Why an aircon coil ices up, and what the ice tells you

An aircon works by holding one surface colder than the room and pushing room air across it. That surface is the indoor coil. Heat leaves the air, crosses into the refrigerant inside the tubing, and travels out to the unit on the ledge. A coil running below room temperature is the whole point of the machine.

Water coming off it is a consequence, not a fault. Air here reaches the fins carrying a heavy moisture load, and a cold surface takes part of that load out of it. Droplets gather, slide down the fin faces, and collect in the tray underneath. The drain pipe carries them outside.

All of that depends on the water staying liquid. Draining only works on something that flows. Fin gaps stay open because whatever lands on them keeps moving down and leaves.

Freezing removes that assumption. Take the metal below freezing point and the same moisture still arrives, but it sticks where it lands. Nothing drains, because nothing is liquid any more. A surface that was shedding water begins collecting it instead.

Ice on a coil is unstable in one direction only. Once the first frost forms, everything it causes brings the next layer sooner.

Frost does two jobs at once, and neither helps. It is a poor conductor, so it sits between the passing air and the metal and slows the heat crossing between them. It also fills the channels the air was travelling through, so less air reaches the surface at all.

Both effects starve the refrigerant of the heat it was meant to take on, so it stays colder as it moves through the tubing. Colder metal freezes moisture faster. Heavier frost blocks more of the face. Each turn of that loop is quicker than the one before it.

This is why the failure sounds sudden when a homeowner describes it. A lightly loaded filter or a charge slightly down produces nothing anyone notices for weeks. Then the coil crosses freezing point once, and the runaway finishes the job inside a single long run. The report is of a unit that died overnight. The condition behind it was not overnight at all.

It also explains the false recovery. Switch the unit off, come back later, and it cools properly again, because the frost melted and the fin gaps opened while it sat idle. Nothing was repaired. The loop was reset, and it starts from the same place on the next long run.

Refrigerant needs heat to boil against, and room air is where that heat comes from. Thin out the air and the heat goes with it. What is left inside the tubing runs colder than intended, and part of the coil drops past freezing.

Anything that reduces the stream of warm air belongs on this list. A mesh matted grey with dust. Fin faces packed with the grime that got past the mesh. A blower barrel carrying enough buildup to shift less than it used to. A return grille backed up against a wardrobe. Outlets shut off in a ducted or cassette layout, leaving the remaining ones to carry everything.

Two habits sit here as well, with nothing dirty anywhere in the unit. Running a small room through a cool night leaves very little heat in the air for the coil to work against. Holding the fan at its lowest speed for long stretches arrives at the same place by another route, because fan speed sets how much air crosses the fins.

The order of events is the tell. Vent airflow weakened first, and the cooling complaint followed it. A gas top-up quoted against that sequence has skipped the cheaper half of the diagnosis. Refrigerant added to a coil that is short of air supplies none of the missing heat, so the frost returns on the next long run and the money is gone.

The second route reaches the same frost from the opposite side. Air crosses the coil normally, but there is too little refrigerant in it to carry the heat away, or something upstream is throttling what gets in.

An undercharge does this through pressure. Refrigerant boils at whatever temperature its pressure dictates, and a sealed circuit that has lost part of its charge sits at a lower pressure than it was built for. What remains boils colder. The tubing follows it down, and a stretch of coil goes below freezing while warm air is still arriving in full.

A restriction produces the same effect in one place rather than across the whole coil. A metering device stuck part way shut, or a choked drier, lets too little through into the section behind it. That section runs starved and very cold, and the rest of the coil does almost no work.

Undercharge and restriction both sit on the refrigerant side, and they do not share a repair. A shortage is answered by locating the leak, sealing it, then charging back to the stated weight. A blockage is answered by replacing whatever is throttling flow, and gas poured on top of it deepens the fault instead. Anyone who has narrowed the problem this far can say which of the two they found.

An undercharge is a leak until somebody proves otherwise. A sealed circuit does not use refrigerant up the way an engine uses oil. If the charge is low, it left through somewhere, and a top-up that never names the exit point is selling the same fault back. Ask what was located, not what was added.

Where the frost sits is the cheapest evidence in the whole diagnosis, and it is usually gone before anyone with tools arrives. Photograph it first, then switch off.

An even sheet across the whole face of the coil points at the air side. Refrigerant is being distributed properly along the tubing, and the surface is uniformly short of the warm air it needed. Frost that covers everything evenly is describing what was missing from the room, not from the circuit.

Frost concentrated at one point argues the other way. A patch that begins at a pipe joint, or a band that starts in one spot and fades along the run, means the cold is arriving unevenly. Either flow is being metered at that spot, or there is not enough refrigerant to reach past it.

Frost showing outdoors on the thick insulated line, with the indoor coil face clear, carries its own meaning. Cold has travelled past the coil rather than being absorbed inside it, which happens when the coil is given too little heat or too little flow to boil off what it receives. Read outdoor frost as a reason to open the indoor unit, not as an outdoor fault of its own.

Neither pattern closes the question on its own. Grime lying unevenly on the fins will ice unevenly, and a leak left long enough eventually frosts a wide area too. The pattern decides which test goes first. It does not stand in for the test.

Running a unit with an iced coil puts the compressor at risk, and that is the part of this worth taking seriously. Refrigerant is meant to arrive back at the outdoor unit as gas, having finished boiling inside the coil. A blocked coil does not finish that job. Liquid travels back down the suction line into a pump built for vapour, and what follows is mechanical damage rather than a service item.

Fan mode is the move. Turn cooling off, leave the indoor fan running, and let room air lift the frost off the fins. The unit is not being repaired by this. It is being made readable.

Wait for a state, not a clock. The coil is ready when no frost is left anywhere on the fins, the pipe joint at the side feels dry, and water has stopped arriving in the tray. A half-thawed coil measures like a half-blocked coil, and every reading taken against it points somewhere wrong.

Expect water while it clears. A full sheet of ice releases more at once than the tray underneath can pass along, so put a towel or a basin below the indoor unit before it starts. A wet wall after an icing episode is usually the melt itself rather than a separate drain fault.

What follows the thaw is a real test, and it needs a coil that is working. Vent airflow gets judged against the condition of the filter, the fins and the blower. Refrigerant readings get taken against the operating conditions of that moment. Neither means anything on a coil that is still part frozen, which is the honest reason the thaw comes first.