Aircon horsepower: what HP, BTU and kW each tell you

Horsepower on an aircon does not describe the motor inside it. It is a size bracket the trade settled on, and manufacturers place a model into a bracket by convention. A system sold as 1HP holds no one-horsepower motor, and no maker claims it does.

The convention started from something real. One mechanical horsepower is about 746 watts of input power. Drive a compressor at that input, at the efficiency an older fixed-speed system reached, and the unit pulls roughly 9,000 BTU/h of heat out of a room. That pairing stuck. The cooling output kept the horsepower label long after motors stopped lining up with it.

Efficiency has moved on since. The label has not. A modern inverter compressor drawing the same input shifts far more heat than the old ratio assumed, so the bracket now describes output rather than the motor behind it. HP survived because it is short, easy to say over the phone, and every supplier here already speaks it.

That is why the vocabulary matters more in Singapore than the physics does. Marketplace listings, contractor quotes and neighbour recommendations all trade in HP. A homeowner asks for a 1.5HP for the master bedroom and gets understood immediately. The shorthand works well enough to book a site visit. It stops working the moment two quotes need comparing.

The label also travels further than the unit it was attached to. A household replacing an ageing system will often ask for the same HP the old one carried, because that is the only figure anyone remembers about it. The bracket gets copied forward. Whatever reasoning produced it originally, sound or not, comes along unexamined.

Two of these three units convert exactly and one does not. BTU/h and kW both measure heat removed per hour, so they translate cleanly: one kW is about 3,412 BTU/h. HP has no such relationship with either. It maps onto them by market habit, and the map is indicative.

The table below is the version the Singapore market runs on. Treat it as a lookup for what a bracket usually means, not as arithmetic. The figures describe where a bracket sits, and the section after this one covers how far a real model can drift from that centre.

The ratio drifts across the ladder, which is the clearest evidence HP is a convention. One HP maps to about 9,000 BTU/h. Multiply that by 1.5 and the answer is 13,500, yet the market calls 12,000 BTU/h a 1.5HP unit. The step up to 2HP jumps by 6,000 BTU/h, and the step to 2.5HP by another 6,000. The brackets follow what manufacturers chose to build, not a formula.

Smaller and larger brackets exist either side of that range. A 0.75HP head appears on multi-split systems and in small bedrooms, at roughly 7,000 BTU/h. Refrigeration tons turn up on commercial work, where one ton is exactly 12,000 BTU/h. A ton converts cleanly because it was defined as a quantity of heat. HP was never defined that way, which is the whole reason it needs a table.

Listings rarely give all three figures, so one number usually has to be translated. A product page quoting only BTU/h can be divided by 3,412 to reach kW. A quote naming only a bracket gives neither, and the table converts it into the range worth asking about. Whichever figure arrives first, push the conversation toward the two that were measured.

Kilowatts are what the paperwork actually uses. Manufacturer spec sheets, energy labels and the national register all state capacity in kW, because that is the unit the testing standards are written in. BTU/h shows up on product listings and in conversation, largely through American habit. HP is the one figure in the set that appears nowhere in the technical record. It lives entirely in the retail layer above it.

Two quotes both marked 1.5HP can differ by more cooling than a small bedroom draws. The bracket is assigned by the manufacturer. The cooling capacity is measured under test and registered, and those two numbers pull apart more than most buyers expect.

Singapore publishes the evidence, so this is checkable rather than theoretical. NEA registers the models sold here and lists a rated cooling capacity for each one. Thirty-six registered models across sixteen brands sit in the band the market calls 1.5HP. Their rated capacity runs from 3.24 kW at the bottom to 3.73 kW at the top, which is roughly 11,100 to 12,700 BTU/h.

A spread of that size lands on the room rather than on paper. Two households both buy a 1.5HP. One takes a model near the top of the band and one takes a model near the bottom. The second room settles less easily through a hot afternoon, and its unit runs closer to full output to hold the set point. Neither buyer was misled. They compared brackets instead of capacities.

Ask for the model number before weighing two quotes against each other. A supplier who cannot give one is quoting a bracket rather than a machine. A model number resolves to a registered capacity, and that figure reads the same no matter who is selling it. Comparing on model numbers takes one question and removes the guesswork the bracket introduces.

The smallest bracket carries the widest spread, which is where it does the most damage. Registered models in the 1HP band run from 2.30 kW to 2.92 kW. In percentage terms that is a wider gap than any bracket above it, and it sits exactly where HDB bedrooms and study rooms get specified. A room at the bottom of that band is receiving close to a fifth less cooling than a neighbour who bought the same nominal size.

Multi-split systems widen the gap again. A head on a System 3 or System 4 carries its own rated capacity, but the outdoor unit sets a ceiling on what every head can draw at once. Run three rooms together and each one gets less than its individual rating suggests. A 1.5HP head on a shared system and a 1.5HP single split are different propositions even when the label matches.

The rating label carries the number HP stands in for. On a wall unit it usually sits behind the front flap, along the right-hand edge of the casing, or on the underside near the drain outlet. Open the flap the way a filter clean starts. The label is a printed sticker, not a moulded marking on the plastic.

Two kW figures appear on most labels and they are not interchangeable. Cooling capacity is the heat the unit removes from the room. Rated input, sometimes printed as power input, is the electricity it draws to do that. On a 1.5HP wall unit the first sits near 3.5 kW and the second nearer 1 kW. Reading the smaller number as the unit's size makes a healthy machine look far weaker than it is.

The outdoor unit carries a label of its own, normally on the side panel facing away from the wall. On a single split the two labels should agree on system capacity. On a multi-split the outdoor label states what the condenser can deliver in total, which is lower than the heads connected to it add up to. That shortfall is designed in, not a fault.

Model numbers hint at capacity once the pattern is clear. Several brands embed the kW figure multiplied by ten, so a model ending 25, 35 or 50 points at 2.5 kW, 3.5 kW or 5.0 kW. Others embed BTU/h in thousands, so 9, 12 or 18 in the name points at 9,000, 12,000 or 18,000 BTU/h. The scheme differs by brand, which is why the printed label settles the question and the model number only narrows it.

Labels fade, and on older installations they sometimes lose the argument to a paint roller. Where the indoor sticker is unreadable, try the outdoor unit instead. It sits out of reach of decorating and its label is printed on tougher backing. Failing both, the model number stamped into the casing or written on the original installation paperwork resolves against the manufacturer's published figures.

What the label leaves out matters as much as what it states. It gives capacity, input, refrigerant type and charge weight. It says nothing about how the unit is performing today. A system with a fouled coil or a low charge still shows its factory numbers on the sticker. The label sets the benchmark, and a technician's readings say how far the unit has drifted from it.

HP works as shorthand and fails as a comparison. It narrows a conversation fast, and everyone in the market understands it. Once two specific systems are on the table, the useful number is the rated cooling capacity in BTU/h or kW, because that one was measured rather than assigned.

Working out what a room needs is a separate question from what the units mean. Sun exposure, glass area, ceiling height and daily usage all move the answer, and the BTU sizing method covers that ground properly. The BTU calculator turns those inputs into a starting figure. This page settles what the numbers mean, and that one settles which number the room needs.

For a unit already on the wall, the label answers something the HP figure cannot. A room that never settles may hold a correctly sized unit that is underperforming, which is a service question rather than a capacity one. Confirm the installed capacity first. It rules out one of the two explanations before anyone starts quoting a replacement.

Replacement is where a copied bracket does the quiet damage. Matching the new system to the old one's HP feels like the safe move, and it usually goes unchallenged because nobody has a reason to reopen it. If the original unit was specified well, the bracket carries a sound decision forward. If it was never right, the same shortfall gets bought a second time and blamed on the new machine. The label alone cannot tell those two cases apart.

Treat an HP figure on a quote as a starting point that still needs a model number attached to it. A quote naming a bracket and no model has not committed to a capacity yet. Two suppliers can both write 2HP, deliver systems separated by most of a kilowatt of cooling, and both be honest about the label they used.