Heat pump explained: the aircon circuit run in reverse

Mechanically, no. Nothing inside an aircon manufactures cold. The machine holds a compressor, two coils and a metering device, and its whole job is carrying heat from one side of a wall to the other. Cold is what a room has left once the heat has been taken out of it.

That description doubles as the definition of a heat pump. The name states a mechanism, not a product class, so a cooling-only unit on a Singapore ledge already qualifies. Anything performing heat rejection onto a ledge or into a corridor is pumping heat for a living, whatever the box was labelled.

What a machine sold under the heat pump name adds is a choice of direction. The indoor coil absorbs heat while the unit cools. Send the flow the other way and that same coil releases heat instead, while the outdoor coil starts collecting it from the air. The two coils trade jobs. Neither gets rebuilt to do it.

A sealed system is what makes the trade possible at all. The loop stays closed and the charge stays put, so flow direction is the one variable left to play with. Reversing it leaves the machine the same machine. All that moves is which side of the wall ends up warm.

None of this is visible from the room. A reversible unit cooling a bedroom in Singapore behaves exactly like a cooling-only one, because it is doing exactly the same thing. The difference lives in one valve and in the paperwork, never in the air coming out of the louvres.

One valve does the reversing, and the rest of the circuit has to be built to survive both directions. Daikin's service manual for its heat pump range draws the cooling cycle and the heating cycle with an identical parts list. Only the valve state differs between the two diagrams: energised for one direction, de-energised for the other.

The metering device shows the constraint most plainly. Daikin's manual describes a restrictor orifice that seats itself in cooling, forcing refrigerant through a metered hole in its centre. In heating the same orifice moves back off its seat and lets refrigerant pass unmetered around the outside. One component, two behaviours, settled by which way the flow arrives at it.

Filtration has the same problem to solve. A heat pump parts list names a bi-flow filter drier in place of the single-direction type, because flow reaches it from either end. The part exists because the circuit no longer has a fixed inlet.

The compressor needs its own protection for a related reason. A suction accumulator sits ahead of it on that same parts list. Copeland, which builds compressors for this duty, warns that liquid refrigerant flooding back during defrost or under varying loads can dilute the oil, and treats the accumulator as the guard against it. A cooling-only circuit never produces the flip that creates the risk.

Defrost is why the direction gets flipped back mid-run in a cold climate. Running in heating, the outdoor coil is the one absorbing heat, so it ices over and has to be cleared. Daikin's manual describes defrosting as jointly controlled by a defrost control board and a defrost thermostat. Nothing in Singapore outdoor air ever brings that sequence into play.

Copeland's guidance covers the moment of the flip itself. When the valve changes flow direction abruptly during defrost, suction and discharge pressures move outside the compressor's normal operating envelope. The sound the compressor makes through that transition is described as normal, and the suggested remedy is to stop the compressor while the valve moves. None of that sequence ever runs on a machine built to cool only.

Count what the reversible build carries as a result. A valve, a two-way metering arrangement, a bi-flow drier, an accumulator and a defrost control, all mounted on a circuit whose single requirement is to stay closed. Each addition brings its own joints, and every joint has to hold pressure without help.

No home here needs the heating mode, and the local rules are written on that assumption. NEA's minimum energy performance standards for split-type air-conditioners are keyed to cooling capacity. The thresholds are COP figures taken at full load and at half load, combined into a weighted value that gives the part-load reading the heavier share.

Heating does not appear in that table anywhere. A machine's ability to heat is not measured here, not labelled, and not compared between models. Where a spec sheet quotes a heating capacity, that figure was printed for a market with winters and reached you inside a document written for several countries at once.

So the honest answer to whether a flat should have a heat pump is that the question does not arise. The unit already on the ledge pumps heat for a living. Paying for the reverse direction buys a capability with no season to run in.

Narrowness is the useful finding here, and it deserves saying plainly. On a Singapore spec sheet the word carries close to no buying signal, so it should not be weighed against capacity, siting or tick rating. Treat it as a feature and you end up ranking models on a specification nobody in the household will use.

The reverse direction is not free either, even left unused. It brings the parts listed above and the failure modes attached to them, in exchange for a mode the climate never calls for. That can still be a reasonable position when a range is only built in the reversible form, which happens often enough. It is worth recognising as a trade instead of an upgrade.

The water heater is the one household product where a heat pump becomes a real Singapore purchase decision. NEA extended mandatory energy labelling and minimum performance standards to household water heaters from 1 April 2025. Its regulated list names heat pump water heaters as their own type, scoped to a rated input power of 6 kW or less and a tank holding 500 litres or less.

Efficiency on that label is scored as Uniform Energy Factor, defined under Singapore Standard 646. The minimum permitted standard is a UEF of 0.40. Tick bands climb from there, and the top band of five ticks requires a UEF above 1.00.

That top band is where the mechanism becomes readable on a label. NEA has refused test reports for instantaneous water heaters claiming a UEF above 1.00, on the stated ground that the thermal energy received by the water cannot surpass the input electrical energy. An element that generates heat is capped at 1.00 by that argument.

A machine that moves heat has no such ceiling. Most of what it delivers into the tank was already sitting in the surrounding air, and only the moving of it drew current. The same reasoning that holds a resistive element under 1.00 is what lets a refrigerant circuit pass it, which is why the top band was written at all.

The scale here is ordinary but not trivial. NEA ranks water heaters third among household energy users, behind air-conditioners and refrigerators. They account for around 10 per cent of what a typical household consumes. Hot water is the one place in a Singapore home where the choice between making heat and moving it lands on a monthly bill.

Standing in front of two tanks, the practical read is short. Where one option reaches a tick count the others structurally cannot, that gap is physics surfacing on a label. It records that the machine inside moves heat instead of generating it.

A reversing valve introduces a fault that a cooling-only machine cannot have, and it is the expensive kind to get wrong. Daikin's service manual states the valve can leak internally, with discharge gas passing into the suction side inside the valve body. It then warns that compound gauges give the same symptoms as bad compressor valves or broken scroll flanks.

Copeland reports the same confusion from the compressor's side. Where flow rates and pressure drop across the valve fall too low for it to shift, the guidance says the compressor appears not to be pumping. Two very different parts produce one gauge picture, and one of them is the costliest item on the machine.

So when a quote names compressor replacement on a machine that reverses, ask what ruled the valve out. The answer should be a measurement. Daikin's manual sets the acceptable difference between true suction and the suction line after the valve at no more than 4 degrees, and notes the centre tube should always be the cold one. Somebody either took that reading or did not.

One noise on a reversible machine is normal and still gets misread. Copeland's guidance says the valve solenoid should be wired so the valve does not reverse when the system shuts off. Where it is allowed to reverse, pressures equalise back through the compressor and can turn it slowly until they settle. The guidance is explicit that this does no harm to compressor durability, though it does produce an unexpected sound after switch-off.

Most units in Singapore homes never carry the part at all, which puts a different question first. Before anyone argues about which component failed, establish which build is bolted to the ledge. A cooling-only machine has no reversing valve to blame, and a diagnosis reaching for one has gone wrong earlier than it looks.

Manufacturers make that harder than it needs to be. Daikin's installation manual for one split range covers heat pump and cooling-only models in a single book. It itemises what separates them. A drain plug ships with the reversible build alone, a facility setting switch appears only on the cooling-only board, and the trial-operation steps are written out separately for each.

Shared documentation is what keeps the word arriving in front of owners who have no use for it. Shared model families and one shared circuit finish the job. Reading it correctly means treating the word as a claim about which build you own, then confirming that claim before any part gets priced.