Why a Bigger Air Conditioner Is Not a Better One

A load calculation is an engineering estimate of how much heat your house gains on a hot day and loses on a cold one. Heat is counted in BTUs, or British thermal units, and the answer comes out as BTUs per hour. The standard method is called Manual J, published by the Air Conditioning Contractors of America. It works room by room and takes in wall and attic insulation levels, the area and type of every window, which direction each wall faces, ceiling height, how leaky the house is, how many people live there, and where the ductwork runs. A ton of cooling is simply 12,000 BTUs per hour, which is why equipment gets described in tons.

The cooling number has two halves, and this is the part most homeowners never hear. Sensible load is heat you feel as temperature. Latent load is the heat carried in water vapor, meaning the moisture the system has to condense out of the air. A proper calculation reports both, because in a Mid-Atlantic summer the latent half is a large share of the total. The calculation is also run against a design temperature, which is roughly the outdoor condition exceeded only a small fraction of hours in a normal year, not the hottest afternoon ever recorded. Sizing to the record day guarantees an oversized machine for the rest of the season.

The old rule of thumb, roughly one ton for every 500 to 600 square feet, treats two houses of the same size as the same house. They almost never are. A 1,600 square foot bungalow from the 1920s with original single pane windows, uninsulated walls and a drafty attic has a very different load than a 1,600 square foot house built in the 1990s with insulated walls and double pane glass. The two can differ by half. That rule of thumb was built around older, leakier housing stock, so applying it to a house that has since been insulated and air sealed produces a system that is badly oversized.

The load also moves over time, which is why a size chosen decades ago is not evidence of anything. Adding insulation in the attic lowers it. Air sealing the top plates and the rim joist lowers it. Replacing windows lowers it, and so does adding shade or an awning on the west facing glass that bakes all afternoon. Things move it the other way too: an addition off the back, a finished basement, a converted garage, a wall of new glass, or taking down the big oak that used to shade the roof. If your house has changed since the last system went in, the correct size has changed with it.

Short cycling is when equipment satisfies the thermostat quickly, shuts off, and restarts a few minutes later, over and over. An oversized system does this by design. If it can deliver an hour of cooling in ten minutes, then ten minutes is all it will run. That sounds efficient and is the opposite. Starting is the hardest thing a compressor does: it draws a large surge of current, oil has to redistribute, and the contactor and start components take a beating every time. A machine that starts six times an hour instead of two ages roughly three times as fast in the parts that matter most.

The rest of the house feels it too. Air takes time to travel through a duct system, so the rooms farthest from the air handler are just beginning to get their share when the blower shuts off. Those rooms stay warm in summer and cool in winter no matter where you set the thermostat. Efficiency suffers as well, because rated performance assumes the equipment is running in a steady state. A coil needs several minutes to get properly cold, and a furnace heat exchanger needs time to come up to temperature. A system that never gets there spends a disproportionate share of its run time in its least efficient minutes, and the electric bill shows it.

An air conditioner dries the air as a side effect of cooling it. Warm indoor air passes over the evaporator coil, which sits well below the dew point of that air, and water condenses on the cold metal the way it beads on a glass of iced tea. That condensate drains away, and the moisture is gone from the house. The important detail is that this takes time. In the first minutes of a cycle the coil is still warming down and barely wet, so almost no water is removed. Real dehumidification happens in the back half of a long run, once the coil is cold and running wet.

An oversized system never gets to the back half. It hits the setpoint in seven minutes and shuts off, and worse, any condensate still clinging to the coil evaporates straight back into the airstream when the blower keeps running afterward. The result is a house at 72 degrees sitting at 60 percent relative humidity, which feels worse than 75 degrees at 48 percent. Clammy air makes people reach for the thermostat, and lowering the setpoint only makes the cycles shorter. In August around here, with outdoor dew points parked in the low 70s, this is the single most common comfort complaint in an otherwise healthy system.

The first reason is time. A room by room load calculation means measuring windows, counting them, checking insulation depth in the attic, looking at the duct system and entering all of it. Reading the model number off the old unit and ordering the same tonnage takes two minutes. The trouble is that the old unit was frequently oversized itself, so the error gets copied forward from one equipment generation to the next. A house can be on its third system and its second set of windows and still be running the size somebody guessed at in 1978.

The second reason is that the risk is lopsided. Nobody ever calls a contractor to complain that the house cooled down too fast. They do call when the system cannot keep up on a 95 degree afternoon, and that call is expensive and unpleasant for everybody. So the cautious move looks like rounding up. Add a safety factor on top of a calculation that already used conservative assumptions and you can land a full ton over. Modern equipment is less forgiving of that habit than old equipment was, because higher efficiency ratings and variable capacity compressors depend on long, steady run times to deliver what they are rated for.

Ask for the load calculation itself, in writing, before you sign anything. A real one runs several pages. It lists the rooms individually, states the indoor and outdoor design temperatures used, shows the insulation values and window areas it assumed, and separates the sensible and latent cooling numbers. Somebody should have walked your house to produce it: up in the attic, down in the basement, measuring window openings, looking at where the ducts run. If a proposal arrives the same afternoon as the first phone call with a tonnage already on it, that number came from somewhere other than your house.

A few things are worth treating as warnings. The size came off the old nameplate. The size came from square footage. The equipment was selected before anyone looked at the attic or the ductwork. The contractor will not share the calculation. Or the calculation exists but a large safety factor was added on top of it, which quietly undoes the whole exercise. It is also fair to ask whether the duct system was evaluated, because correctly sized equipment on a duct system that cannot move the air will still disappoint you.

Do not assume a smaller answer is a mistake. It is common for an honest calculation to come back a half ton below what is currently installed, and that is usually correct rather than risky. It is equally fine for the answer to match what you have, as long as it was reached independently. If you are planning insulation, air sealing or window work in the next couple of years, say so, because that changes the number and it is better to hear about it before the equipment is ordered than after.