BTU, tons, watts and the number that connects them
A cooling capacity in BTU per hour says how much heat the unit moves. The watts it draws to do that is capacity divided by the efficiency rating on the label. A 12,000 BTU/hr unit rated 10 draws 1,200 W. The same 12,000 BTU/hr in a modern unit rated 15 draws 800 W for identical cooling. That single division is the whole calculation, and everything after it is bookkeeping about hours and rates.
| Capacity | Also written | Draw at rating 10 | Draw at rating 15 |
|---|---|---|---|
| 5,000 BTU/hr | small window unit | 500 W | 333 W |
| 12,000 BTU/hr | 1 ton | 1,200 W | 800 W |
| 24,000 BTU/hr | 2 tons | 2,400 W | 1,600 W |
| 36,000 BTU/hr | 3 tons | 3,600 W | 2,400 W |
The ratings are not interchangeable. EER and CEER are measured at a single set of test conditions and describe the machine on a hot day. SEER is averaged across a simulated cooling season that includes mild weather, so it is a larger number for the same equipment. Using SEER here produces an average draw across the season, which is the right figure for a monthly total and the wrong one for asking what the unit does during a heat wave.
Hours running is not hours switched on
A fixed-speed compressor is either on at full power or off. When the room reaches the thermostat setting it stops, and the fan often keeps going. A unit that is switched on for twelve hours might have the compressor engaged for six of them, and the six is what this calculator wants. The fan alone typically draws a small fraction of the compressor and is not worth modelling separately unless you are chasing the last dollar.
Variable-speed and inverter equipment behaves differently enough that entering the same hours would be wrong. Rather than cycling on and off, it slows down and holds the room at temperature at a fraction of rated power. Enter the hours it is cooling and pick a duty figure that reflects how hard it works. The practical consequence is the opposite of the old advice: switching an inverter off when you leave a room for an hour and back on when you return can cost more than leaving it alone, because bringing a warmed room back down runs the compressor near full output.
What actually moves the number
Four things, roughly in order. Run hours, which is mostly a function of how much heat gets into the space. The efficiency rating, which is fixed once you own the machine. The rate, which is not yours to set. And the thermostat setting, which changes run hours by something like four to five percent per degree Fahrenheit in ordinary conditions.
Everything else on the usual list of tips works by reducing run hours. Cleaning or replacing a clogged filter restores airflow the compressor was fighting against. Keeping the outdoor coil clear and out of direct sun lets it reject heat with less run time. Closing blinds on the sunny side removes heat before it arrives. A ceiling fan does not cool the room at all, but moving air makes a given temperature feel cooler, which lets the setting go up a couple of degrees, which is where the saving comes from.
Why the answer is not a bill
This page produces the incremental cost of running one machine: kilowatt-hours multiplied by a rate. A bill adds a customer charge that does not move with usage, delivery charges, riders and tax. If your utility charges a higher rate above a monthly threshold, the cooling kilowatt-hours sit on top of everything else in the house and are billed at whatever rate applies up there, which is higher than the average. If you are on a time-of-use plan, an afternoon of cooling in a peak window can cost several times what the same kilowatt-hours cost overnight, and a single average rate cannot show that.
The best defence against all of it is to use your own all-in rate rather than a headline energy rate: take the bill total, divide by the kilowatt-hours on it, and enter that. It folds the fixed charges and taxes into a per-unit figure that is at least honest about what a marginal month costs you.
Sizing is a separate question, and getting it wrong costs more than the rate does
An oversized air conditioner cools the air quickly, satisfies the thermostat and shuts off before it has removed much moisture, leaving a room that is cold and clammy and a compressor that short-cycles. An undersized one runs continuously and never quite gets there. Neither problem shows up in this calculation, because both are about matching capacity to the space rather than about the cost per kilowatt-hour. If the question behind your visit was what size to buy, that is worked out from room area, sun exposure, ceiling height and occupancy on the AC BTU sizing calculator, and it is worth settling before the running cost matters at all.
Questions people ask
How much does it cost to run a window air conditioner for a month?
Take the BTU per hour off the label, divide by the efficiency rating to get watts, multiply by hours per day and days, divide by 1,000 for kilowatt-hours, then multiply by your rate. A 12,000 BTU/hr unit rated 10 draws 1,200 W; at eight hours a day for 30 days that is 288 kWh, which at 16 cents comes to $46.08. Change any of those four inputs and the answer moves proportionally, which is why a single figure quoted for "a window unit" is not usable. The rate alone varies by more than threefold across the country.
Is it cheaper to leave the air conditioner running all day or turn it off?
For a fixed-speed unit, turning it off while nobody is home is cheaper. The house warms up, and a warmer house loses less heat to the outside than a cold one, so total run time falls even counting the longer run needed to cool it back down. For a variable-speed or inverter system the answer often flips, because that equipment is most efficient holding a temperature at low output and least efficient at the full-power pull-down a cold start demands. Setting back a few degrees rather than switching off entirely works well for both.
What is the difference between SEER and EER, and which should I enter?
EER and CEER are efficiency at a single hot test condition; SEER is a weighted average across a simulated cooling season that includes mild days. For the same machine the SEER number is larger. Enter whichever number is on your label and read the result accordingly: a SEER figure gives you a reasonable season-average draw and therefore a reasonable monthly total, while an EER or CEER figure gives you the draw on a hot day and will overstate a month that included mild weather. Neither is wrong, they answer different questions.
Does raising the thermostat by one degree really save a fixed percentage?
It saves run time, and four to five percent per degree Fahrenheit is a reasonable rule of thumb, but it is a rule of thumb rather than a property of your house. The saving comes from the smaller temperature difference between inside and outside, so it is largest when the gap is small and shrinks in extreme heat. It also depends on insulation, sun exposure and humidity. The honest way to check is to run the same setting for two comparable weeks, read the meter at each end, and compare kilowatt-hours per day rather than trusting any percentage.