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

Slim duct mini-split concerns

gm155 | Posted in Mechanicals on

Hello, I’ve been getting all sorts of conflicting advice about using slim ducts. House is new construction in Massachusetts. second floor has three smallish bedrooms, about 600 sq ft in total. Heat and cooling load calcs support using an 18k slim to hit the three bedrooms, which aI have a contract with a firm to install. However, a couple of other firms have stated adamantly that the cfm rating of the unit is too low and we won’t be happy. Thoughts?

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

    On a load per square foot basis an 18K unit seems ~2x or more oversized even for a code-min house. Is it the heating numbers, or is it the cooling load that is driving the sizing?

  2. GBA Editor
    Martin Holladay | | #2

    I'm not familiar with "slim ducts." There is a brand of PVC covers to hide linesets called "Slimduct" -- but I guess that's not it.

    It looks like you are referring to a fan-coil unit from Fujitsu called the Halcyon Slim Duct. Did I guess correctly?

    -- Martin Holladay

  3. gm155 | | #3

    I'm not sure which is driving the sizing, but the push back I've gotten is due to the CFM rating on the slim duct. The concern being that the unit wouldn't perform adequately.

  4. gm155 | | #4

    Martin, yes, exactly. We were hoping to avoid having a wall unit in each bedroom by using an 18k Fujistu slim duct in the attic to hit the three bedrooms on the second floor.

  5. user-626934 | | #5

    Many installers shy away from these units because the available fan power is relatively small and requires a more careful duct design compared to a "traditional" residential air handler/furnace. I've designed and tested the Fujitsu version on houses/zones up to about 2,000sf, so 3 bedrooms and 600sf is certainly within the realm of reason with a good duct design. I would encourage you to ask for a guarantee from the installer that the air handler, as installed, will move the design airflow within a certain tolerance (+/-15% is ACCA and Energy Star tolerance...I like it closer to +/-5%).

    And, I certainly agree with 18k unit for even a code-minimum 600sf in MA sounds very high.

    Also, I'm assuming (based on MA location) that the attic is sealed and insulated at the roofline, right?

  6. gm155 | | #6

    Second floor plans attached. Theres actually about 830 sf on the second floor in total. The bedroom sq footage is smaller. I'm assuming the heat load is what led to the 18k unit being specified. This is almost a code-min house. We're spraying open cell for most of it.

    I'd like to go with the ducted mini-split for the three bedrooms but I'll go with three wall units if the ducted unit won't perform well. John, you mentioned you've used them up to 2000 sf? Could this have been one unit? We had wanted to use one 18k unit for the three bedrooms. Any thoughts on whether or not the CFM will be adequate? I'm open to switching to three wall units , but would rather not.

  7. Expert Member
    Dana Dorsett | | #7

    "I'm assuming the heat load is what led to the 18k unit being specified."

    Don't assume- get a hard copy print out of the load calculations. If they were done by HVAC contractors, (sadly) at least 9 times out of 10 the calculations are complete junk.

    The 18RLFCD puts out about 20,000 BTU/hr into a 70F room at typical MA 99% outside design temps (in the 0F - 13F range.) At +17F it delivers 21,600 BTU/hr at full speed per the submittal sheet:

    Assuming the 20K, number, for 600' of conditioned space that becomes (20,000/600=) 33 BTU/hr per square foot, which is on the order of ~3x what I would expect from a pretty tight but code minimum-R house in MA. The floor plan doesn't show a crazy exterior shape with lots of corners, nor does it show a ridiculous window/floor ratio, either of which would elevate the load numbers. Without taking a WAG on the construction & U-factors I'd expect it to come in between 9000-13000 BTU/hr @ 0F outside for heating the entire 830', less than 10,000 BTU/hr @ 0F for the 600 square feet.

    The cfm of these units are fine, assuming the contractor designs ducts rather than hacking them in by crude rules of thumb. But the oversize factor is insane. It's worth spending the $500-1000 for a AGGRESSIVE Manual-J load calculation from a qualified third party, not a HVAC contractor, not some DIY web tool, or you'll end up buying 2-3x too much equipment that will provide less comfort & efficiency. A certified Professional Engineer, or a RESNET rater should be able to perform the calculations, and provide a detailed room by room loads, complete with the U-factors and ventilation/infiltration numbers used. Tell them up front you are expecting them to be aggressive rather than conservative in all their assumptions, (as directed in the Manual-J documents!!)

    With a proper Manual-J it's pretty likely that the load will be within range of the 9RFLCD, or worst case, the 12RLFCD:

    ^^^^^12,000 / 600= 20 BTU/hr per square foot heating, WELL above typical code min ratios, cooling capacity might be close (probably isn't but run the numbers.)

    ^^^Even more oversized for the likely heating load, could be overkill for the cooling load too, but probably not 3x oversized.

    Run the real numbers.

  8. Anon3 | | #8

    spending the $500-1000 for a load calc???

    A Midea 20 SEER 30,000 btu mini split costs $1300 with free shipping.

  9. Expert Member
    Dana Dorsett | | #9

    And why would you throw such a ridiculously oversized piece of kit at this? (@#8)

    Right sizing it is buying comfort & efficiency, and you don't get there without credible load calculations.

  10. gm155 | | #10

    I do want to get an independent manual j calculation done. If you can recommend someone, please do, thanks.

  11. Expert Member
    Dana Dorsett | | #11

    Got a ZIP code?

  12. gm155 | | #12

    02631 - Brewster, MA

  13. Expert Member
    Dana Dorsett | | #13

    And outfit in Falmouth did a reasonable load calcfon a project I was involved with in MV getting on two years ago. I had no direct contact them, but used their room-by-room calc to spec the mini-splits, and as luck would have it, a copy of the Manual-J is still on this computer! IIRC they're not dirt cheap, but they're not gold-plated either. (Some of the contractors on the Cape & Islands clearly have yacht payments to make! :-) )

    Contact info on the Manual-J PDF is:

    Heyoka Solutions, LLC
    19 Howes Lane
    Falmouth, MA 02540

    Phone: 866-389-8578
    Fax: 508-444-8737

    [email protected]

  14. Anon3 | | #14

    Midea Premier Hyper 9,000 BTU 27.7 SEER Ductless Mini Split Heat Pump costs under a thousand with free shipping.

  15. GBA Editor
    Martin Holladay | | #15

    Paul Raymer of Hekoya Solutions is a smart guy. I first interviewed him many years ago, when he used to work at Tamarack Technologies.

    -- Martin Holladay

  16. Expert Member
    Dana Dorsett | | #16

    The minimum heating capacity of the MCHSU-09PHH2 & MEHSU-09CHD2 3/4 ton ducted Midea is 4500 BTU/hr, over half the likely heat load, and thus won't modulate NEARLY as much of the time as a Fujitsu 9RLFCD slim duct unit, which modulates down to 3100 BTU/hr.

    In heating dominated Brewster MA the SEER is somewhat irrelevant, due to the very low annual cooling degree days. (No matter which way the wind blows in Brewster it's across low-temperature water, and the sensible cooling hours are very modest.) With an HSPF of 10.5 the Midea mini-duct unit is no match for the Fujitsu's HSPF 12.2, especially, and with the high minimum modulation relative to the load it will be cycling most of the time, and unlikely to even hit it's already lower test numbers. And it's tested SEER is only 19, not 27.7 as alleged in response #14:

    Local distributor/installer support would also be an issue in this location. At any price point it's risky to use as a primary heating system without local product support.

  17. Anon3 | | #17

    Anyone that spends $1000 for a calculation when the equipment costs $1000 is probably the same kind of people that spends $10,000 for HVAC instead of $2000.

  18. josh_in_mn | | #18

    Anyone who installs equipment without knowing that it'll work correctly and efficiently is probably wasting money and time. There's more to this equation than just the cost of the equipment. After all, a roaring bonfire in the middle of the room costs practically nothing and will heat the room just fine.

  19. Expert Member
    Dana Dorsett | | #19

    The installed cost of a ~$1000 ducted Midea will be WELL over $2000 in Massachusetts- count on it! I'd be stunned (and possibly worried about the duct design) if it came in under $3K.

    It's also highly unlikely that a 3/4 ton ducted Fujistu would come in anywhere near $10K, but it might if one opted for one of the gold-plated contractors.

    At MA residential retail electricity rates the $600-700 difference in hardware cost between the 3/4 ton ducted Midea and 3/4 ton ducted Fujitsu will be recovered in the higher operating efficiency well before the Fujitsu warranty is up. Even at the tested HSPF the Fujitsu delivers 16% more heat per kwh, and the fact that the minimum modulated output of the Midea is 50% higher than the Fujitsu and more than half the likely design heat load means it won't make it's numbers. Rather than 16% higher operating cost more like a 22-25% increased operating cost, maybe more, depending on where the load numbers actually come in. The Midea has less than a 2.5:1 turn down ratio (compared to 6:1 for the Fujitsu), which means a fairly narrow band of load size where it can actually hit it's numbers.

    Proper load calculations should be performed in ANY new construction house prior to specifying the HVAC equipment. It's required by law in California (often ignored). It isn't required in MA, but it's still a good idea, especially when installing modulating equipment such as condensing boilers or mini-splits. Unfortunately most modulating equipment installed in MA seems to be oversized to the point of never or rarely modulating (as would absolutely have been the case with the original 30,000 BTU Midea suggestion). Sure, oversizing the hell out of it still heat & cools the place, but it's the opposite of comfort, and it takes a toll on efficiency.

  20. gm155 | | #20

    So, maybe there is special hvac contractor yacht payment pricing on the cape, because I've been getting quotes of 7-8k before rebates for the Fujistu 18RLFCD, with ducts run to three bedrooms. I did ask the contractor I'm working with what his load calcs came to and he replied:

    "For the second floor unit I came up with 13,576 BTU's for heating and .7 tons for cooling. In my view the 18 is still the best pick as the 12 would be short on heating capacity"

    Thoughts? Also, I emailed Paul Raymer a couple of days ago. No response yet.

  21. user-626934 | | #21

    It doesn't sound like your contractor has reviewed the Fujitsu design manual thoroughly. The 12k model can meet those (questionable) loads with room to spare.

    The 12k model running at the rated 383cfm has a heat output of over 17,000Btu/hr at your 99% design temperature (15.6F)...15,000Btu/hr at -5F.

    On top of that, 383cfm is probably a little on the low side. Unless there were moisture control problems in the summertime, I would normally run the airflow on these a little higher...more like 425cfm, which gives nearly a 10% boost in capacity over the engineering data.

  22. user-6811621 | | #22

    @ Dana,

    It does seem the 18RLFCD is overkill, but the 9RLFCD and 12RLFCD also have the same 3100 Btu/hr min. turndown ratio. If turndown ratios are the same, are there other performance penalties in oversizing other than slightly reduced HSPF and SEER ratings? Just trying to learn here..

  23. Expert Member
    Dana Dorsett | | #23

    $7-8K all -in (including ducts) isn't at all ridiculous for the 1.5 tonner. At $8K it may be a bit high, but still nowhere near $10K. In competitive bidding you might get it down to the $5-6K range, but there are so few installers of ducted mini-splits it's hard to get multiple bids. My "...if it came in under $3K" comment would be even for a cheap 3/4 ton Midea hacked in with flex duct using undocumented uncertified labor, and an "office" that consists of the cell phone in the pickup truck.

    The 18RLFCD alone runs ~$2230 at internet pricing, not counting any of the mounting brackets/pads, refrigerant lines and other ancillary hardware, and not counting the ducts. Call it $2500-2800 in direct material cost to the contractor, and the installed price would be on the order of 2x that, sometimes more with a complicated install. The 9RLFCD is ~$600 cheaper than the 18RLFCD (and $600 more than a 3/4 ton ducted Midea) but the rest of the stuff doesn't scale.

    The 13,576 BTU/ht estimate for heating load for the entire second floor is on the very high side for the whole 830' ( that's over 16 BTU/ft,whereas I'd expect tight new construction code-min to come in something like 9-11 BTU/ft at the +13F outside design temp). It's in the ridiculously high range for the ~600' of space that it's serving directly.

    The bigger compressor and bigger blower of the 1.5 tonner uses additional start up power when cycling, but it's true that when it's modulating there isn't much penalty at all, and the larger coils may even offer a slight efficiency advantage. Even if with perfect modeling and perfect installation it's possible to show a slight efficiency edge, it will NEVER make up the difference in upfront cost of the oversized unit. If you're going to punt and go big, the 12RLFCD is already more than sufficient overkill.

    The key is to find a duct designer who can guarantee the air flow. Worries about adequate air flow from the rules- of thumb guys are probably what's driving them to the bigger unit (or to tell you even the 1.5 tonner isn't enough, and you won't be happy.)

  24. user-626934 | | #24

    A possible issue with the larger unit is summertime moisture control. Indoor coil temperature at low compressor, low indoor fan will be quite a bit different between the two models, with much less moisture removal form the 18k at lowest compressor speed compared to the 12k. Low indoor fan speed on the 18k will move about 425-450cfm, versus 275-300cfm on the 12k model.

    Granted, this is less of a concern in Brewster, MA than it would be down here in VA....

    Also, I agree with Dana on pricing. I use 3x to 3.5x internet equipment-only price as a ballpark estimate for what installed systems cost in my area. This would put it pretty close to the $7k-8k range for the 18k model...should be more like $6k-$7k for the 12k model.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |