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

Radiant Floor Cooling?

airfix | Posted in Mechanicals on

We are in climate zone 6a in a fairly dry climate. Outside humidity varies between 20% and 35% I would say on those dry days. Our home is going to of a good enough design with R-30 walls and R-60 roof with ACH50 around 1.5 and I’m unsure if I will go with triple or double glazed windows.

I’ve read a lot of the pro’s and cons about using radiant floor heating in a well designed home and I understand that we probably should go with forced air. There are a number of reasons why I am still looking at radiant floor heating most of which are not technical justifications but more because we really really want it. A friend has a house with a similar spec and he loves his radiant floors and says they are not overkill.

Having made a decision on radiant floor heating does it make sense to go with radiant floor cooling? Is this even a thing in a residential home? I understand that we will have to be careful about set points to ensure we don’t have condensation build up on the floors. We are planning polished concrete floors on the basement and on the first floor (understand extra structural support required for first floor). There will be no wood or installed carpet, we will likely have a few rugs here and there. IT seems like radiant floor heating combined with a cooling solution might make the decision to go radiant a little more justified.

Are there any articles or companies I should read or look into?
Here is something I found that made me think of doing it:


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. user-2310254 | | #1


    Does your friend have a tight or a leaky home? If it is leaky, he may find the environment more comfortable because the radiant system has to work pretty hard to reach the set point. On cooling, I have read a number GBA post on the topic. My recollection is it can work but condensation is a concern.

  2. Jon_R | | #2

    You just need to ensure that the temperature of the floor remains above the dew point. If this doesn't occur naturally, could do this with DOAS or dehumidification in an open floor plan.

    Chiltrix makes a heat pump and fan coils (for dehumidification) that could work.

  3. Expert Member
    Dana Dorsett | | #3

    Radiant cooling works well in dry climates with mostly negative latent loads, but gets more complicated to control when humidity is higher and condensation on the plumbing manifolds etc needs to be managed. Some amount of latent cooling with air coils may need to be included if the design water temps are otherwise too low.

    Radiant cooling ceiling or wall panels is more effective per square foot than radiant floor cooling due to the stratification issue with cold floors pooling the cool air near the floor, rather than convecting to the air.

    Of course it all starts with the load calculations, from which the design water temps for doing it solely with the floor can be derived.

  4. Yupster | | #4

    Robert Bean @ would be the expert you would want to contact about this. His website has loads of resources on hydronic space conditioning. He's well known as an expert in radiant cooling. I've always wanted to design a residential hydronic system with radiant cooling, but they are very rare and I haven't had the opportunity. You might consider using an air to water heat pump like what is offered by Ecologix. It can produce chilled and heated water. Lot's of unique stuff going on with a system like this, best get someone with lots of experience in hydronic system design to do your design.

  5. scsiguy | | #5

    We're in the final two months of building a Bensonwood home using an air to water heat pump for radiant floor heating and cooling in Boulder, CO (Zone 5B). Given the building's specs (measured .6 ACH50, 9.5 Btuh/sqft heating, 5 Btuh/sqft sensible cooling, slightly negative latent cooling load) we expect the system to unobtrusively keep the house at a stable temperature. We won't have toasty floors, but we'll have no noise, no drafts, and be comfortable walking around bare foot year round.

    Getting the system designed and installed has been challenging. As others have mentioned, radiant cooling in homes is rare. In our area, I couldn't find anyone outside of commercial outfits that had installed an ATW system - and the commercial guys had no interest in a project our size. I ended up contacting John Siegenthaler (the radiant text book author) who connected me to Harvey Ramer for design assistance. Harvey's design package was detailed enough for my GC's plumber to build and commission the system. But getting my GC and plumber comfortable enough with the design to commit to doing it took a lot of work on my part. I had to completely understand the components of the system, walk them through how the system would function, and be available regularly to answer questions and perform inspections.

    The end result looks great. We'll have our refrigerant line set and fire it up for the first time next week.

    This is not the most cost effective way to heat or cool a tight, thermally efficient home. Mini-splits or central forced air could have worked for a fraction of the installed cost of our system. We decided to do it because we've never been thrilled with any forced-air system in any home we've lived in, and using the floor avoided ducting, radiators, fan coils, etc. that would have impacted our floor plan.

    I'm happy to provide more details if there is interest, but the basics are a Nordic ATW heat pump driving a hydronic buffer tank which is mixed down for delivery to loops set in Warmboard-R. A Tekmar 406 controller manages everything including dew point control for water temps when in cooling mode. The max range of water temps is 55F - 95F. The Nordic can handle the full heating load at a little below the design outside temp, but we have a small electric boiler for backup just in case. Unlike in DX/forced air heat pump setups, this "full load sizing" is possible because there is a buffer tank in the system to prevent short cycling the equipment.

  6. Jon_R | | #6

    Justin, I'm curious why it was necessary to mix down the temperature (reducing heat pump efficiency as to compared to directly producing the needed temp).

  7. user-2890856 | | #7

    Certainly contact Robert Bean as stated above . You will find no one more qualified for this type task . PERIOD .

    A properly designed radiant heating / cooling system will not present any condensation issues .

  8. Jon_R | | #8

    That article makes some good points about radiant cooling being increasingly effective in energy efficient buildings.

    a) as sensible load goes down, large radiators don't have to be as cold - so less chance of condensation

    b) latent load goes down with better air sealing - so less dehumidification needed.

    Maybe we will see more dehumidifiers that work properly at the ~100 CFM flows seen in residential ventilation systems (ie, DOAS).

  9. scsiguy | | #9

    In our current application, the buffer tank temperature and the mix temperature will have similar set points. Divergence can occur though as we tune the system to limit cycling of the heat pump. Given this is a two-stage heat pump sized for the heating load, and our cooling load is much smaller, I expect the buffer tank may at times end up below the dew point controlled water temp that will be sent to the floors. The "mix down", via a variable speed pump to the floor zones, just ensures any temperature overshoot never reaches a surface where condensation can occur. From an efficiency standpoint, we'll still be operating the heat pump with moderate ELT temperatures (at least 10F better than the values used by the manufacturer to compute the COP).

    There is one other case where the tank and mix temps can diverge. This is in a system that combines fan coils with floor cooling. Fan coils want much lower temps that the floor so they can provide latent cooling. In our system, we have pre-provisioned PEX runs to an upstairs soffit so we can add a coil should it ever become necessary. This would serve as second-stage cooling driven off of the tank loop, while still allowing the mix loop to supply properly tempered water to the floors.

  10. Jon_R | | #10

    Thanks Justin. Systems can be designed to produce two temperatures by having the heat pump alternate between two different temperature buffer tanks (ie, one for fan coils and one for large radiators). It's not clear to me that the controls to do this are readily available.

  11. airfix | | #11


    Thanks for the detailed response. We're in a very similar climate to you and it just seemed to make sense that we can do radiant floor cooling. We really only need cooling for about 4-6 weeks in the middle of summer and so I'm looking for a cost effective way of doing that.

    Firstly assuming you were already on the market for hydronic radiant floor heating how much extra do you think you had to pay as a percentage for design, procurement and installation of the radiant floor cooling?

    How did you run the spacing of the tubes? Did you let the cooling loads drive your spacing?

    How do you deal with the stratifying of the air mass? We are planning on having ceiling fans in the bedrooms and the living room and my hope is that those will be enough to move the air around.

    I assume you have a basement with living space above a framed structure. What did you use as your finish floor material. I've been toying with using polished and stained concrete on our main level but that adds a lot of structural cost and complexity. Tiling our whole main level seems cost prohibitive so I'm interested in what you did.


  12. scsiguy | | #12


    The cost delta for adding cooling is difficult to answer. Depending on the load calc for your building and the efficiency you're looking for, there may be fairly inexpensive ways to get radiant heating to work for you (e.g. a gas water heater, a circulator, some simple controls, and in-slab or staple-up tubing).

    We had already decided to go for an all electric house with the plan to eventually add PV and hit net-zero. With that as a starting point, the heat source had to be a heat pump, and heat pumps are most efficient at moderate temps (our system will be calling for 85F or lower water for the majority of the heating season). Low water temps work best in low thermal mass radiant floors/panels or you can use fan coils optimized for lower temps. This brings us to the first big jump in cost for our system. We used Warmboard-R radiant panels to achieve low thermal mass and fast response. Material and labor, just for installing the panels, ran $9.30sqft.

    Warmboard, unlike staple-up - the typical method used in our area - made it possible for us to consider cooling. The cost delta to add cooling from this point should have been nominal: a few thousand dollars for slightly more advanced controls, a little extra copper "near boiler" piping, insulating the pipes to prevent condensation, and maybe one extra circulator pump. But executing it required finding and hiring an outside designer (guessing 4 - 5x the cost for a simple system), ~20% more labor as my plumber learned the controls and the little details that are different when handling chilled water (e.g. more limits on circulator orientation, different pump volute insulation details), and lots of my time.

    The cost delta for you will also depend on your choice of heating fuel. A chiller/heat-pump is more expensive than an AC unit, but that extra cost is offset if you use it for heating too.

    Tubing spacing in Warmboard is fixed at 12". Your designer should be able to guide you on the right spacing/emitter type based on your load calc, min/max water temps for your equipment, and flooring type.

    Stratification reduces the effectiveness of radiant floor cooling. The math says our system should be able to reject enough Btus even without air circulation, but we do have ceiling fans in all bedrooms, the living room and the family room.

    In our house, all living space is above grade. The main floor sits above a well insulated (R20 slab, R30 walls) crawl space.

    All flooring is either slate tile or white oak. The wood flooring has more thermal resistance than the tile. This increases response time and somewhat reduces the cooling efficiency but our cooling load is low enough for it to work.

    Both the main and second floors are insulated. Insulation is required for any radiant floor application. This was a sizable omission in our budget. My GC used OSB strips to keep the insulation tight to the subfloor. For the second floor, where the insulation fills the joist space, these run perpendicular to and are nailed to the joists. For the first floor, the insulation only fills a little over half of the joist space. Here the OSB strips run down the middle of the joist bays and are held in place by "OSB blocking" that rests on the bottom lips of the TJIs.

  13. user-2890856 | | #13

    No radiant works without insulation on the opposite side of area to be heated . Heat does not rise , hot air does . If the resistan
    ce below the radiant tubing , plates , etc is less than the resistance above very little heat will travel to the space meant to be heated . It's all about resistance , heat energy , like water or air is lazy and stupid and will travel the path of least resistance . Insulation is REQUIRED

  14. user-2310254 | | #14


    What a nice looking house. And how did you manage such a clean construction site? But boy, the heating and cooling must have been a PITA.

  15. Jon_R | | #15

    Warmboard with tile above it and wood subfloor below it already has less resistance above. And it's not all about R ratios - air heated from above tends to sit right there - providing less heat to the room below than the ratio suggests. The opposite happens with cooling.

    If you want a radiant floor to also serve as a radiant ceiling for the space below, insulation gets in the way.

  16. Jon_R | | #16

    > Insulation {between floors} is required for any radiant floor application.

    Was this verified with modeling or is this just standard practice?

    > Warmboard, unlike staple-up - the typical method used in our area - made it possible for us to consider cooling.

    Would also be interested in further details on this. My guess is that staple-up could work, especially if left uninsulated. But the upper floor would require some cooling assistance (perhaps with fan coils, but those may be needed for some dehumidification anyway).

  17. scsiguy | | #17

    I found that building a house is very much like sausage making. You really don't want to watch. But if you do, you can't stop yourself from trying to help clean up the mess. Our GC has also done a great job keeping the site clean. In one of the pictures I attached above, you can see the Ram Board that was used to keep dust and mud off the Warmboard through portions of rough-in and drywall.

    Several aspects of the build have been stressful, but that is because we wanted things out of the norm: Bensonwood shell package, multipurpose vs. separate fire sprinkler system, the first DWHR system installed in Boulder county, Zehnder ERVs, radiant heating and cooling, an ATW heat pump, etc. We didn't know if we could find trades willing to learn and take on a challenge. Building is booming here so GCs and trades can be picky. After striking out with 3 others, I feel extremely lucky to have found our GC. He has been a great partner and his subs have been amazing. It wouldn't have been possible otherwise.

    The only sad thing is that my GC is retiring after our job. Not because we have been hard to work with (at least so he tells me :-)), but because he would have retired 5 years ago if not for the 2008 recession. He has learned so much on our project, it's a shame he won't be using that knowledge to help additional clients.

  18. user-2890856 | | #18

    Jon R ,
    Let me know how that works out . Downward losses are not really what you want . You probably know better than I though .

  19. charlie_sullivan | | #19

    Great to hear you found such a good GC. Sad that his his newly developed skills won't be put to more use. Maybe he can become a mentor to some younger GCs. Or start answering questions on GBA as a hobby.

    Coming back the the broader question of when radiant cooling works, a leaky house has a lot of latent load, and when you tighten it up a lot of that goes away. But then when you go beyond that point, you can keep reducing the sensible load with better insulation, and it can come back around to being mostly latent load from the outside air coming in through the ERV or HRV, and the moisture introduced by the occupants. If you aren't doing any active dehumidification, ERV doesn't help you vs. HRV in that regard either. So in thinking about really high performance houses, it seems a lot more likely that you could get by with DOAS and not other cooling than that you could get by with radiant cooling and no dehumidification.

    The latest Zhender models seem to have DOAS options available. I also really like the idea of using solar desiccant dehumidificaiton for DOAS.

    One more comment: I think the biggest challenge in doing chilled water cooling is convincing the plumbers that you need meticulous insulation to avoid condensation, at least on any parts of the system that use water cooler than the dew point.

  20. scsiguy | | #20

    My thinking was similar to yours before I spoke to the designer. He had two concerns: zone isolation between the two floors, and that the air space would increase thermal mass and thus increase system response time. I don't know if this was explicitly modeled between the 1st and second floor, but the design calcs do include a 2.25Btuh/sqft loss to the unconditioned crawl from the 1st floor. (This assumed R19, but we went with R24 because that was the closest to R19 we could get with Rockwool's 24" 'steel stud' bats that fit our TJI spacing).

  21. tschmelcher | | #21

    Justin, can you give us all an update on your system? I am considering installing a very similar system and I would really like to hear if you are happy with yours. Particularly whether you are satisfied with the cooling performance of radiant floors.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |