Aircon oversizing: why bigger is not the safer choice

A buyer choosing capacity is protecting against one failure and cannot see the other. Everyone knows what an aircon that will not cool a room looks like. Almost nobody has a name for what happens when a machine carries more output than the room asks for. The risk with a name wins, long before any figure gets compared.

The fear behind it is usually borrowed rather than lived. A buyer has heard about the bedroom that never cooled, or slept in one, and the memory is specific. The machine ran all night, the room stayed warm, the bill arrived anyway. That memory does the sizing. It sets the direction before the room is measured.

The advice pointing the same way deserves a second look. Going up one step is the recommendation with the least downside for whoever gives it. A room that cools too fast produces no callback. A room that stays warm produces one, plus an argument about who pays. A supplier who rounds up is not necessarily acting in bad faith. They are avoiding the only outcome that reliably comes back to them.

The two parties end up managing different risks. The supplier is managing the risk of a complaint. The household is managing hot afternoons for the whole working life of the machine. Those two jobs point at different capacities, and only one of the parties has to live with the answer.

The errors also differ in how long they survive uncorrected. An undersized room gets escalated, because somebody is uncomfortable in a way they can describe to a contractor. Too much capacity produces a household that quietly drops the temperature, buys a fan, and puts a dehumidifier in the corner. Nothing in that sequence points back at the capacity decision, so the purchase never gets revisited.

A machine with more output than the room needs reaches the set temperature early and stops. That is the whole mechanism. Dropping the temperature is fast work. Taking the moisture out is slow work, so a run that ends early has completed the quick half of the job and abandoned the other half. What repeated stopping and starting then does to the machine belongs to the short cycling guide.

Singapore is where that trade lands hardest. The complaint here is rarely that a room is warm. It is that the room is cold and still feels wrong, that bedding never dries out, that the wardrobe has taken on a smell. A bedroom can sit at 24°C and be unpleasant to sleep in, and no thermostat has any way to report that.

The experience arrives in a sequence, and the good part comes first. The first hot afternoon is genuinely impressive. The room drops faster than the old unit ever managed and the purchase feels confirmed. What follows is slower and much harder to attribute. The temperature stops holding steady and starts moving around the setting. The outdoor unit becomes something the household notices, because it keeps stopping and starting instead of settling into a run.

None of that reads as a sizing problem from inside the flat. Damp air reads as a servicing problem. A room that swings reads as a remote or a sensor problem. An outdoor unit that keeps restarting reads as a fault worth quoting for. Each of those gets paid for in turn. None of them reaches the thing producing the pattern.

Three costs move in the same direction, and the buyer is only watching one of them. The equipment costs more, which is expected and usually accepted as the price of the padding. The second and third costs are not expected at all, and together they are the larger number.

Running cost is the first surprise. The claim is that more output means less time running, and less time running means a smaller bill. That breaks down once the room is too small for the output on offer. The machine spends its working life starting rather than running, and starting is the expensive part of a compressor's day. The short cycling guide sets out why.

Comfort is the second surprise, and nobody prices it into the purchase. A household living with damp air does something about it. A dehumidifier goes into the bedroom. The setting gets dropped a degree, then another. A fan runs alongside the unit that was bought to make fans unnecessary. Each of those is a purchase and a monthly cost, made to compensate for capacity that has already been paid for.

Buying bigger also commits the flat to things a quotation never itemises. Output comes with physical bulk, and bulk has to fit into space that is usually spoken for already.

Some rooms genuinely need a large figure, and none of the above is an argument for buying small. Undersizing is a real error with real consequences, and a room that cannot hold its setpoint is a poor place to have saved money. What separates the two is where the large figure came from.

A large figure that came out of a calculation is not oversizing. It is correct sizing for a demanding room. A large figure that came out of an instinct is oversizing, even on the occasions when it lands near the right answer. The same machine can be the correct choice in one flat and the wrong one in the flat above it.

Several conditions lift a genuine requirement well past anything the floor plan suggests. A unit under the roof takes heat down through the ceiling for hours after the sun has gone. Broad west glass with an open outlook delivers its worst dose late, on top of air that is already at its warmest. A living area that opens onto a kitchen, or onto a corridor nobody shuts, is cooling more than the plan shows. What each of those adds is set out in the heat load guide.

Distinguishing the two cases is straightforward if the question is asked plainly. A demanding room can be described. Somebody can point at the glass, the storey, the open kitchen, and explain why the number came back high. Padding cannot be described that way, because there is nothing to point at. It exists to cover the possibility that the description was wrong.

The oversizing decision is rarely made in the abstract. It gets made at the moment a calculated requirement lands between two models that actually exist. Domestic split units climb in fixed jumps, from 9,000 BTU/h to 12,000, then 18,000, then 24,000, and a room's requirement falls wherever it falls. The gap between two rungs is where padding gets added, because rounding upward is the only move anybody suggests out loud.

One rung up is usually defensible. The jumps are coarse, the calculation carries a tolerance of its own, and a requirement sitting just under a model is not meaningfully different from it. Two rungs up is a different decision, and it is the one producing every symptom described above. Establish which one is on the table, since both get described the same way at the counter.

There is a move most buyers are never offered. When a requirement lands a little above a model, the load can often be brought under it instead of the capacity being lifted over it. Film on the west glass takes a real amount off. So does a door that gets shut, a filament fitting swapped for one that runs cool, and a desktop machine moved out of a bedroom. Each of those lowers what the room asks for permanently, and none carries a monthly cost.

Package systems make the choice harder, because it gets made for several rooms at once. A System 4 is quoted as one bundle, so the indoor heads arrive in whatever mix the package offers rather than the mix the flat needs. The smallest bedroom tends to be where that lands badly. It is also the room where damp air gets noticed first, by people trying to sleep in it.

The answer is a calculation, and asking to see one is reasonable. A number on its own is not a calculation. The inputs behind it are window direction, which storey the unit sits on, ceiling height, how the room is occupied, and the equipment running inside it. A quotation naming only a model has skipped the question rather than answered it, and padding is what the skipping gets hidden behind.

Both errors are real, and the target is neither the smaller machine nor the larger one. It is the figure the room produces once somebody works it out properly. A supplier doing that work will say so when the answer comes back high, and will be able to show why. That is a different conversation from one which starts at a step up and reasons backwards.