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

Community and Q&A

Radiant floor vs. minisplits?

clayhufford | Posted in General Questions on

I have a 800 sq. ft. home in Asheville, NC, and am looking to put on a 550 sq. ft. addition. I have been looking into all options of radiant floor heat vs mini splits for our heating and get stuck between preference and practicality. I am adding three floors with each floor being about 190 sq. ft.

I would love to put radiant heat in the pad of basement floor and mini splits in second and third floor, but sense that may be a bad idea since both have high up front cost. Don’t need AC in basement however, so feel it is a waste to put a mini split there, and warm concrete in the winter would be nice!
Is this a bad idea?

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. Dana1 | | #1

    Getting to the right options starts with a room by room heating & cooling load analysis. A 190' per floor the individual rooms probably have heating & cooling loads WAY below the output of a mini-split, but run the numbers and let us know what they are. Mini-splits while modulating, do not have infinite turn-down ratios. Oversizing by more than ~50% for the peak loads leads to diminishing efficiency & comfort.

    It may be possible to run the above grade floors as a single zone with reasonable temperature balance, but the load characteristics of basements are different enough that it's usually better to make it a separate zone.

    A possible mini-split solution to the upper floors MIGHT be a dedicated 3/4 ton single-zone mini-duct cassette (eg. Mitsubishi SEZ/SUZ, using very short duct runs to split the output/return between the multiple zone.

    The water temp and total heat requirements for basements are typically so low that a hydronic loop running off the hot water heater may make the most sense, depending on the water heater type & fuel. But the heating system for the existing house is probably more than 2x oversized, and may be extendable to manage the added load of the addition.Unless it's the world's leakiest house the heating load of that 800' house is probably under 15,000 BTU/hr, and the cooling load under 12,000BTU/hr. It's worth running the room by room load numbers on the original house too, which may steer the solution a number of different ways, depending on what the current mechanical systems are.

  2. GBA Editor
    Martin Holladay | | #2

    I'm not sure why you keep repeating the opinion that oversizing a ductless minisplit by more than 50% results in diminished efficiency. Kohta Ueno's research (and, if my memory serves me well, the research project in the Pacific Northwest) show that sizing a ductless minisplit at 200% of the design heating load often results in the best efficiency.

    These ductless minisplits have excellent efficiency under part load conditions. See “Long-Term Monitoring of Mini-Split Ductless Heat Pumps in the Northeast.”

  3. clayhufford | | #3

    My current house it heated by 1.5 ton heat pump that i belive max's out about 900 sq ft.
    The addition is a basement with two floors above while existing house is basement
    With one floor above. If i do 3/4 ton with ducts for two new floors that may work. Then run radiant in the basement. Will also look into most efficient water heater for domestic and 190 sq ft of heated floors.

  4. Dana1 | | #4

    They are not infinitely modulating. At a capacity 200% of the instantaneous load, sure they're GREAT. But if sized for capacity of 200% of the PEAK load it means that at the average load they don't modulate much, but cycle on/off, spending much of their time in standby mode. The hit in efficiency at 200% oversizing is slim to none- it's a slight bend in the curve sloping slowly downward due to wasted idling power, not a cliff, and in terms of comfort it's less than ideal.

    For not-super-insulated houses the total cfm of 200% oversizing has both noise and wind/draft impacts. At 125-150% oversizing it's in the sweet zone- it's still modulating even at both the peak and average load. At 200% oversizing for peak, many mini-splits would not be modulating at the average wintertime load, and oversizing above that they would almost NEVER modulate. In Ueno's paper the heat loads of the houses were so low that 200% oversizing couldn't be avoided, but if you don't have to oversize that much you shouldn't, and you certainly shouldn't be overzing by 400% or more, which would be the case here.

    A 3/4 ton mini-split like the Mitsubishi MSZ/MUZ FH09NA has a max capacity of over 12,000 BTU/hr @ +17F, which is only 1 degree warmer than Ashville's 16F 99% outside design temp). Serving 190 square feet of code-min addition in Asheville would be more like 400-600% oversizing. This unit has an unusually large turn-down ratio- enough that it would modulate a bit on design-day, but most of the time it would be cycling on/off any time the temps were above 30F or so (which is below the mean temperature for Ashville even in January). The minimum output at 47F is 1600 BTU/hr, for a load that is going to be less than half that at 47F:

    A 3/4 ton mini-duct cassette would still be oversized for the 380' (2 floors of 190') but by a much less egregious amount. It would be running somewhere between 1/2-3/4 of full power at 16F. I would need the real heat load numbers to be more precise, but it'll be under 6000 BTU/hr total for the two rooms. That would be a more appropriate fit, despite having only about a 3/1 turn down ratio. (Note, the "rated capacity" at +17F in the submittal sheet means it's modulating at 810watts of input, out of a possible 1000 watts)

    And it would be only one compressor for the two floors. A multi-split with 6000 BTU/hr heads would run into minimum load issues on the compressor end and would have even more cycling on the compressor end.

  5. user-3368124 | | #5

    This is more question than answer, I am having the same conversation with myself daily regarding minisplits, electric boilers etc.. location, Sturbridge MA

    We have fully insulated slab, 800sqft with radiant that is loved beyond words... except it is heated by a 140 degree OWB cut to very low flow., so it needs a back up if I want to go away.

    Here is complication, we have a 40 panel PV system - that gives 110% of all electrical usage. Now can I have an electric boiler, say a 7kw ??? I might have a $50 - 75 electric bill ... ? Is there a good, better best brand ?

    Then I might also add - can I utilize a Heat Pump HW heater of any style - it would be isolated from living space in a separate building UG that is always 60 degrees with air leakage. ??

    Thank You - I have enjoyed the site very much over the last several months. Look forward to any feedback.

  6. Dana1 | | #6

    Clayton Hufford writes: "My current house it heated by 1.5 ton heat pump that i belive max's out about 900 sq ft."

    That would be very unusual, which is why it's important to run the actual numbers. Heat/cooling loads are not a function of the square footage of conditioned floor space, but rather the square footage & R values/U-factors/solar gain of the exterior surfaces of the house. If the addition adds 200 square feet of west-facing high solar gain glass it might tip beyond the cooling capacity of a 1.5 ton heat pump, but it's highly unlikely that the addition puts the heating load @ +16F beyond the heating output of a 1.5 ton heat pump (unless this is some 30-40 year old R12 refrigerant antique, in which case it's worth replacing anyway.)

    An addition typically replaces some amount of below-code wall/window/roof area with a larger amount of code or better wall area. A 2x4 framed R/11 wall has a U-factor of about 0.1, whereas a code-min 2x6/R20 wall comes in at about U0.065. Similarly clear-glass double-panes run about U0.5, to a current code-min U0.34. So even if the amount of wall area and window of the addition is 50% larger than the old section of wall that got pushed out (or became partition-wall), the net effect on the heating load is ZERO. If the amount of new exterior area is 2-3x that of the original section of wall & window there is a net up-tick in heat load, but it is not at the same average BTU/hr per square foot of the original. If you build the addition to 1.5x better-than-code (usually still cost effective) you may even be reducing the total heat load of the house.

    So, you really DO have to run the heat load calculation. If you have the model name/number of the existing heat pump, look up the extented temperature output numbers to estimate it's capacity at +16F.

    Howard Kelley writes: "Now can I have an electric boiler, say a 7kw ??? I might have a $50 - 75 electric bill ... ? Is there a good, better best brand ?"

    You really should start a new thread, but 7000 watts for an 800' house seems like a ridiculously high 24,000BTU/hr for a load at 0F (a likely 99% temperature bin for Sturbridge MA). Unless the house is unusually leaky or has a lot of single pane glass the true load is probably half that or less. Like anything else, do a legitimate heat load calculation, the numbers will dictate the range of possible backup solutions.

    Electric boilers are pretty rugged- can't say one brand stands out amongst the other (for either good or ill.) As a backup system the simplest dumbest and smallest unit that actually covers your load makes sense- no outdoor reset or super-modulating design required.

    The wood boiler is almost certain to be ridiculously oversized, and unless it is coupled to a large buffer tank or something the efficiency of the wood boiler is going to be really low.

    "Then I might also add - can I utilize a Heat Pump HW heater of any style - it would be isolated from living space in a separate building UG that is always 60 degrees with air leakage. ?? "

    Too many abbreviations- can't wrap my head around "UG" at the moment. But when you start pumping heat out of a building with a hot water heater, the odds that it will still stay at 60F are pretty iffy. Is it passive solar gain or something else that keeps it at 60F even on days when the high temps are barely into positive single-digits (like it might be come this Tuesday.)

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |