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
Q&A Spotlight

A Shortcut for Whole-House Air Conditioning

To save money, a homeowner wants to use a fan with a thermostat to circulate cool air

Fans instead of more minisplits: In order to bring conditioned air to a first floor office and two small bedrooms on the second floor, a GBA reader is considering thermostatically controlled fans instead of additional minisplit heads. Floor plan courtesy Paul.

Paul lives in a 1700-sq.-ft., two-story house in Massachusetts where he plans to install heat pumps. The upper and lower levels share an open space under a vaulted ceiling, essentially connecting the air volumes of the first and second floors. And that’s given Paul an idea for a low-budget way of circulating cool air throughout the house.

“I want to maintain a comfortable temperature in the home office and two seldom used bedrooms,” Paul writes in this recent Q&A post. “All three of these rooms are small–about 120 sq. ft. It would be expensive to equip each of them with its own heat pump. Could I put a fan through the wall of each room to mix the air in these rooms with the air in the main living area?”

He plans to keep the main living area at about 70°F with ductless minisplits. In each of the three small rooms, he would install a wall fan that would be activated with a thermostatic switch.

“By mixing the air in the smaller rooms with the air in the main living area, the temperature in all the rooms would presumably stay at the temperature called for the thermostats on the heat pumps in the main living area,” he says.

As anyone heard of this approach? And are off-the-shelf thermostats for fans available? Those are the questions for this Q&A Spotlight.

The idea should work

An unnamed GBA reader named User-5946022 (who we will call User 59) cites an article by Carl Seville in which Seville discusses a similar tactic in his own home. Seville used a Panasonic 190 cfm fan near a minisplit head on the second floor to duct conditioned air into two bedrooms at the front of the house.

The fan runs whenever the minisplit…

GBA Prime

This article is only available to GBA Prime Members

Sign up for a free trial and get instant access to this article as well as GBA’s complete library of premium articles and construction details.

Start Free Trial


  1. Expert Member
    KOHTA UENO | | #1

    I'll echo Jake Staub's sentiment of "a redistribution fan will be a marginal system, with marginal results," based on BSC's previous research on ductless heat pumps and simplified heating and cooling. The giant brick of a report that we did under the Building America contract is this one:

    BA-1407: Long-Term Monitoring of Mini-Split Ductless Heat Pumps in the Northeast

    However, a hopefully more digestible version is the slide deck from a presentation covering key results:

    Minisplit Heat Pumps: Lessons from the Field

    Simplified space conditioning starts on slide 25—takeaways include transfer fans being only of marginal use, single point works really well with super-insulated houses when the doors between rooms are open, bonus rooms and two story buildings with a single head are a problem, and going over ~900 sf per head is where we saw problems in the field.

    Specifically to this idea of "grabbing cold air" and redistributing to those upper floor bedrooms: check out the case study described in BA-1407 under "6.4 Thermal Buoyancy Effects (Use of Single Mini-Split Heat Pump
    on First Floor)." Our builder said, "Hey, the second floor heat pump barely runs in winter, maybe we can get away with a single head on the first floor." Sure... it worked for the winter... and then come summer, he started getting comfort callbacks. Because warm air rises, and cool air sinks... so no space conditioning to the second floor. And that *did* include a transfer fan to upstairs bedrooms... but it was evidently not enough to work. The builder ended up doing a very expensive retrofit of a MSHP head on the upper floor.... lesson learned.

    Even worse: *where* will the upper floor bedrooms be "grabbing" cool air? If they are in the wall of a two-story space, they will probably be gathering 75-80F air, rather than 70 F air (closer to floor level), due to thermal buoyance and stratification in a two-story space.

  2. jameshowison | | #2

    I would love to read a GBA case study of an Airzone install.

  3. frontrange | | #3

    I think this really highlights one of the major drawbacks of ductless mini-splits - no distribution. The ductless mini-split is a great tool when distribution isn't needed, but trying to force it into scenarios where distribution is needed is trying to fit a square peg in a round hole. Just use the right tool for the job (e.g. a ducted system).

    1. Expert Member
      MALCOLM TAYLOR | | #4


      I agree although I think the answer is very climate dependent - and that's why Carl Seville finds he is able to make do with transfer fans in his newly built Pretty Good House. Around me here in the PNW, where cooling isn't (presently) and issue, ductless mini splits are often being put in as replacements for a central heat source, which in the past was often a wood stove. Because in our climate a well insulated house doesn't see much temperature variation between rooms, a bit of supplemental baseboard heat can easily even things out.

      1. frontrange | | #5

        Agree. The supplemental baseboard heat is of course a heat distribution system. And distributing heat is generally easier to do because there are more options (e.g. hot water or electric, baseboards or panels), but cooling is pretty much limited to a mechanical heat pump for residential housing.

        But I often see things like "just leave the bedroom doors open", which always makes me feel sorry for the next home owner, who will probably end up considering options like blowing air with fans as in this article.

  4. Expert Member


    "The supplemental baseboard heat is of course a heat distribution system"

    I hadn't th0ught of it like that, but of course you are right - and as you say cooling is a tougher nut to crack.

  5. Expert Member
    CARL SEVILLE | | #7

    Since I seem to be the subject of a bit of disparagement, I'll chime in here. The transfer fan does do the job in my house, but it is a very unique situation, and would not likely be useful in a standard home. My house is extremely air tight (<1 ACH50) well insulated, and very compact. The fan move a small amount of air into two bedrooms when operating. Most of the time, with the doors open, the fan doesn't serve much purpose. When the doors are closed, it can help avoid big temperature swings between the rooms and hallways. I don't claim to be able to handle the advanced math but my most clear experience was when my wife was ironing in a room with the door closed in the summer. The transfer fan was not switched on and the room was several degrees warmer than the rest of the house. After turning the fan on, the temperature began to even out. That said, in retrospect I wish I had installed compact ducted HVAC systems. That would be in the next house, but very unlikely I will build another one. BTW, for reference, my home is in CZ 3, and our temperatures are generally pretty mild. Humidity is our biggest issue.

    1. Expert Member
      MALCOLM TAYLOR | | #8


      The advice on GBA has to be widely applicable. Unfortunately that means it misses the outlier situations where other solutions work.

      I live in a rural area dominated by logging. The debris left behind that isn't salvaged as firewood is typically piled up and burned when the rains come. That completely alters the equations around carbon sequestration and air pollution. The advice GBA gives about including wood heat in a house makes sense pretty well everywhere but here.

      If what you came up with works in your situation, I don't think you need to be apologetic about it.

  6. maine_tyler | | #9

    I've always been confused by the 'air doesn't move a lot of heat' concept since that is what an HVAC system does: moves heat via air. Is the difference just the amount of air able to be moved via ducts vs fans? Somehow I don't feel like the distinction is typically made evident (at least in layman terms).

    Trying to make sense of the technical explanation by Jake:
    "Q = 1,200 / (1.08*5) = 222 CFM or approximately 2,200 CFM/ton!"

    I'm sure it's obvious to HVAC types, but what happened here? What's the conversion from the 222CFM of air needed to this 2,200 CFM/ton?

    In my head (mistakenly, I'm sure) it goes like this:
    — A 120 sq. ft. room with 8' high ceilings has a volume of 960 cu. ft.
    — If using a 190 CFM fan like Carl, the room will see total air change every 5 minutes, under ideal conditions.
    — Changing out the warmer air with 70 degree air every 5 minutes should keep the room cool unless the heat gain in 5 minutes is able to raise 70 degree air to an uncomfortable level WITHIN that 5 minutes. (Sound reasoning? Not sure.)
    — The energy needed to raise 960 cu. ft. of air by 5 degrees F is about 92 BTU's. (I can come back later to show work if needed)
    — There are assumed to be 100 BTU's added every 5 minutes at 1,200 BTU/h rate. ("assume that a room of Paul’s has exterior walls with typical windows. In which case, 10 Btu/hour-sq.ft...." Does this assume exterior walls on all sides?)
    — So in 5 minutes, the mass of air would just about raise 5 degrees (this ignores other objects in the room), but in the same 5 minutes the air has completely changed over.

    Getting a little hand wavy here, but doesn't that imply a sort of equilibrium in which the temperature would remain rather constant at around 75 degrees?

    It's more than likely I made a math error there, or if not, I'm getting lost in concept somewhere.

    1. Expert Member
      MALCOLM TAYLOR | | #10


      This discussion does a good job of laying out the difficulties:

      1. maine_tyler | | #11


        I guess I take issue (read 'don't understand') the approach laid out by Martin and here. The low specific heat of air and the low delta T is cited as evidence that not enough heat can be moved. Perhaps that is the case, but I struggle to intuit it. If a room's air is being turned over completely every 5-7 minutes, it would seem like the new 70 degree air isn't imparting energy via a delta T as suggested at all, but rather replacing it. Convection not radiation or conduction. So I'm not getting how the delta T plays in as suggested. The delta T will come into play at the heat pump coil in the other room. Perhaps the equation cited be Martin and by Jake here can be better explained and that would likely clear up my confusion.

        I think I've done a halfway decent job outlining the way it makes sense in my brain, and would be happy to hear someone directly refute it because I have no doubt something must be wrong-headed about my approach.

        And I'm not trying to argue that fans are a good option vs ducted systems as I can see that there are likely many real world challenges and 'unideal' scenarios (like perhaps changing all the air every 5-7 minutes is just not realistic in the real world without dedicated ducts). I'm trying to clarify the mathematical/scientific argument used to invalidate the idea.

        1. Expert Member
          MALCOLM TAYLOR | | #12


          I think part of that is because the whole air in the room isn't being turned over. It's being mixed and that mix removed. If you add a cup of water to four cups of cranberry juice and then take out a cup, you can repeat the process four times, but you don't end up with clear water in the jug.

          If things worked the way you suggest, and you turned off the heat in a tight house of say 1 ACH 50, it would be whatever the outside temperature was in one hour.

          1. maine_tyler | | #14

            "If things worked the way you suggest, and you turned off the heat in a tight house of say 1 ACH 50, it would be whatever the outside temperature was in one hour."

            That's an interesting point (with the caveat that it's at 50 pascals—it would actually be pretty interesting to see how many degrees a house DOES change in an hour at 50 pascals of pressurization). It makes me realize how little I understand about the intersection of thermodynamics and fluid mechanics.

            It would seem the nature of the air exchange drastically affects the energy exchange (which to me suggests a purely delta-T driven equation would be lacking accuracy if so dependent on fluid dynamics?).

            Would maximizing mixing reduce the temperature change, whereas having a clear entry and exit path on opposite sides/stories of a house increase the temperature change? Turbulent vs laminar flow?

            On the extreme end, if we picture an air conditioned house as filled with water (representing cool air) and we:
            ‣ introduce an opening in the envelope at the bottom and top of the vessel that allows 1 gallon-per-minute of water to pass,
            ‣ and if the house is able to hold 60 gallons,
            ‣ in one hour the house would be entirely filled with air.
            ‣ but this process happens under complete stratification and probably is considered a laminar flow regime.

            But real world, your example is more appropriate (two similar fluids that don't perfectly stratify) where the injection point creates turbulence and mixing. It's pretty tough to make sense of the energy exchange at that point, but it makes total sense that a house filled with cranberry juice wouldn't be flushed clear in an hour even with enough total fluid flow.

        2. frontrange | | #13

          The thermodynamics are more complex then that. Before the fan is turned on, heat is entering the room (e.g. through the exterior surfaces) and is exiting the room (e.g. through the interior surfaces). If the rate of heat entering is greater than the rate of heating leaving, the temp in the room will rise. The rising temp decreases the rate of heat entering and increases the rate of heat leaving the room until an equilibrium temp is reached. (Or vice versa if the rate of heating leaving is higher.)

          When the fan is turned on, the air isn't actually being changed out every so many minutes. Instead what's happening is that cool air is entering the room, mixing with the air in the room, and then that mixed air is leaving the room. Because the entering air is cooler than the exiting air, heat is leaving the room and the rate that it leaves is driven by the temp difference between the entering and exiting air.

          As the fan exchanges cooler air for warmer air, the average air temperature will decrease which causes the temp differences to change: the heat entering will have a larger temp diff and rate of heat entering the room will increase, the heat leaving the room through the interior walls will have a smaller temp diff and the rate of heat leaving the room will decrease, and the heat leaving the room due to the air exchange will have a smaller temp diff and the rate of heat leaving will decrease. And the room will reach a new equilibrium temp below the previous equilibrium temp but above the temp of the air entering the room.

          You'll always have this issue as you cool the room down, it will get more and more "difficult" to cool it further and simultaneously "easier and easier" to heat it back up.

          1. maine_tyler | | #15

            "Because the entering air is cooler than the exiting air, heat is leaving the room and the rate that it leaves is driven by the temp difference between the entering and exiting air."

            Does this assume a theoretical 100% mixing? In the real world, is near perfect mixing actually a pretty close approximation of reality? I can see that my previous thinking assumed pretty much 0% mixing, which is obviously not reality.

            I am still baffled by this:
            " = 222 CFM or approximately 2,200 CFM/ton!"

            To me, that's the part where the magician covers the box with linen and spins it around a few times before revealing that the body actually wasn't cut in two.

            * Found a good explanation of the 1.08 multiplier:

            ** "temp difference between the entering and exiting air" hmm, maybe part of my hang up is on which delta T to use. Martin mentions the delta T "between the main room and the bedrooms." Presumably the temp 'of the bedroom' is only the same as the 'exiting air' if there has been complete mixing. The highest delta T would occur if no mixing occurs, which does line up with your statement and my current thinking.

          2. frontrange | | #17

            "Does this assume a theoretical 100% mixing?"

            No, the only assumption I'm making is that there is some amount of mixing. But if you wanted to calculate numbers, you'd need to model the mixing.

            "In the real world, is near perfect mixing actually a pretty close approximation of reality?"

            Depends on the scenario, but I'm guessing it's going to generally be a decent amount of mixing but not perfect.

            Minor point, but if there was no mixing that can lead to the largest or the smallest temp difference depending on the locations of the inlet and outlet. (For smallest imagine the inlet is pointed at the outlet.) There's a lot of variables in real world scenarios that can be difficult to model correctly, which is why it's difficult to know exactly what is going to happen especially when the desired heating/cooling input is close to the undesired gains/losses.

          3. maine_tyler | | #18


            makes sense, thanks.

    2. Tim_O | | #16

      2200/ton comes from the math of 222CFM = 1,200 BTU. Break that down, 1CFM moves 5.405 BTU. 12,000 BTU in a ton, 12,000/5.405 =~ 2200CFM.

      It is worthwhile in my opinion to do at least a basic heat load on the room. Using a Borst calculator to do some rough math, a 120sqft room with only a north facing wall, may only need 300btu. And worth considering, bedrooms are often used at night where the cooling load would be lower yet. Even a warm August night here in Michigan would drop that 300btu quite a bit more.

      Finally, all this CFM talk doesn't take into account the heat loss radiated through interior walls from the 75* bedroom to the 70* cooled area. Again, using some rough calculations, 0.5" drywall = R 0.5, times both sides of the wall. Counting the studs and 3.5" airgap does add more R value though. So wall assembly has an R value closer to 2.5. So for our 5* temp delta = 2 BTU/sqft.

      My example is a 10x10 room on the 1st floor in the north east corner of a "Not Very Great" tract house located near Detroit in Climate zone 5. With a 0.33 Uvalue window on both the North and East walls. Maybe a seldom used bedroom or an office?
      Sensible heat gain is around 500btu mid day. Assuming one interior wall is shared with the house at 80sqft. ~ 160 btu would be lost from the 75* bedroom to the 70* living room through this wall (assuming the other wall is to another 75* bedroom). Leaving only 340 btu needing to be transferred by the fan, now you only need a 63CFM fan, which would be much more reasonable. Just a computer fan will do it, and you can even get one with cool LEDs.

      Of course, all these numbers grow much bigger the less your DeltaT between the living area and your bedroom is. You may need to drop the living room to 67* if you like sleeping in a 70* bedroom. And your south facing bedroom may need a mini split or exterior shades. In a Pretty Good House, that 500 BTU would probably be less than half of that. Also worth noting, I didn't do anything with the ceiling, I just assumed another 75* bedroom was above for simplicity.

      My point being, it's worth running the numbers.

      1. maine_tyler | | #19

        Thanks superman. The math makes sense, though I suppose I'm still hazy on what the significance of citing that number is.
        I thought the CFM per ton mattered mostly in terms of what's passing over the coil (for latent vs sensible load reasons) so I'm not sure what the relevance is to room-to-room exchange fans. But that's ok, I will continue to investigate.

        1. Tim_O | | #20

          I think the per ton number was added more for dramatic effect than anything else. In the math done by the original post, 222 CFM was the number that matters.

Log in or become a member to post a comment.



Recent Questions and Replies

  • |
  • |
  • |
  • |