLFCD systems on a regular basis (hundreds of them in service) primarily for low load/high performance homes many of them located in (even more humid than NJ) Virginia, where his business is located. His take on their latent load handling in response #4 in this thread is:

    "Dehumidification is reasonable in cooling mode, quite good in dry mode....but watch out for duct and register condensation in dry mode, since the supply air temperature will be quite cold."

    How well it does with an ARU-xxRLF or AUUxxRLF married to a multi-split may differ in the particulars, but the engineers at Fujitsu clearly know how to make DRY mode work.

  12. joshdurston | | #26

    Dana, makes a good point. There are so many variables that aren't reflected in the submittal data. Mini splits behave differently than conventional split AC units in that they are capable of holding a suction pressure set point and super-heat value over a wider range of operating conditions. What this means, is the de-humidification performance is going to vary, and as the wetbulb goes up the the SHR ratio will improve (go down) quite a bit.
    I only see the SHR being an issue if you like to keep you space really cold. If you keep you space in the 60's a 64f wetbulb is going to feel like a cave, but in the mid 70's
    The FH09 seems to have a 64degF wetbulb inflection point, if the wetbulb is lower de-humidification rate goes down drastically. But has the humidity and wetbulb climbs above 64f the dehumidification improves drastically.
    For me that's fine since I only cool my space 75f to 78f, humidity would have to be above 55%rh to exceed a wetbulb of 64f. If my space was 69F the humidity might be over 70% which would be uncomfortable.
    Mini splits also have some secret sauce that isn't reflected in the performance data due to difficulties in testing systems with a lot of variables.

    Dry mode improves the dehum ratio, but may not provide the same pints/hr as cooling under a higher load. So if it's hot and humid use cooling mode, but cool and humid use dry, or potentially cool mode at a forced low fan speed. I don't think it makes sense to use a standalone dehumidifier until dry mode starts to overcool the space because of the sensible load associated with dehumidifiers.

    1. Jon_R | | #33

      Note that you are referring to a table for rated load. At lower loads, the numbers for most m-splits (not saying anything about that specific one) get far worse.

      Even ignoring that critical issue, say someone needs a SHR of .70 (common) and wants 77F/55% inside (also common). The table shows that that unit won't do it (SHR is a much too high .84). That leaves about 1/2 of the moisture removal to a dehumidifier (an issue IMO).

  13. jeffesonm | | #27

    Thanks all for the thoughtful and researched discussion.

    Just to give some perspective, we've been getting by with barely any air conditioning for the past 5 years we've owned the home. We run the bedroom window unit at night probably ~4 weeks/year, and the giant 1980s window unit in the dining room ~2 weeks/year. The rest of the summer we use the whole-house fan at night, close up the windows during the day, and wish it was less hot and humid inside (and outside.) Which is of course why I'm doing the mini split thing here.

    I'm just adding the context to give some sense of our goals/criteria. In the winter we use the wood stove to heat almost exclusively, so we deal with pretty drastic temp swings. We will definitely be 'cool it to 75-78' type of people and, if on the two days of the year it breaks 100 I can only get the house down to 80, I won't be sad. But if I can increase the overall summer comfort, provide some reasonable level of cooling and de humidification, not send my electric bill through the roof and not destroy my bank account, I will call this project an overwhelming success.

    That being said, I'm leaning towards the Fujitsu 18k multi split with concealed duct units indoors.

    Living area would be the 12k unit, installed in the basement right near the staircase that leads to the basement. Depending on space I would run the supply ducting up inside the staircase or build a chase inside the cabinet thingy in the living room that shares a back wall with the staircase. I could then have two registers come out just below the ceiling directing air into the living room, one into the dining room and one into the kitchen, utilizing the open space at the very top of the stairwell.

    The dining room and den were an addition and down below there's a block wall separating that part of the house. I'd rather not cut through it so using the stairwell area would let me get one register over to that side. I was thinking maybe one large central return at the other end of the living area. Those are just openings between LR/den and kitchen/dining, no doors, so hopefully air would circulate through those.

    Bedrooms would be served by the 7k unit (or 9k? it's actually cheaper...) installed in the basement. I would run supply ducting up through closets and registers into each bedroom just below the ceiling. For returns I would do a vent in each along the floor, and was thinking just near the end of the door swing area would be a good space unlikely to obstructed by furniture.

    The 7k/9k/12k units all list a max static pressure of 0.36wg which seems easily achievable on the 12k side given the extremely short runs, and reasonably achievable on the 7k/9k side. I would use rigid ducting on all, round if possible, rectangle if not. I am now reading all things about duct design and have one question in this regard.

    The 12k unit puts out 383 cfm on high which seems in line with the 400 cfm/ton sizing ratio I see thrown around. The 7k/9k units list 324/353 cfm respectively, which seems disproportionately large relative to the BTU's they need to move. Why is that, and what should I use for my duct sizing calculations? Column K in my spreadsheet divides the total cfm for each unit, allocated to each room based on its relative proportion of load for that zone.

    Also, any other general thoughts on this, newest of the plans?

    1. Jon_R | | #30

      > 12k unit puts out 383 cfm on high which seems in line with the 400 cfm/ton

      Which may or may not produce a low enough SHR at rated load (300 cfm/ton would be a better new standard). And the point of much of the above is that you should also care about cfm/ton at lower than design loads.

  14. NEplumber | | #28

    When choosing a multi head unit with dehumidification in mind, the biggest concern is the lowest modulation of the outdoor unit for cooling. If you dont hit the minimum modulation of the compressor, regardless what the CFM the indoor unit spans, you will not dehumidify.
    ie, some of the late model Mitsubishi 3 & 4 ton unit have a 28% minimum modulation while the 5 ton of the same style has a 38%. This is also why moisture removal and sensible heat ratio are not specified for multi head units.
    For multiple rooms you may want to consider a ducted unit as the smallest of indoor machines are generally too large for bedrooms.
    We are a licensed NJ contractor so if your interested in more info, please contact us through the web page at


    1. jeffesonm | | #40

      Thanks Moe, I am a dedicated DIY guy but maybe I'll call you guys to cleanup my mess :)

      J/K (hopefully)... a few years back I successfully replaced my boiler and re-zoned some absurdly long loops so hoping this foray into hvac is equally successful.

      1. NEplumber | | #43

        All for DIYers. We have a few clients that have us spec, they install, and we pull a vacuum on the system and adjust charge.

        Let us know how the install goes!

  15. Jon_R | | #29

    > If you dont hit the minimum modulation of the compressor, regardless what the CFM the indoor unit spans, you will not dehumidify.

    Of course you can always hit the minimum modulation of the compressor - by cycling. And this often produces better dehumidification than designs that modulate the compressor down lower without a corresponding decrease in interior CFM/ton.

    1. NEplumber | | #32

      Incorrect, and I have downsized many VRF com pressures (specifically city multi units) due to this issue.
      If the compressor cycles, that means it's off. if it's off, the coil isnt cold. If the coil isnt cold, theres no moisture removal.

      For dehumidification purposes you want a system that is always dehumifying or you end up like my clients, with a 70 degree home and 65%+ humidity.

      1. Jon_R | | #34

        Note "designs", referring to various equipment. My statements are correct. Some equipment will produce better latent removal when modulating vs cycling - some will do worse.

        > If the coil isnt cold, theres no moisture removal.

        And no sensible heat removal either. Which soon causes it to switch on - and then there is both.

      2. joshdurston | | #35

        Having a lower min modulation on the compressor doesn't dehumidify any more if the fan doesn't go low enough.
        I have a mitsubish FH09 and wish the min cooling mod was higher or adjustable, min CFM is about 130 and min cooling is about 1500btu/hr. Under low loads I find I often have to crank it down temporarily to dehumidify, and then turn it back up to avoid overcooling. If I let it do it's thing under lower loads it maintains setpoint but I end up with 63%rh.
        If I was in Arizona I probably wouldn't care, but in Southern Ontario I need dehumidification. Under higher modulations it will remove a lot of moisture (I clocked 1 pint in 30minutes yesterday evening), but a lower modulation it can run all afternoon and barely pull out a 1/8th of a pint. I actually kinda wish I would've went with the FH12 instead of FH09 due to similar airflow ratings but higher min BTU for better dehum.

        1. Jon_R | | #36


          > wish I would've went with the FH12 instead of FH09...for better dehum

          And yet for some reason people think over-sizing reduces dehumidification. And that "try it" is a valid design method.

          > I often have to crank it down temporarily

          Ie, increase latent removal by forced cycling at higher output. An inconvenience that manufacturers should fix. And a nice example of my point in #29.

  16. jeffesonm | | #39

    I understand the concerns but don't see a lot of great options to mitigate them:

    1) Get manufacturers to produce units with an SHR more conducive to humidity control at low loads - Yes that would be great, I don't see it happening in the next two weeks
    2) Oversize units so they cycle and run lots of cooling at low airflow so more humidity is removed in each cycle - I don't like the idea of oversizing everything, seems exactly contrary to doing a manual J, getting a high efficiency, modulating system etc.
    3) Use cooling equipment to cool the house and get a separate dehumidifier to remove humidity - This makes sense to me, but I don't really want to buy another $1k unit, more ducting, and more ongoing energy costs.
    4) Undersize units so they have to run closer to full load and therefore better remove humidity - This is what I'm going with. At higher loads it should work as designed, and at lower loads I'll use the Dry feature and hope for the best.

    Jon R - I understand 'try it' isn't a great approach, which is why I'm here counting BTUs, calculating friction rate and so forth. At some point I need to buy and install some equipment, and see what happens, and I've reached that point.

    If the humidity continues to be a problem after implementing option 4, I can always supplement with option 3.

    1. Expert Member
      Akos | | #41

      Adjusting air flow with mini splits is not that hard (see attached). With a ducted unit it is even easier with registers. All modern split units have a temperature sensor in the coil and will not freeze up.

      Just adjust the air flow on those low load days until you get the right amount of dehumidification.

      1. joshdurston | | #44

        Just tried this and it actually worked pretty well, I was able to get the discharge temp down from about 14.5degC to 11.5degC (fan on lowest setting) by blocking the inlet with some magazines.

        Bit of hack, but works well if you're willing to fiddle with things. Most the time leaving the fan speed fixed to the lowest setting gives adequate dehum capacity without blocking the air flow. Last week was abnormal though with outdoor dewpoints in the mid 70F range.

        I run a portable dehum in my basement (set to 55%rh), it keeps my house from getting too humid.

        More incentive to tighten things up and properly air seal my recessed lighting.

  17. Jon_R | | #42

    In terms of low cost practical advice, consider installing dampers on every register/branch to give an additional manual control over cfm/ton (and balance).

    I expect you will install a portable dehumidifier and with open interior doors, it will provide good comfort (although with some additional energy use over AC with better latent removal).

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |