Aircon temperature split: how to measure it in your flat

Most published temperature split figures come from North American service manuals, written around design conditions Singapore never sees. Those conditions assume indoor air carrying far less moisture than a flat here carries on an ordinary afternoon. Read your unit against that benchmark and a healthy machine looks like a failing one.

Converting the figure to Celsius does not repair it. The problem is not the unit of measurement, it is the assumed air. The benchmark was set for a room at moderate humidity, and Singapore rooms are not that room.

A cooling coil does two jobs at the same time. It lowers the temperature of the air passing over it, and it condenses moisture out of that same air. Both draw on the same capacity. The more moisture the coil has to strip, the less capacity is left to show as a temperature drop, and a thermometer only sees what is left.

That leaves no single figure worth publishing as a pass mark. Settled wall units in Singapore flats tend to read somewhere around 8°C to 12°C between the air drawn in and the air blown out. That range describes what the reading usually looks like in practice here. It is not a published standard, and no manufacturer specifies it. The same unit sits near the bottom of it on a wet afternoon and near the top late at night, with nothing wrong either time.

Use one thermometer for both readings. A cheap probe can read a degree or two off true, which is enough to swamp the answer. The same instrument used twice carries that error into both numbers, and subtracting cancels it. Two thermometers stack their errors instead.

Fix the conditions before measuring anything. Run the unit in cooling mode. Pick a fan speed and leave it there, and set a temperature well below what the room is reading. Auto fan changes the airflow underneath you, and the split moves with it. Near the set point an inverter unit throttles down and the split collapses, which is correct behaviour rather than a fault.

Let the system settle before recording anything. The reading taken in the first minutes of a run is not the reading the same unit gives once the room has stopped changing. Wait until the air at the outlet stops getting colder, then measure.

Read the incoming air at the intake grille, on the top or front face of the indoor unit, in the stream being drawn in. Read the outgoing air in the middle of the outlet, held in the airstream, clear of the louvre and clear of the coil. Give each reading time to settle before writing it down. Keep the probe out of direct sun and off any metal surface, and keep it outside the casing. Nothing here requires opening the unit.

• •Mode and fan speed, written down rather than remembered
• •The set point, and what the room was reading at the time
• •Whether the unit had just started or had been running a while
• •Time of day, and whether it was raining or sunny outside
• •Which unit, if the flat has more than one

Most of the spread between one reading and the next comes from the conditions, not the machine. The same unit can hand you two very different numbers within the hour, both of them honest.

Dry mode belongs in its own category. It deliberately runs low airflow and cycles the compressor to pull moisture without over-cooling, so the reading it produces has no relationship to a cooling-mode reading. Measure in cooling mode every time.

The practical consequence is that a reading only means something next to another reading taken the same way. Two numbers from different fan speeds, or one from a fresh start and one from a settled room, say nothing when set side by side.

A large split on a room that still will not cool points at airflow rather than capacity. The air at the louvre feels very cold, but there is not much of it. The coil is pulling plenty of heat out of a small volume of air, so the total heat leaving the room stays low.

The usual causes sit in the air path. A loaded filter, a blower wheel furred with grime, or a coil face packed solid all restrict the air and widen the reading. Ice on the coil or the pipework belongs in the same group, because the ice itself blocks the path, and the split widens further as the room gets worse.

If refrigerant is proposed on a wide split, ask what the airflow check found first. Low refrigerant narrows the split. It does not widen it. A wide split with a hot room is an airflow story, and adding gas to it will not change what the room feels like.

The split tells you whether capacity is down. It does not tell you why. A narrow reading is equally consistent with low refrigerant, a fouled coil, an expansion valve that has stuck, or a compressor no longer pumping properly. Separating those needs pressures, temperatures taken at the outdoor unit, and electrical readings under load.

The negative answer is often the more valuable one. If the split is healthy and the room still will not hold temperature, the unit is doing its job. The problem sits in the room instead. Afternoon sun through glass, a door open to an unconditioned corridor, a kitchen in use, or a unit sized for a smaller space all produce a warm room and a working aircon.

The strongest signal available without tools is the unit measured against itself. A reading of 9°C means little on its own. The same unit reading 9°C where it used to read 12°C, at the same fan speed and the same hour of the day, is a genuine change and worth reporting. Take the reading while the unit is still working well, note the conditions beside it, and the number becomes evidence later instead of a guess.