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

Maximal minisplit efficiency

Corey_Haddad | Posted in Mechanicals on

A few things I have gathered reading here and elsewhere on mini-split efficiency

-A larger compressor, operating at the lower end of its range is more efficient. (But not too far down, as at some point, efficiency drops off again)
– The compressor still shouldn’t be too big, or short cycling can occur, reducing efficiency
– Maintaining a smaller temperature difference is easier, so higher COP

I live in an old, drafty, and poorly insulated farmhouse near Olympia, WA. Currently the main source of heat is wood stove, with space heaters used to supplement. Needless to say, the interior of the house experiences very large temperature swings, from being in the 40’s on frozen winter mornings, to well into the 80’s when the fire really gets cooking. Eventually the house will be well insulated and sealed up, but that’s some time off. In the meantime, I am looking to get a mini-split to help the house not be quite so cold when the wood stove isn’t going.

Here’s how I’d like to program the heat pump to operate:
Default: run at 25% speed always (25% being a proxy for max efficiency)
except when indoor temperature is less that 50°F , then use up to 75% of speed to get back to and hold 50°.
and use all available capacity to hold 40°
if the temperature reaches 80°, turn off, and only turn back on at 70°.

My goal with the above program is to get the highest COP most of the time as a baseline input of heat into the house. Trying to maintain a set point is a costly proposition, and I don’t mind a variety of indoor temperatures. The wood stove will be utilized when temperatures above what 25% of the heat pump’s output can generate are desired.

Turning off at 80° and only back on at 70° is to avoid short-cycling. A 10° range seems not short. And why 80°? If the heat pump ever managed to make the interior of the house 80° in the winter running at only 25% speed, it would be a gift I’d happily accept.

Is my theory correct here?

And is there any mini-split out there that can accept this kind of input? I can write some python code and program and Arduino if that’s what would be needed.

If direct control or limits on compressor speed would not be possible, perhaps it would work to write a program that continually moved the set point, keeping it 2° above the actual indoor temperature? Might that ‘trick’ the unit’s firmware into thinking it was close to it’s goal, and therefore run the compressor at a low speed?

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. Jon_R | | #1

    Yes, it it likely that the mini-split will run at low power when the temperature is close to the set point. You may need to allow it to occasionally reach set point to avoid the "I" factor in PID controls.

    I expect that air sealing/testing would take less time and provide better results than your planned optimization.

    1. Expert Member
      Dana Dorsett | | #8

      >"I expect that air sealing/testing would take less time and provide better results than your planned optimization."


      Air sealing is cheap stuff too- usually by far the best bang/buck for a building that would be described as "... old, drafty, and poorly insulated farmhouse ...".

  2. walta100 | | #2

    From the way you describe your home and how you operate it to me it seems unlikely a mini will rarely be running at low speed.

    When you use the wood stove the mini will be off. When the wood stove goes out by the time the mini understand it needs to run it is likely to be too late for it to spend much time on low it will switch to high trying to heat you leaky uninsulated house. And when you light the stove it is back to off again.

    Minis work best at a constant temp setting in a tight well insulated building where they can find a speed that just replaces that is escaping the building and run 24/7.

    Oversized equipment and trying to make big indoor temperature changes is a bad plan for a heat pump system.

    I say money spent insulating and air sealing will make you more comfortable than new equipment.


  3. Corey_Haddad | | #3

    Jon R - that makes sense.

    Walta - The problem I am anticipating is battling the 'brains' of the mini-split. They are programmed with a certain use case in mind. I may not have described it clearly, but I am hoping to have the mini operate at low speeds most of the time, resulting in a defined contribution of heat, at a high COP, without respect to set point under most circumstances. The tradeoff I am willing to make is to allow the house to fluctuate as the ambient temperature changes.

    In thinking more since I posted, maybe the simplest way to approximate that would be to write a program that can interface with the head unit, and which could adjust the set point to keep it a fixed amount above ambient. I say 'simplest' because it looks like there are already software libraries out there that can turn an Arduino into a programmable remote, using an IR diode that points at the head unit. The mini thinks you are sitting there pushing buttons on the remote.

    I would have my own outdoor temp sensor feeding into the program. When the wood stove brings the temperature higher than the current set amount above ambient, the mini would turn off. As the house eventually drops back through the set point, the mini would turn on again. I'm not sure about the delay you speak of Walta. Can you say more about that? It seems a standard situation that indoor temperature is high enough to exceed set point, so mini is off, such as during a sunny day, and then as night falls, the temperature steadily drops until heat is needed. If I were the mini, I'd turn on slowly in that situation, and slowly ramp-up as needed to hold the set point. And again, since my set point will be a fixed amount above ambient (perhaps 20f or 25f), this should be easy for a larger mini to maintain at a slower speed.

  4. walta100 | | #4

    It would be a waste of time trying to talk you out of the idea that 300,000 BTU wood stove and the 24,000 BTU mini split are a good team.

    I wish you luck with your plan.


  5. Jon_R | | #5

    Consider a wood boiler, a water tank and an air-water heat pump. Chiltrix HPs allow you to control compressor output directly (via MODBUS).

    1. Corey_Haddad | | #6

      Jon -
      Good tip on the Chiltrix!
      For the moment I am just going to install a single-head, air sourced mini as temporary solution, so provide some heat to the house when the wood stove isn't burning. Some years from now, after the foundation is fixed, and the house insulated and sealed up, I'll do a quality install on a zoned system. I like the idea of high-temp hydronic system, as it would be silent to operate. Seems the options are currently quite limited in the US, but maybe that will change in the future.

      The water tanks I've seen so far seem pretty small in terms of buffer capacity. Do you know of any systems that can accommodate a large tank as a 'thermal battery'? Capturing heat during the warmest part of the day and releasing later is an appealing concept.

      1. Jon_R | | #7

        You can contruct an open tank (say from concrete) as large as you want. A polyethylene tank might be good up to 140F.

  6. Jon_R | | #9

    Where did you get max efficiency at 25%?

    Review the 3rd graph below, where you see that in one case, efficiency is good from 15% to 100% rated load with the peak around 70% load.

    Given how little time is spent there, even the 0-15% efficiency (cycling operation) is fine.

    1. Corey_Haddad | | #10

      I don't know where I got 25%... but chart #3 is nice. Very encouraging.
      Maybe I just read that lower compressor speeds are more efficient, at least until a certain point.

      Another question for you - 'short cycling' is clearly not desirable. But how do you have any sense of how short is short? If the cycle was comprised of 1 hr intervals for each on and off period, would there be much loss in that scenario?

      1. Jon_R | | #11

        True, many don't realize that oversized equipment doesn't necessarily cause short cycling (eg, variable capacity or thermal mass) and can even improve efficiency and AC latent performance. See below for "achieve CD values of 0.05 or less", which means that with some equipment, even short cycles are pretty much not an efficiency issue (particularily when measured in $).

        Also note that even perfectly sized, variable capacity equipment will short cycle when loads are very low. And is likely to have poor latent performance much of the time.

        Need a generic answer - for best efficiency, design for 10+ min on cycles under most conditions.

  7. creativedestruction | | #12

    When you eventually do the air sealing and insulating that’s "some time off", your peak loads will change significantly. You may then need a different minisplit setup, at which point this calculus exercise will feel a bit wasted.

    Weatherstripping and caulk are cheap. Equivalent COP = 1000. I round up.

    1. Corey_Haddad | | #14

      Yes, the mini-split I just ordered is a small, single-head unit. It's an interim solution.
      Many of the walls are not insulated, and strong winter winds push right through the ship-lap siding enough to close or open interior doors. Before doing the insulation, air gap sealing, and putting in double-pane windows, the foundation has to be redone. So yes, 'some time off' is how it is.

      Eventually, once all that stuff if complete, I would do proper calculations and hope to install a hydronic system that can actually deliver heat to each room, and do so silently.

  8. davidsmartin | | #13

    For many years my Vermont house was off grid and I heated with a wood stove. On cold nights I was too lazy to put more wood in the stove in the middle of the night and when I got up in the morning the house was colder than I would have liked. Several years ago I connected to the grid and installed a mini-split to supplement my wood stove. I just leave the thermostat at 68 degrees. When the wood stove is cranking during the day, the mini-split is barely working. On cold nights, the mini-split might come on at 3AM and keep the house comfortable. Then in the morning when I feed the wood stove the mini-split gradually lowers its output.

    The second use of the mini-split is during the shoulder seasons when a fire in the wood stove would rapidly overheat the house. The third use is to keep the house warm when I go away for several days.

    I don't know if this is the most efficient use of the mini-split, but I don't really care as it keeps the house comfortable. The cost of the electricity is not excessive and I use somewhat less firewood.


  9. Corey_Haddad | | #15

    David - that's great to hear. Those are my three envisioned used cases as well.
    I'm looking forward to a future without 45F winter-morning temperatures in the kitchen.

  10. davidsmartin | | #16

    Corey -- I should add one hint. If you put the mini-split too close to the wood stove and/or too close to the ceiling it will often be confused about how much heat it should be putting out. The only solution for me was to add a remote thermostat some distance from the wood stove. For Fujitsu that was an expensive option but it made a big difference in the ability of the mini-split to maintain the desired temperature and to avoid short cycling.


Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |