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

Minisplit power consumption

Trevor_Lambert | Posted in Green Products and Materials on

What would be a typical electrical power consumption of a 9kBTU ductless mini-split? Does 700-800W seem like about right?

Is there a simple way to extrapolate the consumption from the EER or SEER and the BTU rating? I can’t really wrap my head around SEER; it’s not intuitive to me the way COP is. Why don’t they just use COP anyway?

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
    BILL WICHERS | | #1

    BTU/EER = watts, approximately. But use the name plate of the unit you want to buy to be sure of the units power consumption.

    9,000BTU / 15 EER gives approximately 600 watts. But, don’t trust that for anything important like sizing an electrical circuit to supply the unit. EER is measured at only one set of operating conditions (basically a specific indoor and outdoor temperature and humidity), and will be different if the unit is operated under some other set of conditions.

    COP is an instantaneous measure of efficiency relating output energy to input energy and is essentially more general than the EER.

    SEER attempts to specify an EER but using a varied set of operating conditions likely to be encountered over a full season of system use.

    Air conditioners have to “work harder” under certain conditions, basically a high thermal differential. Air conditioners work by moving heat against a thermal gradient, from cold to hot. Since this is against the natural direction the energy wants to flow, we have to put energy in (in this case, electrical energy), to do what we want. The larger the differential, the more work has to be done. It takes more energy to cool a room to 70 degrees on a 100 degree day than it does on an 80 degree day, for example. Also, you will get less effective cooling output if the room is very humid since some of the air conditioner’s efforts are wasted dehumidifying (this is the difference between “sensible” and “latent” heat). EER/SEER try to allow for this, COP does not.


  2. Expert Member
    Dana Dorsett | | #2

    The COP varies quite a bit with both temperature and modulation level- there is no single COP number to assign that would have any relevance. Even the HSPF is a funky sort of number, since it presumed duty cycle with some COP curve test at +47F and +17F. (It's messier than that.)

    There is no such thing as "typical", since the instantaneous COP depends on the capacity and modulation range relative to the outdoor temperature variations.

    The AHRI submittal sheets usually spell out the minimum and maximum power draw, and the COP at the temperature & modulation level tested for at least the+47F and +17F test temperatures. eg:

    If you look at the " Heating at 47°F " data and the "Heating at 17°F" data,

    max power, at +47F the FH09 draws 1470 watts

    max power at +17F it draws 1440 watts

    At the "rated" modulation level of 10,900BTU/hr @ +47F it draws 710 watts.

    At the "rated" modulation of 6700 BTU/hr @ +17F it draws 600 watts.

    If you then look at the "Efficiency" rows it lists the COP under a few conditions, including "COP at 17°F in Maximum Capacity" (which is at a higher modulation level, and lower COP than at the "rated" modulation output at which the HSPF was calculated.)

    The the as-used COP and power level in a given installation will vary from the "rated" COPs b quite a bit. As a general rule at any given temperature the COP will be higher at lower modulated output level, lower at a higher output. At 30F+ the difference in COP between max & min is large, but the difference shrinks as it gets colder outside. The fact that it's substantially more efficient when running at a lower modulation level is why a "set and forget" approach ends up using fewer kwh than deep overnight setbacks:

    At max speed at +47F most mini-splits don't do better than a COP of 2-2.5. The "rated" COP of the FH09 is 4.5 (at about 0.6x it's max capacity at that temp), so it uses about half the kwh per BTU than it would on a recovery ramp running full-out. But the COP at it's minimum output level @ 47F is probably higher than that COP 4.5 number, for an even bigger difference between letting it modulate all night long vs. ramping back up at high speed/low efficiency.

    When it's cycling on/off due to lack of load loses quite a bit of efficiency to compressor spin-up. Even though it's steady state COP efficiency is quite high at minimum output, the spin-up from a cold start is a big bite, which is why you want to size them to where they can modulate over most of the range of load & temperature.

  3. kjmass1 | | #3

    If you are asking how much it costs to run- here is one benchmark.

    Over 4 months this summer in Boston, I averaged .31 kw/hr in cooling costs running 19-30k worth of heads hooked up to non-modulating Fujitsu multi systems. About $1.50/day @ $.20 kw/hr for May-August billing periods with multiple weeks approaching 100 degrees. I imagine a single 9k would be a fraction of that. My worst month was .54 kw/hr cooling.

    These systems are crazy efficient compared to my window ACs which approached 1.5 kw/hr.

  4. Expert Member
    Dana Dorsett | | #4

    The units kw/hr isn't a standard unit used in the industry or anywhere else. Kilowatts are a rate: 1kw= 1000 joules/second. A second is 1/3600 of an hour, so...

    1kw/hr would be 3.6million joules of energy.

    $1.50/day @ $0.20/kwh is 7.5kwh/day or ( divided by 24 h/day=) an average of 0.3125 kw (not kw/hr) which is 312.5 watts of average power consumption. I suspect you really meant kw, not kw/hr.

    1. kjmass1 | | #6

      Thanks that was I typo I will fix. Just trying to put a relative cost in perspective for Trevor, if indeed that is what he is looking for.

  5. Expert Member
    Peter Engle | | #5

    Trevor didn't mention whether he was trying to figure out his "typical" energy use or to size equipment. If it is sizing, you've got to use the nameplate data. The manufacturers look at all operating regimes, and provide information on the minimum required ampacity of the circuit, maximum overcurrent protection, etc. Those are real installation requirements, and they are somewhat independent of how much power the unit draws at any given time, though startup in hot weather is probably its closest operating condition. The specifications also include all of the required safety margins, wire temperatures, and a bunch of other factors that go into it.

    If you're trying to figure out your seasonal energy costs, SEER is an attempt to get there, but as mentioned above, that's just a SWAG based on a standardized set of varying operating conditions. Use it for guidance, but don't bet the farm.

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |