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

Ducted minisplit sealing

Gerald Pehl | Posted in Mechanicals on

Hello GBA Community.

My passive house architect specified a ducted mini-split system, and figuring out where to house the indoor unit raised the question of how to seal it- it’s mounted tight to a surface, so I’m not sure what to do with the long edge that’s pretty much inaccessible. Also, I understand that mastic is the optimal way to seal duct seams, but what is the best way to seal the connection between the supply and return plenums (plena?) and the evaporator? Lastly, it seems that the three pieces will fit together like a puzzle, and that the evaporator won’t be easily removable for service when it’s all done.

Attached is an illustration of the design for reference.

Thanks in advance for any thoughts!

And big thanks for all the advice I’ve benefited from in the past, too.

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

    Mastic seal & tape the plenum and other less-accessible parts that are going onto the wall to the unit before it's mounted. From the vertical orientation I can only assume this is a Fujitsu RLFCD(?) like this one:

    On parts that may need to be swapped out for service aluminum tape is usually the better choice. But to reliably seal it when re-installing just the one piece of the puzzle is a bit challenging.

    BTW: The coil in the cassette is only an evaporator when it is being used for cooling. It's a condenser during the heating season.

  2. Gerald Pehl | | #2

    Thanks Dana- yes, the indoor unit is the ARU18RLF. And yes, I understand that it's not an evaporator when it heats, but then what is the generic term for it?

    Can I ask what type of tape you like to use over the mastic?

  3. Gerald Pehl | | #3

    After thinking about this some more, I'm hoping for a little more input:

    - For reference, the longest branch duct will be about 20' of oval - 6" round equivalent.

    - How concerned with gentle transitions should I be- do I need the angled transitions in the vertical parts of the supply and return plenums? should I have turning vanes for _all_ of the takeoffs, or do I need turning vanes at all?

    - Do factory seams need to be sealed? Like the closing corners on the custom plenums, and the long seam on the oval duct?

    - What kind of insulation products are typically used? What are the products used in the photo that Dana's attached? what kind of R-value should I be shooting for?

    - I'm assuming dampers are used for balancing, but can I use motorized dampers for zoning, or do the mini split not like that?

    - I'd appreciate any reactions to my drawing- thanks again!

    1. Expert Member
      Akos | | #5


      The length of the ducts do not matter as much as the transitions. Your unit does about 0.3" wg, which is pretty decent but not near a standard AC so you do have to pay attention to losses. Also be careful with oval ducts, a 6" oval is about the same at 5" round.

      I would be careful with your two horizontal to vertical transitions, those should be swept bends not sharp as shown.

      Turning vanes definitely help (adding in one on a badly done furnace return reduced the pressure loss by almost 0.1"). You can also you use side bends for your side takeoffs. Lets say your outlet is an 8x32 duct, put one 8x8 side elbow, in the middle a short section of 8x16 duct and another 8x8 side elbow. The 8x16 duct can then feed two more 8x8 side elbows. Connect your oval ducting then to the side elbows. Less of a make and more of a buy type of setup.

      Ducted minisplits like to see a know fixed pressure loss which you normally program into the unit. Motorized dampers might not be the best, you might be able to add in one or two as long as you don't reduce the overall flow in the system too much. Balancing dampers are a must for properly commissioning the setup.

      Up here in the great white north we don't insulate duct runs inside the conditioned envelope.

  4. Walter Ahlgrim | | #4

    Will all your duct work be inside the conditioned space? If not I say make a new plan. If so in my opinion insulation and duct sealing are not going to change your operating costs, if insulation and duct sealing have any effect at all the farthest room may get slightly more air but but it is unlikely to be perceptible.

    The duct work in post 1 is great it is big and short. 20’ of 6 inch is not. Mini split duct work must be designed very carefully as the blowers are not very powerful. Insist in a computer model with static pressures consistent with your blower.

    Where is the air filter?


  5. Gerald Pehl | | #6

    Thanks very much for the comments Akos

    - understood about oval duct- I actually did pressure calculations, although I had to estimate with the plenum pieces since they're not standard shapes/sizes.

    - I don't follow your comment regarding vertical to horizontal transitions- the inner corner is sharp, but the outer corner is a smooth radius- can you please elaborate?

    - My thoughts on insulating were driven by the photo that Dana linked to- it got me thinking that condensation is a possibility and that in heating season the area where the unit is housed could get over-hot. Maybe they're not realistic concerns- does proper function means that heat is being transferred at a fast enough rate?

    - And thanks for the thoughts on dampers- it now occurs to me that zoning could create states where static pressure is very high.

    1. Expert Member
      Akos | | #8


      With low static pressure setup, you generally want to stick to smooth bends, radius on both inside and outside. When deciding on the fitting to use, good to look at loss chart like this:
      Smooth bend is 10' equivalent loss, sharp bend is 70'.

      The one area I've had condensation issues with here is sweat on the drain line, good to put some pipe insulation on that especially if it runs into the attic. Also be careful with the insulation on the AC lines going to the unit, even a small gap on the insulation and condensation will drip.

      The filter the unit comes with mostly works to keep large particles out but the coil and blowers will get dusty over time, adds a bit more maintenance. If you have the space and pressure budget, an inline pleated filter is a good thing. Larger is better, they last longer and create less pressure drop. For reference, the mitsubishi filter box comes with something like a 12x36" filter.

  6. Gerald Pehl | | #7

    Hello Walter- thanks very much for your thoughts.

    - Yes, the entire system is in conditioned space- super insulated passive house retrofit spec.

    - Especially thanks for the comments on sealing and insulating- I'm comfortable with the math and science behind all of this, but it's more difficult to know where various effects become important.

    - Also thanks for asking about the filter- I hadn't got to that detail yet. The unit came with a synthetic mesh filter- on the order of 0.5 to 1.0 mm openings. I haven't read the documentation closely on this point, but it appears that the filter is intended to attach to the return aperture of the unit, but it would not be easy to access under the current design. My preference would be to put a filter behind the return grille and service it by removing the grille every few months. Which begs the question of what kind of material it should be- I would assume it shouldn't introduce anymore resistance than the one they included. Thanks in advance for any further thoughts.

  7. Jerry Liebler | | #9

    To get a filter that actually works and can be powered by the ducted minisplit means it needs to be BIG! Can you use a 24"x 24" return grill? If so the air velocity will be 1/4 of the CFM and around 100 feet per minute. OTOH if your return grill is 6" x24" the flow rate through the grill will be numerically equal the the CFM. Filtrete brand filters have pressure drop specified at 300 feet per minute and the merv 12 ones are about 0.2" wc. With the 6"x24" the return flow rate will be about 400 feet/minute the filter will use all your fan power leaving none for the ducts ! However if you have the 24" x24" situation the pressure loss will be about 1/9x0.2 or 0.022"wc hardly noticed.

    1. Gerald Pehl | | #10

      Hello Jerry- yes I can go just about as large as I want, and so I will. The bigger question for me is what kind of filtering I need- I didn't see anything in the design guide, and so it's not clear if the mesh that is included is sufficient.

  8. Jerry Liebler | | #11

    The mesh offers some protection against dust buildup on the coil & and; blower but IMHO you'll really want more filtration and cleaner indoor air quality. I'm opting for MERV 12 pleated filters. The 2'x2'x1" size is popular and suitable "filter- grills" and filters are readily available and would be my choice.

  9. Gerald Pehl | | #12

    Thanks again Jerry. We have a ventilation system for air quality, so my primary concern is maintenance of the indoor unit- keeping it free of dust and debris to optimi performance.

  10. Jerry Liebler | | #13

    Typically the ventilation flow rate is much smaller than the heating-cooling flow rate and it's filters are more specific, expensive and inaccessible. So my preference is to use the HVAC to do most of the filtering.

  11. Gerald Pehl | | #14

    Jerry- good points on all :-)

    thanks again!

  12. Expert Member
    Dana Dorsett | | #15

    The picture I linked to was of an installation in California, the system designed by a guy named Larry Waters. It looks like the short plenum and cassette (= "...the generic term...") are double or triple wrapped in the bubble-pack type radiant barrier, which is good for about R2 per wrap. The ducts are all standard pre-insulated flex duct. The installer runs duct-blaster tests to prove air tightness, and claims most of his company's installations are tight enough to be nearly unmeasurable leakage even at the "0th ring".

    Duct tightness IS important- velocity & pressure have to be preserved in order to guarantee proper mixing and volumes at both minimum & maximum blower speed. so sealing even the factory seams of hard piped ducts is "worth it." Larry Waters refers to the supply registers as "nozzles", with throw and angle carefully selected for each room to ensure reasonable mixing & distribution even at the lowest speed cfm of the cassette.

    Apparently the fully installed price of that particular installation for the 1.5 ton Fujitsu was about $13K, including the Manual J & ducts & design work, installation & commissioning. That's considerably more expensive than it costs in my area, but I don't live in the richer parts of the eastern San Francisco Bay area, where the overhead costs to the contractors are higher than some other locations. There are probably more details to be gleaned from Larry Waters' comments in the discussion of that house on this bit o' bloggery.

  13. Gerald Pehl | | #16

    Thanks very much Dana- the expertise on this forum is incredible.

    'cassette'- somehow I had gotten the impression that term referred to the finished indoor unit(s) of duct_less_ system. Thanks for the enlightenment.

    Especially thanks for your comments on duct sealing.

    Do you think insulating, beyond sealing, is important? The photo appears to be a conditioned space, so it makes me wonder if air flow through the system is slow enough so that heat/cold can actually build up somewhat right at the cassette?

    It's interesting to me that a high end installation uses flex duct, and it appears that it's totally common, but I've been scared off of flex duct by some reading I've done:

    Allison A. Bailes III, PhD;

    But also, I just don't think flex duct is made from very robust materials- I don't like the stress I have to put on it in order to pull the folds out of it- I've handled plenty of runs where the connections failed because the material itself isn't strong enough to withstand the force of pulling out the folds, AND a key use of flex duct is for bends which inherently introduce folds on the short side of the bend, which is the most critical area for creating turbulence, if I understand correctly. Obviously some of this is conjecture- I've never seen any formal studies or anything.

    That said, I look forward to being corrected by the experience on the list :-)

    Lastly, can you elaborate on supply registers? Can you point me to a discussion or article on the topic? I don't feel like I see many options in the big box home improvement stores.

    Thanks again!

    1. Expert Member
      Dana Dorsett | | #18

      >"Do you think insulating, beyond sealing, is important?"

      That's often a code requirement even in conditioned space, the primary rationale being limiting the risk of copious condensation on the outside of the duct during the cooling season. That issue is somewhat moot for modulating mini-ducted systems, but the local letter of the code still has to be followed.

      I too am not partial to flex duct, and prefer well designed hard pipe whenever possible. There are too many ways to cut corners and screw it up with flex duct. The crews need to be trained on the issue for it to always work out well.

      John Semmelhack (who posts here often, and is a big fan of the ducted Fujitsus) is a better resource for what's out there for diffusers/registers with different throw patterns. Check out the diminutive very-directional register depicted in this Fujitsu installation it posted pics on in response #14:

      What you won't see in a well designed mini-duct cassette system is a big rectangular grill type diffuser /register, which won't have sufficient velocity for good mixing at low cfm.

    2. Ethan Foley | | #19

      A great source for an overview of the process for selecting supply registers is this Engineering Guide to Air Distribution
      You will probably have to go to a local hvac supplier to find a source for diffusers and grilles with performance data and multiple sizes. The big box stores around here just carry 4x10 floor registers and stamped metal grilles.

      1. Expert Member
        Dana Dorsett | | #20

        Thanks for that resource!

  14. Expert Member
    Peter Engle | | #17


    Re-read Allison Bailes articles. They do provide the results of formal studies about using flex ducts, bends, etc.

    Flex duct is not automatically a bad thing, but if you follow Allison's advice, they've got to be pulled tight and used primarily for straight runs. Elbows should still be done in sheet metal. IMHO, the real value in flex ducts is that you have insulation and seamless ductwork in one package. You can also use it for small realignments without much friction penalty. But real changes in direction are still best done in sheet metal. The caveat being that if the system is well designed and installed, the friction from a few elbows might make no difference at all. This is why actually calculating friction losses is so important. until you run the numbers, you have no idea whether you're in good shape or not.

  15. Jerry Liebler | | #21

    Are " carefully selected nozzles" really necessary to assure adequate mixing where a ducted mini split indoor unit is used? Probably if the supply and return are right next to each other as they are in an un-ducted situation (why use ducted units in this situation). However, if the ducts supply to the perimeter and the return is in the " middle" of the area being served, I'd say not so much because substantial mixing will occur even at the very low velocity flow through the room, from supply to return, because that flow will be disrupted by any furniture or other "objects" in the room and inevitably result in substantial "mixing". There is no attention to this "detail" in conventional residential forced air systems and consumers accept the performance as adequate, however the "conventional" installation has supply along the perimeter and returns, more or less, centered. With a minisplit the CFM/BTU is quite similar to more conventional HVAC but the fan pressure is much lower so bigger ducts are needed. The bigger ducts make the linear flow rate lower resulting in smaller pressure losses and the lower linear flow rate continues right out of out of the supply "registers" but does it really affect mixing? I think not materially provided the supply to the perimeter, central return topology is maintained! In other words I'm going to use conventional (big box store) supply registers at 1/3 to 1/4 the typical linear flow rates and I expect acceptable mixing! My mind could be changed by hard data that suggests otherwise.

    1. Jon R | | #22

      Yes, supply face velocity really does effect throw, mixing and stratification and is far more important than registers location. But I don't have a link to good data showing this.

      1. Expert Member
        Dana Dorsett | | #23

        >"Yes, supply face velocity really does effect throw, mixing and stratification and is far more important than registers location."

        That's right!

        And it's far more important in low performance building envelopes like this one, which will have very cool wall surfaces on the exterior walls than it would be in high-performance houses. In a PassiveHouse it hardly matters at all, but this house is on the other end of the envelope-efficiency spectrum

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |